Intramolecular dearomative 1,4-addition of silyl and germyl radicals to a phenyl moiety

Abstract

Herein, we present that the radicals [Ph₃PC(Me)EMes₂]˙ (2^Si and 2^Ge) can be generated from the α-silylated and α-germylated phosphorus ylides Ph₃PC(Me)E(Cl)Mes₂ (1^Si and 1^Ge) through one-electron reduction with Jones’ dimer (^MesNacNacMg)₂ in benzene. Although isolation of the free radicals was not possible, the products of the intramolecular addition of the radicals to a phenyl substituent of the phosphorus moiety, followed by subsequent reaction with 2^Si or 2^Ge to the isolated species 3^Si and 3^Ge, respectively, were observed. This transformation witnesses a dearomative 1,4-addition of tetryl radical species to the phenyl scaffold in a stereoselective anti-fashion.

---

Radical species exhibit unique electronic properties and reactivities owing to their unpaired electrons, rendering them the focus of intense interest for decades. The Ph₃C˙ radical (Gomberg radical) is probably the best-studied radical of group 14. It was long thought that, in solution, this species is in equilibrium with its symmetrical dimer, i.e. hexaphenylethane.¹ Further studies showed that this symmetrical dimer is only formed in the case of the heavier analogues of group 14,² whereas the Ph₃C˙ radical actually attacks one phenyl ring of another Ph₃C˙ molecule in para position resulting in an unexpected dearomatized regioisomeric σ-dimer (Scheme 1).³

Scheme 1 Equilibrium of Gomberg-type radicals in solution.

The heavier Gomberg-type radicals (Mes₃E˙) of silicon, germanium and tin bearing bulky mesityl substituents (Mes) were reported by Lappert and co-workers.⁴ The authors described dimerization reactions or reactions with the solvent once the radicals were generated. The germyl radical showed the longest half-life time (t_1/2 > 24 h at 0 °C) because the dimerization reaction is suppressed by steric hindrance compared to the much larger tin atom (t_1/2 = 20 s at 0 °C), and the reaction with the solvent is much slower than for the silicon (t_1/2 = ca. 20 s at 0 °C) analogue. The decomposition products were later described as a mixture of Si–H group-containing compounds, polymeric material and other products.⁵

The decay of such radicals was studied in detail by Ingold and co-workers. They found a similar reaction known for the Gomberg radical, where the Ph₃Si˙ radical adds to a phenyl moiety of another molecule of Ph₃SiH in the para position.⁶ This reactivity has been described before for the smaller alkyl-substituted radicals such as Me₃Si˙ and Et₃Si˙, which readily react with benzene.⁷ Pears and co-workers studied the addition of Me₃Si˙ and Me₃Ge˙ radicals to benzene in detail. Irradiation of (Me₃Si)₂Hg in benzene gave the homolytic aromatic substitution products such as phenyltrimethylsilane (20%), 2,5-cyclohexadienyltrimethylsilane (12%) and hexamethyldisilane (27%), along with silylated biphenyls, phenylcyclohexadienes and bis(cyclohexadienyls). For the similar reaction involving (Me₃Ge)₂Hg, they obtained Me₆Ge₂ (98%), PhGeMe₃ (1%) and 2,5-cyclohexadienyltrimethylgermane (traces). This was ascribed to the significantly weaker Ge–C bond, whose formation is less favourable compared to the that of the dimer (Me₃Ge)₂ (Scheme 2a).⁸ The chemistry of silyl and germyl radical species and their reactivity have been summarized in numerous review articles over the past decades.⁹

Scheme 2 (a) Typical reactivity of R₃E˙ (E = Si, Ge) with benzene. (b) Recent reports of the dearomative 1,4-additions of main group element radicals to benzene. (c) Dearomative 1,4-addition of phosphorous ylide stabilized silyl and germyl radicals to a phenyl moiety reported in this work.

