Reversible, room-temperature C-C bond activation of benzene by an isolable metal complex

: The activation of C–C bonds is of fundamental interest in the construction of complex molecules from petrochemical feedstocks. In the case of the archetypal aromatic hydrocarbon benzene, C–C cleavage is thermodynamically disfavoured, and is brought about only by transient highly reactive species generated in situ . Here we show that the oxidative addition of the C–C bond in benzene by an isolated metal complex is not only possible, but occurs at room temperature and reversibly at a single aluminium centre in [(NON)Al] - (where

From the seminal work of Kekulé onwards, the cyclic structure of benzene has played a central role in rationalizing structure and reactivity in organic chemistry. 1,2he recognition of enhanced thermodynamic stability for the planar C 6 H 6 molecule arising from -electron delocalization proved critical to the development of the concept of aromaticity, 3,4 which in turn has been used to rationalize why much of the chemistry of benzene (and related compounds) proceeds with net retention of the unsaturated C 6 core. 51][12][13][14][15][16][17][18][19] Oxidative activation of the C-C bond in benzene is therefore known only in conjunction with transient, highly reactive species which are generated in situ (such as carbenes or triplet silylenes). 20-23A well-known example is the Buchner ring expansion reaction, which classically involves the reaction of an unactivated arene with a carbene derived from a diazo-acetic ester by a thermal, photolytic or transition-metal catalysed protocol. 20,21e recently reported the dimeric potassium aluminyl compound 1 (Scheme 1), which acts as an unprecedented aluminium nucleophile (e.g. in reactions with carbon-and metal-centered electrophiles), and which effects the formal oxidative addition of the C-H bond in benzene to give the corresponding aluminium(III) phenyl hydride. 24,25DOSY NMR measurements show that the dinuclear structure of 1 determined in the solid state by X-ray diffraction is retained in benzene-d 6 solution.With the aim of investigating whether the chemistry of this system is dependent on its degree of aggregation, we set out to synthesize a related mononuclear compound by the addition of the K + sequestering agent [2.2.2]cryptand (4,7,13,16,21,24- The compound so generated, [K(2.2.2-crypt)] [(NON)Al] (2; where NON = 4,5-bis(2,6-diisopropylanilido)-2,7-di-tert-butyl-9,9-dimethylxanthene), can be shown by X-ray crystallography to be monomeric in the solid state (Figure 1), featuring no short Al … Al or K … Al contacts (< 7 Å).The geometry at the aluminium centre is very similar to that found in 1, 24    While 1 reacts with benzene over a period of 96 h at 60 o C (via C-H bond cleavage), 24 the corresponding reaction with 2 proceeds at room temperature (in the dark), with conversion occurring over a period of 48 h into a single species which can be isolated in ca.95% yield (Scheme 2).This new compound (3) can be shown by microanalysis to be the product of the uptake of one molecule of benzene by 2, and is characterized in solution by three new signals in the 1 H NMR spectrum, at chemical shifts ( H = 7.21, 7.00, 5.97 ppm) characteristic of alkenic C-H bonds. 27The structure of 3 is confirmed unambiguously by X-ray crystallography (Figure 1), and is shown to result from C-C bond activation of the C 6 H 6 molecule at a single aluminium centre to give a sevenmembered AlC 6 H 6 metallacycle.The Al-C bond lengths [both 1.980(2) Å] are in line with other structurally characterized vinyl-aluminium species, 28 and the C-C bond lengths [1.343(2), 1.442(3), 1.352(3), 1.451(3) and 1.352(3) Å] are consistent with alternating C=C double and C-C single bonds within the metallacycle. 261][22][23] By contrast, here we show by experimental and computational means that cleavage of the C-C bond in the C 6 H 6 molecule by 2 is completely reversible, facilitated by the tractable (but reactive) nature of 2 and a thermally accessible barrier to dissociative loss of benzene from 3. Exchange of the C 6 H 6 fragment within the metallacycle component of 3 with C 6 D 6 can be brought about at temperatures in the range 60-80 o C to generate the per-deuterio analogue 3-d 6 (Scheme 2).In situ NMR monitoring of the reaction as a function of temperature allows the activation parameters associated with the exchange process to be evaluated.The values determined for H ‡ and S ‡ (119.3(2.2) kJ mol -1 and 23.9(6.5)J mol -1 K -1 ) are consistent with a dissociative mechanism for benzene substitution, and the magnitude of the free energy barrier G ‡ (111 kJ mol -1 at 80 o C) is in line with that calculated by DFT methods for the transition state leading to ejection of the C 6 H 6 moiety from 3 (vide infra).
As a further probe of the reversibility of benzene addition to 2, we examined the reactivity of both 2 and 3 towards naphthalene (Scheme 3).We hypothesized that chemical reversibility (i.e.facile release of benzene from 3) would be signaled by reactivity towards a molecular probe which is indistinguishable from that of the 'naked' aluminyl complex 2. In the case of 2, reaction with naphthalene at 80 o C generates a mixture of two isomeric products resulting from formal C-H oxidative addition at aluminium; the 2-and 1-naphthyl products (4a/b) are formed in a 3:2 ratio (Scheme 3 and ESI). 29Under analogous conditions, the reaction of 3 with naphthalene yields a near identical mixture of the same two products, thereby presenting further evidence that the C-C addition of benzene to 2 is chemically reversible under mild conditions.Scheme 3. Reactions of 2 and 3 with naphthalene at 80 o C to give a mixture of C-H activation products resulting from attack at the 1 and 2 positions. 29igure 2. DFT-calculated (DFT-D3, PBE0, Def-TZVP) pathways for benzene C-C and C-H bond activation by 2'.For computational efficiency the backbone t Bu groups of the NON ligand were replaced by methyl groups in all calculations.The 'cut-down' versions of compounds 2, 3 and 5 so modeled are labeled 2', 3' and 5'.
