Oxaphosphiranes: isolable phosphorus-containing epoxide rings

Abstract

Combining different heteroatoms in an epoxide-type ring having Lewis basic and acidic characteristics is challenging as it creates an increasing number of polar bonds and high ring strain energy. The first examples of isolable oxaphosphiranes, i.e., epoxide rings with a phosphorus atom, have been synthesized using a facile and effective protocol starting from [pentacarbonyl{dichloro(trityl)phosphane}molybdenum(0)] (trityl = CPh₃), tert-butyllithium and commercially available fluorinated benzaldehydes. Reactions with various acids and bases will be described. Theoretical results unveil a singlet carbene-like FMO situation at phosphorus in oxaphosphirane, which is used to explain reaction mechanisms.

---

Introduction

Epoxides¹ (I) (Fig. 1), cyclopropanes containing one oxygen instead of a carbon atom in the ring, are versatile reagents,² and their oxidation can serve as a protecting strategy for alkenes. For example, due to their polar C–O bonds and high ring strain energy (RSE), ring opening reactions can be initiated easily, which makes them highly potent monomers in polymer synthesis.³ Owing to the rich chemistry of epoxides, investigating heavier analogues and those possessing a second ring heteroatom has been a subject of great interest. Although small inorganic ring systems have been intensively explored,⁴ examples of heavy analogues of epoxides are still scarce. It is known that the presence of an additional heteroatom in the ring has a significant influence on its stability and reactivity. This is illustrated by a series of main group element-containing epoxides of groups 14–16 also starting with structurally confirmed oxasiliranes (II). The latter were exclusively accessed through the reaction of a bulky dialkylsilylene with a ketone,⁵ but their chemistry is almost unknown.

Fig. 1 Epoxide (I) and heteroatom-containing analogues (II–VI); lines denote ubiquitous substituents.

The latter might be due to the high RSE value (estimated through the additive method based on bond contributions as 33.61 (ref. 6) and 34.22 (ref. 7) kcal mol⁻¹; accurately computed through homodesmotic reactions as 38.70 (ref. 6) kcal mol⁻¹). In contrast, oxaziridines (III) are easily accessed and, hence, well established in organic synthetic chemistry.⁸ For example, oxidizing imines using peroxy acids is one of the common methods to prepare derivatives of III.⁸ Dioxiranes (V) are also available, e.g., as dimesityl derivatives synthesized by photooxidation of an in situ generated carbene in air, as reported by Sander, but the compound is stable only below −20 °C.⁹ Again, including a third row element from group 16 diminishes the thermal stability and, hence, oxathiirane (VI) is known only in argon matrices at low temperature.¹⁰

The search for phosphorus homologues of oxaziridines, i.e., oxaphosphiranes (IV), started with metal complexes such as VII, reported by Mathey et al. in 1990,¹¹ relying on the oxidation of end-on phosphaalkene pentacarbonyltungsten(0) complexes. Despite being a synthetic breakthrough, the multi-step protocol was too tedious to allow for broader studies. Even the subsequent new protocol, reported by Streubel¹² – the trapping of an in situ generated terminal phosphinidene complex – turned out to be only a slight improvement. The real breakthrough came in 2007 when we significantly improved the methodology¹³ by using Li/Cl phosphinidenoid complexes and a great variety of aldehydes and ketones, even bearing other functional groups.¹⁴ In the following decade, numerous transformations became accessible, thus illustrating the great synthetic potential of the oxaphosphirane ring system.¹⁵

However, the masking of the phosphorus lone pair via coordination to a transition metal center precluded the exploration of the chemistry of unligated derivatives. The search for unligated oxaphosphiranes VI started with Ramirez in 1968, when hexafluoroacetone reacted with phosphanes to yield dioxaphospholanes.¹⁶ Later, the pathway was investigated theoretically, thus showing that intermediate σ⁵λ⁵-oxaphosphiranes are likely to be involved.¹⁷ In 1993, Röschenthaler et al. reported the [2 + 1]-cycloaddition of an iminophosphane to hexafluoroacetone, thus forming a σ⁴,λ⁵-oxaphosphirane; it should be noted that no structural confirmation was provided.¹⁸ Very recently, we demonstrated that 1,3-di-tert-butyl-oxaphosphirane can be obtained from a pentacarbonylmolybdenum(0) complex via a thermal reaction with diphenylphosphinoethane (DPPE).¹⁹ The product was obtained exclusively as toluene solution, thus precluding various interesting investigations, and no X-ray structure could be obtained. Therefore, we restarted the search for a novel method to access isolable, unligated oxaphosphiranes that would allow for broad investigations. Furthermore, we hypothesized that the ring carbon atom should possess fluorine-containing substituents, based on the conclusions of a theoretical study.²⁰

Herein, we describe a new facile protocol for oxaphosphirane liberation from metal(0) complexes using N-methylimidazole under thermal conditions. Its molecular solid-state structure is discussed and the bonding was analysed using high-level theoretical investigations. Reactions of the isolated oxaphosphiranes with various reagents such as water (H₂O and D₂O), tetrachloro-ortho-benzoquinone (TOB), an acid (HCl) and a base/nucleophile (an N-heterocyclic carbene) are reported.

