Photochemical formation and reversible base-induced cleavage of a phosphagallene

Abstract

The reactivity of Cp*Ga (Cp* = C₅Me₅) towards phosphanylidenephosphoranes of the type ^ArTerP(PMe₃) (^ArTer = ^DipTer 2,6-(2,6-iPr₂C₆H₃)₂C₆H₃), ^TipTer 2,6-(2,4,6-iPr₃C₆H₂)₂C₆H₃ was investigated. While no thermal reaction was observed (in line with DFT results), irradiation at 405 nm at low temperatures resulted in the formation of phosphagallenes ^DipTerP = GaCp* (1a) and ^TipTerP = GaCp* (1b) accompanied by release of PMe₃. When warming the reaction mixture to ambient temperatures without irradiation, the clean re-formation of ^ArTerP(PMe₃) and Cp*Ga in a second-order reaction was observed. Upon removal of PMe₃, 1a and 1b were isolated and fully characterized. Both derivatives were found to be labile and decomposed to the phosphafluorenes 2a and 2b, indicating generation of the transient phosphinidene ^ArTerP along with Cp*Ga. First reactivity studies show that CO₂ and H₂O cleanly reacted with 1a, affording ^DipTerPCO (3) and ^DipTerPH₂ (4), respectively.

---

Introduction

Heavier main group element–element multiple bond systems are no longer lab curiosities and tremendous effort has been devoted to establish new bonding patterns and to utilize the transition-metal like reactivity of these systems in terms of small molecule activation and functionalization.^1–11 Our group is interested in E¹³–E¹⁵ multiple bond systems,¹² which are the iso-valent electronic heavier analogs of C–C multiple bond systems and have been proposed as intermediates in Metal–Organic Chemical Vapour Deposition (MOCVD) processes, which are based on utilizing molecular single-source E¹³–E¹⁵ precursors.^13,14 The synthesis of E¹³–E¹⁵ multiple bonds is challenging due to adjacent Lewis acidic E¹³ and Lewis basic E¹⁵ atoms, resulting in a tendency to oligomerize.^15–17 Just recently, our group isolated phospha- and arsaalumenes ^DipTerPn = AlCp* and ^TipTerP = AlCp* (^DipTer = 2,6-(2,6-iPr₂C₆H₃)₂–C₆H₃; ^TipTer = 2,4,6-(2,46-iPr₃C₆H₃)₂–C₆H₃; Cp* = C₅Me₅; Pn = P (C), As; Fig. 1)^18,19 by combining the pnictinidene transfer reagents ^ArTerPn(PMe₃)²⁰ (Ar = Dip, Tip) with (Cp*Al)₄ at 80 °C.^21,22 The heavier analogs of phosphaalumenes, phosphagallenes, have also been realized synthetically just recently, utilizing phosphanyl- or gallaphosphaketenes in the reaction with (^DipNacnac)Ga (^DipNacnac = HC[C(Me)NDip]₂) facilitating CO cleavage and formation of [(S)P]–P=Ga(^DipNacnac) (A; [(S)P] = (H_nCNDip)₂P; n = 1, 2)²³ or (^DipNacnac)Ga=P–Ga(Cl)(^DipNacnac) (B, Fig. 1),²⁴ respectively. Reactivity studies revealed that B reacted directly at the Ga=P-bond and for example the reversible insertion of two CO₂ molecules, [2 + 2] cycloadditions with isocyanates and carbodiimides,²⁵ and 1,2-oxidative additions of E–H bonds were reported.²⁶ Interestingly, the phosphanylphosphagallene A showed mainly dipolar frustrated Lewis-pair (FLP) type 1,3-reactivity towards E–H bonds,²⁷ and with CO₂ the formation of a five-membered ring species with a P₂GaCO core was described.²³ We reasoned that the combination of Cp*Ga,^28,29 with ^ArTerP(PMe₃) (Ar = Dip, Tip) would afford the phosphagallenes ^ArTerP = GaCp* by analogy with our strategy to access phospha- and arsaalumenes. Here we show that this transformation can be achieved photochemically and that in the presence of PMe₃ the reaction is fully reversible, indicating a facile and unusual ligand exchange at a phosphinidene center (Scheme 1).

