Variable Ca‐Caryl Hapticity and its Consequences in Arylcalcium Dimers

Abstract The dimeric β‐diketiminato calcium hydride, [(DippBDI)CaH]2 (DippBDI = HC{(Me)CN‐2,6‐i‐Pr2C6H3}2), reacts with ortho‐, meta‐ or para‐tolyl mercuric compounds to afford hydridoarylcalcium compounds, [(DippBDI)2Ca2(μ‐H)(μ‐o‐,m‐,p‐tolyl)], in which dimer propagation occurs either via μ2‐η1‐η1 or μ2‐η1‐η6 bridging between the calcium centers. In each case, the orientation and hapticity of the aryl units is dependent upon the position of the methyl substituent. While wholly organometallic meta‐ and para‐tolyl dimers, [(DippBDI)Ca(m‐tolyl)]2 and [(DippBDI)Ca(p‐tolyl)]2, can be prepared and are stable, the ortho‐tolyl isomer is prone to isomerization to a calcium benzyl analog. Computational analysis of this latter process with density functional theory (DFT) highlights an unusual mechanism invoking the generation of an intermediate dicalcium species in which the group 2 centers are bridged by a toluene dianion formed by the formal attachment of the original hydride anion to the initially generated ortho‐tolyl substituent. Use of a more sterically encumbered aryl substituent, {3,5‐t‐Bu2C6H3}, facilitates the selective formation of [(DippBDI)Ca(μ‐H)(μ‐3,5‐t‐Bu2C6H3)Ca(DippBDI)], which can be converted into the unsymmetrically‐substituted σ‐aryl calcium complexes, [(DippBDI)Ca(μ‐Ph)(μ‐3,5‐t‐Bu2C6H3)Ca(DippBDI)] and [(DippBDI)Ca(μ‐p‐tolyl)(μ‐3,5‐t‐Bu2C6H3)Ca(DippBDI)] by reaction with the appropriate mercuric diaryl. Conversion of [(DippBDI)Ca(H)(Ph)Ca(DippBDI)] to afford [{{(DippBDI)Ca}2(μ2‐Cl)}2(C6H5‐C6H5)], comprising a biphenyl dianion, is also reported. Although this latter transformation is serendipitous, AIM analysis highlights that, in a related manner to the ortho‐tolyl to benzyl isomerization, the requisite C–C coupling may be facilitated in an “across dimer” fashion by the experimentally‐observed polyhapto engagement of the aryl substituents with each calcium.


Introduction
Although it is now more than 120 years since the discovery of Grignard's eponymous reagents, [1] organomagnesium compounds continue to prevail as some of the most generally useful and widely employed organometallics in chemical synthesis. [2]In contrast, and despite sporadic reports of analogous RCaX species dating from a similar time period, [3] organic derivatives of calcium remained largely neglected throughout much of the 20 th century and it is only during the past 20 years that a defined and characteristic chemistry has started to emerge. [4]The groups of Anwander and Westerhausen have provided particularly notable recent contributions through their respective development and structural elucidation of dimethylcalcium and a wide variety of di-and monoarylcalcium reagents. [5]eports of subsequent onward reactivity for such species, however, remain relatively sparse, [5a,6] most likely due to their limited tractability in comparison to magnesiumderived systems.
In parallel with these advances, the last 20 years have seen the more general emergence of calcium-based reagents, primarily as Earth-abundant and inexpensive vectors for homogeneous catalysis. [7]Central to these latter efforts has been the use of a variety of multidentate uninegative supporting Scheme 1. Use of compound 1 in the a) preparation of alkyl calcium complexes and the subsequent nucleophilic alkylation of benzene.b) synthesis of the calcium phenyl (3) and its reaction with aryl bromides to prepare biaryl molecules.
