摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, Pages 1389–1394 1389 Co-ordination of heavier alkali metals by polyimido antimony(III) ligands; syntheses and structures of [{Sb2(NC6H11)4}2M4] (M 5 Li or Na), Li[{(Me2N)Sb(Ï-NC6H11)2}2Sb] and M[{(C6H11NH)- Sb(Ï-NC6H11)2}2Sb]?2thf (M 5 K or Rb) Alan Bashall,b Michael A. Beswick,a Christopher N. Harmer,a Alexander D. Hopkins,a Mary McPartlin,b Michael A. Paver,a Paul R. Raithby a and Dominic S. Wright *,†,a a Chemical Laboratory, University of Cambridge, Lensfield Road, Cambridge, UK CB2 1EW b School of Applied Chemistry, University of North London, Holloway Road, London, UK N7 8DB The in situ reactions of MCH2Ph (M = Na, K or Rb) with C6H11NH2 in toluene followed by the addition of the appropriate stoichiometric quantity of Sb(NMe2)3 gave the new heterobimetallic antimony(III)/alkali metal complexes [{Sb2(NC6H11)4}2Na4] and M[{(C6H11NH)Sb(m-NC6H11)2}2Sb]?2thf (M = K or Rb).Comparison of the crystal structures of these species with those of the lithium complexes [{Sb2(NC6H11)4}2Li4] and Li[{(C6H11NH)- Sb(m-NC6H11)2}2Sb] reveals that the geometries of these heterobimetallic cages are fundamentally dictated by the rigidity of the [Sb2(NC6H11)4]22 and [{(C6H11NH)Sb(m-NC6H11)2}2Sb]2 anions.Recently we have shown that the antimony(III) polyimido and phosphinidine anions [{(Me2N)Sb(m-NR)2}2Sb]2 I,1b [Sb2(NR)4]22 II (R = C6H11),1a and [Sb(ER)3]32 III [E = N, R = C6H11,1a But or 2,4-(MeO)2C6H3;1g E = P, R = C6H11 1e] can be prepared by the stepwise metallation reactions of dimethylamidoantimony( III) reagents and primary amido and phosphido lithium complexes.1 The resulting heterometallic SbIII/Li cage compounds are versatile precursors for further cage assembly, the polyimidoantimony(III) anions remaining intact in transmetallation and co-complexation reactions with various main-group and transition-metal sources.2 For example, the transmetallation reaction of [{Sb(NC6H11)3}2Li6] with Pb- (C5H5)2?tmen [tmen = (MeNCH2)2] gives the polyhedral SbIII/ PbII cage [{Sb(NC6H11)3}2Pb3],2a whereas the reaction of [{Sb(NC6H11)3}2Li6] with KOBut results in the co-complex [{Sb(NC6H11)3}2Li6]?3KOBut.2c The accommodation of metal ions of very diVerent sizes by the [Sb(NR)3]32 ligand owes much to the flexibility of the antimony bridgehead, whose N]Sb]N bond angles can adjust in order to satisfy the co-ordination requirements of the metal ion incorporated.2d This situation is in marked contrast to analogous silicon(IV) ligand systems, [RSi(NR)3]32, where the angles at the silicon bridgehead are constrained by the more rigid sp3 hybridisation and by the absence of a lone pair.3 In order to provide a broader assessment of the coordination behaviour of the other antimony(III) polyimide ligands [the monoanion (type I) and dianion (type II)] and to examine the extents to which modification in their Sb]N cores may occur with the varied ionic radii of the co-ordinated metals, we present here a structural study of the co-ordination of the heavier alkali metals with these ligands.The new complexes [{Sb2(NC6H11)4}2Na4] 2, containing the [Sb2(NC6H11)4]22 dianion, and M[{(C6H11NH)Sb(m-NC6H11)2}2Sb]?2thf (M = K 4 or Rb 5), containing the [{(C6H11NH)Sb(m-NC6H11)2}2Sb]2 monoanion, have been structurally characterised. Comparison of these complexes with the lithium analogues [{Sb2(NC6- H11)4}2Li4] 1 1a and Li[{(Me2N)Sb(m-NR2)}2Sb] 3,1b whose structures have been communicated previously, gives new insights into the formation and stability of these species and of † E-Mail: dsw1000@cus.cam.ac.uk the geometric flexibility of their monoanion and dianion ligands.Results and Discussion The preparation of the lithium complex 1 was achieved by the in situ reaction of the antimony(III) dimer [{(Me2N)Sb- (m-NC6H11)}2] [ prepared from the 1: 1 reaction of C6H11NH2 with Sb(NMe2)3] with (LiNHC6H11)n (1 : 1 monomer equivalents, respectively) in toluene (Scheme 1).1a Yields of up to 80% of this species can be obtained in large-scale preparations, thus providing a readily accessible starting material for the investigation of the co-ordination chemistry of the [Sb2- (NC6H11)4]22 anion.The new complex 2 was prepared by the one-pot reaction of a mixture of C6H11NH2 with NaCH2Ph in toluene (2 : 1 equivalents) with Sb(NMe2)3 (1 equivalent) in ca. 50% yield.The reaction is assumed to occur in a similar way to that producing 1, by the initial formation of [{(Me2N)Sb (m-NC6H11)}2] which then reacts with NaNHC6H11. Attempts to prepare 2 by the exchange reaction of 1 with NaOBut proved unsuccessful owing, we assume, to competing incorporation of Sb N Sb N N Sb N NR1R2 NR1R2 C6H11 C6H11 C6H11 C6H11 N Sb N Sb C6H11 C6H11 C6H11N NC6H11 ER Sb RE ER – 2– 3– R1 = R2 = Me R1 = H, R2 = C6H11 R = C6H11, But, 2,4-(MeO)2C6H3; E = N R = C6H11; E = P I II III1390 J.Chem. Soc., Dalton Trans., 1998, Pages 1389–1394 2 Sb(NMe2)3 1 2C6H11NH2 24Me2NH [{(Me2N)Sb(m-NC6H11)}2] 2LiNHC6H11 ��� [{Sb2(NC6H11)4}2Li4] 1 2Me2NH 1 2 Sb(NMe2)3 1 4C6H11NH2 1 2NaCH2Ph 1– 2[{Sb2(NC6H11)4}2Na4] 1 2C6H5Me 1 6Me2NH 2 Li1[Sb(NHC6H11)4]2 1 2 Sb(NMe2)3 Li[{(Me2N)Sb(m-NC6H11)2}2Sb] 1 4Me2NH 3 3 Sb(NMe2)3 1 6C6H11NH2 1 MCH2Ph thf M[{(C6H11NH)Sb(m-C6H11)2}2Sb]?2thf 1 C6H5Me 1 9Me2NH M = K 4, Rb 5 Scheme 1 the metal alkoxide into the cage of 1 2c (only a highly soluble, impure material could be isolated by removal of the solvent in this case).The preparation of the lithium monoanion complex 3 was performed by the reaction of the primary amido antimony(III) intermediate Li1[Sb(NHC6H11)4] [prepared in situ from SbCl3 and LiNHC6H11 (1: 4 equivalents)] with Sb(NMe2)3 (1:2 equivalents, respectively).1b The crystallisation of 3 from toluene is very temperamental and as a result of its high solubility the yield of crystalline material is usually low (up to 32%).However, yields of up to ca. 60% can be obtained by precipitation of the complex with hexane. Unlike the stepwise process used in the preparation of 3, the preparations of the new monoanion complexes 4 and 5 were achieved directly by the simple in situ reactions of Sb(NMe2)3 (3 equivalents) with a mixture of C6H11NH2 (6 equivalents) and MCH2Ph (M = K or Rb) (1 equivalent). This route provides the cleanest and without doubt the best route to the related [{(C6H11NH)Sb(m-NC6H11)2}2Sb]2 monanion. Although it is not clear how the spiro Sb]N framework of the monoanion comes about from this reaction, Norman and co-workers 4 have recently shown that the reaction of SbCl3 with LiNHR (1: 3 equivalents) gives the imidoantimony( III) dimer [{(RNH)Sb(m-NR)}2] (R = 2,6-Me2C6H3).The related complex [{(C6H11NH)Sb(m-NC6H11)}2], generated from the reaction of [{(Me2N)Sb(m-NC6H11)}2] with excess of C6H11H2, is a likely intermediate in the formation of the [{(C6- H11NH)Sb(m-NC6H11)2}2Sb]2, which can be conceived to occur by the equilibration reaction of [{(C6H11NH)Sb(m-NC6H11)}2] with the known dianion [Sb2(NC6H11)4]22 (Scheme 2).However, it should be noted that repeated attempts to prepare various neutral dimers similar to [{(C6H11NH)Sb(m-NC6H11)}2], by the reactions of [{(Me2N)Sb(m-NR)}2] (R = 2-MeOC6H4) with R9NH2, have so far failed, owing to the apparently low reactivity of the terminal Me2N groups with primary amines (only the unchanged species [{(Me2N)Sb(m-NR)}2] being isolated).The initial characterisation of all of the complexes 1–5 was made using a combination of 1H NMR spectroscopy and elemental analyses (C, H and N). The presence of C6H11 groups, whose CH2 protons appear as a set of broad overlapping multiplets in the region d 0.5–2.5, in these complexes makes their definitive characterisation diYcult on the strength of basic techniques. In particular, for the monoanion complexes 4 and 5 only very weak and broad N]H stretching bands are observed in their IR spectra and no N]H proton could be identified unequivocally in their room-temperature 1H NMR spectra, despite the presence of terminal C6H11NH groups (as later revealed by X-ray structural determinations).The structural characterisation of these complexes was therefore of primary [{(C6H11NH)Sb(m-NC6H11)}2] 1 1– 4[{SH11)4}2M4] thf M[{(C6H11NH)Sb(m-NC6H11)2}2Sb]?2thf Scheme 2 importance, especially in respect of the assessment of potential structural variation induced in the cage frameworks by the incorporation of alkali-metal cations of vastly diVerent ionic radii.The structures of complexes 1–5 were determined using lowtemperature X-ray crystallography. Those of 1 and 3 have previously been communicated and will only be discussed here by way of comparison with the new complexes 2, 4 and 5. Table 1 lists key bond lengths and angles for 2.For comparison, data for 4 and 5 are given collectively in Table 2. The comparison of the structures of complexes 1–5 allows the assessment of the way in which the [Sb2(NC6H11)4]22 dianion and [{(RNH)Sb(m-NR)2}2Sb]2 monoanion can adjust to accommodate diVerent alkali-metal cations. The structure of [{Sb2(NC6H11)4}2Li4] 1 (Fig. 1) is that of a molecular cage which can be regarded as being formed by the association of two [{Sb2(NC6H11)4}Li2] cubane halves.1c This formulation is supported by the dissociation of the complex into its cubane fragments in arene solutions and by the observation of discrete cubane units for the related bismuth complex [{Bi2(NBut)4}2- (Li?thf)2], in which solvation of the Li1 cations intercepts the formation of the larger dimeric cage arrangement.1c The overall structure of 1 is similar to that of [AlH(NPri)]8 5 and [AlMe(NMe)]8.6,7 The [Sb2(NC6H11)4]22 dianions co-ordinate the Li1 cations of the core using their m-NC6H11 [2.00(2)– 2.14(2) Å] and pendant NC6H11 groups [1.94(2)–2.03(2) Å], resulting in similar, highly irregular planar geometries for the lithium centres [range N]Li]N 88.0(8)–135.6(1)8, sum of N]Li]N average 353.88].The Li]N bonds throughout the core of 1 are typical of amidolithium complexes.8 In addition, a-C]H interactions occur with adjacent C6H11 groups which (in eVect) reinforce the association of the cubane halves of the core {2.53(2)–2.63(2) Å; cf. 2.60–2.70 Å in [LiN(CH2Ph)2]3 9}.The structure of the sodium complex [{Sb2(NC6H11)4}2Na4] 2 has similar features to those observed in 1. There are two independent chemically identical molecules of 2 in the asymmetric unit which diVer marginally in their bond lengths and angles (one of which is shown in Fig. 2). Despite the similarity with 1 in terms of its composition, the accommodation of the larger alkali-metal cations by the [Sb2(NC6H11)4]22 dianions in 2 has a profound eVect on the geometry of the cage.The most obvious result is the adoption of a planar, rhombic Na4 arrangement at the centre of the cage (with alternating Na ? ? ? Na ? ? ? Na angles of average 93.5 and average 86.48 and with the mean deviation from the Na4 plane of 0.05 Å), rather than the tetrahedral pattern that is present in 1. As a consequence of the greater ionic radius of Na1 and of the correspondingly longer Fig. 1 Structure of [{Sb2(NC6H11)4}2Li4] 1J. Chem. Soc., Dalton Trans., 1998, Pages 1389–1394 1391 Table 1 Selected bond lengths (Å) and angles (8) for complex 2 Molecule 1 Molecule 2 Sb(1)]N(2) Sb(1)]N(7) Sb(2)]N(2) Sb(2)]N(7) Sb(1)]N(5) Sb(2)]N(6) Sb(3)]N(1) Sb(3)]N(8) Sb(4)]N(1) Sb(4)]N(8) Sb(4)]N(4) Sb(8)]N(3) Na(1)]N(5) Na(1)]N(3) Sb](m-N)]Sb (m-N)]Sb](m-N) (m-N)]Sb]N SbNNSb dihedral * Average 2.14(2) 2.12(2) 2.09(2) 2.12(2) 1.95(2) 1.99(2) 2.11(2) 2.17(2) 2.10(2) 2.13(2) 1.97(2) 2.00(2) 2.28(3) 2.25(2) 94.8 * 77.0 * 99.4 * 140.3 * Na(1)]N(1) Na(1)]N(1) Na(2)]N(6) Na(2)]N(2) Na(2)]N(8) Na(2)]N(2) Na(3)]N(4) Na(3)]N(5) Na(3)]N(1) Na(3)]N(7) Na(4)]N(6) Na(4)]N(4) Na(4)]N(7) Na(4)]N(8) Na ? ? ? Na N(5,6)]Na]N(3,4) N(2,7)]Na]N(1,8) Na ? ? ? Na ? ? ? Na 2.95(3) 3.18(3) 2.30(2) 2.31(2) 2.97(3) 2.74(3) 2.26(2) 2.33(3) 2.92(3) 2.89(3) 2.37(2) 2.36(2) 2.90(2) 2.78(3) 3.16 * 154(1)–164(1) 126.0(8)–131.1(7) 85.5(5)–94.2(5) Sb(5)]N(15) Sb(5)]N(16) Sb(6)]N(15) Sb(6)]N(16) Sb(5)]N(9) Sb(6)]N(11) Sb(7)]N(14) Sb(7)]N(10) Sb(8)]N(14) Sb(8)]N(10) Sb(7)]N(13) Sb(8)]N(12) Na(5)]N(9) Na(5)]N(12) Sb](m-N)]Sb (m-N)]Sb](m-N) (m-N)]Sb]N SbNNSb dihedral 2.14(2) 2.15(2) 2.13(2) 2.12(2) 1.99(2) 2.00(2) 2.11(2) 2.12(2) 2.09(2) 2.10(2) 1.96(2) 1.99(2) 2.24(3) 2.30(3) 94.8 * 77.7 * 99.6 * 141.2 * Na(5)]N(14) Na(5)]N(15) Na(6)]N(12) Na(6)]N(11) Na(6)]N(15) Na(6)]N(10) Na(7)]N(11) Na(7)]N(13) Na(7)]N(10) Na(7)]N(16) Na(8)]N(9) Na(8)]N(13) Na(8)]N(14) Na(8)]N(16) Na ? ? ? Na N(11,9)]N] N(12,13) N(15,16)]Na] N(10,14) Na ? ? ? Na ? ? ? Na 3.23(3) 2.90(3) 2.29(2) 2.25(3) 2.96(3) 3.23(3) 2.32(2) 2.36(2) 2.77(3) 2.80(2) 2.37(2) 2.37(2) 2.65(3) 2.98(2) 3.16 * 151(1)–169.3(9) 129.3(7)–135.3(7) 86.7(4)–93.8(4) alkali metal–nitrogen bonds, the [{Sb2(NC6H11)4}Na2] halves of the molecule no longer resemble cubane fragments.The strain induced by the complexation of the larger cations results in greater puckering in the [Sb(m-NC6H11)]2 rings of the [Sb2(NC6H11)4]22 dianions, which are splayed open in order to engage the Na1 cations using their m-NC6H11 and pendant NC6H11 groups.The major advantage of this more open arrangement is that the Na1 cations ultimately obtain a greater co-ordination number than is observed for the Li1 cations of 1, with each being bonded to a m-NC6H11 and pendant NC6H11 group of the [Sb2(NC6H11)4]22 dianions in the molecule. The metal core arrangement and the mode of co-ordination of the metal centres by the m-NC6H11 and pendant NC6H11 groups in 2 are similar to those in the copper(I) complex [{Sb2(NC6H11)4}2- Cu4], where a central square-planar Cu4 core is stabilised by two [Sb2(NC6H11)4]22 dianions.2a,b However, this similarity does not stem from similar ionic sizes [i.e.Cu1 0.91 Å; cf. Na1 1.10 Å and Li1 (four-co-ordinated) 0.73 Å],10 but rather from the preference for approximately linear co-ordination of the copper(I) centres by the pendant NC6H11 groups of the dianion (N]Cu]N average 168.78; cf. N]Na]N average 159.68) and from the formation of weak Cu ? ? ? Cu interactions (average 2.57 Å; Fig. 2 Structure of [{Sb2(NC6H11)4}2Na4] 2 cf. 2.56 Å in copper metal 10). The possibility of Na ? ? ? Na bonding can be discounted in 2.8 The molecular arrangement of complex 2 can be understood in terms of the compromise between the bonding demands of the [Sb2(NC6H11)4]22 dianions and the co-ordination requirements of the Na1 cations. The more rigid bonding demands of the [Sb2(NC6H11)4]22 dianion units clearly dominate this balance, as can be seen from the similarity of the bond lengths and angles observed in the [Sb2(NC6H11)4]22 dianions of 1 and 2 and from the large rearrangement in the imidoalkali metal core geometries.The Na1 cations of 2 have extremely irregular, squared-based pyramidal co-ordination geometries. Although the Na]N bonds made with the pendant NC6H11 groups of the [Sb2(NC6H11)4]22 dianions fall in the expected range [2.24(3)– 2.37(2) Å] found in amidosodium complexes,8b,11 the bonds made with the m-NC6H11 groups are unusually long and highly irregular [2.65(3)–3.19(2) Å].These are best described as weak co-ordinative interactions. The molecular structures of Li[{(Me2N)Sb(m-NC6H11)2}2Sb] 3 1b (Fig. 3), K[{(C6H11NH)Sb(m-NC6H11)2}2Sb]?2thf 4 and Fig. 3 Structure of Li[{(Me2N)Sb(m-NC6H11)2}2Sb] 31392 J. Chem. Soc., Dalton Trans., 1998, Pages 1389–1394 Table 2 Selected bond lengths (Å) and angles (8) for complexes 4 (M = K) and 5 (M = Rb)* Sb(1)]N(1A) Sb(1)]N(1B) Sb(2)]N(1A) Sb(2)]N(1B) N(1A)]Sb(1)]N(1B) N(1A)]Sb(1)]N(1BI) N(1A)]Sb(1)]N(1AI) N(1B)]Sb(1)]N(1BI) N(1A)]Sb(2)]N(1B) N(1A)]Sb(2)]N(1C) N(1B)]Sb(2)]N(1C) Sb(1)]N(1A)]Sb(2) 4 2.219(6) 2.091(5) 2.000(6) 2.047(6) 73.5(2) 88.2(2) 151.5(2) 100.5(3) 79.3(2) 94.1(3) 97.4(3) 99.9(2) 5 2.224(6) 2.086(6) 1.991(6) 2.039(6) 73.0(2) 89.0(2) 151.8(4) 101.0(4) 79.1(2) 94.4(3) 98.0(3) 100.0(2) Sb(2)]N(1C) N(1B)]M N(1C)]M O]M Sb(1)]N(1B)]Sb(2) O]M]O N(1B)]M]N(1BI) N(1C)]M]N(1CI) N(1C)]M]N(1B) N(1C)]M]N(1BI) O(1D)]M]N(1C) O(1D)]M]N(1CI) O(1D)]M]N(1BI) 4 2.092(7) 2.916(6) 2.941(6) 2.788(8) 102.7(2) 82.3(4) 66.9(2) 160.4(3) 64.1(2) 98.8(2) 96.5(2) 98.3(2) 108.6(2) 5 2.088(7) 3.019(7) 3.060(7) 3.023(8) 103.1(3) 86.4(4) 63.5(2) 154.0(3) 61.6(2) 95.3(2) 93.2(2) 105.8(2) 106.9(2) * Symmetry transformation used to generate equivalent atoms: I 2x 1 1, y, 2z 1 ��� .Rb[{(C6H11NH)Sb(m-NC6H11)2}2Sb]?2thf 5 (Fig. 4) all contain similar imidoantimony(III) monoanion ligands which consist of two fused Sb2N2 rings sharing a central four-co-ordinate (10e) square-based pyramidal antimony centre.The alkali-metal cations are co-ordinated in a similar way in all of these species, by the terminal amide ligands of the antimony(III) monoanions and by two of the m-NC6H11 imido groups within the [Sb3N4] cores. The Li1 cation of 3 adopts a highly distorted tetrahedral geometry (N]Li]N range 87.4–143.68), while the additional coordination of the K1 and Rb1 cations by two thf molecules in 4 and 5 (each of which is disordered over two 1 : 1 sites) results in distorted octahedral geometries for these ions [N(1C)]K] N(1CI) 160.4(3), N(1B)]K]N(1BI) 66.9(2), O]K]O 82.3(4)8 in 4; N(1C)]Rb]N(1CI) 154.0(3), N(1B)]Rb]N(1BI) 63.5(2), O]Rb]O 86.4(4)8 in 5].Despite the obvious diVerences in the steric demands of the terminal NMe2 and NHC6H11 substituents present in complexes 3, 4 and 5 and the diVering co-ordination numbers and ionic radii of the alkali-metal cations in these species, the geometries of their imidoantimony(III) anions are extremely similar. The pattern of (short, medium and long) Sb]N bond lengths and N]Sb]N angles within these units largely reflects the electronic and bonding demands of the antimony(III) Fig. 4 Structure of Rb[{(C6H11NH)Sb(m-NC6H11)2}2Sb]?2thf, illustrating the structural pattern found in the isomorphous complexes (M = K 4 or Rb 5) centres. The longest Sb]N bonds occur at the axial positions of the four-co-ordinate (10e) antimony(III) centres [Sb(2)] N(21,31) average 2.23 Å in 3;1b Sb(1)]N(1A) 2.219(6) and 2.224(6) Å in 4 and 5 respectively], with Sb]N bonds of intermediate lengths being found at the equatorial positions [Sb(2)]N(11,41) average 2.12 Å;1b Sb(1)]N(1B) 2.091(5) and 2.086(6) Å in 4 and 5 respectively], and with the shortest Sb]N bonds occurring at the terminal, three-co-ordinate (8e) antimony( III) centres [Sb]m-NC6H11 average 2.02 Å in all the complexes]. The more asymmetrical structure of the [{(Me2N)Sb- (m-NC6H11)2}2Sb]2 anion of 3 presumably results from strain induced by the accommodation of the smaller Li1 cation into the cage arrangement and from the presence of stronger alkali metal–nitrogen bonds which can compete more eVectively for the electron density on the NC6H11 groups.In this connection, the only major diVerence in the geometries of the imidoantimony( III) anions of 3–5 is in the N]Sb]N angle between the equatorial NC6H11 groups of the central four-co-ordinate Sb atom [N(11)]Sb(2)]N(41) 92.5(2)8 in 3; cf.N(1B)]Sb(1)] N(1BI) 100.5(3) and 101.0(4)8 in 4 and 5 respectively]. In K- [{(C6H11NH)Sb(m-NC6H11)2}2Sb]?2toluene,2a in which the K1 cation is only loosely solvated by toluene C]H? ? ?K1 interactions, not only is the [{(C6H11NH)Sb(m-NC6H11)2}2Sb]2 anion almost identical in terms of its bond lengths and angles to that present in 4, but a similar N]Sb]N angle between the equatorial NC6H11 groups of the four-co-ordinate antimony centre [100.8(2)8] is observed.The expansion of this angle is directly related to the increased size of the co-ordinated alkalimetal cations (Li1 0.73, K1 1.33, Rb1 1.48 Å 10) which are chelated by the equatorial NC6H11 groups, and presumably this results in a reduction in strain within the more symmetrical antimony(III) anion arrangements found in the heavier alkalimetal complexes.Conclusion The structural investigation presented here provides a more detailed understanding of the co-ordination behaviour, flexibility and bonding demands of imidoantimony(III) monoanions, of the type [{(R1R2N)Sb(m-NC6H11)2}2Sb]2 (R1, R2 = Me; R1 = H, R2 = C6H11), and of the imidoantimony(III) dianion ligand, [Sb2(NC6H11)4]22. Comparison of the Li1 complex [{Sb2- (NC6H11)4}2Li4] 1 with the Na1 analogue [{Sb2(NC6H11)4}2Na4] 2 and Li[{(Me2N)Sb(m-NR)2}2Sb] 3 with M[{(C6H11NH)Sb- (m-NR)2}2Sb]?2thf (M = K 4 or Rb 5) illustrates that the structures of these heterobimetallic antimony(III)/alkali metal cages depend on a subtle balance between the bonding demands of the antimony and alkali-metal centres. The greatest deformations in the antimony anion geometries occur in the Li1 complexes, where the alkali metal–nitrogen bonding is strongest.However, overall the more rigid requirements of SbIII dominateJ. Chem. Soc., Dalton Trans., 1998, Pages 1389–1394 1393 Table 3 Crystal data for complexes 2, 4 and 5 Chemical formula M Crystal size/mm T/K Crystal system Space group a/Å b/Å c/Å b/8 U/Å3 Z Dc/g cm23 m/mm21 q Range/8 F(000) Reflections collected Independent reflections (Rint) R1, wR2 [F > s(F)]* (all data) Peak and hole/e Å23 2 C48H88N8Na4Sb4 1356.22 0.30 × 0.10 × 0.08 150(2) Monoclinic P21/a 24.560(4) 18.46(1) 25.187(5) 91.28(2) 11 413(7) 8 1.579 1.942 2.54–20.00 5440 11 085 10 631 (0.021) 0.076, 0.173 0.254, 0.252 1.129, 21.211 4 C44H82KN6O2Sb3 1131.51 0.16 × 0.20 × 0.40 223(2) Monoclinic C2/c 23.616(3) 11.067(1) 20.777(2) 104.233(7) 5263(1) 4 1.428 1.645 1.78–25.00 2304 9491 4637 (0.040) 0.048, 0.115 0.085, 0.139 0.888, 20.491 5 C44H82N6O2RbSb3 1177.88 0.20 × 0.32 × 0.40 223(2) Monoclinic C2/c 22.733(4) 11.298(1) 21.075(4) 103.52(2) 5263(2) 4 1.487 2.485 1.84–25.00 2376 5570 4630 (0.067) 0.051, 0.097 0.111, 0.121 0.889, 20.717 * R1 = S||Fo| 2 |Fc||/S|Fo|, wR2[Sw(Fo 2 2 Fc 2)2/SwFo 4]� �� , w = 1/[s2(Fo 2) 1 (xP)2 1 yP], P = (Fo 2 1 2Fc 2)/3.the essentially ionic alkali metal–nitrogen frameworks in these species, as is illustrated by the predominant rigidity of the Sb]N cores. Larger metal ions can be accommodated by the [Sb2(NC6H11)4]22 dianion by maximising the M]N bonding with the m-N and terminal N groups and by deformation about the N ? ? ? N axis of the Sb2N2 ring. In the monoanions the complexation of ions with greater ionic radii is achieved almost entirely by expanding the chelating N]Sb]N angle at the central Sb.Experimental General preparative techniques All the reactions were undertaken under dry, O2-free argon using a vacuum line and standard inert-atmosphere techniques. 12 Tetahydrofuran, Et2O and toluene were dried by distillation over sodium–benzophenone and hexane was distilled over Na. Cyclohexylamine was dried using molecular sieves (13X). Complexes 1–5 were isolated and characterised with the aid of a N2-filled glove-box (Miller-Howe, fitted with a Belle internal circulation system).Melting points were determined using a conventional apparatus and sealing samples in capillaries under N2. Elemental analyses (C, H and N) were performed by first sealing samples in air-tight aluminium boats (1–2 mg) prior to analysis using a Perkin-Elmer 240 Elemental Analyser. Proton NMR spectra were recorded on a Bruker WH 250 MHz spectrometer, using the NMR solvents as internal reference standards.The syntheses of [{Sb2(NC6H11)4}2Li4] 1 and Li[{(Me2N)Sb(m-NC6H11)2}2Sb] 3 have been communicated previously [see refs. 1(a) and 1(b)]. The compounds MCH2Ph (M = Na, K, or Rb) were prepared by the reactions of MOBut with LiBun in toluene (by deprotonation of C6H5Me with the MBun initially formed), the reagents being isolated as orangered amorphous materials in quantitative yields. Syntheses Complex 2. A solution of C6H11NH2 (1.72 cm3, 15 mmol) in toluene (10 cm3) was added at 278 8C to a suspension of NaCH2Ph (0.98 g, 7.5 mmol) in toluene (10 cm3).The mixture was warmed to room temperature and stirred (5 min) to give a slightly cloudy, brown solution. A standardised solution of Sb(NMe2)3 (3.75 cm3, 7.5 mmol, 2.0 mol dm23 in thf) was added to the cooled solution at 278 8C. The resulting solution was stirred (20 min) and allowed to warm to room temperature. Filtration (porosity 3, Celite) gave a brown solution. Reduction of the filtrate under vacuum to ca. 10 cm3 gave a pale green precipitate which was heateion. Storage at room temperature (24 h) gave light yellow crystalline rods of 2; 1.2 g (50%); decomp. 130 8C; 1H NMR (125 8C, 250 MHz, C6D6) d 0.8–2.5 (overlapping multiplet, C6H11 groups) (Found: C, 42.4; H, 6.6; N, 7.9. Calc. for [{[Sb2(NC6H11)4]Na2}n]: C, 43.5; H, 6.5; N, 8.2%). Complex 4. A solution of C6H11NH2 (1.72 cm3, 15 mmol) in toluene (10 cm3) was added at 278 8C to a suspension of KCH2Ph (0.33 g, 2.5 mmol) in toluene (10 cm3).The mixture was warmed to room temperature and stirred (5 min) to give a red-brown solution. A standardised solution of Sb(NMe2)3 (3.75 cm3, 7.5 mmol, 2.0 mol dm23 in thf) was added to the cooled solution at 278 8C. The resulting solution was stirred (20 min) and allowed to warm to room temperature. Filtration (porosity 3, Celite) gave a red solution. The toluene was removed under vacuum and Et2O (10 cm3) added. Addition of thf gave initial precipitation of a colourless solid which redissolved upon addition of further thf (ca. 2 cm3). Crystals of 4 were grown by storage (12 h) of this solution at 215 8C; 1.55 g (54%); decomp. 130 8C; IR (Nujol) >3000vw (br) cm21 (N]H str.); 1H NMR (125 8C, 250 MHz, C6D6), d 0.8–2.5 (overlapping multiplet, C6H11 groups) (Found: C, 46.2; H, 7.1; N, 7.4. Calc. for {K[Sb3(NC6H11)4(NHC6H11)2]?2thf}n: C, 44.3; H, 7.7; N, 7.7%). Complex 5. A solution of C6H11NH2 (1.72 cm3, 15 mmol) in toluene (10 cm3) was added at 278 8C to a suspension of KCH2Ph (0.42 g, 2.5 mmol) in toluene (10 cm3).The mixture was warmed to room temperature and stirred (5 min) to give a deep red solution. A standardised solution of Sb(NMe2)3 (3.75 cm3, 7.5 mmol, 2.0 mol dm23 in thf) was added to the cooled solution at 278 8C. The resulting solution was stirred (20 min) and allowed to warm to room temperature. Filtration (porosity 3, Celite) gave a dark brown-red solution. The toluene was removed under vacuum and Et2O (10 cm3) added.Addition of thf gave initial precipitation of a colourless solid which redissolved upon addition of further thf (ca. 2 cm3). Crystals of complex 5 were grown at room temperature (12 h); 1.73 g (59%); decomp. 130 8C; IR (Nujol) >3000vw (br) cm21 (N]H str.); 1H NMR (125 8C, 250 MHz, C6D6) d 0.8–2.5 (overlapping1394 J. Chem. Soc., Dalton Trans., 1998, Pages 1389–1394 multiplet, C6H11 groups) (Found: C, 44.4; H, 7.1; N, 7.4.Calc. for {Rb[Sb3(NC6H11)4(NHC6H11)2]?2thf}n: C, 44.8; H, 7.1; N, 7.1%). X-Ray crystallography Crystals were mounted directly from solution under argon using a perfluorocarbon oil which protects them from atmospheric O2 and moisture.13 The oil freezes at reduced temperatures and holds the crystal static in the X-ray beam. Data for complexes 1, 2 and 3 were collected on a Stoe-Siemens AED four-circle diVractometer and for 4 and 5 on a Siemens P4 diffractometer. The structures of all the complexes were solved by direct methods and refined by full-matrix least squares on F2 (SHELXTL14).In the isomorphous crystals of 4 and 5 the alkali metal-co-ordinated thf ligands and one cyclohexyl ring are all disordered over two sites of approximately 0.5 occupancy. Details of the structure refinements for 2, 4 and 5 are shown in Table 3. CCDC reference number 186/902. See http://www.rsc.org/suppdata/dt/1998/1389/ for crystallographic files in .cif format. Acknowledgements We gratefully acknowledge the EPSRC (M.A. B., C. N. H., M. McP., P. R. R., D. S. W.), the Leverhulme Trust (M. A. B.), Electron Tubes, Ruislip (UK) (A. D. H.), The Royal Society (P. R. R., D. S. W.) and Jesus College, Cambridge (fellowship for M. A. P.) for financial support. References 1 (a) R. A. Alton, D. Barr, A. J. Edwards, M. A. Paver, M.-A. Rennie, C. A. Russell, P. R. Raithby and D. S. Wright, J. Chem. Soc., Chem. Commun., 1994, 1481; (b) A. J. Edwards, M.A. Paver, M.-A. Rennie, C. A. Russell, P. R. Raithby and D. S. Wright, Angew. Chem., 1994, 106, 1334; Angew. Chem., Int. Ed. Engl., 1994, 33, 1277; (c) D. Barr, M. A. Beswick, A. J. Edwards, J. R. Galsworthy, M. A. Paver, M.-A. Rennie, C. A. Russell, P. R. Raithby, K. L. Verhorevoort and D. S. Wright, Inorg. Chim. Acta, 1996, 248, 9; (d ) M. A. Paver, C. A. Russell and D. S. Wright, Angew. Chem., 1995, 107, 1077; Angew. Chem., Int. Ed. Engl., 1995, 34, 1545; (e) M. A. Beswick, C.N. Harmer, A. D. Hopkins, M. A. Paver, P. R. Raithby, A. E. H. Wheatley and D. S. Wright, Chem. Commun., 1997, 1897; ( f ) M. A. Beswick, N. Choi, A. D. Hopkins, M. McPartlin, M. A. Paver and D. S. Wright, Chem. Commun., 1998, 261; ( g) M. A. Beswick, N. Choi, C. N. Harmer, A. D. Hopkins, M. McPartlin, M. A. Paver, P. R. Raithby and D. S. Wright, Inorg. Chem., in the press. 2 (a) M. A. Beswick, C. N. Harmer, M. A. Paver, P. R. Raithby, A. Steiner and D. S. Wright, Inorg.Chem., 1997, 36, 1740; (b) D. Barr, A. J. Edwards, S. Pullen, M. A. Paver, P. R. Raithby, M.-A. Rennie, C. A. Russell and D. S. Wright, Angew. Chem., 1994, 106, 1960; Angew. Chem., Int. Ed. Engl., 1994, 33, 1875; (c) A. J. Edwards, M. A. Paver, M.-A. Rennie, C. A. Russell, P. R. Raithby and D. S. Wright, Angew. Chem., 1995, 107, 1088; Angew. Chem., Int. Ed. Engl., 1995, 34, 1012; (d ) A. Bashall, M. A. Beswick, C. N. Harmer, M. McPartlin, M. A. Paver and D. S. Wright, J. Chem.Soc., Dalton Trans., in the press. 3 D. J. Brauer, H. Bürger, G. L. Liewald and J. Wilke, J. Organomet. Chem., 1985, 287, 305; D. J. Brauer, H. Bürger and G. L. Liewald, J. Organomet. Chem., 1986, 308, 119; M. Veith, A. Spaniol, J. Pöhlmann, F. Gross and V. Huch, Chem. Ber., 1993, 126, 2625. 4 S. C. James, N. C. Norman, A. G. Orpen and M. J. Quayle, J. Chem. Soc., Dalton Trans., 1996, 1455; see also, M. Noltemeyer, H. W. Roesky, H. Schmidt and U. Wirringa, Inorg. Chem., 1994, 33, 4607. 5 G. Del Piero, M. Cesari, G. Perego, S. Cucinella and E. Cernia, J. Organomet. Chem., 1977, 129, 289. 6 S. Amirkhalili, P. B. Hitchcock and J. D. Smith, J. Chem. Soc., Dalton Trans., 1979, 1206. 7 See also, M. Veith, Chem. Rev., 1990, 90, 3. 8 (a) W. N. Setzer and P. v. R. Schleyer, Adv. Organomet. Chem., 1985, 24, 353; (b) R. E. Mulvey, Chem. Soc. Rev., 1991, 20, 167; (c) K. Gregory, P. v. R. Schleyer and R. Snaith, Adv. Inorg. Chem., 1991, 37, 47 and refs. therein. 9 D. R. Armstrong, R. E. Mulvey, G. T. Walker, D. Barr and R. Snaith, J. Chem. Soc., Dalton Trans., 1988, 617; D. Barr, W. Clegg, R. E. Mulvey and R. Snaith, J. Chem. Soc., Chem. Commun., 1984, 285. 10 F. A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry, 5th edn., Wiley, New York, 1988; J. E. Huheey, Inorganic Chemistry, 3rd edn., Harper International, London, 1983. 11 For example, see P. C. Andrews, W. Clegg and R. E. Mulvey, Angew. Chem., 1990, 102, 1480; Angew. Chem., Int. Ed. Engl., 1990, 29, 1440; W. Clegg, M. MacGregor, R. E. Mulvey and P. A. O’Neil, Angew. Chem., 1992, 104, 74; Angew. Chem., Int. Ed. Engl., 1992, 31, 93; N. P. Lorenzen, J. Kopf, F. Olrich, U. Schümann and E. Weiss, Angew. Chem., 1990, 102, 1481; Angew. Chem., Int. Ed. Engl., 1990, 29, 1441; K. Gregory, M. Bremmer, P. v. R. Schleyer, N. P. Lorenzen, J. Kopf and E. Weiss, Organometallics, 1990, 9, 1485. 12 D. F. Schriver and M. A. Drezdon, The Manipulation of Air- Sensitive Compounds, 2nd edn., Wiley, New York, 1986. 13 T. Kottke and D. Stalke, J. Appl. Crystallogr., 1993, 26, 615. 14 SHELXTL PC, version 5.03, Siemens Analytical Instruments, Madison, WI, 1994. Received 25th November 1997; Paper 7/08496J

