Alkyne insertion at a nickel–aluminum heterometallic complex: implications for pyridine alkenylation

Abstract

Cooperative Ni–Lewis acid catalysis has emerged as an effective strategy for the selective C–H functionalisation of heteroarenes, yet the elementary steps following C–H activation remain poorly understood. In particular, direct experimental insight into the migratory insertion step, which governs C–C bond formation and potentially product regioselectivity, has been lacking. Herein, we demonstrate that a heterobimetallic {Ni–Al} complex derived from the C–H activation of a pyridine undergoes efficient and selective insertion of alkynes under mild conditions. This reaction provides direct access to structurally well-defined intermediates relevant to cooperative pyridine functionalisation catalysis. A broad range of internal alkynes undergo rapid insertion to furnish five-membered {Ni–Al} metallacycles featuring alkenyl fragments that are σ-bound to aluminum and π-bound to nickel. Kinetic and mechanistic studies establish that alkyne insertion is non-reversible and that both steric and electronic effects govern the rate of insertion and the coordination mode of the bridging pyridyl ligand. Oxidatively-induced reductive elimination affords a corresponding vinylpyridine product, establishing the competence of these species in C–C bond formation. These findings provide the first direct experimental characterisation of migratory insertion intermediates relevant to cooperative {Ni–Al}-catalysed pyridine functionalisation and define mechanistic principles governing substrate coordination, insertion, and product formation.

Cooperative Ni–Lewis acid catalysis has emerged as an effective strategy for the selective C–H functionalisation of heteroarenes, yet the elementary steps following C–H activation remain poorly understood.

Introduction

Aromatic N-heterocycles are ubiquitous structural motifs across the pharmaceutical and functional materials industries – methods for the selective functionalisation of these cores is an important objective for synthetic chemistry.¹ Transition-metal-mediated C–H activation has emerged as a powerful strategy for the modification of N-heterocyclic scaffolds.² However, many established systems rely on second- and third-row transition metals, whose limited abundance and high cost can restrict broad applicability.³ In contrast, achieving comparable reactivity and selectivity using first-row transition metals remains challenging due to their distinct electronic structures and in some cases, less predictable reactivity via single-electron processes.⁴ Cooperative transition metal/Lewis acid catalysis has emerged as a powerful strategy for the selective functionalisation of pyridines and related heteroaromatics (Scheme 1a).^5–7

Scheme 1. (a) {Ni–Al} co-catalysis for pyridine upgrading reactivity; (b) probing the insertion step. Important positions are numbered and described throughout the text.

Seminal reports by Nakao and co-workers have demonstrated {Ni–Al} and {Ni–Zn} complexes as catalysts for the alkenylation of pyridines.⁸ Substantial advances have been made in catalyst design, enabling selective functionalisation at multiple positions of the pyridine framework through modulation of the transition metal, Lewis acidic metal, and supporting ligand environment (Scheme 1).^6,7,9 Subsequent developments have expanded the scope of these transformations beyond alkenylation to include alkylation, hydrocarbamoylation, and hydroheteroarylation reactions using both nickel- and cobalt-based catalyst systems.^10–13 These studies have established cooperative catalysis as a general platform for overcoming the intrinsic challenges associated with the activation and selective functionalisation of heteroaromatic substrates with a first row metal.

Mechanistic investigations accompanying these developments have focused predominantly on understanding the nature of the C–H activation event and the origins of site-selectivity. Experimental and computational studies have provided substantial evidence supporting ligand involvement in the cleavage of the C–H bond. While C–H activation in these systems is commonly formulated through an oxidative-addition pathway, subsequent computational analysis of the Nakao–Hiyama pyridine alkenylation system suggested that ligand-to-ligand hydrogen transfer (LLHT) mechanism also provides a plausible pathway for C–H cleavage.¹⁴ Kinetic isotope effect measurements, isotopic labelling experiments, stoichiometric studies, and density functional theory calculations have been used extensively to establish whether C–H activation is reversible and whether it contributes to the turnover limiting step.^7,12,13 In several cobalt-catalysed systems, the C–H activation has been characterised as a non-reversible and kinetically important process based on the observed primary kinetic isotope effect and computational studies, consistent with C–H cleavage contributing to the turnover limiting step.^12,13 Conversely, studies of nickel-based catalysts for pyridine functionalisation have also revealed cases in which C–H activation proceeds reversibly and exhibits negligible kinetic isotope effect, indicating that it is neither turnover limiting nor selectivity-determining.⁵

