Introduction

Since the first reports on the isolation of free singlet carbenes, stabilised by their incorporation in N-heterocycles1, N-heterocyclic carbenes (NHCs) have become a fundamental cornerstone in molecular chemistry and beyond2,3,4,5,6,7,8,9,10,11. Whilst in transition metal chemistry they are known to enhance reactivity at the metal1,2,3,4,5,6,7,8,9,10,11,12, in the field of main group chemistry they are typically used as an ambiphilic platform to stabilise low oxidation state and highly reactive compounds, acting as supporting ligands that do not directly participate in further reactivity2,3,4,5,6,7,8,9,10,11,12. In Group 14 chemistry, NHCs have been shown to allow for the isolation of the first dihalosilylenes, independently reported at the same time by Filippou, Roesky, Stalke and co-workers13,14, which serve as soluble and readily available precursors of silicon( + II), with plenty of derivatives having been synthesized in subsequent years (I, Fig. 1)2,3,4,5,6,7,8,9,10,11,12. This is just one representative example, illustrating the impact of NHCs on Group 14 chemistry. Since then, for silicon and germanium, significant progress has been made in stabilising low-coordinate species and even allowed for the development of open-shell species2,3,4,5,6,7,8,9,10,11,12.

Fig. 1: Notable literature examples and research scope.
figure 1

I: Landmark NHC-stabilised dihalosilylenes; II: first examples of NHC-ligated stannylenes.

Full size image

On the contrary, silicon and germanium’s heavier congener tin has received significantly less attention2,3,4,5,6,7,8,9,10,11,12. This can be attributed to poorer overlap between the NHC and the heavier tetrel elements, and consequently renders the formation of stable NHC–tin adducts more difficult compared to the lighter analogous15. Furthermore, the example of parent tetrylenes E(14)H2 exhibiting decreased Lewis acidity when descending Group 14 can be referenced, making the formation of stable NHC–stannylene complexes more difficult compared to their lighter congeners16.

Pioneering examples of NHC-stabilised stannylenes were reported by Kuhn et al. and Weidenbruch and co-workers, dating back to 199517,18. They synthesized the dichloro- and diaryl-substituted stannylenes Cl2Sn(IiPr2Me2) and Tipp2Sn(IiPr2Me2), respectively (IiPr2Me2 = 1,3-diisopropyl-4,5-dimethylimidazol-2-ylidene; Tipp = 2,4,6-iPr3-C6H2) (II, Fig. 1). This breakthrough opened the door to the isolation of numerous additional examples within this compound class, ranging from bis-hypermetallyl-substituted to transition metal-based derivatives19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44.

All of these tin-containing examples share the characteristic that the NHC acts as an inert stabilising spectator ligand, not participating in any subsequent chemistry. As with any rule, there are exceptions. In this case, a comparatively small number of examples with other main group elements and transition metals have indeed shown follow-up reactivity, involving either formal C–H, C–N, or C–C bond activations; and ring opening, ring expansion, or protonation pathways2,3,4,5,6,7,8,9,10,11,12.

Here, we report the effects of NHC-ligation to heteroleptic stannylenes, which induce enhanced reactivity motifs, where the NHC is not merely an innocent spectator ligand. This led to formal inter- and intramolecular C(sp3)–H activations at the backbone methyl groups of the chosen NHCs, leading to a diverse series of products. We show that these reactions are critically dependent on the vicinal steric hinderance of both the chosen NHC and the terphenyl ligands around the tin centre, leading to the selective formation of stannylene heterocycles and a series of macrocyclic multinuclear tin complexes.

