Isolable C‐Phosphonio‐Boratavinyl Anion Authors: Huihui Xu, Aurora Rodríguez‐Álvarez, Md Jabed Hossain, Antoine Baceiredo, Eddy Maerten, Nathalie Saffon‐Merceron, Vicenç Branchadell, Tsuyoshi Kato Published: 2026 ABSTRACT ABSTRACTA persistent C‐phosphonio boratavinyl anion 1 has been successfully synthesized. This boron‐containing vinyl‐lithium 1, featuring a vinyl carbon with a negative charge (−1.59), considerably higher than that of carbon analogue 7 (−0.76), exhibits strongly nucleophilic and basic properties, as demonstrated by the results of rapid reactions with H2 (10 bar, 15 min or 1 bar, 2 h at RT) and C6F6 (1 equiv., 15 min at −80°C). The reactions of boratavinyl anion 1 with electrophiles offer a convenient route for the introduction of a borata‐alkene‐based bulky and σ,π‐double‐donating substituent, enabling to efficiently stabilize reactive intermediates such as halogenogermylenes. Furthermore, the reaction with CO2 takes place at the borata‐alkene moiety, rather than at the P‐ylide function, resulting in the double insertion of CO2 into the B─C bond to afford a unique cyclic P‐ylide 13.Keywords: alkylidene phosphoranes, borataalkenes, germylenes, vinyl lithium, ylides Introduction Olefins consist of a C═C double bond function (I) whose transformation reactions and applications in various fields have long been the subject of research (Figure 1). In parallel, isoelectronic analogues of olefins, with one or two carbon atoms replaced by other elements, have also been extensively investigated for their analogical utility, but also for their original electronic properties and chemical behavior [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18]. Among them, borata‐alkenes (II), which are boryl‐stabilized carbanions, are a unique class of alkene analogues with a B═C double bond polarized toward the carbon atom [19]. Due to their anionic character and the strong bond polarization, they present a nucleophilic character at the carbon atom [19] and a particular hybrid coordination pattern between olefins and donor–acceptor ligand [20, 21]. Furthermore, the synthetic usefulness of borata‐alkenes II has also been demonstrated by their “boron‐Wittig olefination” of carbonyl compounds [22]. Nevertheless, unlike olefins, whose reactive anionic derivatives, vinyl anions IV, are well developed as versatile synthetic tools [23, 24, 25, 26], research on related borata‐vinyl anions (V) remains scarce. Indeed, to our knowledge, only one example of borata‐vinyl anion VI, featuring two boryl groups, has been described by Berndt et al. in 2001 [27], while their reactivity remains unexplored to date. The fundamental difficulty with borata‐vinyl anions synthesis stems from the excessive negative charge on the central carbon atom with a formal dianionic character due to the strongly polarized B═C bond, in addition to the vinyl anion moiety. Therefore, substituents providing powerful stabilization of such an excessive negative charge on the carbon atom are needed. Olefins (I), borata‐alkenes (II), P‐ylides (III), and their anionic derivatives (IV–VIII). A phosphonium group efficiently stabilizes an adjacent carbanionic center by forming an ylide bond [28, 29] (III in Figure 1). As evidenced by its high stabilizing capacity, phosphonium‐stabilized dianionic carbon species, namely anionic phosphonium ylides VII, have been described as stable and isolable species [30, 31, 32, 33, 34]. The unique reactivity of these anionic ylides allows access to novel compounds [35, 36], as demonstrated most recently by Gessner's group [37, 38, 39, 40, 41, 42, 43, 44]. In this context, we were interested in combining the borata‐vinyl anion (V) with a P‐ylide function (III), expecting an enhanced stability of the elusive anionic species VIII. Such a compound VIII can also be regarded as an anionic borate‐vinylidene phosphorane or a phosphine‐stabilized borate‐vinylidene anion. Several stable vinylidene phosphoranes [45, 46, 47] and their heavier