View PDF VersionPrevious ArticleNext Article
 
DOI: 10.1039/C5RA22169B (Paper) RSC Adv., 2016, 6, 16131-16140

Binuclear nickel carbonyls with the small bite chelating diphosphine ligands methylaminobis(difluorophosphine) and methylenebis(dimethylphosphine): formation of Ni[double bond, length as m-dash]Ni double bonds in preference to ligand cleavage†

Ling Liuab, Xiuhui Zhang*ab, Zesheng Li*ab, Qianshu Libc and R. Bruce King*cd
aKey Laboratory of Cluster Science, Ministry of Education of China, School of Chemistry, Beijing Institute of Technology, Beijing 100081, China. E-mail: zhangxiuhui@bit.edu.cn; zeshengli@bit.edu.cn
bState Key Laboratory of Explosion Science and Technology, Beijing Institute of Technology, Beijing 100081, China
cMOE Key Laboratory of Theoretical Chemistry of Environment, Center for Computational Quantum Chemistry, South China Normal University, Guangzhou 510006, China. E-mail: rbking@chem.uga.edu
dDepartment of Chemistry and Center for Computational Chemistry, University of Georgia, Athens, Georgia 30606, USA. Tel: +1-706-542-1901

Received 22nd October 2015 , Accepted 20th January 2016

First published on 25th January 2016

Abstract

The structures and thermochemistry of the triads of binuclear nickel carbonyl complexes (bid)Ni2(CO)n (n = 6, 5, 4) and (bid)2Ni(CO)n (n = 4, 3, 2) of the small-bite bidentate chelating diphosphines CH3N(PF2)2 and (Me2P)2CH2 have been investigated using density functional theory. The lowest-energy structures of the carbonyl-richest (bid)Ni2(CO)6 and (bid)2Ni2(CO)4 structures have long Ni⋯Ni distances indicating the lack of direct nickel bonds. Similarly, the lowest energy structures of the intermediate (bid)Ni2(CO)5 and (bid)2Ni2(CO)3 systems have Ni–Ni distances of ∼2.7 Å and intact diphosphine ligands. Furthermore, the lowest energy structures of the carbonyl-poorest (bid)Ni2(CO)4 and (bid)2Ni2(CO)2 systems have shorter Ni[double bond, length as m-dash]Ni distances of ∼2.5 Å suggesting formal double bonds and retain the intact diphosphine ligands. This contrasts with the previously studied binuclear iron carbonyls [CH3N(PF2)2]Fe2(CO)6 and [CH3N(PF2)2]2Fe2(CO)4 for which ligand cleavage to separate CH3NPF2 and PF2 units rather than Fe[double bond, length as m-dash]Fe double bond formation occurs in the lowest energy structures. The experimental [(Me2P)2CH2]2Ni2(CO)4 structure with the boat form of the NiPCPNiPCP eight-membered lies ∼0.5 kcal mol−1 in energy below the higher energy isomer with the chair form of the NiPCPNiPCP ring at the M06-L/TZP level of theory.

1. Introduction

Transition metal complexes of small bite bidentate phosphines of the general type Z(PX2)2 (Z = one atom bridging group; X = alkyl, aryl, or halogen) are of interest in forming four-membered chelate rings with a single metal atom or five-membered chelate rings with an M–M unit (Fig. 1). Since five-membered chelate rings are generally more favorable than four-membered chelate rings, such small bite bidentate phosphines are of particular interest in stabilizing metal–metal bonded binuclear complexes. In a search for small bite bidentate phosphines that are also strong back-bonding ligands like carbon monoxide and PF3, methylaminobis(difluorophosphine), MeN(PF2)2, is of particular interest1 because of its facile synthesis from methylamine, phosphorus trichloride, and antimony trifluoride2 and its resemblance to the strong back-bonding ligand PF3. Despite the electron releasing properties of the bridging methylamino group, MeN(PF2)2 is still a sufficiently strong back-bonding ligand to form the white volatile air-stable zerovalent metal derivatives [MeN(PF2)2]3M (M = Cr, Mo, W) containing four-membered carbon-free NP2M chelate rings analogous to the corresponding hexacarbonyls M(CO)6.3,4
image file: c5ra22169b-f1.tif
Fig. 1 Chelating of small bite bidentate diphosphines to single metals and to binuclear M2 units. For the chelate small bite bidentate diphosphines discussed in this paper X = F and Z = MeN for MeN(PF2)2 and X = Me and Z = CH2 for (Me2P)2CH2.

In addition to forming the mononuclear [MeN(PF2)2]3M (M = Cr, Mo, W) complexes containing four-membered NP2M chelate rings, the MeN(PF2)2 ligand also forms binuclear metal carbonyl complexes with five-membered NP2M2 chelate rings (Fig. 1). For example, reaction of Fe3(CO)12 with excess CH3N(PF2)2 in boiling tetrahydrofuran initially gives the yellow diiron hexacarbonyl derivative [CH3N(PF2)2]2Fe2(CO)6, shown by X-ray crystallography to have a structure consisting of two square pyramidal P2Fe(CO)3 units linked by the two ligands without an iron–iron bond (Fig. 2).5 This structure can be considered as two five-coordinate L2Fe(CO)3 units related to the well-known Fe(CO)5 and linked by the MeN(PF2)2 ligand leading to an eight-membered FePNPFePNP ring. This product was relatively unstable evolving carbon monoxide in solution to give the diiron pentacarbonyl derivative [CH3N(PF2)2]2Fe2(CO)4(μ-CO), which could also be obtained directly by reaction of CH3N(PF2)2 with Fe3(CO)12 under more forcing conditions. This product is related to the well-known Fe2(CO)9 by replacing four carbonyl groups with two bridging MeN(PF2)2 ligands forming a pair of fused five-membered NP2Fe2 chelate rings sharing the Fe–Fe edge.


image file: c5ra22169b-f2.tif
Fig. 2 Binuclear [CH3N(PF2)2]2Fe2(CO)n derivatives that have been synthesized and structurally characterized by X-ray crystallography.