anions (L) as spectator ligands in the study of heteroleptic derivatives, LCaX (where X =, e.g., amide, phosphide, alkyl, hydride).While the primary requirements of L are to enhance LCaX solubility in non-coordinating solvents and suppress otherwise deleterious Schlenk-type equilibration, its steric demands must be appropriately and conveniently perturbed to maintain an appropriate level of kinetic reactivity at the Ca-X bonded unit.The -diketiminate (BDI) class of ligand has, thus, been particularly prominent in these advances. [8]While recent developments have progressed to more sterically encumbered variants, [9] the Dipp BDI ligand ( Dipp BDI = HC{(Me)CNDipp} 2 , where Dipp = 2,6i-Pr 2 C 6 H 3 ) has proved a reliable and readily accessible workhorse that has facilitated a wide variety of new chemistry.Reaction of the otherwise base-free hydride complex, [( Dipp BDI)CaH] 2 (1), with terminal alkenes has, for example, allowed access to a broad palette of dimeric -alkyls, [( Dipp BDI)CaR] 2 . [10]10a,11] More recently, we have reported that the -phenyl derivatives, [( Dipp BDI)Ca(-H)(-Ph)Ca( Dipp BDI)] (2) and [( Dipp BDI)Ca(-Ph)] 2 (3) are accessible via sequential reactions of compound 1 with Ph 2 Hg (Scheme 1b). [12]6b] In this contribution, we report our investigation of the synthesis of alternative hydridoarylcalcium complexes, documenting a structural dependence between  1and  6 -aryl coordination, synthetic access to unsymmetrical calcium aryl complexes of the form, [( Dipp BDI)Ca(-Ar)(-Ar')Ca( Dipp BDI)], as well as the serendipitous isolation of a biaryl-coordinated calcium species and an assessment of the potential significance of variable aryl ligand hapticity for the future development of this chemistry.

Results and Discussion
Although the dimeric solid-state structures of compounds 2 and 3 were both propagated by Ca- 2 -C Ph -Ca interactions, the bridging phenyl substituents of the two compounds adopted contrast-ing orientations with respect to the calcium centers. [12]Whereas the solitary phenyl group of 2 was effectively orthogonal to the Ca-H-Ca-C Ph least squares plane (88.4°), the corresponding angle subtended with the plane defined by the Ca-C Ph -Ca-C Ph heterocycle in compound 3 was only 24.88°.This latter unusual feature was attributed by density functional theory (DFT) calculations to a stabilizing interaction between the ortho-CH bonds of the phenyl substituents and the coordinatively unsaturated calcium atoms.As an initial investigation of the potential generality of this ortho-CH interaction, therefore, compound 1 was reacted with half an equivalent of (o-tolyl) 2 Hg at room temperature in benzene (Scheme 2).In a manner reminiscent of that observed during the syntheses of compounds 2 and 3, this reaction resulted in an immediate effervescence of H 2 gas alongside the deposition of mercury metal.Two new sets of aryl environments could be identified in a 7:3 ratio in the resultant 1 H NMR spectrum, which exhibited aromatic resonances at  H 6. 79, 6.72, 6.60, 6.27 and 5.82,  5.65, 5.55 ppm, respectively (Figure S7, Supporting Information).The significantly lower frequency shifts and the apparent higher symmetry of the latter species, along with a corresponding 2H methylene signal observed at 1.90 ppm, allowed its tentative assignment as a benzylic calcium environment, which had apparently formed in competition with the initially envisaged o-tolyl derivative. [13]This supposition was subsequently confirmed by fractional crystallization of the reaction mixture and X-ray diffraction analysis of both the o-tolyl mixed hydride, [( Dipp BDI)Ca(H)(otolyl)Ca( Dipp BDI)] (4), and its benzylic isomer (5).
The introduction of a methyl group into the ortho-C position of compound 4 evidently disrupts any potential for the symmetrical C-H•••Ca interactions that were a defining feature of compound 3 in the solid state.The resultant structure (Figure 1a), thus, comprises two differentiated calcium environments, where Ca1 is coordinated by two N-Ca contacts [2.341(1) and 2.355(2) Å], a bridging hydride ligand [2.0825(5) Å] and an  6 -arene interaction to the o-tolyl moiety with a centroid to calcium distance of 2.5276(11) Å. Ca2 is also coordinated to two -diketiminate nitrogen atoms [2.353(2) Å] and a bridging hydride interaction [2.1927(6) Å], but displays a close contact with C32 [2.526(3) Å], identifiable as the ipso-carbon delivered by the original o-tolyl mercuric reagent.Although the calcium coordination environments of compound 5 are broadly comparable to those of 4 (Figure 1b), the  6 -arene to  Ca1 centroid distance of 2.6150(9) Å is significantly longer.This feature may be attributed to the Ca2-C30 -bond [2.6036 (18) Å] that provides the interaction between calcium and the now benzylic C30 carbon, but which is elongated in comparison to the Ca-C(sp 2 ) -aryl interaction observed in 4.9b,12-14] Although we believe complex 5 is formed via isomerization from compound 4, we have been unable to study this transformation experimentally due to the longer-term solution instability of both compounds.For this reason, a DFT study at the B3PW91-D3 level of theory was performed on the possible mechanistic pathway involved in this isomerization process.The optimized structures of 4 (4 opt ) and 5 (5 opt ) are shown in Figure S34 (Supporting Information).Calculated bond distances are listed in Table 1 and are compared with those experimentally observed.