Fig. 1 Known phosphagallenes^23,24 and -alumenes¹⁸ and the target of this study.

Scheme 1 Reversible formation of ^DipTerPGaCp* (1a). Images of the reaction solution before (bottom left) and after irradiation (bottom right) along with the UV absorption spectra of ^DipTerP(PMe₃) and 1a (in n-hexane, bottom middle).

Results and discussion

Synthesis and kinetic studies

Our investigations started with combining Cp*Ga and ^DipTerP(PMe₃) in toluene-d₈, giving no discernible reaction at room temperature or when heated to 100 °C over a period of 72 h according to ¹H and ³¹P NMR spectroscopy. Theoretical studies at the B3LYP-D3/def2-TZVP level of theory (cf. p. S34, ESI†) revealed an endergonic Cp*Ga for PMe₃ substitution to give ^DipTerP = GaCp* (1a). Irradiation of the NMR sample with an LED at 396 nm, which matches the longest wave-length absorption of ^DipTerP(PMe₃) (λ_max,exp = 381 nm, λ_max,calcd = 410 nm), gave a color change from yellow to green within minutes (Scheme 1). Excitation of Cp*Ga can be excluded as it shows an absorption only in the UV-region (λ_max,exp = 246 nm, λ_max,calcd = 241 nm cf. Table S3†). In the absence of light, the green color faded and only ^DipTerP(PMe₃) and Cp*Ga were detected by NMR spectroscopy, indicating the formation of the desired phosphagallene 1a in reversible fashion. Recently, the reversible phosphinidene transfer in a stannaphosphene has been reported.³⁰ We next irradiated a toluene-d₈ solution containing ^DipTerP(PMe₃) and Cp*Ga in a 1 : 1 ratio inside the NMR spectrometer using a laser diode (λ = 405 nm, 140 mW).

The formation of 1a was traced by ¹H and ³¹P NMR spectroscopy at room temperature, showing two new resonances in the ³¹P NMR spectrum at −62.9 ppm (PMe₃) and −104.8 ppm for 1a, which corresponds well with the theoretically predicted value for 1a (δ(³¹P)_calcd = −105.2 ppm). After irradiation for 2 h a dynamic steady state between 1a, PMe₃ and ^DipTerP(PMe₃), Cp*Ga was reached. When switching off the laser the concentration of 1a and PMe₃ decreased, and the starting materials were re-formed. Irradiation at −20 °C gave 1a nearly quantitatively, effectively suppressing the thermal back reaction. The thermal reverse reaction was then studied at 10, 15, 20 and 25 °C using the same sample repeatedly, indicating a fully reversible system. Using the concentrations derived from the relative integrals of 1a, PMe₃, Cp*Ga and ^DipTerP(PMe₃) in the ¹H NMR data, the reaction kinetics of the thermal reverse reaction were investigated. The reaction follows 2nd order kinetics. However, the concentration profiles could be described reasonably well using two different models, one including an associative (nucleophilic attack of PMe₃ at 1a in a S_N2-type reaction), the other a dissociative mechanism (with the dissociation reaction 1a ⇌ ^DipTerP + Cp*Ga as its first step and reaction with PMe₃ in a subsequent step; cf. p. S30†).Thus, the experimental data alone do not allow a definitive statement regarding the mechanism of the reaction. Yet, since the dissociative mechanism would include free phosphinidenes, albeit in very low concentrations, it is certainly less likely. Nonetheless, both models gave similar activation barriers (ΔG^‡ ≈ 90 kJ mol⁻¹, Table S6†) when analyzing the temperature dependence of the rate constants using transition state theory (TST).

