Amino-functionalised metal xanthates

Introduction

While the chemistry of metal xanthates has been the subject of study for many years, driven by applications in both analytical chemistry and metal extraction (Rao, 1971; Magee, 1973; Donaldson, 1976; Rao and Ramakrishna, 2003), recent years have seen a renaissance in the area resulting from the use of metal xanthates as single-source precursors (SSPs) to metal sulphide materials (Nair et al., 2002; Koh et al., 2003; Han et al., 2007; Clark et al., 2011; Liu et al., 2011; Rath et al., 2013; Kociok-Köhn et al., 2014). Remarkably, despite reports of ca. 320 structures, the overwhelming majority incorporate simple S₂COR groups, where R is a common hydrocarbon unit (Me, Et, etc.). The most complex ligand system described is arguably the bis-ligand S₂COCH₂(CH₂OCH₂)₃CH₂OCS₂ (as a dimeric nickel derivative) (Beer et al., 2003). Cases where R incorporates a donor functionality are extremely rare (R=CH₂CH₂OMe; Abrahams et al., 1988a,b; Edwards et al., 1990a,b; Chen et al., 2003a,b; Haiduc et al., 2004). CH₂(CH₂OCH₂)₃CH₂OH (Beer et al., 2003) and, to our knowledge, only one with a pendant N-donor (N-benzylpiperidyl) have so far been incorporated (Rajput et al., 2013). Such an approach would, however, open opportunities for the synthesis of heterobimetallic species, which could be exploited as SSPs to form mixed-metal materials. We now report some initial findings in this area of xanthate synthesis.

Results and discussion

Synthesis of 1 followed the generic procedure reported by Limura et al. (1972) for sodium xanthate formation involving deprotonation of the chelating alcohol 2-{[2-(dimethylamino)ethyl]-methylamino}ethanol (Hdmem) with NaH before addition of CS₂ [Eq. (1)].

1 is a pale yellow solid, formed in good yield (92%) and characterised most clearly by the presence of a characteristic ¹³C resonance of the significantly deshielded S₂CO at 232.8 ppm. The structure of 1 is shown in Figure 1 and is only the second structure of a sodium xanthate reported, the other being NaS₂COEt.H₂O (Mazzi et al., 1969); selected X-ray parameters and geometric data are given in Tables 1 and 2 respectively. Sodium is surrounded by a S₃N₂O coordination sphere, which is highly irregular in geometry, with S(2) approximately trans to N(2′), while S(1) is opposite S(1′) and O(1′). There is a large vacancy in the coordination sphere opposite N(1′) [next nearest contact to sodium is with C(8) at >3.5 Å]. The xanthate acts as a κ²-S,S chelate to one metal and a κ²-S,O chelate to a neighbour. Thus, S(1) and S(2) chelate the sodium in a broadly symmetrical manner [2.9897(8), 2.8012(8) Å, respectively], although the bond to S(1) is somewhat longer due to its μ₂-bridging interaction with a neighbouring metal [3.0886(8) Å]. The Na-O interaction is significant [2.3896(13) Å], while the remaining interactions are a symmetrical κ²-N,N chelate from the tail of the dmem functionality.

Figure 1:

The asymmetric unit of 1 showing the labelling scheme used; thermal ellipsoids are at the 50% probability level. Symmetry operations: ′ 3/2-x, 1/2+y, z; ″ 3/2-x, y-1/2, z; ′″ x, 1+y, z.

As far as we are aware, this is the first example of a metal xanthate with a pendant N-donor functionality. Indeed, metal xanthates with any pendant functional group are rare and seem to be limited to O-donors such as K[S₂COCH₂(CH₂OCH₂)₃CH₂OH] (Beer et al., 2003), Cd(S₂COCH₂CH₂OMe)₂.2.2′-bipy (Chen et al., 2003a), and [Et₄N][Cd(S₂COCH₂CH₂OMe)₃] (Abrahams et al., 1988b).

Attempts to use 1 to form d¹⁰ Group 12 salts were unsuccessful. In reactions with either ZnCl₂.6H₂O, Zn(NO₃)₂.6H₂O, or CdCl₂.2.5H₂O, the solution initially turned yellow, suggestive of metal xanthate formation, followed by precipitation of an insoluble powder that did not melt and is tentatively assumed to be the metal sulphide (ZnS, CdS). It would appear that the initially formed M[S₂COCH₂CH₂N(Me)CH₂CH₂NMe₂]₂ readily decomposes, and some indication as to how and why this might ensue comes from the isolation of small amounts of crystalline by-product from the soluble residue of the reaction involving hydrated CdCl₂ [Eq. (2)].

The nature of the by-product (2) was identified by X-ray crystallography (Figure 2) as the 1,1,4-trimethylpiperazinium salt of the dmem-xanthate anion present in 1. The formation of 2 plausibly arises from the mechanism shown in Scheme 1.