Results

Synthesis and characterisation of 2

Expanding on our findings that the heteroleptic terphenyl-/amido-stannylenes of the type ArTerSn{N(SiMe3)2} (ArTer = MesTer = 2,6-(2,4,6-Me3C2H6)2C6H3 (1a); ArTer = DippTer = 2,6-(2,6-iPr2C6H3)2C6H3 (1b); ArTer = TippTer = 2,6-(2,4,6- iPr3C6H2)2C6H3 (1c)) can formally activate C(sp3)–H bonds through the elimination of the corresponding amine hexamethyldisilazane (HN(SiMe3)2)45,46, we sought to investigate how the introduction of strong Lewis bases, such as NHCs, might impact the observed activation pattern and, ultimately, whether this could trigger the activation of the NHC. We hypothesized that NHC ligation renders the tin centre more electron-rich, thereby further weakening the tin–nitrogen bond while sterically repelling the trimethylsilyl groups. This, in turn, renders theses complexes more reactive towards C–H bond scission with the leaving group HN(SiMe3)2, thereby enhancing reactivity at the tin–nitrogen bond (Fig. 1, bottom). Hence, we initiated our study by treating 1a with stoichiometric quantities of the NHC IMe4 (1,3,4,5-tetramethylimidazol-2-ylidene). A rapidly advancing reaction was observed in benzene at room temperature, as evidenced by 1H nuclear magnetic resonance (NMR) spectroscopy, indicating complete consumption of both starting materials (Fig. S10)47. This implied that the corresponding adduct MesTerSn(IMe4){N(SiMe3)2} (2a) had been formed quantitatively (Fig. 2A). To our initial surprise, the characteristic signal of free HN(SiMe3)2 (δ1H = 0.10 ppm in C6D6) was observed with its intensity slowly increasing over time (cf. Fig. S16). Therefore, and to obtain 2a in good isolated crystalline yields of 61%, the product was purified immediately by crystallisation of 2a from Et2O at -30 °C.

Fig. 2: Synthesis and structures of 2a-c, 3a, and 4b (polymorph a), and 4 c. A.
figure 2

Reactivity of 1a-c, towards IMe4 to yield 2a-c, the tetramer 3a, and the dimers 4b,c. Conditions (i) benzene, r.t.; (ii) benzene, r.t.–80 °C. B–D: Molecular structures of 2a, 3a, and 4a determined by single crystal X-ray crystallography. Anisotropic displacement parameters are drawn at the 50% probability level. Hydrogen atoms have been omitted where necessary for clarity, and in the case of 3a terphenyl substituents have also been omitted for clarity.

Full size image

Yellow block-shaped crystals obtained via this method were suitable for single crystal X-ray diffraction (SCXRD) and verify the formation of the NHC-stabilised stannylene 2a (Fig. 2C). 2a crystallises in the monoclinic space group P21/n with the tin atom being in a distorted trigonal pyramidal coordination environment (Σ∢Sn = 301.4°). The dihedral angle between the planes defined by N1–Sn1–C8 and N2–C1–N3 is approximately 70°. The Sn–CNHC bond length of 2.3255(9) Å falls within the expected range for stannylene–NHC adducts. For comparison, it is slightly elongated relative to other terphenyl-substituted stannylenes cf. the tin(II)-monohydride MesTerSn(IMe4)H (2.2804(11) Å)33, and slightly shorter than in the dichlorostannylene Cl2Sn(IPr) (2.341(8) Å; IPr = 1,3-bis(2,6-diisopropylphenyl)imidazol-2-ylidene)22. The Sn–Namido bond length (2.1990(8) Å) is elongated compared to 1a, as well as the only other structurally characterised amido-substituted, NHC-stabilised stannylene IPrSn(Cl)N(H)Dipp (2.1142(16) Å)26. As another structural feature, the IMe4 ligand and one mesityl group of the terphenyl moiety are aligned almost coplanar with a centroid–centroid distance of 3.48 Å, indicative of what is generally referred to as π–π-interactions (Fig. S37)48,49.

In solution, this interplay seems to be maintained as shown by observed separate signals for both mesityl groups (Fig. S11). In the 119Sn NMR spectrum of 2a a singlet signal with a chemical shift of δ119Sn = –15.7 ppm is observed (Fig. S14), thus being significantly shifted towards higher field when compared to the starting material 1a (δ119Sn = 1192.3 ppm; Fig. S1) and also when compared to other terphenyl- and amido-substituted NHC-stabilised stannylenes (IPrSn(Cl)N(H)Dipp (δ119Sn = –93.2 ppm); MesTerSn(IMe4)H (δ119Sn = –349.4 ppm))26,33.