ISSN:1477-9226    

DOI:10.1039/a708496j

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, Pages 1395–1396 1395 Solvated electrons: electron paramagnetic studies of solutions of lithium in ethylamine* Martyn C. R. Symons a and Fatai A. Taiwo b a Department of Chemistry and Physics, Faculty of Applied Sciences, De Montfort University, The Gateway, Leicester, UK LE1 9BH b Department of Pharmacy, Faculty of Applied Sciences, De Montfort University, The Gateway, Leicester, UK LE1 9BH Lithium metal dissolves in ethylamine to give stable blue solutions the EPR spectra of which showed the presence of a singlet assigned to solvated electrons, together with a nine-line spectrum assigned to Li1 solv–e2 solv ion pairs.The hyperfine splitting of ca. 2.5 G is assigned to four equivalent 14N nuclei. There is no detectable coupling to lithium nuclei. In the region of 50% diglyme there is extensive displacement of two of the four amine ligands. Thus, in addition to the solvated electron centres, two ion pairs were detected, one showing hyperfine coupling to four 14N nuclei and one to two 14N nuclei.The hyperfine splitting remained constant at ca. 2.5 G. This result strongly supports a model in which it is the lithium ligands that give rise to coupling in the ion pairs. The diglyme displaces two amines at a time. There was no increase in the singlet, so the displacement of the second pair of amine ligands is not favourable. One of the most remarkable chemical entities is the solvated electron.This species, in various forms, is unique in that, at least for most models, there is no ‘central’ set of nuclei around which the electrons occupy specific orbitals.1 [The major exception is the model that places the unpaired electrons into orbitals centred on one (or more) solvent molecules, H2O~2 being a key example.2,3] The earliest known system is that of dilute alkali metals in liquid ammonia.4,5 The resulting blue solutions are remarkably stable, but unfortunately the EPR spectra comprise single narrow lines which provide no structural information. However, solutions in amines, ethers and other aprotic solvents, though far less stable, have EPR spectra with some very informative hyperfine features, as summarised in Table 1.For the amines, hyperfine coupling to 23Na1 and other metal nuclei is large, and increases markedly on heating.6 Only lithium fails to give any metal splitting, but, in contrast, the lithium solutions uniquely give hyperfine features from solvent 14N nuclei.7,8 This is a unique result, and it was sometime before a model that could explain it was constructed.This model is essentially an ion pair (solvent shared ion pair 9), in which one or two solvent molecules are shared. This is just one of many proposals for the structure of ion pairs involving solvated electrons. An alternative model comprises solvent-shared and more distant ion pairs,10 and expanded metal units.7,11,12 The former comprise fully solvated cations adjacent to fully solvated electrons. These are in contact for ‘solvent shared’ units, but more distant units are also considered.The latter comprise normal solvated cations, with the electron moving in an expanded orbital centred on the cation. Our results rule out the former since these units are ‘encounter’ ion pairs with very low lifetimes. They also rule out the latter, since no hyperfine coupling to the lithium cations was obtained.However, the solvent-shared model is strongly supported by the present results. In this novel entity, shown in Fig. 1, the Li1 ions are tetrahedrally solvated by four amine molecules (this is the normal mode of solvation for Li1 ions) and the adjacent electron is solvated by about six solvent molecules, one or two of which are bridged between Li1 and the electron. (In the following discussion we invoke two bridging solvent molecules, but there may only be one.) * Non-SI unit employed: G = 1024 T.Results and Discussion EPR Spectral analysis In the present extension we have analysed the EPR spectra for solutions of lithium in ethylamine–diglyme (2,5,8-trioxanonane) mixtures in the hope of detecting a change in the nitrogen-14 splitting from four equivalent nitrogen nuclei (for the experimental methods used see ref. 8). The solutions became progressively unstable as the concentration of diglyme increased and ca. 56% diglyme was the upper limit for EPR studies.The absence of metal hyperfine coupling in marked contrast with the results for the other alkali-metal cations shows Fig. 1 Representations of the lithium cation–electron ion-pair species solv solv solv solv solv Li+ solv solv solv e– solv solv solv Li+ solv solv e– solv solv (a) (b) Table 1 Hyperfine coupling to alkali-metal cations in amine solvents together with estimated spin densities Nucleus 7Li 23Na 39K 85Rb 133Cs Hyperfine coupling (MHz) 0 8.86 18.6 29.2 26.6 28.2 36.2 100.2 134.6 212.5 317.2 T/8C 230 110 129 230 110 150 230 110 150 110 % Atomic character 0 1.0 2.1 3.3 11.5 12.2 15.7 9.9 13.3 21.0 13.81396 J.Chem. Soc., Dalton Trans., 1998, Pages 1395–1396 that these units must be separated ion pairs and not centrosymmetric units. Since the solvated electron centres for non-ion-pair units give a narrow singlet EPR spectrum, the interacting solvent molecules must be the shared molecules not the remainder, which will still exchange rapidly with bulk solvent.Since four equivalent 14N nuclei are detected, these must be the primary solvent shell for Li1 cations, which, because of the strong bonding to lithium, have long lifetimes on the EPR time-scale. This concept is fully supported by the results for solutions containing diglyme. These molecules solvate the Li1 cations strongly, in a chelated unit. Hence only two ethylamine molecules remain, and only two bridging molecules with strong interaction with the unpaired electron are detected.Hence we suggest that for the pure ethylamine solution the Li(EtNH2)4 1 units rotate relative to the electron, so that all four become equivalent. When the chelating diglyme molecules add this still occurs, but now there are only two 14N nuclei with hyperfine coupling. A typical EPR spectrum for a 56% diglyme solution is shown in Fig. 2. This comprises a singlet from e2 solv, a nine-line set Fig. 2 (a) First-derivative EPR spectrum for the lithium–electron ion pair in a 56% diglyme–44% ethylamine solution, showing 14N hyperfine features for the (EtNH2)4 and (EtNH2)2 complexes.(b) A computer simulation using a ratio of 4 : 3 for the (EtNH2)4 and (EtNH2)2 complexes; A(14N) is invariant at 2.5 G. (c) A computer simulation of the narrow singlet for solvated electrons not in ion-pair species. (d) A computer simulation of the five-line spectrum for (EtNH2)2 complexes. (e) A computer simulation of the nine-line spectrum for (EtNH2)4 complexes.The best fit was obtained with a linewidth of 0.5 G from the Li(EtNH2)4 1 solvates, but also a quintet of lines assigned to the unit (diglyme)Li(EtNH2)4 1. The best simulation requires a ratio of 4 : 3 for the Li(EtNH2)4 1 solvates to the (diglyme)Li(EtNH2)2 1. The contribution from the narrow singlet is quite noticeable. In the simulation for Fig. 2 we include the narrow singlet that is obtained for dilute solutions of lithium, together with the nine-line spectrum obtained for concentrated solutions in pure ethylamine, and the five-line spectrum that grows in with added diglyme.In arriving at this fit we used a range of widths, greater and smaller than those finally used which gave the best fit of all. These results strongly support the model proposed, con- firming that the amine and the diglyme are co-ordinated to the lithium ions, but that only two amine molecules bridge to the electrons. The contrast with the results for the other alkali metals is remarkable, with a complete switch from 14N hyperfine splitting to hyperfine coupling to the metal nuclei.Thus it seems probable that for the latter species the electrons move extensively into the cation–solvent system, partially occupying the outer s orbitals of the cations. The spin densities on these cations increase markedly on heating, and on increasing the bulk of the alkyl groups. However we were unable to detect a clear change with temperature for the lithium solution.These lithium solutions represent the limit in which this movement of the electrons to the cations is almost totally blocked by the strengths of the lithium–solvent bonds. However it is this strength that makes solvent exchange relatively slow, so that well resolved spectra result. Unfortunately, our results do not distinguish between one or two bridging solvent molecules. However, in view of the long lifetime required for the ion-pair units, two seems more probable.This is supported by comparison with our pyrrolidine results. Comparison with irradiated pyrrolidine When pyrrolidine glasses are exposed to ionisation radiation at 77 K they become deep blue, and display a strong EPR spectrum containing nine hyperfine features from four equivalent 14N nuclei.13 This centre was identified as an electron trapped at a solvent cavity comprising four suitably located solvent molecules. It is noteworthy that the 14N hyperfine splitting was 5 G, which is exactly twice the value found for the lithium solutions discussed above.This again nicely fixes the number of shared ethylamine molecules as two. Thus, since all four of the liganded amines interact with the electron, ‘rotation’ must be rapid. Hence the hyperfine coupling should be 50% of the true coupling. This gives 5 G for the true value, exactly equal to the pyrrolidine value.13 References 1 M. C. R. Symons, Chem. Soc. Rev., 1976, 5, 337; F. S. Dainton, Chem. Soc. Rev., 1975, 4, 323. 2 H. F. Hameka, G. W. Robinson and C. J. Maesden, J. Phys. Chem., 1987, 91, 3150. 3 T. R. Tuttle and P. GraceVa, J. Phys. Chem., 1971, 75, 843. 4 W. Weyl, Pogg. Ann., 1864, 121, 601; 123, 350. 5 C. A. Krauss, J. Am. Chem. Soc., 1908, 30, 1323. 6 R. Catterall, M. C. R. Symons and J. W. Tipping, J. Chem. Soc. A, 1967, 1234. 7 K. Bar-Eli and T. R. Tuttle, J. Chem. Phys., 1964, 40, 2508. 8 R. Catterall, M. C. R. Symons and J. W. Tipping, Pure Appl. Chem., 1970, 317. 9 T. GriYths and M. C. R. Symons, Mol. Phys., 1960, 3, 90. 10 R. Catterall, M. C. R. Symons and J. W. Tipping, J. Chem. Soc. A, 1966, 1529. 11 R. Catterall and P. P. Edwards, J. Phys. Chem., 1975, 79, 3010. 12 R. Catterall and M. C. R. Symons, J. Chem. Soc., Dalton Trans., 1972, 139. 13 W. T. Cronenwett and M. C. R. Symons, J. Chem. Soc. A, 1968, 2991. Received 16th January 1998; Paper 8/00471D

ISSN:1477-9226    

DOI:10.1039/a800471d

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, Pages 1397–1402 1397 Dynamic processes in platinum(II)–adenosine complexes. Preparation, NMR spectroscopic characterisation and crystal structure of isomeric PtII(dien)–adenosine complexes Jorma Arpalahti,*,†,a Karel D. Klika,a Reijo Sillanpää a and Raikko Kivekäs b a Department of Chemistry, University of Turku, FIN-20014, Turku, Finland b Department of Chemistry, University of Helsinki, FIN-00014, Helsinki, Finland The distribution of aquated PtII(dien) between the adenosine N1 and N7 sites has been found to be aVected in aqueous solution not only by lowering the pH, but also by equilibration of the mixture of PtII and nucleoside at high temperature.In the former case PtII predominantly binds to the N7 site (1), whereas the latter procedure significantly increases the amount of the N1-bound complex (2) at the expense of the N7 isomer. Both complexes exhibit two 195Pt resonances at ambient temperature, but 1H and 13C NMR spectra recorded in D2O revealed only signals with expected multiplicities for the adenine moiety.However, in [2H7]dmf the aromatic protons and the dien NH proton resonances were split into two components at 250 8C and the C6]NH2 protons were no longer equivalent in both complexes. This suggests restricted rotation about the platinum–nucleoside bonds and/or conformational changes of the dien ligand. Also rotation of the exocyclic amino group about the C6]N6 bond appears to be restricted at low temperatures.Both complexes were crystallised as perchlorate salts and their crystal structures determined. Based on X-ray data, further disturbance in the platinum(II) co-ordination sphere may be induced by close proximity of the exocyclic amino group of the nucleobase and PtII. The versatile metal-ion binding capability of N9-substituted adenine nucleobases has stimulated a great deal of research in the past two decades.1,2 Of the diVerent binding modes reported, the ring nitrogens N1 and N7 are the predominant binding sites for various metal ions,1–4 but co-ordination to the exocyclic N6 amino group upon displacement of a proton has also occasionally been reported.5 The N3 binding mode appears to be very rare in 9-substituted adenines and becomes signifi- cant only when both N1 and N7 are sterically blocked, e.g.in 6,6,9-trimethyladenine.6 The distribution of metal ions between the major binding sites N1 and N7 depends on the pH, metal, and additional ligands co-ordinated to the metal.1 Much attention has been paid to the co-ordination of various platinum(II) compounds to adenine derivatives, in particular, since the adenine N7 site is one of the major targets in DNA for anticancer platinum drugs.7 For example, monofunctionally binding PtII(dien) distributes almost equally between the N1 and N7 sites in neutral adenosine, while the faster reacting palladium analogue favours the N1 site by 5: 1.8 With 59-AMP (adenosine 59- monophosphate) PtII(dien) slightly favours the N7 site.9 Similarly, various aquated cis-platinum(II) diamines show a slight preference for the adenosine N7 site,10 as does cis-[PtCl2- (NH3)2].11 Although selective N7 platination in 9-substituted adenines may be achieved in strongly acidic solution, i.e.when the N1 site is blocked by a proton,1,10 no such procedure that gives N1-platinated species as a major product appears to have been reported.In this work we show that prolonged treatment of aquated PtII(dien) with an excess of adenosine at ca. 85 8C in aqueous solution considerably increases the amount of N1-bound species at the expense of the N7 isomer. The chromatographically purified N1- and N7- bound platinum–adenosine complexes were characterised by 1H, 13C and 195Pt NMR spectroscopy, which show that in both cases PtII exists in two slightly diVerent environments. Both complexes were crystallised as perchlorate salts and their crystal structures determined.To our knowledge they represent † E-Mail: jorma.arpalahti@utu.fi the first examples of X-ray structurally characterised platinum(II)–adenosine complexes. Experimental Materials Adenosine (Ado) was from Sigma and used as received; [Pt- (dien)I]I and its aqua derivative were prepared as previously described.12 For the preparation of [Pt(dien)(Ado-N7)]21 1, adenosine (1.8 mmol) was dissolved in 1 M HNO3 (2 cm3), followed by addition of aquated PtII(dien) (0.36 mmol).After stirring for 3 d at room temperature, NaOH (1.9 mmol) was added and the solution concentrated to about 0.5 cm3. The residue was fractionated on a preparative RP-18 column (40 mm, 30 × 200 mm) using 10% methanol in aqueous 0.1 M NaClO4 (pH 3) as eluent. The combined fractions of 1 were concentrated to about 0.5 cm3, and the excess of electrolyte was removed by refractionation with 15% methanol. Upon cooling, the concentrated solution of 1 gave 190 mg (67%) of [Pt(dien)(Ado-N7)][ClO4]2?H2O 1a as colourless prisms, which had a strong tendency to dim during crystallisation (Found: C, 21.8; H, 3.3; N, 14.4.Calc. for C14H28Cl2N8O13Pt: C, 21.5; H, 3.6, 14.3%). For the preparation of [Pt(dien)(Ado-N1)]21 2, adenosine (5.2 mmol) was suspended in water (50 cm3) and to this was added aquated PtII(dien) (0.56 mmol) in water (15 cm3). The mixture was stirred in a stoppered flask for 5 d at 80–85 8C, after which most of the unchanged adenosine was precipitated upon cooling the mixture with ice.The mixture was filtered and the filtrate, after concentration to about 2 cm3 on a rotary evaporator, was passed through the RP-18 column using 10% methanol in aqueous 0.1 M NaClO4 (pH 3) as eluent. The combined fractions of 2 were then concentrated to about 3 cm3, which gave 230 mg of thin colourless needles upon cooling at 14 8C. About 150 mg of this product were dissolved in water (1 cm3), assisted by gentle warming, and [Pt(dien)(Ado-N1)][ClO4]2?2.77H2O 2a (the amount of crystal water is based on X-ray analysis refinement) slowly crystallised at room temperature as thick, colourless plates which very easily lost water of crystallisation1398 J.Chem. Soc., Dalton Trans., 1998, Pages 1397–1402 (Found for dried substance: C, 21.8; H, 3.5; N, 14.6. Calc. for C14H26Cl2N8O13Pt: C, 22.0; H, 3.4; N, 14.7%). For X-ray analysis, the mother-liquor was removed and the crystals were immediately suspended in paraYn to prevent decomposition.A large plate was cut into pieces under paraYn, and a selected piece sealed in a capillary. NMR studies The NMR measurements were carried out in D2O and [2H7]dmf (1H) or in H2O/D2O (13C and 195Pt) at diVerent temperatures ranging from 250 to 125 8C [2H7]dmf and from 125 to 160 8C (D2O or H2O/D2O). Spectra were acquired primarily on a JEOL Alpha 500 spectrometer equipped with a 5 mm tunable probe operating at 500.16 MHz for 1H, 125.78 MHz for 13C and 107.21 MHz for 195Pt, or a JEOL Lambda 400 spectrometer equipped with a 5 mm CH probe operating at 399.78 MHz for 1H and 100.54 MHz for 13C. The 1H and 13C spectra were referenced internally to sodium 4,4-dimethyl-4- silapentanesulfonate (dss), assigned as d 0.015 for proton and 0 for carbon, and the 195Pt spectra externally to [PtCl4]22 (dPt 21625 from [PtCl6]22).One-dimensional high-resolution proton spectra were acquired with normal single-pulse excitation, 458 flip angle, pulse recycle time of 9 s and spectral widths of 8 kHz consisting of 65 k data points (digital resolution 0.11 Hz per point), zero- filled to 128 k prior to Fourier transformation.Spectra were also acquired with presaturation of the water signal (32 k data points and 5 s of presaturation for a total pulse recycle time of 9 s). Exchange spectroscopy (EXSY) diVerence spectra were acquired using irradiation times of 6–8 s and 8 k data points, zero-filled to 128 k and with 1 Hz exponential weighting applied prior to Fourier transformation.Double quantum filtered (DQF) COSY was acquired in the phase-sensitive mode with spectral widths appropriately optimised from the onedimensional spectra, and processed with zero-filling (×2, ×4) and exponential weighting (1 Hz) applied in both dimensions prior to Fourier transformation. One-dimensional carbon spectra were acquired with normal single-pulse excitation, 458 flip angle, pulse recycle time of 3.5 s and spectral widths of 34 kHz consisting of 64 k data points (digital resolution of 0.52 Hz per point), zero-filled to 128 k and with 1 Hz exponential weighting applied prior to Fourier transformation.The DEPT spectra (90 and 1358) were acquired under similar conditions, but with the pulse recycle time reduced to 3 s. The CH shift correlation and heteronuclear multiple bond correlation (HMBC) spectra were acquired in magnitude mode with spectral widths appropriately optimised from the one-dimensional spectra and processed with zero- filling (×2, ×4), 2p/3-shifted sinebell functions, and exponential weighting (1 Hz) applied in both dimensions prior to Fourier transformation.One-dimensional platinum spectra were acquired with normal single-pulse excitation, 708 flip angle, pulse recycle time of 0.8 s and spectral widths of 118 kHz consisting of 32 k data points (digital resolution 3.6 Hz per point), zero-filled to 64 k and with 50–100 Hz exponential weighting applied prior to Fourier transformation.Crystallography All X-ray data were collected on a Rigaku AFC5S diVractometer at ambient temperature using Mo-Ka radiation (l = 0.710 69 Å). The intensity of three standard reflections remained constant throughout the measurement in both cases. The intensities of the reflections were corrected for Lorentzpolarisation and absorption (empirical) eVects. The structures were solved by standard Patterson and Fourier-diVerence methods and refined by full-matrix least-squares calculations employing the TEXSAN program package 13 (1a) or SHELXL 93 (2a).14 In 1a all atoms except hydrogen and the water oxygen were refined with anisotropic thermal parameters, whereas in 2a only Pt and Cl were anisotropically refined.Further details of the structure refinement are given in Table 3. CCDC reference number 186/891. See http://www.rsc.org/suppdata/dt/1998/1397/ for crystallographic files in .cif format.Results and Discussion Distribution of PtII between the adenosine N1 and N7 sites can be conveniently followed by HPLC. As seen in Fig. 1, the N7 binding mode of [Pt(dien)(H2O)]21 predominates at pH 2 because the N1 site is blocked by a proton.1,10 At higher pH coordination to the N1 site becomes significant, and at pH 5 the apparent N1:N7 ratio based on peak areas at 260 nm is ca. 2 : 3. We have reported earlier that certain 3d metal ions slightly prefer the N7 over the N1 site in 9-substituted adenines.15 However, attempts to block the N7 site by 3d metal ions were not successful.Only in the presence of a large excess of Ni21 ions the amount of N1-platinated species increased; other metal ions tested (Co21, Cu21) were found to be ineVective. By contrast, prolonged treatment (5 d) of an aqueous solution of [Pt(dien)- (H2O)]21 and adenosine at ca. 85 8C significantly increased the amount of N1-bound species at the expense of the N7 isomer.According to HPLC analysis the initial N1:N7 ratio of 2 : 3 slowly increases to ca. 3 : 1 during this treatment. This process parallels the behaviour of the 104–105 times faster reacting palladium(II) analogue 8 and reflects the greater thermodynamic strength of the Pt]N1 bond over the Pt]N7 bond consistent with the diVerence in basicity of these sites. In an equimolar mixture of PtII and adenosine a third (minor) product also appears in the reaction.It is tentatively assigned as the N1,N7-diplatinated species because its concentration could be greatly reduced by employing an excess of the nucleoside.1 This suggests that the N7 to N1 isomerisation proceeds via breaking and reformation of platinum–nucleobase bonds without participation of the C6]NH2 group in the process. NMR Studies The 1H, 13C and 195Pt chemical shifts for complexes 1 and 2 in D2O are listed in Tables 1 and 2. The assignments of the 1H and 13C resonances were based on COSY, C]H correlation, DEPT, HMBC, and by comparison to adenosine.Downfield shifts of Fig. 1 The HPLC elution profiles of mixtures of aquated PtII(dien) and adenosine ([Pt]T : [L]T = 1 : 5) after 2 h of mixing at pH 2 (A), at pH 5 (B), at pH 5 in the presence of Ni21 ([Ni]T : [L]T = 5 : 1) (C) and mixture (B) after 5 d (D). Eluent: 3% MeOH in 0.05 M NaClO4 (pH 3), flow rate 1.0 cm3 min21; LiChrospher RP-18 column (4 × 250 mm, 5 mm)J.Chem. Soc., Dalton Trans., 1998, Pages 1397–1402 1399 Table 1 Proton chemical shifts for complexes 1, 2 and free adenosine * N N N N NH2 Pt H2N NH H2N O CH2OH 1 2 3 7 8 9 1¢ 2¢ 3¢ 4¢ 5¢ dH Compound 1 2 Adenosine H2 8.383 8.65 (br) 8.146 H8 8.901 (br) 8.421 8.283 H19 6.156 (d) [J(H19H29) 5.3] 6.086 (d) [J(H19H29) 5.9] 6.029 (d) [J(H19H29) 6.1] H29 4.786 (t) [J(H19H29) = J(H29H39) 5.0] 4.766 (t) [J(H19H29) = J(H29H39) 5.5] 4.769 (dd) [J(H19H29) 6.1; J(H29H39) 5.3] H39 4.434 (t) [J(H29H39) = J(H39H49) 4.7] 4.428 (dd) [J(H29H39) 5.2; J(H39H49) 3.8] 4.424 (dd) [J(H29H39) 5.3; J(H39H49) 3.4] H49 4.329 (q) [J(H39H49) = J(H49H59) = J(H49H50) 3.5] 4.274 (td) [J(H39H49) = J(H49H50) 3.9; J(H49H59) 2.9] 4.292 (q) [J(H39H49) = J(H49H50); J(H49H50) 3.2] H59 3.963 [d(AB)d] [J(H59H50) 12.9; J(H49H59) 2.7] 3.900 [d(AB)d] [J(H59H50) 12.8; J(H49H59) 3.0] 3.921 [d(AB)d] [J(H59H50) 12.9; J(H49H59) 2.8] H50 3.882 [d(AB)d] [J(H59H50) 13.1; J(H49H50) 3.6] 3.835 [d(AB)d] [J(H59H50) 12.8; J(H49H50) 4.0] 3.839 [d(AB)d] [J(H59H50) 12.9; J(H49H50) 3.6] CH2 of dien 3.26–3.34 (2 H) 3.06–3.16 (4 H) 2.90–2.98 (2 H) 3.24–3.33 (2 H) 3.02–3.13 (4 H) 2.86–2.95 (2 H) * In ppm from dss.Spectra recorded in D2O at ambient temperature. Coupling constants in Hz in square brackets.1400 J. Chem. Soc., Dalton Trans., 1998, Pages 1397–1402 Table 2 Carbon-13 and 195Pt chemical shifts for complexes 1, 2 and free adenosine a dC Compound 1 2 Adenosine dPt 22861 (0.3) b 22875 (0.7) 22925 (0.4) 22936 (0.6) C2 156.55 156.31 155.20 C4 150.74 149.80 151.08 C5 119.60 122.40 121.76 C6 156.93 158.35 158.27 C8 145.16 144.45 143.27 C19 92.20 91.13 91.07 C29 76.78 76.68 76.43 C39 72.80 73.02 73.38 C49 88.62 88.29 88.55 C59 63.78 64.00 64.25 CH2 of dien 56.71; 53.14 56.53, 53.16 a In ppm from dss.Spectra recorded in D2O at ambient temperature. b The values in parentheses refer to the relative intensity of the signal. H8 (1) and H2 (2) of co-ordinated adenosine are in accordance with N7- and N1-platinated species, respectively, analogous to those observed earlier for the corresponding complexes of 9-methyladenine.16–18 Moreover, all signals of the adenosine moiety in 1 and 2 show the expected multiplicity, whereas the proton resonances of the dien group are rather complicated although the eight protons in both cases show the 2:4:2 pattern typical for PtII(dien) complexes.16 By contrast, in the 195Pt NMR spectra both compounds exhibit two platinum resonances near d 22900.This chemical shift is consistent with a PtN4 coordination sphere with a tridentate dien group.6 For 1 the two signals diVer by 14 ppm and for 2 by 11 ppm, significantly less than the diVerence of about 60 ppm between the major resonances of 1 and 2 suggesting two only slightly diVerent environments for PtII in the two complexes. Increasing the temperature up to 60 8C causes merging of the 195Pt signals and considerably simplifies the dien region in both cases, suggesting that exchange between the platinum(II) environments becomes fast on the NMR time-scale at higher temperature.However, the merging of the platinum(II) signals may also be due to broadening and/or the significant temperature dependency of the resonances (a ca. 20 ppm downfield shift is observed on going from 25 to 60 8C). Changing the solvent to [2H7]dmf and lowering the temperature revealed several interesting aspects in the 1H NMR spectra of complexes 1, 2 and free adenosine, as shown in Fig. 2. First, the aromatic proton resonances of both 1 and 2 are clearly split into two components at 250 8C, whereas those of uncomplexed adenosine remain as singlets. Most notable is the large diVerence in the chemical shift of the H8 protons between the two species of 1. Secondly, the C6]NH2 protons of uncomplexed adenosine are no longer equivalent, which indicates restricted rotation about the C6]NH2 bond (the signals remain split at 225 8C but have coalesced at 125 8C).For 1, both adenosine NH2 proton Fig. 2 Downfield section of the 1H NMR spectra of complexes 1, 2 and free adenosine (A) in [2H7]dmf at 250 8C. Notation: (a) dien NH, assignment made from its observable coupling to the dien CH2 (con- firmed with COSY or homodecoupling experiments); (b) adenosine NH2, assignment made from one-dimensional EXSY spectra which indicated them to be exchangeable at 225 or 250 8C. Assignment of the two base protons was left by default and distinction between them by comparison with the D2O spectra.Unknown impurities are denoted as (i) resonances are further split into two components giving a total of four signals for the NH2 group. In 2, instead, one of the NH2 protons remained as a singlet down to 250 8C. (The small probability of accidental overlap of signals from diVerent protons rather than signals from the same proton in diVerent conformers being isochronous is disregarded.) Thirdly, for both complexes the dien NH proton resonance is also split into two components at 250 8C.Like the NH2 protons in free adenosine, the NH2 and dien NH proton resonances of 1 and 2 remain split at 225 8C, but have coalesced at 125 8C. Owing to overlap, the dien NH2 and CH2 proton resonances could not be seen to be discernibly split, but they also showed dramatic broadening associated with dynamic processes. Accordingly, most spectra (i.e. all nuclides) showed, to varying extents, dynamic processes primarily due to slow rotation about the platinum–nucleoside bond and/or conformational change of the dien ligand (not discernible in practice).Usually platinum(II)–nucleobase complexes exhibit rapid rotation about the platinum–nucleobase bond on the NMR timescale, unless the other ligands in the platinum(II) co-ordination sphere (usually amines) are bulky.19,20 With adenine derivatives a few studies have reported restricted rotation about the Pt]N7 or Pt]N1 bond, e.g.in cis-[PtL2(59-AMP-N7)2] even when L2 is not bulky,19 and in cis-[Pt(PMe3)2(made)2]21 (made = 9-methyladenine), 21 respectively. In these cases all nuclides studied (1H and 195Pt for the former and 1H, 13C, 15N, 31P and 195Pt for the latter) indicated restricted rotation. On the other hand, a single 195Pt resonance but a very complex set of resonances for the dien group and splitting of the resonances for H2, H8 and N9(CH3) into two components for [Pt(dien)(tmade-N3)]21 (tmade = 6,6,9-trimethyladenine) have been attributed to conformation changes of the dien ligand.6 In addition to the expected change of the labile protons (NH, OH) and the slow rotation of the Pt]N1 (N7) bond, slow rotation of the C6]NH2 bond was also indicated in both 1 and 2.In dmf exchange of the labile protons could be slowed, and then eVectively stopped at 225 8C; at 250 8C the platinum rotation could also be eVectively stopped, but not the C6]NH2 rotation which was still comparatively fast.Crystal structures of complexes 1a and 2a Selected structural features are shown in Fig. 3 (1) and compiled in Table 4 (2). The molecular structure of the cation 1 depicted in Fig. 3 confirms N7 platination of adenosine. The Pt]N bond lengths are normal, but the angles of the metal co-ordination sphere deviate from the ideal square-planar geometry, as is usual with PtII(dien)X compounds.6,16,22,23 In fact, the platinum(II) coordination sphere of 1 is very similar to that of the corresponding 9-methyladenine complex.16 The lattice water molecule forms weak hydrogen bonds to the sugar hydroxyl groups O(39) [2.85(2)] and O(59) [2.96(3) Å].The packing of 1a is stabilised by hydrogen bonds involving the hydroxyl groups of the sugar moiety and perchlorate oxygens [O(29) ? ? ? N(1i) 2.80(2), O(39) ? ? ? O(8ii) 2.80 (2), O(4) ? ? ? O(59iii) 2.93(2), O(2) ? ? ? N(12iv) 2.96(2) Å; symmetry operations (i) 2��� 2 x, 1 2 y, ��� 1 z, (ii)J.Chem. Soc., Dalton Trans., 1998, Pages 1397–1402 1401 Table 3 Crystal data and data collection parameters for the structures of [Pt(dien)(Ado-N7)][ClO4]2?H2O 1a and [Pt(dien)(Ado-N1)][ClO4]2? 2.77H2O 2a Empirical formula M Crystal system Space group a/Å b/Å c/Å a/8 b/8 g/8 U/Å3 Z Dc/g cm23 F(000) m/mm21 Crystal size/mm q Range/8 Index ranges Reflections collected Independent reflections (Rint) Data, restraints, parameters Goodness of fit Final R indices (all data) Largest diVerence peak and hole/e Å23 1a C14H28Cl2N8O13Pt 782.42 Orthorhombic P212121 (no. 19) 13.520(6) 14.23(1) 12.914(4) 2484(4) 4 2.092 1536 5.999 0.20 × 0.20 × 0.20 1.83–27.53 0 < h,k < 18, 0 < l < 17 3230 2145, —, 339 1.96 (on F) R = 0.054, R9 = 0.040 [I > 3s(I)] 1.75, 21.98 2a C14H31.54Cl2N8O14.77Pt 814.32 Triclinic P1 (no. 1) 12.167(4) 14.644(2) 8.746(2) 105.170(10) 108.40(2) 70.97(2) 1375.6(6) 2 1.966 803 5.376 0.24 × 0.20 × 0.16 1.83–27.53 0 < h < 15, 218 < k < 19, 211 < l < 10 6631 6627 (0.0437) 6626, 97, 359 1.041 (on F2) R1 = 0.0441, wR2 = 0.1064 [I > 2s(I)] R2 = 0.0671, wR2 = 0.1167 0.941, 20.899 21 1 x, y, z, (iii) ��� 1 x, ��� 2 y, 1 2 z, (iv) 1 2 x, 2��� 1 z, ��� 2 z].In fact, the size of the unit cell of 1a is comparable to that of [Pt(dien)(Guo-N7)][ClO4]2 (Guo = guanosine),22 and also similar is that base stacking is not an important packing factor in 1a. Fig. 4 shows the two crystallographically diVerent cations of complex 2. Interestingly, the units show pseudo-symmetry between the base moieties and platinum(II) atoms that is obeyed also by the perchlorate groups. In both units the Pt]N bond lengths and the angles of the platinum co-ordination sphere are normal and similar to those of 1. The dihedral angles between the PtN4 co-ordination plane and the adenine moiety are identical within the estimated standard deviations (e.s.d.s) in both units [75.8(2)8]. By contrast, diVerences between the two units can be seen in the sugar conformation.The unit denoted as A has an anti (-ac) conformation [torsion angle O(49a)]C(19a)] N(9a)]C(4a) 2168.0(7)8], while the B has a high-anti (-sc) Fig. 3 Molecular structure of [Pt(dien)(Ado-N7)]21 cation. Selected interatomic distances (Å) and angles (8) with estimated deviations in parentheses: Pt]N(7) 2.05(1), Pt]N(10) 2.06(1), Pt]N(11) 1.98(2), Pt]N(12) 2.00(2) and Pt ? ? ? N(6) 3.52(2); N(7)]Pt]N(10) 95.4(6), N(7)]Pt]N(11) 175.3(7), N(7)]Pt]N(12) 94.7(6), N(10)]Pt]N(11) 84.2(7), N(10)]Pt]N(12) 169.0(7) and N(11)]Pt]N(12) 86.0(7) conformation [torsion angle O(49b)]C(19b)]N(9b)]C(4b) 281.2(8)8].24 In both units large thermal ellipsoids for the sugar hydroxyl groups indicate structural disorder.In particular, the O(59) atom is disordered and can even adopt two positions in unit B, which also shows partial disorder in the dien group. The packing of 2 is stabilised by moderate hydrogen bonding between the exocyclic amino group and N7 of the neighbouring molecules [N(6a) ? ? ? N(7b) 2.963(8), N(6b) ? ? ? N(7a) 2.976(8) Å].Lattice water forms hydrogen bonds to sugar oxygens and dien nitrogens [O(59c) ? ? ? O(20) 2.61(3), x, y, z 1 1; O(39b) ? ? ? O(20) 2.68(1), N(12a) ? ? ? O(17) 2.87(1) Å]. Also the ring nitrogen N3 may participate in hydrogen bond formation with lattice water [N(3a) ? ? ? O(18) 2.90(1), x, y, z 2 1; N(3b) ? ? ? O(17) 2.99(1) Å; x 1 1, y, z 1 1].Unfortunately, crystallographic data do not give an unambiguous answer for the observed splitting of 195Pt resonances. It is highly unlikely that the two crystallographically diVerent cations of complex 2 could give rise to two platinum resonances, since similar splitting of platinum(II) signals is seen Fig. 4 The two crystallographically diVerent cations of complex 2. Dashed lines represent proposed hydrogen bonds1402 J. Chem.Soc., Dalton Trans., 1998, Pages 1397–1402 also for 1. Rather, it is tempting to attribute two platinum(II) environments in 1 and 2 to the exocyclic amino group, which is known stericaboth N1 and N7 sites.15 Hover, the angles C(5)]C(6)]N(6) and N(1)]C(6)]N(6) do not indicate that the C6]NH2 group and PtII are mutually pushing away each other. In 1 the former is 126(2)8 and the latter 115(2)8, but the diVerence is within limits of 3s due to the poor quality of the crystal. In 2 the corresponding angles are rather diVerent to the two units, viz. 117.9(6) and 119.5(8) in unit A, and 127.5(7) and 116.1(7)8 in unit B, respectively. Also the distances between PtII and N(6) are quite large; 3.52 Å in 1 and 3.25 (unit A) and 3.16 Å (unit B) in 2, suggesting no direct interaction between these atoms.Although hydrogen atoms were not located in the structure determination, the distances above give an estimation of the shortest possible separation for PtII and the C6]NH2 hydrogen, viz. 2.79 Å for 1, 2.83 2 (unit A) and 2.67 Å (unit B) for 2, assuming that the NH is directed towards Pt. Similar short M]H]N contacts typical of hydrogen rather than agostic bonds have recently been documented for a number of platinum(II) complexes.25 Thus, an intramolecular hydrogen bond-like interaction of Pt]H]N cannot be excluded in either case. This type of interaction may also be aVected by the slow rotation about the platinum– nucleoside bond and/or conformational change of the dien ligand. Conclusion The distribution of PtII between the adenosine N1 and N7 sites can be aVected in aqueous solution not only by lowering the pH of the solution, but also by equilibration of the mixture of PtII and nucleoside at high temperature. In the former case, PtII predominantly binds to the N7 site, whereas the latter procedure significantly increases the amount of the N1-bound complex at the expense of the N7 isomer.Both complexes exhibit two 195Pt resonances at ambient temperature and the 1H and 13C NMR spectra recorded in D2O show indications of exchange phenomena through the broadening of 1H and 13C resonances of those nuclei spatially close to PtII. Mirroring the splitting of the platinum signal in D2O for both compounds, in [2H7]dmf the dien NH, adenosine NH2 (now non-equivalent) and both aromatic proton resonances were all shown to be split accordingly at low temperatures (some only at 225 8C, all at 250 8C) for both 1 and 2.The one exception to this was one of the NH2 protons in 2, which remained as a singlet down to 250 8C. In Table 4 Selected interatomic distances (Å) and angles (8) of complex 2a with estimated deviations in parentheses Pt]N(1) Pt]N(10) Pt]N(11) Pt]N(12) Pt ? ? ? N(6) N(1)]Pt]N(11) N(1)]Pt]N(10) N(1)]Pt]N(12) N(10)]Pt]N(11) N(11)]Pt]N(12) N(10)]Pt]N(12) Cation A 2.037(7) 2.064(7) 2.041(8) 2.085(7) 3.246(6) 174.8(3) 91.7(3) 99.7(3) 83.2(3) 85.4(3) 167.1(3) Cation B 2.050(7) 2.018(8) 2.005(9) 2.043(8) 3.168(6) 172.6(3) 93.9(3) 97.2(3) 84.0(3) 84.9(3) 168.9(4) dmf, exchange of the labile protons could be slowed, and then eVectively stopped at 225 8C; at 250 8C the platinum rotation could also be eVectively stopped, but not the C6]NH2 rotation which was still comparatively fast.The restricted rotation about the platinum(II)–nucleoside and C6]NH2 bonds may also aVect the disturbance to the platinum(II) co-ordination sphere induced by an intramolecular hydrogen bond-type interaction of N6]H? ? ? Pt, assuming coplanarity between the C6]NH2 group and the base moiety. In both 1 and 2 the minimum distance between a NH2 proton and PtII is within the range of 2.2–3.25 Å given in literature for this type of interaction.25a Acknowledgements We thank COST group D8-004-96 for support and stimulating discussions. References 1 B.Lippert, Prog. Inorg. Chem., 1989, 37, 1. 2 R. B. Martin, in Metal Ions in Biological Systems, eds. H. Sigel and A. Sigel, Marcel Dekker, New York, 1996, vol. 32, ch. 3. 3 H. Sigel, N. A. Corfu, L.-N. Ji and R. B. Martin, Comments Inorg. Chem., 1992, 13, 35. 4 J. Arpalahti, in Metal Ions in Biological Systems, eds. H. Sigel and A. Sigel, Marcel Dekker, New York, 1996, vol. 32, ch. 10. 5 J.-P. Charland, M. Simard and A. L. Beauchamp, Inorg. Chim. Acta, 1983, 80, L 30; J.-P. Charland, M. T. Phan Viet, M. St.-Jacques and A. L.Beauchamp, J. Am. Chem. Soc., 1985, 107, 8202; F. Zamora, M. Kunsman, M. Sabat and B. Lippert, Inorg. Chem., 1997, 36, 1583. 6 C. Meiser, B. Song, E. Freisinger, M. Peilert, H. Sigel and B. Lippert, Chem. Eur. J., 1997, 3, 388. 7 J. Reedijk, Chem. Commun., 1996, 801. 8 M. C. Lim and R. B. Martin, J. Inorg. Nucl. Chem., 1976, 38, 1915. 9 R. B. Martin, Acc. Chem. Res., 1985, 18, 32. 10 J. Arpalahti and P. Lehikoinen, Inorg. Chim. Acta, 1989, 159, 115. 11 K. Inagaki, M. Kuwayama and Y. Kidani, J. Inorg. Biochem., 1982, 16, 59. 12 J. Arpalahti and P. Lehikoinen, Inorg. Chem., 1990, 29, 2564. 13 TEXSAN-TEXRAY, Single crystal structure analysis software, version 5.0, Molecular Structure Corporation, The Woodlands, TX, 1989. 14 G. Sheldrick, SHELXL 93, Program for the Refinement of Crystal Structures, University of Göttingen, 1993. 15 J. Arpalahti and E. Ottoila, Inorg. Chim. Acta, 1985, 107, 105. 16 F. Schwarz, B. Lippert, H. Schöllhorn and U. Thewalt, Inorg. Chim. Acta, 1990, 176, 113. 17 J. H. J. den Hartog, H. van den Elst and J. Reedijk, J. Inorg. Biochem., 1984, 21, 83. 18 R. Beyerle and B. Lippert, Inorg. Chim. Acta, 1982, 66, 141. 19 M. D. Reily and L. G. Marzilli, J. Am. Chem. Soc., 1986, 108, 6785. 20 A. T. M. Marcelis, J. L. van der Veer, J. C. M. Zwetsloot and J. Reedijk, Inorg. Chim. Acta, 1983, 78, 195. 21 L. Schenetti, A. Mucci and B. Longato, J. Chem. Soc., Dalton Trans., 1996, 299. 22 R. Melanson and F. D. Rochon, Can. J. Chem., 1979, 57, 57. 23 J. F. Britten, C. J. L. Lock and W. M. Pratt, Acta Crystallogr., Sect. B, 1982, 38, 2148; G. Frommer, H. Preut and B. Lippert, Inorg. Chim Acta, 1992, 193, 111. 24 W. Saenger, Principles of Nucleic Acid Structure, Springer, New York, 1984, ch. 4. 25 (a) W. Yao, O. Eisenstein and R. H. Crabtree, Inorg. Chim. Acta, 1997, 254, 105; (b) D. Braga, F. Grepioni, E. Tedesco, K. Biradha and G. R. Desiraju, Organometallics, 1997, 16, 1846. Received 12th December 1997; Paper 7/08955D