A limitation in the ability to synthesize MeN(PF2)2 complexes is the susceptibility one of its P–N bonds to undergo cleavage to give separate PF2 and MeNPF2 units that each can bridge the pair of central metal atoms. This process occurs particularly when reactions of MeN(PF2)2 with metal carbonyls are carried out under forcing conditions in an attempt to replace all or almost all of the carbonyl groups with MeN(PF2)2 ligands. Thus the product CH3N(PF2)2]4Fe2(CO) obtained by photolysis of Fe3(CO)12 with excess MeN(PF2)2 is not a maximally substituted Fe2(CO)9 derivative but instead (CH3N[double bond, length as m-dash]PF2)[μ-CH3N(PF2)2]3Fe2(μ-PF2)(CO) in which one of the CH3N(PF2)2 ligands has undergone phosphorus–nitrogen bond cleavage to form separate CH3NPF2 and PF2 units (Fig. 3).6 Similarly, the photolysis of (η5-C5H5)2Fe2(CO)2(μ-CO)2 with excess CH3N(PF2)2 leads to complete substitution of the four carbonyl groups with two CH3N(PF2)2 ligands, similar to the Cr(CO)6/CH3N(PF2)2 photolysis noted above. However, the product is shown by X-ray crystallography to be [μ-CH3N(PF2)2](μ-CH3NPF2)(μ-PF2)Fe2(η5-C5H5)2 in which the two η5-C5H5Fe units are bridged by three different groups, namely an intact CH3N(PF2)2 ligand, a CH3NPF2 unit, and a PF2 unit (Fig. 3).7 The bridging CH3NPF2 + PF2 combination donates a total of six electrons to the dimetal system rather than only four electrons from an intact CH3N(PF2)2 unit and is thus likely to arise in unsaturated metal complexes. Theoretical work on [MeN(PF2)2]2Fe2(CO)n derivatives (n = 6, 5, 4) predicts the experimental structures (Fig. 2) for n = 6 and 5 but an Fe2(CO)4(μ-CH3NPF2)2(μ-PF2) structure with separate bridging PF2 and CH3N(PF2)2 units.8


image file: c5ra22169b-f3.tif
Fig. 3 Examples of metal complexes in which an MeN(PF2)2 ligand has split into separate MeNPF2 and PF2 units.

The other small bite chelating bidentate diphosphine of interest is methylenebis(dimethylphosphine) (Me2P)2CH2, in which each phosphorus atom is bonded only to carbon atoms. This small bite bidentate diphosphine is the opposite of MeN(PF2)2 in being a strongly basic ligand with limited back-bonding ability. The small size and high basicity of (Me2P)2CH2 lead to metal carbonyl complexes having a relatively electron-rich and sterically unencumbered central metal atom. This is most obviously reflected in an increased air sensitivity of metal carbonyl complexes of (Me2P)2CH2.9

Homoleptic binuclear nickel carbonyls Ni2(CO)n has never been synthesised, although they have been studied by density functional theory.10 However, small bite bidentate diphosphine ligands are able to stabilize binuclear nickel carbonyls.11 Thus the reaction of Ni(CO)4 with the small bite biphosphine (Me2P)2CH2 has been reported to yield [(Me2P)2CH2]2Ni2(CO)4 with two (Me2P)2CH2 molecules serving as bridging Ni(CO)2 groups forming an eight-membered Ni2P4C2 ring12 (Fig. 4) related to the [MeN(PF2)2]2Fe2(CO)6 structure in Fig. 2. The reaction of [C5H5NiCO]2 with (Me2P)2CH2 in boiling tetrahydrofuran results in displacement of the cyclopentadienyl rather than carbonyl groups to give yellow [(Me2P)2CH2]3Ni2(CO)2.9 A related [MeN(PF2)2]3Ni2(CO)2 complex of methylaminobis(difluorophosphine) has also been synthesized.4


image file: c5ra22169b-f4.tif
Fig. 4 Experimentally known binuclear nickel carbonyl complexes of the small bite chelating diphosphines MeN(PF2)2 and (Me2P)2CH2.

The present paper explores the chemistry of binuclear nickel carbonyl complexes of MeN(PF2)2 and (Me2P)2CH2 using density functional theory (DFT). This research has included the metal complex triads [MeN(PF2)2]Ni2(CO)n (n = 6, 5, 4), [MeN(PF2)2]2Ni2(CO)n (n = 4, 3, 2), [(Me2P)2CH2]Ni2(CO)n (n = 6, 5, 4), and [(Me2P)2CH2]2Ni2(CO)n (n = 4, 3, 2). The stoichiometries of these triads were chosen to cover possible structures without a nickel–nickel bond and with nickel–nickel bonds if the ligands remain intact and the nickel atoms have the favored 18-electron configuration.13–17 In contrast to the previous study8 on binuclear CH3N(PF2)2 iron carbonyl complexes, nickel–nickel bonds in all the low-energy structures were found in preference to phosphorus–nitrogen or phosphorus–carbon bond cleavage of the ligands.

2. Theoretical methods

Electron correlation effects were considered by using density functional theory (DFT) methods, which have been suggested as a practical and effective computational tool, especially for organometallic compounds.18–25 Thus, in the present study, the computations were carried out by two selected DFT methods implemented in the Gaussian 09 program package,26 namely B3LYP, BP86, and M06-L. The B3LYP method is the hybrid HF-DFT method using a combination of the three-parameter Becke functional (B3) with the Lee–Yang–Parr (LYP) generalized gradient correlation functional.27,28 The BP86 method combines Becke's 1988 exchange functional (B)29 with Perdew's 1986 gradient corrected correlation functional (P86).30 The local density functional, M06-L,31,32 is designed to capture the main dependence of the exchange–correlation energy on local spin density, spin density gradient, and spin kinetic energy density. It is parametrized to satisfy the uniform-electron-gas limit and to have good performance for both main-group chemistry and transition metal chemistry.

The calculations were performed using the double-ζ plus polarization (DZP) basis sets. The DZP basis sets used for carbon, nitrogen, oxygen, fluorine, and phosphorus atoms add one set of pure spherical harmonic d functions with orbital exponents αd(C) = 0.75, αd(N) = 0.80, αd(O) = 0.85, αd(F) = 1.0, and αd(P) = 0.6 to the standard Huzinaga-Dunning contracted DZ sets33,34 and are designated (9s5p1d/4s2p1d) for the carbon, nitrogen, oxygen and fluorine atoms and (11s7p1d/6s4p1d) for the phosphorus atom. For hydrogen, a set of p polarization functions αp(H) = 0.75 is added to the Huzinaga-Dunning DZ set. The loosely contracted DZP basis set for nickel is the Wachter's primitive set35 augmented by two sets of p functions and a set of d functions contracted according to Hood, Pitzer, and Schaefer designated (14s11p6d/10s8p3d).36

The geometries of all structures were fully optimized using the DZP B3LYP and DZP BP86 methods. The fine grid (75 radial shells, 302 angular points)37 was the default for evaluating integrals numerically, and the tight (10−8 hartree) designation was the default for the self-consistent field (SCF) energy convergence. The finer (120, 974) integration grid37 was used only to re-examining the small imaginary vibrational frequencies.

