Hydrosilylation and hydrogermylation of white phosphorus

Abstract

The development of efficient, direct strategies for the transformation of white phosphorus (P₄) into useful monophosphorus compounds, as alternatives to the current wasteful and hazardous indirect processes, remains a significant challenge. Encouragingly, recent reports have shown that the reduction of P₄ with organotin hydrides and subsequent functionalisation with electrophiles allows for the efficient synthesis of an array of industrially relevant monophosphines in a ‘one-pot’ manner. However, despite the practical and conceptual simplicity, the appeal of this method is limited by the inherent toxicity of most organotin derivatives. Here, we address this problem through experimental and computational studies of the reactivity of lighter and less toxic hydrogermane and hydrosilane homologues of organotin hydrides (R₃EH, E = Ge or Si) towards P₄. These hydroelementation reactions can be employed to directly transform P₄ into useful monophosphorus compounds, in a simple ‘one-pot’ fashion similar to the original organotin-based systems.

---

Introduction

White phosphorus, P₄, is by far the most reactive and industrially relevant allotrope of the element. It serves as the synthetic precursor for all commercially valuable and academically important organophosphorus compounds (OPCs).^1–4 Current industrial routes for converting P₄ into these useful OPCs involve indirect multi-step processes such as (oxy)chlorination reactions with chlorine gas (Cl₂) to produce phosphorus chloride intermediates (PCl₃/PCl₅/POCl₃), or disproportionation reactions under acidic or basic conditions to generate phosphine gas (PH₃). The desired OPCs are then obtained by functionalisation of these intermediates with nucleophiles or by hydrophosphination reactions of unsaturated organic compounds, respectively (Scheme 1a).^1–6 As a result, significant efforts have been made to develop alternative strategies for the functionalisation of P₄ that avoid the use of such hazardous reactants and intermediates.^7–11

Scheme 1 (a) Current state-of-the-art routes for the transformation of P₄ into P₁ products. (b) Previously reported hydrostannylation of P₄ and subsequent treatment with electrophiles to afford P₁ compounds directly from P₄. (c) Hydroelemention of P₄ with lighter main group hydrides, reported herein. E⁺ represents a generic electrophile.

Notably, in recent years various strategies have been reported for converting P₄directly into one or more P₁ products.^12–14 These remain at early stages of development, but include the degradation of P₄ by divalent silicon species,¹⁵ photocatalytic reactions,^16,17 electrochemical degradation,¹⁸ and oxidative ‘onioation’ of P₄.¹⁹ Among these new approaches, our group recently reported a simple method that can convert both white and red phosphorus (P_red) into stannyl phosphines (Bu₃Sn)_xPH_3−x (x = 0–3) using the ‘classical’ radical reagent Bu₃SnH and initiation by light or a chemical radical initiator such as azobis(isobutyronitrile) (AIBN).^20,21 The resulting stannylphosphine mixture serves as a P³⁻ source, affording industrially relevant monophosphorus compounds upon treatment with various electrophiles in a ‘one-pot’ protocol (Scheme 1b). Moreover, simple procedures have been developed for the closed-loop recycling of the tin hydride reagent, thus minimising organotin waste.^20,21

Unfortunately, the use of most common organotin compounds raises fundamental concerns around toxicity that cannot be fully mitigated, even by these recycling strategies (although some much less toxic organotin derivatives are known, albeit typically at the cost of significantly higher molecular weight).²² The replacement of R₃SnH with more benign options has long been an important goal for synthetic free-radical chemistry, and many alternative reagents have been proposed for a variety of other chemical transformations.^23–26 Particularly appealing in the context of P₄ reduction are lighter group 14 hydrides, which in other contexts are known to undergo reactions analogous to those of R₃SnH, although they display an intrinsically lower propensity for homolysis of the element-hydrogen bonds.²⁷

Here, we present the first examples of hydrogermylation and hydrosilylation of P₄. Utilizing LED light irradiation, these reactions produce mixtures of germanyl- and silylphosphanes that can be transformed into valuable P₁ products via simple one-pot procedures (see Scheme 1c). Our findings significantly expand the application of hydroelementation as a general method for synthesizing useful organophosphorus compounds in a single reaction step from white phosphorus.