analogues [48], as well as other base‐stabilized vinylidene derivatives [49, 50, 51, 52], have been reported to date. A Lewis acid‐stabilized vinylidene has also been recently synthesized by Martin et al. [53]. Here, we would like to report the synthesis of the first persistent and isolable C‐phosphonio boratavinyl anion 1 with high nucleophilic and basic properties. Result and Discussion The first attempt of synthesis of 1 starts by the reaction of borataalkene anion 2 with (chloro)(diamino)phosphine 3 followed by deprotonation using potassium hexamethyldisilazide (KHMDS) to generate a C‐phosphino borata‐alkene anion 4 (Scheme 1) (Supporting Information). The addition of isopropyl iodide to 4 at RT affords the corresponding phosphonium‐substituted derivative 5 (isolated yield from 2: 94%) which is a potential protonated precursor of C‐phosphonio boratavinyl anion 1. However, in contrast to the case of anionic ylides VII, which can be conveniently generated by deprotonation of the corresponding protonated precursors using MHMDS (M = Li, Na, and K) as bases, no reaction of 5 with KHMDS or even with nBuLi in THF was observed at RT. These results suggest an excessively low acidity of the central C─H group in 5 probably due to the adjacent electropositive boron atom. Synthesis of C‐phosphonio‐boratavinyl anion 1. Therefore, to generate desired anion 1, we have considered an alternative transmetallation reaction from halogenated derivative of 6 (Scheme 1). Thus, compound 5 was transformed into the corresponding C‐brominated product 6 with N‐bromosuccinimide (NBS) (isolated yield: 42%). The lithiation of 6 using an excess of nBuLi (2.5 equiv.) at −80°C in THF selectively affords C‐phosphonio boratavinyl anion 1. In the 31P‐NMR spectrum, only one new signal was observed at 42.3 ppm, upfield shifted relative to that of the precursor 6 (62.8 ppm). It is important to use an excess of nBuLi and to warm up the reaction mixture to RT for a short period of time (5 min.) in order to complete the transformation of the formed nBuBr via a Wurtz‐type coupling to octane [54]. A 1:1 complex formed between 1 and LiBr, was isolated as a yellow powder by the evaporation of solvent at −30°C under vacuum and washing of residue with pentane at the same temperature (yield: 75%). The boratavinyl anion 1 is stable bellow −30°C not only in solution but also in the solid state. Nevertheless, yellow crystals of LiBr‐free anion 1 suitable for the x‐ray diffraction analysis were obtained as yellow crystals from a Et2O/pentane solution at −80°C. The 11B‐NMR spectrum exhibits a broad signal at 39.7 ppm. In the 13C‐NMR spectrum, the signal corresponding to the central carbon atom appears as a broad doublet signal at 98.4 ppm (1 J CP = 95.0 Hz), which is downfield shifted compared to those observed for the corresponding protonated and brominated products (60.2 ppm for 5 and 72.3 ppm for 6) and for anionic ylides VII (−5.5–44.6 ppm) [30, 31, 32]. In the 1H‐NMR spectrum, at −80°C, two signals at 6.50 and 6.68 ppm corresponding to aryl protons for each of the two mesityl groups were observed similarly to the C‐phosphonio borata‐alkene 5. This suggests a vinyl‐lithium‐type bent structure [55, 56, 57, 58, 59] of BCP fragment in 1 instead of linear allenic‐type structure (1‐linear, Figure 3). However, these two signals start to broaden around −10°C and become one singlet signal at RT, suggesting a rapid cis‐‐trans isomerization of borata‐alkene moiety occurring at RT. Such a phenomenon in 1H‐NMR spectroscopy was not observed for other related borata‐alkenes 5 neither at RT or higher temperatures up to 105°C. Wiberg bond orders and the NPA charge at the central carbon atom (in parentheses) of C‐phosphonio borata‐alkene 5, C‐phosphonio boratavinyl anion 1, and its carbon‐analogue 7. The structure of 1 reveals a bent structure with a B─C1─P angle of 146.3(2)° with a lithium ion, coordinated by two THF molecules, in contact mainly with the central C1 atom [2.076(5) Å] as well