Molecular structure of novel heterobimetallic thiolate [Cu(PPh₃)]₄(ZnCl)₂(SEt)₆.6THF

In the search for efficient absorber materials for photovoltaic cells derived from earth-abundant elements, Cu₂ZnSnS₄ (CZTS) has attracted significant attention (Ito and Nakazawa, 1988; Ramasamy et al., 2012); cells with efficiencies, which have increased from ca. 7% to in excess of 11%, have now been reported (Guo et al., 2010; Todorov et al., 2010, 2013; Barkhouse et al., 2012; Repins et al., 2012; Shin et al., 2013).

While several methods for forming CZTS are known (Ramasamy et al., 2012; Kociok-Köhn et al., 2014a and references therein), deposition in thin film form has been problematic. We (Kociok-Köhn et al., 2014a) and others (Ramasamy et al., 2011) have noted the specificity of conditions required to bring at least three precursors (one for each metal, which may, or may not, incorporate the chalcogen) together to generate CZTS in the correct stoichiometry. One way to simplify this problem is the use of precursors that embody more than one metal, which thereby reduces the overall number of precursors involved in the process. To this end, we have previously reported the synthesis of several M-Cl-M′ species (M=Cu, Zn, Sn inter alia), which could plausibly act as starting points for the synthesis of M-S-M′-containing precursors (Kociok-Köhn et al., 2014b). Unfortunately, to date, our attempts to produce such species by salt metathesis of the M-Cl-M′ reagents with thiolate anions have been unsuccessful. However, we now wish to report an initial success from the combination of (Ph₃P)₂CuCl and [(EtS)₃Zn]Na and the structural characterisation of the first Cu^I-S(R)-Zn^II linkage.

[Cu(PPh₃)]₄(ZnCl)₂(SEt)₆ (1) was synthesised from the reaction of Na[Zn(SEt)₃], prepared by partially following the methodology used to prepare [(EtS)₆Zn₂]Na₂, (Watson et al., 1985) and Cu(PPh₃)₂Cl:

4(Ph3P)2CuCl+2[(EtS)3Zn]Na→-4Ph3P, -2NaCl(i)   toluene, (ii)   THF[(Ph3P)Cu]4(ZnCl)2(SEt)6(1)

Although the reaction was carried out at 1:1 reagent stoichiometry, this is not translated to the ratio of metals in the final product. Moreover, the reaction involves (perhaps unsurprisingly) loss of Ph₃P and also, but less anticipated, migration of a halogen from the softer copper to the harder zinc.

Crystals of 1.6THF were isolated from the solution, but microanalysis implies that the molecules of solvation are easily lost from the solid when the initially prepared compound was pumped dry. Spectroscopic data are unexceptional, and despite NMR integrals revealing the ratio of EtS:Ph₃P ligands, the structure is revealed only by crystallography (Figure 1; Table 1). 1 consists of two Cu₂ZnS₃ rings, each in a distorted boat conformation, joined co-facially via two Cu-S [Cu(2)-S(1) 2.3650(11) Å] and two Zn-S [Zn(1)-S(1) 2.3836(11) Å] bonds. This results in a cage embodying both six-membered Cu₂ZnS₃ and four-membered CuZnS₂ heterocycles. In addition to bonds to sulphur, each zinc is bonded to chlorine [Zn(1)-Cl(1) 2.2505(11) Å], while each copper has a Ph₃P group external to the cage [Cu(1)-P(1) 2.2087(10) Å; Cu(2)-P(2) 2.2214(10) Å]. Cu(1) is three-coordinate in a trigonal planar environment, while both Cu(2) and Zn(1) are four-coordinate and tetrahedral, although the formation of fused rings requires significant bond angle distortion [angles about Zn(1) 98.45(4)–118.24(4)°; Cu(1) 108.76(4)–124.73(4)°; Cu(2) 97.75(4)–114.35(4)°]; the shorter bond lengths to Cu(1) in comparison to Cu(2) reflect this coordination number difference. S(1) and S(2) act in a μ₃-bridging mode, while S(3) is only μ₂ and has the shorter bond lengths to its neighbouring metals. Despite the two distinct copper (and hence phosphorus and thiolate) environments, separate signals are not seen in either the ³¹P or ¹³C NMR spectra, suggesting some structural fluxionality in solution.

Figure 1:

The asymmetric unit of 1 showing the labelling scheme used; thermal ellipsoids are at the 30% probability level.

P(1) and C(7) are hidden behind C(40). Symmetry operation: 1-x, -y, 2-z.

1 incorporates the first examples of structurally characterised Cu^I-S(R)-Zn^II linkages. Indeed, only one other structure in the Cambridge Crystallographic Database establishes the Cu^I-S-Zn^II array, namely, [(Ph₃P)Cu(μ-SCOR)₃Zn(PPh₃)] (R = Ph, thiophene), which incorporates μ₂-S thiocarboxylate ligands (Singh et al., 2013).

An analogous approach in search of a similar bimetallic Cu-Sn species only generated a known compound. Thus, [(PhS)₃Sn]Na, prepared by adding three equivalents of NaSPh to SnCl₂ in MeOH following published methodology (Dean et al., 1985), was added to Cu(PPh₃)₂Cl in CH₂Cl₂ in a 1:1 ratio. The resulting white precipitate was recrystallised from toluene, generating, over time, red crystals of Sn[Cu(PPh₃)]₂(SPh)₆.2CH₃C₆H₅, in which the tin had oxidised; the product was identified crystallographically. Sn[Cu(PPh₃)Cu]₂(SPh)₆.3H₂O has been reported by others (Wang et al., 2008) from the reaction of CuCN, PPh₃ and Sn(SPh)₄ (1:1:1).