Synthesis and characterisation of 3

Although the elimination of HN(SiMe3)2 already occurs at room temperature in solution, 2a is stable for a prolonged time in the solid state under inert conditions (at least four weeks at –30 °C and one week at room temperature), allowing for the investigation of its reactivity. Complete consumption of 2a accompanied by the release of HN(SiMe3)2 is realised by heating a solution of 2a to 80 °C for 3 h, which is accompanied by the formation of a suspension from the initially clear solution (Fig. S15).

While this manuscript was in its final stages, the Inoue group also reported the isolation and characterisation of 2a, observing that this compound “is unstable in solution and decomposes into an intractable mixture of products after four hours at room temperature”50. As expected, the analytical data obtained by both Inoue and co-workers, as well as by us, are in excellent agreement, and a detailed discussion is presented here to support its categorisation within the context of this work. Notably, they further report on the use of 2a as a hydrosilylation catalyst for aldehydes and ketones.

Upon slow evaporation of benzene solutions, SCXRD suitable colourless block-shaped crystals can be repeatedly obtained, revealing the formation of the tetranuclear, twenty-membered macrocycle 3a as a result of consecutive C(sp3)–H-activations of the NHC backbone methyl groups (Fig. 2A, B). 3a crystallises as a benzene solvate in the triclinic space group P\(\bar{1}\) and can be regarded as self-assembled tetramer of a MesTerSnCH2NHC building block. The shape of the macrocycle can be described as a “bowl”, with two opposite “NHC-walls” facing inwards and outwards to varying degrees (Fig. S39). The distances between the tin atoms allow conclusions to be drawn about the approximate size of the “bowl” and are 9.33 Å (Sn1•••Sn3) and 6.15 Å (Sn2•••Sn4). Each tin atom is coordinated to three carbon atoms in a trigonal pyramidal coordination environment with C–Sn–C angles between 89° and 105°. All carbon–tin distances are in a narrow window of on average 2.26 Å (Sn–CH2), 2.27 Å (Sn–CMesTer) and 2.28 Å (Sn–CNHC) and, accordingly, slightly shortened when compared to the starting material 2a (2.2809(9) Å (Sn–CMesTer), 2.3255(9) Å (Sn–CNHC)). Consequently, the NHC ligand in 2a undergoes transformation into a unique formal hybrid NHC / mesoionic N-heterocyclic olefin (mNHO) ligand in 3a, which serves to bridge the tin atoms51,52. mNHOs have recently emerged as exceptional ligands for main-group and transition-metal centres due to their outstanding donor properties and are typically generated through a methylation / deprotonation sequence, starting from abnormal NHCs51.

Having found this unusual route towards macrocyclic main group systems based on a MesTerSnCH2NHC building block generated upon amine release from an NHC-stabilised heteroleptic stannylene, we wanted to investigate how the choice of the ancillary ligands, namely the terphenyl substituent and the NHC, influence the macrocycle formation. Accordingly, DippTerSn{N(SiMe3)2} (1b) and TippTerSn{N(SiMe3)2} (1c) were reacted with IMe4 in benzene or toluene at room temperature, leading to the immediate formation of the NHC-adducts DippTerSn(IMe4){N(SiMe3)2} (2b) and DippTerSn(IMe4){N(SiMe3)2} (2c), as verified by multinuclear NMR spectroscopy (δ119Sn = -27.4 ppm (2b), -20.0 ppm (2c); Figs. S18 and S21) (Fig. 2A).

