Donor Acceptor Complexes between the Chalcogen Fluorides SF₂, SeF₂, SeF₄ and TeF₄ and an N‐Heterocyclic Carbene

Abstract

The majority of binary chalcogen fluorides are fiercely reactive and extremely difficult to handle. Here, we show that access to crystalline donor‐acceptor complexes between chalcogen difluorides (sulfur, selenium) and tetrafluorides (selenium, tellurium) with the N‐heterocyclic carbene (NHC) 1,3‐bis(2,6‐diisopropylphenyl)imidazol‐2‐ylidene (IPr) is possible conveniently and safely without the need to generate the highly unstable EF₂ (E=S, Se) or the very toxic and corrosive SeF₄.

Donor‐acceptor complexes between the carbene 1,3‐bis(2,6‐diisopropylphenyl)imidazol‐2‐ylidene (IPr) and EF₂ (E=S, Se) or EF₄ (E=Se, Te) were conveniently synthesized. The compounds were structurally investigated by single crystal X‐ray diffraction and a suite of density functional theory (DFT) methods were employed in order to elucidate fine electronic details of the chemical environment around the central chalcogen atoms.

Amongst the many known chalcogen fluorides, only the tetrafluorides EF₄ and hexafluorides EF₆ (E=S, Se, Te) are readily accessible and structurally characterized. Electron diffraction studies of SF₄, ^[1] SeF₄ ^[2] and TeF₄ ^[3] reveal that they adopt pseudo trigonal bipyramidal structures with C_2v symmetry in the gas phase. These so‐called ‘seesaw’ structures are consistent with the valence shell electron pair repulsion (VESPR) model and the results of quantum chemical calculations. ^[4] Owing to their inherent Lewis acidity, the same species are aggregated in condensed phase. The solid‐state structures of SF₄, ^[5] SeF₄ ^[6] and TeF₄[ ⁶ , ⁷ ] established by single crystal X‐ray diffraction show an increasing degree of aggregation of the molecular entities by bridging fluorine atoms, which correlates well with increasing atomic number. In nonpolar solvents, the aggregation is retained in concentrated solutions, whereas monomers exist in high dilution. According to ¹⁹F NMR studies, the pseudo trigonal bipyramidal structures of these SF₄, ^[8] SeF₄ ^[9] and TeF₄ ^[10] monomers are fluxional with the axial and equatorial fluorine atoms being in rapid exchange at room temperature on the NMR scale, which has been attributed to Berry pseudorotation. At around −100 °C, the exchange can be slowed down for SF₄ and SeF₄, but not for TeF₄. In polar solvents or the presence of donor molecules, Lewis pair complexes are formed, some of which have been isolated and fully characterized. For SF₄, these include complexes with triethylamine, ^[11] pyridine, 4‐methylpyrdine, DMAP, ^[12] the mono‐protonated bipyridinium ion, ^[13] THF, cyclobutanone, DME ^[14] DABCO and urotropine. ^[15] For TeF₄, these include THF, toluene[ ¹⁰ , ¹⁶ ] dioxane, DME ^[16] and OPR₃ (R=Me, Ph). ^[17] In earlier work, it was concluded that TeF₄ in certain donor solvents would undergo electrolytic dissociation into [TeF₃(solvent)_n]⁺ cations and the [TeF₅]^– anion, however, this was never unambiguously proven. ^[18] For SeF₄, no donor acceptor complexes have been structurally characterised yet.[ ¹⁹ , ²⁰ ] In recent years, the bond situation in donor acceptor complexes of EF₄ (E=S, Se, Te) with ammonia, amines and pyridines was investigated by quantum chemical calculations.[ ²¹ , ²² , ²³ ] One of these studies also addressed the stability of complexes between EF₆ (E=S, Se, Te) with ammonia, which was predicted to be very weak. The reaction of TeF₆ with trimethylamine proceeded with reduction at tellurium and gave rise to the formation of [Me₂NCH₂NMe₃][TeF₅]. ^[24] The inherently unstable difluorides SF₂ ^[25] and SeF₂ ^[26] were generated photochemically and characterized in an argon matrix, ^[27] whereas TeF₂ is still unknown. For bulk SF₂, an equilibrium was established with the mixed valent FSSF₃, which, however, irreversibly converts into SSF₂ and SF₄.[ ²⁸ , ²⁹ ] N‐heterocyclic carbenes (NHCs) are indispensable tools for the stabilization of low‐valent and Lewis acidic main group compounds. ^[30] Surprisingly, there is only one claim of a donor acceptor complex between an NHC and a chalcogen fluoride, namely, (I ⁱ Pr₂Me₂)SF₂, which, however, was not structurally characterized (see below). ^[31] The lack of such donor acceptor complexes is even more surprising as related complexes involving higher chalcogen halides, such as (NHC)EX₂ (E=S, Se, Te; X=Cl, Br, I), are abundantly known[ ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ ] and have found applications in organic synthesis.[ ³⁸ , ³⁹ ]