To investigate the possible mechanism of the thermal reverse reaction further, we performed a series of semi-empirical GFN2-xTB computations³¹ as well as DFT calculations at the B3LYP-D3/def2-TZVP level of theory (cf. p. S46†).^32–34 We could in fact identify both an associative as well as a dissociative reaction pathway. The associative S_N2-type mechanism involves an activation barrier of ΔG^‡ = 80.1 kJ mol⁻¹ (, TS2, Scheme 2), in good agreement with the experimental data. A second potential associative pathway, involving a weak donor–acceptor complex between PMe₃ and 1a was also identified, however this path is associated with a significantly higher barrier (ΔG^‡ = 121.4 kJ mol⁻¹, TS1, Scheme 2) and is therefore not in line with the experimental values. The free dissociation energy of 1a into PMe₃ and singlet ^DipTerP, on the other hand, was estimated at +107.2 kJ mol⁻¹ (note a rather high uncertainty for this value due to the singlet biradical character of ^DipTerP, which is not well represented within DFT), which puts this first reaction step of the dissociative pathway in a reasonable energetic range. Assuming a low activation barrier for this dissociative pathway, it lies in a similar energy window as the S_N2-type substitution with PMe₃, and thus could be a (minor) contributing factor to the thermal reverse reaction. Note that singlet ^DipTerP would only be formed in situ, as it is predicted to possess a triplet ground state (ΔE_S–T = 45.5 kJ mol⁻¹). Singlet phosphinidenes have previously been shown to be electrophilic and irreversible ligand exchange reactions were described.^35–37

Scheme 2 Computed reaction pathway for the thermal reverse reaction of 1a with PMe₃ (B3LYP-D3/def2-TZVP, c° = 1 mol L⁻¹).

Characterization

Next, we sought to isolate 1a rather than generating it in situ.

A flask containing a 1 : 1 mixture of ^DipTerP(PMe₃) and Cp*Ga in toluene was connected to a receiving flask via a U-shaped glass tube and the headspace was evacuated (Scheme 3, top). After irradiating the mixture for 4 h (λ_LED = 396 nm) a color change to deep green was observed and the receiving flask was then cooled to ca. −70 °C to slowly evaporate PMe₃ and toluene into the receiving flask, while the product crystallized in the reaction flask. After washing with n-heptane at −78 °C, 1a was isolated as a dark turquoise solid in good yield (77%). X-ray quality crystals of 1a were grown from a saturated n-hexane solution at −30 °C.

Scheme 3 Isolation of 1a and 1b, its thermal decomposition and reactivity of 1a towards CO₂ and H₂O.

The ¹H NMR data of 1a are similar to those of ^DipTerPAlCp* (C) with three characteristic signals for the Dip-groups in the alkyl region and a singlet resonance for the Cp*-substituent, indicating η⁵-coordination, or at least a fast sigmatropic rearrangement in solution. The ³¹P NMR signal of 1a (δ(³¹P)(C₆D₆) = −104.8 ppm; δ(³¹P)(C₇D₈) = −105.5 ppm) is significantly deshielded compared to C (cf. δ(³¹P) = −203.9 ppm). The Ga–P stretching vibration in 1a at 448 cm⁻¹ (calcd 439 cm⁻¹) is red-shifted compared to the Al–P stretch in C at 558 cm⁻¹ in the IR spectrum, while the lowest energy absorption in the UV-vis (λ_max,exp = 620 nm, λ_max,calcd = 635 nm) is similar to that in ^DipTerAsAlCp* (λ_max = 590 nm).⁵

1a crystallizes in the monoclinic space group P2₁/c and is isomorphous with C with a P–Ga distance of 2.2104(5) Å (cf.C 2.2113(6); ∑r_cov(P=Ga) = 2.19 Å)³⁸ and a C1–P1–Ga1 angle of 105.02(5)°(Fig. 2, left), minimally narrower than in C, resulting in a close contact between Ga1 and one of the flanking Dip-groups of the ^DipTer-moiety (d(Ga1–C19) = 2.8869(15) Å).¹⁷ The Ga–C_Cp* distances range from 2.1526(18) to 2.5202(17), with two short and three rather long contacts, indicating that in the solid state the Cp* substituent is not perfectly η⁵-coordinated.

Fig. 2 Molecular structures of 1a (left), 1b (middle) and 3 (right). ORTEPs drawn at 50% probability. Selected bond lengths (Å) and angles (°) of 1a: P1–Ga1 2.2104(5), Ga1–C19 2.8869(15); C1–P1–Ga1 105.02(5); 1b: P1–Ga1 2.2176(5), Ga1–C7 2.8555(16); C1–P1–Ga1 104.44(6); 3: P1–C31 1.6833(12), C311–O1 1.1559(14); C1–P1–C311 105.17(5), P1–C31–O1 162.45(10).