ISSN:1477-9226    

DOI:10.1039/a708955d

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1998, Pages 1403–1410 1403 Rhodium carbonyl complexes containing pyridine; crystal structure of an unusual octahedral rhodium(I) complex [Rh2(Ï-CO)3Cl2(py)4] Brian T. Heaton,* Chacko Jacob and Jeyagowry T. Sampanthar Chemistry Department, University of Liverpool, Liverpool, UK L69 3BX The products resulting from the progressive addition of pyridine (py) to a solution of [Rh2(m-Cl)2(CO)4] 1 have been found to depend both upon the solvent and the atmosphere (CO or N2).In CH2Cl2 under N2, cis- [Rh(CO)2Cl(py)] 2, [Rh(CO)2Cl(py)2] 3 and [Rh2(m-CO)3Cl2(py)4] 4 were obtained successively; under CO, 4 was converted into 3 and under N2 disproportionation of 4 slowly occurred to give 3 and trans-[Rh(CO)Cl(py)2] 5 which reacted with CO to give 3. In more polar solvents (thf or MeOH), 1 reacted under N2 to give the lightly solvent-stabilised complex [Rh(CO)2Cl(solv)] 6 (solv = thf a or MeOH b) and, in the presence of AgClO4, cis- [Rh(CO)2(solv)2]1 7 (solv = thf a or MeOH b); additionally, when solv = MeOH there was spectroscopic evidence for the formation of [Rh2(m-CO)x(MeOH)y]21 8 (x = 2, y = 4 or x = 3, y = 6) which reacted with CO to give [Rh(CO)2(MeOH)2]1.Complex 7 reacted with py to give successively cis-[Rh(CO)2(py)2]1 9 and [Rh(CO)(py)3]1 10; under CO 10 was converted into 9. The stereochemistry of all the above complexes has been established through a combination of IR and multinuclear (13C, 15N, 103Rh) NMR measurements and X-ray crystallography for 2 and 4.Analogous reactions have been carried out using trans-[Rh2(m-Cl)2(CO)2(CxHy)2] 11 (CxHy = C2H4 a or C8H14 b) under a nitrogen atmosphere and spectroscopic measurements on the reaction of 11a with pyridine were consistent with the successive formation of [Rh2(m-Cl)2(CO)2(C2H4)2(py)2] 12, which lost ethylene and rearranged to give [Rh2(m-CO)2Cl2(py)2] 13. Reaction of 11b with pyridine gave immediately trans-[Rh2(m-Cl)2(CO)2(py)2] 14 and both 13 and 14 further reacted with pyridine to give 5.Compared to the plethora of data available on mono- and dinuclear rhodium complexes containing both CO and P-donor ligands which are catalytically active, there are few reports of analogous rhodium carbonyl complexes containing N-donor ligands. A common cause of catalytic deactivation with Rh/PR3 homogeneous catalysts is P]C bond cleavage and progressive replacement of phosphines by N-donor ligands should increasingly alleviate this problem.Successful examples using this approach are illustrated by the following. (a) The complex [Rh2{(NH2NH)2CH2}(PPh3)4]21, which results from the reaction of Rh(NO3)3–PPh3–N2H4 in MeOH, is used commercially as a more selective eYcient hydrogenation catalyst than [RhCl(PPh3)3].1 (b) Dissolution of [RhCl(PPh3)3] in pyridine selectively catalyses the dimerisation of enolisable Cn aldehydes into saturated dimers, C2n monoaldehydes, which are useful in the perfumery industry.2 (c) The water gas shift reaction is catalysed by an aqueous pyridine (or substituted pyridine) solution of RhCl3 3 and rhodium carbonyl clusters in the presence of N-heterocyclic ligands exhibit high catalytic activity under mild conditions; 4 these are much more eVective than other transition metal/cluster catalysts.4 Fachinetti et al.5 examined these reactions in much more detail; they were able to isolate and identify [Rh(CO)n(py)4 2 n] [Rh5(CO)13(py)2] (n = 2 or 3) and proposed a mechanism for the catalytic cycle.This paper concentrates on the preparation and characterisation of rhodium complexes containing CO and N-heterocyclic ligands. Previously reported complexes of this type have often been formulated only on the basis of IR, 1H NMR and elemental analysis, with few reports of structural characterisation from 13C NMR and/or X-ray crystallography, and the proposed formulations are sometimes contradictory.The complex [Rh2(m-Cl)2(CO)4] 1 undergoes bridge cleavage on reaction with pyridine to give cis-[Rh(CO)2Cl(py)] 2,6 and consistent with this formulation has two equally intense bands due to n(CO) in the IR spectrum 7 and two equally intense doublets in the 13C NMR spectrum.8 However, further addition of py to 1 has been claimed by Hieber et al.6 to give [Rh2- (m-Cl)2(CO)4(py)4], on the basis of elemental analysis and diamagnetism, whereas later studies by Lawson and Wilkinson 9 reformulated this product as a neutral five-co-ordinate complex [Rh(CO)2Cl(py)2].We show later that these structural formulations are both incorrect and this product is actually [Rh2- (m-CO)3Cl2(py)4]. Similarly, we have reinvestigated the nature of the products formed on the progressive addition of py to trans- [Rh2(m-Cl)2(CO)2(CxHy)2] (CxHy = C2H4 a or C8H14 b).10 In both this reaction and that of 1 with pyridine we have made extensive use of multinuclear NMR measurements [including 15N NMR measurements using the insensitive nucleii enhanced by polarisation transfer (INEPT) pulse sequence,11 which we have found to be most useful] structurally to characterise the products and we find that the nature of the product often varies on changing either the solvent or reaction atmosphere (N2 or CO).Results and Discussion [Rh2(Ï-Cl)2(CO)4] 1 1 py The nature of the products resulting from the incremental addition of py to complex 1 depends on the solvent and we now report on the nature of the products formed on reaction in lowpolarity (e.g.CH2Cl2) versus polar solvents (e.g. thf, MeOH). (a) Reaction in CH2Cl2. The products resulting from the incremental addition of py to complex 1 in CH2Cl2 under a nitrogen atmosphere are shown in Scheme 1. The first formed product 2 has been reported 6 and characterised by IR 7 and 13C NMR spectroscopy.8 In order to build up a spectroscopic and structural data base of Rh/CO/py complexes, we have measured the 15N NMR spectrum (see Table 1) and also determined the crystal structure (see Fig. 1). The complex cis-[Rh(CO)2Cl(py)] 2 crystallises with two molecules in the asymmetric unit; there is no intermolecular coordination between these two molecules which show only marginal diVerences in bond lengths/angles. Selected bond lengths/ angles of the two molecules are given in Table 2. The geometry around rhodium is square planar and, consistent with the1404 J.Chem. Soc., Dalton Trans., 1998, Pages 1403–1410 solution spectroscopic data, there are two cis-carbonyls. The dihedral angle between the plane of the pyridine ring and the rhodium(I) square plane is 1418 and is very similar to that found for cis-[Ir(CO)2Cl(py)] (1438).13 Further addition of py to complex 1 or 2 (Rh:py = 1:2) gives a new compound for which microanalytical data suggest the empirical formula [Rh(CO)2Cl(py)2] 3 and, since it is a non-conductor in both CH2Cl2 and MeOH solution, can be formulated as a five-co-ordinate neutral complex.For this formulation there are five possible isomers (see Scheme 2) and a combination of IR and 13C/15N NMR measurements allows 3 to be assigned unambiguously to isomer A. Thus, there are two equally intense doublets in the 15N-{1H} spectrum at 223 K, a sharp doublet in the carbonyl region of the 13C NMR spectrum which remains unchanged from 297 to 193 K and two strong n(CO) bands in the IR spectrum (see Table 1).Scheme 1 Products resulting from the progressive addition of py to [Rh2(m-Cl)2(CO)4] 1 in CH2Cl2 Rh Cl Rh CO Cl CO Rh Cl N Rh Rh Cl N N Rh Cl N N N N OC 2 py OC OC OC OC N2 C O OC Cl OC C O N2 N2 1 py 4 py 3 4 excess py N2 CO Fig. 1 Crystal structure of cis-[Rh(CO)2Cl(py)] 2 Attempts to obtain crystals of complex 3 suitable for X-ray analysis failed but further evidence for the formation of the five-co-ordination has been obtained by addition of 1 equivalent of NBun 4Cl to a solution of cis-[Rh(CO)2(py)2]ClO4 which, on the basis of NMR measurements, gives exclusively 3 [equation (1)].cis-[Rh(CO)2(py)2]1 NBun 4Cl AgClO4 [Rh(CO)2Cl(py)2] (1) 3 It should be noted that Fachinetti et al.5 proposed, on the basis of IR evidence, that addition of 1 equivalent of [N(PPh3)2]Cl to a pyridine solution of cis-[Rh(CO)2(py)2][BPh4] under a CO atmosphere gives cis-[Rh(CO)2Cl(py)] 2. However, since the bands due to n(CO) for 2 are almost identical to those of 3 we suspect that, on the basis of our NMR measurements, their product should have been formulated as 3 rather than 2 [see equation (1)].Over ca. 12 h in CH2Cl2 solution complex 3 does not lose CO under a nitrogen atmosphere and does not lose pyridine under a CO atmosphere. However, in the presence of a trace of pyridine and under a nitrogen atmosphere it loses CO immediately to give the air-stable yellow solid [Rh2(m-CO)3Cl2(py)4] 4 which has been prepared previously6,9 but wrongly formulated.We have been able fully to characterise this compound both crystallographically and spectroscopically. The structure is shown in Fig. 2 and selected bond lengths and angles are given in Table 3. It has a pseudo-two-fold axis which is coincident with the unique bridging carbonyl vector, C(21)O(1). As a result there are two diVerent types of bridging carbonyls in the ratio 1 : 2 and two diVerent types of pyridines in the ratio 2 : 2 with two pyridines being trans to the unique carbonyl, C(21)O(1), and the other two equivalent pyridines trans to C(22)O(2) and C(23)O(3) respectively. This structure, which is retained in solution, see below, is very unusual and to our knowledge is the only example of a dinuclear rhodium(I) complex containing an octahedral array of ligands about the metal.Consistent with each rhodium having 17 electrons, a Rh]Rh bond is formed and because of the compact geometry the Rh]Rh distance is very short (2.566 Å), comparable to distances found for dinuclear rhodium(I) A-frame complexes, e.g. 2.612 Å in [Rh2(m-CO)(m-h2- Scheme 2 Possible isomers of [Rh(CO)2Cl(py)2] 3 Rh py Cl py Rh py py Cl Rh CO py Cl Rh Cl CO py Rh CO CO Cl OC OC py py py OC OC OC OC py A B C D E Fig. 2 Crystal structure of [Rh2(m-CO)3Cl2(py)4] 4J. Chem. Soc., Dalton Trans., 1998, Pages 1403–1410 1405 Table 1 Spectroscopic data for mono- and di-nuclear rhodium(I) carbonyl complexes containing pyridine IR 13C NMR 15N NMR Complex 2 cis-[Rh(CO)2Cl(py)] Solvent CH2Cl2 CH2Cl2 n& (CO)/cm21 2089s, 2015s, 2089s, 2014s b Solvent CH2Cl2 C6H5Me T/K 223 5 213 5 d(13CO)a 183.9(1) 180.2(1) 185.6(1) c 181.3(1) c 1J(Rh]CO)/ Hz 66.2 74.9 67 73 Solvent CH2Cl2 T/K 223 d(15N)a 2143.8 1J(Rh]N)/ Hz 15.9 3 [Rh(CO)2Cl(py)2] CH2Cl2 2089s, 2015s CH2Cl2 CH2Cl2 297 223 182.6 182.1 69.6 70.6 CH2Cl2 223 5 2128.6(1) 2128.3(1) 16.8 17.6 4 [Rh2(m-CO)3Cl2(py)4] CH2Cl2 1835s, 1814s CH2Cl2 223 5 213.2 (t, 2) 211.7 (t, 1) 28.7 27.6 5 trans-[Rh(CO)Cl(py)2] CH2Cl2 CH2Cl2 1962 1961 d CH2Cl2 CH2Cl2 297 223 184.5 183.7 83.9 84.7 CH2Cl2 CH2Cl2 297 213 2152.2 2153.3 18.6 18.6 6a cis-[Rh(CO)2Cl(thf)] thf 2083s, 2011s thf thf 297 183 5 181.5 185.4(1) 181.4(1) 74.3 73.8 71.6 6b cis-[Rh(CO)2Cl(MeOH)] MeOH 2087s, 2014s MeOH MeOH 297 193 5 182.2 184.5(1) 180.6(1) 74.6 78.9 70.6 7a cis-[Rh(CO)2(thf)2]1 thf 2102s, 2029s thf 223 181.6 76.6 7b cis-[Rh(CO)2(MeOH)2]1 MeOH 2105s, 2031s MeOH 223 181.2 75.7 8 [Rh2(m-CO)x(MeOH)y]21 (x = 2, y = 4 or x = 3, y = 6) MeOH 1865 MeOH 223 198.4 (t) 30.1 9 cis-[Rh(CO)2(py)2]1e CH2Cl2 thf py 2104s, 2043s 2098s, 2033s 2100s, 2038s CH2Cl2 thf 223 297 182.2 184.5 69.2 69.4 thf 297 2153.4 19 10 [Rh(CO)(py)3]1 thf py 1991 1993 f thf MeOH 223 223 189.8 191.0 78.9 78.2 11a trans-[Rh2(m-Cl)2(CO)2- (C2H4)2] CH2Cl2 C6H6 2022 2010 d CH2Cl2 223 178.8 77.6 11b trans-[Rh(m-Cl)2(CO)2- (C8H14)2] 12 [Rh2(m-Cl)2(CO)2(C2H4)2- (py)2] 13 [Rh2(m-CO)2Cl2(py)2] 14 trans-[Rh2(m-Cl)2(CO)2- (py)2] CH2Cl2 CH2Cl2 Nujol CH2Cl2 2008 2089s, 2006s 1789 2015 CH2Cl2 CH2Cl2 CH2Cl2 297 223 213 182.2 184.7 187.4 82.3 68.4 72.7 CH2Cl2 213 2127.3 13.5 a Figures in parentheses are relative intensities; all signals appear as doublets except as shown.b Ref. 7. c Ref. 8. d Ref. 10. e d(Rh) 73. f Ref. 12. NC5H4PPh2-2)2Cl2] 14 and 2.639 Å in [Rh2(m-CO)(m-NC5H4- PPh2-2)2(h2-H2BH2)2].15 Complex 4 is insoluble in most polar solvents (e.g.thf, MeOH, Me2CO and MeCN) and non-polar solvents (e.g. C6H6, MeC6H5) but slightly soluble in CH2Cl2. Using a freshly prepared solution in CH2Cl2 the spectroscopic data are entirely consistent with the solid-state structure. Thus, there are two almost equally intense n(CO) bands in the bridging region and the 13C Table 2 Selected bond lengths (Å) and angles (8) for the two molecules in the asymmetric unit of cis-[Rh(CO)2Cl(py)] 2 Rh(1)]Cl(1) Rh(1)]N(1) Rh(1)]C(1) Rh(1)]C(2) Rh(2)]Cl(2) Rh(2)]N(2) Rh(2)]C(8) Rh(2)]C(9) O(1)]C(1) O(2)]C(2) O(3)]C(8) O(4)]C(9) Cl(1)]Rh(1)]N(1) Cl(1)]Rh(1)]C(1) Cl(1)]Rh(1)]C(2) N(1)]Rh(1)]C(1) N(1)]Rh(1)]C(2) C(1)]Rh(1)]C(2) 2.347(4) 2.122(7) 1.84(1) 1.81(1) 2.344(3) 2.114(7) 1.84(1) 1.83(1) 1.12(1) 1.14(1) 1.13(2) 1.13(2) 91.3(2) 177.5(3) 87.4(4) 91.2(4) 175.7(5) 90.1(5) N(1)]C(3) N(1)]C(7) N(2)]C(10) N(2)]C(14) C(3)]C(4) C(4)]C(5) C(5)]C(6) C(6)]C(7) C(10)]C(11) C(11)]C(12) C(12)]C(13) C(13)]C(14) Cl(2)]Rh(2)]N(2) Cl(2)]Rh(2)]C(8) Cl(2)]Rh(2)]C(9) N(2)]Rh(2)]C(8) N(2)]Rh(2)]C(9) C(8)]Rh(2)]C(9) 1.35(1) 1.34(1) 1.35(1) 1.35(1) 1.37(1) 1.38(2) 1.38(1) 1.37(1) 1.35(1) 1.38(1) 1.39(2) 1.36(1) 90.6(2) 176.6(3) 87.2(5) 92.8(4) 177.7(5) 89.4(6) NMR spectrum consists of two triplets in the ratio 1 : 2 due to the bridging carbonyls (see Table 1); it was unfortunately impossible to obtain 15N NMR data because of the low solubility and long collection times required.Under an atmosphere of CO complex 4 is converted quantitatively into 3 whereas, under a nitrogen atmosphere, it undergoes slow disproportionation [equation (2)].The resultant Table 3 Selected bond lengths (Å) and angles (8) for [Rh2(m-CO)3- Cl2(py)4] 4 Rh(1)]Rh(2) Rh(1)]Cl(1) Rh(1)]N(1) Rh(1)]N(2) Rh(1)]C(21) Rh(1)]C(22) Rh(1)]C(23) Rh(2)]Cl(2) Rh(2)]Rh(1)]Cl(1) Rh(2)]Rh(1)]N(1) Rh(2)]Rh(1)]N(2) Rh(2)]Rh(1)]C(21) Rh(2)]Rh(1)]C(22) Rh(2)]Rh(1)]C(23) Cl(1)]Rh(1)]N(1) Cl(1)]Rh(1)]N(2) Cl(1)]Rh(1)]C(21) Cl(1)]Rh(1)]C(22) Cl(1)]Rh(1)]C(23) 2.566(5) 2.462(8) 2.185(3) 2.198(3) 2.021(3) 2.002(3) 1.997(3) 2.472(9) 121.62(2) 126.41(7) 125.39(7) 50.31(9) 49.94(9) 50.58(9) 90.24(7) 91.79(7) 91.75(9) 91.62(9) 171.97(9) Rh(2)]N(3) Rh(2)]N(4) Rh(2)]C(21) Rh(2)]C(22) Rh(2)]C(23) C(21)]O(21) C(22)]O(22) C(23)]O(23) N(1)]Rh(1)]N(2) N(1)]Rh(1)]C(21) N(1)]Rh(1)]C(22) N(1)]Rh(1)]C(23) N(2)]Rh(1)]C(21) N(2)]Rh(1)]C(22) N(2)]Rh(1)]C(23) C(21)]Rh(1)]C(22) C(21)]Rh(1)]C(23) C(22)]Rh(1)]C(23) 2.188(3) 2.223(3) 2.011(3) 1.995(3) 2.016(3) 1.155(4) 1.164(4) 1.164(4) 91.19(9) 91.4(1) 176.3(1) 96.5(1) 175.6(1) 92.0(1) 92.4(1) 85.3(1) 83.8(1) 81.4(1)1406 J.Chem. Soc., Dalton Trans., 1998, Pages 1403–1410 [Rh2(m-CO)3Cl2(py)4] CH2Cl2 N2 [Rh(CO)2Cl(py)2] 1 4 3 trans-[Rh(CO)Cl(py)2] (2) 5 solution shows, in addition to the expected carbonyl IR bands and 13C NMR resonances due to 3, a band at 1962 cm21 and a doublet in the 13C NMR spectrum [d 184.5, 1J(Rh]CO) 83.9 Hz)] of half the intensity of the other carbonyl resonance due to 3; we attribute these additional bands to trans-[Rh(CO)- Cl(py)2] 5.This formulation is further substantiated by bubbling CO through this solution which results in the formation of 3 and complete disappearance of the n(CO) band/13C NMR resonance attributed to 5, equation (3). However, the reverse is not observed on bubbling N2 through a solution of 3. trans-[Rh(CO)Cl(py)2] CH2Cl2/CO [Rh(CO)2Cl(py)2] (3) 5 3 (b) Reaction in polar solvents (MeOH, thf).Dissolution of complex 1 in either MeOH or thf under N2 results in bridge cleavage to give the lightly stabilised solvento-complex 6 [equation (4)]. It is surprising that this reaction has not [Rh2(m-Cl)2(CO)4] solv/N2 2 cis-[Rh(CO)2Cl(solv)] (4) 1 6a solv = thf 6b solv = MeOH previously been reported since both the IR and 13CO resonances of 6 are significantly diVerent to the analogous data for 1 in CH2Cl2 or CHCl3 solution (see Table 1). Unambiguous evidence of the formulation of 6 comes from the variabletemperature 13CO NMR spectra of [Rh2(m-Cl)2(13CO)4] at low temperature (see Fig. 3) which for both MeOH and thf show two equally intense resonances (see Table 1); at room temperature there is rapid solvent exchange giving rise to a time-averaged carbonyl chemical shift [equation (5)]. The lowtemperature chemical shift diVerences between CaO and CbO for solv = thf or MeOH is rather similar but the coalescence temperature of the two doublets is 213 and 203 K for MeOH and thf respectively.This implies that thf is slightly more labile than MeOH. Unfortunately, isolation of 6a and 6b proved impossible. Addition of 2 equivalents of AgClO4 to solutions of complex 1 in MeOH or thf under a nitrogen atmosphere produces slightly diVerent results. The cleanest reaction occurs in thf to give cis-[Rh(CO)2(solv)2]1 7a (solv = thf) which has IR and NMR data entirely consistent with this formulation. A similar reaction occurs in MeOH to give predominantly 7b but there is also evidence for the formation of minor amounts of [Rh2(m-CO)x- (MeOH)y]21 8 as evidenced by a weak band at 1865 cm21 in the IR spectrum and the presence of a triplet at d 198.4 [1J(Rh]CO) 30.1 Hz] in the low-temperature 13C NMR spectrum.On the basis of the spectroscopic evidence we cannot presently be sure of the values of x and y which are likely to be either x = 2 and y = 4 or x = 3 and y = 6. However, bubbling CO through a MeOH solution containing 8 results in the immediate transformation into 7b.Rh Cl S OCb OCa Rh S Cl OCb OCa solv solv = thf or MeOH (5) Addition of pyridine (Rh :py = 1 : 2) to a solution of complex 7a in thf or 7b in MeOH under CO gives the same product, [Rh(CO)2(py)2]1 9 and the IR data of this complex are in agreement with previously reported results.12 We have recorded the 15N-{1H} NMR spectrum of 9 which, as expected, consists of one doublet (see Table 1). This complex can also be prepared by bubbling CO into a solution of cis-[Rh(diene)(py)2]1 (diene = cod or nbd) in thf.However, it should be noted that 9 is unstable under N2 and is transformed overnight into presently unidentified products. Nevertheless, immediate addition of 1 equivalent of py to a solution of 9 in thf or MeOH under N2 produces a new band in the IR spectrum at 1991 cm21 and the 13C-{1H} NMR spectrum of the solution at 223 K shows a new doublet at d 189.8 [1J(Rh]CO) 79 Hz]; we attribute these to [Rh(CO)(py)3]1 10 (see Table 1).Reaction of 10 with CO immediately gives 9. The reactions of [Rh2(m-Cl)2(CO)4] 1 with py in polar solvents (thf or MeOH) are summarised in Scheme 3. trans-[Rh2(Ï-Cl)2(CO)2(CxHy)2] 11 (CxHy = C2H4 a or C8H14 b) 1 py Mixing equimolar amounts of [Rh2(m-Cl)2(CO)4] and [Rh2- (m-Cl)2(CxHy)4] (CxHy = C2H4 or C8H14) in CH2Cl2 solution has previously been reported 10 on the basis of IR measurements to give exclusively [Rh2(m-Cl)2(CO)2(CHy)2] 11 (CxHy = C2H4 a or C8H14 b).This result is a little surprising since a statistical distribution of products might have been expected but we only find the presence of one carbonyl doublet in the 13C NMR spectrum both at room and low temperature (see Table 1) consistent with the presence of only the trans isomer. Previous IR and 1H NMR measurements 10 suggest that addition of pyridine to a solution of complex 11a in CH2Cl2 solution under N2 (Rh:py = 1 : 1) results in retention of C2H4 and formation of [Rh2(m-Cl)2(CO)2(C2H4)2(py)2] 12.We have confirmed these measurements and, in addition, the presence Fig. 3 Variable-temperature 13C NMR spectra of cis-[Rh(13CO)2Cl- (MeOH)] 6a in MeOH; r.t. = room temperatureJ. Chem. Soc., Dalton Trans., 1998, Pages 1403–1410 1407 of one doublet [d 184.7, 1J(Rh]CO) 68 Hz] in the 13C NMR spectrum is consistent with the product shown in Scheme 4. On standing the above solution under a reduced nitrogen pressure a very insoluble black precipitate 13 is formed.Microanalysis of 13 is consistent with the empirical formula [{Rh(CO)Cl(py)}n] and the mass spectrum is consistent with n = 2 as reported previously. However, IR data on 13 had not been reported previously and we find that the IR spectrum in Nujol contains no bands in the terminal carbonyl region and only one strong n(CO) band at 1789 cm21. As a result 13 must be formulated as shown in Scheme 4 rather than trans-[Rh2(m-Cl)2(CO)2(py)2] as previously reported.10 Bubbling CO through a suspension of 13 in CH2Cl2 gives cis-[Rh(CO)2Cl(py)] 2 and addition of py to 12 (Rh:py = 1 : 2) results in the formation of 5.This has been reported previously10 on the basis of IR and elemental analysis and our 13C and 15N NMR data (Table 1) are entirely consistent with this formulation. Retention of cyclooctene does not occur on addition of pyridine to complex 11b (Rh:py = 1 : 1) under a nitrogen atmosphere.In this case immediate loss of cyclooctene occurs to give trans-[Rh2(m-Cl)2(CO)2(py)2] 14. It should be noted in this case that there is only spectroscopic evidence for terminal COs (see Table 1) and surprisingly we have no evidence for rearrangement of 14 to 13 or vice versa. Complex 14 reacts with py (Rh:py 1 : 1) to give 5 which reacts further with an excess of pyridine in MeOH to give the ionic species [Rh(CO)(py)3]Cl 10. Experimental All manipulations involving solutions or solids were performed under an atmosphere of N2 or CO as appropriate, using standard Schlenk-line techniques.All solvents were dried and distilled under N2 immediately prior to use. Deuteriated solvents were stored over activated molecular sieves (4 Å) for at least 24 h prior to use. Manipulations using 13CO were carried out using standard vacuum-line techniques. The 13C NMR spectra were obtained on Bruker WM200, 250 and AMX400 spectrometers using internal SiMe4 as reference, Scheme 3 Products resulting from the reaction of [Rh2(m-Cl)2(CO)4] 1 with pyridine in polar solvents (thf or MeOH) cis-[Rh(CO)2(solv)2]+ [Rh2(m-CO) x(solv) y]2+ CO solv = MeOH 2py solv = thf or MeOH OC cis-[Rh(CO)2(py)2]+ N2 + 1 py [Rh(CO)(py)3]+ 9 7 [Rh2(m-Cl)2(CO)4] 1 N2 solv (solv = thf, a; or MeOH b) cis-[Rh(CO)2Cl(solv)] 6 N2 1AgClO4 2AgClO4 solv = thf N2 8 ( x = 2, y = 4 or x = 3, y = 6) 10 thf, a; MeOH, b) (solv = MeOH) (solv = CO – 1 py 15N INEPT spectra on a Bruker AMX400 spectrometer using the previously published method 16 with chemical shifts referenced to external MeNO2 and 103Rh NMR spectra on a Bruker WM360 spectrometer using a 15 mm NMR tube {7 cm3 of a ca. 0.8 mmol solution containing ca. 25 mg of [Cr(acac)3] as relaxing agent} with shifts referenced to 3.16 MHz (d 0) at such a magnetic field that the protons in SiMe4 resonate at exactly 100 MHz. The IR spectra were obtained in solution with CaF2 windows on a Perkin-Elmer 1720x FTIR spectrometer.The starting materials were prepared using literature methods: [Rh2(m-Cl)2(C2H4)4],17 [Rh2(m-Cl)2(C8H14)4] 18 and [Rh2(m-Cl)2(nbd)2].19 In order to avoid using expensive RhCl3? 3H2O, [Rh2(m-Cl)2(CO)4] was prepared by heating [Rh2(m-Cl)2- (nbd)2] to 100 8C in a stream of CO whereupon the product sublimed, yield 80%. The preparation of [Rh2(m-Cl)2(13CO)4] was carried out by stirring a suspension of [Rh2(m-Cl)2(C2H4)4] in light petroleum (b.p. 40–60 8C) with a slight excess of 13CO.On cooling this solution to 278 8C the product precipitated and removal of the solvent by a syringe left the pure product, yield 65%. Preparation of complexes cis-[Rh(CO)2Cl(py)] 2. Pyridine (83 ml, 1.0 mmol) was added to a solution of [Rh2(m-Cl)2(CO)4] (200 mg, 0.5 mmol) in CH2Cl2 (5 cm3), under a nitrogen atmosphere, to give a pale yellow solution. X-Ray-quality orange crystals of cis-[Rh(CO)2- Cl(py)] were obtained on adding diethyl ether (10 cm3) and leaving the solution to evaporate slowly under a nitrogen atmosphere.The product was filtered oV and dried under vacuum (260 mg, 93%) (Found: C, 30.7; H, 1.8; N, 5.1. C7H5ClNO2Rh requires C, 30.7; H, 1.8; N, 5.1%). [Rh(CO)2Cl(py)2] 3. Pyridine (210 ml, 2.57 mmol) was added to a solution of [Rh2(m-Cl)2(CO)4] (250 mg, 0.64 mmol) (Rh:py = 1 : 2) in CH2Cl2 (3 cm3) under a nitrogen atmosphere to give a yellow solution of the product. Yellow crystals were obtained by layering this solution with light petroleum but unfortunately they were unsuitable for X-ray analysis (395 mg, 87%) (Found: C, 40.5; H, 2.8; N, 7.5.C12H10ClN2O2Rh requires C, 40.9; H, 2.9; N, 7.95%). [Rh2(Ï-CO)3Cl2(py)4] 4. Pyridine (310 ml, 3.86 mmol) was added dropwise to a solution of [Rh2(m-Cl)2(CO)4] (250 mg, 0.64 mmol) in CH2Cl2 (3 cm3) under a nitrogen atmosphere. The IR and 13C NMR measurements showed the formation of [Rh2(m-CO)3Cl2(py)4]. X-Ray-quality yellow crystals were obtained from this solution by leaving it under a slow flow of N2.The product was filtered oV and dried under vacuum (380 mg, 87%) (Found: C, 40.8; H, 2.9; Cl, 10.7; N, 8.3. C23H20Cl2N4O3Rh2 requires C, 40.8; H, 3.0; Cl, 10.5; N, 8.3%). trans-[Rh(CO)Cl(py)2] 5. Pyridine (225 ml, 2.8 mmol) was added to a solution containing [Rh2(m-Cl)2(CO)4] (135 mg, 0.35 mmol) and [Rh2(m-Cl)2(CxHy)4] where CxHy = C8H14 (250 mg, 0.35 mmol) or CxHy = C2H4 (135 mg, 0.35 mmol) (Rh :py = 1 :2) in CH2Cl2 (2 cm3) under a nitrogen atmosphere and addition of diethyl ether gave yellow crystals of the product which were filtered oV and dried under vacuum (370 mg, 87%) (Found: C, 41.0; H, 3.1; N, 8.5.C11H10ClN2ORh requires C, 40.7; H, 3.1; N, 8.6%). cis-[Rh(13CO)2Cl(solv)] 6 (solv 5 thf a or MeOH b). The complex [Rh2(m-Cl)2(13CO)4] (200 mg, 0.5 mmol) was dissolved in thf or MeOH (5 cm3) under a nitrogen atmosphere and stirred for 2 h to give cis-[Rh(13CO)2Cl(solv)]. The IR and variable-temperature 13C-{1H} NMR measurements were consistent with the formulation but attempts to isolate the product through either slow evaporation or layering with either diethyl ether or light petroleum always resulted in reformation of the starting material.1408 J.Chem. Soc., Dalton Trans., 1998, Pages 1403–1410 Scheme 4 Products resulting from the progressive addition of py to an equimolar mixture of [Rh2(m-Cl)2(CO)4] 1 and [Rh2(m-Cl)2(CxHy)4] (CxHy = C2H4 or C8H14) Rh Cl HyCx Cl Rh Rh Cl OC Cl Rh CO N N Rh Cl Cl Rh CO N Rh Cl Cl Rh CO N N N Rh Cl Cl N Rh Cl N Rh N N Rh Cl N N Rh Rh Cl Cl Rh C O OC CO CO OC OC OC OC CxHy + HyCx (C xHy = C2H4, a; C8H14 b) CxHy CxHy 11 (C xHy = C2H4) C2H4 H4C2 12 14 N2 – 2C2H4 2 py N2 HyCx 5 10 MeOH/N2 > 2 py 2 13 + N2 2py OC 2py N2 (C xHy = C8H14) OC CO OC OC cis-[Rh(CO)2(solv)2][ClO4] 7 (solv 5 thf a or MeOH b).Addition of AgClO4 (210 mg, 1.0 mmol) to a solution of [Rh2(m-Cl)2(CO)4] (200 mg, 1.0 mmol) in thf (or MeOH) (5 cm3) under a nitrogen atmosphere gave a yellow solution and a white precipitate (AgCl).The yellow solution was syringed oV and concentrated to 2 cm3. The IR and 13C NMR measurements were consistent with the formulation of the product as cis-[Rh(CO)2(solv)2][ClO4] but attempts to isolate it by layering the solution with either diethyl ether or light petroleum failed. cis-[Rh(CO)2(py)2][ClO4] 9. Method A. The salt AgClO4 (315 mg, 1.5 mmol) was added to a solution of [Rh2(m-Cl)2- (CO)4] (300 mg, 0.75 mmol) in thf (5 cm3) under a CO atmosphere to give a yellow solution with a white precipitate (AgCl).The yellow solution was filtered and concentrated to 2 cm3. Pyridine (240 ml, 3 mmol) was then added under a CO atmosphere followed by diethyl ether (5 cm3) which gave the product as a yellow precipitate; this was filtered oV and dried under vacuum (445 mg, 71%) (Found: C, 34.5; H, 2.4; N, 6.7. C12H10ClN2O6Rh requires C, 34.6; H, 2.4; N, 6.7%). Method B.The salt AgClO4 (275 mg, 1.3 mmol) was added to a thf solution (6 cm3) containing [Rh2(m-Cl)2(nbd)2] (300 mg, 0.65 mmol) and pyridine (210 ml, 2.6 mmol) under a nitrogen atmosphere to give a yellow solution with a white precipitate (AgCl). The yellow solution was concentrated under vacuum to give a yellow precipitate. The precipitate was washed with diethyl ether (2 × 3 cm3), redissolved in thf (5 cm3) and COJ. Chem. Soc., Dalton Trans., 1998, Pages 1403–1410 1409 Table 4 Crystal structure analysis, crystal data and experimental details for complexes 2 and 4 * Formula MA ppearance a/Å b/Å c/Å a/8 b/8 g/8 U/Å3 Z Dc/g cm23 F(000) m(Mo-Ka)/cm21 Crystal dimensions/mm Reflections measured h,k,l Ranges Unique reflections Rmerg Tmax, Tmin Observed reflections [I > 3s(I)] Number parameters refined RR 9 Final diVerence electron density (maximum, minimum)/e Å23 2 C7H5ClNO2Rh 273.48 Red plate crystal 9.784(6) 13.59(1) 7.446(5) 103.73(5) 93.45(6) 108.42(7) 903(1) 4 2.012 528 21.18 0.20 × 0.050 × 0.250 3386 0 to 12, 216 to 16, 29 to 9 3182 0.037 0.74, 1.0 1768 157 0.039 0.039 0.521, 20.54 4 C23H20Cl2N4O3Rh2 677.15 Yellow prism 11.023(1) 13.395(1) 9.012(3) 91.89(2) 94.05(2) 71.570(8) 1259.2(4) 2 1.786 668 15.36 0.20 × 0.10 × 0.20 4661 0 to 13, 216 to 16, 211 to 11 4409 0.013 0.88, 1.0 3765 307 0.022 0.028 0.37, 20.42 * Details in common: triclinic, space group P1� (no. 2). bubbled through the solution for 30 min.The product was isolated by layering the solution with diethyl ether (605 mg, 84%) (Found: C, 34.55; H, 2.4; N, 6.7. C12H10ClN2O6Rh requires C, 34.6; H, 2.4; N, 6.7%). [Rh(CO)(py)3][ClO4] 10. The salt AgClO4 (215 mg, 1.02 mmol) was added to a thf solution (10 cm3) containing [Rh2- (m-Cl)2(CO)4] (200 mg, 0.51 mmol) and pyridine (320 ml, 3.95 mmol) under a nitrogen atmosphere to give a reddish orange solution with a white precipitate. Addition of diethyl ether (20 cm3) to the thf solution gave a reddish orange precipitate which was filtered oV, washed with diethyl ether and dried in vacuum (338 mg, 69%) (Found: C, 42.2; H, 3.6; N, 9.3.C16H15ClN3O5Rh requires C, 41.1, H, 3.2; N, 9.0%). trans-[Rh2(Ï-Cl)2(CO)2(CxHy)2] 11 (CxHy 5 C2H4 a or C8H14 b). Dichloromethane (2 cm3) was added to a mixture of [Rh2- (m-Cl)2(CO)4] (135 mg, 0.35 mmol) and [Rh2(m-Cl)2(CxHy)4] (CxHy = C2H4, 135 mg, 0.35 mmol; C8H14, 250 mg, 0.35 mmol) under a nitrogen atmosphere and spectroscopic measurements confirmed that trans-[Rh2(m-Cl)2(CO)2(CxHy)2] was formed exclusively. No attempt was made to isolate these products which were used directly for further reactions and monitored using IR and NMR spectroscopy.[Rh2(Ï-Cl)2(CO)2(C2H4)2(py)2] 12. Pyridine (82 ml, 2.0 mmol) was added to a solution containing [Rh2(m-Cl)2(CO)4] (100 mg, 0.25 mmol) and [Rh2(m-Cl)2(CO)2(C2H4)4] (100 mg, 0.25 mmol) (Rh:py = 1 : 1) in CH2Cl2 (2 cm3) under a nitrogen atmosphere to give a clear yellow solution.The IR and 13C-{1H} NMR measurements (which showed the presence of co-ordinated C2H4) were entirely consistent with the formation of [Rh2- (m-Cl)2(CO)2(py)2(C2H4)2] (see Table 1). trans-[Rh2(Ï-CO)2Cl2(py)2] 13. Addition of pyridine (82 ml, 2.0 mmol) to a mixture containing [Rh2(m-Cl)2(CO)4] (100 mg, 0.25 mmol) and [Rh2(m-Cl)2(C2H4)4] (100 mg, 0.25 mmol) (Rh:py = 1 : 1) in CH2Cl2 (5 cm3) under a nitrogen atmosphere gave a clear yellow solution.When left under a reduced pressure of N2 this solution formed a violet precipitate of trans-[Rh2- (m-CO)2Cl2(py)2] which was filtered oV and dried under vacuum (181 mg, 72%) (Found: C, 29.4; H, 2.1; N, 5.5. C12H10Cl2N2- O2Rh2 requires C, 29.2; H, 2.4; N, 5.6%). trans-[Rh2(Ï-Cl)2(CO)2(py)2] 14. Pyridine (115 ml, 1.4 mmol) was added to a solution containing [Rh2(m-Cl)2(CO)4] (135 mg, 0.35 mmol) and [Rh2(m-Cl)2(C8H14)4] (250 mg, 0.35 mmol) (Rh:py = 1 : 1) in CH2Cl2 (2 cm3) under a nitrogen atmosphere.The IR, 13C-{1H} and 15N-{1H} INEPTRD (insensitive nuclei enhanced by polarisation transfer, refocussed and decoupled) NMR measurements were entirely consistent with the formation of trans-[Rh2(m-Cl)2(CO)2(py)2] (see Table 1). Crystallography Crystal data, data collection and processing details are given in Table 4. All data were recorded on a Rigaku AFC6S diVractometer at 2120 8C using graphite-monochromatised Mo-Ka radiation, l = 0.710 69 Å and 50 kW sealed-anode generator, scan width between 1.10 1 0.30 tan q for complex 2 and 0.89 1 0.30 tan q for 4, scan speed 48 min21 for both (2 rescans), 2qmax = 508.Three standard reflections were measured after every 150 scans; no significant decay was observed. An empirical absorption correction (y scans 20) was applied by the TEXSAN21 system. The unit cells were determined from diVractometer angles for 25 automatically centred reflections with 2q 7.29–18.768 for 2 and 38.3–44.138 for 4.Structure analysis and refinement. Direct methods by fullmatrix refinement on F. All non-hydrogen atoms were treated as anisotropic and hydrogen atoms placed in geometric ideal positions and assigned isotropic thermal parameters 20% greater than the B equivalent value of the atom to which they were bonded. The weighting scheme was based on counting statiics and included a factor (P = 0.03) to downweight the intense reflections. Plots of Sw(|Fo| 2 |Fc|)2 versus |Fo|, reflection order in data collection, (sin q)/l, and various classes of indices showed no unusual trends.Neutral atom scattering factors were taken from Cromer and Waber.22 Anomalous dispersion eVects were included in Fc. All calculations were performed using the TEXSAN21 software package and ORTEP23 was used to produce Figs. 1 and 2.1410 J. Chem. Soc., Dalton Trans., 1998, Pages 1403–1410 CCDC reference number 186/884. See http://www.rsc.org/suppdata/dt/1998/1403/ for crystallographic files in .cif format.Acknowledgements We thank the Leverhulme Foundation for the award of a Research Fellowship (to B. T. H.), EPSRC for a postdoctoral fellowship (to C. J.), the ORS for an award (to J. T. S.), Dr. D. Reed (University of Edinburgh) for the direct 103Rh NMR measurements, Dr. J. A. Iggo for helpful discussions on 15N NMR measurements and J. V. Barkley for the crystal structures. References 1 B. T. Heaton, C. Jacob, G. K. Monks, M. B. Hursthouse, I. Ghatak, R. G. Sommerville, W. Heggie, P. R. Page and I. Villax, J. Chem. Soc., Dalton Trans., 1996, 61 and refs. therein. 2 B. T. Heaton, J. A. Iggo, C. Jacob, J. Nadarajah, M. A. Fontaine, R. Messere and A. F. Noels, J. Chem. Soc., Dalton Trans., 1994, 2875 and refs. therein. 3 A. J. Pardey and P. C. Ford, J. Mol. Catal., 1989, 53, 247. 4 K. Kaneda, M. Hiraki, K. Sano, T. Imanaka and S. Teranishi, J. Mol. Catal., 1980, 9, 227. 5 G. Fachinetti, G. Focki and T. Funaioli, Inorg. Chem., 1994, 33, 1719. 6 W. Hieber, H. Heusinger and O. Vohler, Chem. Ber., 1957, 90, 2425. 7 P. S. Hall, G. E. Jackson, J. R. Moss, D. A. Thornton, P. F. M. Verhoeven and G. M. Watkins, Spectrosc. Lett., 1993, 26, 1247. 8 A. J. Pribula and R. S. Drago, J. Am. Chem. Soc., 1976, 98, 2784. 9 D. N. Lawson and G. Wilkinson, J. Chem. Soc., 1965, 1900. 10 R. Poilblanc, J. Organomet. Chem., 1975, 94, 241. 11 G. A. Morris and R. Freeman, J. Am. Chem. Soc., 1979, 101, 760. 12 G. Fachinetti, T. Funaioli and P. F. Zanazzi, J. Organomet. Chem., 1993, 460, C34. 13 F. Shafiq and R. Eisenberg, Inorg. Chem., 1993, 32, 3287. 14 J. P. Farr, M. M. Olmstead and A. L. Balch, J. Am. Chem. Soc., 1980, 102, 6654. 15 F. Shafiq and R. Eisenberg, Abstr. Papers Am. Chem. Soc. Meeting, 1992, 204, 427. 16 B. T. Heaton, C. Jacob, W. Heggie, P. R. Page and I. Villax, Magn. Reson. Chem., 1991, 29, S21. 17 R. Cramer, Inorg. Synth., 1974, 15, 14. 18 A. Ent and A. L. Onderdelinden, Inorg. Synth., 1973, 14, 93. 19 J. A. McCleverty and G. Wilkinson, Inorg. Synth., 1966, 8, 211. 20 A. C. T. North, D. C. Phillips and F. S. Mathews, Acta Crystallogr., Sect. A, 24, 351. 21 TEXSAN-TEXRAY, Structure Analysis Package, Molecular Structure Corporation, Houston, TX, 1989. 22 D. T. Cromer and J. T. Waber, International Tables for X-Ray Crystallography, Kynoch Press, Birmingham, 1974, vol. 4, Table 2.2A. 23 C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976. Received 15th October 1997; Paper 7/07435B