as with the C2 atom [2.574 (5) Å] of mesityl ring (Figure 2) [60]. Although bent, a significantly shortened C1─P bond [1.637(2) Å] and a wider B─C1─P angle than those observed for the brominated precursor 6 [C1─P 1.727(2) Å and B─C1─P 136.4 (2)°] are in good agreement with an enhanced B═C═P allenic character resulting from the C1─P ylidic interaction upon lithiation, although less pronounced than that of V (B─C─B: 168.4°–180.0°). The B─C1 bond [1.461(2) Å] is similar to those observed for 6 [1.480(3) Å] and VI (1.465–1.468 Å) [27]. The borata‐alkene fragment is distorted from an ideal planar geometry with a C3─B─C1‐P torsion angle of 26.8(3)°, probably due to an increased steric congestion in addition to the strongly polarized B═C bond [61, 62]. Indeed, a similar but weaker distortion was observed for the less bulky brominated and protonated compounds 6 and 5 (12.5° and 5.3°, respectively). A previously reported C‐phosphino borata‐alkene [Mes2B = C(SiMe3)PF(NiPr2)2] presents an even more strongly twisted structure (31.5°) [63]. Molecular structure of 1 [60]. Thermal ellipsoids represent 30% probability. H and disordered atoms are omitted for clarity. Selected bond lengths [Å] and angles [°]: B─C1 1.461(4), C1─P 1.637(2), B─C2 1.642(3), B─C3 1.606(4), P─N1 1.701(2), P─N2 1.691(2), P─C4 1.857(3), C1─Li 2.076(5), B─Li 2.549(5), C2─Li 2.574(5), B─C1─P 146.3(2). Σ°B = 360.0°. Torsion angle: C3─B─C1─P 26.9(5)°. The optimized structure of 1 [M06‐2X/6‐31G(d)] agrees well with the experimental one (B─C1: 1.463 Å, C─P: 1.633 Å, B─C1─P: 148.69°, C3─B─C─P: 26.91°). Its linear isomer (1‐linear in Figure 3) is not a minimum on the energy surface and could not even be localized as a transition state for the cis– trans isomerization of boratavinyl anion 1. Instead, calculations predict that a cis–trans isomerization proceeds by B═C bond rotation through a twisted transition state structure (1‐twisted in Figure 3). The calculated bond rotation barrier is considerably low (11.0 kcal/mol), which is in good agreement with the experimentally observed rapid interconversion of 1 at RT. This low barrier for the B═C bond rotation can be explained by the stabilization of transition state (1‐twisted) by an interaction between σ‐lone pair on the carbon atom and the boryl group in addition to the π‐accepting effect of phosphonium moiety, which is evidenced by a B─C bond distance (1.472 Å) in 1‐twisted as short as that in 1 (1.462 Å). Not surprisingly, the B═C bond rotation barrier for the corresponding protonated product 5 is calculated to be higher (22.8 kcal/mol). The lowest occupied molecular orbital (LUMO, 0.37 eV) corresponds to the antibonding π*B═C orbital with a higher coefficient at the boron atom mixed with the π‐orbital in the mesityl moiety (Figure 4a). In contrast to the HOMO of 5 corresponding to the π‐allyl type orbital spreading over the P─C─B fragment (Figure 4d), the highest occupied molecular orbitals, HOMO (−5.43 eV) and HOMO‐1 (−5.88 eV), correspond to two well‐distinguished and perpendicularly oriented π‐type orbitals localized at C─P and B─C fragments respectively (Figure 4b,c), indicating its certain allenic character. Indeed, the Wiberg bond order of both C─P and B─C moieties considerably increases upon deprotonation (1.13 and 1.21 for 5 → 1.42 and 1.45 for 1). However, these two orbitals also have a high coefficient at the central carbon atom due to the strong polarization of both π‐bonds towards the carbon atom, thereby increasing the electron density on it. In fact, calculated NPA negative charge at this carbon atom (−1.59) in 1 is considerably elevated and much higher than that calculated for the corresponding carbon analogue, vinylidene phosphorane 7 (−0.76). The corresponding protonated precursor 5 also shows a strongly negatively charged central carbon atom (−1.37). These explain the low acidity of C─H moiety in 5, in contrast to the vinylidene phosphoranes, which are readily generated by the deprotonation of the