Results and discussion

Inspired by the practical and conceptual simplicity of the hydrostannylation of P₄, we anticipated that this radical-based activation of P₄ could be extended to the use of lighter R₃EH (E = Ge or Si). To evaluate the viability of such a process, we performed a computational investigation at the outset of this project. DFT studies at the PBE-D3(BJ)/def2-TZVP level of theory were focused on the first P–P bond cleavage step of the reaction of P₄ with truncated model radicals Me₃E˙ (E = Ge, Si, Sn) and the subsequent hydrogen atom abstraction (HAT) step (Fig. 1a). The HAT step was found to be rate limiting for Me₃SnH.²⁸ Notably, while the addition of Me₃Sn˙ to P₄ is energetically uphill (2.2 kcal mol⁻¹), the same process is barrierless and downhill for both other Me₃E˙ (E = Ge, Si), forming the ‘butterfly’ P₄ radical intermediates (Me₃E)P₄˙. The subsequent HAT step to (Me₃E)P₄˙ from another equivalent of Me₃EH then proceeds over activation barriers that are significantly higher for Si and Ge than for the heavier analogue Sn (23.4, 18.7 and 10.8 kcal mol⁻¹, respectively) consistent with expected differences in the E–H bond strengths (Fig. 1b, for calculated bond dissociation energies see ESI, Table S1†).^24,27,29,30 Nevertheless, all of these barriers should be easily accessible even at room temperature, suggesting that the desired hydro-elementation should be feasible. An analysis of the influence of the substituents in hydrides R₃EH (E = Si, R = SiMe₃, Ph, Me; E = Ge, R = Ph, Me) revealed the expected trend. The lowest barrier for the hydroelementation of the first P–P bond was observed for (Me₃Si)₃SiH. The barriers for Me₃EH are generally higher than for Ph₃EH (see Fig. S2, ESI†).

Fig. 1 (a) Model hydroelementation of the first P–P bond in P₄. (b) Calculated mechanism, via attack of [E]˙ and subsequent HAT (relative free energies in kcal mol⁻¹). [E] = Me₃Si, Me₃Ge or Me₃Sn. For simplicity, stereochemistry is not shown.

Hydrogermylation of P₄

Encouraged by the computational results, we began experimentally with investigation of the hydrogermylation of P₄. Notably, despite the stronger Ge–H bond, the commercially available organogermanes Bu₃GeH and Ph₃GeH have both been used as alternatives to replace organotin compounds in various radical reactions.^23,25,31 We first tested the reactivity of Bu₃GeH towards P₄, under similar conditions to those used for Bu₃SnH.²⁰ Unfortunately, in sharp contrast to the efficient reaction of P₄ with R₃SnH, the ³¹P{¹H} NMR spectra of the reaction mixtures mainly showed unconsumed P₄ at −521 ppm (for full details see ESI, Section 3.1†). However, when Ph₃GeH, which has a weaker Ge–H bond (see ESI, Table S1†),²⁹ was used instead, additional minor resonances were detected at −240.5 and −248.0 ppm. Markedly, significant amounts of orange precipitate had formed at the end of the reactions with both Bu₃GeH and Ph₃GeH, suggesting the formation of red phosphorus or other insoluble polyphosphorus compounds (presumably (R₃Ge)_xP_yH_z; more detailed characterisation is hampered by the insolubility of these species).

Further investigations revealed that longer reaction times and irradiation with near UV LED light (365 nm) favoured the conversion of P₄ into the new species (Scheme 2a).‡ These can be assigned as (Ph₃Ge)₂PH (1; −240.5 ppm) and Ph₃GePH₂ (2; −248.0 ppm), both by analogy to the major products observed in the previous hydrostannylation reactions and by comparison with the chemical shifts reported for related Ge-substituted phosphines.^20,32 This assignment is further supported by ¹H-coupled ³¹P NMR spectra, where these resonances appear as a doublet (¹J(³¹P–¹H) = 185 Hz) and a triplet (¹J(³¹P–¹H) = 181 Hz), respectively (Scheme 2b). Corresponding doublets arising from coupling between ¹H and ³¹P nuclei could also be observed in the ¹H NMR spectra (see ESI, Fig. S7†). A single-crystal X-ray diffraction study confirmed the identity of 1, whose structure is analogous to the recently reported (TerMe₂Sn)₂PH, which is similarly a product of P₄ hydroelementation (Ter = 2,6-Mes₂C₆H₃).²⁸ An additional, minor resonance was always also observed during these investigations, appearing as a small singlet at −125.8 ppm in the ³¹P{¹H} NMR spectra. This signal splits into a doublet in the ¹H-coupled ³¹P NMR spectrum (¹J(³¹P–¹H) = 199 Hz) and is attributed to Ph₃GeP(H)Ph (3) forming as a minor side product (Scheme 2b).§³³ Note that unlike with Bu₃SnH, no significant tertiary phosphine (i.e. (Ph₃Ge)₃P) or PH₃ products were observed.