In order to test the computational accuracy of the above two DFT methods, we initially optimized the experimentally known compound12 [(Me2P)2CH2]2Ni2(CO)4 at the B3LYP/DZP and BP86/DZP levels of theory. The lowest-energy optimized structure was different from the experimental structure. In order to resolve this discrepancy, we reoptimized the low-energy singlet structures in the present paper employing a new local density functional of M06-L with triple-ζ valence quality plus polarization38 (TZP) basis sets using the empirical dispersion correction by Grimme's DFT-D3 method.39 The most stable singlet structure of [(Me2P)2CH2]2Ni2(CO)4 predicted by M06-L/TZP is similar to the experimental structure. In addition, the results predicted by the three functionals are generally in agreement. Therefore, only the M06-L/TZP results for the structure geometries and relative energies are discussed in the text. However, the BP86/DZP method appears to be more reliable for the prediction of ν(CO) frequencies, probably coincidentally.40,41 The results of the B3LYP/DZP and BP86/DZP methods are presented in the ESI.† In addition, the energies of all of the triplet structures were found to be higher than those of the isomeric singlet structures. Therefore the triplet structures are not discussed in this paper but are presented in the ESI.†

Each structure in this paper is designated as PF-XY-EN or PMe-XY-EN, where PF and PMe refer to the ligands MeN(PF2)2 and (Me2P)2CH2, respectively; X refers to the number of ligands; Y refers to the number of carbonyl groups; E refers to the order of relative energies, and N indicates the spin state multiplicity (S = singlet and T = triplet) for the triads [MeN(PF2)2]Ni2(CO)n (n = 6, 5, 4), [MeN(PF2)2]2Ni2(CO)n (n = 4, 3, 2), [(Me2P)2CH2]Ni2(CO)n (n = 6, 5, 4), and [(Me2P)2CH2]2Ni2(CO)n (n = 4, 3, 2). Thus, the global minimum of [MeN(PF2)2]Ni2(CO)6 is designated as PF-16-1S and that of [(Me2P)2CH2]2Ni2(CO)4 as PMe-24-1S.

3. Results and discussion

3.1 Structures with one MeN(PF2)2 ligand

3.1.1 [MeN(PF2)2]Ni2(CO)6. The structure PF-16-1S is predicted to be the lowest energy [MeN(PF2)2]Ni2(CO)6 structure (Fig. 5). In PF-16-1S the two Ni(CO)3 groups are arranged in syn positions relative to the central MeN(PF2)2 ligand. A slightly higher energy singlet [MeN(PF2)2]Ni2(CO)6 structure PF-16-2S, lying 0.8 kcal mol−1 in energy above PF-16-1S, is similar to PF-16-1S except for a anti rather than syn arrangement of the two Ni(CO)3 groups relative to the central MeN(PF2)2 ligand. The singlet [MeN(PF2)2]Ni2(CO)6 structure PF-16-3S, lying 2.3 kcal mol−1 above PF-16-1S, is similar to PF-16-1S and PF-16-2S but with a still different arrangement of the Ni(CO)3 groups relative to the MeN(PF2)2 ligand. All three [MeN(PF2)2]Ni2(CO)6 structures have long Ni⋯Ni distances indicating the lack of direct nickel–nickel bonds. Nevertheless, each nickel atom in each of the three [MeN(PF2)2]Ni2(CO)6 structures has the favored 18-electron configuration. The predicted ν(CO) frequencies from 2163 to 2085 cm−1 in the three [MeN(PF2)2]Ni2(CO)6 structures by the M06-L/TZP method are higher than the experimentally observed range of 2090 to 1985 cm−1 and indicate exclusively terminal CO groups in these structures.5,9 The ν(CO) frequencies from 2069 to 2006 cm−1 predicted by the BP86/DZP method for the [MeN(PF2)2]Ni2(CO)6 structures are closer to the experimental values in accord with expectation.
image file: c5ra22169b-f5.tif
Fig. 5 The three singlet [MeN(PF2)2]Ni2(CO)6 structures. The numbers for bond distances (Å) and the relative energies are shown in parentheses were obtained by the M06-L method. The data in all of the other figures in the present paper have the same arrangement.
3.1.2 [MeN(PF2)2]Ni2(CO)5. Two singlet energetically low-lying [MeN(PF2)2]Ni2(CO)5 structures were found (Fig. 6). The lowest lying [MeN(PF2)2]Ni2(CO)5 structure is the singlet PF-15-1S, having an Ni–Ni distance of 2.632 Å, corresponding to a formal single bond thereby giving each nickel atom the favored 18-electron configuration. This Ni–Ni bond in PF-15-1S is bridged by a single CO group, predicted by the M06-L/TZP method to exhibit a ν(CO) frequency of 1971 cm−1. This bridging ν(CO) frequency in PF-15-1S is significantly lower than the terminal ν(CO) frequencies in PF-15-1S, ranging from 2148 to 2093 cm−1. The other singlet [MeN(PF2)2]Ni2(CO)5 structure PF-15-2S lies 1.1 kcal mol−1 in energy above PF-15-1S. Structure PF-15-2S is similar to PF-15-1S except for the approximate right angle rather than a straight angle between the bridging CO group and the face formed by the Ni–Ni bond and the bridging MeN(PF2)2 ligand.
image file: c5ra22169b-f6.tif
Fig. 6 The two singlet [MeN(PF2)2]Ni2(CO)5 structures.
3.1.3 [MeN(PF2)2]Ni2(CO)4. Two singlet structures were optimized for [MeN(PF2)2]Ni2(CO)4 (Fig. 7). The lowest energy [MeN(PF2)2]Ni2(CO)4 structure, PF-14-1S, has an intact bridging MeN(PF2)2 ligand and a Ni[double bond, length as m-dash]Ni distance of 2.361 Å. The Ni[double bond, length as m-dash]Ni distance in PF-14-1S is ∼0.27 Å shorter than the Ni–Ni single bond distance in the [MeN(PF2)2]Ni2(CO)5 structures discussed above and thus suggests a formal double bond in PF-14-1S. Such a Ni[double bond, length as m-dash]Ni double bond gives each nickel atom in PF-14-1S the favored 18-electron configuration. Two of the CO groups in PF-14-1S are bridging CO groups exhibiting a relatively low ν(CO) frequency of 2002 cm−1 (M06-L/TZP) or 1884 cm−1 (BP86/DZP).
image file: c5ra22169b-f7.tif
Fig. 7 The two singlet [MeN(PF2)2]Ni2(CO)4 structures.

