Concomitant Carboxylate and Oxalate Formation From the Activation of CO₂ by a Thorium(III) Complex

Abstract

Improving our comprehension of diverse CO₂ activation pathways is of vital importance for the widespread future utilization of this abundant greenhouse gas. CO₂ activation by uranium(III) complexes is now relatively well understood, with oxo/carbonate formation predominating as CO₂ is readily reduced to CO, but isolated thorium(III) CO₂ activation is unprecedented. We show that the thorium(III) complex, [Th(Cp′′)₃] (1, Cp′′={C₅H₃(SiMe₃)₂‐1,3}), reacts with CO₂ to give the mixed oxalate‐carboxylate thorium(IV) complex [{Th(Cp′′)₂[κ²‐O₂C{C₅H₃‐3,3′‐(SiMe₃)₂}]}₂(μ‐κ²:κ²‐C₂O₄)] (3). The concomitant formation of oxalate and carboxylate is unique for CO₂ activation, as in previous examples either reduction or insertion is favored to yield a single product. Therefore, thorium(III) CO₂ activation can differ from better understood uranium(III) chemistry.

There has been an international drive to reduce emissions of CO₂ through cleaner energy generation since its identification as a key contributor to global warming.1 In tandem the employment of CO₂ as a C₁ feedstock for fine chemical (by direct insertion into organic molecules)2 and liquid fuel (via reduction to CO for Fischer–Tropsch processes)3 synthesis have rapidly expanded to complement the optimized photosynthetic pathways employed by nature.4 Early d‐transition metal complexes have received most attention for CO₂ activation as their inherent oxophilicity is advantageous in overcoming the considerable thermodynamic and kinetic barriers in this process.5 Similarly, actinides are highly oxophilic, so CO₂ activation by U^III complexes is also developing rapidly6 and proof of concept catalytic processes have been disclosed.7 The mapping of U^III‐mediated CO₂ activation by DFT calculations has provided key insights into possible mechanistic pathways.8 In contrast, Cloke reported the only example of CO₂ activation by a putative Th^III intermediate9 as Th^III small molecule activation is in its infancy.10, 11 Herein we report the first reaction of an isolated Th^III complex with CO₂, and CS₂ for comparative studies.

[Th(Cp′′)₃] (1, Cp′′={C₅H₃(SiMe₃)₂‐1,3}) reacts with 0.5 to 10 equivalents of CS₂ to give [{Th(Cp′′₃)}₂(μ‐κ¹:κ²‐CS₂)] (2) as the only isolable product in 45 % yield (Scheme 1; see the Supporting Information for full details). This reaction is consistent with U^III chemistry as the double reduction of CS₂ by [U(Cp′)₃] (Cp′=C₅H₄SiMe₃) yields [{U(Cp′₃)}₂(μ‐κ¹:κ²‐CS₂)].12 However, 1 reacts with excess CO₂ to give [{Th(Cp′′)₂[η²‐O₂C{C₅H₃‐3,3′‐(SiMe₃)₂}]}₂(μ‐κ²:κ²‐C₂O₄)] (3) in 65 % yield (Scheme 1), in contrast to the U^III reduction of CO₂ by [U(Cp′)₃] to afford [{U(Cp′)₃}₂(μ‐O)] and CO.13 The FTIR spectrum of 3 has absorptions at 1653 cm⁻¹ and 1560 cm⁻¹ that can be attributed to asymmetric C−O stretches of the oxalate and carboxylate groups respectively.14 The reaction of 2 with CO₂ gave a mixture of products including carboxylate (see the Supporting Information).

Scheme 1.

Synthesis of 2 and 3 from 1.

The solid state structures of 2 and 3⋅2C₇H₈ were determined by single crystal XRD (Figure 1. A polymorph 2 b⋅2C₆H₁₄ was also obtained; see the Supporting Information). The (μ‐CS₂)²⁻ unit in 2 was disordered over two positions so only the major component is discussed here. This fragment binds in an asymmetrical μ‐κ¹:κ²‐fashion [S−C: 1.644(11) and 1.717(10) Å; S‐C‐S: 124.4(7)°], in common with the motif seen for [{U(Cp′₃)}₂(μ‐κ¹:κ²‐CS₂)] [S−C: 1.464(19) and 1.831(19) Å; S‐C‐S: 131.7(13)°]12 and similar to that seen for [{U[OSi(OtBu)₃]₃}₂(μ‐κ²:κ²‐CS₂)] [S−C: 1.594(12) and 1.748(11) Å; S‐C‐S: 131.6(8)°].15 The oxalate of 3 has similar metrical parameters to those seen in [{Th(COT^TIPS2)(Cp*)}₂(μ‐κ²:κ²‐C₂O₄)].9 The carboxylate ligand exhibits both C−C and C=C lengths in the C₅ ring and a geminal 3,3′‐disilane. The C−O_carboxylate [1.284(5) and 1.263(5) Å] lengths evidence delocalization about the carboxylate framework, although the binding is asymmetric due to sterics [Th−O_carboxylate 2.400(3) and 2.484(2) Å]. The electronic structures of 2 and 3 were characterized at the DFT level, employing the B3LYP exchange‐correlation functional and a polarized split‐valence basis set for structural optimizations. Structural parameters of 3 were in good agreement with experiment (see the Supporting Information for full details).