Starting from ^TipTerP(PMe₃), ^TipTerPGaCp* (1b, 49%), was synthesized in similar fashion and X-ray quality crystals of 1b were grown from a saturated n-heptane solution at −30 °C, revealing similar structural parameters compared to 1a (Fig. 2, middle), with a short Ga1–P1 atomic distance of 2.2176(5) Å and a similar short Ga1⋯C_Tip contact (d(Ga1–C7) = 2.8555(16) Å) and a C1–P1–Ga1 angle of 104.44(6)°. This shows that the terphenyl moiety has minimal influence on the structure of 1. 1b showed a singlet signal in the ³¹P NMR spectrum at −109.8 ppm in C₆D₆ (calcd −107.9 ppm) and in the ¹H NMR spectrum two characteristic resonances in a 2 : 1 ratio were detected for the methine protons of the ^TipTer-substituent, along with three doublet signals for the iPr-groups and one singlet for the Cp* substituent at Ga.

Surprisingly, even isolated 1a and 1b are thermally labile in solution. Particularly for 1b decomposition was already observed after 10 min in C₆D₆ solution, with new signals at 1.91 and at 0.9 ppm in the ¹H NMR spectrum indicating the formation of Cp*Ga and a phosphafluorene (2b, Scheme 3), previously observed when ^TipTerP(PMe₃) was irradiated at λ = 365 nm.³⁹ Albeit slower, 1a showed a similar behavior in solution, decomposing to give Cp*Ga and the related phosphafluorene 2a. The thermal decomposition in C₇D₈ solution was then studied by ¹H NMR spectroscopy over time, which revealed a 1st order decay of 1a and 1b with half lives of 47.8 and 7.8 h, respectively. This compares well with the theoretically determined free dissociation energy of 1a into ^DipTerP and Cp*Ga (ΔG_diss,calcd = 107.2 kJ mol⁻¹, ; cf. Section in the ESI†) and corroborates the formation of 2a as decomposition product of ^DipTerP. Moreover, the faster decomposition of 1b can be understood in terms of electrophilic attack of the phosphinidene on one of the flanking aryl groups, which should be faster for the more electron-rich Tip-substituent present in 1b.

Electronic structure

The electronic structure of 1a was probed using NBO,^40,41 AIM,^42,43 and ELF analyses.⁴⁴ The P–Ga bonding is best described as a polarized double bond (Scheme 4): The σ bond, which is formed by the 3p_x(P) and 4s(Ga) orbitals, is evenly distributed between both atoms. The π bond, on the other hand, is strongly polarized towards the P atom [∼87% 3p_z(P), ∼13% 4p_z(Ga); resonances type II and III], which agrees with the NBO data for known variants A and B.^23,24 The polarization of the P–Ga double bond is also reflected in the natural charges of P (−0.28e) and Ga (+1.11e) as well as in the Wiberg bond index of 1.31. In structure II-a, there are two formal lone valences at the Ga atom (4p_x and 4p_y orbitals), which are stabilized by donor–acceptor interactions with the s-type LP at the P atom (resonance I) as well as the Cp* ligand (resonances of type b), respectively (see also Scheme S2†). Note that resonance I, to which we attribute only a small weight, should be understood in terms of a non-classical multiple bond, i.e. the LP(P) → 4p_x(Ga) interaction is not a π bond (in-depth description of the electronic structure, p. S40ff†). Inspection of the Electron Localization Function (ELF, Fig. 3) supported the results from NBO analysis, with a LP on the P atom and strongly polarized π-electron density towards the P atom perpendicular to the Ga–P–C_Ter plane, in line with resonance between Lewis structures II and III (Scheme 3).

Scheme 4 Lewis resonance scheme of compound 1a.

Fig. 3 ELF of 1a in the Ga–P–C plane (left) and perpendicular to that plane (right).

Inspection of the Laplacian of the electron density ∇²ρ corroborates charge accumulation near the P atom in the π bonding system, in agreement with the NBO and ELF results. The ellipticity ε at the bond critical point (BCP) of 0.31 indicates double bond character (cf. ∼0.3 for ethylene, ∼0.2 for benzene; Table S15†).