Synthesis and characterisation of 4

As observed for the analogous reaction with the MesTer derivative, the formation of HN(SiMe3)2 already starts straight away and after 1b,c were consumed completely, slow evaporation of benzene solutions led to the formation of colourless and yellow crystals, respectively, suitable for SCXRD, verifying the formation of the ten-membered macrocyclic dimers 4b and 4c of the ArTerSnCH2NHC building block (Fig. 2D and Fig. S30). In analogy to 3a, 4b,c both crystallise in in the triclinic space group P\(\bar{1}\) as benzene solvates with the tin centres being in trigonal pyramidal coordination environments. For 4b, crystals suitable for SCXRD were also successfully obtained without the co-crystallisation of any additional solvent (Fig. S24). Since the solvent-free crystal structure exhibits slight disorder, the structural parameters of the solvent-containing structure are discussed for clarity (both modifications show very good agreement in their structural parameters)47. The tin–carbon distances are further apart when compared to 3a. The shortest bond length is observed for Sn–CDippTer/TippTer (2.2420(9) Å (4b), 2.2444(11) Å (4c)) and the longest bond length is either observed for Sn–CH2 (2.3111(10) Å) in case of 4b or for Sn–CNHC (2.3031(15) Å) in case of 4c. The Sn•••Sn distances are 6.45 Å (4b) and 6.46 Å (4c), respectively, and as a structural feature, the two five-membered imidazolium moieties are not aligned coplanar but show a rather short distance to each other of approximately 3.40 Å (4b) and 3.47 Å (4c) by considering the centroids of the five-membered N-heterocycles.

It is worth mentioning that although 3a and 4b,c can be reproducibly obtained as single crystalline materials, sufficiently clean NMR spectra for the characterisation of both 3a and 4b,c in solution could not be obtained due to their poor solubility in various organic solvents (aliphatic and aromatic hydrocarbons, ethers etc.), especially when the crystalline materials were dried under vacuum. This is likely due to removal of the lattice benzene solvent observed during SCXRD upon drying in vacuo. Information on the bulk purity via elemental combustion analysis repeatedly led to good agreement of the theoretical and measured H and N values, with the C values consistently being too low47.

Notably, 4c demonstrated slightly increased solubility in organic solvents compared to all other herein reported macrocycles, allowing for the detection of its M+ signal by LIFDI mass spectrometry.

Interestingly, when 1a was reacted with IPr – an NHC lacking a backbone methyl group – at temperatures up to 100 °C, only the tin starting material decomposed, while the NHC remained unaffected.

To obtain further insight into the unique formation of 3a and 4b,c, and bonding situation therein, quantum chemical calculations were performed at the BP86-D3BJ/Def2-TZVP (PCM = benzene) level of theory, which includes corrections for both solvation and dispersion effects53,54,55,56,57,58. On barrierless coordination of IMe4, adducts 2a and 2b were found to form with a ΔG298 of −11.6 and −10.7 kcal•mol−1 respectively. Inspection of the Kohn-Sham HOMO indicates that the Sn(II) centre retains its lone pair character, owing to either zwitterionic or donor-acceptor natures of the CNHC–Sn interactions. This is comparable to our previously reported complexes MesTerSn(IMe4)CH2PNR (R = alkyl, aryl)46. Natural Bond Orbital (NBO) analysis confirms this59, providing a two electron NBO for the Sn lone pair. Like 1a, a nitrogen-based lone pair also contributes to the HOMO which allows for the amide to also take part in nucleophilic processes. The virtual orbitals, on the other hand, are different: 1a has a lower-lying LUMO (HOMO–LUMO gap = 2.344 eV), which exhibits p-orbital character at the tin centre, allowing for its donor-acceptor type reactivity. In 2a, the NHC has extinguished this orbital, leading to a larger HOMO–LUMO gap (2.542 eV) and a terphenyl-centred delocalised LUMO. Meanwhile, the LUMO + 1 is localised on the NHC in its archetypal pπ form.

