DALTON FULL PAPER J. Chem. Soc., Dalton Trans., 1999, 1717–1727 1717 Co-ordination chemistry of the [Re(NO)2(PR3)2]1 fragment: crystallographic and computational studies† Heiko Jacobsen, Katja Heinze, Angela Llamazares, Helmut W. Schmalle, Georg Artus and Heinz Berke Anorganisch-chemisches Institut, Universität Zürich, Winterthurerstr. 190, CH-8057 Zürich, Switzerland Received 19th February 1999, Accepted 8th April 1999 The cationic complexes [Re(NO)2(PCy3)2]1 I1 and [Re(NO)2(PR3)2L]1 [L = CO, R = Cy III1; L = C6H5CHO, R = Cy IV1; L = ONRe(NO)(PR3)2H, R = iPr V1] have been synthesized and their structures determined. The counter ion in all cases is [B{3,5-(F3C)2C6H3}4]2.Complex I1 adopts the C2v butterfly geometry, whereas III1 takes on a trigonal bipyramidal (TBP) co-ordination. In IV1 and V1 one of the nitrosyl ligands is strongly bent, and a shape analysis suggests that the co-ordination geometry of the [Re(NO)2(PR3)2L]1 core is best described as tetragonal pyramidal (TP).A computational study based on density functional theory showed how steric eVects due to the ligand L induce the NO bend, and subsequently lead to the change in co-ordination from TBP to TP. Examination of a series of model compounds [Re(NO)2(PH3)2L]1 showed further how the p donor and acceptor properties of the ligand L are reflected in the P–Re–P and N–Re–N angles of the complexes. The nitrosyl ligand 1 plays a special role in transition metal chemistry.It is capable of supporting diVerent oxidation states of the metal center via diVerent co-ordination modes, and has the capability to activate metal–ligand bonds. Prominent examples of the latter are nitrosyl substituted transition metal hydride complexes,2 in which the M–H bond shows an increased reactivity toward alkyne insertion and carbonyl reduction. In the context of structural chemistry and reactivity exploration in this class of compounds, we have prepared a series of mononitrosyl hydrido complexes containing various phosphorus donor ligands, as well as chromium,3 tungsten 4 and rhenium5 centers.The increased hydridicity 2 of these compounds has been probed by the interaction with acidic substrates. 6 We also directed our eVorts towards the synthesis of dinitrosyl hydride derivatives, which should possess even more activated metal–hydrogen bonds. In contrast to their carbonyl analogues [Mn(CO)3(PR3)2H], manganese complexes of the general formula [Mn(NO)2(PR3)2H] undergo facile insertions of polar unsaturated molecules.7 We then set out to extend this chemistry to the related rhenium complexes, and provided synthetic access to compounds of the type [Re(NO)2(PR3)2H].8 During the course of this work we were also able to isolate and characterize the 16 electron fragment [Re(NO)2(PCy3)2]1, as well as a variety of complexes of the type [Re(NO)2(PR3)2L]1.The present paper is mainly concerned with structural aspects of [Re(NO)2(PR3)2L]n1 complexes (n = 0 or 1).In particular, we want to address the question of how structural changes in the [Re(NO)2(PR3)2]1 fragment under co-ordination of a ligand L might provide information about the nature of the Re–L bond. The variation in the co-ordination geometry might further influence the reactivity of the species. The experimental part is complemented by a computational study based on density functional theory (DFT).9 The molecular and electronic structure of 12 model compounds were investigated, as displayed in Fig. 1. The calculations were implemented to support † Supplementary data available: optimized geometries and eigenvalues. For direct electronic access see http://www.rsc.org/suppdata/dt/1999/ 1717/, otherwise available from BLDSC (No. SUP 57540, 4 pp.) or the RSC Library. See Instructions for Authors, 1999, Issue 1 (http:// www.rsc.org/dalton). the results obtained from the X-ray crystallographic analyses, and to provide explanations for the observed structural features.Results and discussion Crystallographic studies We determined the crystal structures of [Re(NO)2(PCy3)2]- [BArF 4], and of the three [Re(NO)2(PR3)2L][BArF 4] complexes with L = CO, C6H5CHO, or ONRe(NO)(PR3)2H. Here, [BArF 4] stands for [B{3,5-(F3C)2C6H3}4]2. The anion is excluded from our discussion, which is focussed on the structural elements of the rhenium fragments which will be analysed together with those of Re(NO)2(PR3)2H.8 We shall refer to the metal fragments as [Re(NO)2(PCy3)2]1 I1, Re(NO)2(PiPr3)2H II, [Re(NO)2- (PCy3)2(CO)]1 III1, [Re(NO)2(PCy3)2(C6H5CHO)]1 IV1, and [Re(NO)2(PiPr3)2{ONRe(NO)(PiPr3)2H}]1 V1.Selected structural parameters for these complexes are presented in Table 1. For in-depth background information a reader should refer to the deposited crystallographic data. [Re(NO)2(PCy3)2]1 I1. A view of the molecular structure of complex I1 in the crystal is displayed in Fig. 2. It can be Fig. 1 The [Re(NO)2(PH3)2L]n1 model complexes n = 0 or 1. L Re ON ON PH3 n+ 1+ 2 3+ 4+ 5+ 6 7+ 8 9+ 10+ 11+ 12+ L – H – CO CN-CH3 PH3 CNNC- CH3 Cl- ON H2CO ONH ON-R' x z y R' = Re(NO)(PH3)2H PH31718 J. Chem. Soc., Dalton Trans., 1999, 1717–1727 Table 1 Selected bond lengths and angles a for [Re(NO)2(PR3)2L]n1 complexes (n = 0 or 1) I1 II III1 IV1 V1 R Cy iPr Cy Cy iPr L — H2 CO C6H5CHO ONR9 b Re–L — 177.93(2) 197.9(6) 218.8(3) 219.9(9) Re–P 245.4(3) 246.2(3) 242.8(2) 242.1(2) 247.4(2) 248.8(1) 248.4(1) 248.8(1) 247.2(4) 248.2(4) Re–N 173.5(10) 176.6(8) 180.4(7) 178.0(7) 179.0(7) 182.5(5) 175.8(4) 181.1(4) 178(1) 180(2) N–O 122.5(12) 118.0(11) 119.3(9) 122.7(9) 119.1(9) 117.6(1) 119.9(5) 120.4(5) 120(1) 117(2) P–Re–P 159.93(8) 153.89(6) 169.62(5) 158.40(4) 162.8(1) Re–N–O 166.9(9) 165.7(9) 173.1(8) 175.4(7) 174.0(6) 176.3(6) 150.9(3) 175.9(4) 158(1) 172(1) N–Re–N 115.9(4) 127.4(3) 121.5(3) 108.8(2) 111.4(6) Other L–Re–N C–O L–Re–N C–O L–Re–N O–N L–Re–N 122.3(3) 110.2(2) 114.6(8) 129.6(3) 108.9(3) 123.6(5) 91.6(1) 159.5(1) 125(2) 97.4(5) 150.7(5) a Distances in pm, angles in (8).b R9 = Re(NO)(PiPr3)2H. obtained almost quantitatively by the reaction of II with [(C6H5)3C][BArF 4] in benzene. The complex can be described as a distorted C2v butterfly fragment, which is obtained on removing one equatorial ligand from an ideal trigonal bipyramidal arrangement. Important geometric features are a P–Re–P angle some 208 smaller than the ideal value of 1808, and an N–Re–N angle close to 1208.The phosphorus atoms are bent away from the NO groups; the nitrosyl ligands themselves are not linearly co-ordinated but show a cisoid bend of about 158. Structures of a variety of [M(NO)2(PR3)2]n1 compounds (n = 0 or 1; M = Fe, Ru, Os, Co, Rh or Ir), are described in the literature.10 All these complexes having electron counts of 18 (or 17),10m exhibit the co-ordination geometry of a distorted tetrahedron,‡ and therefore cannot be compared with I1.However, the crystal structures of two isoelectronic carbonyl compounds are known, namely [Rh(CO)2{P(2,4,6-(MeO)3C6H2)3}2]1 and [Ru(CO)2(PtBu2Me)2].12,13 The latter complex 13 possesses the same C2v butterfly geometry as that of I1, whereas the former12 shows square planar co-ordination. Thus, it was not initially clear which geometry the