Reactivity studies

Next, we tested the reactivity of 1a towards CO₂. Phosphagellenes were previously shown to react with CO₂ to give five-membered ring species²³ or to reversibly insert two CO₂ molecules into the Ga–P bond.²⁴ When a green toluene-d₈ solution of 1a was exposed to an atmosphere of CO₂ a color change to orange was observed after 20 min at room temperature.

A resonance in the ³¹P{¹H} NMR spectrum at −201.3 ppm, along with a doublet in the ¹³C NMR spectrum at 199.9 ppm (¹J_PC = 102.4) indicated the formation of the arylphosphaketene ^DipTerPCO (3), which was verified by SC-XRD experiments on crystals grown from n-hexane at −30 °C. Additionally, a broad feature between 1.65–2.01 ppm in the ¹H NMR spectrum with an integral value of ca. 15 indicated the formation of various Cp*-containing species. The PCO unit in 3 is in-plane with the central aryl ring of the ^DipTer-substituent, and deviates from linearity (⦠(P1–C31–CO) = 162.45(10)°) with short P1–C31 (1.6833(12) Å) and C31–O1 (1.1559(14) Å) distances (Fig. 2, right). Surprisingly, 3 represents the first structurally characterized arylphosphaketene. Phosphaketene, tBu–PCO, was first isolated by Appel and Paulen in 1983 through treatment of tBuP(SiMe₃)₂ with phosgene at −70 °C, with a characteristic ³¹P NMR shift of −180 ppm. In the same year Mes*PCO (Mes* = 2,4,6-tBu₃C₆H₂) was isolated as stable orange crystalline solid,⁴⁵ and has been utilized as phosphinidene transfer reagent.^46,47 Recently, the chemistry of phosphaketenes has seen a resurgance based on Na(dioxane)_xPCO as an easily accessible form of the phosphaethynolate anion [PCO]⁻,⁴⁸ allowing to access heteroatom-substituted E–PCO systems through salt metathesis reactions.^{23,24,35–37,49–51} The deoxygenation of CO₂ to access 3 is reminiscent of the bora-phospha-Wittig reaction, in which phosphaborenes are utilized to make phosphaalkenes.⁵² Due to the presence of Cp*-containing impurities it was not possible to isolate 3 in pure form, however, it can be generated in situ in clean fashion, which will allow further reactivity studies in the future.

With H₂O 1a reacted to give ^DipTerPH₂ (4) quantitatively,⁵³ clearly underlining the potential of 1a to act as an deoxygenation reagent, while also clearly deviating from the reactivity of ^DipTerP(PMe₃) towards H₂O, which afforded the primary phosphine oxide ^DipTerP(O)H₂ instead.⁵⁴

Conclusion

Phosphagallenes have previously been obtained by reacting phosphaketenes with the Ga(I) source ^DipNacnacGa as thermally stable entities. Here we show that ^DipTerPGaCp* (1a) is only formed upon irradiation of ^DipTerP(PMe₃) in the presence of Cp*Ga. Under thermal conditions 1a reacted with PMe₃ back to the starting materials in a 2nd-order reaction, indicating a facile and reversible PMe₃ for Cp*Ga exchange at a phosphinidene. When removing PMe₃ while irradiating the reaction mixture, 1a and 1b were isolated. However, in solution the isolated phosphagallenes dissociate into the phosphinidene ^ArTerP and Cp*Ga. ^ArTerP is highly reactive and formed the phosphafluorenes 2a and 2b. The bonding in 1a was investigated by combined theoretical means (NBO, ELF, AIM), clearly showing a strongly polarized (towards P) P–Ga bond with significant double bond character. Considering the longer half-life of 1a, its reactivity towards CO₂ and H₂O was probed, giving phosphaketene 3 and phosphine 4, respectively. Further reactivity studies to harness the reversible formation of 1a and its lability with respect to phosphinidene release are currently underway.