Results and discussion

Oxaphosphirane molybdenum complexes 3a and 3b and the oxaphosphirane tungsten complex 3c were prepared according to our protocol^14b,21via in situ formed Li/Cl phosphinidenoid complex 2 ²² using fluorinated benzaldehyde derivatives, resulting in moderate yields (50–65%) (Scheme 1).

Scheme 1 Synthesis of [Mo] (3a and b) and [W] oxaphosphirane (3c) complexes.

The molecular structures of 3b and 3c were confirmed by single crystal X-ray diffraction studies (Fig. 2). In comparison with known oxaphosphirane molybdenum²³ and tungsten²¹ complexes with fluorinated aryl C-substituents, the most important differences can be observed in slightly elongated P–M bonds (P–Mo: 2.4872(14) Å (3b) vs. 2.467(2) Å;²³ P–W: 2.4715(13) Å (3c) vs. 2.4630(11) Å (ref. 21)).

Fig. 2 Single crystal structures of 3b and 3c (50% probability) The hydrogen atoms, except the ones at the ring carbons, are omitted for clarity. 3b: selected bond lengths/Å: P–Mo 2.4872(14), P–O1 1.673(4), P–C1 1.802(5), and O1–C1 1.458(6); selected bond angles/°: O1–P–C1 49.5(2), P–C1–O1 60.7(2), and P–O1–C1 69.9(2); 3c: selected bond lengths/Å: P–W 2.4715(13), P–O1 1.669(4), P–C1 1.790(5), and O1–C1 1.481(7); selected bond angles/°: O1–P–C1 50.5(2), P–C1–O1 60.5(3), and P–O1–C1 69.6(3).

Following previous investigations on ligand decomplexation from molybdenum complexes using DPPE as a chelating ligand under thermal conditions,¹⁹ which were successful for the P-t-Bu substituted oxaphosphirane ligand, we achieved the decomplexation of the oxaphosphirane ligands in 3a and b. Unfortunately, here an inseparable mixture of the oxaphosphiranes and non-identified by-products was obtained.

Therefore, an alternative protocol needed to be developed. Inspired by the decomplexation of 1H-phosphirenes using sequentially elemental iodine and 1-methylimidazole, described by Mathey,²⁴ we decided to examine this for 3a and 3b. Heating solutions only with 4 eq. of 1-methylimidazole resulted in a clean liberation of the oxaphosphiranes 4a and 4b (Scheme 2), having almost identical ³¹P-NMR resonances (−26.0 (4a) and −25.5 ppm (4b)).

Scheme 2 Decomplexation of the oxaphosphirane ligands 4a and b.

Compounds 4a and 4b were fully characterized by NMR, mass spectrometry, IR and elemental analysis. Additional column chromatography (Al₂O₃/ether) at low temperature (0 to −10 °C) can be used for purification, but this results in diminished yields (loss: 15–20%). Suitable single crystals for X-ray diffraction measurements of 4b were obtained (Fig. 3), confirming the molecular structure of an unligated oxaphosphirane having the trans-conformation as in the precursor. The molecular structure of 4b contained only the 2S,3R-enantiomer with the two substituents at P and C in the trans-position as identified by the Flack parameter²⁵ of 0.00(18). In comparison with the crystal structure of 3b, only very small differences could be identified, i.e., smaller O1–P–C10 and P–C1 C2 angles (Table 1).

Fig. 3 Single crystal molecular structure of 4b (50% probability). The hydrogen atoms, except the one at the ring carbon, are omitted for clarity. Selected bond lengths/Å: P–O1 1.682(3), P–C1 1.819(4), and O1–C1 1.458(5); selected bond angles can be found in Table 1.

Table 1 Selected bond angles of 4b and 3b

Bond angles/°	4b	3b
O1–P–C1	49.04(16)	49.5(2)
P–C1–O1°	60.57(19)	60.7(2)
P–O1–C1	70.4(2)	69.9(2)
P–C1 C2	118.0(3)	124.5(4)
C1–P C10	110.74(18)	105.3(2)
O1–P–C10	104.99(16)	108.9(2)

---

It is worth highlighting that the original protocol of Mathey²⁴ was not successful in our study.