The singlet [MeN(PF2)2]Ni2(CO)4 structure PF-14-2S lies at the relatively high energy of 14.2 kcal mol−1 above PF-14-1S. The MeN(PF2)2 ligand in PF-14-1S has broken into seperate PF2 and MeNPF2 units by P–N bond rupture. The Ni–Ni distance of 2.602 Å in PF-14-2S is similar to that in the [MeN(PF2)2]Ni2(CO)5 structures PF-15-1S and PF-15-2S (Fig. 6) and thus suggests the formal single bond required to give each nickel atom the favored 18-electron configuration in a [MeN(PF2)2]Ni2(CO)4 structure in which the MeN(PF2)2 ligand has ruptured into separate MeNPF2 and PF2 units.

3.2 Structures with two MeN(PF2)2 ligands

3.2.1 [MeN(PF2)2]2Ni2(CO)4. Two singlet [MeN(PF2)2]2Ni2(CO)4 structures were found (Fig. 8). The lowest energy structure PF-24-1S has C2v symmetry with two intact MeN(PF2)2 ligands bridging the pair of nickel atoms. The long Ni⋯Ni distance of 4.046 Å in PF-24-1S clearly indicates the absence of a direct nickel–nickel bond thereby giving each nickel atom the favored 18-electron configuration. The predicted ν(CO) frequencies from 2141 to 2084 cm−1 indicate exclusively terminal CO groups in PF-24-1S. One of the two terminal CO groups on each nickel atom in PF-24-1S is oriented towards the same side of the eight-membered NiPNPNiPNP ring corresponding to a syn stereoisomer. A slightly higher energy singlet [MeN(PF2)2]2Ni2(CO)4 structure PF-24-2S, lying 1.5 kcal mol−1 above PF-24-1S, is the anti stereoisomer of PF-24-1S with terminal CO groups on opposite sides of the eight-membered ring.
image file: c5ra22169b-f8.tif
Fig. 8 The two singlet [MeN(PF2)2]2Ni2(CO)4 structures.
3.2.2 [MeN(PF2)2]2Ni2(CO)3. Two singlet [MeN(PF2)2]2Ni2(CO)3 structures were found (Fig. 9). In the lowest energy structure PF-23-1S the nickel atoms form an isosceles triangle with the bridging carbonyl carbon. The predicted Ni–Ni distance in PF-23-1S of 2.574 Å can be interpreted as a formal single bond thereby giving each nickel atom the favored 18-electron configuration. The bridging carbonyl group in PF-23-1S has a predicted infrared ν(CO) frequency of 1958 cm−1 (M06-L/TZP) or 1882 cm−1 (BP86/DZP).
image file: c5ra22169b-f9.tif
Fig. 9 The two singlet [MeN(PF2)2]2Ni2(CO)3 structures.

The next singlet [MeN(PF2)2]2Ni2(CO)3 structure PF-23-2S, lying 8.0 kcal mol−1 in energy above PF-23-1S, can be derived from PF-24-1S by removal of one of the terminal CO groups (Fig. 9). The Ni → Ni distance of 2.590 Å in PF-23-2S can be interpreted as a formal dative single bond from the nickel atom bearing two CO groups to the nickel atom bearing only a single CO group. This gives each nickel atom the favored 18-electron configuration. The predicted ν(CO) frequencies from 2094 to 2132 cm−1 (M06-L/TZP) in PF-23-2S indicate exclusively terminal CO groups.

3.2.3 [MeN(PF2)2]2Ni2(CO)2. Two singlet [MeN(PF2)2]2Ni2(CO)2 structures were found (Fig. 10). The predicted Ni[double bond, length as m-dash]Ni distance of 2.393 Å in the lowest energy structure PF-22-1S is ∼0.2 Å shorter than the Ni–Ni single bond distances in the [MeN(PF2)2]Ni2(CO)5 (Fig. 6) and [MeN(PF2)2]2Ni2(CO)3 (Fig. 9) structures thereby suggesting a formal double bond in PF-22-1S. One of the two CO groups in PF-22-1S is a bridging CO group exhibiting a relatively low ν(CO) frequency of 1959 cm−1 (M06-L/TZP). The combination of two intact bridging MeN(PF2)2 ligands and a Ni[double bond, length as m-dash]Ni double bond gives each nickel atom in PF-22-1S the favored 18-electron configuration.
image file: c5ra22169b-f10.tif
Fig. 10 The two singlet [MeN(PF2)2]2Ni2(CO)2 structures.

The relatively high energy [MeN(PF2)2]2Ni2(CO)2 structure PF-22-2S, lying 22.0 kcal mol−1 above PF-22-1S, is similar to PF-22-1S except for a different angle of the terminal CO group relative to the Ni[double bond, length as m-dash]Ni double bond.