corresponding protonated precursors [45, 46, 47]. These results also suggest a highly nucleophilic/basic nature of 1 at the central carbon. Calculated frontier orbitals of 1 (a–c) and 5 (d) at the M062‐X/Def2TZVP//M06‐2X/6‐31G(d) level. Indeed, due to the very high basicity of borata‐vinyl anion 1, it readily transforms, at RT in C6D6, into a 1:1 mixture of dianionic cyclic compound 8, featuring a borata‐alkene‐substituted phosphonium ylide, and protonated product 5 (Scheme 2). The reaction completes within 2 h. The transformation of 1 to 8 probably starts with its tautomerization via the intramolecular deprotonation of ortho‐methyl group of the mesityl substituent by the borata‐vinyl anion to generate the corresponding benzyl anion 9, followed by a nucleophilic attack on the phosphonium center to afford the anionic intermediate 10 (Figure 5). The following deprotonation of 10 by 1 produces the experimentally observed 1:1 mixture between 8 and 5. The same transformation was observed in Et2O and toluene solutions, while the evolution of 1 in THF solution affords a complex mixture probably because of its reaction with THF. Reactivity study of C‐phosphinio‐boratavinyl anion 1. Possible mechanism for the isomerization from 1 to 8. The highly basic and nucleophilic character of 1 was also demonstrated by a fast reaction with H2 (10 bars) at RT in THF to cleanly afford the protonated product 5 and LiH within 15 min (Scheme 2). The use of D2 leads to the formation of the corresponding deuterated product 5‐D confirming the direct reaction between 1 and D2. The concomitant generation of highly active LiH was also confirmed by its immediate reaction with dibenzylketone at RT generating the corresponding lithium enolate which was then reacted with Me3SiCl to afford a silyl enol ether [64]. The reaction is facilitated by the high pressure of H2 (10 bars). Indeed, although the same reaction takes place at 1 bar of H2, it is considerably slower (2 h) and thus becomes competitive with the transformation of the boratavinyl anion 1 to give a mixture of products 5 and 8 (ratio, 8:2). Theoretical calculations on the reaction pathway indicate that the reaction is thermodynamically favored and proceeds with a relatively low energy barrier (ΔG = −9.9 kcal/mol, ΔG ‡ = 20.3 kcal/mol) through a nucleophilic substitution at a hydrogen atom of H2 without involvement of the boron atom. The lithium cation coordinates on the H2 at the side, assisting its cleavage. Only a few highly reactive alkyl anions such as BuLi/TMDA [64] and RCaL [65] (L = NacNac ligand) are able to react with H2 at RT. Attempts to increase the reactivity of 1 by coordination of TMDA failed. Indeed, the addition of TMDA to 1, either in THF or in toluene, at ‐80°C results in an immediate protonation (and a color change of the solution from yellow to colorless), giving 5, which is certainly due to an excessively increased basicity of 1. Boratavinyl anion 1 also reacts with various E–X type electrophiles (Me─I, PhS─SPh, C6F5─F, I─Ph) via nucleophilic substitution or via a metal–halogen exchange reactions to give the corresponding C‐substituted products 11a–d (Scheme 2). The reaction with C6F6 takes place at −80°C and completes in 15 min. With PhI, a metal–halogen exchange reaction takes place, producing the corresponding iodo‐substituted compound 11d and phenyl‐lithium. The addition of CO2 to 1 at −80°C leads to an immediate reaction at the central carbon atom to generate the corresponding CO2 adduct 12. Compound 12, at RT, reacts further with a second CO2 molecule to give phosphonium ylide 13 (Scheme 2). The structure of 13 was confirmed by x‐ray‐diffraction analysis (Figure 6‐left) [60]. This result suggests an in situ generation of cyclic ylide 14, also regarded as a borane‐stabilized anionic ylide, by the isomerization of 12 via the attack of carboxyl anion on the boron atom (Scheme 2). This nucleophilic intermediate 14 reacts then with a second molecule of CO2 to give the final product 13. Molecular structures of 13 (left) and 15 (right) [60]. Thermal ellipsoids represent 30% probability. H and disordered atoms, and solvent molecules are omitted for clarity. Selected bond lengths [Å] and angles [°]: 13: C1─P 1.743(2), C1─C2 1.443(3), C2─O1 1.310(2), O1─B1 1.514(3), C2─O2 1.250(2), O2─Li 1.896(4), B─C4 1.642(3), P─C1─C2 115.6(1), C2─C1─C3 118.8(2), C3─C1─P 124.1(1). 