Pentacoordinate oxaphosphiranes have been previously reported²⁶ and, most often, assumed to be short-lived intermediates. Recently, we obtained first, albeit weak, evidence in the reaction of the P-t-Bu substituted oxaphosphirane with TOB via VT-³¹P{¹H} NMR spectroscopy at low temperature, but the characterization remained incomplete. As pentacoordinate phosphirenes have been accessed via a reaction with TOB,²⁷ we decided to investigate the reaction of 4a with TOB in CH₂Cl₂ (Scheme 3).

Scheme 3 Synthesis of the spirobisbenzodioxaphosphole 5 and the proposed reaction mechanism involving σ⁵λ⁵-oxaphosphirane 6 and σ³λ³-benzodioxaphospholane 7.

First, the reaction with 1 eq. of TOB was carried out at room temperature, and after 2 h, a 1 : 1 mixture of 4a showed product 5 at 1.7 ppm in the ³¹P{¹H}-NMR spectrum. By subsequently adding 1 eq., a full conversion to 5 was observed. Mass spectrometry showed the isotopic pattern of the molecular ion peak (m/z 765.8 [M]˙⁺ (EI)), indicating a structural motif with eight chlorine atoms, and the exact mass corresponded to the known P-CPh₃ substituted spirobisbenzodioxaphosphole.²⁸ VT-³¹P{¹H} NMR studies from −70 °C to r.t. did not reveal an intense signal that could be safely assigned to the intermediately formed pentacoordinate oxaphosphirane 6. However, at −10 °C, a signal with very low intensity at 208.2 ppm can be assigned to the P-CPh₃ substituted dioxaphospholene 7 ²⁸ (δ³¹P{¹H} = 208.5 ppm). The formation of 7 provided further insight into the reaction mechanism, pointing to a preceding formation of the σ⁵λ⁵-oxaphosphirane 8, followed by a loss of the aldehyde. This is followed by the addition of a second TOB unit to yield the final product 5. To gain a deeper understanding of the reaction mechanism, theoretical studies were performed (vide infra).

Quantum chemical (QC) calculations (CPCM_toluene/PWPB95-D3/def2-QZVPP//CPCM_toluene/PBEh-3c) were performed using the model compound 4a′, where the P-CPh₃ (trityl) substituent was replaced by t-Bu. The two individual steps of the reaction with TOB are very exergonic (ΔΔG ≈ −46 kcal mol⁻¹) and proceed over a rather low barrier (ΔΔG^‡ ≈ 19 kcal mol⁻¹), whereas the loss of the aldehyde unit is moderately exergonic with a slightly higher barrier (Fig. 4).

Fig. 4 Computed (CPCM_toluene/PWPB95-D3/def2-QZVPP//CPCM_toluene/PBEh-3c) Gibbs energy profile for the oxidation of model 4a′ with TOB.

In our recent report on the P-t-Bu substituted oxaphosphirane,¹⁹ obtained as solution only, we observed great moisture sensitivity; the latter is in stark contrast to the case of oxaphosphirane complexes. This abovementioned unusual sensitivity of the free oxaphosphirane prompted us to investigate the reactivity of 4a towards water. Oxaphosphirane 4a was dissolved in THF-d₈ and small amounts of water were added (Scheme 4).

Scheme 4 Investigations of the reactivity of 4a towards H₂O (or D₂O), resulting in the formation of diastereomers 8 (or 10) and the primary phosphane oxide 9 (or 11).

