Andryj M.
Borys
and
Eva
Hevia
*
Departement für Chemie, Biochemie und Pharmazie, Universität Bern, 3012 Bern, Switzerland. E-mail: eva.hevia@unibe.ch
First published on 11th January 2023
Whilst low-valent nickelates have recently been proposed as intermediates in Ni-catalysed reactions involving polar organometallics, their isolation and characterisation is often challenging due to their high sensitivity and reactivity. Advancing the synthetic, spectroscopic and structural insights of these heterobimetallic systems, here we report a new family of alkyne supported alkali-metal nickelates of the formula Li4(solv)n(Ar)4Ni2{μ2:η2,η2-Ph–CC–Ph} (where solv = Et2O, THF; Ar = Ph, o-Tol, naphthyl, 4-tBu-C6H4) which can be accessed through the combination of Ni(COD)2, Ph–CC–Ph and the relevant lithium aryl in a 2:1:4 ratio. Demonstrating the versatility of this approach, the sodium and potassium nickelates can also be accessed when using PhNa or via alkali-metal exchange with AMOtBu (AM = Na, K). When employing bulky or structurally constrained aryl-lithiums, mononickel complexes of the formula Li2(solv)n(Ar)2Ni{η2-Ph–CC–Ph} are instead obtained, highlighting the structural diversity of alkali-metal nickelates bearing alkyne ligands. Expanding the catalytic potential of these systems, their ability to promote the catalytic cyclotrimerisation of diphenylacetylene to hexaphenylbenzene was explored, with mononickel compounds bearing electron rich aryl-substituents displaying the best performance.
Despite exhibiting a wealth of unique structural features, low-valent anionic nickelates remained dormant in the literature for several decades and were overshadowed by parallel developments on the applications of nickel complexes in catalysis.21 More recently, however, it has been proposed by theoretical studies that low-valent nickelates may be potential intermediates in certain nickel-catalysed cross-coupling reactions involving polar organometallics,22,23 prompting a renewed experimental interest into these overlooked species.
We have previously explored the rich co-complexation chemistry of Ni(COD)2 and PhLi, which gives a series of lithium nickelates including II (Scheme 1), and demonstrated that these are key intermediates in the nickel-catalysed cross-coupling of aryl ethers.19,24 When further assessing these co-complexation reactions, we found that using a larger excess of PhLi led to the formation of a polynuclear dinickel cluster containing a bridging C6H4 dianion as a result of intramolecular C–H activation of a phenyl substituent (Scheme 1, IV).25 The formation of IV suggests that the homoleptic tri-lithium nickelate “Li3NiPh3(solv)3” is too electron-rich to form, leading to the in situ formation of the π-accepting benzyne-type ligand. Only when moving to organolithiums which themselves could serve as π-accepting ligands, namely lithium aryl acetylides, is it possible to isolate homoleptic tri-lithium nickelates III, which were found to be further stabilised by London dispersion interactions.20 Notably, compound III reacts with PhI to give V in which the cross-coupled product (diphenylacetylene) is trapped and coordinated in a bridging motif between two nickel centres. Since V bears similar structural features to IV, we thus considered whether it was possible to use diphenylacetylene as a simple π-accepting ligand to access new families of alkali-metal nickelates. Herein, we detail synthetic and structural insights into dinickel and mononickelate complexes which can be readily accessed by reacting Ni(COD)2 with two molar equivalents of alkali-metal aryls in the presence of diphenylacetylene. Advancing the structure/reactivity correlations, we also assess the catalytic ability of these novel heterobimetallic complexes to promote the cyclotrimerisation of diphenylacetylene, as well as its insertion into biphenylene.