The computed distances globally reproduce the values obtained experimentally, validating the level of the computational method employed.Variations higher than 0.04 Å between the experimental and computed bond distances were found for some Ca1-C( 6 -arene) bond lengths, providing Ca1-( 6 -arene) centroid distances of 2.593 Å versus 2.5276(11) Å for 4 opt and 2.765 Å versus 2.6150(9) Å for 5 opt .These disparities are ascribed to the presence of intermolecular short contacts within the crystal lattice of both 4 and 5.In order to evaluate the effect of the introduction of a methyl group into the ortho-C position of the arene ligand on the charges of the Ca centers, we computed the natural charges of 4 and 5 via an NBO analysis.As shown in Figure S34 (Supporting Information), the presence of two differentiated calcium environments in compounds 4 and 5 does not affect the charges of the two Ca atoms, which are practically identical in both compounds (+1.75 for Ca1 and Ca2 in 4 opt and +1.76 for Ca1 and Ca2 in 5 opt ).
In order to rationalize the formation of complex 5, we focused on the mechanism involved in the isomerization reaction between 4 and 5.As shown in the Gibbs free energy (enthalpy) profiles in Figure 2, two mechanistic pathways have been computed, displaying either a direct one-step H transfer within the o-tolyl ligand (black profile) or a three-step process involving the bridging Ca 2 (-H) hydride ligand (blue profile).The geometries of all the optimized structures in Figure 2 are shown in Figure S35 (Supporting Information), together with the most significant bond lengths and NPA charges.The formation of 5 is an exergonic process by 15.4 kcal mol −1 , indicating that the benzylic isomer is more stable than its o-tolyl derivative.Starting from compound 4, the direct one-step transfer of one H atom from the methyl group to the ipso-carbon of the o-tolyl ligand requires a transition state (TS1, ∆G = 51.0 kcal mol −1 ) that is unfeasibly high given the applied experimental conditions.Although this possibility may, thus, be discounted, a second computed pathway induced by the nucleophilic attack of the bridging Ca 2 (-H) hydride anion at the ipso-carbon of the o-tolyl anion provides Int2 with a C 7 H 8 2− dianion, which bridges two ( Dipp BDI)Ca + units (Figure 2, blue pathway).This reaction is endothermic by 5.5 kcal mol −1 but invokes a more accessible transition state (TS2) of 32.4 kcal mol −1 .9b,15] Our current deductions, therefore, indicate a wider generality of such processes and, in common with an analogous benzene dianion strontium derivative reported in the same study, [15] the structure of Int2 presents a slightly puckered toward the benzylic coordination mode of product 5 opt , with the formation of the benzylic C-Ca interaction providing the driving force of the reaction.
With compounds 4 and 5 in hand, and to assess the potential generality of the benzyl isomerization process, this synthetic protocol was extended to analogous meta-and para-phenyl methyl substitution.Accordingly, reactions of compound 1 with half an equivalent of (x-tolyl) 2 Hg (x = m-or p-) provided the familiar loss of H 2 via rapid bubbling of the reaction solution and the formation of a mercury precipitate (Scheme 3).Initial assessment of the m-tolyl-derived reaction by 1 H NMR spectroscopy evidenced the generation of two calcium aryl species, which were identified by the appearance of two new iso-propyl methine resonances at  H 3.16 and 2.96 ppm.Three singlet resonances were also observed across the Dipp BDI -methine and Ca--H region (ca.4-5 ppm), with the latter two signals integrating in a relative 1:2 ratio (Figures S12-S14, Supporting Information).Guided by the lack of discrimination apparent in our previous observations of the formation of compounds 2 and 3, we ascribed the two prod-ucts formed as a likely mixture of a calcium meta-tolyl mixed hydride ( 6) and its dinuclear arylcalcium analog, [( Dipp BDI)Ca(mtolyl)] 2 (7) (Scheme 3).In support of this assignment, treatment of the initial reaction mixture with a further 0.5 equivalents of (mtolyl) 2 Hg resulted in the complete disappearance of the hydridic resonance and afforded exclusive access to compound 7, which was characterized by iso-propyl methine and Dipp BDI -methine resonances at  H 2.96 and 4.86 ppm, respectively.