In the present work we show that N‐heterocyclic carbene‐stabilized complexes of SF₂, SeF₂, SeF₄ and TeF₄ can be obtained as crystalline solids by the oxidation of the imidazol‐2‐thione, imidazol‐2‐selenone and imidazol‐2‐tellurone derivatives using xenon difluoride. Thus, the reaction of IPrE (E=S (1S), ^[40] Se (1Se) ^[41] and Te (1Te), ^[42] IPr=1,3‐bis(2,6‐diisopropylphenyl)imidazol‐2‐ylidene) with one or two equivalents of xenon difluoride afforded the chalcogen(II) fluoride complexes IPrSF₂ (2S, Figure 1) and IPrSeF₂ (2Se) as well as the chalcogen(IV) fluoride complexes IPrSeF₄ (3Se, Figure 2) and [IPr₂TeF₃][TeF₅] (3Te, Figure 3). The latter eventually converted over the course of ca. 6 days into the mesoionic complex aIPrTeF₄ (4Te, aIPr=1,3‐bis(2,6‐diisopropylphenyl)imidazol‐4‐ylidene) following a normal to abnormal coordination switch (Figure 3). ^[30] This transformation is accelerated at higher temperatures and can reach completion at 40 °C within 3 days. Attempts to isolate a tellurium(II) fluoride complex, IPrTeF₂, by reacting 1Te with one equivalent of XeF₂ were unsuccessful. Regardless of the stoichiometric ratio between the IPrTe and XeF₂, 3Te was in every case the major product initially forming, in the presence of several other minor side‐products. The reaction of free carbene IPr with one equivalent of TeF₄ gave systematically 3Te in a better purity. The opposite was true in the reaction between IPr and SeF₄ which gave 3Se with a significantly lower purity and yield compared with the oxidation reaction. While investigating methods of purification for 3Se we isolated a few crystals that proved to be [IPrF][SeF₅] (4Se). We determined that this species does not form by the reaction of 2Se and excess XeF₂ but can be rationally obtained by heating 3Se at 80 °C in THF. A stable sulphur tetrafluoride complex, IPrSF₄, could not be obtained by further oxidation of 2S with XeF₂ or by reaction of IPr with SF₄. In solution, all complexes were extremely sensitive to adventitious traces of water and their further purification from traces of [IPrH]⁺ salts or IPrE proved to be frustratingly difficult despite arduous efforts. We speculate also that at least some of these complexes might be able to react to some extent with the borosilicate glass. In solid state however, the compounds seemed to be stable when stored under argon atmosphere at room temperature for at least several weeks.

Figure 1.

Synthesis of 2S and 2Se and molecular structures of 2S and 2Se showing 50 % probability ellipsoids and the numbering scheme of selected atoms. Selected bond lengths [Å] and angles [°] for 2S: S1−C1 1.7298(16), S2−C4 1.7321(15), S1−F1 1.8261(16), S1−F2 1.8188(14), F1−S1−F2 170.98(7), F1−S1−C1 86.05(7), F2−S1−C1 85.40(7); for 2Se: Se1−C1 1.8877(13), Se1−F1 1.9588(9), Se1−F2 1.9122(9), F2−Se1 F1 168.83(4), C1−Se1−F1 82.89(5), C1−Se1−F2 85.99(5).

Figure 2.

