Richard A.
Manzano
and
Anthony F.
Hill
*
Research School of Chemistry, Australian National University, Canberra, Australian Capital Territory ACT 2601, Australia. E-mail: a.hill@anu.edu.au
First published on 7th March 2023
The new fluorocarbynes [M(CF)(CO)2(Tp*)] (M = Mo, W; Tp* = tris(dimethylpyrazolyl)borate) arise from electrophilic fluorination of the lithiocarbynes [M(CLi)(CO)2(Tp*)] with FN(SO2Ph)2. The reactions of [W(CF)(CO)2(Tp*)] with [AuCl(SMe2)] and PhICl2 afford the first μ2-fluorocarbyne complex [WAu(μ-CF)Cl(CO)2(Tp*)] and the first high oxidation state fluorocarbyne [W(CF)Cl2(Tp*)], respectively.
Scheme 1 Synthesis of fluorocarbyne complexes via fluoride abstraction from CF4, trifluoromethyl and difluorocarbene ligands. (i) Ar(s), 8K.1 (ii) 2KC8, –2F−, –CO.2 (iii) [H(SiEt3)2][HCB11Cl11], –Et3SiF.3 |
Beyond the addition of cobalt carbonyl to [Mo(CF)(CO)2(η5-C5Me5)] to afford a μ3-fluorocarbyne capped cluster [MoCo2(μ3-CF)(CO)8(η5-C5Me5)],2a no subsequent reactivity of fluorocarbyne ligands has been reported. This is in marked contrast to the heavier halocarbynes [M(CX)(CO)2(Tp*)] (M = Mo, W; X = Cl, Br; Tp* = hydrotris(dimethylpyrazolyl)borate),5 the synthetic utility of which has been convincingly demonstrated.6–9 This utility centres on nucleophilic substitution of the halogen that may be spontaneous6–9 or palladium-mediated,10 the latter via the intermediacy of μ-carbido complexes [MPd(μ-C)Br(CO)2(PPh3)2(Tp*)].11 Alternatively, lithium/halogen exchange with nBuLi affords, in situ, the lithiocarbynes [M(CLi)(CO)2(Tp*)] (M = Mo, W) that allow the introduction of carbyne substituents in electrophilic form.9,12 This latter approach has allowed the synthesis of carbyne ligands that bear substituents based on boron and all the elements of groups 14–16 as well as a range of transition metals.11,12
Missing from the otherwise complete series of molybdenum and tungsten halocarbyne complexes are the lightest members [M(CF)(CO)2(Tp*)] (M = W 1a, Mo 2a) that might be seen as analogues of Hughes' fluorocarbynes [M(CF)(CO)2(η5-C5R5)]. These are not available via Lalor's original halocarbyne synthesis due to a range of very specific factors peculiar to that system,5b thereby preventing a complete comparative analysis beyond computational interrogation of the hypothetical complexes [Mo(CX)(CO)2(Tp)] (X = F, Cl, Br, I; Tp = hydrotris(pyrazolyl)borate).13 Herein, we now describe (i) the synthesis of the low-valent fluorocarbyne complexes 1a and 2avia a novel strategy; (ii) the reaction of 1a to afford the first example of a μ2-fluorocarbyne complex 3; (iii) oxidation to afford the first isolable1 high-valent fluorocarbyne complex [W(CF)Cl2(Tp*)](4).
Scheme 2 Synthesis of fluorocarbyne complexes via electrophilic fluorination of lithiocarbynes. (i) nBuLi, –nBuBr. (ii) F–N(SO2Ph)2. (iii) [AuCl(SMe2)]. (iv) PhICl2. |
Treating the lithiocarbyne [W(CLi)(CO)2(Tp*)], generated in situ, with an electrophilic source of fluorine, viz. FN(SO2Ph)2,15 affords the new fluorocarbyne complex [W(CF)(CO)2(Tp*)] (1a, 59%). Similar treatment of [Mo(CBr)(CO)2(Tp*)] (2c) affords the molybdenum analogue [Mo(CF)(CO)2(Tp*)] (2a, 77%, Scheme 2).
In addition to spectroscopic and electrochemical data (Table 1) 1a was characterised crystallographically to confirm connectivity, however detailed geometric analysis is precluded by positional disorder of isosteric CO and CF ligands (one contributor shown in Fig. 1 inset), as also recognised for Hughes' complexes [M(CF)(CO)2(η5-C5R5)] (M/R = Mo/Me, W/Me, W/H).2 Accordingly, and not surprisingly, crystals of 1a were found to be essentially isomorphous with those of the radical [W(CO)3(Tp*)].16 Crystals of the molybdenum complex 2a, however, led to a precise structural model free of disorder, geometric data for which indicate that the CF ligand exerts a pronounced trans influence on the unique pyrazolyl donor that is reproduced in the computationally optimised geometries of 1a, 2a and the simpler model complex [W(CF)(CO)2(Tp)] (see ESI†). It is a useful feature of the Tp* ligand that in contrast to cyclopentadienyls for which analogies are entertained, the more clearly octahedral geometry allows both structural and computational interrogation of bonding and implications, e.g., the trans influence. Comparative data for F–C(sp) bonds are surprisingly limited to Hughes' (in all but one case disordered) fluorocarbynes2,3 and the unique fluoroethynyl complex [Ru(CCF)(dppe)(η5-C5Me5)] (C–F = 1.324(4) Å).15a Amongst the spectroscopic data of interest, the carbyne gives rise to a doublet resonance in the 13C{1H} NMR spectrum at δC = 200.8 showing coupling to both 19F (1JCF = 528 Hz) and 183W (1JCW = 274.6 Hz) nuclei, the latter being somewhat larger than typically found for other carbyne complexes of pseudo-octahedral tungsten.5c The 19F{1H} NMR spectrum comprises a single resonance (δF = 45.5) straddled by a doublet due to the 183W isotopomer (2JWF = 111.1 Hz).