Scheme 2 (a) Hydrogermylation of P₄ with Ph₃GeH promoted by near-UV LED irradiation (365 nm). (b) Insets of the ³¹P NMR spectrum for the reaction of P₄ (0.01 mmol) with Ph₃GeH (3 equiv. per P atom) driven by 365 nm LED irradiation. (c) Single-crystal XRD structure of [Ge]₂PH (1). [Ge] = Ph₃Ge. Thermal ellipsoids are shown at 50%. H atoms, except for the one bound directly to P, are omitted for clarity. C atoms are shown in grey, H is white, P in orange, and Ge in dark green.

Under these reaction conditions, the conversion to the main products (1 and 2) was still relatively limited (<54%), likely due to the continuing formation of the insoluble (Ph₃Ge)_xP_yH_z compounds noted above. Gratifyingly, however, using an excess of Ph₃GeH and more concentrated reaction mixtures resulted in significantly improved reaction outcomes with conversions up to 87% (for 1 and 2; 95% including 3), reaction times reduced to 24 hours (see Scheme 3a and ESI, Table S2†), and clear yellowish solutions with no observable precipitates at the end of the reaction. Orange precipitates still form at the beginning of these reactions but later disappear,¶ which may suggest that the initially formed (Ph₃Ge)_xP_yH_z compounds are in fact still available during the overall reaction course, serving as a source of P atoms in the formation of the final products. A similar degradation process has been described for the reaction of P_red with Bu₃SnH.²¹ Consistent with this suggestion, reactions of P_red with Ph₃GeH gave the same hydrogermanylphosphines 1 and 2, albeit with reduced conversion (39%, see ESI, Section 3.4†). Nevertheless, this provides a proof-of-principle confirmation that bench-stable P_red can also be functionalised using germanium hydrides.

Scheme 3 (a) Optimised conditions for the hydrogermylation of P₄ promoted by near-UV irradiation (365 nm). (b) Proposed radical chain mechanism for P₄ hydrogermylation, where [P]–[P] represents a generic P–P bond.

We propose that the hydrogermylation of P₄ follows a radical chain mechanism analogous to the hydrostannylation of P₄ and P_red (Scheme 3b).|| This is also consistent with the calculations noted above. Nevertheless, the observation of significant formation of insoluble (R₃E)_xP_yH_z is a clear point of contrast between R₃GeH and R₃SnH.

To understand why this difference might occur, we considered the aggregation of P₄ to form larger (R₃E)_xP_yH_z moieties computationally. For simplicity, calculations were limited to a single representative step: addition of further P₄ to the initial radical intermediate (Me₃E)P₄˙, which was calculated for all three model systems, E = Sn/Ge/Si. This step would result initially in a P₈ “double butterfly” radical intermediate ((Me₃E)P₈˙), en route to further aggregation and rearrangement steps (Fig. 2). Notably, for all three systems, this step was found to be less energetically demanding than the calculated HAT step that would lead to hydroelementation (cf.Fig. 1b), confirming that aggregation to larger (R₃E)_xP_yH_z fragments is likely to be mechanistically relevant. However, for Me₃Sn˙ the difference in activation barriers is relatively modest (2.2 kcal mol⁻¹; cf.Fig. 1b and 2 or see ESI, Fig. S1†), whereas for Me₃Ge˙ and especially Me₃Si˙ it is much greater (10.6 kcal mol⁻¹ and 13.0 kcal mol⁻¹, respectively). This is mostly due to differences in the HAT step, which is sensitive to changes in the E–H bond strength. In comparison, the identity of the Me₃E substituent has relatively little impact on the barrier to further P₄ addition since it is distant from the active radical site. The results are consistent with the idea that for R₃SnH the desired hydrostannylation is sufficiently kinetically competitive with aggregation to prevent significant formation of large, insoluble (R₃E)_xP_yH_z fragments. In contrast, for R₃GeH and R₃SiH there is a much stronger kinetic preference for aggregation, suggesting that large, insoluble (R₃E)_xP_yH_z aggregates should be more prevalent and that their precipitation should limit the overall reaction rate, consistent with experimental observations. This problem should become more severe as E–H bond strength increases, consistent with the relative performance of Bu₃GeH and Ph₃GeH.

Fig. 2 Calculated aggregation between P₄ and the initial intermediate [E]P₄˙ (relative free energies in kcal mol⁻¹). [E] = Me₃Si, Me₃Ge or Me₃Sn. For simplicity, stereochemistry is not shown.