We have previously demonstrated the reactivity of the Sn–N moiety in 1 towards C–H bonds, resulting in the production of bis(trimethylsilyl)amine and a tin alkyl complex, albeit with mildly activated C–H bonds45,46. Compounds 2, on the other hand, are able to activate aryl-substituted methyl groups, which are typically more difficult to activate. Natural Population Analysis (NPA) found a significant decrease in NPA charge (1.29 to 1.08) at the tin centre, showing higher electron density at the metal (i.e. increased basicity). This is reflected in the Wiberg bond index for the Sn–N interaction, which decreases from 0.45 to 0.38, indicating a reduction in bond strength and thus allowing for the more facile activation across the bond; and supports the observed elongation of the bond measured by SCXRD, vide supra. A model mechanism for the formation of 3a from 1a was calculated using Kohn-Sham Density Functional Theory (DFT), finding an overall Gibbs free energy, ΔG298, of −23.1 kcal•mol−1. This cascade reaction consists of four consecutive C–H activation steps with decreasing activation energy, ΔG‡298, between +18.2 and +5.7 kcal•mol-1 (Fig. 3), which explains why the reaction is facile at room temperature and low temperatures are required to isolate 2a without further activation.

Fig. 3: Computed ring expansion mechanism.
figure 3

DFT-calculated mechanism for the tin amide mediated C–H activation cascade forming 3a at the BP86-D3BJ/Def2-TZVP (PCM = Benzene) level of theory. Gibbs free energies (ΔG298, kcal•mol-1) are given relative to the starting materials. Natural Population Analysis (NPA) charges and Wiberg Bond Indices (WBI) given for 1a and 2a were calculated using NBO-7.

Full size image

The unobserved cyclic dimer, isostructural to 4b, was found to be thermodynamically less favourable than the final tetrameric product with a ΔG298 of −15.3 kcal•mol−147. On the contrary, the difference in energy between 4b and the hypothetical tetramer was found to be less than 1 kcal•mol−1, owing to its higher steric demand. To quantify how NHC coordination induces the observed increased reactivity, a hypothetical reaction of 1a and 2a was modelled by DFT, where 1a activates the C–H bond of the NHC in 2a, analogous of the first C(sp3)–H activation step shown in Fig. 3. This was found to proceed through a much higher energy transition state (ΔG‡298 = 37.4 kcal•mol−1), roughly double the height of the barrier when two equivalents of 2a react, vide supra, to an endergonic product (ΔG298 = 4.4 kcal•mol−1) (Fig. S55).

Synthesis and characterisation of 5

After observing and analysing the impact of the different terphenyl ligands on the reactivity of the respective heteroleptic stannylenes 1a-c towards IMe4, the reactivity of 1a-c towards the slightly larger NHC IiPr2Me2 (1,3-diisopropyl-4,5-dimethylimidazol-2-ylidene) was investigated. Upon equimolar addition of IiPr2Me2 to a solution of 1a in benzene at room temperature, no reaction between the two substrates can be observed by NMR spectroscopy. However, heating the reaction mixture to 60 °C leads to clean conversion to a single species, accompanied by the release of HN(SiMe3)2. By contrast to the formation of the hitherto reported macrocycles 3a and 4b,c, the newly obtained compound exhibits good solubility in both aliphatic and aromatic hydrocarbons, enabling its characterization through multinuclear NMR spectroscopy. Additionally, crystals suitable for SCXRD were obtained from concentrated solutions in n-hexane at -30 °C, revealing that, unlike the smaller carbene, IMe4, IiPr2Me2 induced an intramolecular C(sp3)–H bond scission at the benzylic position of the terphenyl ligand, yielding the carbene-stabilised stannylene heterocycle 5a (Fig. 4).

Fig. 4: Synthesis and structures of 5a, and 6b,c. A.
figure 4

Reactivity of MesTerSn{N(SiMe3)2} (1a), DippTerSn{N(SiMe3)2} (1b) and TippTerSn{N(SiMe3)2} (1c) towards IiPr2Me2 to yield the C–H activation product 5a, and doubly C–H activation products 6b,c. Conditions (i) benzene, r.t. –60 °C; (ii) benzene, r.t.–90 °C. B: DFT-Calculated mechanism for the tuck-in formation of 5a from 1a and IiPr2Me2 at the BP86-D3BJ/Def2-TZVP (PCM = Benzene) level of theory. Gibbs free energies (ΔG298, kcal•mol−1) are given relative to the starting materials. C,D Molecular structures of 5a and 6b determined by single crystal X-ray crystallography. Anisotropic displacement parameters are drawn at the 50% probability level. Hydrogen atoms have been omitted where necessary for clarity, and in the case of 6b terphenyl substituents have also been omitted for clarity.