For pyridine alkenylation, considerably less attention has been devoted to understanding the subsequent migratory insertion step. Migratory insertion is universally invoked as a key elementary process that links C–H activation to C–C bond formation, through reductive elimination, and product generation. While computational studies frequently include insertion within the overall catalytic pathway, few reports directly examine the factors governing insertion reactivity, regioselectivity, or reversibility and the roles that metal sites within heterometallic complexes may play within this elementary step. Experimental studies specifically designed to probe insertion remain, to our knowledge, absent. Moreover, there are no well-characterised or isolated heterometallic intermediates that inform understanding of the insertion step.

This knowledge gap becomes particularly important in catalytic systems where C–H activation has been demonstrated to be reversible. Under such conditions, the position at which C–H cleavage initially occurs need not necessarily determine the final regiochemistry. Instead, selectivity may be established during one or more downstream elementary steps, including migratory insertion or reductive elimination. The present work seeks to address this knowledge gap and interrogates the migratory insertion step in a system directly relevant to pyridine alkenylation.

Here, we report reactivity of a heterobimetallic {Ni–Al} complex, 1 derived from the C–H activation of a pyridine, with alkynes and probe subsequent C(sp²)–C(sp²) bond-forming chemistry of the resulting insertion products. Reaction of 1 with a series of internal alkynes affords well-defined Ni-alkenyl complexes in which the newly formed alkenyl fragment is π-bound to nickel and σ-bound to aluminum (Scheme 2). Evidence supports the reversible coordination of the alkyne to the nickel centre of the heterometallic complex prior to migratory insertion. The coordination environment is responsive to the electronics of the substrate, with C-to-N linkage-isomerisation of the pyridyl group occurring on modification of the alkyne substrate. These insertion reactions provide direct structural representation of intermediates commonly proposed in cooperative {Ni–Al} catalysed heteroarene functionalisation and reveal hidden facets and complexity of a migratory insertion step potentially relevant to catalysis.

Scheme 2. Preparation of alkyne-inserted products 2–5, Dipp = 2,6-diisopropylphenyl.

Results and discussion

Alkyne coordination and insertion

We recently demonstrated that a heterometallic {Ni–Al} complex (Scheme 2) was capable of promoting selective C(sp²)–H activation at the α-position of aromatic N-heterocycles, including pyridine (Py) and 4-dimethylaminopyridine (DMAP), releasing H₂ to afford a µ-(Py) and a µ-(DMAP) (1) complex.¹⁵ In this example, nickel permits non-reversible C–H bond oxidative addition, with the Lewis-acidic aluminum center coordinating to the nitrogen atom of the heterocycle, anchoring the substrate and positioning the adjacent C–H bond for selective activation. Related stoichiometric C(sp²)–H activation of pyridines have been carried out independently by Ozerov,^16–18 Lu,¹⁹ and Nakao²⁰ using heterometallic complexes, though insertion reactivity of the resulting C–H activated product was not explored. Examples of d/p-block cooperative systems that undergo insertion chemistry are also precedented.^21–26

To probe the insertion propensity of 1, a series of internal alkynes commonly used in pyridine alkenylation,^7,8,12 including 2-butyne, 3-hexyne, 4-octyne, and diphenylacetylene (DPA), were surveyed. Mixing of 1 with ≥ 2 equiv. of each alkyne substrate in toluene at room temperature provided clean conversion of 1 to {Ni–Al} heterometallic complexes 2–5 featuring both π-coordinated and π-inserted alkyne fragments (Scheme 2). We suggest that the alkyne adducts 2 to 5 are zwitterionic, comprising an anionic Ni(0) and a cationic Al(iii) center, whose oxidationhave been computationally described and are unchanged from 1.¹⁵ All adducts were obtained as a single Ni–C-bound isomer, with the exception of the products from 3-hexyne and 4-octyne, which produced an inseparable mixture of two Ni–C- and Ni–N-bound regioisomers with a ratio of 7.5 : 2.5 for 3a : 3b and 7 : 3 for 4a : 4b. The reactivity of 1 was also briefly examined using terminal alkynes. In these cases, reactions with phenylacetylene or trimethylsilylacetylene produced intractable mixtures of isomers and attempts to isolate or characterize the products were unsuccessful.