Synthesis of 3Se and its decomposition to 4Se and molecular structures of 3Se and 4Se showing 50 % probability ellipsoids and the numbering scheme of selected atoms. Selected bond lengths [Å] and angles [°] for 3Se: Se1−F1 1.8281(7), Se1−F2 1.8675(7), Se1−C1 1.9668(15), F1−Se1−F2 168.92(3), F1−Se1−F2 168.92(3), F1−Se1−C1 85.62(4), F1−Se1−F1a 89.92(5); 4Se: Se1−F5 1.8123(10), Se1−F2 1.7137(10), Se1−F3 1.8346(10), Se1−F6 1.8307(11), Se1−F4 1.8186(13), F5−Se1−F3 167.93(5), F5−Se1−F6 89.91(5), F2−Se1−F5 84.71(5), F2−Se1−F3 83.25(5), F2−Se1−F4 83.18(7), F4−Se1−F3 89.93(6), F4−Se1−F6 166.51(7).

Figure 3.

Synthesis of 3Te and 4Te and molecular structures of 3Te and 4Te⋅THF showing 50 % probability ellipsoids and the numbering scheme of selected atoms. Selected bond lengths [Å] and angles [°] for 3Te: Te1−C1 2.2955(11), Te1−C1a 2.2883(11), Te1−F2 1.9726(10), Te1−F2a 1.9694(10), Te1−F1a 1.8616(14), C1a−Te1−C1 158.331(12), F2−Te1−F2a 154.750(14), F2−Te1−C1 89.03(4), F2a−Te1−C1 86.18(4), F1a−Te1−F2 103.25(12), F1a−Te1−F2a 77.14(7); for 4Te⋅THF: Te1−F3 1.9875(8), Te1−F1 1.9568(9), Te1−F4 1.9758(9), Te1−F2 1.9786(10), Te1−C2 2.1169(12), F3−Te1 C2 80.60(4), F1−Te1−F3 165.35(4), F1−Te1−F4 88.50(4), F1−Te1−F2 89.18(5), F1−Te1−C2 84.79(4), F4−Te1−F3 88.43(4), F4−Te1−F2 163.95(4), F4−Te1−C2 82.06(4), F2−Te1−F3 89.82(5), F2−Te1−C2 81.92(4).

The ¹⁹F NMR spectra of 2S, 2Se and 3Se displayed singlet resonance signals at δ=−97.1, −157.8 and 3.6 ppm, respectively, signals that were accompanied, in the case of 2Se and 3Se, by ⁷⁷Se satellites (1136 and 255 Hz, respectively) matching the coupling constants observed for the multiplets assigned in the ⁷⁷Se NMR spectra to these species: a triplet resonance at δ=1003.5 ppm for 2Se and a quintet at δ=884.7 ppm for 3Se. Remarkably, the ¹⁹F resonance reported for (I ⁱ Pr₂Me₂)SF₂ (37.7 ppm) ^[31] is at a significantly higher chemical shift compared to 2S. Because of this marked difference and because we became motivated to determine the molecular structure of (I ⁱ Pr₂Me₂)SF₂, which was not reported in the original study, we decided to reinvestigate this compound. Following closely the original procedure of reacting (I ⁱ Pr₂Me₂)SCl₂ with AgF, ^[31] we were, unfortunately, unable to isolate (I ⁱ Pr₂Me₂)SF₂. Although we observed a very weak signal around 37.9 ppm by ¹⁹F NMR, the ¹H NMR spectrum showed the major species to be (I ⁱ Pr₂Me₂)S. Attempting to oxidize (I ⁱ Pr₂Me₂)S with XeF₂ failed as well to give the desired complex, leading us to assume that (I ⁱ Pr₂Me₂)SF₂ might be too unstable to be isolated. The formation of 3Te was confirmed by ¹⁹F NMR spectra which displayed a triplet resonance assigned to the apical fluorine atom (δ=−24.1 ppm) and a doublet resonance signal (δ=−91.5 ppm) for the remaining two basal fluorine atoms of [IPr₂TeF₃]⁺, as well as resonances assigned to the [TeF₅]⁻ counter ion (quintet at δ=−31.8 ppm, apical F; doublet −37.5 ppm, basal F), thus confirming the retention of the solid state structure of 3Te in solution. The ¹²⁵Te NMR spectrum showed multiplet resonance signals corresponding to the cation (doublet of triplets at δ=1078.9 ppm) and the anion (doublet of quintets at δ=1144.4 ppm) with ¹⁹F−¹²⁵Te coupling constants matching the ¹²⁵Te satellites observed for the assigned ¹⁹F resonances. A slow conversion of 3Te into the abnormal complex 4Te (characterized by a singlet resonance signal at −47.0 ppm in the ¹⁹F spectrum and a quintet at 1202.5 ppm in the ¹²⁵Te NMR) was observed. The observation of coupling between ¹⁹F−¹²⁵Te (239 Hz) indicates that the molecular structure of 4Te does not undergo conformational change in solution (at room temperature) as observed previously for R₃POTeF₄ (R=Me, Ph) and remains similar to the solid‐state structure. ^[17]