Scheme 2 Synthesis of dinickelate complexes 1a–f. Isolated crystalline yields given in parentheses. aPrepared by treating 1a with NaOtBu or KOtBu, respectively. bCould not be isolated in pure form. |
Interestingly, we also found that by systematically studying these co-complexation reactions when using ortho-substituted aryl-lithiums, o-Tol-Li and 1-naphthyl-Li, the formation of mononickelate complexes of the formula Li2(solv)n(Ar)2Ni{η2-Ph–CC–Ph} was instead favoured, with the choice of donor solvent playing an important role in dictating the crystallised product. With o-Tol-Li, small quantities of the Et2O solvate (1e) could be crystallised, but samples were always plagued with variable amounts Ni(COD)2 or [(COD)Ni]2{μ2-η2:η2-Ph–CC–Ph}, preventing its isolation in pure form. Contrastingly, performing the reaction with one equivalent of diphenylacetylene in the presence of THF, allowed Li2(THF)2(o-Tol)2Ni{η2-Ph–CC–Ph} (2a) to be reliably prepared and isolated in pure form (Scheme 3). Similarly with 1-naphthyl-Li, small quantities of the THF solvate (1f) could be crystallised, but not isolated in pure form. Performing the reaction with one equivalent of diphenylacetylene in Et2O however, allowed for the reliable isolation of Li2(Et2O)2(1-Naph)2Ni{η2-Ph–CC–Ph} (2b). Further exploring how general the formation of mononickelate alkyne complexes was, the bulky aryl-lithium 2,6-Me2-C6H3-Li also gave Li2(Et2O)2(2,6-Me2-C6H3)2Ni{η2-Ph–CC–Ph} (2c), whilst the structurally constrained aryl-lithium, 2,2′-dilithiobiphenyl, reacted smoothly to give Li2(THF)4(2,2′-biphenyl)Ni{η2-Ph–CC–Ph} (2d). Compound 2d is closely related to the dilithionickelole Li2(THF)4(2,2′-biphenyl)Ni{η2,η2-COD} which was previously reported by Xi and co-workers.28
Scheme 3 Synthesis of mononickelate complexes 2a–d. Isolated crystalline yields given in parentheses. |
Fig. 1 Variable temperature 1H NMR spectra of 1d in THF-d8 illustrating fast or frozen rotation about the Caryl–Ni bonds. |
The di- and mono-nickelate complexes can be further distinguished by assessing the 13C{1H} NMR signal of the acetylene carbon. For compounds 1a–d, this signal is observed in the range of δ 68.6–76.6, which is considerably shielded compared to [(COD)Ni]2{μ2-η2:η2-Ph–CC–Ph} (δ 106.9), and consistent with significant back-donation from the electron-rich Ni centres. Contrastingly, the acetylene carbon signal for compounds 2a–d is observed in the range δ 144.9–150.8, which is considerably deshielded compared to free diphenylacetylene (δ 90.8),29 and slightly shielded when compared to other L2Ni{η2-Ph–CC–Ph} complexes [L = N-heterocyclic carbene (δ 139.3)30 or diphosphine (δ 141.3)].31 The Caryl–Ni ipso-carbon signal for compounds 1a, 1d and 2a–d is observed in the range δ 177.2–191.1, which is similar to other phenyl-nickelate complexes such as II and IV (Scheme 1),19,25 and comparable to the free aryl-lithiums (see ESI† for full spectroscopic details). Notably, the ipso-carbon signal for the phenyl-derivatives becomes more deshielded on moving from Li (1a; δ 182.0) to Na (1b; δ 195.8) to K (1c; δ 203.8), consistent with increased charge density at the carbanionic centre.32
Compound | Distance/Å | Angle/° | Torsion/° | ||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Ni1–C1 | Ni–C2 | Ni2–C3 | Ni2–C4 | Ni⋯M range | Ni1⋯Ni2 | Ni–C6/7 range | C6–C7 | C5–C6–C7 | C8–C7–C6 | C5–C6–C7–C8 | |
a The Ph–CC–Ph is disordered across two positions. | |||||||||||
1a | 1.986(2) | 1.965(1) | 1.986(2) | 1.965(1) | 2.469(3)–2.663(3) | 2.633(3) | 1.947(1)–1.968(1) | 1.394(2) | 127.6(1) | 127.6(1) | 5.0(2) |
1.980(1) | 1.977(2) | 1.980(1) | 1.977(2) | 2.441(1)–2.654(3) | 2.654(3) | 1.936(1)–1.979(1) | 1.388(2) | 127.6(1) | 127.6(1) | 5.8(2) | |
1b | 2.042(8) | 1.957(3) | 2.042(8) | 1.957(3) | 2.825(1)–3.099(1) | 2.8873(7) | —a | 1.38(2) | 125.2(9) | 125.2(9) | —a |
1d | 1.980(2) | 1.983(2) | 1.973(2) | 1.968(2) | 2.395(5)–2.628(4) | 2.7079(8) | 1.948(2)–1.964(2) | 1.392(3) | 127.7(2) | 127.5(2) | 3.0(4) |
1e | 1.958(2) | 1.952(2) | 1.987(2) | 1.965(2) | 2.448(3)–2.758(3) | 2.7906(7) | 1.921(1)–2.013(2) | 1.386(2) | 132.6(1) | 129.8(1) | 20.1(3) |
1f | 1.977(3) | 1.959(3) | 1.977(3) | 1.959(3) | 2.674(5)–2.770(5) | 2.8218(8) | 1.931(3)–2.027(3) | 1.381(3) | 127.2(2) | 127.2(2) | 14.8(4) |
1.953(3) | 1.975(4) | 1.953(3) | 1.975(4) | 2.587(6)–2.795(6) | 2.7606(6) | 1.925(3)–2.034(2) | 1.392(4) | 126.7(2) | 126.7(2) | 21.8(4) |