3.3 Structures with one (Me2P)2CH2 ligand

3.3.1 [(Me2P)2CH2]Ni2(CO)6. Three singlet [(Me2P)2CH2]Ni2(CO)6 structures were found within ∼5 kcal mol−1 of energy indicating a fluxional system. The lowest energy [(Me2P)2CH2]Ni2(CO)6 structure PMe-16-1S has the two Ni(CO)3 groups in anti positions relative to the central (Me2P)2CH2 ligand (Fig. 11). A slightly higher energy singlet [(Me2P)2CH2]Ni2(CO)6 structure PMe16-2S, lying 0.8 kcal mol−1 above PMe-16-1S, is similar to PMe-16-1S except for a syn rather than an anti arrangement of the two Ni(CO)3 groups relative to the central (Me2P)2CH2 ligand. This leads to a shorter, but still long non-bonding Ni⋯Ni distance of 4.599 Å in PMe-16-2S, clearly indicating the lack of a direct nickel–nickel bond. The remaining low-energy [(Me2P)2CH2]Ni2(CO)6 structure PMe-16-3S, lying 4.4 kcal mol−1 above PMe-16-1S, is also a syn stereoisomer similar to PMe16–2S but with a different orientation of the (Me2P)2CH2 ligand relative to the Ni(CO)3 units. The predicted ν(CO) frequencies in the three [(Me2P)2CH2]Ni2(CO)6 structures range from 2052 cm−1 to 2130 cm−1 (M06-L/TZP) indicating exclusively terminal CO groups. Each nickel atom in each of the three [(Me2P)2CH2]Ni2(CO)6 structures has the favored 18-electron configuration.
image file: c5ra22169b-f11.tif
Fig. 11 The three singlet [(Me2P)2CH2]Ni2(CO)6 structures. Hydrogen atoms are omitted for clarity in the structures of the (Me2P)2CH2 nickel carbonyl complexes.
3.3.2 [(Me2P)2CH2]Ni2(CO)5. Two singlet [(Me2P)2CH2]Ni2(CO)5 structures were optimized (Fig. 12). The lowest energy [(Me2P)2CH2]Ni2(CO)5 structure is the singlet PMe-15-1S, having a Ni–Ni distance of 2.607 Å, corresponding to a formal single bond thereby giving each nickel atom the favored 18-electron configuration. This Ni–Ni bond is bridged by both a single CO group and the (Me2P)2CH2 ligand. The bridging ν(CO) frequencies in PMe-15-1S of 1883 cm−1 (M06-L/TZP) are significantly lower than the terminal ν(CO) frequencies, which range from 2034 cm−1 to 2112 cm−1 (M06-L/TZP) in PMe-15-1S in accord with expectation.
image file: c5ra22169b-f12.tif
Fig. 12 The two singlet [(Me2P)2CH2]Ni2(CO)5 structures.

The [(Me2P)2CH2]Ni2(CO)5 structure PMe-15-1S lies a very large 38.5 kcal mol−1 below the next lowest energy isomer PMe-15-2S and thus appears to be a very favorable structure. The [(Me2P)2CH2]Ni2(CO)5 structure PMe-15-2S has one of the P–C bonds in (Me2P)2CH2 cleaved to form separate Me2P and Me2PCH2 units. Each of these units bridges the pair of nickel atoms thereby donating a total of three electrons each to the central Ni2 unit. Thus the combined Me2P + Me2PCH2 units obtained from cleaving a P–C bond in a (Me2P)2CH2 ligand donate a total of six electrons to the Ni2 unit. The long Ni⋯Ni distance of 3.889 Å in PMe-15-2S clearly indicates the lack of a direct nickel–nickel interaction. However, in PMe-15-2S both nickel atoms have the favored 18-electron configuration.

3.3.3 [(Me2P)2CH2]Ni2(CO)4. Two singlet [(Me2P)2CH2]Ni2(CO)4 structures were found (Fig. 13). The lowest energy structure PMe-14-1S has two symmetrical bridging CO groups and a short Ni[double bond, length as m-dash]Ni distance of 2.375 Å similar to the Ni[double bond, length as m-dash]Ni distances in the corresponding [MeN(PF2)2]2Ni2(CO)4 structures interpreted as formal double bonds. This gives each nickel atom in PMe-14-1S the favored 18-electron configuration.
image file: c5ra22169b-f13.tif
Fig. 13 The two singlet [(Me2P)2CH2]Ni2(CO)4 structures.

The [(Me2P)2CH2]Ni2(CO)4 structure PMe-14-1S lies 21.9 kcal mol−1 below the next lowest energy [(Me2P)2CH2]Ni2(CO)4 structure PMe-14-2S and thus appears to be a highly favorable structure. Structure PMe-14-2S has separate bridging Me2P and Me2PCH2 groups derived from the (Me2P)2CH2 ligand and can be derived from the [(Me2P)2CH2]Ni2(CO)4 structure PMe-15-2S (Fig. 12) by simple removal of a terminal CO group (Fig. 13). The Ni–Ni distance of 2.684 Å in PMe-14-2S suggests the formal single bond required to give each nickel atom the favored 18-electron configuration.

3.4 Structures with two (Me2P)2CH2 ligands

3.4.1 [(Me2P)2CH2]2Ni2(CO)4. Two low-energy [(Me2P)2CH2]2Ni2(CO)4 structures were found (Fig. 14). In the lowest-energy structure PMe-24-1S each (Me2P)2CH2 ligand remains intact and bridges the central Ni2 unit to form an eight-membered NiPCPNiPCP ring adopting the boat conformation leading to a structure exhibiting C2 symmetry. The long Ni⋯Ni distance in PMe-24-1S of 4.388 Å indicates the absence of a direct nickel–nickel bond thereby giving each nickel atom the favored 18-electron configuration. The experimental [(Me2P)2CH2]2Ni2(CO)4 structure12 is similar to PMe-24-1S with an essentially identical Ni⋯Ni non-bonding distance of 4.39 Å.
image file: c5ra22169b-f14.tif
Fig. 14 The two singlet [(Me2P)2CH2]2Ni2(CO)4 structures.

The second [(Me2P)2CH2]2Ni2(CO)4 structure PMe-24-2S, lying only 0.5 kcal mol−1 in energy above PMe-24-1S, is similar to PMe-24-1S except for a chair rather than a boat conformation of the eight-membered NiPCPNiPCP ring. This leads to a structure with C2h rather than C2 symmetry (Fig. 14). The nickel coordination geometry in PMe-24-2S is approximately tetrahedral with P–Ni–P angles of 108.1°. The Ni⋯Ni non-bonding distance of 4.654 Å in PMe-24-2S is a significantly longer than that in PMe-24-1S.