Functionalization of Ph₃GePH₂ and (Ph₃Ge)₂PH

After optimising the hydrogermylation of P₄, we next focused on the synthetic utilization of the newly formed Ge-substituted phosphines 1 and 2 in a ‘one-pot’ fashion. Notably, simple acidification of the crude mixture with HCl (in 1,4-dioxane) prompted the cleavage of the P–Ge bonds, resulting in the formation of PH₃ (4, Scheme 4, top) with similar efficiency to that achieved using Bu₃SnH (≥50% conversion, cf. ≥56% via Bu₃SnH). Furthermore, treatment of the crude mixture with benzyl bromide (BnBr) or bromoethane (EtBr) in the presence of base yielded the phosphonium salts [Bn₄P]Br (5) and [Et₄P]Br (6), with the former being isolated at preparative scale in 77% yield (Scheme 4, cf. 80% via Bu₃SnH). These results confirm that the P–Ge bonds in the new P₁ intermediate mixture are sufficiently reactive to allow functionalisation into desirable P₁ compounds although they require longer reaction times and higher temperatures than with (Bu₃Sn)_xPH_3−x.²⁰ Reactions of the crude mixture of 1 and 2 with other representative electrophiles that reacted successfully with (Bu₃Sn)_xPH_3−x, such as paraformaldehyde or acyl chlorides, have so far only resulted in low conversions and/or incomplete reactions (see ESI, Section 3.7†), indicating that this equivalence is not universal.

Scheme 4 Functionalisation of the crude mixture of 1 and 2 derived from P₄ directly into useful P₁ products. Equivalents are defined per P atom. (i) P₄ (0.25 equiv.), Ph₃GeH (3 equiv.), PhMe, 365 nm LEDs, RT, 24 h. (ii) Transformation of P₄ into PH₃ from crude mixture (1 and 2): 10 equiv. HCl (4.0 M in 1,4-dioxane) RT, 1 (h). (iii) Preparation of phosphonium salts [R₄P]Br from crude mixture (1 and 2): 10 equiv. RBr (R = Bn or Et), 2.5 equiv. KHMDS, 100 °C, 3d. [Ge] = Ph₃Ge.

Hydrosilylation of P₄

Having established successful hydrogermylation of P₄, focus was then shifted to the generally cheaper and more readily available hydrosilanes, R₃SiH (cf. the more specialised R₃GeH). Hydrosilanes, however, are known to be poor radical chain reagents due to their relatively strong Si–H bonds.^30,34 Moreover, the calculations discussed above imply that formation of insoluble (R₃Si)_xP_yH_z is likely to be an even greater problem for R₃SiH than for R₃GeH. Indeed, initial reactivity studies confirmed that the combination of R₃SiH (R = Ph, Et) with P₄ under the same conditions previously established for either R₃SnH or R₃GeH leads to either no reaction (see ESI, Section 4.1†) or formation of insoluble (R₃Si)_xP_yH_z compounds only. Even when the hypersilane (Me₃Si)₃SiH, known for its higher reactivity as a radical reagent due to the hyperconjugation effect of the Me₃Si groups,^26,30,35 was used, only trace new resonances (and unreacted P₄) were observed in the ³¹P{¹H} NMR spectra when it was reacted with P₄ (Scheme 5a).

Scheme 5 (a) Reactivity studies of silanes R₃SiH towards P₄ driven by LED irradiation (456 nm or 365 nm). (b) Proposed thiol-assisted hydrosilylation of P₄, where [P]–[P] represents a generic P–P bond.

To overcome similar problems in the radical hydrosilylation of unsaturated C=C bonds, organic chemists have developed the use of simple thiols as hydrosilylation catalysts.^36,37 The origin of this catalysis lies in the poor electronic compatibility between the hydrosilane and C-centred radical (both nucleophilic), and much better electronic match between these species and the sulfanyl radical and thiol (both electrophilic), respectively.³⁸ Given the well-known, broad chemical similarity between P and C, and the fact that P is even slightly more electropositive than C, we speculated that a similar thiol-assisted reaction could facilitate the key HAT process and enable the direct hydrosilylation of P₄ to (R₃Si)_nPH_3−n, as outlined in Scheme 5b.