Single crystal X‐ray diffraction investigations revealed that in solid state 2S, 2Se, 3Se, 3Te and 4Te are distinct monomers. ^[43] In 2S and 2Se the coordination geometry around the chalcogen is a distorted T‐shape, while 3Se, 3Te and 4Te feature distorted square pyramidal geometry arrangements around the chalcogen. The carbene occupies the apical positions in 3Se and 4Te. In the cationic 3Te the carbene ligands coordinate trans with respect to each other in the basal plane while the apical position is occupied by a fluoride. The averaged S−F (1.819(1) Å) and Se−F (1.935(1) Å) bond distances in 2S and 2Se are significantly larger than the values determined by microwave spectroscopy for SF₂ in the gas phase (1.589 Å) ^[44] and SeF₂ in an argon matrix (1.725 Å). ^[27] The F−E−F angles in both 2S (average 171.50(7)°) and 2Se (average 168.84(4)°) deviate from the ideal 180° value as a result of electrostatic repulsion exerted by the lone pairs at the chalcogen. The average C−E bond lengths of 1.731(2) Å (2S) and 1.889(1) Å (2Se) are longer than the ^IPrC=S (1.670(3) Å) and the ^IPrC=Se bonds (1.827(6) Å) ^[45] and are closer to the values observed generally for single C−S (e. g., 1.763(1) Å) ^[46] or C−Se (e. g., 1.927(3) Å) ^[47] bonds in diaryl sulfides and selenides. The Se(IV) complex 3Se features shorter Se−F bond lengths (1.8675(7) and 1.8281(7) Å) and a more elongated C−Se bond length (1.967(2) Å) compared with 2Se but retains a deviation of the diagonal F−Se−F bond angles (168.92(3)°). The Se−F bond lengths observed in 3Se are comparable with those observed for the Se−F bonds contained in the basal plane of the [SeF₅]⁻ anion of 4Se, the thermal decomposition product of 3Se, which has a similar coordination geometry around selenium. This anion features, however, a shortened bond distance between the apical fluoride and selenium (1.714(1) Å). In the cation of 3Te, two different Te−F bond lengths can be observed depending on whether the F atom occupies the apical (1.862(1) Å) or basal (1.973(1) Å) positions and are similar to those observed in the R₃POTeF₄ (R=Me, Ph) complexes. ^[17] The C−Te bonds (2.288(1) Å) are much more elongated compared to 1Te (2.055(3) Å). ^[45] The [TeF₅]⁻ counter anion, ^[24] which has the Te atom in a square pyramidal coordination geometry, was affected by significant positional disorder. The four fluorides in 4Te are disposed in the basal plane. The Te−F bond lengths (average 1.973(1) Å) are comparable to those seen in the cation of 3Te for the basal Te−F bonds, but the C−Te bond is significantly shorter (2.117(1) Å). The diagonal F−Te−F bonds deviate slightly from collinearity (165.35(4)° and 163.95(4)°).