fragment I1 might adopt. Fig. 2 Molecular structure of complex I1. Displacement ellipsoids are shown at the 30% level.Hydrogen atoms are omitted for clarity. Not shown are the counter ion and solvate molecules. ‡ For an orbital analysis of MX2(NO)2 systems by extended Hückel theory compare ref. 10(m). See also ref. 11. The main aspects of the Walsh diagram for the planar D4h into bent C2v transformation are established for ML4 complexes, 14a and Caulton and co-workers 13 have adapted this analysis for compounds of the type M(CO)2(PR3)2. The dz2 orbital is stabilized under bending, because of diminished overlap with the sCO lone pair, and because back bonding into p*CO is now possible.15 The dxz orbital is also stabilized by back donation in the bent geometry.On the other hand, the dyz orbital is strongly destabilized in the bent structure, due to diminished overlap with p*CO, and due to antibonding overlap with the sCO lone pair. The important interactions are shown below [adopted from ref. 13(a)] (see also Fig. 7). The antibonding interaction between dyz and sCO or sNO, respectively, can be reduced by a cisoid bend of the M–C–O or the M–N–O angle.This explains the observed non-linear coordination of the nitrosyl ligands in complex I1. The preferred geometry will be non-planar if the stabilization due to back donation outweighs the destabilizing interactions. The important criterion is the energetic match between the metal donor orbitals and the ligand p*XO (X = C or N) acceptor orbitals.13 In I1 the electron rich metal center Re2I possesses d orbitals which are at relatively high energies.These are energetically well suited for an interaction with the p*NO orbitals. Thus, I1 prefers the C2v butterfly geometry. The same holds for the neutral ruthenium complex [Ru(CO)2(PtBu2Me)2].13 In contrast, the low energy of the d orbitals of RhI in [Rh(CO)2{P(2,4,6- (MeO)3C6H2)3}2]1 decreases the role of back donation. This complex therefore adopts the square planar geometry.12 In a DFT calculation we have tried to optimize the square planar geometry of the model complex [Re(NO)2(PH3)2]1 11.This could only be achieved by employing angular constraints, and enforcing a planar co-ordination environment, which indicates that planar 11 is not a local minimum on the potential energy surface. This hypothetical molecule should have a triplet state, since one of the rhenium d-based orbitals and one combination of NO p* orbitals are accidentally degenerate.[Re(NO)2(PiPr3)2H] II. The preparation and the structure of complex II, as shown in Fig. 3, have already been discussed,8 d z 2 - p*CO y z d yz - sCOJ. Chem. Soc., Dalton Trans., 1999, 1717–1727 1719 and we will only briefly comment on its geometry. The structure is that of a distorted trigonal bipyramidal (TBP). Compared to I1, we observe that the N–Re–N angle opens up by about 118 under co-ordination of the hydride ligand. At the same time, P–Re–P becomes narrower by 68.The bending distortion of the phosphorus donor ligands is well understood.14 Bending of the PR3 groups of [Re(NO)2(PR3)2L]n1 towards the ligand site polarizes the dxz orbital of the metal in the direction of the p accepting nitrosyl ligands, providing better dxz–p*NO overlap, and enhancing the amount of back donation to NO. The degree of back bending of the PR3 is limited by steric repulsion between PR3 and L, and between the phosphorus ligands themselves. The small hydride ligand does not provide much steric hindrance for the bulky PiPr3 group, but it does increase the electron density on the rhenium center.Back bonding to the nitrosyl ligands becomes stronger, and as a consequence the P–Re–P angle decreases. [Re(NO)2(PCy3)2(CO)]1 III1. Reaction of complex I1 with the prototypical p acceptor ligand CO leads to formation of III1. Its molecular geometry in the crystal is displayed in Fig. 4. One of the PCy3 ligands is highly disordered, but was resolved in the course of the structure refinement.The N–Re–N angle is still larger than that in the free fragment I1, but the P–Re–P angle opens up by 108 (see Table 1). Since CO is competing with the NO ligands for back donation, a strong polarization of dxz away from the L site is no longer favorable, and consequently P–Re–P opens up. The fact that CO is competing for electron density manifests itself also in a small N–O bond contraction Fig. 3 Molecular structure of complex II.Displacement ellipsoids are shown at the 40% level. Hydrogen atoms are omitted for clarity, except for the hydride ligand, which is displayed as a sphere with arbitrary size. PH3 PH3 PH3 PH3 x z y [compare d(N–O) in I1 and III1], and a small C–O bond elongation [compare to d(C–O) = 112.8 pm in the gas phase 16]. [Re(NO)2(PCy3)2(C6H5CHO)]1 IV1. Compound IV1 is instantaneously formed when I1 is treated with benzaldehyde. In this complex a new structural motif is introduced.The benzaldehyde L does not bind in a symmetrical, but rather asymmetrical fashion, as can be seen in Fig. 5. The pseudo C2 rotational axis is removed, and only idealized Cs symmetry is retained. Characteristic bond lengths such as Re–N and Re–P are very similar in III1 and IV1, but the co-ordination geometry is very diVerent. The N–Re–N angle is now smaller than that of the free fragment I1. Furthermore, one of the nitrosyl ligands (N2O2 in Fig. 5) is strongly bent forming a Re–N–O angle of 1518.This falls right between the linear co-ordination of the 3e2 donor NO1 (M–N–O 1808) and the bent co-ordination of the 1e2 donor NO2 (M–N–O 1208). Two very diVerent L–Re–N angles are observed, one being close to 908 and the other being about 1608. Thus, the co-ordination geometry of IV1 resembles Fig. 4 Molecular structure of complex III1. Details as in Fig. 2. Fig. 5 Molecular structure of complex IV1. Details as in Fig. 2.1720 J. Chem. Soc., Dalton Trans., 1999, 1717–1727 Table 2 Observed d angles a for [Re(NO)2(PR3)2L]1 complexes, together with values b for idealized polyhedra.Also given are the standard deviations s(G) (see text for definition) Complex Ideal TBP (D3h) III1 IV1 V1 Ideal TP (C4v) d(a1) 101.5 115.5 124.7 120.3 119.8 d(a2) 101.5 111.9 79.1 85.2 75.7 d(a3) 101.5 98.3 127.6 123.8 119.8 d(a4) 101.5 115.3 126.1 122.8 119.8 d(a5) 101.5 110.9 77.1 80.9 75.7 d(a6) 101.5 99.7 127.7 123.5 119.8 d(e1) 53.1 26.6 68.3 65.5 75.7 d(e2) 53.1 52.7 59.0 52.4 75.7 d(e3) 53.1 44.0 5.6 5.9 0.0 s(D3h) 0.0 12.4 26.1 23.2 26.9 s(C4v) 26.9 30.4 7.9 9.7 0.0 a In 8.b From ref. 18(b). more closely that of a tetragonal pyramid (TP) than that of a trigonal bipyramid (TBP). The infrared spectrum shows however, both in solution and in the solid state, a group of peaks in the carbonyl–nitrosyl region that could not be assigned (see Experimental section). We can envisage this structural change as follows:§ co-ordination of benzaldehyde L to the open coordination site of the d8-Re(NO)2(PR3)2 fragment induces a bend in one nitrosyl ligand.This subsequently leads to a formal oxidation of the metal center, resulting in a d6-Re(NO)2(PR3)2L species, which is still co-ordinatively and electronically unsaturated. The preferred geometric arrangement of