Figure 2:

The asymmetric unit of 2 showing the labelling scheme used; thermal ellipsoids are at the 50% probability level. Selected geometric data: S(1)-C(1) 1.6919(17), S(2)-C(1) 1.6768(16), O(1)-C(1) 1.367(2), O(1)-C(2) 1.442(2) Å; O(1)-C(1)-S(1) 113.10(11), O(1)-C(1)-S(2) 120.56(12), S(2)-C(1)-S(1) 126.34(10)°.

Scheme 1:

Chugaev elimination and formation of compound 2.

In effect, this is the known Chugaev decomposition of a xanthate (Tschugaeff, 1900) promoted by the intramolecular N:→C^δ+O interaction resulting from the pendant tertiary amine, which results in a facile deposition of MS from the reaction mixture. The origin of H⁺ in Scheme 1 is, at present, unclear.

The structure of 2 is unexceptional and comprises separated ions, the closest significant contact between which is S(1)…N(3) of 4.14 Å. Key geometric data (Figure 2, Table 2) are typical of other xanthate anions (Tiekink and Haiduc, 2005).

We have also attempted to form M[S₂C(dmem)]₂ by insertion of CS₂ into the polar M-O bond of the alkoxides RM(dmem) (Scheme 2), a strategy that has worked for other metals beyond those of group 1 (Hevia et al., 2002).

Scheme 2:

Synthetic routes to compounds 3 and 4 and attempted Zn and Cd xanthate synthesis.

Alkylmetal alkoxides were prepared by direct combination of R₂M and Hdmem using a protocol we have previously found successful (Johnson et al., 2008a,b; Hollingsworth et al., 2010). In addition to routine spectroscopic characterisation, the structures of both species have been determined (Figures 3 and 4) and add to the limited number of other reported M(dmem) compounds [Zn(dmem)₂ (Goel et al., 1990a), Cu(dmem)₂ (Goel et al., 1990b), Ce₂(OPrⁱ)₆(μ-dmem)₂ (Hubert-Pfalzgraf et al., 1992)], the iron clusters [Fe₇O₄(O₂CPh)₁₁(dmem)₂, Fe₇O₄ (O₂CMe)₁₁(dmem)₂, Fe₆O₂(OH)₄(O₂CBu^t)₈(dmem)₂, and Fe₃O(O₂CBu^t)₂(N₃)₃(dmem)₂] (Bagai et al., 2007), and the bi-metallic species [Fe₂GdO(O₂CBu^t)₂(dmem)₂(NO₃)₃ and Mn₂GdO(O₂CBu^t)₂(dmem)₂(NO₃)₃] (Mukherjee et al., 2010).

Figure 3:

The asymmetric unit of 3 showing the labelling scheme used; thermal ellipsoids are at the 50% probability level. Selected geometric data: Zn-O(1) 1.9916(9), Zn-O(1′) 2.1085(10), Zn-N(1) 2.4911(12), Zn-N(2) 2.2215(12), Zn-C(1) 2.0100(15) Å; O(1)-Zn-O(1′) 80.55(4), O(1)-Zn-N(1) 76.60(4), O(1)-Zn-N(2) 106.53(4), O(1)-Zn-C(1) 129.81(5), O(1′)-Zn-N(1) 146.92(4), O(1′)-Zn-N(2) 88.01(4), O(1′)-Zn-C(1) 110.98(5), N(1)-Zn-N(2) 76.11(4), N(1)-Zn-C(1) 102.02(5), N(2)-Zn-C(1) 122.05(5), Zn-O(1)-Zn′ 99.45(4)°. Symmetry operation: ′ 1-x,-y,1-z.

Figure 4:

The asymmetric unit of 4 showing the labelling scheme used; thermal ellipsoids are at the 50% probability level. Selected geometric data: Cd-O 2.203(4), Cd-O′ 2.306(4), Cd-N(1) 2.574(5), Cd-N(2) 2.445(5), Cd-C(1) 2.171(6) Å; O-Cd-O′ 79.17(15), O-Cd-N(1) 74.06(14), O-Cd-N(2) 103.63(15), O-Cd-C(1) 138.0(2), O′-Cd-N(1) 137.72(14), O′-Cd-N(2) 83.79(14), O′-Cd-C(1) 115.6(2), N(1)-Cd-N(2) 71.72(15), N(1)-Cd-C(1) 106.1(2), N(2)-Cd-C(1) 116.5(2), Cd-O-Cd′ 100.83(15)°. Symmetry operation: ′ 1-x, -y, -z.

Both compounds 3 and 4 adopt dimeric structures incorporating a μ₂-OR bridge and a central M₂O₂ ring. For both metals, tetrameric [RMOR′]₄ cubane structures (M=Zn, Cd) are most common, while dimeric analogues occur when the steric bulk at either zinc or oxygen increases (Johnson et al., 2008b, and references therein). The dimeric structure of 3 is in contrast to tetrameric MeM(OCH₂CH₂NMe₂), where the single N-donor simply serves to open the cubane to a puckered M₄O₄ ring in the case of zinc (Johnson et al., 2008b) or retains the cubane but increases the coordination number at the metal in the case of cadmium (Johnson et al., 2008a). Thus, it is the presence of the two N, N donors that imparts a dimeric structure to 3 and 4 and is a rare example of dimeric [RZnOR′]₂ that embodies five-coordinate metal centres (Groysman et al., 2006; Hung et al., 2008; Lichtenberg et al., 2012) and the first example for cadmium.