A close inspection into the structures of compounds 1e and 1f, which are derived from o-Tol-Li and 1-naphthyl-Li respectively, reveal considerable structural distortion when compared to lithium nickelates 1a and 1d. In the solid-state structure of 1e, all ortho-CH3 substituents are orientated in the same direction towards the coordinated diphenylacetylene unit with Ni–Caryl–Cortho–CH3 torsion angles ranging from 1.5(3)–6.4(3)° (Fig. 2 & Fig. S10†). Most notably, one of the four Et2O molecules coordinates to Li2 via anagostic interactions (i.e. largely electrostatic) from the CH3 group [C–H⋯Li2 = 2.817–3.189 Å], in contrast to the expected oxygen coordination mode observed for Li1, Li3 and Li4. This is likely a consequence of the shorter Li2⋯C2 and Li2⋯C3 distances [2.257(4) Å and 2.197(4) Å, respectively] when compared to the Li3⋯C1 and Li3⋯C4 distances [2.446(4) Å and 2.563(3) Å, respectively]. Anagostic interactions between CH3 groups and Li have been observed in several crystal structures,37,38 including in lithium nickelates,39 but this is limited to intramolecular examples. In addition, there is considerable torsion in the diphenylacetylene unit [C5–C6–C7–C8 = 20.1(3)°, see Table 1 for general label numbering] to enable additional Carene⋯Li2 interactions [2.727(4)–2.750(4) Å].
Fig. 2 Molecular structure of 1e. Thermal ellipsoids shown at 30% probability and hydrogen atoms (except those on Et2O showing anagostic interactions to Li2) omitted for clarity. See Fig. S10† for alternative view. |
In the solid-state structure of 1f (see Fig. S11† for full structure), the 1-naphthyl substituents are similarly all orientated in the same direction towards the coordinated diphenylacetylene unit with Ni–Caryl–Cortho–CH3 torsion angles ranging from 2.9(5)–5.4(5)°. As with 1e, there is also significant torsion in the coordinated diphenylacetylene unit [C5–C6–C7–C8 = 14.8(4)° or 21.8(4)°] to enable additional Carene⋯Li2/3 interactions [2.671(5)–2.998(6) Å]. The central core of 1f shows considerable distortion when compared to 1a (Fig. 3). For example, in 1a, the four ipso-carbons (C1–C4) lie in an approximate plane which sits co-planar below the mean plane of all four lithium atoms (Li1–Li4). In 1f, neither the four ipso-carbons (C1–C4) or the four Li atoms (Li1–Li4) reside in an approximate plane. Whilst Li2 and Li3 are approximately co-planar with Ni1 and Ni2 (∼0.1 Å deviation), Li1 and Li4 sit considerably lower than Li2/Li3 and Ni1/Ni2, as well as the four ipso-carbons (C1–C4). This exposes the Li1/Li4 cations such that two molecules of THF can coordinate to each lithium, in contrast to only one molecule of coordinated Et2O observed in 1a, 1d and 1e. Whilst the identity of the ethereal solvent clearly impacts the isolation and crystallisation of compounds 1e and 1f over their mononickelate analogues 2a and 2b, it is unclear whether the unique solvent coordination in 1e and 1f is the cause or simply a consequence of the observed structural distortions.
Fig. 3 Simplified top and side views of the hexanuclear cores of 1a and 1f illustrating the structural distortion in 1f when compared to 1a. |
In the solid-state structure of mononickelate complexes 2c and 2d (Fig. 4), the Ni centre adopts a pseudo trigonal planar geometry in which the diphenylacetylene coordinates in a η2-fashion. The Ni⋯CC distances are shorter [1.892(1)–1.932(1) Å for 2c; 1.877(2)–1.881(2) Å for 2d] when compared to dinickelate complexes 1a–f (see Table 1). In addition, the CC bond lengths [1.318(2) Å for 2c; 1.324(2) for 2d] are shorter, and the Cipso–CC angles are closer to 180° [134.8(1)–135.4(1)° for 2c; 136.1(1)–137.6(1)° for 2d], indicative of weaker overall back-bonding from Ni to the coordinated diphenylacetylene. The Li⋯Ni distances for 2c and 2d range from 2.408(3)–2.515(3) Å, which is comparable to dinickelate complexes 1a and 1d–f, and other structurally characterised lithium nickelates.19,20,25
Fig. 4 Molecular structure of 2d. Thermal ellipsoids shown at 30% probability. Hydrogen atoms and two molecules of coordinated THF omitted for clarity. |
The catalytic activity of the neutral dinickel olefin complexes [(COD)Ni]2{μ2-η2:η2-Ph–CC–Ph}27 and Ni2COT2 (COT = cyclooctatetraene)50 was first evaluated (Table 2 and Table S1†). Using 5 mol% of dinickel catalyst, only modest conversions (38% and 41%, respectively) of diphenylacetylene was observed after heating to 80 °C for 4 hours (entries 1 and 2). Moving to the lithium nickelate Li4(Et2O)4Ph4Ni2{μ2-η2:η2-Ph–CC–Ph} (1a, entry 3) led to a considerable increase to 74% conversion. Contrasting with other catalytic studies using alkali-metal magnesiates,51–55 no apparent alkali-metal effect was observed, with comparable conversions observed for the sodium nickelate 1b (73%, entry 4) and potassium nickelate 1c (78%, entry 5), indicating that the alkali-metal does not appear to play an active role in catalysis. Using compound 1d as a catalyst gave a slightly improved conversion (85%, entry 6) suggesting that the more electron-rich 4-tBu-C6H4 substituents enhances the catalytic activity when compared to 1a. Supporting this claim, the acetylide substituted dinickel compound Li4(Et2O)4(Ph–CC)4Ni2{μ2-η2:η2-Ph–CC–Ph} (V, Scheme 1) was completely inactive for the catalytic cyclotrimerisation of diphenylacetylene (entry 7), but could catalyse the cyclotrimerisation of more activated terminal alkynes such as phenylacetylene (see ESI† for further details).