Two products were also observed in the reaction of 1 with half an equivalent of (p-tolyl) 2 Hg.In this instance, however, three iso-propyl methine resonances were observed in the resultant 1 H NMR spectrum alongside two differentiated p-tolyl aromatic environments.Although comprehensive spectroscopic assignment was impeded by overlap of the various signals in this NMR spectrum (Figure S19, Supporting Information), these observations implied that a similar mixture of products had again been formed, albeit with an apparent reduction in symmetry across the calcium aryl dimer.After treatment of the reaction mixture with an additional half an equivalent of (p-tolyl) 2 Hg, the resultant 1 H NMR spectrum presented only two of the initial iso-propyl methine signals at 2.91 and 3.09 ppm.Further Dipp BDI resonances ( H 4.86, 4.77 ppm), which integrated in a 1:1 ratio, in addition to a notable absence of any signal that could be assigned to a hydridic environment, supported a formulation of [( Dipp BDI)Ca(ptolyl)] 2 (8).
The formation of colorless single crystals from a cooled (−35 °C) saturated toluene solution of 8, allowed its structural confirmation by X-ray diffraction as the unsymmetrical dimeric structure inferred by NMR spectroscopy (Figure 3b).In contrast to the symmetrical structures of both 3 and 7, the C30-and C35-containing p-tolyl rings are differentiated by their respective orientations.While the C35-containing carbocycle is almost orthogonal to the Ca1-C30-Ca1 1 -C35 least squares plane (82.36°), the C30-containing ring is constrained to lie in a similar plane (30.39°) through the adoption of ortho-C-H•••Ca interactions familiar from the structures of 3 and 7.In this case, however, coordinative unsaturation of the calcium centers is further satisfied by close contacts to the C27/C27 1 methine C-H bonds of two N-Dipp iso-propyl substituents.
Although the structural characterization of compounds 2, 4 and 9 demonstrates the viability of such dinuclear hydrido-phenyl and hydrido-tolyl calcium complexes, their synthesis via the relevant mercuric aryl reagent was invariably accompanied by the corresponding wholly aryl-bridged compounds.Our study progressed, therefore, to the use of (3,5-t-Bu 2 C 6 H 3 ) 2 Hg, expecting that the incorporation of additional steric bulk on the distal side of the aromatic unit would enable an enhanced level of kinetic discrimination.Accordingly, a reaction of compound 1 with half an equivalent of (3,5-t-Bu 2 C 6 H 3 ) 2 Hg provoked the familiar evolution of H 2 gas as well as the slow deposition of mercury metal.Although this process required a longer reaction time of 72 hours to reach completion, assessment at this point by 1 H NMR spectroscopy revealed two characteristic resonances assigned to new Dipp BDI -methine and Ca--H environments at  H 4.76 and 4.58 ppm.These signals emerged in a mutual 2:1 ratio by relative integration, alongside a broad iso-propyl N-Dipp methine signal at 3.11 ppm and two discriminated (9H) tert-butyl resonances.(Figure S22, Supporting Information).These spectroscopic features and, particularly, the observation of an aryl ipso-carbon signal at 177.7 ppm in the corresponding 13 C{ 1 H} NMR spectrum (Figures S23-S25, Supporting Information) were strongly indicative of the exclusive formation of [( Dipp BDI)Ca(H)(3,5-t-Bu 2 C 6 H 3 )Ca( Dipp BDI)] (10, Scheme 4).This deduction was confirmed by a subsequent X-ray diffraction analysis (Figure 4b), which revealed a further dimeric -H--aryl-bridged structure.While the bond lengths and angles about the calcium centers are only marginally perturbed relative to those of the most directly comparable species described above, and the structure maintains the ortho-C-H⋅⋅⋅Ca interactions that provide a distinctive feature in the structures of compounds 3, 7 and 8, the aromatic ligand itself adopts a somewhat more skewed orientation (36.3°) with respect to the plane defined by Ca1, C59 and Ca2.