Data availability

Crystallographic data for 1a, 1b and 3 has been deposited at the CCDC under 2216822–2216824. The datasets supporting this article have been uploaded as part of the ESI.†

Author contributions

J. P., J. B. and C. H.-J. conceptualized the project and designed the experiments. T. T., J. B. and C. H.-J. carried out most of the experimental work. T. T., F. D., J. B., J. P. and C. H.-J. characterized the compounds and analyzed the data. D. M., J. B. and C. H.-J. performed the in situ NMR studies and analyzed the data. J. B. carried out the computational studies. 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.

Acknowledgements

F. D. and C. H.-J. thank the Leibniz Association for funding within the scope of the Leibniz ScienceCampus Phosphorus Research Rostock (https://www.sciencecampus-rostock.de). We wish to thank the ITMZ at the University of Rostock for access to the cluster computer, and especially Malte Willert for his assistance with the queuing system and software installations.

Notes and references

1. L. Weber, Chem. Rev., 1992, 92, 1839–1906.
2. P. P. Power, Chem. Rev., 1999, 99, 3463–3504.
3. Y. Wang and G. H. Robinson, Chem. Commun., 2009, 5201–5213.
4. R. C. Fischer and P. P. Power, Chem. Rev., 2010, 110, 3877–3923.
5. P. P. Power, Acc. Chem. Res., 2011, 44, 627–637.
6. P. Bag, C. Weetman and S. Inoue, Angew. Chem., Int. Ed., 2018, 57, 14394–14413.
7. T. Chu and G. I. Nikonov, Chem. Rev., 2018, 118, 3608–3680.
8. J.-D. Guo and T. Sasamori, Chem.–Asian J., 2018, 13, 3800–3817.
9. C. Weetman and S. Inoue, ChemCatChem, 2018, 10, 4213–4228.
10. R. L. Melen, Science, 2019, 363, 479–484.
11. R. Borthakur and V. Chandrasekhar, Coord. Chem. Rev., 2021, 429, 213647.
12. F. Dankert and C. Hering-Junghans, Chem. Commun., 2022, 58, 1242–1262.
13. S. Schulz, in Advances in Organometallic Chemistry, Academic Press, 2003, vol. 49, pp. 225–317.
14. M. A. Malik, M. Afzaal and P. O'Brien, Chem. Rev., 2010, 110, 4417–4446.
15. M. Kapitein and C. von Hänisch, Eur. J. Inorg. Chem., 2015, 2015, 837–844.
16. M. Kapitein, M. Balmer, L. Niemeier and C. von Hänisch, Dalton Trans., 2016, 45, 6275–6281.
17. S. Nees, F. Fantuzzi, T. Wellnitz, M. Fischer, J.-E. Siewert, J. T. Goettel, A. Hofmann, M. Härterich, H. Braunschweig and C. Hering-Junghans, Angew. Chem., Int. Ed., 2021, 60, 24318–24325.
18. M. Fischer, S. Nees, T. Kupfer, J. T. Goettel, H. Braunschweig and C. Hering-Junghans, J. Am. Chem. Soc., 2021, 143, 4106–4111.
19. S. Nees, T. Wellnitz, F. Dankert, M. Härterich, S. Dotzauer, M. Feldt, H. Braunschweig and C. Hering-Junghans, Angew. Chem., Int. Ed., 2023, e202215838.
20. P. Gupta, J.-E. Siewert, T. Wellnitz, M. Fischer, W. Baumann, T. Beweries and C. Hering-Junghans, Dalton Trans., 2021, 50, 1838–1844.
21. C. Dohmeier, C. Robl, M. Tacke and H. Schnöckel, Angew. Chem., Int. Ed., 1991, 30, 564–565.
22. H. Sitzmann, M. F. Lappert, C. Dohmeier, C. Üffing and H. Schnöckel, J. Organomet. Chem., 1998, 561, 203–208.
23. D. W. N. Wilson, J. Feld and J. M. Goicoechea, Angew. Chem., Int. Ed., 2020, 59, 20914–20918.
24. M. K. Sharma, C. Wölper, G. Haberhauer and S. Schulz, Angew. Chem., Int. Ed., 2021, 60, 6784–6790.