M | R | ν CO cm−1 | k CO N cm−1 | δ WC ppm | 1 J WC Hz | E 0 V |
---|---|---|---|---|---|---|
a Measured in CH2Cl2. b k CK = Cotton–Kraihanzel CO force constant. c Measured in CDCl3. d Relative to ferrocene (E1/2 = 0.460 V cf. Ag/Ag+ = 0), measured in CH2Cl2 with 1.0 M [NnBu4][PF6] at 100 mV s−1; see ESI for cyclic and square-wave voltammograms. n.r. = not reported. | ||||||
W | F (1a) | 1988, 1888 | 15.15 | 200.8 | 275 | 0.795 |
W | Cl (1b) | 1991, 1902 | 15.29 | 205.7 | 259 | 0.830 |
W | Br (1c) | 1994, 1905 | 15.33 | 198.0 | 254 | 0.820 |
W | I (1d)14 | 1992, 1907 | 15.33 | 183.2 | n.r. | — |
Mo | F (2a) | 2002, 1910 | 15.44 | 193.5 | — | 0.865 |
Mo | Cl (2b) | 2005, 1921 | 15.54 | 208.7 | — | 0.825 |
Mo | Br (2c) | 2008, 1924 | 15.59 | 202.5 | — | 0.830 |
Mo | I (2d)5b | 2009, 1927 | 15.62 | n.r. | — | — |
W | H14 | 1986, 1891 | 15.16 | 280.6 | 192 | n.r. |
W | CH3 (ref. 7d and 17) | 1968, 1867 | 14.86 | 289.3 | n.r. | +0.510 |
W | Ph18 | 1969, 1876 | 14.91 | 277.9 | 187 | n.r. |
W | CCtBu10c,19 | 1977, 1886 | 15.05 | 250.6 | 199 | +0.761 |
With the complete series of halocarbynes now in hand, a host of physico-chemical data is available to benchmark the nature of the fluorocarbyne ligand against other halocarbynes and more conventional carbyne ligands in the complexes [W(CR)(CO)2(Tp*)] (R = F, Cl, Br, I, H,14 CH3,17 Ph,18 CCPh,19Table 1) and to validate recent predictions based on the model complexes [Mo(CX)(CO)2(Tp)] (X = F, Cl, Br, I).13
Descending group 17 for the halocarbyne series, the metal centre becomes marginally more electron-poor and less π-basic (increase in νCO and kCO). In the absence of 1a it might seem that the carbyne 13C resonance moves to higher frequency down group 17 showing a normal halogen dependence arising from relativistic spin–orbit coupling.20 The non-conforming position of 1a in the series therefore most likely reflects a competing inverse halogen effect arising from paramagnetic shielding due to magnetic coupling of occupied and unoccupied orbitals, e.g., σ*(C–F) and π*(WC). The oxidation potential is remarkably insensitive to the nature of the halogen, as might be expected from computational studies on the hypothetical series [Mo(CX)(CO)2(Tp)] that show the energy of the metal-based HOMO (orthogonal to the C–X vector) to be essentially in variant.
Whilst μ3-fluorocarbyne complexes are well-known2a,21 there are no previous examples of doubly bridging fluorocarbyne ligands. The addition of the ‘AuCl’ fragment to terminal carbyne complexes has on numerous occasions been shown to afford heterobimetallic μ2-carbynes22 including the recent isolation of bromo- and chlorocarbyne examples.13 Accordingly, the reaction of 1a with [AuCl(SMe2)] was investigated and found to afford the μ2-fluorocarbyne complex [WAu(μ-CF)Cl(CO)2(Tp*)] (3). The resonance for the fluorocarbyne carbon appears at δC = 236.6, to higher frequency of the precursor and displays a reduced coupling to tungsten-183 (1JCW = 143.8 Hz) and fluorine-19 nuclei (1JCF = 454 Hz), consistent with a decrease in s-character for the sp2vs. sp carbon hybridisation and the increase in W–C bond length (1.876(6) Å). The molecular structure (Fig. 2) reveals that the fluorocarbyne ligand adopts an unsymmetrical semi-bridging mode (W1–C1–F1 = 148.7(6)°), as is commonly encountered for carbyne (and carbonyl) ligands bridging between tungsten and d10 metal centres. Therein, the carbyne may be viewed as a Z-type σ-acceptor ligand, i.e., donation from gold(I) to a π-hole on the carbonyl ligand. Similar geometric features are reproduced upon geometry optimisation (ωB97X-D/6-31G*/LANL2Dζ, see ESI†) of the model complex [WAu(μ-CF)Cl(CO)2(Tp)] (3′: WC = 1.893 Å; Au–C = 2.042 Å, W–Au = 2.829 Å, W–C–Au = 91.8° W–C–F = 149.7°) for which Löwdin bond orders (WC = 1.80, AuC = 0.72) suggest that a dimetallacyclopropene description is useful.