Gratifyingly, the reaction of (Me₃Si)₃SiH and P₄ in the presence of an excess of a thiol such as iPr₃SiSH led to the full consumption of P₄ and the formation of a mixture of Si/H-substituted monophosphorus products (Scheme 6), which appeared as singlets between ca. −210 and −280 ppm in the ³¹P{¹H} NMR spectrum. The main species formed were identified based on the ³¹P{¹H} and ³¹P NMR spectra (Scheme 6b and c, also see ESI, Fig. S30 and S31†).^39,40 The major product was the primary phosphine [(Me₃Si)₃Si]PH₂ (7, −266.5 ppm), followed by [(Me₃Si)₃Si]P(H)SiMe₃ (8, −257.2 ppm), Me₃SiPH₂ (9, −236.5 ppm) and (Me₃Si)₂PH (10, −236.3 ppm), which contain Me₃Si groups directly bound to the phosphorus atom, indicating Si–Si bond cleavage within the hypersilane motif. Whether this cleavage occurs primarily before or after the hydrosilylation is presently unclear.⁴¹ Satisfyingly, upon optimisation, hydrosilylation via this procedure yielded the major reduced P₁ products in almost 80% combined yield (Scheme 6d, also see ESI, Table S6†). The hydrosilylation of P₄ was also achieved with other thiol derivatives, RSH (R = Cy, Ad, Ph, 4-MePh), and even with 1,4-cyclohexadiene (1,4-CHD; see ESI, Table S7 and Fig. S32–37†).^42,43 Optimisation studies revealed that the use of substoichiometric amounts of the hydrogen atom donor was detrimental to overall conversions. Additionally, other, less activated silanes R₃SiH (R = Ph, Et, Me₃SiO) were found to be ineffective, even when an excess of additive was used (see ESI, Table S8†).

Scheme 6 (a) Hydrosilylation of P₄ with (Me₃Si)₃SiH in the presence of a hydrogen atom transfer donor, promoted by LED irradiation. Insets of the ³¹P{¹H} (b) and ³¹P (c) NMR spectra for the reaction of P₄ with (Me₃Si)₃SiH (3.5 equiv. per P atom) and iPr₃SiSH (3.5 equiv. per P atom) in hexane and driven by 365 nm LED irradiation, for the full version of the spectra see ESI, Fig. S30 and S31.† (d) Optimised conditions for the hydrosilylation of P₄ in the presence of iPr₃SiSH, promoted by near-UV irradiation.

Functionalization of crude [Si]_nPH_3−n mixture

Finally, we examined the reactivity of the crude phosphine mixture obtained from the hydrosilylation of P₄ in the presence of a HAT donor to access monophosphosphorus compounds in ‘one-pot’ transformations. Simple acidification with HCl (in 1,4-dioxane) resulted in the efficient cleavage of the P–Si bonds and the selective formation of PH₃ (4) with good conversion (≥64%; Scheme 7, top). Additionally, various phosphonium salts including [Bn₄P]Br (5, isolated in 70% yield), [Et₄P]Br (6, 54% conversion), and [Bu₄P]Br (11, 40% conversion, marketed commercially as CYPHOS 442W) were also accessible directly from P₄ after reaction of the [Si]_nPH_3−n mixture with the corresponding alkyl bromides and base (Scheme 7, middle). Furthermore, treatment of the hydrosilylphosphine mixture with paraformaldehyde in EtOH resulted in exclusive formation of the industrially relevant phosphonium salt THPC (tetrakis(hydroxymethyl)phosphonium chloride, 12, 48% conversion) after quenching with HCl (Scheme 7, bottom). Although these new reactions generally require more forcing conditions than when starting from (Bu₃Sn)_xPH_3−x, they demonstrate that the [Si]_nPH_3−n mixture has a similar potential to act as an efficient P₁ precursor and P³⁻ synthon.

Scheme 7 Functionalisation of crude [Si]_nPH_3−n mixture into useful P₁ products directly from P₄. Equivalents are defined per P atom. (i) P₄ (0.25 equiv.), (Me₃Si)₃SiH (2.25 equiv.), iPr₃SiSH or 1,4-CHD (2.25 equiv.), PhMe, 365 nm LEDs, RT, 2d. (ii) One-pot, selective transformation of P₄ into PH₃ from crude [Si]_nPH_3−n: 10 equiv. HCl (4.0 M in 1,4-dioxane) RT, 1 (h). (iii) Preparation of phosphonium salts [R₄P]Br from crude [Si]_nPH_3−n: 10 equiv. RBr (R = Bn, Et or Bu), 2.5 equiv. KHMDS, 100 °C, 3d. (iv) preparation of THPC from crude [Si]_nPH_3−n: EtOH, 12.5 equiv. paraformaldehyde, 50 °C, 2d, then 10 equiv. HCl (4.0 M in 1,4-dioxane), RT, 2h. [Si] = (Me₃Si)₃Si or Me₃Si.