Considering the addition to multiple bonds or aromatic systems, germyl radicals are comparatively less reactive than their silicon counterparts. In some cases, however, such transformations were proven as powerful synthetic tools like the recent intramolecular addition of Ge radicals to an alkyne function reported by Durandetti and co-workers.¹⁰ Dearomative 1,4-additions of radicals to benzene moieties resembling Birch-type reactivity are considered as promising tools for arene transformations.^9c Reports from Harder,¹¹ Jones,¹² or Andrada¹³ showcase this reaction mode for magnesium and aluminum compounds under reductive or photolytic conditions, clearly indicating the corresponding radicals to be the crucial intermediates (Scheme 2b). The Birch-type addition of silicon and germanium compounds to benzene scaffolds was by now achieved by cycloaddition strategies like the [4 + 3] cycloaddition reported by Wesemann¹⁴ or the [2 + 2] cycloaddition reported by Cummins.¹⁵ In our understanding, the dearomative 1,4-addition to phenyl rings is hard to accomplish with germyl radicals due to their low reactivity compared to group 2 or 13 and the relatively low Ge–C bond energies.

Herein, we present the observation of an intramolecular addition of phosphorus ylide-stabilized silyl and germyl radicals to a phenyl moiety at the ylidic P atom, followed by recombination of the formed phenyl radical with another group 14-centered radical (Scheme 2c).

We recently showed that α-silylated and α-germylated phosphorus ylides of the general formula Ph₃PC(R¹)E(Cl)R²₂ (1^E with E = Si, Ge) constitute valuable precursors for accessing the cationic species [Ph₃PC(R¹)ER²₂]⁺.¹⁶ We also became interested in whether the phosphorus ylides are capable of stabilizing group 14 radical species by the interaction of the P–C bond with the open-shell Si or Ge atoms. Studies indicated that for the mesityl-substituted derivatives Ph₃PC(Me)E(Cl)Mes₂ (1^Si and 1^Ge), the radicals [Ph₃PC(Me)EMes₂]˙ (2^Si and 2^Ge) can be generated through reduction with Jones’ dimer (^MesNacNacMg)₂ in benzene, which leads to full conversion of the starting materials within several minutes (Scheme 3).¹⁷ To our surprise, we were not able to isolate the free radicals. Although signals were detected during in situ cw-EPR spectroscopic studies, their intensity was very low, despite the high concentration of the samples. The lack of observable hyperfine couplings prevented a detailed EPR spectroscopic characterization. Although the appearance of EPR active species indicated radical pathways, initialised by one-electron reduction with Jones’ reagent, the main product of the reaction seemed to be of diamagnetic nature.¹⁸In situ NMR spectroscopic studies showed a clean conversion of both starting materials (1^Ge and 1^Si) into one new species (ESI, Fig. S10 and 11†). In the case of germanium, we were able to crystallize the Birch-type adduct 3^Ge from hexanes (Fig. 1; space group P1̄). It appears reasonable to assume that 3^Ge is formed through the intramolecular attack of the initially formed germanium-centred radical 2^Ge at one phenyl moiety of the Ph₃P fragment generating a five-membered ring bearing a radical in the aryl moiety. Afterwards, the phenyl-based radical recombines with another molecule 2^Ge leading to the formation of 3^Ge.

Scheme 3 Reduction of 1^E with [^MesNacNacMg]₂ in benzene at room temperature.

Fig. 1 Molecular structure of 3^Ge; ellipsoids are drawn at the 30% probability level. Hydrogen atoms are omitted for clarity. Selected bond lengths (Å) and angles (°): P1–C1 = 1.672(3), P2–C9 = 1.682(3), Ge1–C1 = 1.914(3), Ge1–C3 = 2.037(3), Ge2–C6 = 2.090(3), Ge2–C9 = 1.901(3), C3–C4 = 1.504(4), C3–C8 = 1.490(4), C5–C6 = 1.497(4), C6–C7 = 1.494(4), C4–C5 = 1.337(4), C7–C8 = 1.329(4); C1–Ge1–C3 = 95.49(12), P1–C1–Ge1 = 111.03(15), C9–Ge2–C6 = 103.71(12), P2–C9–Ge2 = 131.10(18), C3–C6–Ge1 = 125.2, C3–C6–Ge2 = 140.7.