At ambient temperature, no reaction was observed, and after heating to 40 °C, a rather slow conversion to three products, two diastereomers (vide infra) of secondary phosphane oxide 8 (δ³¹P{¹H} = 50.3 and 49.3 ppm) and the P-CPh₃ substituted phosphane oxide 9 (δ³¹P{¹H} = 19.9 ppm), were observed; at 60 °C, the reaction was complete in 4 h. The product mixture was investigated by NMR and IR spectroscopy, and the molecular structure of 8 was confirmed by an X-ray diffraction study. The elemental cell contained both diastereomers and two molecules of CH₂Cl₂. The crystal structure revealed intra- and intermolecular interactions between a total of three formula units of 8 and the solvent, as depicted in Fig. S39 (ESI†). The bond lengths and bond angles are in accordance with those of a literature known P-Mes substituted secondary phosphane oxide.²⁹ To gain further insight into the origin of the hydrogen atoms on the phosphorus center, the reaction was repeated with D₂O (Scheme 4). This resulted in the formation of the deuterated phosphane oxide diastereomers 10 (δ³¹P{¹H} = 50.2 ppm, t*, ¹J_P,D = 73.7 Hz; 49.5 ppm, t*, ¹J_P,D = 74.0 Hz) and 11 (δ³¹P{¹H} = 19.5 ppm, quint*, ¹J_P,D = 74.4 Hz), proving the origin of the hydrogen atoms from the reagent water. The molecular structure of 10 was also confirmed by X-ray crystallography (Fig. 5), and the structure revealed, in contrast to the molecular structure of 8, the formation of a bridging dimer between the PO and the OD groups. The distance between the oxygen atoms of the PO group and the OD group of another formula unit (d(O1⋯O2) = 2.709 Å) is in good accordance with the bridging dimer of a literature known tertiary phosphane oxide³⁰ (d(O1⋯O2) = 2.682(3) Å).

Fig. 5 D-bridging dimer of the molecular structure of 10 (enantiomers 1R,2S and 1S,2R) with co-crystallized CH₂Cl₂ in the single crystal (50% probability level). Hydrogen atoms, except deuterium ones and the hydrogen atom at C1, are omitted for clarity. Selected bond lengths/Å: O1–O2 2.709, P–O1 1.4895(12), P–C1 1.8774(18), and P–C8 1.8746(17); selected bond angles/°: O1–P–C1 111.08(7) and 116.34(7).

Quantum chemical calculations at the working level of theory (THF was employed for the continuum solvation model) using the same simplified compound 4a′ revealed that initial ring opening can proceed through a nucleophilic attack on either the C or P ring atoms, in agreement with the reported behaviour for P-complexed oxaphosphiranes.^14c Similar to the previous report, the attack by two, three and five explicit water molecules was explored. The approach of only one H₂O molecule was not inspected, as the very similar methanolysis reaction of the parent oxaphosphirane with one MeOH molecule was reported to occur over a rather high-lying transition state.³¹ In the case of two H₂O molecules (see Fig. S51†), only ring opening by attack on C was observed, leading to zwitterion l-12, which then readily undergoes an internal acid–base reaction (most likely occurring intermolecularly through a chain of water molecules) to yield the stable phosphinic acid 13 (Scheme 5). When there is excess water, the energy barrier of this ring opening (ΔΔG^‡ = 31.0 kcal mol⁻¹) can be considered as the energy difference between the TS (transition state) and the initial van der Waals complex 4a′·2H₂O. As the starting oxaphosphirane is racemic, the 2R,3S (unlike or ‘u’) configuration was chosen, in agreement with the observed trans-arrangement observed in the single crystal molecular structure of 4b. Ring opening proceeds with inversion only at the C atom, thus resulting in an l-13 diastereomeric pair. The use of three H₂O molecules (see Fig. S51†) favors attack on P, affording the other most stable diastereomeric pair u-13 (ΔΔG_l/u = 1.1 kcal mol⁻¹) with no inversion at either P or C (i.e., the diastereomer of that depicted in Scheme 5) and over a lower barrier (ΔΔG^‡ = 27.0 kcal mol⁻¹). When a chain of five H₂O molecules is included, even lower energy barriers are obtained, with preferential P (ΔΔG^‡ = 18.8 kcal mol⁻¹) over C attack (ΔΔG^‡ = 28.5 kcal mol⁻¹) (Fig. 6), forming u-13 and l-13, respectively. Despite the significantly most favored P-attack in model u-4a′, the larger steric demand of the P-trityl group in 4a should favor C-attack to some extent. The resulting l-13 intermediate was thus used to describe the rest of the mechanistic sequence (Scheme 5 and Fig. 6).

Scheme 5 Proposed mechanism for the reaction of model 4a′ with H₂O. Stereochemistry corresponds to the C-attack of H₂O. All species except 9′ are racemates.

Fig. 6 Computed (CPCM_THF/PWPB95-D3/def2-QZVPP//CPCM_THF/PBEh-3c) Gibbs energy profile for the hydrolysis of model 4a′.