Conclusions

We have described herein our studies into the fundamental reactivity of hydrogermanes and hydrosilanes towards P₄, as part of our mission to overcome safety concerns associated with the use of tributyltin compounds. Satisfyingly, the hydroelementation of P₄ is successful with both families of lighter homologues R₃EH. While our computational and experimental studies both reveal relevant differences in reactivity (which are related to the differing E–H bond strengths), these can be overcome through rational optimisation of the reaction conditions. Furthermore, these hydroelementation reactions yield similarly functionalisable [E]_nPH_3−n mixtures which can serve as P₁ precursors for useful organophosphorus compounds in a ‘one-pot’ reaction. These results clearly show that the desired P₄ functionalisation is not limited to organotin derivatives and is in fact generalisable. Extrapolation from these results implies that extension to other known R₃SnH alternatives should be possible, many of which are known to compete well with R₃SnH in other contexts. Efforts in this direction are currently ongoing in our laboratories. Of particular interest to us is the extension of these results to more robust R₃SnH surrogates, to facilitate their recycling and reuse and/or catalytic application (cf. Bu₃SnH).²⁰ Besides the obvious efficiency benefits, such (pseudo)catalytic approaches would also remove the need for stoichiometric, ‘upstream’ Cl₂ use.**

Author contributions

JC and DJS developed the hydrosilylation and hydrogermylation procedures. MS performed preliminary experiments with different R₃GeH. JC performed the mechanistic calculations, developed the functionalisation of the crude [R₃E]_nPH_3−n mixtures (E = Si, Ge) and the procedures for product isolation and characterisation. DJS and RW conceived, oversaw, and directed the project. JC and DJS prepared the manuscript, with input from all authors.

Data availability

The data supporting the findings of this study are available within the article and its ESI.† Crystallographic information for 1 has been deposited at the CCDC under 2408662.

Conflicts of interest

A patent covering all of the results described herein has been filed (as of 13 February 2020) by the University of Regensburg (EP 20,157,197.3; inventors, D. J. S. and R. W.). The authors declare no other competing interests.

Acknowledgements

We thank Maria K. Uttendorfer for the acquisition and processing of the XRD data. DJS would like to thank the Alexander von Humboldt foundation for the award of a postdoctoral fellowship and the EPSRC for an Early Career Fellowship (EP/V056069/1). RW would like to thank the DFG (WO 1496/12-1, project number 548830090) for financial support.