ISSN:1477-9226    

DOI:10.1039/a707435b

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

DALTON J. Chem. Soc., Dalton Trans., 1997, Pages 1411–1420 1411 Influence of anion on the solution and solid-state structures of some 1 : 2 adducts of silver(I) salts with 1,3-bis(diphenylphosphino)propane Dermawan AVandi,a Susan J. Berners-Price,*,†,b EVendy,a Peta J. Harvey,b Peter C. Healy,b Beate E. Ruch b and Allan H. White c a Jurusam Pendidikan Kimia, FPMIPA IKIP Malang, Jalan Surabaya 6, Malang 65145, Indonesia b Faculty of Science and Technology, Griffith University, Nathan, Brisbane, Queensland 4111, Australia c Department of Chemistry, University of Western Australia, Nedlands, Western Australia 6907, Australia Crystallization of 1 : 2 silver(I) halide/pseudo halide : 1,3-bis(diphenylphosphino)propane (dppp) mixtures from acetonitrile have resulted in the isolation of a novel series of neutral complexes, [AgX(dppp-P,P9)(dppp-P)] (X = Cl, Br, I or CN) containing co-ordinated anion and uni- and bi-dentate dppp ligands. In contrast, the thiocyanate and nitrate complexes precipitate as ionic [Ag(dppp-P,P9)2]X with unco-ordinated anion and bidentate phosphine ligands, a structural type previously found for other 1 : 2 silver(I) : diphosphine complexes. The complexes have been characterized by single-crystal X-ray structure determinations and solid-state 31P crosspolarization magic angle spinning (CP MAS) NMR spectroscopy.The salt [Ag(dppp-P,P9)2]SCN is obtained as crystals suitable for X-ray studies from pyridine, crystallizing as a sesqui-pyridine solvate in the monoclinic space group P21/c with a = 10.691(2), b = 24.75(2), c = 22.360(4) Å, b = 108.38(1)8.The neutral AgXP3 complexes (X = Cl, Br, I or CN) are isomorphous, crystallizing in the monoclinic space group C2/c, with a ª 21.8, b ª 10.3, c ª 45 Å, b ª 958. The solid-state 31P CP MAS spectra are consistent with the structural results; the tetrahedral SCN and NO3 complexes give a similar broad complex multiplet centred at d 26, whereas the spectra of the neutral AgXP3 complexes are interpretable as A2BMX spin systems with signals assignable to the non-co-ordinated (d 223), bidentate (d ca. 13) and unidentate (d ca. 3) phosphorus atoms. For the latter 1J(P]Ag) couplings are in the range 288 Hz (CN) to 367 Hz (Br). Solution 31P NMR studies on these complexes show that both neutral and ionic complexes exist in equilibrium in solution, with the position of the equilibrium dependent on the nature of the anion X. The potential significance to the antitumour activity of bis-chelated 1 : 2 silver(I)–diphosphine complexes is discussed.Like their gold(I) counterparts, certain tetrahedral bis-chelated 1 : 2 silver(I)–diphosphine complexes of the type [Ag(P]P)2]NO3 [where P]P is Ph2P(CH2)2PPh2 (dppe), cis-Ph2PCH]] CHPPh2 (dppey) or Et2P(CH2)2PEt2 (depe)] have been shown to exhibit antitumour activity against i.p P388 leukaemia in mice, as well as antifungal and modest antibacterial properties.1,2 Although the mechanism for the cytotoxicity is not known, tumour cell mitochondria are likely targets for these lipophilic cations 3 and indeed, the complex [Ag(eppe)2]NO3 [where eppe is Ph2P- (CH2)2PEt2] exhibits selective primary antimitochondrial activity in yeast.4 An understanding of the mechanisms responsible for the biological activity depends on a detailed knowledge of the different structural types that exist for silver(I) complexes of bidentate phosphines in the solid state and in solution.Tolazzi and coworkers have studied the thermodynamics of complexation of AgIClO4 with the bidentate phosphines Ph2P(CH2)nPPh2 (n = 1– 3) in dimethyl sulfoxide 5 and propylene carbonate.6 For n = 2 (dppe) and 3 (dppp) at 1 : 2 Ag:P]P ratios the only species present were ionic [Ag(P]P)2]1 complexes involving chelated diphosphine ligands. Similarly, in previous 31P NMR studies 7 in chloroform solutions, only the bis-chelated ionic complexes [Ag(P]P)2]NO3 were observed for the bidentate phosphines dppe, dppp, depe, eppe and dppey; these had greatly enhanced kinetic and thermodynamic stabilities with respect to similar AgP4 complexes containing monodentate phosphines, so that 109,107Ag–31P couplings were resolved in the 31P NMR spectra at ambient temperatures.† E-Mail: S.Berners-Price@sct.gu.edu.au Both ClO4 2 and NO3 2 are non-co-ordinating anions; more relevant to biological systems are complexes of the type that exist in the presence of co-ordinating anions, particularly chloride.Both [Ag(dppe)2]1 and [Ag(eppe)2]1 cations (as nitrate salts) are stable in the presence of an excess of Cl2 ions,1,4 indicating that chloride does not readily displace phosphorus from the AgI co-ordination sphere. Of great interest therefore is a novel class of compounds that we have isolated for 1 : 2 adducts of a number of silver(I) compounds, AgX, with dppp in the context of systematic studies of AgX: dppx type adducts.8 For X = Cl, Br, I or CN in the solid state the compounds exist as neutral complexes of type [AgX(dppp-P,P9)(dppp-P)] with both uni- and bi-dentate dppp ligands and co-ordinated anion.In contrast, the thiocyanate complex is not of this type but, rather, ionic [Ag(dppp)2]SCN. We report here the characterization of these compounds by single-crystal X-ray structure determinations and compare the structures that exist in the solid state to those in solution by comparison of results from solid state 31P cross-polarization magic angle spinning (CP MAS) and solution 31P NMR spectroscopy.We show that the extent of dissociation of the complexes in solution to form cationic [Ag(dppp)2]1 and/or other species depends critically on the nature of the anion X. Experimental Preparation of compounds [AgX(dppp-P,P9)(dppp-P)] (X = Cl, Br, I or CN). All compounds were prepared similarly. The compounds AgX (1.0 mmol) and dppp (0.85 g, 2.1 mmol) were dissolved with warm-1412 J.Chem. Soc., Dalton Trans., 1997, Pages 1411–1420 ing in acetonitrile (ca. 20 cm3). After filtration, the solutions were allowed to stand and cool, depositing colourless crystals of the product. X = Cl: m.p. 124–126 8C (Found: C, 66.8; H, 5.5. C54H52AgClP4 requires C, 67.0; H, 5.4%). X = Br: m.p. 147–149 8C (Found: C, 63.9; H, 5.2. C54H52AgBrP4 requires C, 64.05; H, 5.2%). X = I: m.p. 159–161 8C (Found: C, 61.4; H, 5.1. C54H52AgIP4 requires C, 61.2; H, 4.95%).X = CN: m.p. 115–117 8C (Found: C, 68.9; H, 5.4; N, 1.6. C55H52AgNP4 requires C, 68.9; H, 5.45; N, 1.45%). [Ag(dppp)2]SCN?1.5py (py = pyridine). With AgSCN the same procedure as for the halide and CN compounds yielded a white precipitate. This was dissolved in warm pyridine (ca. 5 cm3) giving a clear solution, which on cooling deposited colourless crystals, m.p. 95–97 8C (Found: C, 67.8; H, 5.3; N, 3.0; S, 3.0. C62.5H59.5AgN2.5P4S requires C, 67.65; H, 5.4; N, 3.2; S, 2.9%).[Ag(dppp)2]NO3. This was prepared essentially according to the literature method,7 by addition of AgNO3 (0.085 g, 0.5 mmol) in water (10 cm3) to a solution of dppp (0.42 g, 1 mmol) in acetone (20 cm3). Colourless microcrystalline material formed from the clear solution on standing. Crystallography Structure determinations. Unique room temperature diffractometer data sets (Enraf-Nonius CAD-4 instrument, T ª 295 K, monochromatic Mo-Ka radiation l = 0.710 73 Å) were measured to 2qmax = 508 yielding N independent reflections, No with I > 3s(I) being considered ‘observed’ and used in the fullmatrix least-squares refinements after Gaussian absorption corrections.Anisotropic thermal parameters were refined for the non-hydrogen atoms, (x,y,z,Uiso)H being included constrained at estimated values. Conventional residuals (on |F|), R and R9 are quoted, statistical reflection weights being derivative of s2(I) = s2(Idiff) 1 0.0004s4(Idiff). Neutral-atom complex scattering factors were employed, computation using the XTAL 3.2 program system implemented by S.R. Hall.9 Pertinent results are given in the figures and tables. Abnormal features/variations in procedures/comments for individual samples are recorded below (‘Variata’). In all figures 20% thermal ellipsoids are shown for the non-hydrogen atoms; hydrogen atoms, where included, have arbitrary radii of 0.1 Å. Atomic co-ordinates, thermal parameters, and bond lengths and angles have been deposited at the Cambridge Crystallographic Data Centre (CCDC). See Instructions for Authors, J.Chem. Soc., Dalton Trans., 1997, Issue 1. Any request to the CCDC for this material should quote the full literature citation and the reference number 186/412. Crystal data. The compounds [AgX(dppp)2] (X = Cl, Br, I or CN) are isomorphous, monoclinic, space group C2/c (C 6 2h, no. 15), Z = 8. X = Cl. C54H52AgClP4, M = 968.3, a = 21.731(9), b = 10.26(1), c = 44.61(2) Å, b = 94.44(4)8, U = 9917 Å3, Dc = 1.30 g cm23, F(000) = 4000, mMo = 6.2 cm21, crystal size 0.22 × 0.46 × 0.16 mm, A*min,max = 1.09, 1.13, N = 8718, No = 2911, R = 0.065, R9 = 0.069.X = Br. C54H52AgBrP4. M = 1012.7, a = 21.709(5), b = 10.289(5), c = 44.920(9) Å, b = 94.76(2)8, U = 9998 Å3, Dc = 1.35 g cm23, F(000) ª 4144, mMo = 13.6 cm21, crystal size 0.23 × 0.42 × 0.05 mm, Å*min,max = 1.06, 1.32, N = 7628, No = 2697, R = 0.067, R9 = 0.071. X = I. C54H52AgIP4, M = 1059.7, a = 21.77(1), b = 10.341(8), c = 45.35(2) Å, b = 95.51(4)8, U = 10 162 Å3, Dc = 1.39 g cm23, F(000) = 4288, mMo = 10.6 cm21, crystal size 0.20 × 0.42 × 0.16 mm, Å*min,max = 1.17, 1.24, N = 8945, No = 4514, R = 0.041, R9 = 0.041.X = CN. C55H52AgNP4, M = 958.8, a = 21.892(9), b = 10.257(9), c = 44.68(2) Å, b = 94.36(3)8, U = 10 003 Å3, Dc = 1.27 g cm23, F(000) = 3968, mMo = 5.7 cm21, crystal size 0.26 × 0.42 × 0.10 mm, Å*min,max = 1.05, 1.19, N = 7443, No = 3334, R = 0.065, R9 = 0.072.Variata. All compounds presented common problems associated with (a) the long c axis (possibly introducing some systematic error by way of uncompensated reflection overlap, despite the use of an extended counter arm), (b) rather weak data and (c) disorder in the terminal phosphorus of the unidentate dppp ligand, and, less well defined and unresolved, its neighbouring atoms; phosphorus populations x, (1 2 x) refined to x = 0.77(1), 0.84(1), 0.899(5), 0.85(1) for the four compounds.Intercomponent P ? ? ? P distances are 1.74(3), 1.76(4), 1.67(2) and 1.76(3) Å, respectively; the location of the second component (Fig. 2) corresponds more plausibly to that expected for inversion of the phosphorus [as in one of the phases of trimesitylphosphine 10 in which the two sites are 1.725(6) Å apart] rather than, for example, the full or partial occupancy of an associated oxygen site of any oxide impurity should it be present. No resonances for phosphine oxide were discernible in the solidstate CP MAS 31P NMR spectra of these complexes.w Scans were used for data measurement. [Ag(dppp)2]SCN?1.5py. C55H52AgNP4S?1.5C5H5N, M = 1109.5. Monoclinic, space group P21/c (C 5 2h, no. 14), a = 10.691(2), b = 24.75(2), c = 22.360(4) Å, b = 108.38(1)8, U = 5614 Å3, Dc (Z = 4) = 1.31 g cm23, F(000) = 2300, mMo = 5.5 cm21, crystal size 0.51 × 0.28 × 0.16 mm, Å*min,max = 1.09, 1.16, N = 9087, No = 5083, R = 0.052, R9 = 0.051. Variata. 2q/q Scan mode. Pyridine thermal motion is high; solvent(2) is disposed about an inversion centre with modelling of the nitrogen, tentatively assigned, as disordered. NMR Spectroscopy Solid-state CP MAS 31P NMR spectra were obtained at ambient temperature on a Varian UNITY-400 spectrometer at 161.92 MHz. Single contact times of 2 ms were used with a proton pulse width of 7.0 ms, a proton decoupling field of 62 kHz and a recycle delay time of 30 s. The samples were packed in Kel-F inserts within silicon nitride rotors and spun at a speed of 5 kHz at the magic angle.Between 60 and 200 free induction decays were collected and transformed with experimental line broadening values of 10–20 Hz. The 31P CP MAS twodimensional correlation (COSY) experiment was recorded using the pulse sequence 11 available in the Varian Solids User Library, and use of the Haberkorn–Ruben (hypercomplex) method for pure-phase quadrature detection in the F1 dimension. The contact time, 1H 908 pulse length and spin-rate were the same as those implemented in the one-dimensional experiment.Typically a total of 200–256 time increments were used in each of which 64 transients were added, with a 5 s recycle delay. Both dimensions were zero-filled to 1 K words and weighted with sine-bell apodization prior to Fourier transformation. Chemical shift data are referenced to 85% H3PO4 via an external sample of solid PPh3 (d 29.9). Solution 31P NMR spectra were recorded on the same instrument on samples dissolved in CDCl3 or CH2Cl2–10% CD2Cl2 (3 cm3) in 10 mm NMR tubes at variable temperatures (297–183 K).Spectra consisting of 8000 data points were acquired using an 8 ms (458) pulse, proton Waltz decoupling and a 1 s relaxation delay. A total of 256 scans were collected and the spectra processed with a 1–5 Hz line broadening. The chemical shifts were referenced to external 85% H3PO4 (d 0) measured at 295 K. Results and Discussion Crystal structures The results of the room-temperature single-crystal X-ray studies are consistent with the formulation of the complexes, in terms of stoichiometry and connectivity, as 1 : 2 AgX:dppp adducts (X = Cl, Br, I, CN or SCN); in all cases, one formulaJ. Chem.Soc., Dalton Trans., 1997, Pages 1411–1420 1413 Fig. 1 Projection of the [Ag(dppp)2]1 cation of the thiocyanate complex down its quasi-2 axis Table 1 Molecular core geometries (angles in 8) for [Ag(dppp)2]SCN?1.5py.Ligand angles (values for ligands a, b respectively) Ag]P(1)]C(1) P(1)]C(1)]C(12) C(1)]C(12)]C(2) Ag]P(1)]C(111) Ag]P(19)]C(121) C(1)]P(1)]C(111) C(1)]P(1)]C(121) C(111)]P(1)]C(121) P(2)]Ag]P(1)]C(1) Ag]P(1)]C(1)]C(12) P(1)]C(1)]C(12)]C(2) P(2)]Ag]P(1)]C(111) 2 P(2)]Ag]P(1)]C(121) Ag]P(1)]C(111)]C(112) Ag]P(1)]C(121)]C(122) 111.6(3), 110.4(2) 115.8(5), 116.2(5) 114.6(7), 115.6(6) 121.0(3), 121.5(2) 113.0(2), 114.9(2) 103.2(3), 102.4(3) 102.5(3), 101.8(4) 103.6(3), 103.3(3) 225.7(2), 231.7(3) 50.3(6), 54.9(6) 282.1(7), 279.7(8) 147.4(2), 2151.6(3) 89.1(3), 82.9(3) 81.7(7), 2105.1(6) 17.7(7), 2167.6(6) Ag]P(2)]C(2) P(2)]C(2)]C(12) Ag]P(2)]C(211) Ag]P(2)]C(221) C(2)]P(2)]C(211) C(2)]P(2)]C(221) C(211)]P(2)]C(221) P(1)]Ag]P(2)]C(2) Ag]P(2)]C(2)]C(12) P(2)]C(2)]C(12)]C(1) P(1)]Ag]P(2)]C(211) P(1)]Ag]P(2)]C(221) Ag]P(2)]C(211)]C(212) 2 Ag]P(2)]C(221)]C(222) 111.8(2), 110.6(2) 113.1(5), 116.4(5) 111.5(3), 121.3(2) 120.3(2), 113.5(3) 104.5(8), 103.2(3) 101.5(3), 102.8(3) 105.8(3), 103.3(3) 29.1(3), 32.6(3) 256.6(5), 257.2(6) 85.4(6), 81.3(7) 145.6(2), 153.4(3) 289.8(3), 282.4(3) 103.9(6), 160.9(5) 211.6(8), 55.5(6) unit comprises the asymmetric unit of the structure, with fourco- ordinate silver(I) atoms.The thiocyanate complex obtained from pyridine solution is ionic, [Ag(dppp)2]1(SCN)2 as a pyridine sesqui-solvate. The [Ag(dppp)2]1 cation represents the first bis[bidentate bis(diphenylphosphino)propane]silver(I) cation structurally characterized.The pyridine appears to have no unusual interactions with cation or anion, but simply occupies lattice voids, with high thermal motion, also evident on the anion and rendering insignificant the recorded geometries of these moieties. The cation has quasi-2, non-crystallographic symmetry, closely approximated by ring torsions (Table 1) and phenyl ring dispositions about one pole; about the other, the relation is more approximate (Fig.1). The bite angles of the two ligands are similar, while interligand P]Ag]P angles about the quasi-symmetry axis are appreciably different: 108.83(7) vs. 120.34(7)8 reflecting a lowering of the symmetry of the complex from D2d. A slight asymmetry in the Ag]P distances is observed for each ligand with differences of ª0.03 and ª0.01 Å for ligands a and b, respectively. The structure of this cation can be compared to the three other reported structures in which silver(I) is bis-chelated by two bidentate diphenylphosphine ligands:12–14 [Ag(dppe)2]NO3, [Ag(dppey)2][Ph3Sn(NO3)2] and [Ag(dppf)2]ClO4 [where dppf is 1,19-bis(diphenylphosphino)- ferrocene].Comparative data are presented in Table 2. For all of these structures, the [Ag(P]P)2] geometry is only slightly distorted from D2d symmetry with the angles 15 qx, qy, qz close to the expected values of 908 (Table 2). The mean Ag]P distances vary from 2.473(7) for [Ag(dppey)2][Ph3Sn(NO3)2] to 2.57(2) Å for [Ag(dppf)2]ClO4, with the value of 2.52(2) Å for the present dppp complex similar to that recorded for [Ag(dppe)2]NO3.These results can be compared also with data obtained for complexes containing the [Ag(PPh3)4]1 cation where for example, for the nitrate complex,16 the Ag]P bond lengths are 2.643(3) (×3) and 2.671(4) Å, with P]Ag]P angles close to the tetrahedral angle. In the Cl, Br, I and CN, complexes, the silver atom is also bonded to two dppp ligands within a discrete mononuclear species but, unlike the SCN complex, one of these is unidentate rather than bidentate, the fourth co-ordination site being occupied by the halide or cyanide group to give a neutral AgXP2P molecule (Fig. 2). The P]Ag]P angle within the bidentate ligand is the smallest angle within the co-ordination sphere,1414 J. Chem. Soc., Dalton Trans., 1997, Pages 1411–1420 Fig. 2 Projection of [AgCl(dppp-P,P9)(dppp-P)] (a) down and (b) normal to the Cl–Ag bond; the disordered P(2b) component is included in both and not noticeably different from the values observed in those of the cation arrays; the ring conformation has approximate m symmetry here, with phosphorus substituents equatorial and axial, the latter to the same side of the ligands (Table 3).The Ag]P (bidentate) bond lengths are also similar, but Ag]P (unidentate), in a more spacious disposition, is shorter. The X]Ag]P angles are similar in magnitude, the P (unidentate)] Ag]P (bidentate) angles being the largest.In all bidentate ligands, angles within the rings are quasi-tetrahedral at the phosphorus and larger at the carbon atoms. About the metal centre minor geometrical differences only are found between Cl, Br and I adducts; the geometry of the cyanide adduct deviates more substantially, Ag]P being elongated by ca. 0.02 Å for the bidentate and 0.035 Å for the unidentate ligand, with the X]Ag]P angle also enlarged, all indicative of increased Ag]X interaction in the cyanide (Table 3).The [AgX(PPh3)3] analogues, considered in comparison, exist in a wide variety of phases;12,17 Ag]P bond lengths in the chloride and bromide complexes usually lie in the range 2.54–2.60 Å; one of the ligands is incipiently dissociating in the iodide and cyanide, with one Ag]P bond 0.2 (X = I) to 0.5 Å (X = CN) longer than the other two. The Ag]X bond lengths in these complexes are similar to those found for the present series, differences being of the order of 2% or less.J.Chem. Soc., Dalton Trans., 1997, Pages 1411–1420 1415 Solid-state 31P NMR spectroscopy The 31P CP MAS solid-state spectra of the complexes [Ag- (dppp)2]X (X = NO3 or SCN) and [AgX(dppp-P,P9)(dppp-P)] (X = Cl, Br, I or CN) are shown in Fig. 3. The spectrum of the SCN complex, precipitated from acetonitrile solution, consists principally of a complex multiplet centred at d 26 [Fig. 3(b)]; the 31P chemical shift is in a similar region to that of [Ag(dppp)2]1 in solution (Fig. 5), but in the solid state the four P atoms are crystallographically distinct, resulting in magnetic inequivalence and hence a more complex multiplet pattern is expected. The spectrum of the nitrate complex [Fig. 3(a)] is almost identical to that of the SCN complex, showing that the multiplet pattern is independent of the anion. However, recrystallization of the SCN complex from pyridine resulted in small changes in the appearance of the 31P CP MAS multiplet (not shown), with further distinct peaks observable at d ca. 23 and 11, the intensities of which were dependent on the conditions of crystallization. This is suggestive of the presence of different phases of [Ag(dppp)2]1, possibly as a consequence of variation in lattice site occupancy by the pyridine solvent. The solid-state 31P CP MAS spectra of the four neutral [AgX(dppp-P,P9)(dppp-P)] complexes are similar [Fig. 3(c)– 3( f )], and distinct from those of the ionic bis-chelated complexes, consisting of three distinct sets of 31P resonances: a broad high-field singlet (d 223), a broad low-field doublet (d ca. 23) and an intermediate multiplet (d ca. 213) whose appearance is anion dependent. These spectral results are consistent with the structural data with the high-field singlet and low-field doublet assignable to the terminal and co-ordinated phosphorus atoms of the unidentate dppp ligand, respectively, and the complex multiplet at d 213 assignable to the phosphorus atoms of the chelated ligand. While the two phosphorus atoms of the chelate ligand are crystallographically distinct, examination of the AgXP3 core geometries (Table 3, Fig. 2) suggests that the immediate chemical environment is similar for both atoms. Making the assumption that the chemical shifts of both of these atoms are the same, the 31P NMR spectrum of each complex would be expected to consist of the A2BM part of two overlapping A2BMX spin systems which arise from spin–spin coupling to each of the two Ag spin ��� isotopes (107Ag, 51.82%; 109Ag, 48.18%).This multiplet pattern Table 2 Comparative core geometries (distances in Å, angles in 8) for tetrahedral [Ag(P]P)2]X complexes a dppeb dppeyc dpppd dppf e Ag]P(1a) Ag]P(2a) Ag]P(1b) Ag]P(2b) ·Ag]PÒ P(1a)]Ag]P(2a) P(1a)]Ag]P(1b) P(1a)]Ag]P(2b) P(2a)]Ag]P(1b) P(2a)]Ag]P(2b) P(1b)]Ag]P(2b) qx qy qz 2.488(3) 2.523(3) 2.527(3) 2.523(3) 2.51(2) 84.5(1) 129.5(1) 128.8(1) 116.0(1) 117.9(1) 83.8(1) 89.8 100.3 91.3 2.472(2) 2.476(2) 2.479(2) 2.463(2) 2.473(7) 84.1(1) 123.8(1) 127.8(1) 117.8(1) 124.2(1) 83.9(1) 85.7 94.0 91.5 2.527(2) 2.501(3) 2.526(2) 2.516(2) 2.52(2) 93.51(7) 120.34(7) 122.99(6) 120.65(6) 108.83(7) 92.39(7) 86.7 96.2 96.5 2.561(2) 2.584(2) 2.549(2) 2.602(2) 2.57(2) 105.71(4) 113.15(4) 117.84(4) 106.81(4) 114.54(4) 98.39(4) 86.0 94.6 91.4 a qz is generally similar to the dihedral angle between ligand planes usually reported for this type of structure. For molecules adopting D2d symmetry, qx = qy = qz = 90.08.Deviation of qz from 908 represents a ‘twisting’ of ligand b relative to ligand a and lowers the symmetry to D2. Values of qx and qy different from 908 represent ‘rocking’ or ‘waggling’ displacements of ligand b with respect to ligand a (ref. 15). b Ref. 12, X = NO3. c Ref. 13, X = Ph3Sn(NO3)2. d This work, X = SCN. In the anion, C]S,N are 1.76(1), 0.68(1) Å; S]C]N is 164(1) Å. e Ref. 14, X = ClO4. can be assigned unambiguously in the 31P solution spectrum of [AgBr(dppp-P,P9)(dppp-P)] (see below) and the solid-state spectra have the expected appearance of A2BM multiplets in which the line widths are too broad (ca. 130 Hz) to resolve the multiplet splittings. The A2BMX spin system is substantiated by two-dimensional 31P CP MAS COSY spectra, as shown for [AgI(dppp-P,P9)(dppp-P)] in Fig. 4. The weak cross-peaks between the multiplets at d 215 (PA2) and d 24 (PB) confirm that these resonances are part of the same spin system.No cross-peaks are observed between the high-field singlet (PM) and the PB multiplet, but the 4JPB]PM coupling is expected to be only very small and was not resolved in the 31P NMR solution spectrum of [AgBr(dppp-P,P9)(dppp-P)] (Table 5). The chemical shift of the high field singlet (PM) is insensitive to the nature of the anion X in the four neutral complexes (Fig. 3) (and similar to the chemical shift of free dppp), consistent with the expected behaviour for the non-co-ordinated phosphorus environment in the unidentate dppp ligand.The relative broadness of the peak is consistent with the observed structural disorder about this atom (see above). Although the A2BMX spin system is second order the 2J(P]P) coupling constant is expected to be of the order of 50–80 Hz (cf. ref. 18), as shown for [AgBr(dppp-P,P9)(dpppO-P)] (Table 5). As this value is smaller than the line widths in the solid-state spectra, the observedspacing of the low-field (PB) doublet corresponds approximately to the value of 1J(Ag]PB).These values are tabulated in Table 4. Little change is observed in passing from the chloride to the bromide complexes. However, the value for the iodide decreases by ca. 5% while that of the cyanide decreases by a further 17%. These trends are consistent with those observed for the analogous series of complexes [AgX- {P(C6H11)3}2] 19 and reflect the similar donating properties of the halogens to silver in these complexes and, as expected, stronger Ag]X interactions in the cyanide complex.As for Fig. 3 The 31P CP MAS spectra of [Ag(dppp)2]X, (a) X = NO3, (b) X = SCN and [AgX(dppp-P,P9)(dppp-P)] complexes, (c) X = Cl, (d) X = Br, (e) X = I and ( f ) X = CN. The neutral complexes are interpretable as A2BMX spin systems (see text). In (a) the peaks at d 0 to 5 are at present unassigned1416 J. Chem. Soc., Dalton Trans., 1997, Pages 1411–1420 Table 3 Molecular core geometries (distances in Å, angles in 8) for [AgX(dppp-P,P9)(dppp-P)] X Cl Br I CNa Ag]X Ag]P(1a) Ag]P(2a) Ag]P(1b) X]Ag]P(1a) X]Ag]P(2a) X]Ag]P(1b) P(1a)]Ag]P(2a) P(1a)]Ag]P(1b) P(2a)]Ag]P(1b) Ag]P(1a)]C(1a) Ag]P(2a)]C(2a) P(1a)]C(1a)]C(12a) P(2a)]C(2a)]C(12a) C(1a)]C(12a)]C(2a) Ag]P(1a)]C(111a) Ag]P(1a)]C(121a) Ag]P(2a)]C(211a) Ag]P(2a)]C(221a) Ag]P(1b)]C(111b) Ag]P(1b)]C(121b) C(1a)]P(1a)]C(111a) C(1a)]P(1a)]C(121a) C(2a)]P(2a)]C(211a) C(2a)]P(2a)]C(221a) C(1b)]P(1b)]C(111b) C(1b)]P(1b)]C(121b) C(2b)]P(2b)]C(211b) b C(2b)]P(2b)]C(221b) b C(111a)]P(1a)]C(121a) C(211a)]P(2a)]C(221a) C(111b)]P(1b)]C(121b) C(211b)]P(2b)]C(221b) b Ag]P(1a)]C(1a)]C(12a) Ag]P(2a)]C(2a)]C(12a) P(1a)]C(1a)]C(12a)]C(2a) P(2a)]C(2a)]C(12a)]C(1a) P(2a)]Ag]P(1a)]C(1a) P(1a)]Ag]P(2a)]C(2a) X]Ag]P(1a)]C(111a) X]Ag]P(1a)]C(121a) X]Ag]P(2a)]C(211a) X]Ag]P(2a)]C(221a) X]Ag]P(1b)]C(1b) X]Ag]P(1b)]C(111b) X]Ag]P(1b)]C(121b) Ag]P(1a)]C(111a)]C(112a) Ag]P(1a)]C(121a)]C(122a) Ag]P(2a)]C(211a)]C(212a) Ag]P(2a)]C(221a)]C(222a) Ag]P(1b)]C(111b)]C(112b) Ag]P(1b)]C(121b)]C(122b) 2.567(5) 2.508(5) 2.512(4) 2.438(5) 105.7(2) 103.3(2) 104.9(2) 93.1(1) 123.2(2) 124.1(1) 108.3(5) 106.7(5) 117(1) 114(1) 115(1) 118.0(7) 120.6(7) 117.7(5) 125.2(5) 115.5(6) 113.3(5) 105.8(8) 98.6(8) 98.7(7) 102.6(7) 102.4(7) 104.8(7) 107(1) 98(1) 103.2(9) 101.8(7) 103.6(7) 99(1) 58(1) 264(1) 279(2) 82(1) 238.6(5) 41.2(5) 2173.9(6) 246.1(8) 2175.4(5) 53.8(7) 54.3(6) 265.4(6) 175.3(6) 45(1) 33(2) 0(1) 297(1) 1(2) 60(1) 2.696(3) 2.511(5) 2.511(5) 2.445(5) 107.0(1) 102.7(1) 103.7(1) 93.4(2) 124.5(2) 123.1(2) 107.4(7) 106.5(5) 118(1) 114(2) 113(1) 117.1(7) 120.1(8) 117.8(6) 123.7(6) 114.8(6) 112.3(6) 105.7(9) 100.1(9) 99.5(8) 103.0(8) 102.1(9) 106.9(9) 106(1) 97(1) 104(1) 102.8(8) 104.8(9) 98(1) 59(1) 264(1) 281(2) 83(2) 237.7(7) 41.3(6) 2174.6(6) 246.5(8) 2177.6(6) 51.6(8) 55.5(7) 262.5(7) 177.9(7) 44(2) 35(2) 1(2) 298(2) 0(2) 62(2) 2.852(1) 2.494(2) 2.514(2) 2.437(3) 107.66(6) 101.66(5) 103.61(5) 93.68(7) 125.74(7) 121.88(1) 107.7(3) 106.5(2) 117.8(5) 114.3(5) 116.4(6) 116.5(3) 121.2(3) 119.1(2) 122.4(2) 115.0(3) 113.7(2) 105.0(4) 99.9(3) 100.4(3) 104.1(4) 102.4(3) 105.9(3) 103.1(4) 98.9(4) 104.2(4) 101.4(3) 103.4(3) 100.8(5) 56.3(6) 263.3(5) 277.6(8) 81.5(7) 237.8(3) 41.4(3) 2176.7(3) 248.2(3) 2179.9(3) 51.7(3) 56.9(3) 261.8(2) 179.4(3) 43.5(7) 38.5(8) 1.1(7) 299.9(6) 2.7(6) 61.1(6) 2.19(2) 2.532(4) 2.536(4) 2.473(4) 109.3(4) 108.5(4) 108.0(4) 91.2(1) 119.1(1) 119.6(1) 106.9(5) 105.5(4) 117.1(9) 114.1(9) 116(1) 121.0(6) 118.9(5) 121.9(4) 122.6(2) 114.5(5) 117.5(5) 104.3(7) 98.6(6) 98.8(6) 103.0(6) 102.1(6) 104.5(6) 104.6(9) 97.9(9) 104.0(7) 101.2(6) 103.2(6) 100(1) 60(1) 266.7(9) 278(1) 82(1) 242.6(5) 45.4(5) 2173.6(7) 242.8(7) 2176.5(7) 51.7(7) 51.8(7) 264.7(6) 173.9(7) 45(1) 28(1) 1(1) 295(1) 22(1) 61(1) a C]N 1.12(2) Å, Ag]C]N 179(1)8.b Major component. the tricyclohexylphosphine complexes, the observed changes in the scalar coupling constants are not reflected in any signifi- cant changes in the Ag]P bond lengths for the halides, other than the small increase noted for the cyanide complex.The Table 4 Solid-state CPMAS 31P NMR parameters for [AgX- (dppp-P,P9)(dppp-P)] complexes. Estimated error in coupling constants ±10 Hz d PA d PB d PM 1J(PB]Ag)/Hz* Cl Br I CN 212.8 213.6 215.1 212.3 21.7 22.7 24.1 23.8 223.3 223.4 223.0 223.4 363 367 348 288 * Estimate only, based on multiplet splitting (Fig. 3). 1J(PB]Ag) is the average of the 1J(109Ag]31P) and 1J(107Ag]31P) coupling constants, see text. high-field multiplet corresponds to the two phosphorus nuclei in the chelated dppp ligand. The appearance of this multiplet is dependent on the nature of X (Fig. 3). For X = I or CN it consists of a broad, asymmetric doublet while for the bromide and chloride complex the spectrum devolves to a pair of overlapping doublets. A possible interpretation is that for X = I or CN the chemical shifts of the two chelated P are essentially the same in which case the observed splittings (ca. 232 and 173 Hz, respectively) will reflect the average 1J(Ag]PA) coupling constant. For X = Br or Cl the greater complexity of the multiplets suggests significant differences in different chemical shifts for the two phosphorus atoms. For X = Br the multiplet has the appearance of two overlapped doublets with a splitting of ca. 250 Hz. This agrees well with the estimate of the average value for 1J(Ag]PA) obtained from the 31P solution spectrum of [AgBr- (dppp-P,P9)(dppp-P)] (Table 5).However, for the Cl complex itJ. Chem. Soc., Dalton Trans., 1997, Pages 1411–1420 1417 Table 5 Solution 31P NMR parameters for 1 :2 Ag:dppp complexes (estimated error in coupling constants ±1 Hz) 21J/Hz Compound Solvent T/K d PA d PB d PM PA–107,109Ag (average) PB–107,109Ag (average) [Ag(dppp)2]X X = SCNa [AgX(dppp-P,P9)(dppp-P)] X = Cl X = Br X = I X = CN (species I) X = CN (species II) [AgBr(dppp-P,P9)(dpppO-P)] f CDCl3 CDCl3 CDCl3 CD2Cl2–CH2Cl2 CD2Cl2–CH2Cl2 CD2Cl2–CH2Cl2 CDCl3 223 253 223 233 183 183 223 24.7 27.9 27.4 d 211.8 212.4 213.0 e 29.1 b 23.3 d b 23.5 22.7 e 1.8 218.3 218.4 218.6 220.8 219.6 35.0 221, 254 (237) (250) c 239, 276 (257) d (250) c (214) c e 239, 276 (257) b 348, 402 (375) d b (245) c e 348, 402 (375) a The [Ag(dppp)2]1 cation is observed also as a dissociation product in solutions of [AgX(dppp-P,P9)(dppp-P)] (X = Cl, Br, I or CN) (see text).In CDCl3 d has a temperature dependence of 20.01 ppm K21. In CD2Cl2–CH2Cl2 d is 25.3 at 233 K. b Overlapped by major doublet of doublets for [Ag(dppp)2]1. c Couplings not fully resolved at this temperature. d The PA and PB multiplets are partially obscured by the [Ag(dppp)2]1 resonance at d 24.7 and also for PA the [AgBr(dppp-P,P9)(dpppO-P)] multiplet at d 29.1 [see Fig. 5(e)]. The chemical shifts are estimated based on the assumption that the coupling constants and the multiplet splitting patterns are identical to the analogous multiplets in [AgBr(dppp-P,P9)(dpppO-P)]. The spectrum simulated with these values gives a reasonable fit to the observed spectrum [Fig. 5(e)]. e Cannot be measured accurately due to overlap with compound I see Fig. 6. f 2J(Pa]PB) = 58 Hz, 4J(PB]PM) = 0 Hz, 6J(PA]PM) = 0 Hz; see simulated spectrum Fig. 5(e). is not possible to obtain an estimate of the value of 1J(Ag]PA) by inspection.Solution 31P NMR studies The results of the structural and solid-state NMR studies described above show that both cationic [Ag(dppp)2]1 and neutral [AgX(dppp-P,P9)(dppp-P)] complexes can be isolated in the solid state with the formation of each dependent on the anion. To investigate which of these structures is favoured in solution, or whether an equilibrium exists between the two (Scheme 1), we recorded variable-temperature solution 31P NMR spectra of the complexes in CDCl3 and/or CD2Cl2– CH2Cl2 solutions.Representative spectra are shown in Fig. 5. For [Ag(dppp)2]SCN in CDCl3 a broad (Dn2� 1 = 350 Hz) 31P resonance (d 25.9) was observed at 297 K which resolved below 273 K into two overlapping doublets (d 25.1) characteristic of the [Ag(dppp)2]1 cation.7 The 1J(107,109Ag]31P) coupling constants (221 and 254 Hz) are the same as reported previously 7 for [Ag(dppp)2]NO3 ahough for the nitrate complex the spin–spin Fig. 4 The two-dimensional 31P CP MAS COSY solid-state NMR spectrum of [AgI(dppp-P,P9)(dppp-P)] couplings are resolved at 295 K.This is consistent with the relative donor strengths of the two anions, as the more weakly co-ordinating nitrate group would be less expected to facilitate rupture of the Ag]P bonds than the thiocyanate. The 31P solution spectra obtained from dissolution of the neutral [AgX(dppp-P,P9)(dppp-P)] complexes were found to be very dependent on the nature of X. For X = Cl in CDCl3, the spectra exhibit a similar behaviour to the SCN complex.A broad (Dn2� 1 = 200 Hz) resonance (d 26.4) at 297 K resolved at 253 K into the two overlapping doublets for [Ag(dppp)2]1 [Fig. 5(a)]. However, unlike the SCN spectra, additional very minor signals were also just visible at this temperature. These had the expected multiplet pattern for [AgCl(dppp-P,P9)- (dppp-P)], with the assumption that the PB multiplet was obscured by the major [Ag(dppp)2]1 resonance (Table 5). The total integrated intensity of these minor peaks was ca. 2% of the intensity of the [Ag(dppp)2]1 31P signals.This result shows that for the chloride complex in CDCl3 solution, the equilibrium (Scheme 1) exists, but that the equilibrium constant lies well to the left, favouring displacement of the anion by the phosphine ligand. An explanation for the exclusive formation of the neutral species in the solid state is that while the cation is the more stable species in solution, the neutral species is more stable in the solid state and preferentially forms under the recrystallization conditions necessary for the preparation of X-ray quality crystals.For X = Br in CDCl3 the 31P NMR spectrum at 297 K consisted of a broad singlet (d 27.5, Dn2� 1 = 200 Hz). At 253 K this had partially resolved into the pair of doublets indicative of [Ag(dppp)2]1, together with a range of minor broad peaks between d 16 and 210. All couplings were fully resolved on cooling the sample to 223 K [Fig. 5(b) and 5(e)]. The minor peaks arise from two distinct pairs of A2BMX spin systems (for each of the 109Ag and 107Ag isotopes) which are assignable to Scheme 11418 J. Chem. Soc., Dalton Trans., 1997, Pages 1411–1420 Fig. 5 Solution 31P NMR spectra of [Ag(dppp)2X] complexes in CDCl3 at 223 K. (a) X = Cl (253 K), (b) X = Br, (c) X = I and (d) X = CN. (e) An expansion of spectrum (b) together with the simulated spectrum analysed as two distinct pairs of overlapping A2BMX spin systems for each of the 109Ag and 107Ag isotopes.Peaks labelled PA2, PB and PM correspond to [AgBr(dppp-P,P9)(dppp-P)] and P9A2 and P9B to the oxidized species [AgBr(dppp-P,P9)(dpppO-P)]; the P9M peak (not shown) is at d 135 (see text). The derived chemical shifts and couplings constants for [AgBr(dppp- P,P9)(dpppO-P)] are given in Table 5. As the PA2 and PB multiplets for [AgBr(dppp-P,P9)(dppp-P)] are partially obscured by the [Ag(dppp)2]1 resonance, the simulated spectrum was obtained by estimating the chemical shifts of the two multiplets and assuming all coupling constants were the same as for the oxidized species.The peak labelled u is unidentified [AgBr(dppp-P,P9)(dppp-P)] and, as well, its oxidation product [AgBr(dppp-P,P9)(dpppO-P)] in which the non-co-ordinated P in the unidentate dppp ligand has been oxidized. There was no evidence for oxidized product in the 31P CP MAS spectrum of [AgBr(dppp-P,P9)(dppp-P)], suggesting that oxidation had occurred in solution in the period between preparation of the samples and acquisition of the spectra.Fig. 5(e) shows a comparison of the observed 31P spectrum at 223 K, together with a simulated spectrum of two overlapped A2BMX spin systems. From this we deduced that the well resolved multiplets [labelled P9B and P9A2, Fig. 5(e)] are due to the chelated (PA2) and unidentate (PB) phosphorus atoms in the oxidized species since the intensity of the signal at d 218.4 is too low to correspond to the PM resonance in the same spin system.A signal with the correct intensity for the oxidized P in [AgBr(dppp-P,P9)(dpppO-P)] was present at d 135.0. The A2 and B multiplets of [AgBr- (dppp-P,P9)(dppp-P)] are partially obscured by the [Ag- (dppp)2]1 31P signal, but are just visible in Fig. 5(e). These have the correct intensity to belong to the same pair of A2BMX spin systems as the PM signal at d 218.4, as shown in the simulated spectrum.The chemical shifts and coupling constants derived from the simulated spectra are listed in Table 5. From peak integrals we estimate that the ring-opened species account for ca. 7% of the species present {[AgBr(dppp-P,P9)(dppp-P)] 2% and [AgBr(dppp-P,P9)(dpppO-P)] 5%}. This is slightly greater than for the chloride complex, reflecting a relative donor strength Br2 > Cl2. However direct comparison is difficult since the equilibrium (Scheme 1) becomes irreversible once the unco-ordinated P is oxidized.When the sample of [AgBr(dppp- P,P9)(dppp-P)] was dissolved in CD2Cl2–CH2Cl2 the bischelated complex was the only species observed in the 31P solution spectrum at 23 K. For X = I in CDCl3 the compound was not completely soluble at 297 K. However, a single 31P resonance was observed which was sharper (Dn2� 1 = 130 Hz) and more shielded (d 29.2) than for the SCN, Cl and Br complexes indicating the occurrence of equilibria between different types of AgI phosphine species.As the solution was cooled the resonance broadened and then at 253 K began to resolve into three set of peaks: two broad doublets at d 20.4 and 211.2, 1J(Ag]P) (average) ca. 380 and 243 Hz, respectively and a very broad resonance at d 26.8. At 223 K these resonances had been replaced by a pair of doublets centred at d 210.8 [1J(Ag]P) ca. 249 Hz, Fig. 5(c)] and a broadened doublet at d 24.7 characteristic of [Ag(dppp)2]1. The 107,109Ag]31P couplings were not resolved above the freezing point of the solvent.No peaks were visible that were assignable to the ring-opened species [AgI(dppp-P,P9)(dppp-P)] but broadened resonances were observed in the region d 0 to 24, indicative of additional species involved in exchange equilibria. In an attempt to identify the species present in solution we recorded 31P NMR spectra of [AgI(dppp-P,P9)(dppp-P)] in CH2Cl2–CD2Cl2 since this allowed spectra to be obtained at lower temperatures. However, different behaviour was observed in this solvent and at 243 K the spectrum resembled that of the chloride complex with the two overlapping doublets characteristic of the [Ag(dppp)2]1 cation and additional minor signals with the expected multiplet pattern for [AgI(dppp-P,P9)(dppp-P)] (Table 5).There was a significantly greater proportion of the ring-opened species (ca. 16% of the total products, based on peak integrals) compared to the BrJ. Chem. Soc., Dalton Trans., 1997, Pages 1411–1420 1419 and Cl complexes consistent with the greater donor strength of I2.The 31P NMR solution spectrum of the cyanide complex in CDCl3 at 297 K was similar to the iodide complex with a single relatively sharp peak (Dn2� 1 = 140 Hz) at d 210.3. On cooling to 263 K, this resolved into a minor broad signal (d 25.5) and a major signal (d 211.1). At 233 K the former had resolved into the typical doublet for [Ag(dppp)2]1 (d 24.9) and a major very broad peak (d 212.1).With further cooling the latter separated into three broad resonances (d 23, 211 and 218) [Fig. 5(d)]. The [Ag(dppp)2]1 ion accounted for only ca. 9% of the total products in solution, based on peak integrals. The temperaturedependent behaviour is consistent with the equilibrium (Scheme 1), in which there is little dissociation of the anion, so that the major species is [AgCN(dppp-P,P9)(dppp-P)]. However, the broadness of the peaks at 223 K is indicative of the occurrence of an additional equilibrium at low temperature.To investigate this further we recorded 31P solution spectra of the complex in CH2Cl2–CD2Cl2 (Fig. 6). The rate of exchange was slower in this solvent so that at ambient temperature two broad resonances were observed (d ca. 26.3 and 214.5). On cooling the low-field resonance resolved into the [Ag(dppp)2]1 pair of doublets with spin–spin couully resolved at 233 K. This species accounted for ca. 38% of the total products at 273 K, but on cooling the intensity gradually decreased until the resonance disappeared altogether just below 193 K.The high-field resonance sharpened initially, then below 273 K broadened and shifted to high frequency. Below 233 K it began to resolve into three broad resonances (d ca. 23, 213 and 219) and as the solution was cooled further these resolved into distinct sets of multiplets. Although the spin–spin couplings were not fully resolved above the freezing point of the solvent the spectrum at 183 K is interpretable as two distinct (but overlapped) A2BMX spin systems.This is seen most clearly for the PM resonance which shows two distinct singlets (d 220.8 Fig. 6 Solution 31P NMR spectra of [AgCN(dppp-P,P9)(dppp-P)] (4 mmol dm23) in CH2Cl2–30% CD2Cl2 at 297, 233, 203 and 183 K. The spectrum at 183 K is assignable as two independent A2BMX spin systems which may arise from two different conformational structures (I and II) (see text and Table 5) and 219.6) with an intensity ratio of 3 : 1; multiplets with the correct intensity ratio for the second A2BMX spin system are visible, although partially overlapped with the major PA2 and PB multiplets.The chemical shifts and coupling constants for the two A2BMX spin systems are listed in Table 5. Although the 1J(PB]Ag) and 1J(PB]Ag) coupling constants of the major species are slightly larger than estimated from the 31P CP MAS spectrum of [AgCN(dppp-P,P9)(dppp-P)] the chemical shifts for the three non-equivalent P environments are similar to those observed in the solid state and it seems reasonable that each of the species observed in solution is [AgCN- (dppp-P,P9)(dppp-P)] with, however, the substituents on the dppp ligands frozen into two distinct conformational structures.Conclusion Structural dislocation is observed in complexes of 1 : 2 AgX: dppp stoichiometry which can be isolated in both cationic [Ag(dppp)2]X and neutral [AgX(dppp-P,P9)(dppp-P)] forms in the solid state, depending on the anion X.To date, the neutral form only has been isolated from acetonitrile for X = Cl, Br, I or CN, whereas for SCN2 and NO3 2 the cationic complex only is isolated. In solution, the two forms exist in an equilibrium involving displacement of the anion by the uncoordinated phosphine ligand. For Cl2 the cationic form is favoured almost exclusively, consistent with the results of previous studies 1,4 on the related complexes [Ag(dppe)2]NO3 and [Ag(eppe)2]NO3 which showed that Cl2 does not readily displace P from the Ag co-ordination sphere.For the bromide complex similar behaviour was observed although the observation of significant amounts of ring-opened phosphine oxide complex [AgBr(dppp-P,P9)(dpppO-P)] in chloroform solution highlights the ease of opening of the chelated diphosphine, and subsequent oxidation, in the presence of the co-ordinating anion. For X = I, a greater proportion of the neutral complex exists in solution, consistent with the increased donor strength of I2, and competing dissociative equilibria are also evident in CDCl3 solutions. For X = CN, the neutral complex is favoured, with little formation of the cationic complex.Since the antitumour properties of metal diphosphine complexes may be related to the uptake of lipophilic cations into mitochondria, studies of this type are of fundamental importance to a meaningful interpretation of structure–activity relationships to establish whether the bis-chelated cationic species [Ag(dppp)2]1 will be the major species present in vivo for the different complexes of 1 :2 AgX: dppp stoichiometry.Acknowledgements We acknowledge support of this work by the Australian Research Council and the Australian National Health and Medical Research Council (R. Douglas Wright Award to S. J. B.-P.). We thank Dr. David Rice (Varian, Palo Alto) for assistance with the two-dimensional CP MAS COSY pulse sequence. References 1 S. J. Berners-Price, R. K. Johnson, A. J. Giovenella, L. F. Faucette, C. K. Mirabelli and P. J. Sadler, J. Inorg. Biochem., 1988, 33, 285. 2 S. J. Berners-Price and P. J. Sadler, Struct. Bonding (Berlin), 1988, 70, 28. 3 S. J. Berners-Price and P. J. Sadler, Coord. Chem. Rev., 1996, 151, 1. 4 S. J. Berners-Price, D. C. Collier, M. A. Mazid, P. J. Sadler, R. E. Sue and D. Wilkie, Metal-Based Drugs, 1995, 2, 111. 5 P. Di Bernardo, G. Dolcetti, R. Portanova, M. Tolazzi, G. Tomat and P. Zanonato, Inorg. Chem., 1990, 29, 2859. 6 A. Del Zotto, P. Di Bernardo, M. Tolazzi, G. Tomat and P. Zanonato, J. Chem. Soc., Dalton Trans., 1993, 3009.1420 J. Chem. Soc., Dalton Trans., 1997, Pages 1411–1420 7 S. J. Berners-Price, C. Brevard, A. Pagelot and P. J. Sadler, Inorg. Chem., 1985, 24, 4278. 8 Effendy and A. H. White, unpublished work. 9 S. R. Hall, H. D. Flack and J. M. Stewart, The XTAL 3.2 Reference Manual, Universities of Western Australia, Geneva and Maryland, 1992. 10 F. J. Blount, D. Camp, R. D. Hart, P. C. Healy, B. W. Skelton and A. H. White, Aust. J. Chem., 1994, 47, 1631. 11 G. Wu and R. E. Wasylishen, Organometallics, 1992, 11, 3242. 12 C. S. W. Harker and E. R. T. Tiekink, J. Coord. Chem., 1990, 21, 287. 13 C. Franzoni, G. Pelizzi, G. Predieri, P. Tarasconi and C. Pelizzi, Inorg. Chim. Acta, 1988, 150, 279. 14 M. C. Gimeno, P. G. Jones, A. Laguna and C. Sarroca, J. Chem. Soc., Dalton Trans., 1995, 1473. 15 J. F. Dobson, B. E. Green, P. C. Healy, C. H. L. Kennard, C. Pakawatchai and A. H. White, Aust. J. Chem., 1984, 37, 649. 16 P. F. Barron, J. C. Dyason, P. C. Healy, L. M. Engelhardt, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1986, 1965. 17 L. M. Engelhardt, P. C. Healy, V. A. Patrick and A. H. White, Aust. J. Chem., 1987, 40, 1873. 18 S. Attar, N. W. Alcock, G. A. Bowmaker, J. S. Frye, W. H. Bearden and J. H. Nelson, Inorg. Chem., 1991, 30, 4166. 19 G. A. Bowmaker, Effendy, P. J. Harvey, P. C. Healy, B. W. Skelton and A. H. White, J. Chem. Soc., Dalton Trans., 1996, 2449. Received 12th November 1996; Paper 6/07692K