15: B─C1 1.518(4), C1─P 1.726(3), C1─Ge 1.925(3), Ge─X 2.460(1), B─C2 1.617(4), B─C3 1.615(4), C1─Ge─X 108.6(1), B─C1─P 131.7(2), B─C1─Ge 103.9(2), Ge─C1─P 124.1(1). Torsion angle [°]: Ge─C1─B─C3 155.3(2). Boratavinyl anion 1 also readily reacts with GeCl2•1,4‐dioxane complex to afford a stable monomeric borata‐alkene‐substituted halogenogermylene 15 in 73% yield (Scheme 2). Due to the presence of LiBr in the isolated powder form of 1, germylene 15 was isolated as an 80/20 mixture of brominated and chlorinated derivatives, as shown by x‐ray diffraction analysis of crystals obtained from an ether/pentane (1:1) solution at −30°C. The structure of 15 exhibits a monomeric nature and a quasi‐planar geometry of the main framework (Figure 6, right) [60]. This suggests a π‐interaction between the π‐donating borata‐alkene moiety and the germylene center. Indeed, the C1─Ge bond [1.925(3) Å] is significantly shorter than those observed for previously reported (aryl)(chloro)germylenes (1.989–2.023 Å) [66, 67], confirming its multiple bond character. This was also indicated by the 13C‐NMR signal corresponding to the C1 atom in 15 (132.3 ppm, J PC = 79.5 Hz) significantly downfield‐shifted compared to those of related borata‐alkene derivatives 5, 6 and 11 (48.2–72.3 ppm). A similar bonding situation was observed for the chlorogermylene bearing a π‐donating carbodiphosphorane ligand [C─Ge: 1.954 Å] [68]. Of note, the germylene angle [X─Ge─C1: 108.6(1)°] is more open than those of others [91.9–101.3°] [66, 67, 68], probably for both steric and electronic reasons. This suggests a remarkable not only π‐ but also σ‐donating substituent effects of the borata‐alkene moiety. Compound 15 can also be regarded as a mono‐germanium analogue of alkyne stabilized by the coordination of a phosphine ligand to the C atom, whereas in the case of previously reported stable germynes, a phosphine ligand coordinates to the Ge center to form a stable complex presenting a carbene‐like property [69]. Conclusion In conclusion, a persistent C‐phosphonio boratavinyl anion 1 was successfully synthesized. This boratavinyl anion 1, featuring a vinyl carbon with an excessively high negative charge, exhibits remarkable nucleophilic and basic character. This was demonstrated by the results of rapid reactions with H2 at RT and with C6F6. The reactions of borata‐alkene anion 1 with electrophiles also offer a convenient route for introduction of a bulky and σ,π‐double donating substituent, enabling to effectively stabilize reactive intermediates. Moreover, the reaction of 1 with CO2 takes place at the borata‐vinyl anion moiety, rather than at the P‐ylide function, resulting in a double CO2 insertion into the B─C bond to afford a unique cyclic P‐ylide 13. Further studies on the applications of this unprecedented borata‐vinyl anion 1 in synthesis are under active investigation. Author Contributions Huihui Xu: investigation and writing – review and editing. Aurora Rodríguez‐Álvarez: investigation, writing – review and editing. Md Jabed Hossain: investigation and writing – review and editing. Antoine Baceiredo: supervision and writing – review and editing. Eddy Maerten: supervision and writing – review and editing. Nathalie Saffon‐Merceron: formal analysis, investigation, writing – review and editing. Vicenç Branchadell: formal analysis, investigation, funding acquisition, writing – review and editing. Tsuyoshi Kato: writing – original draft, supervision, funding acquisition, conceptualization. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Supporting File: anie73770‐sup‐0001‐SuppMat.docx.