a d6-MX5 fragment is tetragonal pyramidal. The origin of this distortion will be analysed at a later point. [Re(NO)2(PiPr3)2{ONRe(NO)(PiPr3)2H}]1 V1.The last complex we include in this section can be described as an adduct of the type [I1]V1[II]V1 (the nomenclature [X]Y stands for a fragment having the structure of X but with the geometric parameters as found in the molecule Y). It is prepared by the reaction of [(C6H5)3C][BArF 4] on II in a 1 : 2 ratio. The geometry in the crystal is displayed in Fig. 6. The oxygen of one of the nitrosyl groups of II is apparently a Lewis base strong enough to interact with other Lewis acids, such as I1 or BF3.8 The geometry of the [I1]V1 fragment is similar to that of IV1, and might also be described as a tetragonal pyramid.The bend of one of the NO ligands, however, is not as prominent as in IV1, and the L–Re– N angles are also somewhat closer to the value of 1208 of the ideal trigonal bipyramid (see Table 1). The geometry of the [II]V1 fragment is related to that of II. A major diVerence is an even smaller P–Re–P angle of 1418. One of the nitrosyl oxygen Fig. 6 Molecular structure of complex V1. Displacement ellipsoids are shown at the 20% level. Hydrogen atoms are omitted for clarity, except for the hydride ligand, which is displayed as a sphere with arbitrary size. Not shown is the counter ion. § The principal orbital interactions for five- and six-co-ordination have been investigated by HoVmann and co-workers in a classical series of papers, see refs. 15 and 17. atoms of [II]V1 functions as a Lewis base, which leads to electron depletion at this particular nitrosyl ligand.This in turn can be counteracted by an eVective back donation, which is made possible by the narrowing of the P–Re–P angle (see above). We have found a similar eVect for the complex [Re(H)(NO)- (NOBF3)(PiPr3)2].8 The Re and the bridging NO are not coplanar; the Re1–O3–N3–Re2 dihedral angle amounts to 1398. It was mentioned that the co-ordination geometries of both complexes IV1 and V1 are closer to a TP than to a TBP coordination. To put this argument on more quantitative grounds, we follow the approach of Muetterties,18 and obtain a measure of shape for these aggregates by means of the dihedral angles d formed by the normals to adjacent faces of a given polytopal form.The three five-co-ordinated molecules described for the first time in this work can then be compared to the idealized geometries of a D3h trigonal bipyramid and a C4v tetragonal pyramid, as shown below [adopted from ref. 18(b)]. The molecules are oriented such that the phosphorus ligands occupy the A1 and A2 positions. For III1, the CO ligand is chosen to occupy the E2 position, whereas for IV1 and V1 the bent nitrosyl ligand is placed at E2. The results of the shape analysis for the rhenium complexes together with values for the ideal co-ordination polyhedra as defined by Muetterties 18 are collected in Table 2. The values of the shape determining angles d(en), especially that of d(e3), and the fact that two of the d(an) angles, namely d(a2) and d(a5), are significantly smaller than the remaining members of the set, all indicate that IV1 and V1 are indeed closer to the C4h-TP in coordination geometry.For III1, the d(an) angles span a smaller range of values, and its co-ordination geometry is related to that of the D3h-TBP. Also, the standard deviations s(G), eqn. (1), lead to the same conclusion that the co-ordination s(G) =÷1 9 o9 n = 1 ((dn)exp 2 (dn)G)2 (1) polyhedra for IV1 and V1 match closer the tetragonal pyramid, and that III1 can be described as a trigonal bipyramid (compare Table 2).Computational studies We divide the twelve model complexes as presented in Fig. 1 into two groups. Symmetric complexes 11–91 are characterized by ligands L, which possess higher symmetry than Cs, whereas E3 E1 E2 A1 A2 E1 A2 E3 a1 a3 a6 a4 e3 e1 e2 a5 a2 C4v D3h A1 E2J. Chem. Soc., Dalton Trans., 1999, 1717–1727 1721 Table 3 Optimized geometries a for Cs-symmetric [Re(NO)2(PH3)2L]n1 model complexes (n = 0 or 1) Complex 11 23 1 41 51 67 1 89 1b L —H 2 CO CNCH3 PH3 CN2 NCCH3 Cl2 ON Re–L — 172.4 200.9 206.5 248.8 210.7 213.8 249.2 216.6 Re–P 245.7 239.1 247.0 245.4 245.6 242.5 245.7 243.3 247.9 247.7 Re–N 180.3 182.5 183.9 183.0 182.4 182.2 181.5 180.7 181.1 180.7 N–O 118.3 120.3 117.6 118.3 118.4 119.9 118.7 120.4 118.2 118.1 P–Re–P 160.8 151.7 174.0 173.2 176.1 160.9 173.5 161.2 172.6 Re–N–O 161.9 174.1 175.2 172.8 169.9 171.7 168.3 166.6 166.5 167.0 N–Re–N 118.5 126.6 125.9 123.1 121.6 121.6 116.9 115.1 110.9 Other C–O C–N H–P–H C–N N–C O–N Re–O–N 115.3 117.2 98.2 117.6 116.3 119.4 179.5 a Distances in pm, angles in 8.b Unrestricted calculation without symmetry constraints on the doublet state. Table 4 Optimized geometries a for asymmetric [Re(NO)2(PH3)2L]1 complexes Complex 101 trans-111 cis-111 121 L H2CO ONH ONH ONR9 b Re–L 222.0 211.2 211.2 222.2 Re–P 246.1 247.4 247.6 245.0 Re–N 179.3 184.6 182.5 181.7 180.8 185.9 179.3 183.9 N–O 118.2 119.3 118.2 118.6 119.1 118.3 119.1 119.5 P–Re–P 167.7 176.8 169.9 160.6 Re–N–O 176.9 147.8 164.6 172.6 145.8 180.0 177.7 149.4 N–Re–N 108.5 121.5 109.6 108.2 Other C–O L–Re–N O–N O–N–H L–Re–N O–N O–N–H L–Re–N O–N L–Re–N O–N–R9 123.5 95.3; 156.2 126.6 104.4 115.6; 122.9 125.2 107.1 156.9; 93.5 126.0 154.3 97.5 126.0 a Distances in pm, angles in 8.b R9 = Re(NO)(PH3)2H. in the asymmetric complexes 101–121 the ligands are considered to be mirror symmetric.Selected geometric parameters are presented in Tables 3 and 4, respectively. If we compare the optimized geometries of compounds 11, 2 and 31 with the crystal structures of I1, II and III1 we find reasonable agreement between experiment and theory. The general trends are well reproduced in the calculations, e.g. a shortening of Re–P and an increase in N–Re–N when going from I1(11) to II(2). The Re–N separation is generally overestimated in the calculations by about 6 pm.As a consequence, due to a reduced back bonding, the N–O distance falls somewhat short in comparison to the experiment. Surprisingly, the simple model phosphine PH3 reproduces the co-ordination geometry of the phosphorus ligands extremely well, especially where the Re–P distances are concerned. Also, 101 and 121 seem to be good models for complexes IV1 and V1, respectively. The calculation predicts the asymmetric co-ordination with two diVerent nitrosyl ligands as found in the experiment. The co-ordination geometry of the nitrosyl ligand is in satisfactory accordance to the crystal structure, and the angles P–Re–O and N–Re–N are also close to within 28.Influence of the phosphorus donor. The reasonable close agreement between the calculated and observed P–Re–P angle of the symmetric complexes suggests that this parameter is not overly dependent on the nature of the R group of the PR3 ligand. Instead, the right polarization of the dxz orbital needed to achieve an optimum ratio of back bonding between the NO and L ligands to first order determines the degree of PR3 bending (see above).The diVerent donor capability however influences the electron densities at the Re, and therefore to a