The above reactions of 1 were readily monitored by disappearance of the hydride ¹H NMR resonance at δ_H = −3.51 ppm, together with the evolution of resonances for free 1,5-cyclooctadiene (COD). All alkyne adducts have lower symmetry relative to complex 1. While 1 exhibits reflection symmetry across the plane of the DMAP ring (C_s-symmetry), complexes 2–5 are C₁-symmetric. For the alkyne adducts 2–5, the most characteristic feature in their ¹H NMR spectra is the vinyl proton signal with chemical shift values of δ_H = 3.76 ppm for 2; 3.68 ppm for 3a; 3.57 ppm for 3b, 3.67 ppm for 4a; 3.54 ppm for 4b, and 4.22 ppm for 5. Using ¹H–¹³C HSQC spectroscopy, the corresponding vinyl carbon (C⁶) signals were unambiguously assigned at δ_C = 80.0 ppm for 2; 94.4 ppm for 3a; 101.8 ppm for 3b; 92.7 ppm for 4a; 99.7 ppm for 4b and 82.1 ppm for 5. Ni–C⁵ coordination was also confirmed by ¹H–¹³C HMBC NMR spectroscopy for 1, C⁵ has a ¹³C chemical shift of δ_C = 214.64 ppm whereas complexes 2, 3a, 4a, and 5 exhibit C⁵ resonances at comparable chemical shifts (δ_C = 222.60 ppm for 2; 222.18 ppm for 3a; 222.06 ppm for 4a; 212.86 ppm for 5). Complexes 3b and 4b, however, exhibit a C⁵ resonance at δ_C = 183.99 and 184.10 ppm, respectively. Analysis by ¹H–¹H NOESY NMR spectroscopy confirmed 3a/3b and 4a/4b as regioisomers which differ in the coordination mode of the pyridyl group to Ni. 3a/4a display a clear C¹(H)–BDI(CH) contact (BDI = β-diketiminate), whereas 3b/4b exhibit a C⁴(H)–BDI(CH) contact (Fig. S26). These observations suggest that 3a/4a feature C coordination and 3b/4b N coordination of the pyridyl group to nickel. Single-point energy calculations at the BP86-D3(BJ)/def2-TZVP/SMD(benzene) level were performed on optimized geometries derived from the solid-state geometries of 4a and 5 as starting structures. The corresponding N–Ni linkage isomers were generated by exchanging the positions of the pyridyl nitrogen and dimethylamino-groups, and their energies were compared with those of the experimentally observed structures. 4a was calculated to be 2.1 kcal mol⁻¹ lower in energy than the corresponding 4b structure, while 5 was 6.9 kcal mol⁻¹ lower in energy than the corresponding hypothetical N–Ni linkage isomer. The larger energetic preference calculated for 5 is consistent with the absence of detectable linkage isomerism for this complex. This metal–ligand rearrangement may have important implications for reactivity of {Ni–Al} heterobimetallic catalysts, particularly with respect to product release with the observed reorganization likely influencing the mechanism for subsequent elimination steps.

The reaction progress for the formation of 2, 3a/3b, 4a/4b and 5 was monitored by ¹H NMR spectroscopy at 298 K. The time required for full conversion follows a trend of alkyne substrate sterics: conversion of 1 to 2 completes within 15 min, whereas formation of 3a/3b and 4a/4b requires approximately 25 and 40 min respectively, and formation of 5 proceeds much more slowly (4.5 h). All reactions exhibit first-order kinetics in 1, giving rate constants (k_obs) of 2.3 × 10⁻³ s⁻¹ for 2-butyne; 2.1 × 10⁻³ s⁻¹ for 3-hexyne; 1.2 × 10⁻³ s⁻¹ for 4-octyne and 4.6 × 10⁻⁴ s⁻¹ for DPA, establishing a clear inverse correlation between reaction rate and alkyne steric bulk (Fig. 1a). Closer examination reveals that the C–Ni/N–Ni ratio evolves non-monotonically over time for both 3a/3b and 4a/4b, with the C–Ni isomer forming rapidly to a transient maximum before decreasing, concomitant with growth of the N–Ni isomer to its final distribution with an approximately sigmoidal time profile (Fig. 1b). This behaviour is consistent with initial formation of the C–Ni isomer followed by post-insertion equilibration to the N–Ni isomer, rather than parallel independent formation pathways (Fig. 1c). Mechanistically, we consider that the rearrangement may proceed through ionisation of the Ni–C bond followed by rotation of the pyridyl group and N-to-C isomerisation at aluminum. A related process was calculated to occur in the C–H activation of a pyridine with a {Fe–Al} heterometallic complex.²⁷ Observations also suggest that regioselectivity is governed not only by alkyne insertion, but also by subsequent dynamic reorganization within the {Ni–Al} framework.