Given the unprecedented reversibility implied by these experimental studies, DFT calculations were carried out in order to better understand issues of thermodynamics and kinetics relating to the activation of benzene by 2 (Figure 2 and ESI).For computational efficiency the backbone t Bu groups of the NON ligand were replaced by methyl groups in all calculations.These show that both C-C and C-H activation processes by model system 2' are exergonic, but to widely differing degrees (G = -17 and -146 kJ mol -1 , respectively), in line with the loss of aromaticity in the C 6 H 6 molecule on formation of 3'.However, the rate-determining transition state leading to the formation of C-C activated 3' is calculated to be >20 kJ mol -1 lower than that leading to the (phenyl)hydride compound 5', consistent with the observed formation of 3' as the kinetic product.Also consistent with this model is the experimental observation that heating solutions of 2 or 3 in benzene to 80 o C over an extended period leads to the formation of the thermodynamic product 5 (although the reaction is not clean, generating a mixture of products thought to result from competing activation of the NON ligand backbone and the cryptand framework). 30echanistically, the formation of 3' from 2' proceeds in two steps via a bicyclic intermediate (Int), resulting from initial 2][33] The transition states associated with the formation of this intermediate and its further activation to give 3' are calculated to be similar in energy (+97 and +90 kJ mol -1 ), and in each case lower that calculated for C-H activation (120 kJ mol -1 ).These calculated barriers are consistent with the experimental observation of selective C-C activation by 2 at room temperature.The high-lying nature of the aluminiumcentred lone pair (the highest occupied molecular orbital, HOMO) for the monomeric [(NON)Al] -ion presumably facilitates the initial cycloaddition step, with the significant stabilization of this orbital in dimeric K +bridged [K(NON)Al] 2 (i.e.1), 24 potentially underpinning differences in reactivity between 1 and monomeric 2 toward benzene.Experimental data are also consistent with the idea of a facile initial [2+1] cycloaddition at the aluminyl centre: the reactions of 1 or 2 with ethene proceed rapidly at room temperature to generate metallacyclopropane complexes containing the [(NON)Al( 2 -C 2 H 4 ) 2 ] -unit in essentially quantitative fashion.In the case of the dimeric species [K(NON)Al( 2 -C 2 H 4 ) 2 ] 2 (6) obtained from 1, structural authentication by X-ray crystallography reveals a C-C bond length for the coordinated C 2 fragment [1.598(2)Å] which is significantly activated compared to ethene, and more consistent with a C-C single bond (see ESI). 26 In the case of benzene activation, the corresponding bicyclic metallacyclopropane represents an intermediate species (i.e.Int), from which a 6π electrocyclic ring-opening process then generates the seven-membered heterocycle present in the product 3'.
Given the ability of 2 to cleave C-C bonds under ambient conditions, we postulated that this system could be used in an unprecedented synthetic approach to access acyclic derivatives (for example containing a linear cis, cis, cis triene fragment) from benzene by reactions with electrophiles.Accordingly, as proof of principle, the reactions of 3 with tin halides (such as Me 2 SnCl 2 ) can be shown to proceed at room temperature to generate di-tin species bridged by an acyclic C 6 H 6 fragment, in which the stereochemistry of the alkene units has been retained from the benzene precursor (Scheme 4).The structures of both the aluminium- (7) and tin-containing products (8) from this reaction have been confirmed crystallographically (see ESI).Scheme 4. Generation of an acyclic product containing a cis, cis, cis triene fragment by the reaction of 3 with a tin electrophile.
In conclusion, we show here that (i) that a highly electron rich (but well characterized) 'bottle-able' aluminium complex can be used to oxidatively activate the C-C bond in benzene; (ii) that this process is reversible; and (iii) that it can be used to ultimately synthesize functionalized acyclic compounds from benzene.

ASSOCIATED CONTENT
Supporting Information.Full synthetic and characterizing data for new compounds, representative NMR spectra, details of crystallographic and computational studies (PDF).CIFs for X-ray crystal structures (CCDC: 1899005-1899011).

The 1 H
NMR spectrum of this compound in tetrahydrofuran-d 8 solution shows resonances at chemical shifts different from those characteristic of dimeric 1 (together with signals for the cryptand ligand which are shifted from those of free [2.2.2]cryptand), consistent with the existence in solution of separated [K(2.2.2crypt)] + and [(NON)Al] -ions.

Scheme 2 .
Room temperature C-C bond activation of benzene by 2 to give metallacycle 3; facile exchange of the C 6 H 6 fragment of 3 in the presence of C 6 D 6 .