certain extent the geometric arrangement of the ligands in the yz plane. This might explain the somewhat larger deviation between theory and experiment in the Re–N distances. We further checked the influence of the P–Re–P angle on the co-ordination of the NO and L ligands by restricted geometry optimizations for complexes 11–8, in which P–Re–P was fixed at 150 and 1708, respectively. In all cases, only marginal geometric diVerences compared to the fully optimized species were found.The potential energy surface for the P–Re–P bend is very shallow, and the angle bending does not require much energy. As an example, we provide two cases, beginning with 2, the angle P–Re–P fixed at 1708.For the 188 distortion from the calculated equilibrium geometry, an energy of only 11 kJ mol21 is needed. The average deviation between selected bond distances and angles amounts to 0.4 pm and 1.08. For 51 fixed at 1508, narrowing the P–Re–P requires 38 kJ mol21. The selected angles change on average by 0.68, and the bond distances (Re–L excluded) by 0.3 pm. The Re–L bond in 51 (1508) is elongated by 3.2 pm. This is easily explained by keeping in mind that diminishing the P–Re–P angle leads to a polarization of dxz away from L, and thus to a reduced back bonding to the PH3 ligand in equatorial position.This in turn weakens and lengthens the P–Re bond. Additional information on structures and energies of the restricted geometry complexes can be found in SUP 57540. The P–Re–P size allows us to weigh the importance of pxz back bonding to L. For [Re(NO)2(PR3)2L] complexes in which P–Re–P is about the same size or smaller than in the Re(NO)2(PR3)2 fragment, this back-bonding interaction is of no or only minor importance.This is naturally the case for L = H2, 2, and Cl2, 8, but also for CN2, 6. On the other hand, when P–Re–P is substantially larger than in the free fragment, pxz back bonding is of importance, as it is for L = CO, 31, CNCH3, 41, NCCH3, 71, and also for PH3, 51. This argument is based on qualitative considerations, and it does not allow one to infer direct correlation between the amount of p back bonding and the P–Re–P angle.Re-dxz Interactions with other ligand based orbitals, as well as interactions involving Re-dyz, further influence P–Re–P and the amount of p back donation to the equatorial ligands. Dependence of the angle N–Re–N on the nature of L. We already mentioned the important orbital interactions which determine the size of N–Re–N in relation to the problem of the ground state geometry of [Re(NO)2(PCy3)2]1 I1. We will now1722 J. Chem. Soc., Dalton Trans., 1999, 1717–1727 Table 5 Composition of the three highest occupied orbitals of the complexes 2, 31 and 8 at their equilibrium geometry Complex Symmetry Re (%) NO (%) L (%) 2 31 8 1b1 1a1 1b2 1b1 1a1 1b2 1b1 1a1 1b2 px 5 pz 11 py 9 pz 9 py 9 pz 6 py 7 dxz 48 dz2 7 dyz 31 dxz 60 dz2 18 dyz 34 dxz 37 dz2 19 dyz 19 dx2 2 y2 8 dx2 2 y2 5 dx2 2 y2 7 N px 9 N py 8 N pz 15 N px 5 N py 8 N pz 12 N px 7 N py 9 N py 6 N pz 13 N pz 13 N pz 6 N pz 15 O px 24 O py 16 O pz 29 O px 14 O py 15 O pz 22 O px 19 O py 25 O py 11 O pz 21 O pz 22 O pz 13 O pz 22 H s 9 C px 3 C s 3 C py 5 Cl px 32 Cl pz 12 Cl py 13 O px 6 C pz 1 O py 7 discuss this structural parameter in more detail, and address the question how diVerent types of ligands L may influence N–Re– N in [Re(NO)2(PR3)2L]n1 complexes.We have chosen L to be a s donor, p acceptor, or a p donor ligand. The representative model compounds which we will analyse in detail are then 2, 31 and 8. The highest three occupied molecular orbitals for these complexes are displayed in Fig. 7. The basic composition of these MOs is similar in all three compounds, and a detailed breakdown is presented in Table 5. The metal contribution to the HOMO-2, 1b1, is mainly Re-dxz. Back bonding to the p*NO,xz orbitals increases, when P–Re–P is diminished, and 1b1 is lowered in energy. There is no contribution from H2 to 1b1 in 2. For 31 a p*CO,xz acceptor orbital is combined with Re-dxz in a bonding fashion, whereas for 8 we have antibonding interaction between the metal based orbital and a filled px,Cl2 orbital.Not shown in Fig. 7 are contributions of the PH3 ligand to 1b1. Their importance has been discussed in the previous sections. In orbital 1a1 back bonding occurs from the metal Re-dz2 to the p*NO,yz orbitals. Again, the overlap increases when P–Re–P becomes smaller. The contributions from L to 1a1 are in all cases s antibonding; the ligand orbitals involved are sH2 2, sCO 31 and pz,Cl2 8.Lastly, back bonding to p*NO,yz is also possible from Re-dyz, as realized in 1b2. In contrast to 1a1, the overlap is now lessened when P–Re–P decreases. As in the case of 1b1, there is no con- Fig. 7 Sketches of the three highest molecular orbitals of complexes 2, 31 and 8. Not shown are contributions due to the phosphorus donor ligands. H Re N N O O CO Re N N O O Cl Re N N O O 1 b1 1 a1 1 b2 y z 2 1 b1 1 a1 1 b2 3+ 1 b1 1 a1 1 b2 8 tribution from H2, 2, a bonding interaction with p*CO,yz, 31, and an antibonding interaction with py,Cl2, 8.Our analysis shows that two metal d orbitals compete for back bonding to the p*NO,yz orbitals, namely dz2 in 1a1 and dyz in 1b2.¶ However, these interactions show a diVerent nature in their dependence on the angle N–Re–N. In the former case the metal–ligand overlap increases when N–Re–N decreases, whereas for the latter the opposite trend holds. The relative importance of these two interactions will determine the size of N–Re–N.The Walsh diagram along the N–Re–N bending mode for the three highest occupied orbitals for complexes 2, 31 and 8 is displayed in Fig. 8. The energy curve for orbital 1b1 looks similar in all three cases; the interaction of dxz with p*NO,xz is mainly influenced by the phosphorus donors, and only to a minor degree by the nature of the ligand L. This orbital serves as a reference point for the comparison of the relative energies of the orbitals amongst the diVerent systems (due to the cationic nature of 31, its orbitals are at considerably lower energies than those of 2 and 8).The energy dependence of 1a1 and 1b2 follows the expected trend in all three cases, but there are some important diVerences. For 2, we find orbital crossing of 1a1 and 1b2 around 1208, close to the value of N–Re–N in the “free” fragment 11. Distortion of this angle leads to a stabilization of the HOMO-1, which is 1a1 when N–Re–N decreases, or 1b2 when N–Re–N increases.The orbital coeYcient of the metal based d orbitals in 1a1 is smaller when compared to Re-dyz in 1b2. In the case of 1a1, this is due to the antibonding interaction between the d orbitals and sH2. Consequently, back bonding is more eYcient in 1b2, and when complex 2 is formed from the fragments N–Re–N will open up to lower the energy of 1b2, and to increase this particular interaction. We encounter a similar situation for complex 31. Again, we see the destabilizing s interaction in 1a1, which leads to orbital crossing at around 1208.Again, the metal d contributions are smaller in 1a1 than in 1b2, so that an increase in N–Re–N maximizes the bonding energy. The