The complex 3 is unstable in solution, gradually depositing elemental gold (accompanied by some reformation of 1a) to provide a new purple species identified as the high-valent fluorocarbyne complex [W(CF)Cl2(Tp*)] (4). Once recognised as such, complex 4 could be more expediently and directly prepared from 1a and iodobenzene dichloride. Key spectroscopic features of note for 4 include a resonance for the carbyne carbon (δC = 224.9) that is moved to higher frequency of that for 1a, with a similar 1JCF coupling (512.8 Hz) but augmented 1JWC coupling (316 Hz). The 2JWF coupling (105.1 Hz) observed in the 19F{1H} NMR spectrum is similar to that observed for 1a. Taken together, these suggest retention of octahedral geometry at tungsten and sp-hybridisation at the carbyne carbon. Oxidative interconversion of ‘Fischer’ and ‘Schrock’ classes of carbyne is well established23 but in this case is gratifying in that 4 represents a connection between the exotic matrix-isolated ‘Schrock-type’ carbynes X3MCF (X = F, Cl)1 and tractable low-valent compounds such as 1a, 2a and Hughes' [M(CF)(CO)2(η5-C5R5)] (M = Cr, Mo, W; R = H, Me). For simplicity, the frontier molecular orbitals of interest for the model complexes [W(CF)(CO)2(Tp)] (1a′) and [W(CF)Cl2(Tp)] (4′) are shown in Fig. 3 (those for the full and sterically congested molecules 1a and 4 are provided in the ESI† alongside those for the matrix molecules X3WCF, X = F, Cl, calculated at the same level of theory). In the case of 1a, the HOMO is primarily associated with metal carbonyl back-donation (‘dxy’ taking ‘z’ as the W⋯F vector), explaining why the oxidation potentials (Table 1) are rather insensitive to changes in the carbyne substituent. The HOMO − 1 and HOMO − 2 comprise the two orthogonal π-components of the WC multiple bond for 1a′. For 4′, the manifold of frontier orbitals is rather similar with the notable exception that the ‘dxy’ orbital is now vacant and available for nucleophilic attack. This leaves the near degenerate pair of WC π-bonding orbitals as the highest in energy, as is also the case for the Schrock-type molecules X3WCF. Notably, for 1a′, while the LUMO has both metal and carbyne–carbon character, the real complex 1a is not especially susceptible to hydrolysis (cf. difluorocarbene complexes), easily surviving chromatographic purification without any identifiable formation of the anticipated hydrolysis product [WH(CO)3(Tp*)]. The situation for the high-valent 4′ is intriguing in that, by analogy with chloro- and bromo-carbynes, nucleophilic substitution of the fluoride would seem plausible. Furthermore, oxidation state has been shown to profoundly enhance the susceptibility of difluorocarbene ligands towards nucleophilic attack, as demonstrated for the zero- and divalent ruthenium complexes [Ru0(CF2)(CO)2(PPh3)2]24cf. [RuIICl2(CF2)(CO)(PPh3)2], respectively.25 High-valent Schrock-type carbyne complexes of the form [W(CR)X2(Tp′)] (X = Cl, Br; Tp′ = Tp, Tp*)23c,26 are generally prone to hydrolysis via nucleophilic attack (H2O/HO−) at the metal to provide [W(CHR)Cl(O)(Tp′)]. It is therefore noteworthy and somewhat surprising that complex 4 is comparatively robust, air stable and not readily hydrolysed.
Fig. 3 Frontier orbitals (ωB97X-D/6-31G*/LANL2Dζ/gas phase) of interest for the complexes [W(CF)(CO)2(L)] and [W(CF)Cl2(L)] (L = Tp, red, Tp* blue). |
In conclusion, electrophilic fluorination of carbido ligands provides a viable route to fluorocarbyne complexes, building upon previous examples of the iodination of terminal carbido ligands14,27 and the chlorination or bromination of bridging μ2-carbido complexes.28 The ‘Fischer-type’ fluorocarbyne [W(CF)(CO)2(Tp*)] may be oxidised to a rare example23d of a high oxidation state halocarbyne [W(CF)Cl2(Tp*)] that is best viewed as a ‘Schrock-type’ carbyne.
Footnote |
† Electronic supplementary information (ESI) available: Synthetic procedures, spectroscopic, computational and crystallographic data. CCDC 2226627 and 2226628. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d3sc00261f |
This journal is © The Royal Society of Chemistry 2023 |