Fig. 1. The kinetics of alkyne adducting reactions. (a) Normalized first-order kinetics for consumption of 1 at 298 K; (b) reaction monitoring of 3-hexyne adducting; (c) 3a/3b isomeric ratio over time.

Notably, during the initial stages of forming 5, two separate hydride resonances were observed. One corresponds to unreacted 1 (δ_H = −3.51 ppm), while the second resonance at δ_H = −3.90 ppm is tentatively assigned to an intermediate species (5–Int), which is proposed to contain only 1 equiv. of DPA π-coordinated to the nickel center (Scheme 2). 5–Int is consumed following an additional 90 min of reaction time (Fig. S94). A similar intermediate (4–Int) was observed during the formation of 4a/4b, where an additional hydride signal at δ_H = −3.12 ppm appears at early reaction times. Notably, the signal of 4–Int decays rapidly, disappearing within approximately 10 min following addition of 4-octyne to 1, which is significantly faster than the analogous process observed for 4 (Fig. S94). Similar intermediate hydride resonances were also briefly observed for 2–Int (δ_H = −3.16 ppm) and 3–Int (δ_H = −3.12 ppm). This observation suggests that alkyne binding and subsequent insertion occur in a stepwise manner, with the second binding event being kinetically hindered and its rate strongly dependent on substrate steric effects.

To conduct a more detailed study on the transiently produced hydride species, bis(trimethylsilyl)acetylene (bTMSA) was employed as a more sterically demanding substrate, with the aim of further disfavouring the insertion event (Scheme 3). In agreement with this, the reaction between 1 and excess bTMSA proceeds sluggishly at room temperature, with a substantial amount (40%) of 1 remaining even after 7 days. Under these conditions, 6 is formed, which retains a hydride resonance at δ_H = −2.57 ppm, along with signals corresponding to free 1,5-cyclooctadiene (COD). These observations are consistent with ligand substitution process, in which bTMSA coordinates to the nickel center displacing COD without insertion or hydride consumption. The rate of formation of 6 can be significantly enhanced by heating 1 with excess bTMSA (2 equiv.) at 80 °C, reaching full conversion within 18 h. Analysis by ¹H–¹³C HMBC NMR spectroscopy confirms that Ni–C⁵ coordination is retained in 5, with the C⁵ resonance observed at δ_C = 205.15 ppm.

Scheme 3. Reactivity of the {Ni–Al} system 1 in competition experiments.

Complex 6 represents an arrested intermediate in the insertion step and is consistent with the formulation of n–Int (Scheme 2). It appears bTMSA suppresses migratory insertion and allows for observation of the system at a stage preceding Al–C bond formation. Complex 6 additionally offers a well-defined model for interrogating the factors that govern the transition from simple π-coordination to insertion and highlights how subtle changes in alkyne sterics can divert the reaction pathway entirely. Despite its steric bulk, complex 6 is not altogether inert toward alkyne insertion. Treatment of 6 with 2 equiv. of a different alkyne, 4-octyne or DPA, leads exclusively to a 7 : 3 mixture of 4a:4b or 5, respectively (Scheme 3). These observations are consistent with the lability of bTMSA, with its steric encumbrance inhibiting insertion though still allowing displacement by other (less hindered) alkynes.