picture emerged so far changes, when considering the p donor Cl2. In complex 8 both orbitals 1a1 and 1b2 undergo antibonding interaction with occupied pCl2 orbitals; 1b2 is significantly destabilized when compared to 1a1, and the orbital crossing occurs at an angle of around 1358, far from the free fragment.At the N–Re–N value of 11, orbital 1a1 now provides the main backbonding interaction, so that in this case N–Re–N is diminished, to maximize overlap and bonding energy. To sum up our analysis, we might say that in [Re(NO)2- (PR3)2L]n1 complexes, when L is a pure s donor or a p acceptor, the value of N–Re–N is larger than that of the free fragment [Re(NO)2(PR3)2]1. In contrast, if L is a p donor, we expect to find a decrease in N–Re–N. This might allow us to judge the relative importance of p acceptor vs.p donor interaction. From ¶ Strictly speaking, the metal d contribution in 1a1 is a mixture of dz2 and dx2 2 y2, and in complex 2 both components are of equal importance. For the sake of convenience, we keep referring to dz2 in the 1a1 case since only this orbital participates in the s antibonding interaction with L; further details are to be found in Table 4.J. Chem. Soc., Dalton Trans., 1999, 1717–1727 1723 the values presented in Table 2, we see that, in addition to H2 and CO, NCCH3, PH3 and CN2 also show an increase in N–Re–N.Interestingly, for the acetonitrile complex we find a smaller value for this angle, which might indicate that in this case p-acceptor interaction is of only minor importance. The same holds true for the isonitrosyl ligand NO. Before continuing our discussion, we should explain why we included the somewhat unusual isonitrosyl ligand in the list of our model compounds. Initially, we wanted to find a simple model for complex V1 in order to investigate the nature of the NO bend and the unusual co-ordination geometry.To probe the influence of p donation on the co-ordination geometry of the nitrosyl ligands, we provided for starting geometries 8 and 91, in which the Re(NO)2(PR3)2 fragment adapted a similar arrangement to that found in the crystal structures of IV1 and V1. All attempts to optimize such an asymmetric structure, however, converged to the symmetric co-ordination geometry of 8 or 91.This was a first indication that no orbital eVect is probably responsible for the particular co-ordination geometry of V1. We then extended our calculations to the asymmetric complexes 101–121, and also considered steric eVects in our analysis. These results are presented in the next paragraph. Fig. 8 Walsh diagram along the N–Re–N bending mode for (a) complex 2, (b) 31 and (c) 8. Origin of the NO bend.The formaldehyde compound 101 already provides a good model of the benzaldehyde complex IV1. The calculation satisfactorily reproduces the main structural features of the experimentally determined structure. One of the NO ligands is bent by about 308, and the co-ordination geometry falls between TBP and TP (see data in Tables 1 and 4). The calculations on the hypothetical nitroso hydride 19 complex 111 provide an initial clue as to why one of the NO ligands deviates from a linear co-ordination geometry.For HNO two diVerent co-ordination geometries are possible, the first in which the hydrogen points away from the metal fragment, trans- 111, a second in which it is directed toward one of the NO ligands, namely cis-111. As can be seen from the data in Table 4, trans-111 adopts a co-ordination geometry close to that of TBP, Fig. 9 Molecular structures along the transformation pathway trans- 111 æÆ cis-111. The PH3 groups are omitted for clarity. See text for further details.1724 J.Chem. Soc., Dalton Trans., 1999, 1717–1727 whereas cis-111 displays the distorted TBP–TP arrangement, as found in 101 or IV1 (see also Fig. 9). To analyse the origin of this distortion we performed calculations for hypothetical molecules on the pathway trans-111Æ cis-111. Beginning with the fully optimized geometry of trans- 111, we introduce a hydrogen flip by a 1808 rotation around the ON axis of the nitroso hydride ligand, while keeping all other geometric parameters fixed.We then allow for the NO bend to adapt to the value of cis-111. Finally we let the complex relax to the fully optimized asymmetric geometry cis-111. This transformation is illustrated in Fig. 9. The corresponding orbital energy diagram of the highest three occupied orbitals is presented in Fig. 10. In light of this analysis it appears as though the symmetric cis structure should be the most stable one, and the geometry distortion to the final structure of cis-111 should not seem obvious.As anticipated, no orbital eVect is clearly responsible for the observed modification in complex geometry when the coordination of the HNO ligand is changed from trans to cis. We extended our analysis also to include steric eVects, and essentially decomposed the total bonding energy TBE of a given molecule into components due to repulsive steric interaction, DE0, and attractive orbital interaction, DEint.20 The energy decomposition along the pathway trans-111Æcis-111 is presented in Fig. 11. The energy contributions of trans-111 are set at zero. Hydrogen flipping leads to an energetic stabilization due to electronic interactions. As can be observed in Fig. 10, all the three dxz, dyz and dz2 orbitals are lowered in energy.|| However, we also find a considerable increase in steric repulsion, so that, as a net eVect, the hydrogen flip destabilizes the molecular arrangement by 20 kJ mol21. The NO bend now decreases the steric repulsion from 78 to 66 kJ mol21.The orbital interaction energy however is diminished, since the now partially oxidized metal center has an unfavorable TBP co-ordination geometry. In the last step the geometry relaxes from TBP to TP, which eVectively enhances the electronic interaction and further reduces the steric repulsion. Our analysis shows that the hydrogen of the HNO ligand, when pointing towards one nitrosyl group, leads to an increase in DE0. Bending of the aVected NO minimizes steric repulsion, and further rearrangement to the TP geometry maximizes electronic interaction.The same structural element, a hydrogen pointing toward a nitrosyl ligand, can be found in the case Fig. 10 Orbital energy diagram for the three highest occupied orbitals along the path trans-111 æÆ cis-111. || We adopt a simplified classification of the orbitals according to the Re-d contributions. To not confuse the reader, we prefer to keep the nomenclature as it was established for C2v symmetry, although the correct classification for the HOMO to HOMO-2 should be dx2 2 y2, dxy and dyz.Furthermore, in some cases we have substantial mixing between dx2 2 y2 and dxy. of the formaldehyde or benzaldehyde ligands in 101 or IV1, respectively. This hydrogen then induces the same structural changes discussed for the hypothetical nitroso hydride complex 111. The last question we want to address is whether or not steric repulsion is also responsible for the geometric distortion encountered in complex V1.To this end, we performed a bonding analysis of the model compound 121, by building up the final complex from the constituting fragments 11 and 2, eqn. (2). The energy associated with eqn. (2) is the so-called H(NO)(PH3)2ReNO 1 [Re(NO)2(PH3)2]1 æÆ 2 11 [H(NO)(PH3)2ReNOÆRe(NO)2(PH3)2]1 (2) 121 bond snapping energy BEsnap,21 since the fragments have already been promoted from their ground state geometry to the one they adopt in the final complex; BEsnap can again be broken down into steric and electronic contributions, eqn.