Entry | Catalyst | Conversiona (%) |
---|---|---|
a Spectroscopic conversion of diphenylacetylene monitored using hexamethylbenzene as an internal standard. b Reaction performed in THF-d8. | ||
1 | 5% [(COD)Ni]2{PhCCPh} | 38 |
2 | 5% Ni2COT2 | 41 |
3 | 5% Li4(Et2O)4Ph4Ni2{PhCCPh}, 1a | 74 |
4 | 5% Na4(THF)6Ph4Ni2{PhCCPh}, 1b | 73 |
5 | 5% K4(THF)4Ph4Ni2{PhCCPh}, 1c | 78b |
6 | 5% Li4(Et2O)4(4-tBu-C6H4)4Ni2{PhCCPh}, 1d | 85 |
7 | 5% Li4(Et2O)4(CCPh)4Ni2{PhCCPh}, V | Trace |
8 | 10% Li2(THF)2(o-Tol)2Ni{PhCCPh}, 2a | >95 |
9 | 10% Li2(Et2O)2(1-Naph)2Ni{PhCCPh}, 2b | 74 |
10 | 10% Li2(Et2O)2(2,6-Me2-C6H3)2Ni{PhCCPh}, 2c | >95 |
11 | 10% Li2(THF)4(2,2′-biphenyl)Ni{PhCCPh}, 2d | 26b |
Full conversion (>95%) of diphenylacetylene was observed when using Li2(THF)2(o-Tol)2Ni{μ2-η2:η2-Ph–CC–Ph} (2a, entry 8) or Li2(Et2O)2(2,6-Me2-C6H3)2Ni{μ2-η2:η2-Ph–CC–Ph} (2c, entry 10) as the catalyst, demonstrating that the mononickelate complexes show superior catalytic activity compared to the dinickelate complexes 1a–d. The electron-deficient derivative Li2(Et2O)2(1-Naph)2Ni{μ2-η2:η2-Ph–CC–Ph} (2b) showed reduced conversions (74%, entry 9), again illustrating how the electronic properties of the aryl-substituents can influence catalytic activity.
Interestingly, when using Li2(THF)4(2,2′-biphenyl)Ni{μ2-η2:η2-Ph–CC–Ph} (2d) as the catalyst, only low conversions of diphenylacetylene (26%, entry 11) were achieved, and characteristic signals consistent with 9,10-diphenylphenanthrene could be observed by 1H NMR spectroscopy. The formation of this product has been previously reported when treating (L)nNi(2,2′-biphenyl) [where L = (Et3P)2 or iPr2PCH2CH2PiPr2] complexes with diphenylacetylene,31,56 and this could be upgraded to catalytic regimes when using biphenylene and diphenylacetylene in the presence of O2.31 Compound 2d was also found to catalyse the insertion of diphenylacetylene into the strained C–C bond of biphenylene (Scheme 4), however this process is slow (10 mol%, 40 hours, 80 °C), particularly when compared to a Ni(NHC)2 catalyst reported by Radius (2 mol%, 30 min, 80 °C).30 Attempts to spectroscopically identify or isolate possible intermediates for the [2 + 2 + 2] cyclotrimerisation of diphenylacetylene, or insertion of diphenylacetylene into biphenylene failed.
Footnote |
† Electronic supplementary information (ESI) available: Full synthetic details, crystallographic information, and NMR spectra. CCDC 2221333–2221339. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d3dt00069a |
This journal is © The Royal Society of Chemistry 2023 |