Competition experiments

To further probe the influence of alkyne electronics on the migratory insertion step, competition experiments were conducted. 1 was treated with a mixture of 2 equiv. of DPA and 4-octyne, resulting in exclusive formation of the DPA-derived insertion product 5 (Scheme 3). This is attributed to the stronger π-accepting capacity of DPA.²⁸ To assess the possible reversibility of alkyne insertion, a mixture of 4a and 4b was also treated with 2 equiv. of DPA. This reaction resulted in displacement of 1 equiv. of 4-octyne and formation of a single regioisomeric product containing the inserted 4-octyne fragment derived from the starting material, together with a π-coordinated DPA ligand. Notably, although 4-octyne dissociation from nickel readily occurs, the inserted 4-octyne fragment remains intact, indicating that the insertion step does not readily reverse under the reaction conditions tested. Moreover, the identity of the externally added π-bound alkyne (i.e., DPA versus 4-octyne) dictates the observed N–Ni versus C–Ni regiochemistry preference, giving the DPA π-substituted product (7) with >99% C–Ni coordination. This demonstrates that alkyne coordination can induce a regioisomeric switch independent of the insertion event (Scheme 3).

X-ray crystallographic data

Single crystals suitable for scXRD analysis of complexes 3a, 4a, 5, 6, and 7 were obtained by layering saturated toluene solutions with n-pentane followed by storage at −35 °C for 3–5 days (Fig. 2). Structural analysis of these complexes revealed all contained a five-membered {Ni–C–N–Al–E} (E = C for 3a, 4a, 5, and 7, E = H for 6) metallacycle. For complexes 3a, 4a, 6 and 7, this metallacycle is nearly planar. In contrast, complex 5 displays a notable distortion, with the nickel atom displaced by 0.892 Å from the least-squares plane defined by the remaining ring atoms. A similar trend is observed for the orientation of the DMAP ring relative to the metallacycle: these fragments are nearly coplanar in 6 which explains the chemical equivalency of its ligand 2,6-diisopropylphenyl protons in the solution-state due to C_s-symmetry. Similar coplanarity between DMAP and the metallacycle was also observed for 4a and 7, whereas in 5 they exhibit significant structural distortion with a dihedral angle of 19.1°.

Fig. 2. Solid-state structures of 4a, 5, 6, and 7 and the atomic numbering of the core structure (ellipsoids drawn at 40% probability; hydrogen atoms have been omitted for clarity). Inset shows the core structures of 4a, 5, 6, and 7 with measured bond lengths (Å) of the core metallacycles show in red.

Examination of the alkyne-derived fragments reveals that in 4a, 5, and 7, two substrate molecules are associated with each complex. One alkyne remains π-coordinated to the Ni center, while the second undergoes insertion between Ni and Al, interacting with Ni through its vinylic π-bond. A typical C C bond length is approximately 1.20 Å; however, in the three complexes examined, the π-coordinated alkynes display significantly elongated bond lengths: 1.283(5) Å in 4a, 1.265(3) Å in 5, and 1.296(2) Å in 7 (Fig. 2).²⁹ Additionally, the coordinated alkynes exhibit pronounced bending from the typical 180° geometry of alkynes, with bond angles ranging from 142.0(3)° to 152.8(2)°. These distortions suggest greater sp²-character and strong M(dπ)-to-alkyne(π*) interactions. Interestingly, the alkyne bending in 6 is asymmetric (≈9° difference), which is rationalized by the fact that one TMS group of the C C bond is closer to the DMAP ring, while the other is closer to the significantly less hindered µ-H group.

The inserted alkyne fragments, which form the π-bound vinyl moieties, also have metrics that deviate from those of canonical alkenes. Whereas a typical C C bond length is approximately 1.34 Å, the C C bonds in 4a and 5 are substantially elongated to 1.399(4) Å and 1.426(3) Å compared to this typical value, respectively.²⁹ Furthermore, significant pyramidalisation of the alkene carbons is observed, with calculated pyramidalisation angle (φ) of C⁶ and C⁷ being 31.5° & 18.7° for 4a and 27.6° & 8.8° for 5. These structural distortions are consistent with π-coordination of the alkene to the nickel center, which effectively lowers the bond order and perturbs the alkene's ideal planar geometry.