3.4.2 [(Me2P)2CH2]2Ni2(CO)3. Only one low-energy singlet structure was found for [(Me2P)2CH2]2Ni2(CO)3 (Fig. 15) at the M06-L/TZP level. This structure PMe-23-1S has a Ni–Ni distance of 2.621 Å, corresponding to a formal single bond. This Ni–Ni bond is bridged by a single CO group, predicted by the M06-L method to exhibit a ν(CO) frequency of 1877 cm−1. This bridging ν(CO) frequency is significantly lower than the terminal ν(CO) frequencies of 2027 and 2041 cm−1.
image file: c5ra22169b-f15.tif
Fig. 15 The single [(Me2P)2CH2]2Ni2(CO)3 structure.
3.4.3 [(Me2P)2CH2]2Ni2(CO)2. Three singlet [(Me2P)2CH2]2Ni2(CO)2 structures were found (Fig. 16). The lowest energy [(Me2P)2CH2]2Ni2(CO)2 structure PMe-22-1S may be derived from the [(Me2P)2CH2]2Ni2(CO)2 structure PMe-23-1S by removal of a terminal CO group. The Ni[double bond, length as m-dash]Ni distance of 2.490 Å in PMe-22-1S is ∼0.131 Å shorter than the formal Ni–Ni single bond in PMe-23-1S and thus can be interpreted as the formal double bond required to give each nickel atom the favored 18-electron configuration. The bridging CO group in PMe-22-1S exhibits the expected low ν(CO) frequency of 1862 cm−1 (M06-L/TZP) whereas the terminal CO group exhibits a significantly higher ν(CO) frequency of 2026 cm−1 in accord with expectation. The [(Me2P)2CH2]2Ni2(CO)2 structure PMe-22-3S, lying 8.6 kcal mol−1 in energy above PMe-22-1S is similar to PMe-22-1S except that the terminal CO group in PMe-22-1S becomes a semibridging CO group.
image file: c5ra22169b-f16.tif
Fig. 16 The three [(Me2P)2CH2]2Ni2(CO)2 structures.

A higher energy singlet [(Me2P)2CH2]2Ni2(CO)2 structure PMe-22-2S, lying 3.5 kcal mol−1 above PMe-22-1S, has exclusively terminal CO groups (Fig. 16). The Ni–Ni distance of 2.602 Å in PMe-22-2S suggests a formal single bond giving one nickel atom an 18-electron configuration and the other nickel atom a 16-electron configuration.

3.5 Thermochemistry

The reaction energies listed in Table 1 are based on the lowest energy singlet state structures. The dissociation energies for the loss of one CO group from the carbonyl richest complexes of the triads, namely PF-16-1S, PF-24-1S, PMe-16-1S, and PMe-24-1S, range from 13.5 kcal mol−1 to 25.5 kcal mol−1. This process of dissociation converts one of the remaining terminal CO groups to a bridging CO group and introduces a formal Ni–Ni single bond of length ∼2.7 Å. Thus the formation of the Ni–Ni single bond compensates for the two electrons lost upon removal of the CO group.
Table 1 Dissociation energies (in kcal mol−1) for the successive removal of carbonyl groups from binuclear nickel carbonyl derivatives [MeN(PF2)2]Ni2(CO)n (n = 6, 5), [MeN(PF2)2]2Ni2(CO)n (n = 4, 3), [(Me2P)2CH2]Ni2(CO)n (n = 6, 5) and [(Me2P)2CH2]2Ni2(CO)n (n = 4, 3) at the M06-L/TZP level
Process M06-L
MeN(PF2)2Ni2(CO)6 (PF-16-1S) → MeN(PF2)2Ni2(CO)5 (PF-15-1S) + CO 18.5
MeN(PF2)2Ni2(CO)5 (PF-15-1S) → MeN(PF2)2Ni2(CO)4 (PF-14-1S) + CO 28.5
[MeN(PF2)2]2Ni2(CO)4 (PF-24-1S) → [MeN(PF2)2]2Ni2(CO)3 (PF-23-1S) + CO 13.5
[MeN(PF2)2]2Ni2(CO)3 (PF-23-1S) → [MeN(PF2)2]2Ni2(CO)2 (PF-22-1S) + CO 35.4
[(Me2P)2CH2]Ni2(CO)6 (PMe-16-1S) → [(Me2P)2CH2]Ni2(CO)5 (PMe-15-1S) + CO 22.1
[(Me2P)2CH2]Ni2(CO)5 (PMe-15-1S) → [(Me2P)2CH2]Ni2(CO)4 (PMe-14-1S) + CO 25.4
[(Me2P)2CH2]2Ni2(CO)4 (PMe-24-1S) → [(Me2P)2CH2]2Ni2(CO)3 (PMe-23-1S) + CO 25.5
[(Me2P)2CH2]2Ni2(CO)3 (PMe-23-1S) → [(Me2P)2CH2]2Ni2(CO)2 (PMe-22-1S) + CO 38.9

The corresponding successive dissociation energies for the loss of one terminal CO group from the intermediate complexes of each triad, namely PF-15-1S, PF-23-1S, PMe-15-1S, and PMe-23-1S, range from 25.4 kcal mol−1 to 38.9 kcal mol−1. The CO dissociation energies of PF-23-1S and PMe-23-1S are very high (35.4 kcal mol−1 and 38.9 kcal mol−1). For comparison, the reported CO dissociation energies are 37 ± 2, 41 ± 2, and 25 ± 2 kcal mol−1 for Cr(CO)6, Fe(CO)5 and Ni(CO)4, respectively.42 The high dissociation energies reflect the strength of the metal–carbonyl bonds. During this process of CO dissociation a Ni[double bond, length as m-dash]Ni double bond is formed with intact ligands bridging the pair of nickel atoms. This is different from the iron carbonyl complexes [CH3N(PF2)2]Fe2(CO)n (n = 7, 6) [CH3N(PF2)2]2Fe2(CO)n (n = 5, 4) where P–N bond rupture of one of the bridging CH3N(PF2)2 ligands is energetically preferred over shortening of the iron–iron distances to give formal Fe[double bond, length as m-dash]Fe double bonds.8 The bridging CO groups become more asymmetrical in all the complexes except for PMe-14-1S.

4. Conclusion

The nickel carbonyl complexes of the strong back-bonding fluorophosphine ligand CH3N(PF2)2 can be grouped into the two triads [CH3N(PF2)2]Ni2(CO)n (n = 6, 5, 4) and [CH3N(PF2)2]2Ni2(CO)n (n = 4, 3, 2) depending on the CH3N(PF2)2/Ni ratio. In the lowest energy structures in both of these triads, all of the CH3N(PF2)2 ligands bridge the Ni2 unit rather than chelate to a single nickel atom to form a four-membered NiPNP chelate ring. The lowest energy carbonyl-richest structure of each triad has a long Ni⋯Ni distance indicating the lack of a direct nickel–nickel bond. The lowest energy structure of the intermediate member of each triad has a Ni–Ni distance of ∼2.7 Å suggestive of a formal single bond. The lowest energy structure of the carbonyl-poorest member of each triad has a shorter Ni[double bond, length as m-dash]Ni distance of ∼2.5 Å suggestive of a formal double bond. All of the lowest energy structures have intact CH3N(PF2)2 ligands and the favored 18-electron configuration for each nickel atom.