The P1–C1 bond in 3^Ge (1.672(3) Å) is 10 pm shorter than the P2–C9 bond (1.682(3) Å), which is as long as the P1–C1 bond (1.683(3) Å) in 1^Ge indicating the preservation of the ylidic function.¹⁶ The bond lengths in the bridging six-membered ring (see caption of Fig. 1) clearly show the cyclohexadienyl character of the former aromatic moiety. The distances between the bridging cyclohexyl ring and the germanium atoms are quite similar (Ge1–C3 = 2.037(3) Å; Ge2–C6 = 2.090(3) Å). The angles P1–C1–Ge1 = 111.03(15)° and P2–C9–Ge2 = 131.10(18)° differ by 20° from each other (cf. P1–C1–Ge1 = 126.85(11) for 1^Ge).¹⁶

The ³¹P{¹H} NMR spectra of 3^Si and 3^Ge show two signals at δ_³¹P = 18.3 and 20.3 ppm (3^Si) and 18.1 and 18.5 ppm (3^Ge), respectively, which support the formation and stability of the unsymmetrical dimeric structures in solution. For 3^Si, two doublets with chemical shifts of δ_²⁹Si = −11.8 and −34.7 ppm and ²J_PSi coupling constants of 41.7 and 31.9 Hz were detected in the ²⁹Si NMR spectrum. In the ¹H NMR spectra, two distinct sets of signals for the mesityl and methyl moieties of 3^Si were detected. A set of multiplets for the cyclohexyl protons of the bridging ring is found at δ_¹H = 8.14, 7.38, 7.08 and 6.95 ppm for 3^Si. The most characteristic signal at 8.14 ppm presents the ortho proton at the C5 atom. This is in good accordance with our previously reported ortho-substituted phosphorus ylides.¹⁹ Similar spectra were obtained for 3^Ge, but the mesityl groups give four sets of signals, which could be due to a hindered rotation.

The fact that such reactivity with aromatic substrates was not reported so far for germyl and silyl radicals naturally raised the question on a conceivable equilibrium between dimeric (closed shell) and monomeric (radical) species. We therefore investigated solutions of 3^Si and 3^Ge in toluene-d₈ by variable temperature (VT) NMR experiments (293–363 K). However, both compounds showed no change in the ¹H NMR spectra at elevated temperatures (ESI, Fig. S8 and 9†), despite thermal activation of some hindered rotations along the sterically crowded scaffolds, especially for the mesityl moieties. This suggested that within the probed temperature range, the dimerization of the radicals 2^E and INT2^E (Scheme 3) is not reversible enough to form energetically unfavored monomer species in NMR-detectable amounts. Additionally, we challenged this question by more sensitive EPR spectroscopy, as especially 2^Ge was expected to show significant hyperfine coupling to germanium due to large calculated spin density on the Ge atom (ESI, Fig. S12 and 13†). However, both compounds 3^E are EPR-silent up to 80 °C, confirming the NMR observations discussed above that the dimerization of the radicals 2^E and INT2^E is strongly exergonic.

To support these experimental findings, we computed the reaction pathway of the formation of 3^Si and 3^Ge at the PCM-M06-2X/def2-TZVPP//M06-2X/def2-SVP level. As can be seen from Fig. 2, the germanium-centred radical 2^Ge evolves into the radical intermediate INT2^Ge in an endergonic transformation (ΔG = 8.7 kcal mol⁻¹) through the transition state TS-Ge, which is associated with the intramolecular addition of the radical to the adjacent phenyl group attached to the phosphorous. The computed barrier for this process is rather low (ΔG^‡ = 10.8 kcal mol⁻¹), which is mainly due to the close proximity of the reactive phenyl group to the Ge radical as a consequence of the occurrence of a stabilizing π–π interaction with one of the mesityl groups. This is confirmed by the relatively short distance between the centroids of both aryl groups (3.56 Å) and by the NCIPLOT method, which clearly shows the existence of such π–π interaction (see inset in Fig. 2). In sharp contrast, the analogous intermolecular reaction involving benzene (solvent) proceeds with a much higher barrier of 29.0 kcal mol⁻¹ indicating that this alternative reaction is not competitive. Indeed, the NMR spectroscopic studies of the reduction reaction showed also no reaction between the intermediary formed radicals 2^Si and 2^Ge with the solvent benzene-d₆. Subsequent recombination with another molecule 2^Ge in a barrierless process leads to the highly exergonic formation of 3^Ge. The analogous process involving 2^Si is favoured from both kinetic (ΔG^‡ = 6.7 kcal mol⁻¹) and thermodynamic (ΔG = −34.3 kcal mol⁻¹) points of view, which is not surprising considering the higher stability of the Si–C bond compared to the Ge–C bond.

