Crystal structure of poly[(μ3-thiocyanato-κ3 N:S:S)(trimethylphosphine sulfide-κS)copper(I)]

The thiocyanate ions bind the CuII atoms covalently, forming infinite –Cu—SCN—Cu– chains parallel to the a axis. Two crystallographically independent chains propagate in opposite directions, and are held together in a ribbon arrangement by long bonds between CuII atoms in the first chain and thiocyanate S atoms in the second.


Chemical context
The synthesis and metal coordination reactions of phosphine sulfides is of continuing interest (Sues et al., 2014;Tiedemann et al., 2014). The title compound was synthesized by Tiethof et al. (1974) as part of an early series of studies on the coordination chemistry of copper(I) with these sulfur ligands, which established the importance of trigonal-planar coordination for copper(I), then still rare. Indeed, the structure of the cation in [Cu(Me 3 PS) 3 ]ClO 4 (Eller & Corfield, 1971) was the first example of trigonal-planar coordination in a monomeric copper complex. ISSN 1600-5368 Use of the pseudohalide thiocyanate in the synthesis of coordination compounds is a well-used route in the design of polymeric structures. Early papers on copper thiocyanate polymers involving amine adducts include Raston et al. (1979) and Healy et al. (1984). More recent studies include papers on the optical properties of self-assembled amine copper(I) thiocyanate complexes by Niu et al. (2008) and Miller et al. (2011), as well as studies on magnetic properties of a number of similar copper(II) complexes by Machura et al. (2013).

Structural commentary
The previously determined structure of [Cu(Me 3 PS)Cl] 3 (Tiethof et al., 1973) consists of a six-membered ring of alternating Cu and S atoms, with trigonal-planar coordination for the Cu I atoms completed by bonds to a Cl atom. It was noteworthy that the Me 3 PS phosphine sulfide ligands bridged the Cu I atoms to form the ring, and not the chlorine atoms as might have been expected. The structure of [Cu(Me 3 PS)SCN] was undertaken to determine whether this trimeric structure persisted in the presence of the thiocyanate ligand.
The present work determined that trimethylphosphinecopper(I) thiocyanate crystallizes as a one-dimensional polymer, rather than as the discrete trimers found for the chloride analog. Thiocyanate ions bind to two separate copper(I) atoms through Cu-N and Cu-S bonds. In the crystal, the two Cu I atoms are related by translation, which leads to the formation of infinite -Cu-SCN-Cu-chains parallel to the a axis. Although the Cu-N-C angles are approximately linear, the Cu-S-C angles are bent considerably, as expected (see Table 1). Each Cu I atom is also coordinated to a terminal Me 3 PS group via a Cu-S bond.

Figure 1
The ribbon structure of the title polymer, with displacement ellipsoids drawn at the 50% level, showing Cu2 in position A. Hydrogen atoms are omitted.

Figure 2
The alternate ribbon structure of the title polymer, showing the environment of Cu2 in position B, with ellipsoids at the 50% level. Hydrogen atoms omitted.