These two triads of CH3N(PF2)2 nickel carbonyl complexes can be compared with analogous triads of CH3N(PF2)2 iron carbonyl complexes having one more CO group per metal atom in order to retain the favored 18-electron configuration of the metal atoms. In this way the [CH3N(PF2)2]Ni2(CO)n (n = 6, 5, 4) and [CH3N(PF2)2]2Ni2(CO)n (n = 4, 3, 2) triads are the analogues of the [CH3N(PF2)2]Fe2(CO)n (n = 8, 7, 6) and [CH3N(PF2)2]2Fe2(CO)n (n = 6, 5, 4) triads. The lowest energy structures for the carbonyl-richest and intermediate members of the CH3N(PF2)2 iron carbonyl triads have the same patterns as those for the corresponding nickel carbonyl derivatives, i.e., non-bonding Fe⋯Fe distances for the carbonyl-richest system and formal Fe–Fe single bonds of length ∼2.7 Å for the intermediate members. However, in the lowest energy structures of the carbonyl-poorest members of the CH3N(PF2)2 iron carbonyl triads, phosphorus–nitrogen bond cleavage occurs to give separate CH3NPF2 and PF2 bridging groups which donate a total of six electrons to the central Fe2 unit. Such structures retain the ∼2.7 Å Fe–Fe single bond distances of the intermediate members of each CH3N(PF2)2 triad, which is sufficient to give each iron atom the favored 18-electron configuration.

The highly basic small bite bidentate ligand (Me2P)2CH2 was also included in this theoretical study of binuclear nickel carbonyl complexes for comparison with the strong back-bonding CH3N(PF2)2 ligand. The preferred types of (bid)Ni2(CO)n structures were found to be similar for both ligands despite the major differences in their basicity and back-bonding. Thus [(Me2P)2CH2]Ni2(CO)n and [(Me2P)2CH2]2Ni2(CO)n structures in which cleavage of a ligand P–C bond occurs to form separate bridging Me2PCH2 and Me2P are high-energy structures relative to isomeric structures with intact (Me2P)2CH2 ligands.

The single experimentally realized compound discussed in this paper is [(Me2P)2CH2]2Ni2(CO)4, first reported9 as the monomer “[(Me2P)2CH2]Ni(CO)2” but shown by Pörschke and co-workers12 by X-ray crystallography to be the dimer. The experimental structure for [(Me2P)2CH2]2Ni2(CO)4 is very similar to PMe-24-1S with a boat conformation for the NiPCPNiPCP eight-membered ring. This structure lies less than ∼0.6 kcal mol−1 in energy below the similar higher energy structure PMe-24-2S [(Me2P)2CH2]2Ni2(CO)4 with a chair conformation for the eight-membered ring.

Acknowledgements

The authors are indebted to the Chinese National Natural Science Foundation (20903010, 21373025), the Beijing Municipal Natural Science Foundation (2132035), Beijing Higher Education Young Elite Teacher Project (YETP1226), Opening Project of State Key Laboratory of Explosion Science of Technology (Beijing Institute of Technology) (ZDKT12-03), and the U.S. National Science Foundation (Grant CHE-1057466) for support of this research.