ISSN:1477-9226    

DOI:10.1039/a607692k

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

DALTON J. Chem. Soc., Dalton Trans., 1997, Pages 1421–1427 1421 Syntheses and characterisation of rhenium-(III) and -(V) complexes containing aminophosphine ligands. Crystal structures of [ReVOCl2(OMe)L] and [ReIIICl3L(PPh3)], L = o-(diphenylphosphino)-N,N9-dimethylaniline Francesco Tisato,*,†,a Fiorenzo Refosco,a Cristina Bolzati,a Aldo Cagnolini,a Stefano Gatto b and Giuliano Bandoli b a ICTIMA-CNR, C.so Stati Uniti, 4-35020 Padova, Italy b Dipartimento di Scienze Farmaceutiche, Università di Padova, via F.Marzolo, 5-35131 Padova, Italy Substitution reactions of labile rhenium(V) precursors with o-(diphenylphosphino)-N,N9-dimethylaniline, L, gave a series of monooxorhenium(V) complexes of general formula [ReOX2Y(L)] (X = Y = Cl 1; X = Cl, Y = OMe 2) and [ReOX2(OEt)L] (X = Cl 3, Br 4 or I 5). These complexes, which contain only one aminophosphine chelate, have a distorted-octahedral geometry, as evidenced by a crystal structure determination of 2. Reactions conducted in basic media, or treatment of 2 in the presence of the weakly co-ordinating trifluoromethanesulfonic acid, gave the bis(chelate) cationic complexes [ReO2L2]1 6 and [ReCl2L2]1 7, respectively. Similar ligand-exchange reactions between the lower-oxidation-state precursor [ReIIICl3(MeCN)(PPh3)2] and L or L9 [o-(diphenylphosphino)aniline] still produced mono-substituted aminophosphine complexes, i.e.[ReIIICl3L(PPh3)] 8 or [ReIIICl3L9(PPh3)] 9 in which there is a meridional arrangement of chloride ligands along with a cis-phosphorus co-ordination.The crystal structure of 8 was determined. On the contrary, treatment of [ReIIIX3(MeCN)(PPh3)2] (X = Cl or Br) with L9 in the presence of co-ordinating hydrohalogenic acid gave bis-substituted cationic complexes [ReX2L92]1 (X = Cl 10 or Br 11). In spite of the paramagnetism of the low-spin d4 ion, all of the rhenium(III) complexes can be conveniently characterised in solution by proton NMR spectroscopy.Bidentate chelates comprising a ‘soft’ tertiary phosphine and a ‘hard’ anchoring donor such as an amine nitrogen or a carboxylate oxygen are of interest for the synthesis of metal complexes utilised for various purposes, e.g. in catalysis 1,2 and pharmaceutical development.3 In both cases the application of this class of compounds is based on the hemilability of this chelate entity, which may open reversibly to accommodate suitable incoming groups. Chelating anionic P,O ligands are successfully applied, for example, in the nickel-catalysed oligomerisation of ethylene in the Shell higher olefin process (SHOP).4 The reactivity of o-(diphenylphosphino)-N,N9-dimethylaniline, abbreviated L, toward Group VIII metals has been investigated since 1965.5 Complexes of NiII, CoII, PdII and PtII have both the diphenylphosphino and dimethylamino arms of the chelate co-ordinated and are of general formula [MX2L].5–7 On this basis, it has been proposed that L is a useful ligand in homogeneous catalysis being able (i) to stabilise low-oxidationstate complexes through phosphorus co-ordination, (ii) to confer a high nucleophilicity upon the complex through nitrogen co-ordination and (iii) to generate co-ordinatively unsaturated complexes.8 The co-ordination of other Group VIII metals such as RhI,8 IrI 9 and RuII 10 with L has also been investigated and a number of crystal structures determined.9–11 In addition, the five-co-ordinate precursor [RuCl2(PPh3)L] binds H2S reversibly to generate the co-ordinatively saturated complex [RuCl2- (PPh3)L(SH2)].12 Very recently, chelate ring opening in bis(aminophosphine) complexes of PtII has been achieved by appropriate choice of the substituents on N and P via a selective and reversible binding to the DNA base guanine under biologically relevant conditions of pH and chloride concentration.3 Thus, while the complex containing the primary amino–tertiary phosphine ligand † E-mail: tisato@ictr04.ictr.pd.cnr.it H2N(CH2)2PPh2 has a bis(chelate) structure in solution and in the solid state, the related complex containing the tertiary amino–tertiary phosphine Me2N(CH2)2PPh2, based on solutionstate 31P and 195Pt NMR signals, exhibits a pattern consistent with an equilibrium between the bis(chelate) and cis- [PtCl{Me2N(CH2)2PPh2-N,P}{Me2N(CH2)2PPh2-P}]Cl species.Preliminary cytotoxicity tests on these bis-chelated and ring-opened complexes provide interesting results when compared to those of [PtCl2(NH3)2] (cisplatin).3,13 This paper reports on the synthesis and characterisation of some complexes of ReIII and ReV containing L.The coordination of only one chelate in a neutral compound appears to be preferred, as demonstrated by the molecular structures of two representative complexes: [ReVOCl2(OMe)L] and [ReIIICl3L( PPh3)]. Nevertheless, the bis-substituted cationic complexes, [ReVO2L2]1 and [ReIIICl2L2]1, have also been obtained. Other rhenium complexes containing the primary amino– tertiary phosphine o-(diphenylphosphino)aniline (L9) have been prepared for comparison purposes.This study is a continuation of our investigation on the synthesis and chemistry of complexes of MIII and MV (M = Tc or Re) of unsymmetrical phosphinoamine ligands 14–16 with the aim of designing prototype molecules useful for the development of radiopharmaceuticals based on the 99mTc and 186/188Re isotopes.17 We have shown that L9 affords a series of bis-substituted complexes in which the nitrogen donor can behave as an amino, amido or imido group depending on the reaction conditions employed.14–16 Although the versatility of this chelate is challenging from the chemical and stereochemical points of view, the possible application of this class of compounds as potential radiopharmaceuticals is precluded by the high number of species produced.A drastic reduction in the species involved has been achieved by enhancing the chelating ability of a bidentate aminophosphine ligand, e.g.by linking the amine nitrogens of two L9 with a1422 J. Chem. Soc., Dalton Trans., 1997, Pages 1421–1427 propylene chain, giving a tetradentate P2N2 macrocycle,18 or by functionalising the primary amino group of L9 to obtain a tertiary amino–tertiary phosphine chelate. Results and Discussion Synthesis The bidentate phosphinoamine L reacts with [ReOCl4]2 in methanol at room temperature to give the light blue monooxorhenium( V) derivative [ReOCl3L] 1, which is converted into violet alkoxo species [ReOCl2(OR)L] (R = Me 2 or Et 3) in the appropriate refluxing alcohol.In addition, by treating the pentavalent rhenium precursor [ReOX2(OEt)(PPh3)2] (X = Cl, Br or I) with L in refluxing ethanol a series of compounds of general formula [ReOX2(OEt)L] (X = Cl 3, Br 4 or I 5) is produced. Conversely, basic media, e.g. sodium ethoxide, induce transformation to the bis-substituted dioxo species [ReO2L2]1 6.Another bis-substituted yellow species, namely [ReCl2L2]1 7, is produced by dissolving 2 in dichloromethane–methanol solutions in the presence of water and an excess of trifluoromethanesulfonic acid. Reaction of the substitution labile rhenium(III) complex [ReCl3(MeCN)(PPh3)2] with an excess of L in refluxing methanol does not lead to the expected bissubstituted rhenium(III) species just described, but to the neutral monosubstituted orange complex mer-[ReCl3L(PPh3)] 8.The isostructural rhenium(III) complex mer-[ReCl3L9(PPh3)] 9 is obtained in a similar way, whereas the cations [ReX2L92]1 (X = Cl 10 or Br 11) are prepared by treatment of [ReCl3- (MeCN)(PPh3)2] with 2 equivalents of L9 in the presence of an excess of the appropriate concentrated hydrohalogenic acid. Characterisation Monooxorhenium(V) complexes prepared in this work have been characterised by (i) elemental analyses, which are in agreement with the proposed formulations, (ii) conductivity measurements, which show values consistent with uncharged complexes, (iii) spectroscopic measurements, i.e.IR, NMR, positive-ion FAB mass spectrometry and the crystal structure determination of a prototypical complex [ReOCl2(OMe)L] 2. The IR spectra of complexes 1–5 and [ReOCl3L9] 12 show a very strong absorption in the 915–992 cm21 range, which is assigned to the n(Re]] O) stretching vibration. A remarkable bathochromic shift of this vibration is observed on going from the chloro (982–992 cm21) to the corresponding alkoxo compounds (915–940 cm21) (trans to the oxo moiety; see also NMR).In addition, a number of absorptions characteristic of the phenylphosphine moiety of L is observed. No uncoordinated NMe2 groups are present in the molecules, as indicated by the absence of the band at 2780 cm21 which is typical of free NMe2 groups.19 The IR spectrum of the dioxo derivative 6 exhibits the asymmetric O]] Re]] O stretching vibration at 821 cm21, whereas the 1 : 1 electrolytic nature of this complex is established by conductivity measurement. The formulation of the rhenium(V) complexes was further assessed by the positive ion FAB mass spectra which show the molecular ion peaks together with several fragment ions corresponding to loss of chloride and/or alkoxo groups, as summarised in Table 1.Proton NMR spectra establish that both the phosphino and the amino arms are co-ordinated also in solution.Thus, the signal due to the N-methyl resonance in free L is split into two singlets, upon co-ordination, the N-methyl groups being fixed in two magnetically inequivalent environments8 (see below). As seen in Table 1, the N-methyl resonances are significantly shifted downfield upon co-ordination, whereas the aliphatic portion of the alkoxo group is shielded. This contrasting effect, which is the result of a combination of the electron density located along the Re]] O axis and of the acidity of the Re]] O31 core,20,21 allows the assignment of stereochemistry in these complexes; thus, the alkoxo groups are ligated trans to the oxo moiety and receive electron density from, and the chelate co-ordinated in the equatorial plane donates electron density to, the acid metal centre.In addition, the more deshielded N-methyl group is syn oriented with respect to the oxo linkage (anisotropic contact shift), and the less deshielded one is anti positioned facing the alkoxo group.Similarly, the 31P signal of the equatorially coordinated chelate moves downfield from d 216.2 for free L to d 1.2–2.3 for the oxo-alkoxo derivatives. All of the rhenium(III) complexes have been characterised by (i) elemental analyses, (ii) conductivity measurements, (iii) spectroscopic measurements, i.e. IR, scanning electron microscopy (SEM)-EDX, positive ion FAB mass, NMR (see Table 2) and a crystal structure determination for 8.Their IR spectra show a number of absorptions characteristic of the phenylphosphino moieties, but the n(Re]] O) bands are absent. The SEM-EDX analyses are in agreement with the 1 : 2 : 3 Re :P:X (X = Cl or Br) ratios proposed for [ReX2L9]X and the 1 : 2 : 2 : 1 Re :P: Cl :S ratios for [ReCl2L2][CF3SO3] complexes. Octahedral rhenium(III) complexes are paramagnetic, as they usually possess a low-spin d4 electronic configuration; the two unpaired electrons are responsible for the spreading of the proton NMR signals over a wide ppm range.Nevertheless, successful observation of 1H NMR spectra is due to the extremely short electron-spin-relaxation time (T1e < 10211 s) of the rhe- Table 1 Spectroscopic data for rhenium(V) complexes NMRa (d) 1H IR (cm21) UV/VIS Positive-ion FABb Compound n(Re]] O) NMe OR 31P-{1H} l/nm (e/dm3 mol21 cm21) (m/z) L1 [ReOCl3L] 2 [ReOCl2(OMe)L] 3 [ReOCl2(OEt)L] 4 [ReOBr2(OEt)L] 5 [ReOI2(OEt)L] 6 [ReO2L2][BPh4] L9 12 [ReOCl3L9] — 982 940 915 918 918 821c — 992 2.59 (s) 4.07 (s), 3.43 (s) 4.07 (s), 3.25 (s) 4.09 (s), 3.26 (s) 4.19 (s), 3.28 (s) 4.27 (s), 3.28 (s) 3.18 (s) 4.12 (s) d 4.11 (s) d 2.63 (s) 2.93 (m), 0.49 (t) 2.89 (m), 0.48 (t) 2.84 (m), 0.40 (t) 216.2 (s) 216.6 (s) 1.2 (s) 1.3 (s) 2.3 (s) 1.6 (s) 9.7 (s) 222.2 (s) 29.2 (s) 680 (30), 350 (950), 275 (sh) 530 (40), 360 (750), 275 (sh) 520 (40), 360 (600), 275 (sh) 520 (40), 370 (600), 250 (sh) 530 (60), 395 (1300), 255 (sh) 475 (80), 360 (1000), 270 (sh) 614 (20, M1), 579 (100, M1 2 Cl), 544 (18, M1 2 2Cl) 610 (37, M1), 575 (100, M1 2 Cl), 544 (18, M1 2 Cl 2 OMe) 624 (40, M1), 589 (100, M1 2 Cl), 544 (35, M1 2 Cl 2 OEt) a Chemical shifts with multiplicities in parentheses: s = singlet, t = triplet, m = multiplet.b Relative intensities in % in parentheses. c n(O]] Re]] O). d NH.J. Chem. Soc., Dalton Trans., 1997, Pages 1421–1427 1423 Table 2 Spectroscopic data for rhenium(III) complexes 1H NMR (d) a UV/VIS l/nm Positive-ion FAB Compound a (d) b (t) g (t) d (d) NCH3 (s) a (d) b (t) c (t) (e/dm3 mol21 cm21) (m/z) 7 [ReCl2L2][CF3SO3] b 8 [ReCl3L(PPh3)] 9 [ReCl3L9(PPh3)] 10 [ReCl2L92]Cl 11 [ReBr2L92]Br 17.36 16.59 16.59 16.75 15.73 9.45 9.33 9.85 9.27 9.38 10.16 8.63 8.29 8.52 8.73 14.64 12.12 11.77 12.41 12.44 10.30 11.00 — — — 15.93 12.74, 15.18 c 13.30, 14.58 c 14.70 15.15 8.47 8.10, 8.74 c 8.04, 9.19 c 8.86 8.75 8.73 8.66, 8.11 c 7.93, 8.33 c 7.62 7.68 350 (3900), 285 (sh), 260 (9000) 460 (sh), 350 (3500), 285 (sh), 255 (9900) 445 (sh), 380 (2500), 315 (sh), 285 (sh) 450 (sh), 390 (5400), 345 (sh), 290 (sh) 860 (20, M1), 825 (100, M1 2 Cl) 812 (100, M1), 777 (45, M1 2 Cl), 742 (55, M1 2 2Cl) 902 (100, M1), 822 (58, M1 2 Br), 742 (3, M1 2 2Br) a Protons a, b, g, d, a, b and c as in Fig. 1. b In CD3CN solution. c Signal arising from the PPh3 coligand. nium(III) centre.22 The spectra of all [ReX2L2]1 or [ReX2L92]1 (X = Cl or Br) derivatives exhibit two sets of aromatic signals arising from the four phenyl groups attached to the phosphorus atoms (A2B2C system) and from the two phenyl rings interposed between the P and N donors (ABCD system).The coupling of the aromatic protons is evidenced in Fig. 1 for complex 7. Moreover, the proton spectra of 8 and 9 show a further set of aromatic protons due to the presence of the monodentate PPh3 group. As outlined in Table 2, all of the proton signals are shifted downfield with respect to the corresponding resonances of the unco-ordinated phosphinoamines, and the larger deshielding is experienced by those protons (a and d) close to co-ordinated P and N donors.Similar rhenium(III) complexes having one PPh3 coligated in an octahedral arrangement, as in [ReCl3(CNCMe3)2(PPh3)], fac-[ReCl3(PhCOCOPh)(PPh3)] and mer-[ReCl3(py)2(PPh3)] (py = pyridine), exhibit aromatic proton signals of the triphenylphosphine group in the same range as that observed for our [ReCl3(P]N)(PPh3)] derivatives,‡,23 indicating that the ReIIICl3(PPh3) group constitutes a rather stable moiety.Cyclic voltammetric studies on representative rhenium(III) complexes were performed in dichloromethane solutions. The electrochemistry of the bis-substituted [ReX2L92]1 complexes 10 (X = Cl) and 11 (Br) is dominated by the quasi-reversible 24 ReIII]ReII couple (E82� 1 at 21.06 and at 20.94 V for 10 and 11, respectively). As expected on the basis of the p-acid character of halides, the dibromo derivative is easier to reduce than the corresponding dichloro compound by 120 mV, in agreement with previously reported data on analogous [ReIIIX2(P]P)2]1 species [P]P = 1,2-bis(dimethylphosphino)ethane].25 Moreover our complexes, which contain a P2N2X2 donor set, are far more difficult to reduce (ca. 500 mV) than the above-mentioned compounds, 25,26 in which a larger number of phosphorus atoms in the co-ordination sphere (P4X2) ensures a better p-back bonding and stabilises ReII relative to ReIII.25 Conversely, the cyclic voltammogram of the monosubstituted complex [ReCl3L- (PPh3)] shows a quasi-reversible ReIII–ReIV wave at 10.25 V.No other redox processes are observable in the detectable window. Crystal structures The ReV and ReIII in complexes 2 and 8 (Figs. 2 and 3, respectively) exhibit approximately octahedral co-ordination geometry, the maximum deviations from regularity occurring at the angles subtended by the nitrogen and phosphorus donors of the L ligands.In effect, the neutral chelate co-ordinates in the ‡ The aromatic protons of the phenyl groups of co-ordinated PPh3 fall at d 13.42, 7.95 and 7.61 for [ReCl3(CNCMe3)2(PPh3)] in C6D6 solutions; 23 analogously, they occur at d 13.54, 8.85 and 8.49 for mer- [ReCl3(py)2(PPh3)] and at d 14.258.69 and 8.49 for fac-[ReCl3- (PhCOCOPh)(PPh3)] in CDCl3 solutions. bidentate mode via phosphorus [Re]P 2.367(4) in 2 and 2.388(3) Å in 8, respectively] and nitrogen atoms [Re]N 2.28(1) and 2.34(1) in 2 and 8, respectively] (Table 3). The fivemembered chelate ring thus formed, i.e.RePCCN, is roughly planar within ±0.08 Å in 2 (pertinent torsion angles in the range 29.7 to 11.88) but severely rippled up to ±0.20 Å in 8 (torsion angles from 227.7 to 23.68), the PCCN ring forming a dihedral angle with the square plane about the Re atom of 9.38 in 2 and 28.78 in 8. Bidentate co-ordination of the ligand L has been previously observed in the structures of four-co-ordinate Table 3 Selected bond lengths (Å) and angles (8) Complex 2 Re]Cl(1) Re]P Re]O(2) P]C(1) P]C(16) N]C(2) N]C(8) Cl(1)]Re]Cl(2) Cl(1)]Re]O(1) Cl(1)]Re]O(2) O(1)]Re]O(2) P]Re]N Re]P]C(1) Re]P]C(16) Re]O(2)]C(9) Re]N]C(8) Cl(1)]Re]P Cl(2)]Re]O(1) Cl(2)]Re]O(2) Cl(1)]Re]N Complex 8 Re]Cl(1) Re]Cl(3) Re]P(2) P(1)]C(1) P(1)]C(15) P(2)]C(27) N]C(2) N]C(8) Cl(1)]Re]Cl(2) Cl(1)]Re]Cl(3) Cl(1)]Re]P(1) Cl(1)]Re]P(2) Cl(1)]Re]N Cl(2)]Re]Cl(3) Cl(2)]Re]P(1) Cl(2)]Re]P(2) Cl(2)]Re]N Cl(3)]Re]P(1) Cl(3)]Re]P(2) Cl(3)]Re]N 2.383(4) 2.367(4) 1.84(1) 1.82(1) 1.83(1) 1.47(2) 1.51(2) 88.3(2) 94.5(3) 91.4(3) 173.2(4) 82.4(3) 101.0(4) 120.2(4) 111(1) 108.5(8) 96.2(1) 95.1(4) 88.5(3) 178.4(3) 2.427(3) 2.337(4) 2.423(3) 1.80(1) 1.83(1) 1.84(1) 1.46(2) 1.46(2) 90.9(1) 90.9(1) 170.4(1) 85.1(1) 91.8(3) 175.7(1) 86.7(1) 94.9(1) 83.5(3) 90.9(1) 89.2(1) 92.5(3) Re]Cl(2) Re]O(1) Re]N P]C(10) O(2)]C(9) N]C(7) O(1)]Re]N Re]P]C(10) C(1)]P]C(16) Re]N]C(2) Cl(2)]Re]P P]Re]O(1) P]Re]O(2) Cl(2)]Re]N O(2)]Re]N C(1)]P]C(10) C(10)]P]C(16) Re]N]C(7) Re]Cl(2) Re]P(1) Re]N P(1)]C(9) P(2)]C(21) P(2)]C(33) N]C(7) Re]P(1)]C(1) Re]P(1)]C(9) Re]P(1)]C(15) Re]P(2)]C(21) Re]P(2)]C(27) Re]P(2)]C(33) Re]N]C(2) Re]N]C(7) Re]N]C(8) P(1)]C(1)]C(2) N]C(2)]C(1) 2.438(5) 1.70(1) 2.28(1) 1.80(1) 1.39(2) 1.52(2) 86.4(4) 114.1(4) 105.0(6) 113.5(8) 173.8(1) 88.8(3) 87.1(3) 93.0(3) 87.6(4) 108.5(6) 106.9(6) 107.7(8) 2.336(4) 2.388(3) 2.34(1) 1.82(1) 1.83(1) 1.83(1) 1.53(2) 101.0(4) 125.9(4) 115.1(4) 113.6(4) 117.0(4) 115.0(4) 112.1(7) 108.2(7) 112.3(8) 118.4(9) 120(1)1424 J.Chem. Soc., Dalton Trans., 1997, Pages 1421–1427 Fig. 1 Two-dimensional COSY-90 1H NMR contour map of complex 7 in CD3CN over the region d 7.00–19.00 [IrCl(CO)L],9 five-co-ordinate [RuCl2L(PPh3)]?C6H6 10 and sixco- ordinate [IrClH{CH2NH(Me)C6H4PPh2}L]?0.5CH2Cl2.11 The bidentate ligands show P ? ? ? N distances of 3.06 and 3.00 Å in 2 and 8, respectively, which is shortened to 2.95, 2.95 and 2.89 Å in the three compounds mentioned above.The corresponding ‘bite’ angles are 82.48 in 2, 78.68 in 8, and 84.8, 79.6 and 81.88 in the other three species. In all of these complexes the angles about the phosphorus and nitrogen atoms have an approximately tetrahedral arrangement. The methyl groups at the nitrogen atoms are symmetrically placed above and below the equatorial co-ordination plane by 1.19 and 21.28 Å in 2, while they are 0.73 and 21.34 Å further from the plane in 8.The structures of complexes 2 and 8 are roughly superimposable (Fig. 4) with a weighted root-mean-square (r.m.s.) deviation, derived from the BMFIT program,28 of 0.10 Å, when the fitting is performed using the equatorial atoms. In 2 the metal atom is displaced from the mean plane of the Cl2PN donor atom set by 0.05 Å towards the oxo O(1) atom and in 8 it is 0.03 Å from the mean Cl2PN plane towards Cl(3). The Re atom in the co-ordination polyhedron of 2 is 1.25 Å from the O(2), N, Cl(2) plane and 21.13 Å from the O(1), P, Cl(1) plane, the angle between the two triangular faces being 5.38.The corresponding values for 8 are 1.40, 21.25 Å and 4.78. In 2 the chloride ligands occupy mutually cis positions with a Cl(1)]Re]Cl(2) angle of 88.3(2)8, while in 8 the chlorine atoms adopt a meridional arrangement. The Re]Cl bonds trans to P [Re]Cl(2) in 2 and Re]Cl(1) in 8 are of comparable length, 2.438(5) and 2.427(3) Å, respectively] but longer than that trans to the amino-nitrogen in 2 [2.383(4) Å] and than those involving the mutually trans positions in 8 [2.336(4) and 2.337(4) Å].This result confirms the lower trans influence of the amino relative to the phosphine group.29 On the other hand, the ReV]P and ReIII]P(1) bonds trans to Cl are only slightly different [2.367(4) and 2.388(3) Å, respectively], while the ReIII]P(2) distance, trans to the aminoJ.Chem. Soc., Dalton Trans., 1997, Pages 1421–1427 1425 nitrogen donor, is somewhat longer [2.423(3) Å]. The shortening of the ReV]N distance trans to Cl(1) [2.28(1) Å] compared to the ReIII]N bond length trans to P [2.34(1) Å] may be ascribed to the greater trans influence exhibited by an organophosphine with respect to a chloride ligand. The mean value of 2.31 Å represents the greatest value for an Re]N distance, dramatically greater than the mean of 2.00 Å observed for the Re]N (amido) bonds in [ReO(L9 2 H)2X] (X = OEt or Cl).15 A similar trend has been encountered for Pt]Namino vs.Pt]Namido distances in four-co-ordinate cis/trans-[PtL92]21 and [Pt(L9 2 H)2] complexes,30 but in those cases the difference is restricted to 0.08 Å. In any case, the extreme length of the Re]N bonds parallels that found, for example, in trigonal-bipyramidal [NiCl2L(PMe2Ph)] and [NiCl2L(PMePh2)] complexes [2.353(4) and 2.325(8) Å, respectively], where the dimethylamino nitrogen of L still faces a phosphorus donor of a tertiary-phosphine coligand.31,32 This distance shortens to 2.056(8) Å in the squareplanar compound [NiClL(PMePh2)]1,32 indicating that a lesscrowded configuration allows for a stronger metal–nitrogen interaction. Experimental Physical measurements Elemental analyses (C, H, N) were performed on a Fisons model EA 1108 elemental analyser.Only for rhenium(III) compounds, spot or selected area analyses, to determine the Re :P:X: S (X = Cl or Br) ratios, were performed by integral counting of the back-scattered X-ray fluorescence radiation from a Philips model XL 40 scanning-electron microscope Fig. 2 An ORTEP27 view of complex 2 showing the atom-labelling scheme; thermal ellipsoids are drawn at 50% probability and hydrogen atoms are omitted for clarity Fig. 3 An ORTEP27 diagram of complex 8 with 50% probability ellipsoids equipped with an EDX model data station following the procedure described elsewhere.14 Infrared spectra were recorded on a Mattson 3030 Fourier-transform spectrometer (4000–400 cm21) using KBr pellets, 1H and 31P NMR spectra on a Bruker AC-200 instrument using SiMe4 (1H) as internal and 85% aqueous H3PO4 (31P) as external reference and UV/VIS spectra in CH2Cl2 using a Cary 17D spectrophotometer (750–220 nm).Conductivity measurements were made in MeCN at 25 8C using a Metrohm Herison E518 conductometer. Fast-atom bombardment mass spectra in the positive mode were recorded by using 3-nitrobenzyl alcohol or glycerol matrices on a VG ZAB-2F spectrometer.Xenon was used as the primary beam gas, and the ion gun was operated at 8 keV (ca. 1.28 × 10215 J). Data were collected over the mass range m/z 100–1000 at 0.7 s per scan. Cyclic voltammetry measurements were performed on a BAS (Bioanalytical System Inc.) CV-IB cyclic voltammograph at room temperature under an atmosphere of nitrogen by using a conventional three-electrode cell. A platinum-disc electrode (area ca. 1023 cm2) was used as the working electrode, a platinum wire as the counter electrode and a silver wire as a quasi-reference electrode. The measurements of the complexes (ca. 5 mmol dm23) were done in dry and degassed CH2Cl2 solutions (Aldrich gold label) with 0.2 mol dm23 NBu4ClO4 as supporting electrolyte. Potentials were calculated at the midpoint between the anodic and cathodic peaks of the cyclic voltammetric waveform scanned at 0.2 V s21, and were referenced internally to the ferrocenium–ferrocene couple.24,33 All electrontransfer processes measured in this work are monoelectronic, as qualitatively established by comparison of the current peaks of an equimolar amount of the internal ferrocenium–ferrocene standard.Materials Unless otherwise stated, all chemicals and solvents were of reagent grade used without further purification. The compounds [NBu4][ReOCl4] and [ReOX2(OEt)(PPh3)2] were prepared as reported elsewhere 34 starting from fine metal rhenium powder (a gift from H.C. Starck GmbH, Goslar, Germany), which was oxidised to ReVII prior to use. The compounds L and L9 were prepared according to the literature.35,36 Syntheses [ReOCl3L] 1. The salt [NBu4][ReOCl4] (96 mg, 0.19 mmol) was dissolved in MeOH (10 cm3). To the pale green solution solid L (58 mg, 0.19 mmol) was added with stirring at room temperature. The solution turned blue and a precipitate was suddenly deposited. After 30 min the pale blue solid was filtered off, washed with MeOH (5 cm3) and Et2O (10 cm3), and dried under vacuum (yield 90 mg, 78%) (Found: C, 39.4; H, 3.3; N, 2.4.C20H20Cl3NOPRe requires C, 39.1; H, 3.3; N, 2.3%). Molar conductivity, LM = 11 W21 cm2 mol21. Fig. 4 Superimposition of the structures of complexes 2 and 8 (- - -)1426 J. Chem. Soc., Dalton Trans., 1997, Pages 1421–1427 Table 4 Structure determination summary for complexes 2 and 8* 2 8 Empirical formula M Colour, habit Crystal size/mm Space group a/Å b/Å c/Å U/Å3 Z Dc/Mg m23 m/mm21 F(000) 2q Range/8 hkl Ranges Independent reflections Observed reflections [Fo > 4s(Fo)] Weighting scheme, w21 Number of parameters refined Final R, R9 (observed data) Goodness of fit Largest difference peak, hole/e Å23 C21H23Cl2NO2PRe Violet parallelepipeds 0.30 × 0.40 × 0.10 Pbca 15.967(7) 14.976(6) 18.658(10) 4462(4) 8 1.815 5.8 2368 4.0–45.0 0–17, 0–16, 0–20 2879 1611 s2(F ) 1 0.0023F 2 134 0.042, 0.050 0.78 1.00, 21.54 C38H35Cl3NP2Re Orange parallelepipeds 0.15 × 0.20 × 0.30 Pna21 19.636(7) 17.570(5) 10.172(6) 3510(3) 4 1.628 3.8 1704 4.0–50.0 0–23, 0–20, 0–12 3291 2522 s2(F ) 1 0.0015F 2 211 0.035, 0.043 0.83 1.33, 20.69 * Details in common: orthorhombic; Siemens R3m/V diffractometer; Mo-Ka radiation (l = 0.710 73 Å); w–2q scans; variable scan rate 2.49–14.658 min21 in w; scan range 1.058 1 Ka; stationary crystal–stationary counter at beginning and end of each scan, each for 25% of total scan time; two standard reflections every 150; SHELXTL PLUS;37 full-matrix least squares minimising Sw(|Fo| 2 |Fc| )2; data to parameter ratio 12 : 1.[ReOCl2(OR)L] (R = Me 2 or Et 3). The salt [NBu4]- [ReOCl4] (136 mg, 0.27 mmol) was dissolved in MeOH (10 cm3) (or EtOH) and solid L (166 mg, 0.54 mmol) was added with stirring. The mixture was refluxed for 3 h until a clear violet solution gave a violet precipitate. After cooling, the solid was filtered off, washed with MeOH (3 cm3) (or EtOH) and Et2O (10 cm3) and dried under vacuum.Complex 2: yield 130 mg (79%) (Found: C, 41.5; H, 3.9; N, 2.5. C21H23Cl2NO2PRe requires C, 41.4; H, 3.8; N, 2.3%); LM = 8 W21 cm2 mol21. Complex 3: yield 83% (Found: C, 41.5; H, 3.9; N, 2.5. C22H25Cl2NO2PRe requires C, 42.35; H, 4.05; N, 2.25%); LM = 17 W21 cm2 mol21. [ReOX2(OEt)L] (X = Cl 3, Br 4 or I 5). These complexes were prepared in EtOH as described for 2 and 3 but using different starting materials: [ReOCl2(OEt)(PPh3)2] for 3, [ReOBr2( OEt)(PPh3)2] for 4 and [ReOI2(OEt)(PPh3)2] for 5. Complex 4: yield 64% (Found: C, 37.5; H, 3.7; N, 2.1.C22H25Br2NO2PRe requires C, 37.1; H, 3.55; N, 1.95%); LM = 20 W21 cm2 mol21. Complex 5: yield 51% (Found: C, 33.5; H, 3.5; N, 1.8. C22H25I2NO2PRe requires C, 32.75; H, 3.1; N, 1.75%); LM = 23 W21 cm2 mol21. [ReO2L2][BPh4] 6. The salt [AsPh4][ReOCl4] (105 mg, 0.14 mmol) was suspended in MeCN (5 cm3) and solid L (44 mg, 0.14 mmol) was added with stirring.The mixture was left to stand at room temperature for 5 min until it was green. Additional L (44 mg, 0.14 mmol) and an ethanolic solution (2 cm3) containing 0.2 mol dm23 sodium ethoxide were added. The solution darkened and was stirred overnight at room temperature. Then the solvent was removed by a gentle stream of dinitrogen and the residue treated with water–dichloromethane (1 : 1, 6 cm3). The organic phase was separated, reduced in volume and treated with ethanol–hexane (10 cm3).After vigorous stirring a cream solid precipitated. It was washed twice with hexane (5 cm3) and dried under vacuum. Yield 36 mg (28%) (Found: C, 68.3; H, 6.0; N, 2.5. C68H70BN2O2P2Re requires C, 67.7; H, 5.85; N, 2.3%). LM = 186 W21 cm2 mol21. [ReCl2L2][CF3SO3] 7. Complex 2 (50 mg, 0.08 mmol) was dissolved in a CH2Cl2–MeOH (1 : 1, 5 cm3) in a test-tube. Water (1 cm3) and an excess of neat trifluoromethanesulfonic acid (0.5 cm3) were added. The tube was capped and after several days yellow crystals separated.They were filtered off and dried under vacuum (yield 11 mg, 14%) (Found: C, 50.4; H, 5.0; N, 2.6; S, 3.0. C45H50Cl2F3N2O3P2ReS requires C, 50.3; H, 4.7; N, 2.6; S, 3.0%). LM = 197 W21 cm2 mol21. [ReCl3L(PPh3)] 8. The complex [ReCl3(MeCN)(PPh3)2] (83 mg, 0.10 mmol) was suspended in MeOH (10 cm3). Solid L (59 mg, 0.19 mmol) was added with stirring and the mixture refluxed under a dinitrogen atmosphere for 24 h.The hot yellow-orange solution containing a precipitate was filtered and the orange powder washed with MeOH (3 cm3) and Et2O (10 cm3) and dried under vacuum. Yield 60 mg (67%) (Found: C, 53.8; H, 4.4; N, 1.7. C38H35Cl3NP2Re requires C, 53.05; H, 4.10; N, 1.65%). LM = 15 W21 cm2 mol21. [ReCl3L9(PPh3)] 9. The complex [ReCl3(MeCN)(PPh3)2] (94 mg, 0.11 mmol) was suspended in EtOH–benzene solution (1 : 1, 10 cm3). Solid L9 (61 mg, 0.22 mmol) and an excess of 37% HCl (0.5 cm3) were added with stirring and then refluxed for 2 h.The solution turned red-brown and, after cooling, a brick-red solid was filtered off, washed with cold EtOH and Et2O. Yield 30 mg (32%) (Found: C, 52.0; H, 3.6; N, 1.6. C36H31Cl3NP2Re requires C, 51.9; H, 3.75; N, 1.7%). [ReX2L92]X (X = Cl 10 or Br 11). These two complexes were prepared in a similar way, detailed for 10. The complex [ReCl3(MeCN)(PPh3)2] (171 mg, 0.20 mmol) was suspended in EtOH–benzene solution (1 : 1, 10 cm3). Solid L9 (110 mg, 0.40 mmol) and an excess of 37% HCl (0.5 cm3) were added with stirring.This mixture was refluxed for 18 h and a brick-red solid was filtered off after cooling (see above). The yellow filtrate was reduced in volume by a gentle stream of dinitrogen until a yellow solid precipitated. This powder was filtered off, washed with EtOH (2 cm3) and Et2O (10 cm3), and dried under vacuum. Yield 52 mg (30%) (Found: C, 51.7; H, 4.0; N, 3.2. C36H32Cl3N2P2Re requires C, 51.05; H, 3.8; N, 3.3%).Complex 11 was prepared starting from [ReBr3(MeCN)(PPh3)2] and an excess of 48% HBr. Yield 24% (Found: C, 44.9; H, 3.5; N, 2.6. C36H32Br3N2P2Re requires C, 44.1; H, 3.3; N, 2.85%). [ReOCl3L9] 12. Solid L9 (189 mg, 0.68 mmol) and an excess of 37% HCl (1 cm3) were added to a solution containing [NBu4][ReOCl4] (100 mg, 0.17 mmol) dissolved in MeCNJ. Chem. Soc., Dalton Trans., 1997, Pages 1421–1427 1427 (5 cm3) with stirring at room temperature. A pale blue solid was quickly deposited and was filtered off, washed with EtOH (5 cm3) and Et2O (5 cm3), and dried under vacuum.Yield 85 mg (85%) (Found: C, 37.2; H, 2.7; N, 2.5. C18H16Cl3NOPRe requires C, 36.9; H, 2.75; N, 2.4%). Crystallography Room-temperature crystallisation over 2 d gave violet crystals of complex 2 which were washed with CH2Cl2 and dried. Crystals of 8 suitable for X-ray analysis were grown from a CH2Cl2 solution layered with EtOH. For both structures the structure determination data are summarised in Table 4.The structures were solved by conventional heavy-atom methods, the coordinates of the Re atoms being determined from the Patterson function and those of the remaining non-hydrogen atoms from subsequent Fourier syntheses. Anisotropic thermal parameters were applied only to the Re, Cl and P atoms; all H atoms were included in calculated positions and refined anisotropically using a riding model. The correct configuration of 8 was ascertained by the h test 38 [11.07(12)].Absorption correction was performed using the y-scan method39 for fine reflections at c ca. 908 and the refinement procedure was based on F. Atomic coordinates, thermal parameters, and bond lengths and angles have been deposited at the Cambridge Crystallographic Data Centre (CCDC). See Instructions for Authors, J. Chem. Soc., Dalton Trans., 1997, Issue 1. Any request to the CCDC for this material should quote the full literature citation and the reference number 186/415. Acknowledgements Mr.Antonio Aguiari and Mrs. Anna Moresco are gratefully acknowledged for their skilled technical assistance. References 1 W. Keim, Angew. Chem., Int. Ed. Engl., 1990, 29, 235; M. C. Bonnet, F. Dahan, A. Ecke, W. Keim, R. P. Schultz and I. Tkachenko, J. Chem. Soc., Chem. Commun., 1994, 615; G. J. Britovsek, W. Keim, S. Mecking, D. Sainz and T. Wagner, J. Chem. Soc., Chem. Commun., 1993, 1632. 2 P. Braunstein, Y. Chauvin, S.Mercier, L. Saussine, A. DeCian and J. Fisher, J. Chem. Soc., Chem. Commun., 1994, 2203; P. Braunstein, J. Pietsch, Y. Chauvin, S. Mercier, L. Saussine, A. DeCian and J. Fisher, J. Chem. Soc., Dalton Trans., 1996, 3571. 3 A. Habtemariam and P. J. Sadler, Chem. Commun., 1996, 1785. 4 E. F. Lutz, J. Chem. Educ., 1986, 63, 202. 5 H. P. Fritz, I. R. Gordon, K. E. Schwarzhans and L. M. Venanzi, J. Chem. Soc., 1965, 5210. 6 H. P. Fritz and K. E. Schwarzhans, J. Organomet. Chem., 1966, 5, 103. 7 T. B. Rauchfuss, F. T. Patino and D. M. Roundhill, Inorg. Chem., 1975, 14, 652. 8 T. B. Rauchfuss and D. M. Roundhill, J. Am. Chem. Soc., 1974, 96, 3098. 9 D. M. Roundhill, R. A. Bechtold and S. Roundhill, Inorg. Chem., 1980, 19, 284. 10 D. C. Mudalige, S. J. Rettig, B. R. James and W. R. Cullen, J. Chem. Soc., Chem. Commun., 1993, 830. 11 E. Farnetti, G. Nardin and M. Graziani, J. Chem. Soc., Chem. Commun., 1989, 1264. 12 B. R. Martin, 31st ICCC Conference, Vancouver, 1996, abstract PL2, p. 4. 13 J. Reedijk, Chem. Commun., 1996, 801. 14 F. Refosco, C. Bolzati, A. Moresco, G. Bandoli, A. Dolmella, U. Mazzi and M. Nicolini, J. Chem. Soc., Dalton Trans., 1991, 3043. 15 F. Refosco, F. Tisato, G. Bandoli, C. Bolzati, A. Dolmella, A. Moresco and M. Nicolini, J. Chem. Soc., Dalton Trans., 1993, 605. 16 F. Refosco, F. Tisato, A. Moresco and G. Bandoli, J. Chem. Soc., Dalton Trans., 1995, 3475. 17 M. Nicolini, G. Bandoli and U. Mazzi, Technetium and Rhenium in Chemistry and Nuclear Medicine, SGEditoriali, Padova, 1995, vol. 4. 18 F. Tisato, F. Refosco, A. Moresco, G. Bandoli, A. Dolmella and C. Bolzati, Inorg. Chem., 1995, 34, 1779. 19 G. D. Meakins and R. D. Hill, J. Chem. Soc., 1958, 229. 20 F. Tisato, C. Bolzati, A. Duatti, G. Bandoli and F. Refosco, Inorg. Chem., 1993, 32, 2042. 21 F. Tisato, F. Refosco, E. Deutsch and M. Nicolini, Technetium and Rhenium in Chemistry and Nuclear Medicine, eds. M. Nicolini, G. Bandoli and U. Mazzi, SGEditoriali, Padova, 1995, vol. 4, p. 155. 22 H. J. Keller and K. E. Schwarzhans, Angew. Chem., Int. Ed. Engl., 1970, 9, 196; R. R. Conry and J. M. Mayer, Inorg. Chem., 1990, 29, 4862; X. L. R. Fontaine, E. H. Fowles, T. P. Layzell, B. L. Shaw and M. Thornton-Pett, J. Chem. Soc., Dalton Trans., 1991, 1519; S. D. Orth, J. Barrera, M. Sabat and W. D. Harman, Inorg. Chem., 1993, 32, 594; L. E. Helberg, S. D. Orth, M. Sabat and W. D. Harman, Inorg. Chem., 1996, 35, 5584. 23 J. C. Bryan, R. E. Stenkamp, T. H. Tulip and J. M. Mayer, Inorg. Chem., 1987, 26, 2283. 24 R. R. Gagne, C. A. Koval and G. C. Lisensky, Inorg. Chem., 1980, 19, 2854. 25 J. R. Kirchoff, W. R. Heineman and E. Deutsch, Inorg. Chem., 1987, 26, 3108. 26 F. Refosco, F. Tisato, G. Bandoli and E. Deutsch, J. Chem. Soc., Dalton Trans., 1993, 2901. 27 C. K. Johnson, ORTEP, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976. 28 S. C. Nyburg, Acta Crystallogr., Sect. B, 1974, 30, 251. 29 G. J. Organ, M. K. Cooper, K. Henrich and M. McPartlin, J. Chem. Soc., Dalton Trans., 1984, 2377. 30 M. K. Cooper, J. M. Downes, H. J. Goodwin and M. McPartlin, Inorg. Chim. Acta, 1983, 76, L157; M. K. Cooper, J. M. Downes, H. J. Goodwin, M. McPartlin and J. M. Rosalsky, Inorg. Chim. Acta, 1983, 76, L155. 31 L. Crociani, F. Refosco, F. Tisato, S. Gatto and B. Corain, Inorg. Chim. Acta, 1996, 249, 131. 32 L. Crociani, F. Refosco, F. Tisato, A. Dolmella, S. Gatto and G. Bandoli, Z. Kristallogr., submitted for publication. 33 G. Gritzner and J. Kuta, Electrochim. Acta, 1984, 29, 869. 34 G. Rouschias, Chem. Rev., 1974, 74, 531 and refs. therein. 35 T. B. Rauchfuss and D. M. Roundhill, J. Am. Chem. Soc., 1974, 96, 3098. 36 M. K. Cooper, J. M. Downes, P. A. Duckworth, M. C. Kerby, R. J. Powell and M. D. Soucek, Inorg. Synth., 1989, 25, 129. 37 G. M. Sheldrick, SHELXTL PLUS, version 4.2, Siemens Analytical Instruments, Madison, WI, 1990. 38 D. Rogers, Acta Crystallogr., Sect. A, 1981, 37, 734. 39 A. C. T. North, D. C. Phillips and F. S. Mathews, Acta Crystallogr., Sect. A, 1968, 24, 351. Received 7th November 1996; Paper 6/07584C