Full size image

The monomeric NHC-ligated stannylene 5a crystallises in the triclinic space group P\(\bar{1}\). The tin atom is in a distorted pyramidal coordination environment (C1–Sn1–C12 98.96(7)°, C1–Sn1–C29 99.26(7)°, C12–Sn1–C29 80.83(7)°). The Sn–CNHC bond length of 2.316(2) Å is in good agreement to the other herein reported monomeric stannylene 2a (2.3255(9) Å). The tin atom is embedded in a six-membered all-carbon heterocycle, thus representing an NHC-stabilised tin-carbon heterocycle. In this context the Sn–C(sp2) and Sn–C(sp3) bond lengths are almost identical (2.2470(19) Å and 2.2474(18) Å, respectively) and are exceeding the single bond covalent radii by approximately 0.10 Å (Σcov(Sn–C) = 2.15 Å60,61). 5a exhibits a 119Sn NMR chemical shift of δ119Sn = –164.4 ppm which is in good agreement with literature examples, e.g. the NHC-stabilised diaryl substituted stannylene Tipp2Sn(IMe4) (δ119Sn{1H} = -160.7 ppm)31.

The intramolecular C(sp3)–H activation found in the formation of 5a was investigated using DFT (Fig. 4B). This showed that the initial coordination of the larger carbene ligand was stabilising by just ΔG298 = − 5.6 kcal•mol−1, half the thermodynamic impact than its smaller analogue. This produced optimised intermediate is sterically restrained, with an even weaker Sn–N bond (Wiberg bond index = 0.34) and in addition, one of the methyl groups from the terphenyl moiety in the vicinity of the activated Namido. Abstraction of the proton then proceeds with a ΔG‡298 of +22.5 kcal•mol−1, in line with requirement to heat the reaction, producing the amine and heterocycle 5a with an ΔG298 of −10.9 kcal•mol−1. Since stannylenes have been shown to be capable of oxidative addition62, we also explored the two-electron redox pathway computationally, where oxidative addition of the C–H bond results in a Sn(IV) alkyl-hydride intermediate. This was found to be energetically much less favourable than the reported redox-neutral pathway (ΔG‡298,OxAdd = +41.4 kcal•mol−1), producing the Sn(IV) species with an ΔG298 of +19.5 kcal•mol−1.

Synthesis and characterisation of 6

Finally, we investigated the reactions of ArTerSn{N(SiMe3)2} (Ar = Dipp (1b), Ar = Tipp (1c)) with IiPr2Me2, the mixtures with the largest vicinal steric hinderance around the metal (Fig. 4). Upon mixing 1b,c and the respective NHC in C6D6 at room temperature, no reactions could be observed within 16 hours. Reactivity was however enabled by heating to reflux, as verified by the detection of small amounts of HN(SiMe3)2 by 1H NMR spectroscopy (cf. Fig. S34). Further heating of the reaction mixtures caused the solutions to slowly turn cloudy, and two doublets in close proximity to each other (δ1H = 1.11-1.15 ppm) in addition to a new heptet signal (δ1H = 2.90 ppm) were observed in the 1H NMR spectrum, corresponding to free ArTerH (Figs. S26 and S27)63,64. 1b,c were completely consumed after several days of heating. Notably, significant amounts of IiPr2Me2 remained unreacted when stoichiometric amounts were used in initial attempts. Gratifyingly, these reaction conditions repeatedly led to the precipitation of yellow crystals suitable for SCXRD when starting from 1b,c, which could be separated by filtration. The bulk purity of these materials was confirmed by elemental combustion analysis of the corresponding yellow powders after drying under vacuum.