References

1. D. E. C. Corbridge, Phosphorus. Chemistry, Biochemistry and Technology, Elsevier, 2000.
2. J. Svara, N. Weferling and T. Hofmann, Phosphorus Compounds, Organic, Ullmann's Encycl. Ind. Chem., 2012, 27, 20–50.
3. H. Diskowski and T. Hofmann, Phosphorus, Ullmann's Encycl. Ind. Chem., 2012, 26, 725–746.
4. N. Weferling, S. M. Zhang and C. H. Chiang, Commercial Organophosphorus Chemicals: Status and New Developments, Procedia Eng., 2016, 138, 291–301.
5. G. Bettermann, W. Krause, G. Riess and T. Hofmann, Phosphorus Compounds, Inorganic, Ullmann's Encycl. Ind. Chem., 2012, 27, 1–18.
6. A. R. Jupp, S. Beijer, G. C. Narain, W. Schipper and J. C. Slootweg, Phosphorus recovery and recycling-closing the loop, Chem. Soc. Rev., 2021, 50, 87–101.
7. J. E. Borger, A. W. Ehlers, J. C. Slootweg and K. Lammertsma, Functionalization of P₄ through Direct P–C Bond Formation, Chem. – Eur. J., 2017, 23, 11738–11746.
8. N. K. Gusarova and B. A. Trofimov, Organophosphorus chemistry based on elemental phosphorus: advances and horizons, Russ. Chem. Rev., 2020, 89, 225–249.
9. B. M. Cossairt, N. A. Piro and C. C. Cummins, Early-Transition-Metal-Mediated Activation and Transformation of White Phosphorus, Chem. Rev., 2010, 110, 4164–4177.
10. M. Caporali, L. Gonsalvi, A. Rossin and M. Peruzzini, P₄ Activation by Late-Transition Metal Complexes, Chem. Rev., 2010, 110, 4178–4235.
11. M. Scheer, G. Balázs and A. Seitz, P₄ Activation by Main Group Elements and Compounds, Chem. Rev., 2010, 110, 4236–4256.
12. D. J. Scott, Recent Breakthroughs in P₄ Chemistry: Towards Practical, Direct Transformations into P₁ Compounds, Angew. Chem., Int. Ed., 2022, 61, e2022050.
13. J. Hu, W. Liu and W. X. Zhang, Direct functionalization of white phosphorus by organolithium reagents to organophosphorus compounds, Phosphorus, Sulfur Silicon Relat. Elem., 2022, 197, 398–407.
14. Y. Liu, X. Chen and B. Yu, Sustainable Photo- and Electrochemical Transformation of White Phosphorous (P₄) into P₁ Organo-Compounds, Chem. – Eur. J., 2023, 29, e202302142.
15. Y. Wang, T. Szilvási, S. Yao and M. Driess, A bis(silylene)-stabilized diphosphorus compound and its reactivity as a monophosphorus anion transfer reagent, Nat. Chem., 2020, 12, 801–807.
16. U. Lennert, P. B. Arockiam, V. Streitferdt, D. J. Scott, C. Rödl, R. M. Gschwind and R. Wolf, Direct catalytic transformation of white phosphorus into arylphosphines and phosphonium salts, Nat. Catal., 2019, 2, 1101–1106.
17. P. B. Arockiam, U. Lennert, C. Graf, R. Rothfelder, D. J. Scott, T. G. Fischer, K. Zeitler and R. Wolf, Versatile Visible-Light-Driven Synthesis of Asymmetrical Phosphines and Phosphonium Salts, Chem. – Eur. J., 2020, 26, 16374–16382.
18. Y. Mei, Z. Yan and L. L. Liu, Facile Synthesis of the Dicyanophosphide Anion via Electrochemical Activation of White Phosphorus: An Avenue to Organophosphorus Compounds, J. Am. Chem. Soc., 2022, 144, 1517–1522.
19. M. Donath, K. Schwedtmann, T. Schneider, F. Hennersdorf, A. Bauzá, A. Frontera and J. J. Weigand, Direct conversion of white phosphorus to versatile phosphorus transfer reagents via oxidative onioation, Nat. Chem., 2022, 14, 384–391.
20. D. J. Scott, J. Cammarata, M. Schimpf and R. Wolf, Synthesis of monophosphines directly from white phosphorus, Nat. Chem., 2021, 13, 458–464.
21. J. Cammarata, D. J. Scott and R. Wolf, Hydrostannylation of Red Phosphorus: A Convenient Route to Monophosphines, Chem. – Eur. J., 2022, 28, e2022024.
22. N. J. Snoeij, A. A. J. van Iersel, A. H. Penninks and W. Seinen, Toxicity of triorganotin compounds: Comparative in vivo studies with a series of trialkyltin compounds and triphenyltin chloride in male rats, Toxicol. Appl. Pharmacol., 1985, 81, 274–286.
23. P. A. Baguley and J. C. Walton, Flight from the Tyranny of Tin: The Quest for Practical Radical Sources Free from Metal Encumbrances, Angew. Chem., Int. Ed., 1998, 37, 3072–3082.
24. B. C. Gilbert and A. F. Parsons, The use of free radical initiators bearing metal-metal, metal-hydrogen and non-metal-hydrogen bonds in synthesis, J. Chem. Soc., Perkin Trans. 1, 2002, 2, 367–387.
25. W. R. Bowman, S. L. Krintel and M. B. Schilling, Tributylgermanium hydride as a replacement for tributyltin hydride in radical reactions, Org. Biomol. Chem., 2004, 2, 585–592.
26. C. Chatgilialoglu, (Me₃Si)₃SiH: Twenty Years After Its Discovery as a Radical-Based Reducing Agent, Chem. – Eur. J., 2008, 14, 2310–2320.
27. C. Chatgilialoglu and M. Newcomb, Hydrogen Donor Abilities of the Group 14 Hydrides, Adv. Organomet. Chem., 1999, 44, 67–112 For comparison of the E–H bond strength and the rate constants for H-abstraction with carbon-centred radicals (e.g. alkyl radicals) from R₃EH (E = Si, Ge, Sn), also see refs: 23, 28 and 29.
28. J. Cammarata, F. F. Westermair, P. Coburger, D. Duvinage, M. Janssen, M. K. Uttendorfer, J. Beckmann, R. M. Gschwind, R. Wolf and D. J. Scott, Unravelling White Phosphorus: Experimental and Computational Studies Reveal the Mechanisms of P₄ Hydrostannylation, Angew. Chem., Int. Ed., 2024, 63, e202408423.
29. C. Chatgilialoglu and M. Ballestri, Hydrogen Donor Abilities of Germanium Hydrides, Organometallics, 1999, 18, 2395–2397.
30. C. Chatgilialoglu, Structural and Chemical Properties of Silyl Radicals, Chem. Rev., 1995, 95, 1229–1251.
31. H. Yorimitsu and K. Oshima, Recent advances in the use of tri(2-furyl)germane, triphenylgermane and their derivatives in organic synthesis, Inorg. Chem. Commun., 2005, 8, 131–142.
32. D. M. Schubert and A. D. Norman, Synthesis of (Trimethylsilyl)- and (Trimethylgermyl)allylphosphine, Inorg. Chem., 1985, 24, 1107–1109.
33. J. Escudié, C. Couret and J. Satgé, Reactions d'insertion du phénylphosphinidène dans diverses liaisons germanium–hétéroélément, Recl. Trav. Chim. Pays-Bas, 1979, 98, 461–466.
34. C. Chatgilialoglu, Organosilanes in Radical Chemistry, John Wiley & Sons Ltd, 2004.
35. C. Chatgilialoglu and J. Lalevée, Recent applications of the (TMS)₃SiH radical-based reagent, Molecules, 2012, 17, 527–555.
36. H. S. Dang and B. P. Roberts, Polarity-reversal catalysis by thiols of radical-chain hydrosilylation of alkenes, Tetrahedron Lett., 1995, 36, 2875–2878.
37. B. P. Roberts, Polarity-reversal catalysis of hydrogen-atom abstraction reactions: Concepts and applications in organic chemistry, Chem. Soc. Rev., 1999, 28, 25–35.
38. J. J. A. Garwood, A. D. Chen and D. A. Nagib, Radical Polarity, J. Am. Chem. Soc., 2024, 146, 28034–28059.
39. V. Cappello, J. Baumgartner, A. Dransfeld and K. Hassler, Monophosphanes and diphosphanes with the hypersilyl substituent, Eur. J. Inorg. Chem., 2006, 4589–4599.
40. K. X. Bhattacharyya, S. Dreyfuss, N. Saffon-Merceron and N. Mézailles, P₄ functionalization by hydrides: Direct synthesis of P–H bonds, Chem. Commun., 2016, 52, 5179–5182.
41. M. Mohamed and M. A. Brook, Can. J. Chem., 2000, 78, 1357–1362.
42. I. Chatterjee and M. Oestreich, B(C₆F₅)₃-catalyzed transfer hydrogenation of imines and related heteroarenes using cyclohexa-1,4-dienes as a dihydrogen source, Angew. Chem., Int. Ed., 2015, 54, 1965–1968.
43. H. Bauer, K. Thum, M. Alonso, C. Fischer and S. Harder, Alkene Transfer Hydrogenation with Alkaline-Earth Metal Catalysts, Angew. Chem., Int. Ed., 2019, 58, 4248–4253.