Figure 1.

Molecular structure of a) 2 and b) 3.2C₇H₈ with selected atom labelling and displacement ellipsoids set to 30 % probability level. Hydrogen atoms, minor disorder components and lattice solvent omitted for clarity.

We postulated that 3 forms via a [{Th(Cp′′)₃}₂(μ‐CO₂)] intermediate that is analogous to 2. The bulky Cp′’ ligands hinder the elimination of CO and the formation of [{Th(Cp′′)₃}₂(μ‐O)],16 so a second molecule of CO₂ reacts with the (μ‐CO₂)²⁻ fragment to give an oxalate. There are many examples of sterically demanding ligands promoting oxalate formation over a μ‐oxo or carbonate in f‐block CO₂ activation.8, 17 Subsequent insertion of CO₂ into a Th–Cp′’ moiety and silyl/proton migration yields 3. The insertion of CO₂ into lanthanide‐Cp^R bonds to form carboxylates has been postulated not to require steric strain to proceed.18 Additional experiments were performed to probe the mechanism of formation of 3 (see the Supporting Information for full details). A Toepler pump was used to react 1 with 1 or 2 equivalents of CO₂ or ¹³CO₂ at −78 °C, and 3/3‐¹³C was the only identifiable product by ¹H and ¹³C{¹H} NMR spectroscopy in all cases. The reaction of 1 with supercritical CO₂ was monitored by ¹H NMR spectroscopy, and comparison with an authentic sample showed the formation of 3. Minor products in all reaction mixtures could not be identified. In situ FTIR spectroscopy was used to monitor the conversion of 1 to 3 at −78 °C in methylcyclohexane with stoichiometric CO₂. No intermediates could be detected but the CO_2(g) absorption at 2338 cm⁻¹ diminished on slow warming to room temperature, coincident with the ingress of an oxalate absorption at 1653 cm⁻¹ that is seen in the FTIR spectrum of crystalline 3. The experiment was repeated with ¹³CO₂ and the oxalate absorbance of 3‐¹³C was observed at 1609 cm⁻¹, consistent with reduced mass considerations.

Given that no intermediates could be detected experimentally, we performed DFT studies to rationalize this unusual mechanism. Figure 2 shows the calculated enthalpy reaction profile for the formation of 3, with the double reduction of CO₂ to give a μ‐κ¹:κ²‐CO₂ dinuclear Th^IV complex the proposed first step based on the analogous CS₂ reaction as well as CO₂ reactivity reported with other actinide complexes.8, 12, 15, 17 The oxalate formation invokes nucleophilic attack of a CO₂ molecule by a dimetalloxycarbene intermediate [{Th(Cp′′₃)}₂(μ‐κ¹:κ¹‐CO₂)] (C2) in a carbenic fashion, which has previously only been seen in d‐block CO₂ activation for Ti^IV.19 No pre‐interaction is required between the Th^IV centers and the second CO₂ molecule, which is in contrast with all previous examples of Sm^II[17a–c] and U^III[17d] oxalate formation. The resultant cis‐μ‐κ¹:κ¹‐C₂O₄ transition state (TS1) is one of several possible conformers that have Δ_r H° values within the estimated error of the calculation (ca. 3 kcal mol⁻¹) of each other, thus we do not comment on this further. Rearrangement of the oxalate to a trans‐μ‐κ¹:κ¹‐binding mode increases the steric demands about the Th^IV centers (C3). The potential energy surface for these rearrangements is very flat and despite our efforts it was not possible to locate a transition state. This leads to insertion of CO₂ at a single position of a Th–Cp′’ moiety at each Th^IV center (TS2) as the silicon centers stabilize negative charge at the beta position, allowing the best overlap with the empty orbital of CO₂. These insertions are accompanied by the rearrangement of the oxalate to a μ‐κ²:κ²‐binding mode. Subsequent proton and silyl group migrations in the dearomatized Cp′′ rings give the observed product 3 at an energetic minimum.

Figure 2.

Computed enthalpy reaction profile for the formation of 3.

To conclude we have shown that although CS₂ activation by [Th(Cp′′)₃] is analogous to that seen for a similar U^III system, the mechanism by which it reacts with CO₂ to form a mixed oxalate/carboxylate product has no precedent in U^III chemistry, in which CO₂ reduction (and subsequent carbonate formation depending on the supporting ligands) predominates. We probed this reaction to show that the oxalate is generated by a mechanism only seen previously in d‐block chemistry, whereas the carboxylate forms via a route only seen before in f‐element chemistry for lanthanide complexes. This shows that Th^III small molecule activation can furnish results that complement and contrast with uranium, lanthanide and d‐transition metals. Future studies will target heteroallene activation by Th^III complexes supported by different ligand systems to test the generality or divergence of these processes.

Experimental Section

Full synthetic details, characterization data and computational data for 2 and 3 is available in the Supporting Information. Additional research data supporting this publication are available from The University of Manchester eScholar repository at DOI: 10.15127/1.302780.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

A. Formanuik, F. Ortu, C. J. Inman, A. Kerridge, L. Castro, L. Maron, D. P. Mills, Chem. Eur. J. 2016, 22, 17976.