Fig. 2 Calculated reaction pathway to the formation of 3^Si and 3^Ge. Relative free energies (ΔG, at 298 K) and bond lengths are given in kcal mol⁻¹ and angstroms, respectively. Values within parentheses indicate the computed spin-densities. Inset: contour plots of the reduced density gradient isosurfaces (density cutoff of 0.03 a.u.) for compound 2^Ge. The green surfaces indicate attractive non-covalent interactions All data have been computed at the PCM-M06-2X/def2-TZVPP//M06-2X-/def2-SVP level.

Conclusively, we observed the dearomative 1,4-additions of the phosphorus ylide-stabilized silicon (2^Si) and germanium (2^Ge) radicals to phenyl moieties resembling Birch-type reactivity. The radicals [Ph₃PC(Me)EMes₂]˙ were generated from α-silylated and -germylated phosphorus ylides Ph₃PC(Me)E(Cl)Mes₂ through one-electron reduction with Jones’ dimer (^MesNacNacMg)₂ in benzene. We observed the reaction because the radicals are (1) sterically shielded, (2) stabilized within the ylide π-system and (3) the barrier for intramolecular attack is low due to activation of the ring by the PPh₃ fragment and the close proximity of the reactive centres.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors acknowledge PD Dr Hans Egold and Prof. Dr Jan Paradies for recording NMR spectra as well as Alexander Allgaier and Prof. Dr Joris van Slageren for recording high temperature EPR spectra. The authors also acknowledge support by the German Research Foundation (DFG) through grant no BR 2169/5-1 (535548579). J. O. W. thanks the Verband der Chemischen Industrie (VCI) for financial support by a Kekulé scholarship. I. F. is grateful for financial support from the Spanish MCIN/AEI/10.13039/501100011033 (grants PID2019-106184GB-I00, PID2022-139318NB-I00 and RED2022-134287-T).