ISSN:1477-9226    

DOI:10.1039/a607584c

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

DALTON J. Chem. Soc., Dalton Trans., 1997, Pages 1429–1433 1429 Six-co-ordinate 1,2-dithiolene complexes of tungsten(II) of the type [W(S]S)(CO)2L2] [S]S = C3S5, benzene-1,2-dithiolate or maleonitriledithiolate; L2 = (PPh3)2, (PEt3)2 or Ph2P(CH2)2PPh2] Paul K. Baker,*,a Michael G. B. Drew,*,b Emma E. Parker,a Neil Robertson a and Allan E. Underhill a a Department of Chemistry, University of Wales, Bangor, Gwynedd LL57 2UW, UK b Department of Chemistry, University of Reading, Whiteknights, Reading RG6 2AD, UK Reaction of the complexes [WI2(CO)3L2] [L2 = (PEt3)2, (PPh3)2 or Ph2P(CH2)2PPh2 = dppe] with 1 equivalent of Na2[S]S] (S]S = C3S5 or maleonitriledithiolate) or H2bdt (benzene-1,2-dithiol) in acetonitrile and ethanol at room temperature afforded good yields of the six-co-ordinate compounds [W(S]S)(CO)2L2] by displacement of the iodide ligands and a carbonyl ligand.X-Ray crystallographic studies were carried out on [W(S]S)(CO)2L2] [S]S = C3S5, L2 = (PEt3)2; S]S = bdt, L2 = (PEt3)2 or L2 = dppe]. The three structures are all six-co-ordinate with geometries intermediate between octahedral and trigonal prismatic but closer to the latter.The importance of transition-metal complexes containing dianionic delocalised sulfur-donor chelating ligands, such as C3S5 22 (4,5-disulfanyl-1,3-dithiole-2-thionate) and mnt (mnt = maleonitriledithiolate), has increased in recent years. Many examples of solid-state materials derived from these complexes have been shown to conduct electric currents, display unusual magnetic properties and have non-linear optical properties.1–3 Although many examples of square-planar complexes have been described such as b-[NMe4][Pd(C3S5)2]2, which is a superconductor at 2 K,4 few examples of organotransition-metal complexes containing C3S5 and mnt are known.Two early examples are the bis(cyclopentadienyl) complexes [M(mnt)- (cp)2] (M = Mo or W),5 and in 1970 Eisenberg 6 comprehensively reviewed 1,1- and 1,2-dithiolato-chelate complexes, including their organotransition-metal complexes.Several molybdenum complexes containing cp or C5Me5 and dithiolene ligands have been recently studied including [Mo(dmit)2(h- C5Me5)] 7 and the charge-transfer salt [Mo(dddt)(cp)2][tcnq] 8 (dddt = 5,6-dihydro-1,4-dithiin-2,3-dithiolate, tcnq = tetracyanoquinodimethane) which display co-operative low-temperature magnetic phenomena. Other mixed cp–dithiolene complexes which have also recently been studied are complexes of Fe,9 Co,10 Ti 11 and Ru.12 Fewer examples exist of complexes containing dithiolene and phosphorus ligands; the adducts [Co(S2C2X2)(cp)L] 10 (L = tributylphosphine or tributyl phosphite) and the cluster [Au10(mnt)2(PPh3)7] 13 being rare examples.The organometallic square-planar platinum complex [Pt(mnt)(CNMe)2]?(NC)2C2S2CNMe has recently been shown crystallographically to demonstrate costacking of neutral planar metal and organic molecules.14 Over the past ten years we have been investigating the chemistry of the highly versatile seven-co-ordinate complexes [MI2- (CO)3(NCMe)2] (M = Mo or W) and their derivatives.15 Continuing our exploration of the chemistry of these and related complexes, in this paper we describe the reactions of the sevenco- ordinate complexes [WI2(CO)3L2] [L2 = (PEt3)2, (PPh3)2 or Ph2PCH2CH2PPh2 (dppe)] with 1 equivalent of Na2[S]S] (S]S = C3S5 or mnt) or H2bdt (bdt = benzene-1,2-dithiolate) to give the novel six-co-ordinate complexes [W(S]S)(CO)2L2].The molecular structures of [W(S]S)(CO)2(PEt3)2] (S]S = C3S5 or bdt) and [W(bdt)(CO)2(dppe)] are also described. Results and Discussion The starting materials for this research, namely [WI2(CO)3L2] [L2 = (PEt3)2 16 (PPh3)2 17 or dppe18] were prepared by treating [WI2(CO)3(NCMe)2] with 2 equivalents of L = PEt3 or PPh3 or 1 equivalent of dppe. The complexes [WI2(CO)3L2] react at room temperature with 1 equivalent of Na2[S]S] (S]S = C3S5 or mnt) or H2bdt to give good yields of the six-co-ordinate complexes [W(S]S)(CO)2L2] 1–9 by the displacement of the iodide ligands and loss of a carbonyl ligand.All the new complexes have been fully characterised by elemental analysis (C, H and N), infrared (Table 1) and 1H NMR spectroscopy (Table 2); 2, 3 and 8 were also characterised by 31P NMR spectroscopy (Table 2). Complexes 2, 5 and 6 were also crystallographically characterised.Although the solid-state crystal structure of the complex [W(bdt)(CO)2(dppe)] 6 occurs as a 0.5CH2Cl2 solvate, the bulk material did not show a 0.5CH2Cl2 solvate in the 1H NMR spectrum and repeated elemental analysis agreed with the nonsolvate formulation. The FAB mass spectra of the complexes [W(C3S5)(CO)2L2] [L2 = (PEt3)2 or dppe] show their molecular ions at m/z 672 [L2 = (PEt3)2] and 834 [L2 = dppe]. Magnetic susceptibility measurements have shown complexes 1–9 to be diamagnetic.The bis(triethylphosphine) complexes are soluble in diethyl ether and chlorinated solvents, whereas the (PPh3)2 and dppe complexes were considerably less soluble but were soluble in polar chlorinated solvents such as CH2Cl2 and CHCl3. All the complexes were stable when stored under nitrogen, however they decomposed in solution when exposed to air. The characteristic absorption bands for the dithiolene ligands are present in the infrared spectra; comparison of the bands arising from the C]] C and the C]S bonds of the metalbound dithiolene and the free dithiolene show only a slight change in frequency.The stretching frequency of the C]S bond in the ligated dithiolene is slightly lower than in the free dithiolene, as would be expected when bound in a bidentate manner to a metal complex. It has been observed that the C]] C stretching mode in square-planar dithiolene complexes reduces in frequency as the charge on the molecule is decreased consistent with a reduction in bonding electron density at this position.The detection of this band at about 1430 cm21 for the complexes described here is suggestive of highly negatively charged dithiolene ligands bearing approximately a full double negative charge.19 The absorption bands in the regions 1937–1961 and 1863–1889 cm21 correspond to the two cis-carbonyl groups attached to the tungsten centre. The capping carbonyl bands observed in the infrared spectra of the tungsten complexes [WI2(CO)3L2] [L2 = (PPh3)2, (PEt3)2 or dppe] at ca. 2000 cm21 are no longer present.This is supported by the crystal struc-1430 J. Chem. Soc., Dalton Trans., 1997, Pages 1429–1433 Table 1 Physical, analytical a and IR b data for the complexes [W(S]S)(CO)2L2] Analysis (%) IR (cm21) Complex Colour Yield (%) C H N n(CO) n(CS) 1 [W(C3S5)(CO)2(PPh3)2] Dark 78 50.0 3.4 — 1943 1062 brown (51.3) (3.8) 1863 2 [W(C3S5)(CO)2(PEt3)2] Brown- 79 30.4 4.6 — 1950 1059 black (30.4) (4.5) 1879 3 [W(C3S5)(CO)2(dppe)] Dark 48 43.6 3.3 — 1961 1057 brown (44.6) (2.9) 1885 4 [W(bdt)(CO)2(PPh3)2] Dark 37 57.8 3.9 — 1955 1031 brown (58.4) (3.7) 1886 1115 5 [W(bdt)(CO)2(PEt3)2] Brown 50 38.8 5.6 — 1941 1028 (39.0) (5.5) 1884 1119 6 [W(bdt)(CO)2(dppe)] Dark 44 53.3 3.9 — 1946 1027 brown (52.4) (3.6) 1873 1128 7 [W(mnt)(CO)2(PPh3)2] Brown 65 55.5 3.7 2.2 1955 1091 (55.8) (3.3) (3.1) 1886 1120 8 [W(mnt)(CO)2(PEt3)2] Brown- 70 34.9 5.0 4.3 1937 1035 black (35.1) (4.9) (4.5) 1855 1113 9 [W(mnt)(CO)2(dppe)] Brown 69 49.2 3.0 3.4 1962 1024 (49.4) (3.1) (3.6) 1889 1154 a Calculated values in parentheses.b Recorded in CH2Cl2 as thin films between NaCl plates. tures (see Figs. 1–3) which show only the two cis-carbonyls to be present in the products 2, 5 and 6. The near-infrared spectrum of [W(C3S5)(CO)2(PPh3)2] 1 shows two absorption bands at 2325 nm and a broad band at 900–700 nm. The latter band is usually observed for C3S5 complexes.The crystal structures contain discrete molecules of complexes 2, 5 and 6, illustrated in Figs. 1, 2 and 3 together with the atomic numbering scheme. The structures of 2 and 5 are very similar, in particular with regard to the geometry of the coordination sphere. The two carbonyl groups are mutually cis albeit with angles in excess of 908 [106.5(4), 108.2(5) in 2 and 107.0(6)8 in 5] (Table 3) and trans to the bidentate C3S5 or bdt ligand, though the angles are very much distorted from 1808.This cis-arrangement of two carbonyl ligands is usually found in dicarbonyl structures because their strong p acidity makes the trans arrangement unfavourable. The two triethylphosphine monodentate ligands are also trans to each other though the angles are significantly distorted from linear [142.3(1), 140.2(1) in 2, 141.1(1) in 5]. The geometry of the co-ordination sphere can be considered to be intermediate between octahedral and trigonal prismatic, though closer to the latter.Table 2 Proton and selected 31P NMR data * for the complexes [W(S]S)(CO)2L2] Complex 1H, d(J/Hz) 31P, d(JWP/ Hz) 1 7.5 (m, 18 H, m-, p-H of Ph), 7.8 (m, 12 H, o-H of Ph) — 2 1.1 (dt, 18 H, JPH = 15.4, JHH = 7.7, CH3), 1.8 (qnt, 12 H, JPH = JHH = 7.0, CH2) 23.0 (s, 87) 3 2.5 (br m, 2 H, CH2), 2.7 (br m, 2 H, CH2), 7.4 (m, 12 H, m-, p-H of Ph), 7.8 (m, 8 H, o-H of Ph) 32.3 (s, 58) 4 7.5 (m, 18 H, m-, p-H of Ph), 7.7 (m, 16 H, o-H of Ph, bdt) — 5 1.4 (dt, 18 H, JHH = 11.4, JPH = 5.6, CH3), 2.4 (qnt, 12 H, JPH = JHH = 7.1, CH2), 7.7 (m, 4 H, bdt) — 6 2.5 (br m, 4 H, CH2), 7.3 (m, 20 H, Ph), 7.7 (m, 4 H, bdt) — 7 7.4–7.7 (m, Ph) — 8 1.3 (dt, 18 H, JHH = 11.6, JPH = 5.8, CH3), 2.5 (qnt, 12 H, JPH = JHH = 7.1, CH2) 28.6 (s, 87) 9 2.4 (br m, 2 H, CH2), 2.5 (br m, 2 H, CH2), 7.4 (m, 20 H, Ph) — * Spectra run in CDCl3 (125 8C): s = singlet, br = broad, dt = doublet of triplets, m = multiplet, qnt = quintet.Distortions in six-co-ordinate dicarbonyl complexes of molybdenum(II) and tungsten(II) have recently been analysed. 20,21 Apart from the structures only slightly distorted from the ideal forms of octahedron and trigonal prism, there is a complete range of intermediate geometries characterised by the twist angle between the two triangular faces. This is 608 for the octahedron and 08 for the trigonal prism. These angles are 23, 238 in 2 and 208 in 5. These structures conform to the pattern of geometries noted in refs. 20 and 21. By contrast in the structure of complex 6 there is a bidentate phosphine ligand so that the phosphorus atoms are mutually cis subtending an angle of 76.2(1)8 at the metal atom. The two carbonyl groups are mutually cis but subtending an angle of 76.6(7)8 at the metal. According to the classification described above,20,21 when the angle between carbonyl groups is less than 908, then the structure will be octahedral, but that is not the case here. Indeed the structure is unique in that the angle between carbonyl groups is significantly less than 908 but yet the structure is significantly distorted towards the trigonal prism.The twist angle in this structure is only 128, so this structure is very close to being an ideal trigonal prism. The bond lengths in the three structures are as expected. In 2 and 5, tungsten–carbonyl distances are 1.94(2)–1.97(2), W]S Fig. 1 Structure of complex 2. Thermal ellipsoids are drawn at 40% probability. The hydrogen atoms are shown with small arbitrary radiiJ.Chem. Soc., Dalton Trans., 1997, Pages 1429–1433 1431 2.386(3)–2.405(3) in 2 and 2.383(3), 2.405(4) Å in 5 and W]P 2.458(3)–2.467(3) in 2 and 2.475(3), 2.483(3) Å in 5. There seems to be little difference between the effects therefore of C3S5 and bdt on the co-ordination sphere. In 6, while the geometry is different because of the bidentate phosphorus ligand, there are no major differences in dimensions with W]C 1.95(2), 1.99(2), W]S 2.369(4), 2.392(4) and W]P 2.472(5), 2.504(4) Å. The 1H NMR spectra of complexes 1–9 conform with the structures shown in Figs. 1–3.The 31P-{H} NMR spectra of the complexes [W(S]S)(CO)2L2] [S]S = C3S5, L2 = (PEt3)2 2 or dppe 3; S]S = mnt, L2 = (PEt3)2 8] show a singlet resonance with tungsten satellites, see Table 2. Complex 1 is six-co-ordinate, having lost the capping carbonyl observed in the original diiodo complex [WI2(CO)3- (PPh3)2]. Attempts to reintroduce a third carbonyl ligand by saturating a solution of complex 1 in dichloromethane with carbon monoxide were unsuccessful.In conclusion, we have successfully prepared and characterised the first examples of mono(dithiolene) complexes of tungsten( II), and are currently exploring their non-linear optical properties. Experimental The preparation and purification of complexes 1–9 were carried out under an atmosphere of dry nitrogen using Schlenk-line Fig. 2 Structure of complex 5.Details as in Fig. 1 Fig. 3 Structure of complex 6. Details as in Fig. 1 techniques. Ethanol was dried over magnesium–iodine and distilled prior to use. Dichloromethane was dried over P2O5 and distilled before use. The complexes [WI2(CO)3L2] [L2 = (PPh3)2,16 (PEt3)2 17 or dppe18], Na2[C3S5] 22 and Na2[mnt] 23 were prepared according to the literature methods, and all the chemicals were obtained from commercial sources. Elemental analyses (C, H and N) were recorded on a Carlo Erba Elemental Analyser MOD 1106 (using helium as the carrier gas) and 1H and 31P NMR spectra on a Bruker AC/250 spectrometer (1H referenced to tetramethylsilane, 31P to 85% H3PO4). Infrared spectra were recorded on a Perkin-Elmer 1600 FTIR spectrophotometer, the near-infrared spectrum on a Beckman DK-2A ratio-recording spectrophotometer. Magnetic susceptibilities were recorded on a Johnson-Matthey magnetic susceptibility balance.The FAB mass spectra were recorded on a VG-Autospec Instrument using Cs1 ions at 25 kV bombarded onto a sample dissolved in a 3-nitrobenzyl alcohol matrix target (EPSRC Mass Spec.Instr., Swansea). Table 3 Dimensions (distances in Å, angles in 8) in the metal coordination spheres of the three structures Complex 2 Molecule A Molecule B W(1)]C(200) 1.958(10) 1.967(11) W(1)]C(100) 1.971(12) 1.936(13) W(1)]S(2) 2.386(3) 2.397(3) W(1)]S(1) 2.397(3) 2.405(3) W(1)]P(3) 2.460(3) 2.467(3) W(1)]P(4) 2.466(3) 2.458(3) C(200)]W(1)]C(100) 106.5(4) 108.2(5) C(200)]W(1)]S(2) 155.4(3) 157.4(3) C(100)]W(1)]S(2) 87.5(3) 87.0(3) C(200)]W(1)]S(1) 90.0(3) 88.0(3) C(100)]W(1)]S(1) 155.2(3) 155.5(3) S(1)]W(1)]S(2) 84.26(10) 83.78(9) C(200)]W(1)]P(3) 78.4(3) 77.9(3) C(100)]W(1)]P(3) 79.6(3) 78.6(3) S(2)]W(1)]P(3) 124.79(10) 122.51(9) S(1)]W(1)]P(3) 86.06(9) 87.30(11) C(200)]W(1)]P(4) 78.4(3) 80.4(3) C(100)]W(1)]P(4) 79.0(3) 77.1(3) S(2)]W(1)]P(4) 84.79(11) 87.11(8) S(1)]W(1)]P(4) 123.22(10) 124.86(11) P(3)]W(1)]P(4) 142.28(9) 140.22(10) Complex 5 W]C(100) 1.969(11) W]S(2) 2.405(4) W]C(200) 2.01(2) W]P(3) 2.475(3) W]S(1) 2.383(3) W]P(4) 2.483(3) C(100)]W]C(200) 107.0(6) S(1)]W]P(3) 84.23(11) C(100)]W]S(1) 152.2(4) S(2)]W](P3) 126.52(13) C(200)]W]S(1) 90.6(4) C(100)]W](P4) 78.2(3) C(100)]W]S(2) 90.0(4) C(200)]W](P4) 79.3(4) C(200)]W]S(2) 153.4(4) S(1)]W](P4) 127.17(11) S(1)]W]S(2) 82.46(12) S(2)]W](P4) 84.44(12) C(100)]W]P(3) 78.9(3) P(3)]W](P4) 141.09(12) C(200)]W]P(3) 77.8(4) Complex 6 W]C(200) 1.95(2) W]S(2) 2.392(4) W]C(100) 1.99(2) W]P(3) 2.472(5) W]S(1) 2.369(4) W]P(4) 2.504(4) C(200)]W]C(100) 76.6(7) S(1)]W]P(3) 130.3(2) C(200)]W]S(1) 89.6(5) S(2)]W](P3) 84.9(2) C(100)]W]S(1) 146.2(5) C(200)]W](P4) 75.7(5) C(200)]W]S(2) 133.2(5) C(100)]W](P4) 113.5(4) C(100)]W]S(2) 84.6(4) S(1)]W](P4) 92.00(14) S(1)]W]S(2) 82.84(14) S(2)]W](P4) 150.3(2) C(200)]W]P(3) 131.3(5) P(3)]W](P4) 76.23(14) C(100)]W]P(3) 79.3(5)1432 J.Chem. Soc., Dalton Trans., 1997, Pages 1429–1433 Table 4 Crystal data and structure refinement for complexes 2, 5 and 6 2 5 6?0.5CH2Cl2 Empirical formula C17H30O2PS6W C20H34O2P2S2W C34.5H29ClO2P2S2W M 673.59 616.38 820.94 Crystal system Triclinic Triclinic Monoclinic Space group P1� P1� P21/n a/Å 11.117(9) 8.223(7) 8.047(8) b/Å 12.776(9) 11.102(7) 18.581(14) c/Å 17.901(14) 14.391(8) 23.070(14) a/8 91.95(1) 97.21(1) b/8 91.09(1) 101.11(1) 93.22(1) g/8 94.06(1) 98.47(1) U/Å3 2534 1259 3444 Z 4 2 4 Dc/Mg mm23 1.766 1.626 1.583 m/mm21 5.127 4.893 3.675 F(000) 1332 612 1620 Crystal size/mm 0.15 × 0.20 × 0.25 0.20 × 0.35 × 0.25 0.25 × 0.25 × 0.10 q Range for data collection/8 5.46–24.72 2.66–24.03 2.08–25.20 hkl Ranges 213 to 12, 214 to 14, 0–20to 9, 212 to 12, 0–15 0–9, 221 to 21, 227 to 27 Reflections measured 12 665 4725 9500 Independent reflection (Rint) 6499 (0.0303) 2873 (0.0293) 5461 (0.0700) Data, restraints, parameters 6499, 0, 501 2873, 0, 251 5461, 0, 388 Goodness of fit on F2 1.111 1.094 0.992 a, b in weighting scheme * 0.0448, 23.659 0.130, 3.598 0.158, 0.000 Final R1, wR2 indices [I > 2s(I)] 0.0451, 0.1085 0.0497, 0.1527 0.0838, 0.2073 (all data) 0.0568, 0.1158 0.0550, 0.1603 0.1471, 0.2458 Largest difference peak and hole/e Å23 1.052, 21.136 0.813, 21.824 2.464, 21.731 * w = 1/[s2(Fo 2) 1 (aP)2 1 bP], where P = (Fo 2 1 2Fc 2)/3.Preparations [W(C3S5)(CO)2(PPh3)2] 1. To sodium metal (0.032 g, 1.4 mmol) dissolved in dry ethanol was added 4,5-bis(benzoylthio)- 1,3-dithiole-2-thione (0.28 g, 0.3 mmol) with stirring under a stream of dinitrogen.The mixture was stirred under N2 for 45 min. The deep red solution was added dropwise with stirring to a solution of [WI2(CO)3(PPh3)2] (0.72 g, 0.7 mmol) in acetonitrile (10 cm3) and stirred for 16 h. The solvent was removed in vacuo and the dark, oily powder resolvated in dry, degassed CH2Cl2, filtered and the solvent removed in vacuo. This procedure was repeated several times to ensure removal of sodium iodide.The oily product was recrystallised from dry dichloromethane and hexane to give [W(C3S5)(CO)2(PPh3)2] 1 (yield = 0.52 g, 78%. See Table 1 for physical and analytical data. Complexes 2 and 3 were prepared in a similar manner. Analytically pure crystals suitable for X-ray crystallography were grown by resolvating 2 in CH2Cl2–hexane (4 : 1) at 0 8C. [W(bdt)(CO)2(PPh3)2] 4. To a stirring suspension of H2bdt (0.04 g, 0.3 mmol) in acetonitrile (5 cm3) was added dropwise a solution of [WI2(CO)3(PPh3)2] (0.29 g, 0.3 mmol) in acetonitrile (10 cm3).The reaction was stirred for 16 h, filtered and the solvent removed in vacuo. The product was washed with acetone and recrystallised from CH2Cl2–acetone to yield [W(bdt)(CO)2(PPh3)2] 4 (yield = 0.13 g, 37%). Complexes 5 and 6 were prepared in a similar way, and puri- fied by recrystallisation from acetonitrile. Analytically pure crystals suitable for X-ray crystallography were grown by resolvating 5 in CH2Cl2–hexane (4 : 1) and storing the solution at 0 8C.Analytically pure crystals suitable for X-ray crystallography of 6 were grown in a similar manner. [W(mnt)(CO)2(PPh3)2] 7. To a stirring solution of [WI2(CO)3(PPh3)2] (2.81 g, 2.7 mmol) in acetonitrile (10 cm3) was added a solution of Na2[mnt] (0.5 g, 2.7 mmol) in acetonitrile (20 cm3) dropwise under dinitrogen. The mixture was stirred for 4 h and the solvent removed in vacuo. The dark brown oily product was resolvated in dry dichloromethane, filtered and the solvent removed in vacuo.This was repeated several times. The product was recrystallised from acetonitrile to give [W(mnt)(CO)2(PPh3)2] 7 (yield = 1.59 g, 65%). Complexes 8 and 9 were prepared in a similar manner. X-Ray crystallography Crystal data for complexes 2, 5 and 6 are given in Table 4, together with refinement details. Data for all three crystals were collected at 293(2) K with Mo-Ka radiation (l 0.710 73 Å) using the MAR Research image-plate system.The crystals were positioned at 75 mm from the image plate. Ninety five frames were measured at 28 intervals with a counting time of 2 min. Data analysis was carried out with the XDS program.24 The three structures were solved by heavy-atom methods using SHELXS 86.25 In 2 there were two molecules in the asymmetric unit with equivalent dimensions. For all three structures the non-hydrogen atoms were refined anisotropically and hydrogen atoms, included in calculated positions, were refined isotropically. In 6 there was a disordered dichloromethane solvent molecule refined with 50% occupancy.One chlorine occupied two possible positions, each given 25% occupancy. The solvent atoms were given isotropic thermal parameters. In all three structures hydrogen atoms were positioned geometrically and given thermal factors of 1.2 times those of the atom to which they were bonded. Methyl groups were refined as rigid groups.Empirical absorption corrections were carried out.26 The final refinements were carried out on F2 using SHELXL.27 Atomic coordinates, thermal parameters, and bond lengths and angles have been deposited at the Cambridge Crystallographic Data Centre (CCDC). See Instructions for Authors, J. Chem. Soc., Dalton Trans., 1997, Issue 1. Any request to the CCDC for this material should quote the full literature citation and the reference number 186/429. Acknowledgements E.E. P. thanks the EPSRC for a studentship, N. R. for a postdoctoral fellowship and we thank the EPSRC and the University of Reading for funds for the image-plate system. We also thank the EPSRC Service at Swansea for FAB mass spectra.J. Chem. Soc., Dalton Trans., 1997, Pages 1429–1433 1433 References 1 P. Cassoux, L. Valade, H. Kobayashi, A. Kobayashi, R. A. Clark and A. E. Underhill, Coord. Chem. Rev., 1991, 110, 115. 2 A. T. Coomber, D. Beljonne, R. H. Friend, J.L. Bredas, A. Charlton, N. Robertson, A. E. Underhill, M. Kurmoo and P. Day, Nature (London), 1996, 380, 144. 3 C. S. Winter, S. N. Oliver, J. D. Rush, C. A. S. Hill and A. E. Underhill, J. Mater. Chem., 1992, 2, 443. 4 A. Kobayashi, R. Kato, A. Miyamoto, T. Naito, H. Kobayashi, R. A. Clark and A. E. Underhill, Chem. Lett., 1991, 2163. 5 M. L. H. Green and W. E. Lindsell, J. Chem. Soc. A, 1967, 1455. 6 R. Eisenberg, Prog. Inorg. Chem., 1970, 12, 295. 7 M. Fourmigué and C.Coulon, Adv. Mater., 1994, 6, 948. 8 M. Fourmigué, C. Lenoir, C. Coulon, F. Guyon and J. Amaudrut, Inorg. Chem., 1995, 34, 4979. 9 S. Inomata, H. Takano, K. Hiyama, H. Tobita and H. Ogino, Organometallics, 1995, 14, 2112. 10 K. Shimuzu, H. Ikehara, M. Kajitani, H. Ushijima, T. Akiyama and A. Sugimori, J. Electroanal. Chem. Interfacial Electrochem., 1995, 396, 465. 11 F. Guyon, M. Fourmigué, P. Audebert and J. Amaudrut, Inorg. Chim. Acta, 1995, 239, 117. 12 K. Y. Yang, R. L. Verran, S. G. Bott and M. G. Richmond, J. Coord. Chem., 1996, 38, 75. 13 G. M. T. Cheetham, M. M. Harding, J. L. Haggitt, D. M. P. Mingos and H. R. Powell, J. Chem. Soc., Chem. Commun., 1993, 1000. 14 N. G. Connelly, J. G. Crossley, A. G. Orpen and H. Salter, J. Organomet. Chem., 1994, 480, C12. 15 P. K. Baker, Adv. Organomet. Chem., 1996, 40, 45. 16 P. K. Baker, A. I. Clark, M. G. B. Drew, M. M. Meehan, E. E. Parker, R. L. Richards and A. E. Underhill, unpublished work. 17 P. K. Baker and S. G. Fraser, Inorg. Chim. Acta, 1986, 116, L1. 18 P. K. Baker and S. G. Fraser, Inorg. Chim. Acta, 1987, 130, 61. 19 C. W. Schlaepfer and K. Nakamoto, Inorg. Chem., 1975, 14, 1338. 20 T. E. Burrow, D. L. Hughes, A. J. Lough, M. J. Maguire, R. H. Morris and R. L. Richards, J. Chem. Soc., Dalton Trans., 1995, 1315. 21 P. Kubác¡ek and R. Hoffmann, J. Am. Chem. Soc., 1981, 103, 4320. 22 G. Steimecke, H. J. Sieler, R. Kirmse and E. Hayer, Phosphorus Sulfur, 1979, 7, 49. 23 A. Davison and R. H. Holm, Inorg. Synth., 1967, 10, 8. 24 W. Kabsch, J. Appl. Crystallogr., 1988, 21, 916. 25 SHELXS 86, G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467. 26 N. Walker and D. Stuart, Acta Crystallogr., Sect. A, 1983, 39, 158. 27 SHELXL, G. M. Sheldrick, Program for Crystal Structure Refinement, University of Göttingen, 1993. Received 28th November 1996; Paper 6/08056A