Figure 3
The Cu2 ellipsoids after and before the disordered model was introduced. Displacement ellipsoids are drawn at the 50% level.
Two crystallographically independent chains propagate in opposite directions, and are held together in a ribbon arrangement by long Cu-S bonds between the chains. While no disorder was seen in the first chain, each Cu I atom in the second chain appears to be disordered over two positions, Cu2A and Cu2B, 0.524 (4) Å apart, with occupancy factors of 64.7 (6)% and 35.3 (6)%, and slightly different coordination spheres (Fig. 3).
Cu I atoms in the first chain bind to thiocyanate sulfur atoms in the second, with Cu1-S4 = 2.621 (1) Å . Also, one of the disordered Cu I atoms in the second chain forms a long bond to the phosphine sulfide of the first chain, with Cu2B-S1 = 2.702 (5) Å , forming another link between the chains (Figs. 1 and 2), and a ladder arrangement that is seen also in one of the structures in Healy et al. (1984) and in Niu et al. (2008). The Cu-Cu distances across the chain are 3.656 (3) Å for Cu1-Cu2A and 3.351 (5) Å for Cu2B.
In the current structure, the two independent Me 3 PS groups are non-equivalent: the group in the second chain, C21-C23, P2 and S2, is terminal, while that in the first chain, C11-C13, P1 and S1, forms an asymmetric bridge between Cu1 and the minor component atom Cu2B. This may explain the observation of two different P S stretching bands in the infra-red spectrum, see below. The two thiocyanate groups are also nonequivalent, with both S3 and S4 bonded to Cu and N atoms, but S4 forming an additional long bond to Cu1. The nonequivalent groups do not show significant differences in geometry, however (Table 1).
The geometry around Cu1 atoms, in the first chain, is distorted tetrahedral, with angles involving the long Cu1-S4 bond much less than ideal, and the S1-Cu1-N3 angle between the phosphine sulfide and the thiocyanate N atom increased to 133.19 (9) . The geometry around the disordered Cu I atom in the major site, Cu2A, is in a distorted trigonalplanar configuration, with the S2-Cu2A-N4 angle between the phosphine S and the thiocyanate N atoms again opened out, to 137.01 (15) . Atom Cu2B has an irregular tetrahedral configuration. The geometry at the three-coordinated sulfur atoms S1 and S4 is trigonal-pyramidal rather than trigonalplanar, with the sum of the angles at S1 = 303.2 , while at S4 the sum is 294.8 for angles involving Cu2A and 281.0 for angles with Cu2B.

Supramolecular features
A packing diagram viewed down the a * axis is shown in Fig. 4. There are no strong interactions between the chains, and all intermolecular contacts appear normal. The shortest intermolecular contacts are H13AÁ Á ÁH23B(1 + x, y, z), at 2.53 Å , and H12AÁ Á ÁH12A(1 À x, 1 À y, 1 À z) at 2.57 Å . All other HÁ Á ÁH contacts are greater than 2.7 Å .

Database survey
Entry CMPSCU in the Cambridge Structure Database (CSD) is taken from the abstract of our presentation at the 1973 Winter Meeting of the American Crystallographic Association. No coordinates were given.
A search of the database with the fragment Cu-S-C N-Cu fingered 100 analyzable structures with 164 thiocyanate groups. The average thiocyanate geometries were: C N = 1.152 (17), S-C = 1.65 (2)Å ; S-C N = 178.2 (14) . Corresponding parameters in the present structure are indistinguishable from these average values. A much greater spread is seen in average parameters involving Cu, reflecting the diversity of chemical interactions in these structures. For example, average values for Cu-S distances are 2.5 (2) Å , with a range from 2.20 to 3.12 Å .

Synthesis and crystallization
Details of the synthesis and characterization of the title compound are given in Tiethof et al. (1974), which describes the preparation and characterization of a series of copper(I) complexes with tertiary phosphine sulfide, phosphine selenide, and arsine sulfide ligands. Solid LiSCN (0.59 mmol) was stirred with 7 mL of a solution of 0.63 mmol of [Cu(Me 3 PS) 3 ]BF 4 in acetonitrile for 30 min. The resultant solid was collected, washed with ether, dried in vacuo, and characterized by C, H, and N elemental analysis. The infra-red spectrum of a solid sample in a Nujol mull gave bands attributed to P S stretching at 543 and 546 cm À1 . These frequencies are similar to the frequency of 540 cm À1 observed for [Cu(Me 3 PS) 3 ]BF 4 , where the phosphine ligands are terminally bonded to copper as in the present structure, and significantly different from the P S frequency of 564 cm À1 observed for the free ligand, Me 3 PS. Packing of the title complex, viewed along the a * axis, with ellipsoid outlines at 30% probability.