References

1. M. Gomberg, Ber. Dtsch. Chem. Ges., 1900, 33, 3150–3163.
2. (a) H. Gilman and G. E. Dunn, J. Am. Chem. Soc., 1951, 73, 5077–5079; (b) H. Gilman and C. W. Gerow, J. Am. Chem. Soc., 1955, 77, 5509–5512; (c) L. C. Willemsens and G. J. M. Van Der Kerk, J. Organomet. Chem., 1970, 21, 123–130; (d) H.-U. Buschhaus, W. P. Neumann and T. Apoussidis, Liebigs Ann. Chem., 1981, 1190–1197.
3. H. Lankamp, W. T. Nauta and C. MacLean, Tetrahedron Lett., 1968, 9, 249–254.
4. (a) M. J. S. Gynane, M. F. Lapper, P. Riviere and M. Riviere-Baudet, J. Organomet. Chem., 1977, 142, C9–C11; (b) M. J. S. Gynane, M. F. Lappert, P. I. Riley, P. Rivière and M. Rivière-Baudet, J. Organomet. Chem., 1980, 202, 5–12; (c) C. Drost, J. Griebel, R. Kirmse, P. Lönnecke and J. Reinhold, Angew. Chem., Int. Ed., 2009, 48, 1962–1965.
5. W. P. Neumann and K.-D. Schultz, J. Chem. Soc., Chem. Commun., 1982, 43–44.
6. C. Chatgilialoglu, K. Ingold, J. Lusztyk, A. Nazran and J. Scaiano, Organometallics, 1983, 2, 1332–1335.
7. (a) C. Eaborn, R. Jackson and R. Pearce, Chem. Commun., 1967, 920a–920a; (b) H. Sakurai and A. Hosomi, J. Am. Chem. Soc., 1971, 93, 1709–1714; (c) P. Wong, P. Marriott, D. Griller, D. Nonhebel and M. Perkins, J. Am. Chem. Soc., 1981, 103, 7761–7763; (d) C. Chatgilialoglu, K. Ingold and J. Scaiano, J. Am. Chem. Soc., 1983, 105, 3292–3296.
8. S. W. Bennett, C. Eaborn, R. A. Jackson and R. Pearce, J. Organomet. Chem., 1971, 28, 59–72.
9. (a) M. F. Lappert and P. W. Lednor, in Advances in Organometallic Chemistry, 1976, pp. 345–399; (b) C. Lichtenberg, Chem. – Eur. J., 2020, 26, 9674–9687; (c) H. Zhu, S. Fujimori, A. Kostenko and S. Inoue, Chem. – Eur. J., 2023, 29, e202301973; (d) A. Hinz, J. Bresien, F. Breher and A. Schulz, Chem. Rev., 2023, 123, 10468–10526; (e) P. P. Power, Chem. Rev., 2003, 103, 789–810; (f) T. Stuyver, B. Chen, T. Zeng, P. Geerlings, F. De Proft and R. Hoffmann, Chem. Rev., 2019, 119, 11291–11351; (g) B. Tang, J. Zhao, J.-F. Xu and X. Zhang, Chem. Sci., 2020, 11, 1192–1204; (h) K. Chandra Mondal, S. Roy and H. W. Roesky, Chem. Soc. Rev., 2016, 45, 1080–1111; (i) Z. Feng, S. Tang, Y. Su and X. Wang, Chem. Soc. Rev., 2022, 51, 5930–5973; (j) Q.-H. Xu and B. Xiao, Org. Chem. Front., 2022, 9, 7016–7027; (k) A. Hinz and F. Breher, in Organogermanium Compounds, ed. V. Y. Lee, John Wiley & Sons, 2023, pp. 339–360; (l) V. Y. Lee, Organosilicon Compounds: Theory and experiment (synthesis), Academic Press, 1st edn, 2017, pp. 231–287.
10. (a) S. Kassamba, A. Perez-Luna, F. Ferreira and M. Durandetti, Chem. Commun., 2022, 58, 3901–3904; (b) S. Kassamba, M. Reboli, A. Perez-Luna, F. Ferreira and M. Durandetti, Org. Chem. Front., 2023, 10, 3328–3335.
11. T. X. Gentner, B. Rösch, G. Ballmann, J. Langer, H. Elsen and S. Harder, Angew. Chem., Int. Ed., 2019, 58, 607–611.
12. D. D. L. Jones, I. Douair, L. Maron and C. Jones, Angew. Chem., Int. Ed., 2021, 60, 7087–7092.
13. D. Mandal, T. I. Demirer, T. Sergeieva, B. Morgenstern, H. T. A. Wiedemann, C. W. M. Kay and D. M. Andrada, Angew. Chem., Int. Ed., 2023, 62, e202217184.
14. R. H. Kern, M. Schneider, K. Eichele, H. Schubert, H. F. Bettinger and L. Wesemann, Angew. Chem., Int. Ed., 2023, 62, e202301593.
15. A. Velian, W. J. Transue and C. C. Cummins, Organometallics, 2015, 34, 4644–4646.
16. F. Krämer, M. Radius, A. Hinz, M. E. A. Dilanas and F. Breher, Chem. – Eur. J., 2022, 28, e202103974.
17. S. J. Bonyhady, C. Jones, S. Nembenna, A. Stasch, A. J. Edwards and G. J. McIntyre, Chem. – Eur. J., 2010, 16, 938–955.
18. C. Jones, Nat. Rev. Chem., 2017, 1, 0059.
19. F. Krämer, J. Paradies, I. Fernández and F. Breher, Nat. Chem., 2024, 16, 63–69.

---

Footnote

† Electronic supplementary information (ESI) available. CCDC 2293352. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d4dt00089g

---