References

  1. R. B. King, Acc. Chem. Res., 1980, 13, 243–248 CrossRef CAS.
  2. J. F. Nixon, J. Chem. Soc. A, 1968, 2689–2692 RSC.
  3. R. B. King and J. Gimeno, J. Chem. Soc., Chem. Commun., 1977, 142–143 RSC.
  4. R. B. King and J. Gimeno, Inorg. Chem., 1978, 17(9), 2390–2400 CrossRef CAS.
  5. M. G. Newton, R. B. King, M. Chang and J. Gimeno, J. Am. Chem. Soc., 1977, 99(8), 2802–2803 CrossRef CAS.
  6. M. G. Newton, R. B. King, M. Chang and J. Gimeno, J. Am. Chem. Soc., 1978, 100(1), 326–327 CrossRef CAS.
  7. M. G. Newton, R. B. King, M. Chang and J. Gimeno, J. Am. Chem. Soc., 1978, 100(5), 1632–1633 CrossRef CAS.
  8. G. Li, L. Zhou, X. Zhai, Q.-S. Li, Y. Xie, R. B. King and H. F. Schaefer, New J. Chem., 2013, 37(10), 3294–3302 RSC.
  9. R. B. King and K. S. Raghuveer, Inorg. Chem., 1984, 23(16), 2482–2491 CrossRef CAS.
  10. S. I. Ignatyev, H. F. Schaefer and R. B. King, J. Am. Chem. Soc., 2000, 122(9), 1989–1994 CrossRef.
  11. R. J. Puddephatt, Chem. Soc. Rev., 1983, 12(2), 99–127 RSC.
  12. K. R. Pörschke, Y. H. Tsay and C. Krüger, Inorg. Chem., 1986, 25(12), 2097–2099 CrossRef.
  13. I. Langmuir, Science, 1921, 54(1386), 59–67 CAS.
  14. D. M. Bose, Z. Phys., 1926, 219 CrossRef.
  15. F. Reiff, Z. Anorg. Allg. Chem., 1931, 202, 375 CrossRef CAS.
  16. N. V. Sidgwick and R. W. Bailey, Proc. R. Soc. London, Ser. A, 1934, 144, 521 CrossRef CAS.
  17. P. Pyykkö, J. Organomet. Chem., 2006, 691, 4336–4340 CrossRef.
  18. M. Brynda, L. Gagliardi, P. O. Widmark, P. P. Power and B. O. Roos, Angew. Chem., Int. Ed., 2006, 45, 3804–3807 CrossRef CAS PubMed.
  19. N. Sieffert and M. Bühl, J. Am. Chem. Soc., 2010, 132, 8056–8070 CrossRef CAS PubMed.
  20. P. Schyman, W. Lai, H. Chen, Y. Wang and S. Shaik, J. Am. Chem. Soc., 2011, 133(20), 7977–7984 CrossRef CAS PubMed.
  21. R. D. Adams, W. C. Pearl, Y. O. Wong, Q. Zhang, M. B. Hall and J. R. Walensky, J. Am. Chem. Soc., 2011, 133(33), 12994–12997 CrossRef CAS PubMed.
  22. R. Lonsdale, J. Olah, A. J. Mulholland and J. N. Harvey, J. Am. Chem. Soc., 2011, 133(39), 15464–15474 CrossRef CAS PubMed.
  23. L. Crawford, D. J. Cole-Hamilton, E. Drent and M. Bühl, Chem.–Eur. J., 2014, 20(43), 13923–13926 CrossRef CAS PubMed.
  24. H. Zhekova, M. Krykunov, J. Autschbach and T. Ziegler, J. Chem. Theory Comput., 2014, 10(8), 3299–3307 CrossRef CAS PubMed.
  25. M. Bühl and H. Kabrede, J. Chem. Theory Comput., 2006, 2(5), 1282–1290 CrossRef PubMed.
  26. M. J. Frisch, et al., Gaussian 09, Revision D.01, Gaussian, Inc., Wallingford, CT, 2009 Search PubMed (see ESI† for details).
  27. A. D. Becke, J. Chem. Phys., 1993, 98(7), 5648–5652 CrossRef CAS.
  28. C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37(2), 785–789 CrossRef CAS.
  29. A. D. Becke, Phys. Rev. A: At., Mol., Opt. Phys., 1988, 38(6), 3098–3100 CrossRef CAS.
  30. J. P. Perdew, Phys. Rev. B: Condens. Matter Mater. Phys., 1986, 33(12), 8822–8824 CrossRef.
  31. D. G. Truhlar and Y. Zhao, Theor. Chem. Acc., 2008, 120(1–3), 215–241 Search PubMed.
  32. Y. Zhao and D. G. Truhlar, J. Chem. Phys., 2006, 125(19), 194101 CrossRef PubMed.
  33. T. H. Dunning, J. Chem. Phys., 1970, 53(7), 2823–2833 CrossRef CAS.
  34. S. Huzinaga, J. Chem. Phys., 1965, 42(4), 1293–1302 CrossRef.
  35. A. J. H. Wachters, J. Chem. Phys., 1970, 52(3), 1033–1036 CrossRef CAS.
  36. D. M. Hood, R. M. Pitzer and H. F. Schaefer, J. Chem. Phys., 1979, 71(2), 705–712 CrossRef CAS.
  37. B. N. Papas and H. F. Schaefer, J. Mol. Struct.: THEOCHEM, 2006, 768(1–3), 175–181 CrossRef CAS.
  38. P. L. Barbieri, P. A. Fantin and F. E. Jorge, Mol. Phys., 2006, 104(18), 2945–2954 CrossRef CAS.
  39. S. Grimme, J. Antony, S. Ehrlich and H. J. Krieg, J. Chem. Phys., 2010, 132, 154104 CrossRef PubMed.
  40. V. Jonas and W. Thiel, J. Phys. Chem., 1995, 102(21), 8474–8484 CrossRef CAS.
  41. I. Silaghi-Dumitrescu, T. E. Bitterwolf and R. B. King, J. Am. Chem. Soc., 2006, 128(16), 5342–5343 CrossRef CAS PubMed.
  42. A. W. Ehlers and G. Frenking, Organometallics, 1995, 14(1), 423–426 CrossRef CAS.

Footnote

† Electronic supplementary information (ESI) available: Tables S1 to S4: the ν(CO) vibrational frequencies (cm−1) for the [MeN(PF2)2]Ni2(CO)n (n = 6, 5, 4), [MeN(PF2)2]2Ni2(CO)n (n = 4, 3, 2), [(Me2P)2CH2]Ni2(CO)n (n = 6, 5, 4), and [(Me2P)2CH2]2Ni2(CO)n (n = 4, 3, 2) singlet structures at the M06-L/TZP level. Tables S5 to S8: infrared active ν(CO) vibrational frequencies (cm−1) predicted for the singlet and triplet sturctures of [MeN(PF2)2]Ni2(CO)n (n = 6, 5, 4), [MeN(PF2)2]2Ni2(CO)n (n = 4, 3, 2), [(Me2P)2CH2]Ni2(CO)n (n = 6, 5, 4), and [(Me2P)2CH2]2Ni2(CO)n (n = 4, 3, 2) singlet and triplet structures at BP86/DZP and B3LYP/DZP levels. Tables S9 to S20: total energies (E, in hartree), relative energies (ΔE, in kcal mol−1), numbers of imaginary vibrational frequencies (Nimg) at the M06-L/TZP level for the singlet structures of [MeN(PF2)2]Ni2(CO)n (n = 6, 5, 4), [MeN(PF2)2]2Ni2(CO)n (n = 4, 3, 2), [(Me2P)2CH2]Ni2(CO)n (n = 6, 5, 4), and [(Me2P)2CH2]2Ni2(CO)n (n = 4, 3, 2). Tables S21 to S33: total energies (E, in hartree), relative energies (ΔE, in kcal mol−1), spin square values 〈S2〉, number of imaginary vibrational frequencies (Nimg) at the B3LYP/DZP and BP86/DZP levels of the theory for the singlet and triplet structures of [MeN(PF2)2]Ni2(CO)n (n = 6, 5, 4), [MeN(PF2)2]2Ni2(CO)n (n = 4, 3, 2), [(Me2P)2CH2]Ni2(CO)n (n = 6, 5, 4), and [(Me2P)2CH2]2Ni2(CO)n (n = 4, 3, 2). Fig. S1: the one triplet structure found for [(Me2P)2CH2]Ni2(CO)4 at the M06-L/TZP level of theory. Fig. S2 to S13: optimized singlet and triplet structures for [MeN(PF2)2]Ni2(CO)n (n = 6, 5, 4), [MeN(PF2)2]2Ni2(CO)n (n = 4, 3, 2), [(Me2P)2CH2]Ni2(CO)n (n = 6, 5, 4), and [(Me2P)2CH2]2Ni2(CO)n (n = 4, 3, 2) by the B3LYP and BP86 methods. Table S34: theoretical Cartesian coordinates (in Å) for the singlet structures at the M06-L/TZP level. Table S35: theoretical Cartesian coordinates (in Å) for the singlet and triplet structures at BP86/DZP level; complete Gaussian 09 reference (ref. 26). See DOI: 10.1039/c5ra22169b
This journal is © The Royal Society of Chemistry 2016
Click here to see how this site uses Cookies. View our privacy policy here.