Refinement details
Initial refinements with anisotropic displacement parameters for all non-hydrogen atoms and constrained hydrogen atom parameters converged smoothly to R = 0.0315 for F 2 >2, but a difference Fourier synthesis at this stage showed unacceptable features, with a hole of À1.0 e/A 3 and two peaks of 0.7 e/A 3 near Cu2, while there were no significant peaks or holes near Cu1. In addition, the temperature factors for Cu2 indicated an ellipsoid much elongated compared to that for Cu1 (Fig. 3). In case these features were related to systematic anisotropies that might have existed in the data collection, a trial was made to apply a smoothly varying scale factor by a 12 parameter model with XABS2 (Parkin et al., 1995). This had no significant effect on either the difference Fourier map or the R values, and the trial was abandoned. Instead, a model with Cu2 disordered equally between two positions was refined, which converged at R = 0.0307 for F 2 >2, and showed maximum and minimum residual electron densities at 0.71 and À0.81 e/A 3 near Cu2B and Cu2A, respectively, indicating that the sites were not equally occupied. Allowing the occupancy factors to vary led to the final model, with R = 0.0265 for F 2 >2, and residual electron density maxima of 0.29 and À0.31 e/A 3 near S and P atoms. The disordered Cu II atoms sites are 0.524 (4) Å apart, with occupancy factors of 64.7 (6)% and 35.3 (6)%. To facilitate convergence, the U ij for the disordered Cu atoms were constrained to be identical. It is likely that S2 could also be disordered, reflecting bonding to the two different Cu2 sites. We have not pursued attempts to model this.
The two partial copper positions might have represented alternating sites in a larger unit cell with the short a axis doubled. This would have made the disorder an artifact due to the data collection in that only reflections with h = 2n would have been collected. However, inspection of precession photographs of the h0l, h1l and h2l layers did not reveal any indication of doubling of the a axis. Furthermore, if that had been the case, the occupancies of the disorder components would have refined to approximately 0.5 rather than 0.647 (6) and 0.353 (6).
H atoms were constrained to idealized positions with C-H distances of 0.96 Å . The orientations of the methyl groups were determined by calculation of electron density in the toroid that should contain the H atoms of the idealized methyl groups. The U eq values for the H atoms were fixed at 1.2 times the U iso of their bonded C atoms.
Values for the Goodness of Fit (GOOF) near the end of the refinements were rather low, at 0.66, implying that at least some of the estimated values for the data were too high. The factor p in the data processing (Corfield et al., 1973) had originally been set at 0.06, a value that now seemed too large for such a highly refined structure. The values were adjusted to correspond to p = 0.05 with the equation: [(new)/F 2 ] 2 = [(old)/F 2 ] 2 À(0.06 2 À0.05 2 ). In addition, values for 182 very weak reflections, which had been grossly overestimated previously, were set equal to the average value found for the 145 reflections observed with I<0. (These reflections were set to F 2 = 0.) Final refinements with these adjustments to the values raised the value of the GOOF to 0.79 with no significant changes to any parameters.

Poly[(µ 3 -thiocyanato-κ 3 N:S:S)(trimethylphosphine sulfide-κS)copper(I)]
Crystal data Primary atom site location: structure-invariant direct methods Secondary atom site location: difference Fourier map Hydrogen site location: inferred from neighbouring sites H-atom parameters constrained w = 1/[σ 2 (F o 2 )] (Δ/σ) max = 0.001 Δρ max = 0.29 e Å −3 Δρ min = −0.31 e Å −3 Extinction correction: SHELXL97 (Sheldrick, 2008), Fc * =kFc[1+0.001xFc 2 λ 3 /sin(2θ)] -1/4 Extinction coefficient: 0.00128 (9) Special details Geometry. All e.s.d.'s (except the e.s.d. in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell e.s.d.'s are taken into account individually in the estimation of e.s.d.'s in distances, angles and torsion angles; correlations between e.s.d.'s in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell e.s.d.'s is used for estimating e.s.d.'s involving l.s. planes. Refinement. Refinement of F 2 against ALL reflections. The weighted R-factor wR and goodness of fit S are based on F 2 , conventional R-factors R are based on F, with F set to zero for negative F 2 . The threshold expression of F 2 > σ(F 2 ) is used only for calculating R-factors(gt) etc. and is not relevant to the choice of reflections for refinement. R-factors based on F 2 are statistically about twice as large as those based on F, and R-factors based on ALL data will be even larger.