---

Footnotes

† Electronic supplementary information (ESI) available: General experimental procedures, full reaction optimisation, spectroscopic characterisation, computational details, and procedures for isolation of products. CCDC 2408662. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d5qi00869g

‡ Analogous resonances were also observed in the reaction with Bu₃GeH under these conditions, although with lower conversions. For full details see ESI, Section S3.1.†

§ Similar side/decomposition products have been observed in our previous study of the P₄ hydrostannylation system, in the form of R₃SnP(H)R, though in much smaller amounts. The fate of the remaining “Ph₂Ge” fragment whose formation is implied by the presence of the P-bound Ph group is currently unclear.

¶ During the optimisation of the reaction, no signals corresponding to soluble polyphosphorus intermediates were observed in the ³¹P{¹H} NMR spectra.

|| Irradiation with LED light is proposed to induce the formation of an initial R₃Ge˙ or R₃Si˙ radical which initiates the chain reaction. However, the precise details of this initiation are currently unclear and remain under investigation. Proof-of-concept results have also been achieved using chemical radical initators such as AIBN instead of LED irradiation. However, for these systems we have thus far only been able to achieve low conversions (<28% 1 and 2). For full details see ESI, Section S3.5 and S4.7†.

** Synthesis of R₃EH derivatives (E = Si, Ge, Sn) often relies on the use of Cl₂ as a stoichiometric reagent (e.g. to prepare early intermediates such as ECl₄). Thus, stoichiometric use of these reductants to activate P₄ effectively shifts the use of Cl₂ ‘upstream’, rather than eliminating it per se. By contrast, recycling of “R₃EX” byproducts typically does not require additional Cl₂.

---