ISSN:1477-9226    

DOI:10.1039/a608056a

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

DALTON J. Chem. Soc., Dalton Trans., 1997, Pages 1435–1440 1435 Dihydroxotitanium(IV) phosphate: characterisation and studies of n-alkylamine intercalation, ion-exchange properties, and proton conduction Enrique Jaimez and Robert C. T. Slade * Department of Chemistry, University of Exeter, Stocker Road, Exeter EX4 4QD, UK The thermal behaviour and porosity of the dihydroxotitanium(IV) phosphate Ti(OH)2(HPO4)?H2O have been investigated. The intercalation of n-alkylamines, ion-exchange behaviour, and near-ambient-temperature proton conduction have also been studied and are discussed in comparison with the behaviour of layered crystalline titanium(IV) phosphates.Many studies on the synthesis and characterisation of metal(IV) phosphates with lamellar structures have been reported; structural features, physicochemical properties and applications of this rapidly growing class have been the subject of a number of reviews.1–4 Such compounds show a high stability against degradation and in suspension are cation exchangers.5 Their property of intercalation of polar organic molecules, ionic conductivity and potential as catalysts are well known.6–8 Layered titanium(IV) acid phosphates are usually synthesized from a preformed amorphous titanium phosphate, which is further treated with H3PO4(aq) either under reflux or hydrothermally.The crystalline compounds a-titanium phosphate [a-TiP, Ti(HPO4)2?H2O] and g-titanium phosphate [g-TiP, Ti(H2PO4)- (PO4)?2H2O] are well characterised. Titanium phosphates can also be prepared as gels and in intermediate stages of crystallinity, 9 the product obtained depending on the working conditions.10 Recent papers reported the synthesis of two novel hydroxyphosphates of titanium(IV).11,12 In this paper the chemistry of Ti(OH)2(HPO4)?H2O, Ti(OH)P, is further elaborated; the surface area and porosity, n-alkylamine intercalation chemistry, ion-exchange properties and proton conductivity have been investigated.Results and Discussion Characterisation Fig. 1 shows both TG and DTA traces for a sample of Ti(OH)P stored over anhydrous CaCl2(s). Dehydration takes place in three steps, the mass losses associated with the first two (Dm ª 18% at 25–380 8C) being very similar and markedly higher than the third (Dm ª 4% at 380–560 8C). The sample was stored in a dry atmosphere, and surface water loss (at T < 100 8C) was not detected. At 110 8C loss of crystal water occurs, with the formation of the anhydrous compound.At higher temperatures (200–380 8C) hydroxide groups condense, and at 380–560 8C the condensation of hydrogenphosphate groups takes place, giving the pyrophosphate a-TiP2O7. The DTA trace shows three endothermic peaks corresponding to stepwise dehydration and a first-order exothermic peak (ca. 750 8C, Dm = 0) associated with transformation of the pyrophosphate to its high-temperature (cubic) form. Fig. 2(a) shows the X-ray powder diffraction (XRD) profile of Ti(OH)P.A low-angle peak typical of layered materials is present and is assigned to a layer spacing d002 = 10.2 Å. After heating the material overnight at 100 8C the XRD profile of the solid is unchanged (no dehydration has occurred), but treatment at 160 8C overnight leads to dehydration and moves the first diffraction peak to higher angle/lower d spacing [Fig. 2(b)]. The interlayer distance is then d002 = 9.1 Å, which might be associated with anhydrous b-TiP 13,14 (b-TiP is the anhydride Fig. 1 The TG and DTA traces for Ti(OH)P Fig. 2 The XRD profiles for (a) Ti(OH)P, and for solids obtained after treatment at (b) 160 and (c) 900 8C. Phases present are identified as follows: TiP2O7 (*), TiO2 (rutile) (#) and Ti10O19 (s)1436 J. Chem. Soc., Dalton Trans., 1997, Pages 1435–1440 derived from g-TiP). The strongest peak (d = 3.42 Å) for b-TiP is, however, not present in the XRD profile of dehydrated Ti(OH)P.A sample of Ti(OH)P was calcined in air at 900 8C for 10 h. Such treatment of layered crystalline titanium phosphate phases (a- and g-TiP) yields pure titanium pyrophosphate (TiP2O7). The XRD profile of calcined Ti(OH)P, however, also shows peaks associated with titanium oxides [Fig. 2(c)], as expected for a solid with P/Ti < 2. Surface area and porosity Fig. 3 shows the nitrogen adsorption–desorption isotherm (T = 77 K) for Ti(OH)P outgassed at 100 8C. Samples outgassed at different temperatures showed similar isotherms, which indicates that the surface area and porosity (texture) do not vary significantly during the thermal treatment.Dehydration during thermal treatment at 160 8C does not, therefore, affect the texture of the material. The layer spacing in Ti(OH)P is d002 ª 10 Å. Taking into account that the thickness of each layer is ª5 Å (in most titanium phosphates), an interlayer gap of only 5 Å will be present. Isotherms of type IV of IUPAC classification,15 with hysteresis loops of type H3, were found in this work.Such isotherms are characteristic of non-microporous solids, with porosity in the mesoporous range and resulting from defective layer stacking and voids between aggregated particles. The analysis of such isotherms is well established, and a discussion of the methods employed is presented by Gregg and Sing.16 Fig. 4 shows the t-plot (a comparison of the empirical isotherm to the standard isotherm16) typical of samples treated at 160 8C; a straight line at low relative pressures extrapolates to the coordinate origin, consistent with the absence of microporosity. Table 1 compares the values of specific surface area calculated from Brunauer–Emmett–Teller (BET) and t-plot methods, as well as the values of the C parameter in the BET Fig. 3 Nitrogen adsorption–desorption isotherm of Ti(OH)P at 77 K (sample outgassed at 100 8C) Table 1 Specific surface area for Ti(OH)P calculated by BET and t-plot methods as a function of the outgassing temperature, T T/8C SBET/m2 g21 CBET St/m2 g21 25 86 80 87 100 91 102 92 160 90 94 89 model (the low C values reconfirm the absence of microporosity).The pore-size distribution for a sample outgassed at 100 8C can be seen in Fig. 5. The distributions obtained from the adsorption and desorption branches are similar. Both present a maximum at a pore radius, Rp = 74 Å (pore diameter, Dp = 148 Å). The isotherm shape (with a hysteresis loop of type H3, Fig. 3), together with the strong increase in slope at p/p0 > 0.90, suggest that the mesopores are full of liquid nitrogen at that pressure. The increasing adsorption at these higher pressures may be due either to the presence of some macroporosity or a swelling effect due to the nitrogen adsorption. An adequate geometrical model for the analysis of the mesoporosity in this material is not easy to define. The lamellar nature of the atomic-level structure could suggest a porosity with slit shapes, but other factors suggest otherwise.The agreements between (i) the pore-volume distributions obtained from sorption and desorption branches, (ii) cumulative (Fig. 5) and adsorbed volumes and (iii) cumulative and BET surface areas indicate that the cylindrical pore model is the most appropriate in analysis of this mesopore distribution. Fig. 4 The t-plot for Ti(OH)P treated at 100 8C Fig. 5 Cumulative and differential pore volumes of the Ti(OH)P sample treated at 100 8C (cylindrical model)J.Chem. Soc., Dalton Trans., 1997, Pages 1435–1440 1437 Table 2 Microanalytical data and experimental mass loss (TG) at 900 8C for the intercalation compounds, and those calculated from ideal formulae Experimental (%) Calculated (%) Formula Ti(OH)2(H0.5PO4)?0.5MeNH3?H2O Ti(OH)2(H0.5PO4)?0.5C2H5NH3?H2O Ti(OH)2(H0.5PO4)?0.5C3H7NH3?H2O Ti(OH)2(PO4)?C4H9NH3?H2O Ti(OH)2(PO4)?C6H13NH3?H2O Ti(OH)2(PO4)?C7H15NH3?H2O Ti(OH)2(PO4)?C8H17NH3?H2O C 2.55 5.6 8.2 17.55 24.4 27.75 29.9 N 3.4 3.25 3.15 5.35 4.8 4.75 4.3 Loss 25.67 27.27 29.69 43.47 51.02 52.30 54.42 C 2.85 5.5 8.4 17.85 24.25 27.0 29.55 N 3.3 3.2 3.05 5.6 4.7 4.50 4.3 Loss 28.60 30.90 33.20 43.88 49.17 51.46 53.55 n-Alkylamine intercalation The absorption of n-alkylamine vapours by layered phosphates gives rise to air-stable intercalation compounds. According to Casciola et al.17 the intercalation of n-alkylamines in layered phosphates may be considered as an acid–base topotactic solidstate reaction between a solid acid host and a Brönsted-base guest.The intercalation process in Ti(OH)P may be represented by equation (1). The protonation of amines can be demon- Ti(OH)2(HPO4)?H2O 1 RNH2 æÆ Ti(OH)2(HPO4)?xRNH2?yH2O (1) strated by applying spectroscopic techniques. Infrared spectra of the Ti(OH)P–n-butylamine intercalate and of the starting materials are compared in Fig. 6. That of the intercalate can be reproduced as the addition of those of Ti(OH)P and nbutylamine, but with an additional band at 1543 cm21 attributed to protonated amine groups, NH3 1.18 The infrared spectra of all n-alkylamine intercalation compounds are similar, with the presence of NH3 1 always being detected.Intercalation of n-alkylamines into a-TiP is significantly different from that into g-TiP, consequent upon the differing structural chemistries of the hosts. The cross-sectional area of a trans-trans alkyl chain is 18.6 Å2.19 If the free area for each acid centre is higher than that required by an amine molecule, steric factors do not prevent the material becoming saturated with adsorbate.The free area surrounding each phosphate group in a-TiP is 21.6 Å2,20 which permits the accommodation of one amine molecule (A) per phosphate (Ti :A = 1 : 2). A monolayer Fig. 6 Infrared spectra of (a) Ti(OH)P, (b) n-butylamine and (c) Ti(OH)P–n-butylamine intercalate is obtained when only one POH group of every two is engaged in a bond with the amine groups.The arrangement of the guests in intercalation compounds with Ti :A < 1 : 1 leads to the formation of a bilayer since there is insufficient space for the interpenetration of alkyl chains of amines bound to phosphate groups of neighbouring layers. In contrast, the free area of the acid centre (16.4 Å2) 21 in g-TiP is lower than in a-TiP, and steric hindrance should be encountered due to interaction of amines associated with adjacent active centres.Obtaining Ti :A = 1 : 2 in g-TiP is geometrically impossible. In n-alkylamine intercalation of Ti(OH)P the rate of reaction decreases with increasing alkyl chain length. A product of low crystallinity is obtained after 6 h with methylamine, but only after 30 d with octylamine. The XRD profiles show two different behaviours in the intercalation of Ti(OH)P. (i) When methyl-, ethyl- or propyl-amine are intercalated lamellar materials of low crystallinity are obtained after a few contact hours, but their analyses (Table 2) are consistent with incomplete reaction.Furthermore, if these solids are left in the corresponding amine vapour, they transform into amorphous gels and the percentage of the amine in the solid increases. These amines lead to progressive destruction of the layer structure, as opposed to simple intercalation. (ii) The intercalation compounds formed with higher n-alkylamines (pentylamine up to octylamine) are, in contrast, more crystalline and retain that crystallinity independent of contact time.Their analyses (Table 2) are consistent with straightforward intercalation. Table 2 lists formulations for the products obtained, as deduced from elemental (C,N) microanalysis and TG. Short-chain n-alkylamines (methyl up to propyl) give a maximum molar ratio Ti : amine (A) of 1 : 0.5 for initial products, but higher amine contents were obtained for amorphous gels.For higher amines (pentyl up to octyl) the maximum Ti :A ratio obtained was 1 : 1; the intercalation process then involves all acid centres in Ti(OH)P, as in layered a-TiP.22 For those amines giving intercalation compounds (pentyl up to octyl) the variation in the layer spacing, d002, of the intercalates with the length of the amine carbon chain (nc) is consistent with the hypothesis that intercalation takes place with the formation of a double layer of amine molecules in trans-trans conformation almost perpendicular to the Ti(OH)P layer.Table 3 lists d002 values for the intercalates, along with literature data 21,22 for analogous compounds obtained with a- or g-TiP as layered acid host. In this study d002 increases linearly with nc (Fig. 7), following equation (2). The n-alkylamine intercalation d002 = 12.5 1 2.20 nc (2) compounds of the a-MP family (M = Ti, Zr, Sn or Hf) show a similar behaviour.23–25 As the length of an alkyl chain in transtrans conformation increases by 1.27 Å for each additional carbon atom,6 and as the slope of the straight line defining the interlayer distances is higher than 1.27 Å, the amine must be present in the Ti(OH)P as a bimolecular layer.The average inclination angle of the molecules with respect to the sheet is then a = sin21 (2.20/2.54) = 60.08. This angle is similar to that reported previously for the n-alkylamine arrangement in a-TiP1438 J. Chem. Soc., Dalton Trans., 1997, Pages 1435–1440 (58.78),22 and lower than that obtained for intercalation in g-TiP (64.98).21 Ion-exchange capacity With the presence of HOPO3 (acid phosphate) groups in the structure of Ti(OH)P, this material should, by analogy with metal(IV) phosphates,1–6 exhibit cation-exchange properties in neutral and alkaline media.Ion exchange in a phases takes place by diffusion of the cations from the surface of the solids towards the bulk, with an advancing phase boundary (twophase behaviour). The g phases show lower steric hindrance to the diffusion of cations and the process occurs preferentially with a continuous expansion of the interlayer region.4 The theoretical ion exchange capacity of Ti(OH)P (calculated from its formula) is 5.10 mequivalents g21.Studies on the ion-exchange capacity of this solid were performed using model solutions of alkali-metal ions. The adsorption of Na1 and K1 in Ti(OH)P was studied by using a potentiometric titration method.The exchange isotherms and the titration curves as a function of the MOH added are given in Fig. 8. The absolute values for the alkali-metal cation uptake versus model solution acidity are presented in Fig. 9. Addition of MOH to suspensions of Ti(OH)P should, initially, increase the pH of the solution. In the process of ion exchange a certain number of hydroxyl groups (equal to that of the sodium ions exchanged) will be neutralised. If the reaction behaves ideally Fig. 7 Interlayer distance (d002) of intercalation compounds of Ti(OH)P with n-alkylamines [containing 1 mol of amine per mol of titanium(IV)] as a function of the number of carbons, nc, in the alkyl chain. h, Intercalation compounds of Ti(OH)P: d, incomplete reaction and progressive gel formation from Ti(OH)P Table 3 Layer spacing, d002/Å, for intercalated titanium phosphate hosts Intercalated amine Methylamine Ethylamine Propylamine Butylamine Hexylamine Heptylamine Octylamine a-TiP 13.1 14.3 16.9 18.8 23.1 25.6 27.5 g-TiP 13.5 16.2 18.4 20.5 24.6 —— Ti(OH)P 13.0 14.4 16.6 21.5 25.5 27.5 30.5 the equilibrium and initial pH values should be very similar.However, the analysis of M1 ions in solution indicates that a greater number of hydroxyl groups has been used up than the quantity of Na1 held by the exchanger. This is a result of hydrolysis reactions which, depending on the equilibrium pH, can take different forms. Analysis of the data in Fig. 9 shows alkali-metal uptake only at pH > 6, evident in a steep increase of slope in the adsorption curves at pH 6.5–7.0 for both alkali– metal cations.The potentiometric titration curves in Fig. 8 do not show any significant differences and have only one plateau (in the range pH 8–9), despite the large amount of alkali added. In the Na1–H1 ionic substitution the layer spacing d002 (determined by XRD) in the solid was little different after exchange. In the K1–H1 system, however, a second phase with d002 = 11.2 Å was detected during progressive ion exchange.A continuous expansion of the layer spacing was not observed, and the ion-exchange behaviour is, therefore, similar to that shown by a-layered materials (as opposed to g-layered materials). Electrochemical studies: impedance spectra and conductivity The conductivities of the acid phosphates and phosphonates of tetravalent metals are of contemporary interest.4,26 All have in common that the conductivity is strongly dependent on relative humidity and that the host acts as a Brönsted acid toward the water of hydration, which is generally loosely bound in the structure. According to Alberti et al.4 the conductivity of Fig. 8 Titration curves (s, Na1; h, K1) and ion-exchange isotherms (d Na1; j, K1) for Ti(OH)P Fig. 9 Adsorption capacity as a function of equilibrium pH (potentiometric titration with NaOH) for Ti(OH)PJ. Chem. Soc., Dalton Trans., 1997, Pages 1435–1440 1439 layered phosphates at near-ambient temperatures is dominated by surface transport.Proton NMR relaxation studies demonstrated the absence of any translational motion of protonic species through the bulk.27 The impedance plane plots obtained for Ti(OH)P samples consist of a single depressed semicircle (due to the bulk electrolyte resistance) with a low-frequency tail (due to electrode– pellet interfacial impedance). The sample resistance was determined by extrapolation of the high-frequency arc to the real axis, and converted into conductivity (s) via sample geometry.In the temperature range 20–90 8C plots of log10(sT) versus 1/T are linear (Fig. 10) and can be interpreted by using an empirical Arrhenius-type equation [sT = A exp (Ea/RT)]. The Ea and A values are listed in Table 4 for 0–100% relative humidity. The activation energy increases with decreasing relative humidity (and level of hydration). At constant temperature, observed conductivities are some two to three orders of magnitude greater when measurements are carried out at 100 than at 0% relative humidity. Water molecules, therefore, play an important role in the conduction process. The conductivity is also seen to increase by two orders of magnitude when the temperature is increased [at 0% relative humidity, s(25 8C) = 9.7 × 1027 and s(90 8C) = 1.7 × 1025 S cm21; at 100%, s(25 8C) = 1.2 × 1025 and s(90 8C) = 1.1 × 1024 S cm21].Attempts to improve the proton conductivity of layered phosphates and phosphonates by preparing new compounds with Brönsted bases intercalated in the interlayer region or in which functionalised organic radicals replace the hydroxyl of the phosphate groups have met some success.28–30 The a.c. conductivity data for the n-alkylamine intercalates of Ti(OH)P are shown in Fig. 11 and Table 5.The conductivity values recorded for these intercalates are lower than that of the host material, as has been reported in layered phosphates (a and g) for this family of guest molecules.31 The low conductivity of the n- Fig. 10 Conductivity for Ti(OH)P as a function of temperature and of relative humidity [0 (d), 40 (m), 60 (.), 80 (r), 100% (j)] Table 4 Arrhenius parameters for the a.c. conductivity of Ti(OH)P at different relative humidities Relative humidity (%) 0 40 60 80 100 Ea/kJ mol21 52 38 32 29 31 log10(A/S K cm21) 3.23 3.63 2.87 2.73 3.11 alkylamine intercalation compounds is a direct consequence of the basic strength of the guest molecule, the protons being localised on the amine groups (as shown by IR spectroscopy, Fig. 6). Experimental Synthesis All reagents were of analytical grade (Aldrich) used without further purification. A solution of TiCl4 (1 mol dm23, 80 cm3) in HCl(aq) (2 mol dm23) was mixed with H2O2 (30% v/v, 10 cm3) to obtain an intensely coloured titanium complex [Ti(O2)- (OH)(H2O)x]1(aq). Phosphoric acid, H3PO4(aq) (1 mol dm23, 25 cm3), was added. The resulting solution was refluxed at 60 8C in air for 6 d. The white precipitate obtained was filtered off and thoroughly washed with deionised water until the rinses had pH > 3.The solid was air dried at room temperature for 2 d, and then stored at room temperature under different relative humidities, controlled by H2SO4(aq) solutions of different densities (r = 1.00, 1.20, 1.30, 1.40 and 1.50 g cm23 for 100, 80, 60, 40 and 20% relative humidity, respectively) or by CaCl2(s) (0% relative humidity).Instrumental Thermal analyses (TG and DTA) were performed with a Stanton Redcroft STA-781 instrument (heating rate 10 8C min21 in air). X-Ray powder diffraction (XRD) profiles were recorded with a modified Philips PW 1050 diffractometer and Ni-filtered Cu-Ka radiation (l = 1.5418 Å). Infrared spectra were recorded on a Nicolet Magna 550 spectometer, using the KBr disc method. Sample surface area was investigated via adsorption–desorption isotherms (N2 at 77 K) recorded using a Micromeritics Gemini instrument; the vapour pressure of the liquid-nitrogen bath was measured hourly and isotherms were Fig 11 Arrhenius plots of conductivities of Ti(OH)P and its n- CH3(CH2)nNH2 intercalates [n = 3(r), 5(m), 6(d), 7(j) at 0% relative humidity] Table 5 The a.c.conductivity at 25 8C and empirical activation energy (Ea) for Ti(OH)P and its n-alkylamine intercalates at 0% relative humidity Compound Ti(OH)2(HPO4)?H2O Ti(OH)2(H0.5PO4)?0.5C2H5NH3?H2O Ti(OH)2(PO4)?C4H9NH3?H2O Ti(OH)2(PO4)?C6H13NH3?H2O Ti(OH)2(PO4)?C7H15NH3?H2O Ti(OH)2(PO4)?C8H17NH3?H2O Ea/kJ mol21 52 44 44 42 37 54 s/S cm21 9.7 × 1027 6.3 × 1029 6.6 × 1029 2.6 × 1029 1.0 × 1029 1.1 × 10291440 J.Chem. Soc., Dalton Trans., 1997, Pages 1435–1440 repeated to ensure reproducibility. Sample outgassing at 25, 100 or 160 8C was carried out for 5 h using a Micromeritics Flow- Prep 060. Intercalation Intercalation compounds containing n-alkylamines [CH3- (CH2)nNH3, 0 < n < 7] were obtained by placing Ti(OH)P (0.4 g) in an atmosphere saturated with the appropriate n-alkylamine vapour at room temperature for 1–60 d.The intercalated samples were air dried at 50 8C for 12 h, and stored in desiccators at different controlled relative humidities (as above). Ion exchange Adsorption of alkali-metal ions on Ti(OH)P employed model aqueous 0.05 mol dm23 MCl–MOH (M = K or Na) solutions at V:m = 200 : 1 (cm3 : g) at room temperature.In all cases the contact time was 4 d and solutions were periodically shaken. The pH of the model solutions after equilibration with the adsorbent was measured using a Metrohm 713 pH meter. Residual concentrations of ions in solution [Na1(aq), K1(aq)] were measured using an atomic absorption spectrometer (Varian SpectraA 300). Conductivity measurements Electrical impedance measurements utilised a Solartron 1260 impedance analyser and 1270 electrochemical interface programmed via an IBM-compatible computer for data collection and analysis.Impedance spectra were recorded at 10 8C intervals in the range 20 < T/8C < 90, using frequencies 5 Hz–13 MHz and an oscillating voltage of 300 mV. Pellets (13 mm diameter, 0.7–1.2 mm thick) were prepared by pressing 200 mg of material at 40 kN cm22 in a pellet die. The two flat surfaces were painted with conductive silver paint (Electrodag 915, Acheson Colloids) to give ionically irreversible electrodes.The pellet was placed between two copper foils and the sample was spring loaded to ensure good electrode–electrolyte contact. The assembly was placed inside a cell maintained at controlled relative humidity by H2SO4(aq) or CaCl2(s) (see above). The cell was allowed to equilibrate for at least 60 min at each temperature before measurements were taken (this time being found by experience to be much greater than the minimum necessary for temporal stability in impedance spectra).Experiments were repeated at least three times on different samples to ensure reproducibility, and average values are given. Acknowledgements We thank the Commission of the European Communities for funding this work under BRITE/EURAM II programme (contract BRE2-CT93–0535) and for a personal bursary for E. J. (contract BRE2-CT94–3096). References 1 A. Clearfield, in Inorganic Ion Exchange Materials, ed. A. Clearfield, CRC Press, Boca Raton, FL, 1982, p. 1. 2 G. Alberti and U.Costantino, in Inclusion Compounds, eds. J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Oxford University Press, Oxford, 1991, vol. 5, pp. 136–173. 3 A. Clearfield, Chem. Rev., 1988, 88, 125. 4 G. Alberti, M. Casciola, U. Costantino and R. Vivani, Adv. Mater., 1996, 8, 291. 5 G. Alberti, in Recent Developments in Ion Exchange, eds. P. A. Williams and M. J. Hudson, Elsevier Applied Science, London, 1987, p. 223. 6 J. R. García, R. Llavona, M. Suárez and J. Rodríguez, Trends Inorg.Chem., 1993, 3, 209. 7 P. Maireles-Torres, A. Jimenez-Lopez, P. Olveira-Pastor, I. Rodríguez-Ramos, A. Guerrero and J. L. García-Fierro, J. Catal., 1992, 92, 81. 8 D. Bianchi and M. Casciola, Solid State Ionics, 1985, 17, 7. 9 G. Alberti, U. Costantino and M. L. L. Giovagnotti, J. Inorg. Nucl. Chem., 1979, 41, 643. 10 M. Suárez, J. R. García and J. Rodríguez, Mater. Chem. Phys., 1983, 8, 451. 11 Y. J. Li and M. S Whittingham, Solid State Ionics., 1993, 63, 391. 12 A.Bortun, E. Jaimez, J. R. García and J. Rodríguez, Mater. Res. Bull., 1995, 30, 413. 13 E. Kobayashi, Bull. Chem. Soc. Jpn., 1975, 48, 3114. 14 R. Llavona, J. R. García, M Suárez and J. Rodríguez, Thermochim. Acta, 1985, 86, 81. 15 K. S. W. Sing, C. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierotti, J. Rouquerol and T. Simieniewska, Pure Appl. Chem., 1985, 57, 603. 16 S. J. Gregg and K. S. W. Sing, in Adsorption, Surface Area and Porosity, Academic Press, London 1982, chs. 2, 3. 17 M. Casciola, U. Costantino, L. di Croce and F. Marmottini, J. Inclusion Phenom., 1988, 6, 291. 18 L. J. Bellami, in The Infrared Spectra of Complex Molecules, Chapman & Hall, London, 1975. 19 A. I. Kitaigorodsky, in Molecular Crystals and Molecules, Academic Press, New York, 1973. 20 A. N. Christensen, E. K. Andersen, I. G. K. Andersen, G. Alberti, N. Nielsen and M. S. Lehmann, Acta Chem. Scand., 1990, 44, 865. 21 A. Menéndez, M. Bárcena, E. Jaimez, J. R. García and J. Rodríguez, Chem. Mater., 1995, 5, 1078. 22 F. Menéndez, A. Espina, C. Trobajo and J. Rodríguez, Mater. Res. Bull., 1990, 25, 1531. 23 R. M. Tindwa, D. K. Ellis, G. Z. Peng and A. Clearfield, J. Chem. Soc., Faraday Trans. 1, 1983, 545. 24 E. Rodríguez-Castellón, S. Bruque and A. Rodríguez-García, J. Chem. Soc., Dalton Trans., 1985, 213. 25 M. L. Rodríguez, M. Súarez, J. R. García and J. Rodríguez, Solid State Ionics, 1993, 63, 448. 26 K. D. Kreuer, Chem. Mater., 1996, 8, 610. 27 R. C. T. Slade, C. R. M. Forano, A. Peraio and G. Alberti, Solid State Ionics, 1993, 61, 23. 28 M. Casciola, S. Chieli and U. Costantino, Solid State Ionics, 1991, 46, 53. 29 G. Alberti, L. Boccali, M. Casciola, L. Massinelli and E. Montoneri, Solid State Ionics, 1996, 84, 97. 30 E. W. Stein, A. Clearfield and M. A. Subramanian, Solid State Ionics, 1996, 83, 113. 31 M. Casciola, U. Costantino and F. Marmottini, Solid State Ionics, 1989, 35, 67. Received 7th November 1996; Paper 6/07591F

ISSN:1477-9226    

DOI:10.1039/a607591f

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

DALTON J. Chem. Soc., Dalton Trans., 1997, Pages 1441–1445 1441 Molecular-orbital study of a quasi-linear Ru2Mo trinuclear compound with a diamidolene ligand across each metal–metal linkage ‡ Carlo Mealli,*,†,a Andrea Ienco,a Adela Anillo,b Santiago Garcìa-Granda c and Ricardo Obeso-Rosete b a Istituto per lo Studio della Stereochimica ed Energetica dei Composti di Coordinazione (ISSECC-CNR), Via J. Nardi 39, 50132 Firenze, Italy b Departamento de Quimica Organica y Inorganica, Instituto Universitaro de Quimica Organica ‘Enrique Moles’, Universidad de Oviedo, E-33071 Oviedo, Spain c Departamento de Quimica Fisìca y Analitica, Facultad de Julian Claveria, Universidad de Oviedo, s/n-33071 Oviedo, Spain The electronic structure of the trinuclear compound [(Ph3P)(OC)2Ru{m-C6H4(NH)2}Mo(CO)2{m-C6H4(NH)2}- Ru(CO)2(PPh3)], with a diamidolene ligand riding upright over each of the quasi-collinear Ru]Mo linkages, has been investigated by means of the extended-Hückel method and a graphic interface.The ability of the dianionic riders to donate eight electrons to adjacent metals is interpreted, similarly for related Ru2 dimers, in terms of perturbation theory. The molecule’s distortion, from the highest possible C2v to C2 symmetry, was parametrized and the effects on the overall chemical bonding evaluated. The nature of the extended Ru]Mo]Ru linkage is addressed in some detail. The electronic redistribution over the ruthenium and molybdenum atoms is discussed in terms of their limiting oxidation states.Inferences are made as to the structure of a chromium analogue of the MoRu2 compound, as yet unknown. The bidentate diamide [C6H4(NH)2]22 is able to donate four electron pairs when riding upright on binuclear fragments of the type L6M2 (sawhorse). Some of us 1 have previously reported the synthesis and the crystal structure of the dimeric species [(Ph3P)(OC)2Ru{m-C6H4(NH)2-1,2}Ru(CO)2(PPh3)] 1. A derivative with a m-Ph2PCH2PPh2 ligand in place of two carbonyls has also been characterized.2 Similar compounds exist with conjugated diphosphidolene 3 and dithiolene 4 ligands and, usually, the replacement of N with the better donating S or P atoms results in an elongation of the M]M bond.By contrast, the 1,2-dioxolene anions (catecholates) do not seem to support the bridging co-ordination mode which implies the donation of four electron pairs to two metals. A theoretical study of these dimers has just been completed 5 to determine the electronic factors affecting the stereochemistry of the bridge and the strength of the M]M bond as well as to find a rationale for the different behaviour of dioxolenes.During synthetic studies not only the mononuclear five-coordinated precursor [Ru{C6H4(NH)2-1,2}(PPh3)3] was treated with [Ru(CO)3(PPh3)3] to produce the dimer 1, but also its reactivity toward [M(CO)6] (M = Cr, Mo or W) was tested.6 In particular, a novel trinuclear species [Ru2Mo{m-C6H4(NH)2}2- (CO)6(PPh3)2], 2, was synthesized and characterized crystallographically.The overall molecule is significantly distorted from † E-Mail: mealli@cacao.issecc.fi.cnr.it ‡ Non-SI units employed : eV ª 1.60 × 10219 J, cal = 4.184 J. the highest possible C2v symmetry and the three metals do not lie in a perfectly linear arrangement (Ru]Mo]Ru 1678). However, over each metal–metal linkage of ca. 2.70 Å one diamidolene ligand roughly maintains the same upright bridging mode as in the binuclear Ru2 species 1 (a minor 108 tilting of the planar C6N2 skeleton about the N? ? ? N vector and toward the central Mo atom is observed).To a first approximation, since the lateral moieties (Ph3P)(OC)2Ru[(HN)2C6H4)] are almost superimposable in compounds 1 and 2, one could imagine the trimer as arising from the dimer upon replacement of one terminal unit (Ph3P)- (OC)2Ru with the bulkier binuclear fragment (Ph3P)(OC)2Ru- {m-C6H4(NH)2}Mo(CO)2.As mentioned, such an assembly does not necessarily imply that the symmetry planes of the component fragments are maintained in the final complex. Ultimately, the trimer possesses at best a C2 axis bisecting the central (OC)2Mo unit (other significant details of the stereochemistry will be illustrated subsequently). Quite interestingly, a search in the Cambridge Structural Database 7 shows only one other trinuclear species which is electronically and stereochemically comparable with 2, namely the complex [(OC)3Fe(m-ButS)2- Mo(CO)2(m-ButS)2Fe(CO)3] 3.8 In the latter, two independent ButS groups lie almost at the same positions as the nitrogen donors of the chelate diamidolene.Although the less constrained bridges seem to allow a larger molecular flexibility (in particular, the trimetallic unit is more bent with the angle1442 J. Chem. Soc., Dalton Trans., 1997, Pages 1441–1445 Fe]Mo]Fe being 1608), the two structures 2 and 3 are quite comparable.This is confirmed by very fine details such as the non-linearity of the two carbonyl ligands terminally bound to the central molybdenum atom. The ca. 108 bending of the angle Mo]C]O and the fact that both the C and O atoms lie in the same plane as the Mo]Ru vector suggest an incipient semibridging nature. In other words, each Ru or Fe atom appears to exert residual co-ordination ability toward one CO ligand which acts mainly as a two-electron donor toward molybdenum.Intuitively, the deformational trends observed in these structures are meant to improve the strength of the overall metal– ligand and perhaps intermetal bonding. In this paper, we try to interpret for compound 2 how the relative orientation of the two bridging diamidolenes (between themselves and with respect to the L3Ru and L2Mo units) can influence the molecular orbital (MO) picture. From the latter we derive significant pieces of information on the nature of the intermetal bonding itself.In this task we exploit mainly the qualitative concepts of perturbation theory, as they can be developed from the results of extended-Hückel molecular orbital (EHMO) calculations 9 and interpreted with the graphic interface of the package CACAO.10 Results and Discussion A model of the Ru2Mo compound with the highest possible C2v symmetry is shown in I. As mentioned it is related to the dimer 1, since one terminal L3M fragment of the latter is formally replaced by the bulkier fragment Mo(CO)2[m-C6H4(NH)2]RuL3.In the starting geometry I, the Ru]Mo]Ru angle is 1808 and the two L3Ru fragments, as well as the two parallel diamidolene bridges, are mutually eclipsed whereas the (OC)2Mo plane bisects the two molecular halves. The co-ordination geometry at the Mo atom adapts to the trigonal prism highlighted by dashed lines in I. We have found that it is possible to convert this geometry into that of the experimental C2 molecule via a number of combined parametrizations.The most drastic ones are the screw motions a (in opposite directions and up to a maximum of 358) of the two fragments L3Ru[m–C6H4(NH)2] about the corresponding Mo]Ru axes which deviate from collinearity by ca. 138. Moreover, since the vector (hence, the plane) Ru2Mo reorients with respect to the fixed (OC)2Mo plane (the dihedral angle b ranges between 90 and 658), the screw motion of the two Ru-centred fragments (of the type L5M) is best described as that of two anti-propellers (working via combined a and b rotations).Finally, the ca. 108 bending of the planar bridges about the N? ? ? N axes and to the side of the central molybdenum atom is also mimicked. It is noteworthy that, upon the structural rearrangements, the co-ordination at the molybdenum changes from trigonal prismatic to pseudo-octahedral as highlighted in sketch II. Here, the plane of the drawing coincides approximately with the octahedral equatorial plane at the Mo atom: the latter is defined by the two CO ligands (slightly bent with angles Mo]C]O of ca. 1708) and two N atoms of different diamidolenes. The other two N donors are roughly axial as the angle N]Mo]N is not larger than 1538. It is also noteworthy that the Mo]Ru vectors, although non-collinear, pierce two opposite faces of the pseudo-octahedron at molybdenum (each face is defined by the two N atoms of the same chelate and by one CO ligand). The non-linearity of the Ru]Mo]Ru group is of importance to fix the nature of the intermetal bonding network (see below).By combining at subsequent steps all of the motions described above, the interconversion from the ideal (I) into the real system (II) can be monitored in terms of a global deformational co-ordinate. Indeed, it is strategically important to elucidate first the electronic structure of the most symmetric model (C2v) with the various MOs subgrouped in a larger number of orthogonal classes.Subsequently, it becomes easier to follow the topological rearrangement and the mixing of the various orbitals when the symmetry is lowered to C2. Before proceeding to the MO analysis, it is necessary to summarize some electronic features of the dimers of type 1.5 Their understanding is simplified by referring to the classic compounds M2L10 (M = Mn or Re, L = CO) in which two square pyramids are glued together at their bases. Hence, two pairs of eclipsing equatorial carbon monoxides can be replaced by one bridging diamidolene, provided that the residual fragment L6M2 adapts to the sawhorse shown in III.The latter with C2v symmetry is characterized by two sets of three hybrids at each metal atom (these are the typical frontier orbitals of L3M fragments which lie above the t2g set 11). The MO analysis 5 has shown that four of the latter hybrids (considered acceptors) are engaged in bonding–antibonding interactions with symmetry combinations (a1, b1, a2 and b2) of two s and two p levels of the riding dianion.However, one of the four dative bonds, of type b1, is critical. The p-conjugated bridge can barely use a low-lying p-bonding level as a donor (1b1), whereas the involvement of the better energy-located highest occupied molecular orbital (HOMO) (2b1, IV) is complicated by its nodal properties. In fact, the out-of-phase metal s hybrids directed toward the diamidolene ligand (V) are bonding toward the lateral pp orbitals of the nitrogen atoms and are, at the same time, antibonding toward the adjacent and central carbon orbitals.Ultimately, the co-operation of ligand 1b1 and 2b1 levels inJ. Chem. Soc., Dalton Trans., 1997, Pages 1441–1445 1443 the donation of electrons guarantees the fourth bonding interaction between the metals and the bridge. The latter interaction is definitely improved in the presence of donors with diffuse orbitals (diphosphidolene or dithiolene ligands), whereas the contraction and the high electronegativity of the oxygen heteroatoms prevents catecholates from riding on sawhorses.5 Another point which stems from the MO analysis of the dimers 5 is that the two in-plane combinations of the metal hybrids III, not involved in bridge bonding, are responsible for the Ru]Ru single bond.The latter hybrids are typical of squarepyramidal metal fragments and, while their s* combination (b1) lies empty at high energy, the s partner (1a1, VI) is filled for the d7–d7 configuration.The Ru2Mo compound, where one L3Ru fragment is replaced by the more complex unit L3Ru[m-C6H4(NH)2]Mo(CO)2, is also characterized by terminal square pyramids at the Ru. The corresponding s hybrids can now interact with opportune molybdenum orbitals. Thus, it is important to specify these combinations and to ascertain their role. One viewpoint considers ruthenium(II) species with acidic character localized at the empty s hybrids. In this case the neutral molybdenum (d6) could utilize two different lone pairs (filled t2g levels at the octahedral centre) to make dative bonds with the empty ruthenium s hybrids (see below).The opposite viewpoint implies localization of electron density at the two Ru atoms. The ruthenium(0) species, fully saturated in a squarepyramidal environment, would be exceptionally strong s nucleophiles, while the central molybdenum (MoIV) would be the acceptor of two additional metal lone pairs (beside the six from the ligands).The trigonal-prismatic co-ordination of Mo would suit better in this case, but it is clearly avoided. An intermediate hypothesis implies a double electron pairing between a molybdenum(II) species and the ruthenium(I) ones which also characterize the dimer 1. Generally, a mononuclear, six-co-ordinated d4 complex distorts from octahedral to keep diamagnetic, hence avoiding population of the triply degenerate t2g set with four electrons (Jahn–Teller effect).11 Also the structure of the six-co-ordinated monomer [Mo{C6H4(NH)2- 1,2}(CO)2(PPh3)2],12 containing the same set of ligands as that of compound 2, is significantly distorted. While the electronic effects in monomers11 can hardly be extended to polynuclear species which present additional M]M interactions, it is true that neither of compounds 2 or 3 achieves a regular octahedral geometry at the central metal atom.In any case, it should not be forgotten that these structures are highly strained.In assigning the most probable oxidation states of the metals, the trends in the n(C]] ] O) data are also a good source of information. For example the IR wavenumbers of the CO ligands bound to molybdenum are larger for the d4 monomer (1913 and 1826 cm21) 13 than for compound 2 (1853 and 1795 cm21).6 The latter data suggest a less electron-rich Mo atom in the monomer in spite of the strong basicity of the phosphine ligands and of the reduced efficiency of riding diamidolenes when they need to donate four electron pairs.Further support for a more electronrich Mo atom in 2 comes from a comparison between the wavenumbers of the carbonyls bound to the Ru atoms in compound 2 (2022, 2008 and 1955 cm21) 6 and in the dimer 1 (2002, 1968 and 1931 cm21).1 The last values become even smaller for the dimeric derivative with dppm (1972 and ca. 1910 cm21),2 thus confirming the idea of a higher basicity associated with the phosphines when bound to molybdenum.In conclusion, the experimental IR data suggest that the Ru atoms of 2 are more oxidized than the ruthenium(I) species present in the dimer. Significantly, the numerical EHMO results attribute to the Ru atoms a charge definitely more positive in 2 than in the dimer 1 (0.81 and 0.18, respectively). In the preceding paragraphs the possible distributions of the integer oxidation states at the metals have been illustrated. As is often the case in chemistry, limiting models of chemical bonding rarely depict the actual situation but serve as important reference points for the interpretation of the MO calculations.By anticipating here one of the conclusions, the Mo]Ru bonds are better described as weak dative bonds from two nonbonding levels at the central Mo atom (formally d6) into empty acidic hybrids of the Ru atoms (also formally d6). The evolution of the MO levels during the interconversion from the ideal C2v molecule into the experimental one (C2) was followed in a Walsh diagram (not shown).The latter was constructed with the parameters defined in I. Not only the initially insufficient HOMO–LUMO (lowest unoccupied molecular orbital) gap of ca. 0.8 eV increases to >2.0 eV but also the total energy of the system is lowered by ca. 1.2 eV along the pathway: the gain is essentially due to some filled MOs centred at the molybdenum atom. This is not surprising as, in the starting trigonal-prismatic environment, the orbitals xz and yz (antibonding to the ligands) are destabilized with respect to orbital xy (non-bonding).14 Thus for an electron count >d2 rearrangement toward an octahedron is progressively favoured because also xz and yz become prevailingly non-bonding.In the present case, the distortion away from the trigonal prism at molybdenum (see I) reduces the importance of the MoIV/Ru0 combination. Indeed, the pictorial MO analysis supports better the idea of centrifugal rather than centripetal electron flow, consistent with the combination RuII/Mo0. The fragment molecular orbital analysis 15 of the interactions between the central (OC)2Mo unit and the rest of the molecule (the overall model is close to the experimental C2 structure) reveals significantly large overlap populations between the in- and out-of-phase combinations of the ruthenium s hybrids and the molybdenum xz and yz types of orbitals.The drawings in VII present the latter interactions at the antibonding level (the LUMOs of the system).Importantly, the xz and yz orbitals of molybdenum (the x axis is across the page) lie ca. 2 eV deeper in energy than any ruthenium frontier s hybrid (strongly destabilized by the trans axial ligand). According to perturbation theory,16 the two bonding electron pairs are preferentially assigned to molybdenum (i.e. four of its six t2g electrons with the third electron pair occupying the non-bonding xy orbital).Looking at the right-hand part of VII (the central Mo atom lies deeper than the two Ru atoms), it is inferred that a number of structural factors improve the overlap between the two ruthenium s hybrids (pointing toward the observer) and the upper lobes of the xz orbital. Essentially, there is a misalignment of the two trans phosphine ligands (the Ru]P and the Ru]Mo bonds form angles of ca. 1488) and a skewing of the1444 J. Chem. Soc., Dalton Trans., 1997, Pages 1441–1445 RuL6 octahedra (PRu ? ? ?RuP torsion angle ca. 708). Also important is the 138 bending of the Ru]Mo]Ru angle (in the MoFe2 species 3 the bending is as much as 208). Ultimately the distortive trends should favour the overall s(Ru)–xz(Mo)– s(Ru) interaction as the s(Ru) lobes are pulled out of the xz nodal plane. In MO terms the A]B stick bonds of a linear AB2 assembly ensue from interactions which are symmetric and asymmetric with respect to the orthogonal mirror plane (see VIII).By assuming that the central Mo atom uses mainly the centrosymmetric d orbitals (given the six-co-ordination, its atomic orbital pz is largely engaged with the ligands), the linearity of the extended Ru]Mo]Ru linkage could be consistent only with a three-centre two-electron model. Conversely, as shown in IX (the view is orthogonal to that in VII), the Ru]Mo]Ru bending (as well as the reorientation of the ruthenium s hybrids) involves a second filled d orbital (xz) in what becomes an overall four-orbitals four-electron interaction.In conclusion, the structural details and the MO arguments offer jointly a rationale for the intermetal bonding and the preferential oxidation states of the metals. As seen in structures 2, 3 and II, the two CO ligands, terminally bound to molybdenum, are slightly bent so as to adapt to an incipient bridge-bonding mode in spite of the correspondingly long Ru]C separations of ca. 2.70 Å. The situation compares closely to that of the binuclear complex [(OC)3Mo- (m-SC6H4S)Mo(CO)4]22 17 containing the dithiolate ligand as a rider between the fragments (OC)3Mo and (OC)4Mo (see X).One terminal CO ligand of the latter is also bent by ca. 108 and one Mo]C distance is significantly but not dramatically shorter than the other (2.01 and 2.59 Å, respectively). Interestingly, when the two Mo atoms are replaced by two Cr atoms the seventh CO ligand shifts to the fully bridging position, e.g.[(OC)3Cr(m-SC6Cl4CS)(m-CO)Cr(CO)3]22.18 Concerning the potential CrRu2 analogue of MoRu2 compound 2, an alternative structure could be hypothesized. Each Ru]Cr bond could be bridged by one carbonyl ligand beside the diamidolene itself. While different strategies for the synthesis of the chromium derivative are still being elaborated,19 a modelling of the interconversion between the two possible tautomers shows a very flat energy surface. The fluctuations are even smaller than those calculated for the dimeric tautomers (ca. 4 kcal mol21). Moreover, the small HOMO–LUMO gap of only a few tenths of an eV is by itself an indication of the instability of the alternative structure. Simple considerations of electron counting, more than the always questionable quantitative response of the EHMO method, are helpful. One of us (C. M.) has previously proposed, for polynuclear compounds, a simple method which provides the number of M]M bonds (]] m) and the number of free lone pairs at the metal atoms (]] n).20 A knowledge of the latter is important as their repulsion affects the strength of a given M]M linkage.For the CO-bridged and unbridged trinuclear compounds the values of m and n are calculated to be 2 and 7 and 2 and 5, respectively.§ While the number of M]M bonds is that predicted by the effective atomic number rule,21 the seven lone pairs of the trimer with all-terminal CO ligands are consistent with the previous MO considerations.Recalling, that two of the three t2g levels at the central metal atom were considered involved in dative, centrifugal M]M bonding, only the unused lone pair (xy) adds to the double set of t2g nonbonding levels of the Ru atoms. Were the carbonyls bridging, one t2g lone pair from each Ru atom would be involved in socalled back bonding to the carbonyl p* levels, hence the prediction of five metal lone pairs. In the interpretation of the overall MO picture these guidelines are certainly useful.Even though M]M repulsions are lessened in the CO-bridged system, the nature of the direct intermetal M]M bonding is significantly changed. The difficulty in separating the M]M and M](m- CO)]M characters of various MOs recalls the controversial situation observed in [Fe2(CO)9] where the through-bridge coupling seems to overwhelm any direct Fe]Fe interaction.22 In this respect, the drawings of the low-lying LUMOs (XI) are highly indicative. The latter correlate, along the unbridged– bridged interconversion pathway, with the CrRu2 antibonding levels VII. It is evident that an increasing in-phase coupling between the metal orbitals and the CO p* orbital stabilizes these empty levels in the bridged structure (hence the small HOMO–LUMO gap).Indeed, the overall CrRu2 antibonding character of the LUMOs is subtly balanced by the Ru(m-CO)Cr- (m-CO)Ru bridge-bonding character and, eventually, the lack of electrons in these levels may become a serious problem for the stability of the whole system.In conclusion, qualitative considerations seem to favour the unbridged structure 2 also for the CrRu2 trinuclear compound. § A system of two equations is written for the case with terminal CO groups (system A) and for the case with bridging CO groups (system B). The two sets of equations, to be solved for m and n, involve the total number of metal orbitals (9 × 3 = 27) available for the V = 3 metal atoms, the total number of metal–ligand bonds (L = 16 and 18, respectively) and the total valence electron count (T = 50 in each case): system A, 2m 1 n = 9V 2 L = 27 2 16 = 11, 2m 1 2n = T 2 2L = 50 2 32 = 18; system B, 2m 1 n = 9V 2 L = 27 2 18 = 9, 2m 1 2n = T 2 2L = 50 2 36 = 14.J.Chem. Soc., Dalton Trans., 1997, Pages 1441–1445 1445 On the other hand, first-row transition metals are known to favour bridging CO because their d orbitals, too contracted, do not allow sufficient direct M]M interaction at long separations.Quantitatively, the EHMO response cannot be considered reliable and any definite answer to the question requires more sophisticated calculations or, much better, the synthesis and characterization of the species in question. Computational Details All the MO calculations were of the extended Hückel type 9,23 using a weighted-modified Wolfsberg–Helmholz formula.24 The literature Slater atomic orbital parameters were used for Ru,25 Mo,26 and Cr,26 and the standard ones for the main-group elements.The three-dimensional drawings and correlation and/or interaction diagrams were constructed with the program CACAO.10 In general, the structural models were drawn to approximate, at the very best, the geometries of the reported crystal structures. The CACAO input files are available from one of the authors (C. M.) on request. References 1 S. Garcìa-Granda, R. Obeso-Rosete, J. M. Rubio and A. Anillo, Acta Crystallogr., Sect.C, 1990, 46, 2043. 2 A. Anillo, R. Obeso-Rosete, M. A. Pellinghelli and A. Tiripicchio, J. Chem. Soc., Dalton Trans., 1991, 2019. 3 See, for example, M. D. Soucek, C. C. Clubb, E. P. Kyba, D. S. Price, V. G. Scheuler, H. O. Aldaz-Palacios and R. E. Davis, Organometallics, 1994, 13, 1120. 4 See, for example, H. P. Weber and R. F. Bryan, J. Chem. Soc. A, 1967, 182; D. Touchard, J.-L. Fillaut, P. Dixneuf, C. Mealli, M. Sabat and L. Toupet, Organometallics, 1985, 4, 1684. 5 C. Mealli, A. Ienco, A. Anillo, S. Garcìa-Granda and R. Obeso-Rosete, unpublished work. 6 A. Anillo, S. Garcìa-Granda, R. Obeso-Rosete and J. M. Rubio- Gonzalez, J. Chem. Soc., Dalton Trans., 1993, 3287. 7 Cambridge Structural Database System, version 5.11, Cambridge Crystallographic Data Centre, Cambridge, 1995. 8 S. Lu, N. Okura, T. Yoshida and S. Otsuka, J. Am. Chem. Soc., 1983, 105, 7470. 9 R. Hoffmann, J. Chem. Phys., 1963, 39, 1397; R. Hoffmann and W. N. Lipscomb, J. Chem. Phys., 1962, 36, 2179, 2872. 10 C. Mealli and D. M. Proserpio, J. Chem. Educ., 1990, 67, 399. 11 T. A. Albright, J. K. Burdett and M. H. Whangbo, Orbital Interactions in Chemistry, Wiley, New York, 1985. 12 P. Kubácek and R. Hoffmann, J. Am. Chem. Soc., 1981, 103, 4320. 13 A. Anillo, R. Obeso-Rosete, M. Lanfranchi and A. Tiripicchio, J. Organomet. Chem., 1993, 453, 71. 14 R. Hoffmann, J. M. Howell and A. Rossi, J. Am. Chem. Soc., 1976, 98, 2484. 15 R. Hoffmann, H. Fujimoto, J. R. Swenson and C. C. Wan, J. Am. Chem. Soc., 1973, 95, 7644; R. Hoffmann and H. Fujimoto, J. Phys. Chem., 1974, 78, 1167. 16 R. Hoffmann, Acc. Chem. Res., 1971, 4, 1. 17 B. Zhuang, L. Huang, L. He and J. Lu, Inorg. Chim. Acta, 1989, 160, 229. 18 D. Sellmann, M. Wille and F. Knoch, Inorg. Chem., 1993, 32, 2534. 19 R. Obeso-Rosete and A. Anillo, unpublished work. 20 C. Mealli and D. M. Proserpio, J. Am. Chem. Soc., 1990, 112, 5484; C. Mealli, J. A. Lopez, S. Yan and M. J. Calhorda, Inorg. Chim. Acta, 1993, 213, 199. 21 See, for instance, D. M. P. Mingos and D. J. Wales, in Introduction to Cluster Chemistry, Prentice-Hall, Englewood Cliffs, NJ, 1990. 22 C. Mealli and D. M. Proserpio, J. Organomet. Chem., 1990, 386, 203; J. Reinhold, E. Hunstock and C. Mealli, New J. Chem., 1994, 18, 465. 23 R. Hoffmann and W. N. Lipscomb, J. Chem. Phys., 1962, 37, 3489. 24 J. H. Ammeter, H.-B. Bürgi, J. C. Thibeault and R. Hoffmann, J. Am. Chem. Soc., 1978, 100, 3686. 25 D. L. Thorn and R. Hoffmann, Inorg. Chem., 1978, 17, 126. 26 R. H. Summerville and R. Hoffmann, J. Am. Chem. Soc., 1976, 98, 7240. Received 21st October 1996; Paper 6/07191K