(3). The BEsnap = 2[DE0 1 DEint] (3) bond analysis was performed not only for 121, but for a symmetrical compound sym-121 as well, which was constructed by adopting structural features from 121 [geometry of the H(NO)(PH3)2ReNO fragment and the phosphorus donor ligands, N–Re–N] and 91 (O–N–Re). The geometries of both model complexes are shown in Fig. 12, and the results of the bonding analysis are collected in Table 6.For the two coordination geometries the electronic interaction energy is virtually identical. Again, a reduced steric repulsion in 121 favors the Re–L bond in the asymmetric compound by 16 kJ mol21. In this section we have elucidated the role of DE0 in the co-ordination geometry of [Re(NO)2(PR3)2L]n1 complexes. In asymmetric co-ordination geometries one of the NO ligands bends to reduce steric repulsion between [Re(NO)2(PR3)2]1 and the ligand L.Noteworthy is the fact that this bending distortion does not require much energy; DEbend can be estimated as about 20 kJ mol21. The NO ligand seems to be very flexible in adapting to the right co-ordination geometry and eVectively minimizing steric repulsion; this is evident not only in the TBE–TP geometries of IV1 and V1, but also in the strong cisoid bends encountered in I1. Conclusion The co-ordination chemistry of the 16e2 fragment [Re(NO)2- (PR3)2]1 11 has been explored by means of crystal structure analyses and DFT calculations.The ion possesses a C2v butterfly ground state geometry. This arrangement could be Fig. 11 Energy decomposition along the path trans-111 æÆ cis-111. The total bonding energy (d, TBE) is divided into steric (j, DE0) and electronic (r, DEint) contributions. See text for further details.J. Chem. Soc., Dalton Trans., 1999, 1717–1727 1725 rationalized by simple arguments based on orbital interactions, similar to those employed for the isoelectronic compound [Ru(CO)2(PtBu2Me)2].13 The structural changes of the [Re(NO)2(PR3)2]1 under formation of [Re(NO)2(PR3)2L]n1 complexes (n = 0 or 1) have been used to characterize the nature of the Re–L bond.The angle P–Re–P is determined by the competition for p-back bonding between the nitrosyl groups and the ligand L. In symmetric complexes a new orbital eVect was found to determine the size of the N–Re–N angle.When L is a pure s donor or p acceptor the value of N–Re–N is larger than that of 11. In contrast, if L is a p donor, we expect to find a decrease in N–Re–N. In asymmetric complexes it was shown that the driving force in bending of one of the NO groups and the subsequent distortion from a TBP to a TBP–TP is the minimization of steric repulsion. We have also seen that this rearrangement is accompanied by only small changes in the bonding energy, and that the NO ligand is very flexible in adapting its co-ordination geometry to changes in electronic structure or steric influences. This might entail important implications for the chemistry and reactivity of 11.In this work we have investigated the structural and static features of the co-ordination chemistry of the [Re(NO)2- (PR3)2]1 fragment. This study is intended to provide a basis for a better understanding of the reactivity and dynamic features of this transition metal complex. We are currently investigating the potential of 11 as an eVective catalyst in hydrogenation and hydrosilation reactions.22 From a theoretical point of view, the nature of the intramolecular interaction between the bending nitrosyl groups and the ligand L provides an interesting challenge.Further investigations might reveal whether or not intramolecular hydrogen bonding can indeed be related to the phenomenon of NO bending. Fig. 12 Molecular structures of complex 121 and sym-121 in the plane of the NO ligands; PH3 groups are omitted for clarity.Table 6 Bond analyses a for the model complexes 121 and sym-121 DE0 DEint BEsnap 121 38 2213 175 sym-121 54 2212 158 a In kJ mol21. Experimental All operations were carried out under a nitrogen atmosphere using standard Schlenk and glove-box techniques. Solvents were dried over sodium diphenylketyl [THF, Et2O, O(SiMe3)2, hydrocarbons] or P2O5 (CH2Cl2) and distilled under N2 prior to use. The deuteriated solvents used in the NMR experiments were dried over sodium diphenylketyl (C6D6, toluene-d8, THF-d8) or P2O5 (C6D5Cl, CD2Cl2) and vacuum transferred for storage in Schlenk flasks fitted with Teflon stopcocks.All NMR experiments were carried out on a Varian Gemini 300 spectrometer. Chemical shifts are given in ppm. The 1H and 13C-{1H} NMR spectra were referenced to the residual proton or 13C resonances of the deuteriated solvent, 31P chemical shifts externally referenced to 85% H3PO4 sealed in a capillary and inserted into a standard 5 mm NMR tube filled with the deuteriated solvent.The IR spectra were recorded on a Bio-Rad FTS-45 spectrometer. The complex [Re(NO)2(PiPr3)2H] II was prepared according to a reported procedure.9 Benzaldehyde was purchased from Fluka (puriss.), degassed and used without further purification. For the crystal structure analyses, the diVraction data were collected on an image plate detector system (STOE IPDS) for complexes I1 and III1, and on a four circle diVractometer (upgraded Nicolet R3) for IV1 and V1.The X-ray generators were equipped with sealed tubes and graphite monochromators (Mo-Ka, l = 0.71073 Å). All crystals were mounted on glass rods or on top of glass capillaries using silicon grease (IV1, V1) or covered with perfluoro polyether oil (I1, III1). Programs used for cell refinement, data collection and data reduction: CELL,23 EXPOSE,23 INTEGRATE,23 XRED23 and XDISK;24 for absorption correction, numerical,25 XRED (I1, III1), and semiempirical based on y-scan data, XEMP (V1).24 Structure solution was done with SHELXS 9726 (I1, III1) and SIR 92 27 (IV1, V1).Structure refinement was done with SHELXL 9728 (I1, III1) and CRYSTALS 9629 (IV1, V1). All positions of the hydrogen atoms, except for the hydride of V1, were calculated at distances relevant for the measuring temperature, and were placed geometrically for each refinement cycle. Complexes I1 and III1 were refined on Fo 2 using all unique reflections, applying an empirical weighting scheme;28 IV1 and V1 were refined on Fo using reflections with I > s(I ), and a Chebyshev polynomial weighting scheme.30 Molecular graphics were done with PLATON 97.31 CCDC reference number 186/1421.See http://www.rsc.org/suppdata/dt/1999/1717/ for crystallographic files in .cif format. Preparations [Re(NO)2(PCy3)2][BArF 4]. A heterogeneous mixture containing [Re(NO)2(PCy3)2H] (150 mg, 0.189 mmol) and [Ph3C]- [BArF 4] (209 mg, 0.189 mmol) in C6H6 (15 mL) was stirred for 2 h.During this period a dark red oily solid forms. The solvent was removed in vacuo until ca. 3 mL of C6H6 were left, and then pentane was added (15 mL). The liquid was discharged and the solid washed with additional pentane (3 × 15 mL) and dried in vacuo to give 290 mg of I1[BArF 4] (91.9%). Crystals for the X-ray diVraction study were grown by cooling slowly, starting at 90 8C, a saturated C6H6 solution of the complex.IR(Nujol): nNO 1711m and 1649s cm21. 