In both cases, the metals are five-coordinated with an O₂N₂C coordination sphere. The geometry about the metals is distorted, but a τ analysis (3: 0.29; 4: 0.01) (Addison et al., 1984) suggests a square pyramidal description with N(2) at the apex, although, particularly in the case of 4, this belies the asymmetry in the coordination imposed by the incorporation of two five-membered and one four-membered chelate rings about each cadmium centre. In comparison with dimeric [EtZnOC(CH₂NMe₂)₃]₂ (Johnson et al., 2008b), the Zn₂O₂ ring is less regular [Zn-O: 1.9916(9), 2.1085(10) vs. 2.0041(12), 2.0451(12) Å] and the Zn-N bond is weaker [2.4911(12), 2.2215(12) vs. 2.1847(16) Å], although the latter is expected as the zinc coordination number in 3 is higher. Similarly, in comparison with dimeric MeCd[OC(CH₂NMe₂)₃ (Johnson et al., 2008a), the Cd-O bonds are again markedly more asymmetric [2.203(4), 2.306(4) vs. 2.2249(9), 2.2321(9) Å] and the Cd-N interactions longer [2.574(5), 2.445(5) vs. 2.4178(11) Å].

Unfortunately, when CS₂ was added to a hexane solution of either 3 or 4, the outcome was qualitatively the same as noted previously [Eq. (2)], i.e. initial formation of a yellow solution followed by rapid deposition of an insoluble, high-melting solid and soluble material whose NMR spectra again revealed the presence of the 1,1,4-trimethylpiperazinium cation.

While the target M[S₂C(dmem)]₂ (M=Zn, Cd) proved too unstable to isolate, 1 has successfully been used to prepare one other functionalised d¹⁰ metal xanthate, (Ph₃P)₂CuS₂C(dmem) (5):

The structure of 5, shown in Figure 5, closely resembles the many other known (R₃P)₂CuS₂COR′ structures reported (Kociok-Köhn et al., 2014). The structure shows the S₂CO(dmem) ligand bound in a bidentate fashion to a central four-coordinate copper atom. Copper adopts a slightly distorted tetrahedral geometry, with a xanthate bite angle of 121.35(16)°. The Cu-S bond lengths are nearly identical at 2.4110(7) and 2.4247(8) Å, and the C-S bond lengths are identical at 1.692(3) Å. The S-C and C-O bonds [1.692(3) and 1.340(3) Å, respectively] are in good agreement with similar (PPh₃)₂CuS₂COR complexes (Kociok-Köhn et al., 2014). The most significant feature of this structure is the uncoordinated amino-functionalised tail of the dmem moiety which is prevented from coordination to copper by the bulky triphenylphosphine ligands and which offers further opportunity to create bimetallic species.

Figure 5:

The asymmetric unit of 5 showing the labelling scheme used; thermal ellipsoids are at the 50% probability level. Only the major component of the disordered NCH₂CH₂NMe₂ unit is shown. Two molecules of CH₂Cl₂ also present in the lattice have been omitted for clarity. Only the α-carbons attached to phosphorus have been labelled for clarity, the remaining carbon atoms being numbered sequentially around each ring. Selected geometric data: Cu-S(1) 2.4110(7), Cu-S(2) 2.4247(8), Cu-P(1) 2.2573(7), Cu-P(2) 2.2461(7), S(1)-C(1) 1.692(3), S(2)-C(1) 1.692(3), O-C(1) 1.340(3) Å; S(1)-Cu-S(2) 75.20(2), S(1)-Cu-P(1) 103.11(3), S(1)-Cu-P(2) 122.56(3), S(2)-Cu-P(1) 111.40(3), P(2)-Cu-S(2) 107.05(3), P(1)-Cu-P(2) 126.05(3), S(1)-C(1)-S(2) 121.35(16)°.

The difference in stability between 5 and the similarly d¹⁰ zinc and cadmium species M[S₂C(dmem)]₂ is currently unclear. While the bulky Ph₃P groups inhibit N:→Cu coordination, they should not preclude Me₂N:→C₁ [(Eq. (2)] attack. Moreover, while copper(I) xanthates generally have a higher decomposition temperature than other metal xanthates do, this is due to the temperature required to displace the coordinated phosphine groups and not the Chugaev decomposition of the xanthate (Kociok-Köhn et al., 2014). While it might appear that two xanthate ligands are required to generate 2 (not present on 5 but present in M[S₂C(dmem)]₂ and [RMS₂(dmem)]₂), our preliminary experiments of the 1:1 reaction between R₃SnCl (R=Me, Bu) and 1 also lead to decomposition products. One possible factor may be the greater polarising nature of M²⁺ over M⁺, which would encourage nucleophilic attack at C₁. Further experimental work is thus required to delineate the extent of the stability of N-functionalised metal xanthates.

Experimental