Oxidatively-induced reductive elimination

The proximity of the inserted alkene and pyridyl group in 2–5 provides a useful platform from which to consider onward bond-forming reactivity. In principle, coupling of C⁷ and C⁵ would furnish a functionalised 2-alkenylpyridine product akin to those shown in Scheme 1, closing a putative pyridine alkenylation cycle using 1. Indeed, this step (C–C bond reductive elimination) has been shown to be rate-determining by Huang and Ye for the catalytic coupling of pyridines and 1,3-dienes.⁵ In keeping with this, initial efforts to promote direct C–C bond formation either thermally or photochemically from 2–5 were not met with success. Complexes 2–5 may be described as formally containing an Al(iii) center and, in the absence of significant back-bonding, a Ni(0) fragment. This electronic situation renders a direct reductive elimination process intrinsically challenging. Consistent with this, the complexes display notable thermal robustness. As an example, 5 was subjected to heating at 100 °C, 14% of it remained intact even after 24 h, and no productive C–C bond formation was observed under these conditions. Thus, an oxidatively induced pathway was considered for promoting product release. Specifically, oxidation of the Ni(0) center to Ni(ii) using an oxidant was expected to favor reductive elimination, while regenerating a low-oxidation state nickel species (Scheme 4).³⁰

Scheme 4. Oxidatively-induced reductive elimination from 1 to give 8 in 95% yield.

Exposure of 5 to excess I₂ resulted in immediate precipitation of black solid. Following basic aqueous workup. Direct C–C bond formation reaction was achieved stoichiometrically under such condition: the α-alkenylation product 8 was obtained in 95% spectroscopic yield as evidenced by ¹H NMR spectroscopy and high-resolution mass spectrometry (Scheme 4). The Al center is considered redox-innocent with its oxidation state unchanged over this transformation, as evidenced by the formation of the diiodo-aluminum(III) species {(^DippBDI)Al(I)₂}, which was identified by ¹H NMR spectroscopic data. Formation of free DPA was also observed by ¹H NMR spectroscopy, indicating a reaction pathway involving release of the π-bound alkyne fragment. Hence, a plausible mechanism for this reaction involves initial iodination of the aluminum fragment of 5 to generate {(^DippBDI)Al(i)₂}, followed by formation of an intermediate nickel species bearing π-bound DPA and σ-bound DMAP and alkenyl fragments. Subsequent reductive elimination would then produce free DPA and Ni(0). Collectively, these observations indicate that oxidation-induced product release is a viable strategy for C(sp²)–C(sp²) coupling in this system, enabling productive release of the organic fragment. Although the present system is stoichiometric and does not constitute a catalytic cycle, the observed reactivity provides valuable mechanistic insight into the insertion and product-release steps relevant to related catalytic systems, in which an external oxidant would not be required for turnover. These findings therefore provide useful guidance for the design and mechanistic understanding of future catalytic {Ni–Al} systems.

Conclusion

Implications for pyridine alkenylation

Our findings have important implications for pyridine alkenylation, particularly for highly active heterobimetallic {Ni–Al} systems in which closely related intermediates have frequently been proposed but never directly observed (Scheme 5). Complex 1 is derived from the C–H activation of a pyridine, and contains both pyridyl and hydride ligands, and hence represents a well-defined platform to study the intrinsic coordination and insertion chemistry of alkynes in unprecedented detail. The present data demonstrate that the aluminum centre is not merely a spectator ligand but plays an active role in controlling reaction outcome. Following pyridine C–H activation, alkyne coordination likely occurs reversibly at the nickel centre of a {Ni–Al} pyridyl complex. Subsequent migratory insertion into the three-centre two-electron Ni–H–Al bridge proceeds to generate an alkenyl fragment that is π-bound to nickel and σ-bound to aluminum. This step does not reverse readily under the conditions explored and affords insertion products containing both pyridyl and alkenyl fragments poised for C–C bond formation. Notably, the coordination mode of the bridging pyridyl ligand is sensitive to the electronic properties of the substrate bound at nickel, with increased π-acceptor ability promoting C-to-N isomerisation of the pyridyl fragment. This dynamic behaviour may have important implications for the subsequent reductive elimination step by generating multiple reactive isomers with distinct activation barriers for C–C bond formation. The isolation and structural characterisation of these alkenyl insertion products, together with their oxidative conversion to pyridine alkenylation products upon treatment with I₂, provides direct experimental support for elementary steps that have long been postulated in cooperative heterobimetallic pyridine alkenylation catalysis.

Scheme 5. Summary of findings presented in this work.

Author contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: experimental procedures, spectroscopic data, computational, and crystallographic details; X-ray crystallographic data (cif); and computational coordinates (zip, xyz). See DOI: https://doi.org/10.1039/d6sc05792f.

CCDC 2572710–2572714 contain the supplementary crystallographic data for this paper.^31a–e