31P-{1H} NMR (C6D5Cl): d 46.5 (s). 1H NMR (C6D5Cl): d 8.10 (m, br, 8 H, BArF 4), 7.47 (m, br, 4 H, BArF 4) and 2.30–0.6 (m, 66 H, PCy3) (Calc. for C68H78BF24- N2O2P2Re: C, 48.90; H, 4.71; N, 1.68. Found: C, 48.72; H, 4.65; N, 1.57%). Crystal structure determination. The compound crystallizes with one molecule of C6H6 and one molecule of (C2H5)2O per unit cell, which are both disordered via a center of symmetry. Thus, the solvent molecules were refined isotropically.C73H86- BF24N2O2.5P2Re, M = 1746.39, triclinic, space group P1� (no. 2), a = 13.4230(14), b = 17.641(2), c = 17.946(2) Å, a = 101.790(13),1726 J. Chem. Soc., Dalton Trans., 1999, 1717–1727 b = 109.224(12), g = 92.608(13)8, V = 3898.4(0.8) Å3 (5000 reflections used for cell parameter refinement), T = 193 K, Z = 2, m(Mo-Ka) = 1.7 mm21, 218 images exposed using a f oscillation scan mode at constant times of 3.0 min per image. 35186 Reflections measured (qmax = 268), 13856 unique (Rint = 0.0463) which were used in all calculations, 934 parameters in full matrix refinement, final R1 = 0.0738, wR2(F2) = 0.1864.[Re(NO)2(PCy3)2(CO)][BArF 4]. The complex [Re(NO)2- (PCy3)2][BArF 4] (48 mg, 0.0287 mmol) was introduced in a 100 mL flask and C6H6 (10 mL) added. The mixture was placed under 950 mbar of CO and heated at 80 8C for 10 min. Upon cooling to room temperature small yellow crystals started to be formed. The solvent was removed until ca. 1 mL of C6H6 was left, and then pentane was added (10 mL). The solid was subsequently washed with pentane (2 × 10 mL) and dried under vacuum to give 35 mg of III1[BArF 4] (71.3%). Suitable crystals for the X-ray diVraction study and elemental analyses were obtained by recrystallization in CH2Cl2–pentane. IR(CD2Cl2): nCO 2025m; nNO 1717m and 1675s cm21. 31P-{1H} NMR (CD2Cl2): d 23.6 (s). 1H NMR (CD2Cl2): d 7.73 (m, br, 8 H, BArF 4), 7.60 (m, br, 4 H, BArF 4) and 2.40–0.8 (m, 66 H, PCy3). 13C-{1H} NMR (CD2Cl2): d 202.4 (t, CO, JCP = 9.4 Hz) (Calc for C69H78BF24N2O3P2Re: C, 48.80; H, 4.63; N, 1.65. Found: C, 49.11; H, 4.42; N, 1.58%). Crystal structure determination The compound crystallizes with one molecule of CH2Cl2 per unit cell, which is disordered via a center of symmetry. For the solvent molecule, the split Cl atoms were refined anisotropically, whereas the remaining atoms were treated isotropically. C69.5H79BClF24N2O3P2Re, M = 1740.75, triclinic, space group P1� (no. 2), a = 12.9909(12), b = 16.6700(16), c = 19.0701(19) Å, a = 79.633(12), b = 71.952(11), g = 79.440(11)8, V = 3826.4(0.6) Å3 (5000 reflections used for cell parameter refinement), T = 193 K, Z = 2, m(Mo- Ka) = 1.768 mm21, 200 images exposed using a f rotation scan mode at constant times of 1.8 min per image. 48209 Reflections measured (qmax = 308), 20707 unique (Rint = 0.0506) which were used in all calculations, 993 parameters in full matrix refinement.All three cyclohexyl groups bound to P2 are disordered (from diVerence electron density maps), and were refined using the PART option.27 Final R1 = 0.0616, wR2(F2) = 0.1977. [Re(NO)2(PCy3)2(C6H5CHO)][BArF 4]. A slurry of [Re(NO)2- (PCy3)2][BArF 4] (50 mg, 0.0299 mmol) in C6H6 (1 mL) was treated with benzaldehyde (10 mL, 0.0984 mmol). In a few minutes the starting material dissolved and a brown solution was obtained. Pentane was layered over this solution and after 24 h red-brown crystals were collected, washed with pentane (2 × 10 mL) and dried in vacuo yielding 40 mg of IV1[BArF 4](C6H6) (72.2%).IR(CD2Cl2): nCO,NO 1704w, 1668s, 1651 (sh), 1617s, 1611s, 1593s and 1575m cm21. 31P-{1H} NMR (CD2Cl2): d 32.5 (s). 1H NMR (CD2Cl2): d 9.92 (s, 1 H, C6H5CHO), 8.04–7.78 (m, 5 H, C6H5CHO), 7.73 (m, br, 8 H, BArF 4), 7.60 (m, br, 4 H, BArF 4) and 2.30–0.8 (m, 66 H, PCy3). 13C-{1H} NMR (CD2Cl2): d 206.9 (s, br, C6H5COH) [Calc. for C75H84BF24- N2O3P2Re (recrystallized in CH2Cl2–pentane): C, 50.71; H, 4.77; N, 1.58.Found: C, 50.67; H, 4.59; N, 1.58%]. Crystal structure determination. The compound crystallizes with three molecules C6H6 per asymmetric unit. Formula C93- H102BF24N2O3P2Re, M = 2010.79, triclinic, space group P1� (no. 2), a = 13.892(3), b = 18.768(3), c = 19.569(3) Å, a = 97.76(2), b = 107.96(2), g = 102.88(2)8,615.8(1.2) Å3, T = 153 K, Z = 2, m(Mo-Ka) = 1.46 mm21, w scan width 1.68, variable scan speed 2–298 min21, 16060 reflections measured (qmax = 258), 15223 unique (Rint = 0.030) which were used in all calculations, 1194 parameters in full matrix refinement, final R1 = 0.0539, wR(Fobs) = 0.0385.y-Scan reflections for absorption correction were measured, but did not lead to further improvement of the results. Therefore, the uncorrected data set was used in structure refinement. The F atoms for two of the trifluoromethyl groups had to be refined isotropically. One of the cyclohexyl groups appeared to be disorderd as well, and the four C atoms involved had to be split and refined with isotropic displacement parameters. [Re(NO)2(PiPr3)2{ONRe(NO)(PiPr3)2H}]BArF 4].A heterogeneous mixture of [Re(NO)2(PiPr3)2H] (165 mg, 0.299 mmol) and [Ph3C][BArF 4] (163 mg, 0.147 mmol) in C6H6 (15 mL) was stirred for 2 h. During this period an orange solid was formed. The solvent was removed in vacuo until ca. 3 mL of C6H6 were left and then pentane (15 mL) was added.The residue was washed with pentane (3 × 15 mL) and dried in vacuo to give 250 mg of V1[BArF 4] (83.7%). Crystals for the X-ray diVraction study were grown by recrystallization of a diluted solution of the complex from C6H6–pentane. IR(Nujol): nNO 1659s, 1645m, 1627s and 1609s cm21. 31P-{1H} NMR (C6D5Cl): d 54.9 (s, br, 2P) and 43.3 (s, br, 2P). 1H NMR (C6D5Cl): d 8.12 (m, br, 8H, BArF 4), 7.47 (m, br, 4 H, BArF 4), 3.39 (t, br, JHP = 46.8 Hz, Re), 2.30 [m, br, 6 H, P(CHMe2)3], 2.05 [m, br, 6 H, P(CHMe2)3] and 0.95 [m, 72 H, P(CHMe2)3] (Calc.for C68H97BF24N4- O4P4Re2: C, 40.89; H, 4.89; N, 2.80. Found: C, 40.95; H, 4.84; N, 2.78%). Crystal structure determination. The very small crystal size caused high residual electron density of 7.68 e Å23, 0.94 Å away from Re2. C68H97BF24N4O4P4Re2, M = 1997.61, triclinic, space group P1� (no. 2), a = 14.256(2), b = 16.859(2), c = 17.771(2) Å, a = 97.22(1), b = 93.87(1), g = 96.11(1)8, V = 4198.9(0.8) Å3, T = 183 K, Z = 2, m(Mo-Ka) = 3.09 mm21, w scan width 1.28, variable scan speed 2–298 min21, 15369 reflections measured (qmax = 258), 14579 unique (Rint = 0.020), 10254 reflections used in all calculations.Isopropyl groups are disordered; 962 parameters in full matrix refinement, final R1 = 0.1104, wR(Fobs) = 0.081. Owing to the small crystal dimensions, five C atoms of four isopropyl groups, as well as one C atom of the BArF 4 anion, had to be refined isotropically. Computational details All