Attempts to react compound 10 with further equivalents of (3,5-t-Bu 2 C 6 H 3 ) 2 Hg provided no evidence of reaction, presumably as a consequence of the increased steric demands of the dialkylated aryl units.With complex 10 in hand, however, we turned our attention to its potential to act as a synthon for wholly aryl bridged, but asymmetric [( Dipp BDI)Ca(-Ar)(-Ar')Ca( Dipp BDI)] systems.An initial reaction between 10 and 0.5 equiv. of Ph 2 Hg, induced the customary evolution of H 2 and deposition of elemental mercury.Analysis of the resultant solution by 1 H NMR spectroscopy revealed significant asymmetry within the resultant compound (11), albeit only a single Dipp BDI -methine environment at  H 4.77 ppm could be discriminated.Moreover, the aromatic region exhibited characteristic peaks for both the phenyl ( H 6.91, 6.65, 6.54 ppm) and 3,5-t-Bu 2 C 6 H 3 ( H 7.77, 7.55 ppm) ligands, which integrated in 1:2:2 and 2:1 ratios, respectively.These data alongside the identification of two distinctive low field ipso-carbon signals ( C 177.8, 175.7 ppm, identified via HMBC, Figure S29, Supporting Information) in the corresponding 13 C{ 1 H} NMR spectrum, thus, strongly supported the formation of [( Dipp BDI)Ca(Ph)(3,5-t-Bu 2 C 6 H 3 )Ca( Dipp BDI)] (11, Scheme 4).This deduction was validated by single crystal X-ray diffraction, confirming the identity of 11 as a bis-aryl bridged, but asymmetric, calcium -diketiminate complex (Figure 5a).
of 26.6 kcal mol −1 and the formation of the resulting biphenyl dianion Ca product (13 opt ) is an equilibrium reaction (ΔG = 2.1 kcal/mol relative to 3 opt ).Taking into account that the presence of two coplanar phenyl units stabilizes compound 3 opt by 10.1 kcal mol −1 with respect to complex 3', our resultant kinetic barrier of 26.6 kcal mol −1 relative to 3 opt is, thus, perfectly concordant with the value of 17.4 kcal mol −1 relative to 3' opt obtained by the group of Harder.As shown in Figure S36 (Supporting Information), transition state TS4 is facilitated by a further isomer of 3 opt (3'' opt ), which comprises the two phenyl units aligned almost perpendicularly to the C Ph -Ca-C Ph -Ca plane.Complex 3'' opt is located at 10.3 kcal mol −1 , indicating that, while rotation of the first phenyl to form 3' opt is energetically costly (ΔG = 10.1 kcal mol −1 relative to 3 opt ), the rotation of the second phenyl constitutes an equilibrium process (see Figure S37, Supporting Information).On this basis, therefore, the formation of the biphenyl complex ( 13) may plausibly precede the formation of 14, with the computed kinetic barrier of 26.6 kcal mol −1 accounting for the experimentally observed slow reaction time (>7days).
At this point, we considered how the C-C coupling across such bis-phenyl calcium dimers may be facilitated by the establishment of polyhapto--interactions between each aromatic anion and the group 2 metal centers.Starting from 3 the achievement of the TS4 geometry (Figure S36, Supporting Information) requires an adjustment of the initial  1 -engagement of the phenyl anion with calcium to  6 .By analogy with our earlier rationalization of the C-C coupling across calcium acetylide dimers, [17] we propose that the resultant -interactions with Ca induce a polarization across the C 6 unit as a whole and decrease the mutual repulsion of the two carbanionic -carbon centers.