25. M. K. Sharma, C. Wölper, G. Haberhauer and S. Schulz, Angew. Chem., Int. Ed., 2021, 60, 21784–21788.
26. M. K. Sharma, C. Wölper and S. Schulz, Dalton Trans., 2022, 51, 1612–1616.
27. J. Feld, D. W. N. Wilson and J. M. Goicoechea, Angew. Chem., Int. Ed., 2021, 60, 22057–22061.
28. D. Loos and H. Schnöckel, J. Organomet. Chem., 1993, 463, 37–40.
29. P. Jutzi and L. O. Schebaum, J. Organomet. Chem., 2002, 654, 176–179.
30. M. Fischer, M. M. D. Roy, L. L. Wales, M. A. Ellwanger, A. Heilmann and S. Aldridge, J. Am. Chem. Soc., 2022, 144, 8908–8913.
31. C. Bannwarth, E. Caldeweyher, S. Ehlert, A. Hansen, P. Pracht, J. Seibert, S. Spicher and S. Grimme, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 2021, 11, e1493.
32. H. Jonsson, G. Mills and K. W. Jacobsen, in Classical and Quantum Dynamics in Condensed Phase Simulations, 1998, pp. 385–404.
33. F. Neese, F. Wennmohs, U. Becker and C. Riplinger, J. Chem. Phys., 2020, 152, 224108.
34. F. Neese, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 2022, 12, e1606.
35. M. M. Hansmann and G. Bertrand, J. Am. Chem. Soc., 2016, 138, 15885–15888.
36. M. M. Hansmann, R. Jazzar and G. Bertrand, J. Am. Chem. Soc., 2016, 138, 8356–8359.
37. L. Liu, D. A. Ruiz, D. Munz and G. Bertrand, Chem, 2016, 1, 147–153.
38. P. Pyykkö and M. Atsumi, Chem.–Eur. J., 2009, 15, 12770–12779.
39. S. Shah, M. C. Simpson, R. C. Smith and J. D. Protasiewicz, J. Am. Chem. Soc., 2001, 123, 6925–6926.
40. E. D. Glendening, A. E. Reed, J. E. Carpenter, J. A. Bohmann, C. M. Morales, C. R. Landis and F. Weinhold, Theoretical Chemistry Institute, University of Wisconsin, Madison, WI, 2013.
41. F. Weinhold, C. R. Landis and E. D. Glendening, Int. Rev. Phys. Chem., 2016, 35, 399–440.
42. R. F. W. Bader, Atoms in Molecules: A Quantum Theory,Oxford University Press, 1994.
43. T. Lu and F. Chen, J. Comput. Chem., 2012, 33, 580–592.
44. A. D. Becke and K. E. Edgecombe, J. Chem. Phys., 1990, 92, 5397–5403.
45. R. Appel and W. Paulen, Angew. Chem., Int. Ed. Engl., 1983, 22, 785–786.
46. A. H. Cowley, B. Pellerin, J. L. Atwood and S. G. Bott, J. Am. Chem. Soc., 1990, 112, 6734–6735.
47. A. H. Cowley, F. Gabbai, R. Schluter and D. Atwood, J. Am. Chem. Soc., 1992, 114, 3142–3144.
48. J. M. Goicoechea and H. Grützmacher, Angew. Chem., Int. Ed., 2018, 57, 16968–16994.
49. A. R. Jupp and J. M. Goicoechea, J. Am. Chem. Soc., 2013, 135, 19131–19134.
50. D. Heift, Z. Benkő and H. Grützmacher, Dalton Trans., 2014, 43, 5920–5928.
51. K. M. Szkop, A. R. Jupp, H. Razumkov and D. W. Stephan, Dalton Trans., 2020, 49, 885–890.
52. A. M. Borys, E. F. Rice, G. S. Nichol and M. J. Cowley, J. Am. Chem. Soc., 2021, 143, 14065–14070.
53. W. A. Merrill, E. Rivard, J. S. DeRopp, X. Wang, B. D. Ellis, J. C. Fettinger, B. Wrackmeyer and P. P. Power, Inorg. Chem., 2010, 49, 8481–8486.
54. F. Dankert, M. Fischer and C. Hering-Junghans, Dalton Trans., 2022, 51, 11267–11276.

---

Footnote

† Electronic supplementary information (ESI) available: Synthesis and characterization of compounds, NMR spectra, crystallographic, and computational details. CCDC 2216822–2216824. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d2sc06292e

---