calculations were based on the local density approximation (LDA) in the parameterization of Vosko et al.32 with the addition of gradient corrections due to Becke 33 and Perdew34 (BP86), which were included self-consistently (NL-SCF). The calculations utilized the Amsterdam Density Functional package ADF,35 release 2.3.Use was made of the frozen core approximation, and the ns, np, nd and (n 1 1)s shells of the transition metal were described by a triple z-STO basis augmented by one (n 1 1)p function (ADF database IV).The valence shells of the main group atoms were described by a double z-STO basis plus one polarization function (ADF database III). The numerical accuracy 35b,d was set to 5.0, and final gradients were 2.0 × 1023 au Å21 and better. If not mentioned otherwise, calculations were performed under C2v or Cs symmetry constraints. Relativistic eVects were included using a quasi-relativistic approach.36 Acknowledgements This work was supported by the Swiss National Science Foundation (SNSF). Access to the computing facilities of the Rechenzentrum der Universität Zürich is gratefully acknowledged.References 1 G. B. Richter-Addo and P. Legdins, Metal Nitrosyls, Oxford University Press, New York, NY, 1992; M. Feelisch and J. S. Stammler (Editors), Methods in Nitric Oxide Research, Wiley, Chichester, 1996. 2 H. Berke and P. Burger, Comments Inorg. Chem., 1994, 16, 279. 3 A. A. H. van der Zeijden, T.Bürgi and H. Berke, Inorg. Chim. Acta, 1992, 201, 131. 4 A. A. H. van der Zeijden, C. Sontag, H. W. Bosch, V. Shklover,J. Chem. Soc., Dalton Trans., 1999, 1717–1727 1727 H. Berke, D. Nanz and W. von Philipsborn, Helv. Chim. Acta, 1991, 74, 1194; A. A. H. van der Zeijden, V. Shklover and H. Berke, Inorg. Chem., 1991, 23, 4393; A. A. H. van der Zeijden, H. W. Bosch and H. Berke, Organometallics, 1992, 11, 563; 2051. 5 H.-U. Hund, U. Ruppli and H. Berke, Helv. Chim.Acta, 1993, 76, 963; H.-U. Hund, Ph.D. Thesis, Zürich, 1991. 6 S. Feracin, T. Bürgi, V. I. Bakhmutov, I. Eremenko, E. V. Voronstov, A. B. Vimentis and H. Berke, Organometallics, 1994, 13, 4194; V. Bakhmutov, T. Bürgi, P. Burger, U. Ruppli and H. Berke, Organometallics, 1994, 13, 4203; E. S. Shubina, N. V Belkova, A. N. Krylov, E. V. Vorontsov, L. M. Epstein, D. G. Gusev, M. Niedermann and H. Berke, J. Am. Chem. Soc., 1996, 118, 1105; N. V. Belkova, E. S. Shubina, A. V. Ionidis, L.M. Epstein, H. Jacobsen, A. Messmer and H. Berke, Inorg. Chem., 1997, 36, 1522. 7 D. Nietlispach, H. W. Bosch and H. Berke, Chem. Ber., 1994, 127, 2403. 8 D. G. Gusev, A. Llamazares, G. Artus, H. Jacobsen and H. Berke, Organometallics, 1999, 17, 75. 9 T. Ziegler, Chem. Rev., 1991, 91, 651; J. Labanowski and J. Andzelm (Editors), Density Functional Methods in Chemistry, Springer, Heidelberg, 1991; T. Ziegler, Can. J. Chem., 1995, 73, 743; B. B. Laird, R. B. Ross and T.Ziegler, ACS Symp. Ser., 1996, 629. 10 (a) D. M. P. Mingos and J. A. Ibers, Inorg. Chem., 1970, 9, 1105; (b) A. P. Gaughan Junior, B. J. Corden, R. Eisenberg and J. A. Ibers, Inorg. Chem., 1974, 13, 786; (c) V. G. Albano, A. Araneo, P. L. Bellon, G. Ciani and M. Manassero, J. Organomet. Chem., 1974, 67, 413; (d ) B. L. Haymore and J. A. Ibers, Inorg. Chem., 1975, 14, 2610; (e) J. A. Kaduk and J. A. Ibers, Inorg. Chem., 1975, 14, 3070; ( f ) S. Bhaduri and G. M. Sheldrick, Acta Crystallogr., Sect.B, 1975, 31, 897; ( g) B. E. Reichert, Acta Crystallogr., Sect. B, 1976, 32, 1934; (h) J. A. Kaduk and J. A. Ibers, Inorg. Chem., 1977, 16, 3283; (i) A. M. M. Lanfredi, A. Tiripicchio and M. Tiripicchio Camellini, Acta Crystallogr., Sect. C, 1983, 39, 1633; ( j) G. Le Borgne, L. Mordenti, J. G. Riess and J.-L. Roustan, New J. Chem., 1986, 10, 97; (k) H. Li Kam Wah, M. Postel and M. Pierrot, Inorg. Chim. Acta, 1989, 165, 215; (l) V. Munyejabo, J.-P. Damiano, M.Postel, C. Bensimon and J.-L. Roustan, J. Organomet. Chem., 1995, 491, 61; (m) F. L. Atkinson, H. E. Blackwell, N. C. Brown, N. G. Connelly, J. G. Crossley, A. G. Orpen, A. L. Rieger and P. H. Rieger, J. Chem. Soc., Dalton Trans., 1996, 3491. 11 R. H. Summerville and R. HoVmann, J. Am. Chem. Soc., 1976, 98, 7240. 12 S. C. Haefner, K. R. Dunbar and C. Bender, J. Am. Chem. Soc., 1991, 113, 9540. 13 M. Ogasawara, S. A. Macgregor, W. E. Streib, K. Folting, O. Eisenstein and K. G. Caulton, J. Am. Chem. Soc., (a) 1995, 117, 8869; (b) 1996, 118, 10189. 14 T. A. Albright, J. K. Burdett and M.-H. Whangbo, Orbital Interactions in Chemistry, Wiley, New York, NY, 1985, (a) ch. 19.1; (b) ch. 15.4. 15 M. Elian and R. HoVmann, Inorg. Chem., 1975, 14, 1058. 16 K.-P. Huber and G. Herzberg, Molecular Spectra and Molecular Structure, Van Nostrand Reinhold, New York, NY, 1979, vol. IV. 17 R. HoVmann, L. M. Chen, M. Ellian, A. R. Rossi and D. M. P. Mingos, Inorg. Chem., 1974, 13, 2666; A. R. Rossi and R. HoV- mann, Inorg. Chem., 1975, 14, 365; R. HoVmann, J. M. Howell and A. R. Rossi, J. Am. Chem. Soc., 1976, 98, 2484. 18 (a) E. L. Muetterties, Tetrahedron, 1974, 30, 1595; (b) E. L. Muetterties and L. J. Guggenberger, J. Am. Chem. Soc., 1974, 96, 1748. 19 D. H. Mordaunt, H. Flöthmann, M. Stumpf, H.-M. Keller, C. Beck, R. Schinke and K. Yamashita, J. Chem. Phys., 1997, 107, 6603. 20 T. Ziegler and A. Rauk, Theor. Chim. Acta, 1977, 46, 1; Inorg. Chem., 1979, 18, 1558; T. Ziegler, NATO ASI, Ser. C, 1992, 378, 367; F. M. Bickelhaupt, N. M. M. Nibbering, E. M. van Wezenbeek and E. J. Baerends, J. Phys. Chem., 1992, 96, 4864. 21 H. Jacobsen and T. Ziegler J. Am. Chem. Soc., 1994, 116, 3667. 22 A. Llamazares, H. W. Schmalle and H. Berke, in preparation. 23 DS Software, Stoe and Gie, Darmstadt, Germany, 1997. 24 SHELXTL PLUS, Siemens Analytical X-Ray Instruments, Madison, WI, 1994. 25 P. Coppens, L. Leiserowitz and D. Rabinovich, Acta Crystallogr., 1965, 18, 1035. 26 G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467. 27 A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi, M. C. Burla, G. Polidori and M. Camalli, SIR 92, University of Bari, 1992. 28 G. M. Sheldrick, SHELXL 97, University of Göttingen, 1997. 29 D. J. Watkin, C. K. Prout, J. R. Carruthers and P. W. Bettridge, CRYSTALS, Issue 10, Chemical Crystallography Library, University of Oxford, Oxford, 1996. 30 J. R. Carruthers and D. J. Watkin, Acta Crystallogr., Sect. A, 1979, 35, 698. 31 A. L. Spek, Acta Crystallogr., Sect. A, 1990, 46, C34. 32 S. J. Vosko, M. Wilk and M. Nussair, Can. J. Phys., 1980, 58, 1200. 33 A. D. Becke, Phys. Rev. A, 1988, 38, 3098. 34 J. P. Perdew, Phys. Rev. B, 1986, 33, 8822. 35 (a) E. J. Baerends, D. E. Ellis and P. E. Ros, Chem. Phys., 1973, 2, 41; (b) G. te Velde and E. J. Baerends, J. Comput. Phys., 1992, 99, 84; (c) C. Fonseca Guerra, O. Visser, J. G. Snijders, G. te Velde and E. J. Baerends, In Methods and Techniques in Computational Chemistry: METECC-95, eds. E. Clementi and G. Corongiu, STEF, Cagliari, 1995, p. 305; (d ) G. te Velde, ADF 2.1 User’s Guide, Vrije Universiteit, Amsterdam, 1996. 36 T. Ziegler, V. Tschinke, E. J. Baerends, J. G. Snijders and W. Ravenek, J. Phys. Chem., 1989, 93, 3050; G. Schreckenbach, J. Li and T. Ziegler, Int. J. Quantum. Chem., 1995, 56, 477. Paper 9/01384I