The newly obtained compound derived from 1b crystallises in two different crystal habits (block-shaped and needle-shaped crystals), depending on the varying equivalents of co-crystallized benzene molecules. Both forms are highly sensitive, decomposing within minutes even when coated in NVH oil. Due to nearly identical structural data, the parameters of the needle-shaped crystals are discussed because of the better quality of the data set. The SCXRD results perfectly match the observations made while monitoring the reaction by NMR spectroscopy and demonstrate the formation of the tetranuclear tin macrocycle 6b (Fig. 4D). Both polymorphs of 6b crystallise in in the triclinic space group P\(\bar{1}\). The release of both HN(SiMe3)2 and DippTerH, along with the remaining IiPr2Me2 observed during reaction monitoring of the first attempt (vide supra), is confirmed by the connectivity of the macrocycle. The flanking five-membered dimetallacycles in 6b adopt distorted tetrahedral geometries, as evidenced by the τ4 and τ4’ geometry indices of 0.79 and 0.72, respectively, owing to the inherent strain of the heterocycle65,66. This distortion is further indicated by the small Sn1–Sn2–N1 and Sn2–Sn1–C1 angles of 88.30(12)° and 92.59(11)°. The three-coordinate tin atom Sn1 is in a trigonal pyramidal coordination environment (N1–Sn2–Sn1 88.30(12)°, Sn2–Sn1–C1 92.59(14)°, N1–Sn2–C41 98.18(19)°). The Sn1–Sn2 bond length of 2.9055(9) Å exceeds the sum of covalent radii by approximately 0.10 Å (Σcov(Sn–Sn) = 2.80 Å60,61) and is also elongated compared to compounds featuring the same structure of the discussed five-membered ring (2.737(2) Å67 and 2.870(1) Å68). While these literature-known examples formally feature Sn(III)–Sn(III) and Sn(III)–Sn(II) moieties, 6b features a formal Sn(III)–Sn(I) bond. A slight elongation is also observed for the Sn–CNHC bond length of 2.335(6) Å in 6b, compared to the other tin macrocycles reported herein (cf. 2.2988(9) Å in 4a), likely due to the inherently different bonding situation at Sn2. The CH₂NHC units bridge the two five-membered ring systems, completing the tricyclic structure with a formal 12-membered central ring system. The structural data of the TippTer-substituted derivative 6c are, as expected, in very good agreement to those of 6b (Fig. S36).

To further understand both the formation and bonding situation of this unique tricyclic compound, quantum chemical investigations were conducted. The most sterically restricting combination, 1b and IiPr2Me2, is close to thermoneutral (ΔG298 = +0.4 kcal•mol−1) in the formation of the adduct (DippTerSn(IiPr2Me2){N(SiMe3)2}). Whilst a mechanism to the formation of the unique product 6b could not be elucidated, the loss in adduct favourability can be rationalised by the activation of other bonds (such as the loss of the terphenyl moiety). The NHC-induced ligand loss in mono(aryl)tin compounds is known in the literature to stabilise via Sn–Sn bond formation, towards cluster assembly30. From a structure and bonding perspective, 6b is dimeric, composed of two identical components, each containing two bonded Sn centres with very different bonding situations. The most interesting aspect is the Sn–Sn bond itself. NBO analysis describes two inequivalent Sn centres, Sn1 and Sn2: Sn1 is tetracoordinate with an NPA charge of +1.09 (similar to compounds 2), while Sn2 is tricoordinate with an NPA charge almost half of that at +0.64, indicating an extremely electron rich centre. Sn2 also possesses a lone pair of electrons primarily with s character (1.93 e−, s0.89p0.11), in line with 2. The Sn–Sn bond is characterised by two NBOs totalling 1.92 electrons, both polarised towards the tetracoordinate Sn centre Sn1 (64.3%). In this interaction, Sn1 uses a hybrid orbital of s0.35p0.65 composition, whereas Sn2 uses primarily p character (s0.04p0.96). As observed in compounds 2, the NHC substituent is in an imidazolium resonance form, indicating a zwitterionic relationship with Sn2. Based on their NPA charges and connectivity, one could assign the oxidation states of these tin centres as Sn(III) and Sn(I), which can be compared to the structure of the literature-known stannyl-stannylene [[1,2-C6H4{CHP(BH3)Cy2}2]Sn]269.