The bonding within the donor acceptor complexes IPrEF₂, IPrEF₄, [IPr₂EF₃]⁺ and aIPrEF₄ (E=S, Se, Te) was analysed by detailed density functional theory (DFT) calculations, ^[48] energetically by energy decomposition analyses (EDA) ^[49] and other energy descriptors derived from DFT, and electronically by a set of real‐space bonding indicators (RSBI). The overall or instantaneous interaction energy (E_int) between the NHC and chalcogen fluoride fragments ranges from 330–520 kJ mol⁻¹. In all cases, the estimated reorganization energy (E_re; _ca. 190–270 kJ mol⁻¹) is larger than the estimated bond dissociation energy (E_d; ca. 70–260 kJ mol⁻¹), pointing towards a high energy expense upon complex formation caused by the need of structural reorganization of the chalcogen fluoride fragments. The E_int value is higher for S containing complexes than for their Se or Te analogues, which is most likely due to stronger primary S−C bonds, especially reflected in larger orbital attraction (E_orb) values, pointing out bond covalency (Table S5). In S−C bonds, covalent bonding aspects dominate (slightly) over non‐covalent bonding aspects, i. e., the electrostatic attraction (E_elstat), thus E_orb>E_elstat, whereas the opposite is observed for the Se−C and Te−C bonds (E_orb<E_elstat). This is supported by RSBI analysis (see below). Despite higher values for Pauli repulsion for EF₂ than for EF₄ containing complexes, E_int is nevertheless higher for the former as a consequence of higher electrostatic attraction and significantly higher orbital attraction. Including dispersion correction similarly increases Pauli repulsion (E_Pauli), E_elstat, E_orb and consequently also E_int, as the two increased attractive forces overcompensate the increased Pauli repulsion. Notably, E_int and reorganization energies (E_re) decrease for all compounds in the order S>Se>Te (Table S5 and S6), but the fragment sum (E_fs) behaves different, as it shows no clear trend for EF₂ complexes but increases significantly for EF₄ complexes according to S<Se<Te. For 2Se and 3Se, the primary E−F and E−C bonds as well as numerous secondary F⋅⋅⋅C and F⋅⋅⋅H/Cπ interactions are reflected in the AIM (Figure 4a and e) ^[50] bond topologies and the NCI iso‐surfaces (Figure 4b and f). ^[51] The highly polar E−F bonds are characterized by electron density (ED, ρ(r)) values at the bond critical point (bcp) of 0.8–1.5 eÅ⁻³ for S, 0.9–1.3 eÅ⁻³ for Se, and 0.6–1.0 eÅ⁻³ for Te (Table S7). With few exceptions, bond ellipticities (ϵ) are below 0.15, indicating directed bonds with low electron smearing. Particularly high kinetic energy density over ED ratios (G/ρ(r)) larger than 1 point towards dominant ionic bond contributions, whereas equally high (negative) values for the total energy density over ED ratio (H/ρ(r)) indicate dominant covalent bond contributions. For the S−F bonds, G/ρ(r) and H/ρ(r) are typically large and similar, indicating the high relevance of both bonding aspects and resulting in a Laplacian of the ED (∇²ρ(r)) close to zero. For the Se−F and Te−F bonds, non‐covalent bonding aspects start to dominate over covalent bonding aspects (G/ρ(r) > |H/ρ(r)|, clearly positive ∇²ρ(r) value). Accordingly, only for the strong and short S−F contacts corresponding S−F bonding basins are formed in the ELI−D as consequence of covalent bond aspects, although being very small (V_ELI=0.1–0.7 ų) and little populated (N_ELI=0.1–0.8 e). The less polar S−C and Se−C bonds are dominated by covalent bonding aspects (G/ρ(r) < |H/ρ(r)|, clearly negative ∇²ρ(r) value), whereas both bonding aspects are balanced in the Te−C bonds. Complex formation transforms the lone‐pair basin of the carbene C atom into a bonding E−C basin. In this course, the basin volumes (V_ELI) shrinks from 15.1 Å to 3.7–9.5 ų and the localizabilities from 2.53 to 1.83–2.02, indicating increased electron sharing. This is further reflected in the Raub‐Jansen index (RJI), ^[52] which is a measure for bond polarity as it quantifies the relative electron populations within an ELI−D basin being located in adjacent (typically bonding) AIM atomic basins. The RJI approaches 50 % for homopolar covalent bonds and becomes larger than 90 % for high polar dative and ionic bonds. The intermediate regime is occupied by polar‐covalent interactions, such as the E−C bonds (Table S8). Particularly low RJI values are obtained for regular and abnormal complexes, IPrEF₄ and aIPrEF₄. The electron populations (N_ELI), however, are little affected, ranging from 2.2–2.6 in the complexes compared to 2.44 in the free carbene. Notably, they are decreasing for the IPrEF₂ and abnormal complexes aIPrEF₄, but increasing for IPrEF₄ and [IPr₂EF₃]⁺. The secondary F⋅⋅⋅H/C_π interactions in all complexes are weak non‐covalent forces, which is reflected in low ED values typically below 0.1 eÅ⁻³, G/ρ(r) being larger than 0.75 a.u. and H/ρ(r) being positive. In NCI, extended greenish to bluish areas are observed (Figure 4b and f). ^[51] Most E⋅⋅⋅H/C_π bond paths are curved and the bond ellipticities can become as large as 4.4, which is typical for such weak interactions, and results in AIM bond path lengths (d1+d2, d1 and d2 are the distance from atom 1 or 2 to the bcp) larger than the geometric atom‐atom distance. Although each of these interactions are apparently weak, the sum of them significantly contributes to the stabilization energy between the carbene and the EF₄ fragment. For 2Se and 3Se, iso‐surface representations of the ELI−D are displayed in Figure 4c and g. ^[53] For clarity, the non‐bonding lone‐pair basins (LP(E)) and bonding E−C basins are highlighted in solid green in the complexes. For EF₂, adduct formation has a considerable effect on the ELI−D LP basin volumes, which are decreased by 30–50 % (Table S9). This is accompanied by lower electron localizabilities at the attractor position of the LP basins by about 0.3, indicating more pronounced electron sharing with the environment. This is, however, not accompanied by considerable changes in the basin populations, which are between 2.2 and 2.3 e in all cases. To the contrary, EF₄ shows a somewhat unpredictable behaviour: only slightly decreased LP basin volumes in the regular carbene complexes, but drastically enlarged volumes in the abnormal complexes aIPrEF₄, electron localizabilities even increased in some cases, but basin populations varying only to a small extent. With a grid‐step size of 0.05 bohr, these results should not be caused by an unsuitable integration routine, but are most likely due to different structural motifs, especially in the abnormal complexes aIPrEF₄, whereas the two F−S−F planes are about 45° to the central ring plane in the regular carbene structure, one F−Se−F or F−Te−F plane is almost coplanar (the other one almost perpendicular) in the abnormal carbene complexes. The ELI−D E−C bonding basins of the EF₂ adducts show strongly enhanced electron localizabilities in direction of the LP basins (bluish areas), indicating electronic interactions between these two basin types (Figure 4d). In 2S, the E−C and LP basins are still fused at an iso‐value of 1.3 (Figure S22c). The pear‐shaped E−C bonding basins of the EF₄ adducts, however, suggest only weak electronic non‐localized interactions to the S−F bonding or LP(F) basins, visible as ring‐shaped greenish to bluish gradients (Figure 4h and Figures S24–S28). The cationic complexes [IPr₂EF₃]⁺ show both basin interaction types, as they carry three F atoms as well as one lone‐pair in the central EF₃(LP) plane (Figures S29d–S31d).