To further investigate this hypothesis, therefore, we carried out an Atoms in Molecules (AIM) analysis on species 3 opt , 3' opt , TS4 and TS2.As shown in Figure S38 (Supporting Information), the AIM analyses of compounds 3 opt and 3' opt revealed bond critical points (BCPs) located between each Ca atom and both ipso-C atoms of the two phenyl anions as well as a BCP located between the ipso-C atoms of the two bridging phenyl rings.For TS4, on the other hand, the AIM analysis located two BCPs between each Ca atom and one ipso-C of the phenyl anions.In corroboration of the presence of a polyhapto- interaction between the arene and the Ca (see Figure S39, Supporting Information), a BCP was also identified between the ipso-C atoms of the two bridging phenyl ligands along with two cage critical points associated with each bridging phenyl ring and one Ca center.
While BCPs were also observed between the Ca atoms and the ipso-C of the phenyl anion and the hydride ligand of TS2, the presence of a  interaction (see Figure S38, Supporting Information) enables a BCP between the ipso-C atom of the phenyl and hydride ligands and between the phenyl ring and one Ca center.Significantly, while no BCP was detected between the phenyl ligand perpendicular to the C Ph -Ca-C Ph -Ca plane and the Ca center for complex 3' opt , for both TS4 and TS2 the  6 interaction between the phenyl anions and the calcium atoms is evidenced by a cage critical point.This suggests, therefore, that the effective nucleophilic attack of a Ph − (or a H − ) at the second Ph − is likely to be assisted by a phenyl-Ca  interaction, which perturbs the otherwise excessively high mutual repulsion between the two negatively charged ipso-carbon (or ipso-carbon and hydride) atoms, consequently enabling the resultant C-C or C-H coupling reaction.

Conclusions
In summary, we have reported a series of mixed hydrido-aryl calcium complexes, where the disruption of the previously observed ortho-CH-to-calcium interactions affords a tendency toward polyhapto-coordination of the aromatic anion and isomerization to a benzyl species via an unusual "cross dimer" process involving hydride addition to the initially formed aryl substituent.In contrast, the introduction of moderate distal meta-and para-tolyl substitution leads to the maintenance of a more symmetric mode of Ca-C-Ca bridging, albeit with some variation in the relative orientation of the aryl substituents with respect to the Ca-Ca vector.Incorporation of the more sterically influential (3,5-t-Bu 2 C 6 H 5 ) substituent allows selective isolation of a mixed hydrido-arylbridged calcium dimer.In turn, the availability of this species facilitates the first examples of mixed -aryl--aryl' bridged dicalcium complexes.We have also observed the serendipitous formation of the tetra-calcium complex, [{{( Dipp BDI)Ca} 2 ( 2 -Cl)} 2 (C 6 H 5 -C 6 H 5 )], which comprises a biphenyl dianion.Although we have no solid experimental rationale for the mode of C-C bond formation, computational analysis suggests it is a consequence of the variable hapticity, which is a feature of the dimeric derivatives elaborated in this study.This work suggests, therefore, that such "cross dimer" polarization effects may provide the basis of a more generalizable route to both symmetric and unsymmetric biaryl units and related C-X bonded organic products.We are continuing to explore these possibilities.
The Supporting Information contains Experimental details, NMR spectra, X-ray crystallography, and computational details and atomic coordinates for the optimized geometries of the compounds.
Crystallographic data for all compounds have been deposited with the Cambridge Crystallographic Data Centre as supplementary publications CCDC 2260391-2260399 for 4, 5, 7, 8, 9, 10, 11, 12 and 14, respectively.These data are provided free of charge by the joint Cambridge Crystallographic Data Centre and Fachinformationszentrum Karlsruhe Access Structure sservice, www.ccdc.cam.ac.uk/structures.

Figure 2 .
Figure2.Gibbs free energy (enthalpy) profiles at the D3-B3PW91 level of theory (6-311++G** for Ca, 6-311G** N, O and 6-31G**) for the formation of complex 5 (5 opt ) starting from compound 4 (4 opt ).Black: Gibbs free energy (enthalpy) profile of the direct one-step isomerization, involving the transfer of one H atom from the methyl group to the ipso-carbon of the o-tolyl ligand.Blue: Gibbs free energy (enthalpy) profile of the three-step isomerization, involving i) the reaction of the bridging Ca 2 (-H) hydride ligand with the ipso-carbon of the bridging o-tolyl group, followed by ii) the transfer of one hydride from the methyl group to one Ca center and by iii) the isomerization of the bridging C 7 H 7 − anion toward the formation of the benzylic hydride product 5 opt .