ISSN:1477-9226    

DOI:10.1039/a607191k

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

DALTON J. Chem. Soc., Dalton Trans., 1997, Pages 1447–1453 1447 Covalence and spin polarisation in tetraphenylarsonium tetrachloronitridotechnetate(VI) studied by polarised neutron diVraction† Philip A. Reynolds,*,a Brian N. Figgis,b J. Bruce Forsyth c and Francis Tasset d a Research School of Chemistry, Australian National University, Canberra, ACT 0200, Australia b Department of Chemistry, University of Western Australia, Nedlands, WA 6907, Australia c ISIS facility, Rutherford-Appleton Laboratory, Chilton, Didcot, Oxon.OX11 0QX, UK d Institut Laue-Langevin, Avenue des Martyrs, Grenoble 38042, France Polarised neutron diffraction from two orientations of tetragonal single crystals of [AsPh4][TcNCl4] at 1.5 K and a magnetic field of 4.6 T gave a combined data set of 116 magnetic structure factors. These were fitted by a 26- parameter model which includes proton nuclear-spin orientation to give a spin-density model for the formally 4d1 [TcNCl4]2 ion. This model revealed substantial anisotropic p bonding between Tc and Cl resulting in 29(2)% of the spin being delocalised onto the chlorine in-plane 3pp orbitals, 215(1)% onto the nitrogen as a result of spinpolarising the short Tc]N bond, and substantial spin polarisation on the technetium site of 4d and more diffuse density.Three unconstrained Hartree–Fock and density functional calculations with good basis sets failed to reproduce these observations adequately, but an unconstrained density functional calculation with gradient corrections and a relativistic treatment of the core gave encouraging agreement with experiment.Studies of bonding in metal complexes have predominantly been based on spectroscopic techniques. Since magnetic properties arise mainly from the valence electrons they are very sensitive to changes in bonding. Spectroscopic techniques examining magnetic behaviour, such as ESR, are particularly informative. On the other hand, diffraction study of the magnetic properties by polarised neutron diffraction (PND) probes the spatial rather than energetic aspects of the wavefunction and highlights different features of the bonding.This has proved useful for complexes of the first transition series in defining the balance of factors such as covalence, electron correlation, and the influence of spin–orbit coupling, in both ionic and covalent systems.1 We have begun PND studies of complexes involving transition metals of the second and third series, thus far including [Ru(acac)3] 2, (acac = acetylacetonate) tetrachlorobis(Nphenylacetamidino) rhenium,3 and a molybdenum(III) alum.4 The first two have interesting properties which are much modi- fied by intermolecular exchange, while the latter is quite ionic in nature.In the present context, we searched for a highly covalent complex composed of magnetically isolated units, so that we might test molecular theories of bonding in heavier metal systems.In particular, these heavier systems are expected to be more covalent than those of the first transition series, but simultaneously it is well recognised that the results of ab-initio calculations are less reliable. The salt [AsPh4][TcNCl4] is, as we shall see, a more suitable complex for this particular purpose than others have been. First, previous work is reviewed to show why we suppose that this is so. We then introduce the effects of proton nuclear-spin polarisation, a complication which occurs in this material, but the effects of which can be well accounted for in the experiment.Our new PND results are then presented, some new theoretical ab-initio calculations are reported, and we compare them. † Supplementary data available (No. SUP 57216, 4 pp.): magnetic structure factors. See Instructions for Authors, J. Chem. Soc., Dalton Trans., 1997, Issue 1. Non-SI units employed: mB ª 9.27 × 10224 J T21, au ª 4.36 × 10218 J. Previous work on [TcNCl4]2 salts (a) Crystal structure of [AsPh4][TcNCl4] This crystal has been examined by single-crystal X-ray diffraction to obtain structures at 293,5 at 120 K,6 and by neutron diffraction at 28 K.6 There are no phase changes, and the crystal remains tetragonal at all temperatures with a cell a = 12.55(1), c = 7.70(1) Å at 28 K.The cation and anion both exhibit a fourfold symmetry axis, C4 and S4 respectively. The tetraphenylarsonium ion is, as expected, of propeller-like conformation, while the [TcNCl4]2 ion is a square pyramid with a Tc]N distance of 1.625(4) Å and a Tc]Cl distance of 2.335(3) Å with an N]Tc]Cl angle of 103.3(1)8.The cation and anion are linked by H? ? ? Cl contacts. The unit-cell contents are illustrated in Fig. 1, which is a projection down the four-fold axis of this tetragonal crystal. We note that the Tc]N bond is parallel to this axis. (b) Bulk magnetic properties of [AsPh4][TcNCl4] Single-crystal susceptibility and magnetisation measurements have been reported down to a temperature of 4.5 K and up to a magnetic field of 5 T.6 The behaviour is very simple: isotropic with a g factor of 2.00, and a small antiferromagnetic exchange term as defined by the Curie–Weiss plot with q 20.13(5) K.(c) Electron spin resonance on [AsPh4][TcNCl4] Extensive ESR, electron–nuclear double resonance, and electron spin-echo envelope modulation experiments have been performed and interpreted.5,7–9 Both hyperfine and nuclear quadrupole couplings have been extracted from the data.(d) PND on [AsPh4][TcNCl4] Six magnetic structure factors were obtained previously from a small crystal mounted with the c axis perpendicular to the basal plane of the diffractometer.10 In this orientation the spin on Tc and N is not distinguishable, and so a simple one-parameter model with spin on Tc 1 N, and on Cl ligands, was fitted to these data.1448 J. Chem. Soc., Dalton Trans., 1997, Pages 1447–1453 (e) Electronic theory The electronic structures of square-pyramidal MXY4 n2 ions have been extensively studied theoretically by ab-initio methods, and the fundamental qualitative features of the bonding are now thought to be understood.Many of the calculations do not incorporate electron–electron correlation, such as may be introduced by, for example, using configuration interaction or unrestricted methods. We do not pursue these uncorrelated restricted calculations further, since, as will be seen, they cannot even reproduce the qualitative features that are observed in this experiment.There are two unrestricted calculations on the [TcNCl4]2 ion, one a neglect of diatomic differential overlap Hartree–Fock treatment 8 and one using a local density functional method.10 More fundamentally, existing ab-initio calculational packages cannot extract spin densities from multideterminant wavefunctions, so only the unrestricted methods are immediately available for comparison.( f ) Current understanding of the bonding The small magnetic exchange in the crystal suggests that a good model is to consider the crystal wavefunction as constructed essentially from isolated [TcNCl4]2 and [AsPh4]1 units. However, because of electrostatic and other effects these, at least in principle, differ from free-ion wavefunctions. The ab-initio calculations indicate that the naive view of a very strong Tc]N and a weaker Tc]Cl bond is reasonable.Given the large spin–orbit coupling constant for Tc61, ca. 1700 cm21, we would not expect a simple isotropic g tensor of value 2.00 to arise from such a ligand field. Indeed in related molecules there is both anisotropy and departure from 2.00 for the diagonal elements of the g tensor. In this case the isotropy, at least down to 4.5 K, seems to be a fortuitous result of the various covalences and spin–orbit coupling of Tc and Cl.5 However we should be aware that at lower temperatures and high magnetic fields departures from a simple spin ��� Brillouin function may occur in the bulk magnetisation.Electron spin resonance spectroscopy and theory both agree in locating the bulk of the magnetisation in the technetium 4dxy orbital in the ab plane. Both theory and ESR hyperfine coupling values indicate that this is also covalently delocalised onto gand ab plane chlorine 3p orbital. The axes are defined below in the section describing the modelling. A simple inter- Fig. 1 Unit cell of [AsPh4][TcNCl4] in ab projection pretation of the ESR results suggests 20% of the spin is on the chloride ligands. The ESR nuclear quadrupole coupling, and theory, suggest a substantial spin polarisation of the Tc]N bond, such that negative spin resides on the N atom and matching extra positive spin around the technetium atom. While convincing quantitative data may be lacking, the agreement of simple molecular orbital ideas of bonding, more sophisticated ab-initio methods, and the ESR results might lead us to believe that that is essentially a complete description.Thus it was a surprise when the previous PND experiment apparently gave 45(6)% of the spin on the chlorines, more than twice the ESR value. Evidently the situation is more complex than was believed! The new PND results below confirm this. PND on Systems Containing Polarised Proton Nuclear Spins When proton nuclear spins are aligned along the applied magnetic field the effect on the PND experiment is straightforward.Extra interference terms between nuclear spin polarisation and both the nuclear and magnetic structure factors appear in the scattered intensity.11 The nuclear spin–magnetic interference term is small and we shall neglect it. The nuclear spin–nuclear structure factor term appears in the effective magnetic structure factor as an apparent magnetic contribution of 5.402Pp/ cos2 W mB, from each proton site, with a thermally smeared delta function form factor; P is the nuclear polarisation at the proton site, p the neutron polarisation, and W the angle between the normal to the neutron scattering plane and the vertical direction (in the usual experimental configuration).This nuclear spin effect on the PND experiment is due to the neutron–proton nuclear interaction being very different in the singlet and triplet states. A much more common manifestation of this is the very large proton incoherent scattering cross-section compared to other elements.Thus, even if other nuclei are polarised in this crystal to comparable amounts, the scattering change which results is much smaller. The incoherent scattering cross-section for technetium is not known, but is small compared to that for the proton. Nuclear scattering differs from magnetic scattering in that, because it is a contact term, there is no dependence on the angle between the neutron momentum transfer and the direction of polarisation.In the magnetic case there is a strong dependence of the observed flipping ratio on the lifting angle of the diffractometer. This makes magnetic and polarised nuclear scattering experimentally separable. However we do not present this separation, as when it was attempted the data were insuffi- ciently extensive to give reliable results. There is also an extra truly magnetic nuclear polarisation term of 1.52 × 1023Pp mB due to the magnetic scattering from the nuclear magnetic moment, but here this is negligible.This nuclear polarisation effect has not often been observed, but there are two classic experiments which show that the theory is indeed straightforward to apply. Abragam et al. 12 observed it in LiH by brute-force methods, producing proton nuclear spin polarisation of 0.183% (and lithium-6 polarisation of 0.036%, and 7Li of 0.118%) by applying a magnetic field of 2.07 T at a temperature of 1.15 K.We note a typographical error in this paper of a factor of 10 in the quoted polarisations, though this is not carried further through the paper. We have checked all polarisations by appropriate calculations from nuclear moments, and S = ��� , 1 and 3– 2 Brillouin functions. Secondly Leslie et al.13 performed PND experiments on a neodymium-doped lanthanum magnesium nitrate hydrate crystal in which the water protons were oriented by microwave pumping and use of the Overhauser effect.They succeeded in locating protons in the unit cell using the observed flipping ratios. More recently, various PND data on hydrogencontaining materials have required correction, and the subsequent successful refinement of these data indicates the accuracy of this correction process, e.g. ref. 14.J. Chem. Soc., Dalton Trans., 1997, Pages 1447–1453 1449 Experimental A deep orange-red single crystal (60 mg) of [AsPh4][TcNCl4], of prismatic habit, of length 10 mm along c, and ca. 1.5 × 1.5 mm in cross-section, from an earlier preparation and growth,5 was mounted on the D3 lifting-arm polarised neutron diffractometer at the Institut Laue-Langevin in Grenoble. Two sets of measurements were made, one with [1 2 10] and one with [001] parallel to within 28 to the vertical applied magnetic field. The restrictions imposed by the lifting-arm geometry mean that with a single-crystal orientation we can reach reflections up to about 258 out of the diffractometer basal plane.The cell dimensions found [a = 12.55(6), c = 7.69(5) Å] are in excellent agreement with the 28 K result [a = 12.55(1), c = 7.70(1) Å], although less precise due to the relaxed resolution required for efficient PND data collection. The neutron wavelength obtained from the Heussler alloy monochromator was 84.3(1) pm, the polarisation ratios were p1 = 0.965(1) and p2 = 0.974(1), and the flipping efficiency was 1.000(1). An erbium filter was used to reduce l/2 contamination to 0.14(2)%.The flipping ratios, R, of 290 and 270 reflections in the two orientations were measured at an applied magnetic field of 4.61 T at a sample temperature of 1.5 K. Each unique reflection was generally measured at three equivalent positions. The reflections measured were chosen as those with large nuclear structure Bragg intensities, giving relatively unbiased sampling of the magnetic structure factors in the experimentally accessible region of reciprocal space.We assume the magnetic and nuclear structure factors are not closely correlated. The flipping ratios were converted into magnetic structure factors using the neutron diffraction structure from Figgis et al.,6 assuming the collinear magnetisation expected for this magnetically isotropic system. If polarised nuclear scattering is present the derived ‘magnetic structure factors’ are effective numbers, rather than the z components of the magnetic structure factor. The relevant expression is (1) where R is the observed flipping ratio for the FM = FN([p(R 1 1)/(R 2 1)] ± {[p(R 1 1)2/(R 2 1)2] 2 (1/sin2 W)}� �� ) (1) reflection and FN and FM are the nuclear and magnetic structure factors respectively. The data were symmetry averaged to 92 and 68 unique reflections in each orientation.These had 44 reflections in common, which were averaged to a final set of 116 reflections. When polarised protons are present, equivalent reflections at different experimental lifting angles are no longer exactly equivalent, both within and between the two data sets.However test calculations showed that this effect is distinctly smaller than the reduction in error to be made by such averaging. The data error estimates were derived from counting statistics, differences between equivalents, and an estimate of the error in FN. The final data range in (sin q)/l from 0.07 to 0.78 Å21, with Ss(FM)/ SFM = 0.17, and s(FM)/FM as low as 0.03. Results and Refinements We fitted the observed effective magnetic structure factors by a model including both the polarised nuclear and magnetic scattering using a modified version of the program ASRED.15 Polarised nuclear scattering was considered only from the protons.The magnetic scattering has been fitted by multipole models for the magnetisation density. The initial model was given substantial flexibility so that the total magnetisation density could be fitted as well as possible, at the expense of possible large correlations between the fitting parameters.The second model, based on parameters seeming important from subsequent theoretical considerations, with fewer parameters, allows, at the expense of a marginally poorer fit, discussion in chemical terms of the major features contributing to the m density. For the radial dependence of the atomic contributions to the magnetisation density we have used Hartree–Fock solutions 16 for the 4d and 5s orbitals of the 7S state of 4d55s1 Tc1, 3p for 2P Cl0, 2p for 4S N0, and the hydrogen 1s valence function of Stewart et al.17 In addition, to provide sufficient flexibility in the fitting, it was necessary to add a contracted 4d function on the technetium site, 4d9, with radial functions r4d9 of 0.7 r4d.No angular dependence was used on the hydrogen sites, four multipoles on the Cl (all to order 2 assuming mm symmetry), four on N (all to order 3 using S4 crystal site symmetry), and six each on the 4d and 5s technetium function radial dependences (all to order 4 in the C4 crystal site symmetry), and five angular functions for the contracted 4d function on Tc (all the d-like dependences).All the five independent hydrogen sites were constrained to have equal 1s populations and nuclear polarisations. Many of the multipoles have been recast into linear combinations with simple chemical significance without loss of generality. Examples are those representing the dxy and sp hybrid orbitals.This enables us to judge the importance of those populations which simple chemical models predict as being important. Some of the multipoles are not present in these simple chemical schemes and have been retained as unaltered multipoles. The axis systems are: Tc, 1z points along the fourfold axis towards N, 1x perpendicular to the four-fold axis along the projection of the Tc]Cl bond on the ab plane (thus 4dxy is in the ab plane with lobes pointing between the Tc]Cl bonds); N, 1z towards Tc along the four-fold axis [thus (2sp)1 is an sp hybrid with its major positive lobe pointing towards Tc while (2sp)2 points away from Tc]; Cl, 1x along the Cl]Tc bond and 1y in the ab plane perpendicular to this [thus (3sp)1 points its major lobe towards Tc and 3py is a p bonding 3p in the ab plane]. This gives a total of 26 parameters used in the least-squares fit, and thus an observable : parameter ratio of 4.5 : 1.In addition the total magnetisation per unit cell was constrained to 2.00 mB, viz. magnetic saturation for g = 2.00.6 The resulting fit gave the spin-population parameters of Table 1 with agreement factors of R(F) = 0.083, Rw(F) = 0.024 and goodness of fit, c2, of 0.69. The fitted value of the proton nuclear polarisation is 0.278(14)%. A list of observed and calculated magnetic structure factors has been deposited as SUP 57216. In Figs. 2 and 3 we show contour maps of the total magnetisation density in a Cl]Tc]Cl plane and a Cl]Tc]N plane.These were constructed by Fourier summation to a (sin q)/l limit of 1.2 Å21. We tried variation of the model in a number of ways, but none gave significant improvement. They included (1) making all the hydrogen sites independent, (2) increasing the number of multipoles on N and Cl to four and/or allowing radial flexibility by use of a k parameter, (3) introducing mid-bond Tc]N and Table 1 Values of the fitted population parameters for the multipole model (units e).Multipoles are labelled (nm) Atom 4d 4d9 * Tc 4dxy 1.04(15) 20.19(15) 4dxz 1 yz 20.38(18) 0.40(18) 4dx2 1 y2 0.26(15) 20.18(15) 4dz2 0.28(20) 20.24(19) (10) 20.08(2) (30) 20.02(5) 5s (00) 20.21(7) (10) 0.00(4) (20) 0.23(7) (30) 0.09(5) (40) 20.26(8) (44) 0.05(11) Cl (3sp)1 20.017(7) (3sp)2 20.004(4) 3pz 0.021(4) 3py 0.070(5) N (2sp)1 20.06(2) (2sp)2 20.01(1) 2pp 20.06(2) (30) 20.01(3) H 1s 0.002(2) * A function with contracted radius; r4d9 = 0.7 r4d.1450 J.Chem. Soc., Dalton Trans., 1997, Pages 1447–1453 Tc]Cl density functions, (4) introducing an extra hydrogen contracted radial function, (5) introducing populations on carbon and arsenic sites, and (6) independent refinement of the two data sets, i.e. assuming the magnetisation density differs between the two crystal orientations. In our simplified ‘chemical’ model we have been guided by the theoretical results given below. These show no significant 5s or 5p spin components, a polarised very diffuse 4d density, 4d*, while the more contracted 4d density remains unpolarised and atom-like in radial distribution.Thus we removed both the contracted 4d density and the 5sp hybrids on Tc from our model and replaced them by a very diffuse 4d* density function, while retaining the atom-like technetium 4d function. The diffuse 4d* density, starting from the same radial distribution as the theoretical 5s, was allowed to change in radius with a k parameter.We retained both hydrogen parameters, the polarised proton scattering and the 1s term. Also retained was a total of five independent p orbitals on N and Cl, and the (10) multipole on Tc. This gave a total of 18 parameters and a resulting fit of R(F) = 0.092 and c2 = 0.76. The resulting parameters are summarised in Table 2 and given in detail in Table 3. It is clear that this chemical model encompasses the most important features although the more complete model is marginally better in fit.Theoretical Calculations In the theoretical treatment, a convenient benchmark is the scheme of Neuhaus et al.18 who calculated the properties of many oxo- and nitrido-halogeno complexes of Mo, W, Re and Os. We have used exactly their ab-initio scheme to calculate the wavefunction of the [TcNCl4]2 ion. This involves a restricted open-shell Hartree–Fock (ROHF) calculation using the quasirelativistic effective core potential of Hay and Wadt 19 with split-valence technetium basis set 441/2111/31 of Jonas et al. 20 and 3-21G(d) for Cl and 6-31G(d) for N.21 Using GAMESS22 we optimised the geometry in C4v symmetry, giving a Tc]N bond length of 1.557 78 Å [experimental = 1.625(4) Å], Tc]Cl of 2.398 84 Å [experimental = 2.335(3) Å], an N]Tc]Cl bond angle of 102.1658 [experimental = 103.3(1)8] and an energy of 21963.558 78 au. This reproduces the experience of Neuhaus et al. in that the metal]halide bond is calculated too short and the M]N or M]O bond too long.This ROHF calculation cannot, since spin resides in a single molecular orbital, give negative Fig. 2 Modelled spin density in a TcCl2 plane of the TcNCl4 fragment. The positive contours start at 0.01 mB Å23, increasing geometrically by 2� �� for each contour, negative at 20.01 decreasing similarly spin density on the nitrogen atom or anywhere else. However, the latter is one of the main observations of our experiment. Accordingly, as a minimum-level calculation with any hope of agreement with experiment, we performed an unconstrained Hartree–Fock (UHF) calculation, using the experimental molecular geometry.This gave an energy of 21963.550 88 au and some of the spin populations are listed in Table 2. Optimising the geometry gave results very close to those from the ROHF calculation. We also performed a further Hartree–Fock calculation as different as we could from the previous one. This was an allelectron UHF calculation using the MIDI Gaussian basis set with two d-function polarisation functions on Cl and N [exponents 1.028 and 0.257 (Cl) and 1.728 and 0.432 (N) au],23 and gave a total energy of 26073.9642 au at the experimental geometry.We have also quoted in Table 2 the results of the local density functional (LDF) calculation of Figgis et al.,10 which was a non-relativistic, all-electron, calculation with no gradient corrections. The basis set is derived from the numeric solutions of the free ions, and is thus inflexible in the molecular calculation.As an improvement on this we performed an unconstrained calculation at the experimental geometry using the Amsterdam density functional package 24 with the VWN local density approximation, Becke–Perdew gradient corrections, and frozen relativistic atomic cores. The Slater basis is roughly of triplezeta flexibility. We also optimised the geometry and obtained a Tc]N distance of 1.6485 Å, Tc]Cl of 2.370 Å, and an N]Tc]Cl angle of 103.48.These values are more than twice as close to experiment as those calculated by the ROHF method, and, given the possible compressible effects of crystal-packing forces, are perhaps as close as may be expected. Spin popul are listed in Tables 2 and 3. The three 4d-basis function populations have been collapsed to an atom-like 4d and a diffuse 4d*. We calculated an ‘expected’ diffuse atom-like component from the molecular populations of the two contracted d functions and the ratio of the three in the free atom.The difference of this ‘expected’ atom-like population from the actually calculated diffused population we have called the 4d* popu- Fig. 3 Modelled spin density in a TcNCl2 plane of the TcNCl4 fragment. Contours as Fig. 2J. Chem. Soc., Dalton Trans., 1997, Pages 1447–1453 1451 Table 2 Summary of spin populations given by the reduced chemical model fitting of the experiment (units e) and the results from the ab-initio calculations Theory HF HF LDF LDF10 Non-relativistic Relativistic Non-relativistic Relativistic Atom Experiment all-electron pseudo-potential all-electron pseudo-potential Tc 4dxy 0.62(3) 0.88 0.88 0.63 0.64 other 4d 0.13(3) 0.08 0.09 0.28 0.15 5s/p — 0.01 0.00 0.03 0.03 Cl 0.079(6) 0.018 0.020 0.074 0.072 N 20.133(17) 20.051 20.049 20.244 20.124 H 0.005(2) — — — — lation.The latter is thus a measure of the change in shape of the molecular 4d spin at diffuse radii from the free-atom spin shape.It is thus comparable, though not identically defined, to the diffuse function in the fitting of the experimental data. Discussion Magnetisation distribution First we discuss the total magnetisation density, which in this case, can be identified as the spin density since we have an isotropic g tensor of 2.00. From Fig. 2 the majority of the spin density on the TcNCl4 fragment lies in a 4dxy-like distribution on the Tc atom, as a simple ionic crystal-field model predicts.However, in addition, there are clear features due to covalence and spin polarisation, an electron-correlation effect. In Fig. 2 there is noticeable spin density delocalised onto the chlorine inplane 3p orbital. There is little covalence on the chlorine 3ps or out-of-plane 3pp orbitals. In addition in Fig. 3 we see very substantial negative spin density on the nitrogen and positive density on the Tc atom in areas not populated by the 4dxy density.This is clear evidence of spin polarisation of Tc]N bonding orbitals. Lastly, the maps show a distinctly negative area in the Tc]Cl bond in the overlap region. Thus the qualitative interpretation of the maps already requires both covalence and electron correlation to be highly significant. For a more quantitative interpretation we turn to the results of the refinement, which in turn implies a partitioning of spin density between atomic centres.The partitioning is not unique, but in this case the low correlation coefficients between differently atom-centred populations resulting from the relatively contracted radial distributions used in the fit make it quite convincing. The small magnetic exchange present in the crystal is Table 3 Values of the fitted parameters for the reduced chemical model (units e) and the Amsterdam density functional calculation Atom Experiment Theory Tc 4dxy 0.77(2) 0.69 4dxz 1 yz 0.08(2) 0.10 4dx2 2 y2 0.03(2) 0.04 4dz2 0.00(2) 0.03 r4d (k) 1.00(1) — 4d*xy 20.15(2) 20.06 4d*xz 1 yz 0.14(3) 20.01 4d*x2 2 y2 20.18(3) 0.00 4d*z2 0.05(3) 0.00 r4d* (k) 1.30(5) — multipole (10) 20.13(2) — Cl 3px 20.008(5) 20.016 3py 0.069(5) 0.080 3pz 0.018(4) 20.001 d — 20.004 N 2ps 20.012(10) 20.025 2pp 20.122(16) 20.100 dp — 0.001 reflected in the negligible spin density delocalised off the [TcNCl4]2 ion.The spin population on the TcNCl4 fragment is reduced from unity to 0.96(4), with a barely significant 0.04(4) spin in total on the 20 hydrogen sites per Tc.Within the TcNCl4 fragment the overall atom-centred populations on Tc [0.78(2)], Cl [0.28(2)] and N [20.14(1)] demonstrate the covalence in the Tc]Cl bond and the spin polarisation in the Tc]N bond. The result for the chlorine population is in reasonable agreement with the 0.2 spin estimated from the ESR results, given the simple model used in deriving that number. We can now see that the 0.45(6) spin on Cl as estimated from a limited PND data set is mainly a result of the substantial Tc]N polarisation.The fraction of the spin not on the superposed Tc plus N atoms is 0.44 = 0.28/(0.78 2 0.14), in good agreement with the previous PND result.10 The present more extensive data enable us to allow for these extra effects, and to state that 29(2)% of the spin on the [TcNCl4]2 ion is delocalised onto the chlorines with 215(1)% delocalised onto the nitrogen. Thus neither ESR nor the previous limited PND tells the complete story.The anisotropy on the atom centres tells us more. On chlorine the p-bonding orbitals together have net positive spin populations, associated with covalent spin delocalisation, resembling the situation we have previously found, for example in the hexacyanochromate( III) ion.25 However, here we have the added complication of a lower symmetry about the Tc]Cl bond. The in-plane p orbital has a large significant population, while the out-of-plane p population is not significant.This is just what is expected if it is the Tc-centred spin density of the 4dxy orbital that is delocalised. On the nitrogen, which interacts with the formally empty 4dxz, 4dyz and 4dz2 technetium orbitals, the densities are all negative, due to spin polarisation, with p polarisation dominant. The anisotropy and radial dependence around the Tc atom seems more complex. If we restrict the flexibility of the model (Table 3), the dominant feature is a 4dxy positive density of 0.77(2) which is readily understandable as the result of the ligand field. In addition the Tc-centred diffuse d orbitals show significant density, explainable by spinpolarisation effects.The s- and p-bonding Tc]N orbitals show significant negative density on the nitrogen, and there is positive density mainly in the diffuse Tc-centred 4dz2 and 4dxz 1 yz components of these orbitals, 0.05(3) and 0.14(3), with some in the atom-like 4dz2 and 4dxz 1 yz, 0.00(2) and 0.08(2).Thus, consistent with the nitrogen populations, p interaction dominates. The remaining d-centred orbital, 4dx2 2 y2 is s bonding to chlorine, but the diffuse d population of 20.18(3) compared with the atom-like d of 0.03(2) indicates the expected polarisation of the x2 2 y2 density by the nearby xy density. The more flexible model subsumes some of these polarisations into 5sp populations, which may not be realistic, and also via 4d9 allows core polarisation, making an altogether more complicated result less capable of such simple interpretation.However, the overall improvement in fit when diffuse density is modelled a little less flexibly, while core d populations are more so, indicates that core polarisation of the technetium ion is significant.1452 J. Chem. Soc., Dalton Trans., 1997, Pages 1447–1453 Proton-spin polarisation Our empirical modelling has clearly indicated that there is a proton-spin polarisation which is not significantly different in the two crystal orientations nor over the five independent sites, and which does not change over the 10 d period of the PND experiment.For the magnetic field and temperature of our experiment we predict an equilibrium proton polarisation of 0.31%; 0.28(2)% is observed. This corresponds to a spin temperature of 1.7(1) K, agreeing fairly well with the nominal sample temperature of 1.5 K. This agreement provides confidence in the other results of the refinements.We should also note the sensitivity of PND to proton polarisation in which we obtain an error of only 0.014%. Theoretical calculation The unrestricted theoretical calculations, shown in Table 2, all reproduce the qualitative features of the spin density on the [TcNCl4]2 fragment. That is, the localisation of spin into the technetium 4dxy orbital which is covalently bonded to the inplane chlorine p orbitals and the spin polarisation of the s and p systems on the N are both reproduced in all calculations. The covalent delocalisation onto the chlorines in the Hartree–Fock calculations is a factor of 2 too small, while the polarisation of the Tc]N bond is too low by an even larger factor.Given that the MIDI-plus-polarisation calculation employs a relatively extensive basis set, it seems unlikely that the discrepancy arises from inadequacy of the basis set (although this basis is not by any means near to the Hartree–Fock limit).It seems more likely that it is the inadequate treatment of electron correlation inherent in the UHF approximation that is a problem. This conclusion is strengthened by the failure to reproduce the sizeable technetium radial polarisations that we see in the experiment. The multiple p and d functions in the valence region of the 15,9,6/6,3,3 MIDI basis are, in principle, capable of duplicating such spin polarisation. Unfortunately, while configurationinteraction calculations are quite feasible on an ion of this size, available packages are unable to extract spin densities from the resulting better correlated wavefunctions.The simple density functional calculation, by contrast, provides far too much covalence and interatomic spin polarisation. In this calculation the molecular basis is fixed, relatively rigidly, at atom-like functions and the correlation is treated at the lowest density functional approximation. The second density functional calculation is a significant improvement both theoretically and in agreement with experiment. It has both a much more flexible basis, and improved density functional approximations which, among other things, are expected to provide a better treatment of correlation than do the UHF calculations.This improvement seems to provide the key to the better agreement; we need both triple-zeta quality basis sets and a better treatment of electron correlation than the UHF calculation can provide.The use of a relativistic effective core potential may well also be necessary. Even though the interatomic spin polarisations are well described, the intratechnetium polarisation is still not completely correct. Conclusion The experimental spin distribution derived by modelling the PND data from this 4d1 [TcNCl4]2 complex clearly quantifies that obtained from previous limited experiment and qualitative theory, viz.: (a) covalent delocalisation of 29(2)% of the single spin from the technetium 4dxy orbital onto the chlorine in-plane 3pp orbital region; (b) spin polarisation of the Tc]N bond, mainly of p symmetry, resulting in a negative spin population on the nitrogen of 215(1)% of the total; (c) more unexpectedly, there is a complex spin polarisation of d electrons on the Tc which results in a shape change of the formally unpaired spin density such that the diffuse fringes of the d density have opposite sign of spin to the majority; in addition the formally spinpaired d-orbital populations are polarised in ways predictable from their bonding or non-bonding nature. The amount of proton nuclear spin alignment predicted for this high magnetic field and low temperature is also observed in the fit.Theoretical ab-initio calculations must include electron correlation in some way even qualitatively to account for these observations. Our UHF calculations, with reasonable basis sets, reproduce the Tc]Cl covalence and Tc]N spin polarisation qualitatively, but not quantitatively.The spin polarisation on the Tc is not reproduced at all. The inference is that electron correlation must be dealt with better by use of multideterminant wavefunctions and configuration interaction. Simple UHF all-electron or quasi-relativistic effective core potential calculations provide an inadequate description of the bonding in this complex ion. By contrast, a density functional calculation, which includes a flexible basis set and a relativistic effective core potential, reproduces our results quite well, possibly largely because this method also treats electron correlation more adequately than does the UHF method.The spin densities on the ligand are predicted almost quantitatively, while the complex pattern of spin polarisation on the technetium, which the UHF calculations do not predict at all, are now in semiquantitative agreement with experiment. We are comparing differently defined and obtained populations in the theory and experiment, and the measurements are on a crystal while the calculation is on an isolated ion.It is not clear that better agreement should be expected. A better assessment of theory versus experiment would be an Amsterdam density functional band calculation of magnetic structure factors, to compare directly with the experimental values. This is our next task. Acknowledgements We thank the Australian Research Council and the Department of Industry, Trade and Tourism for financial support, and the Institut Laue-Langevin for access to the D3 polarised neutron diffractometer.References 1 B. N. Figgis, J. B. Forsyth, E. S. Kucharski, P. A. Reynolds and F. Tasset, Proc. R. Soc. London, Ser. A, 1990, 428, 113; P. J. Brown, B. N. Figgis and P. A. Reynolds, J. Phys. Condens. Matter, 1990, 2, 5309; C. D. Delfs, B. N. Figgis, J. B. Forsyth, E. S. Kucharski, P. A. Reynolds and M. Vrtis, Proc. R. Soc.London, Ser. A, 1992, 436, 417. 2 P. A. Reynolds, B. Moubaraki, K. S. Murray, J. W. Cable, L. M. Engelhardt, B. N. Figgis and A. N. Sobolev, unpublished work. 3 P. A. Reynolds, B. Moubaraki, K. S. Murray, J. W. Cable, L. M. Engelhardt and B. N. Figgis, J. Chem. Soc., Dalton Trans., 1997, 263. 4 S. P. Best, B. N. Figgis, J. B. Forsyth, P. A. Reynolds and P. L. W. Tregenna-Piggott, Inorg. Chem., 1995, 34, 4605. 5 J. Baldas, J. F. Boas, J. Bonnyman and G. A. Williams, J. Chem.Soc., Dalton Trans., 1984, 2395. 6 B. N. Figgis, P. A. Reynolds, F. K. Larsen, G. A. Williams and C. D. Delfs, Aust. J. Chem, 1996, 49, 663. 7 J. B. Raynor, T. J. Kemp and A. M. Thyer, Inorg. Chim. Acta, 1992, 193, 191. 8 R. Kirmse, K. Kohler, U. Abram, R. Bottcher, L. Golic and E. de Boer, Chem. Phys., 1990, 143, 75. 9 K. Kohler, R. Kirmse, R. Bottcher, U. Abram, M. C. M. Gribnau, C. P. Keijzers and E. de Boer, Chem. Phys., 1990, 143, 83. 10 B. N. Figgis, P. A. Reynolds and J. W. Cable, J. Chem. Phys., 1993, 98, 7743. 11 H. Glättli and M. Goldman, Methods Exp. Phys., 1987, 23C, 241. 12 A. Abragam, G. L. Bacchella, C. Long, P. Meriel, J. Peisvaux and M. Pinot, Phys. Rev. Lett., 1972, 28, 805. 13 M. Leslie, G. T. Jenkin, J. B. Hayter, J. W. White, S. Cox and G. Warner, Philos. Trans. R. Soc. London, Ser. B, 1980, 290, 497 and refs. therein.J. Chem. Soc., Dalton Trans., 1997, Pages 1447–1453 1453 14 A. Zheludev, V. Barone, M. Bonnet, B. Delley, A. Grand, E. Ressouche, P. Rey, R. Subra and J. Schweizer, J. Am. Chem. Soc., 1994, 116, 2019. 15 B. N. Figgis, P. A. Reynolds and G. A. Williams, J. Chem. Soc., Dalton Trans., 1980, 2339. 16 E. Clementi and C. Roetti, At. Data Nucl. Data Tables, 1974, 14, 177. 17 J. M. Stewart, E. R. Davidson and W. T. Simpson, J. Chem. Phys., 1965, 87, 3175. 18 A. Neuhaus, A. Veldkamp and G. Frenking, Inorg. Chem., 1994, 33, 5278. 19 P. J. Hay and W. R. Wadt, J. Chem. Phys., 1985, 82, 299. 20 V. Jonas, G. Frenking and M. T. Reetz, J. Comput. Chem., 1992, 13, 919. 21 W. J. Hehre, R. Ditchfield and J. A. Pople, J. Chem. Phys., 1972, 56, 2257. 22 M. W. Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert, M. S. Gordon, J. H. Jensen, S. Kaseki, N. Matsunaga, K. A. Nguyen, S. J. Su, T. L. Windus, M. Dupuis and J. A. Montgomery, J. Comput. Chem., 1993, 14, 1347. 23 Gaussian basis sets for molecular calculations, eds. S. Huzinaga, J. Andzelm, M. Klobukowski, E. Radzio-Andzelm, Y. Sakai and H. Tatewaki, Elsevier, Amsterdam, 1984. 24 C. Fonseca Guerra, in Methods and Techniques in Computational Chemistry, eds. E. Clementi and G. Corongiu, STEF, Cagliari, 1995; G. te Velde and E. J. Baerends, J. Comput. Phys., 1992, 99, 84; E. J. Baerends, D. E. Ellis and P. Ros, Chem. Phys., 1973, 2, 41. 25 B. N. Figgis, J. B. Forsyth and P. A. Reynolds, Inorg. Chem., 1987, 26, 101. Received 11th November 1996; Paper 6/07650E

ISSN:1477-9226    

DOI:10.1039/a607650e

出版商:RSC    

年代:1997

数据来源: RSC