Figure 4.

RSBI analysis of 2Se and 3Se respectively: (a) and (e) AIM bond paths motif, (b) and (f) NCI iso‐surface at s(r)=0.5, (c) and (g) ELI−D localization domain representation at iso‐value of 1.3, (d) and (h) ELI−D distribution mapped on the E−C ELI−D bonding basin.

In summary, the carbene complexes IPrSF₂ (2S) and IPrSeF₂ (2Se) and IPrSeF₄ (3Se) were obtained as crystalline materials without the need to generate the highly unstable and toxic parent chalcogen fluorides (IPr=1,3‐bis(2,6‐diisopropylphenyl)imidazol‐2‐ylidene). Extending this chemistry to tellurium led us to isolate a metastable ionic complex, [IPr₂TeF₃][TeF₅] (3Te), which slowly converted into the abnormal NHC complex aIPrTeF₄ (4Te). Attempts to obtain analogous NHC complexes of highly reactive SF₄ or (the unknown) TeF₂ (i. e., IPrSF₄ and IPrTeF₂) in a similar way were unsuccessful. The thermodynamic stability of the isolated complexes results from covalent and electrostatic contributions alike.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Dedicated to Professor Ionel Haiduc on the occasion of his 85th birthday

P. Komorr, M. Olaru, E. Hupf, S. Mebs, J. Beckmann, Chem. Eur. J. 2022, 28, e202201023.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.