Discussion

In summary, a synthetic protocol was developed to access a range of well-defined tin macrocycles and a cyclic NHC-stabilised stannylene. The outcomes of the reactions were highly dependent on both the NHC employed and the substitution pattern of the kinetically stabilising terphenyl ligands in the heteroleptic stannylenes used. This revealed a unique non-innocent behaviour of NHCs within the coordination sphere of stannylenes and their transformation into hybrid NHC / mNHO ligands, which bridge the tin centres, emerges as a predominantly observed motif. We are currently envisioning and working on the use of these macrocyclic systems in template-controlled sensing and capturing of appropriate building blocks.

Methods

Detailed descriptions of experimental, spectroscopic, crystallographic and quantum chemical methods and results are given in the Supplementary Information. The authors have cited additional references in the Supplementary Information70,71,72,73,74,75,76,77,78,79.

General

All manipulations of air- and moisture-sensitive materials were carried out using standard Schlenk-line and glovebox techniques (MBraun glovebox with oxygen and water concentrations below 0.1 ppm as monitored by an O2/H2O Combi-Analyzer) under an inert atmosphere of argon.

NMR spectroscopy

NMR spectra were measured in benzene-d6 (C6D6) or toluene-d8 (C7D8) (dried over CaH2, distilled by trap-to-trap transfer in vacuo, degassed by three freeze-pump-thaw cycles and transferred to the glovebox). NMR samples were prepared under argon in NMR tubes with J. Young Teflon valves. NMR spectra were measured on Bruker Avance 400 MHz, 500 MHz and 600 MHz spectrometers. 1H and 13C NMR spectra were referenced internally to residual protio-solvent (1H) or solvent (13C) resonances (C6D6: dH = 7.16 ppm; dC = 128.06 ppm; C7D8: 2.08 ppm; dC = 20.43 ppm). 119Sn NMR spectra were referenced with respect to SnMe4.

Mass spectrometry

LIFDI-(JEOL AccuTOF JMS-T100GCV; inert conditions) and ESI- (Bruker Daltronik micro TOF) MS were measured by the Zentrale Massenabteilung (Fakultät für Chemie, Georg-August-Universität Göttingen).

Elemental analysis

Elemental analyses were obtained from the Analytische Labor (Georg-August-Universität Göttingen) using an Elementar Vario EL 3 analyzer.

X-ray crystallography

Suitable crystals were selected and mounted on a MiTeGen micromount with NVH oil. Single crystal X-ray data were collected from shock-cooled single crystals at 100.00 K on a Bruker D8 VENTURE diffractometer equipped with an Oxford Cryostream 800 low temperature device. The used radiation sources are Incoatec IµS 2.0 or 3.0 microfocus sealed X-ray tubes using mirror optics as monochromators using MoKα radiation (λ = 0.71073 Å) and Bruker PHOTON III detectors.

Computational methods

Geometry optimisations, frequency calculations and PCM solvent corrections were run with Gaussian 16 Revision A.03 using the BP86 functional. For geometry optimisations, all atoms were described with def2-SVP basis sets of Ahlrichs and Weigand. Single point energy calculations were performed on the optimised geometries, at the BP86/def2-TZVP level of theory. Stationary points were fully characterised using analytical frequency calculations as either minima (all positive eigenvalues) or transition states (one negative eigenvalue). IRC calculations and subsequent geometry optimisations were used to confirm the minima linked by the transition states. Energies reported in the text are based on the gas-phase free energies and incorporate a correction for dispersion effects using Grimme’s D3 parameter set with Becke-Johnson dampening (i.e. BP86-D3BJ) as well as solvation (PCM approach) in benzene. Energies are given in atomic units (a.u.) unless otherwise stated. Natural Bond Orbital (NBO) and Natural Localised Molecular Orbital (NLMO) analysis was performed using NBO-7 using single point calculations performed at the BP86/def2-SVP or BP86/def2-TZVP level of theory.

Online content

Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/XX.XXXX/XXXXXX-XXX-XXXXX-X.