The slightly exergonic isomerization of the phosphinic acid moiety [>P–OH] in l-13 to the more stable P-oxide [>P(H)=O] group results in the final (model) product l-8′, through a rather high barrier (ΔΔG^‡ = 50.5 kcal mol⁻¹) O-to-P [1,2]H shift (see Fig. 6). When this O-to-P hydrogen atom transfer occurs with the assistance of one or two water molecules, the energy barrier decreases significantly (ΔΔG^‡ = 28.5 and 26.3 kcal mol⁻¹, respectively) (Fig. 6). Similarly, starting from u-13, the slightly less stable P-oxide product u-8′ (ΔΔG_l/u = 0.9 kcal mol⁻¹) is obtained (the energy profile for this diastereomeric transformation is not shown). The occurrence of two diastereomers of the real P-CPh₃ substituted species 8, as shown in the NMR spectra, is due to the occurrence of competitive C- and P-attack hydrolysis pathways, leading to l- and u-8, respectively.

The intermediate l-13 undergoes transprotonation to yield another more stable betaine l-14 that very readily experiences C–C bond cleavage, furnishing 3,5-difluorobenzaldehyde (Ar^FCHO) and tert-butyl phosphinic acid (R-15). Interestingly, the latter can be formally considered as a product resulting from the addition of H₂O to a phosphinidene unit; the latter is not formed during hydrolysis, but is present/masked in the oxaphosphirane 4a′. This compound 15 isomerizes to the most stable phosphane oxide 9′, either intramolecularly, with a rather high TS (ΔΔG^‡ = 61.3 kcal mol⁻¹), or through one (ΔΔG^‡ = 29.8 kcal mol⁻¹), two (ΔΔG^‡ = 30.8 kcal mol⁻¹) (see Fig. 6) or three (ΔΔG^‡ = 19.6 kcal mol⁻¹) water molecules (Fig. 6). The same outcome results from starting with diastereomeric u-13via the slightly less stable betaine u-14 (ΔΔG_l/u = 1.0 kcal mol⁻¹) (energy profile not shown).

The theory-based indication of a nucleophilic attack of the oxygen atom of water prompted us to study the reactivity towards other ambident reagents. Thus, the reaction of oxaphosphirane 4a with hydrochloric acid in diethyl ether was investigated.^14a,c,32 In contrast to the expected formation of an acyclic product stemming from cleavage of the P–O bond, which was observed before for other oxaphosphirane complexes, the (complete) ring cleavage product 16 was obtained (Scheme 6); this product had been obtained earlier by Plack using a classical reduction route.³³

Scheme 6 Synthesis of the P-CPh₃ chlorophosphane 16.

The ³¹P-NMR spectrum displayed a doublet with a chemical shift of 44.9 ppm and a coupling constant of ¹J_P,H = 189.9 Hz. The signal shoulder (ratio 1 : 3) supports the directly bound chlorine atom to phosphorus, i.e., showing the overlap of the two signals of the two P–Cl isotopomers.

The calculated energy profile for the model reaction of 4a′ (Fig. 7) shows that the first step is the slightly exergonic 1,1-addition of HCl to the P-center, thus exhibiting singlet phosphinidene-like reactivity, furnishing the σ⁵λ⁵-oxaphosphirane intermediate 17. This species displays a proper trigonal bipyramidal arrangement at P with elongated axial P–O (d = 1.740 Å, MBO 0.795) and P–Cl (d = 2.243 Å, MBO 0.685) bonds (O–P–Cl angle – 153.0°). Driven by the relief of ring strain energy, this intermediate undergoes a low barrier exergonic chelotropic [2 + 1]-cycloreversion, leading to the final (model) chlorophosphane 16′via the loss of the aldehyde unit (Fig. 7).

Fig. 7 Computed (CPCM_CH2Cl2/PWPB95-D3/def2-QZVPP//CPCM_CH2Cl2/PBEh-3c) Gibbs energy profile for the reaction of model 4a′ with HCl using Me₂O as a model for Et₂O.

Since the reaction of the oxaphosphirane 4a with the strong acid led to the elimination of the aldehyde unit, its reactivity towards a nucleophile – being also basic at the same time – was explored, hoping to obtain a ring-opened zwitterionic product. When 1,3,4,5-tetramethyl-imidazol-2-ylidene (IMe₄) was allowed to react with the title compound, the aldehyde was eliminated, again. To obtain a more robust product having a greater tendency to crystallize, the primarily formed IMe₄-P-CPh₃18 was treated with W(CO)₅(NCMe) and yielded W(CO)₅ complex 19 (Scheme 7), which was known before. A ³¹P-NMR signal was observed at −37.7 ppm with a ¹J_P–W coupling constant magnitude of 114.08 Hz. The molecular structure of compound 19 in the solid state (Fig. 8) shows the (typical) geometrical features of a (zwitterionic) IMe₄-stabilized phosphinidene complex.

Scheme 7 Reaction of 4a with IMe₄ and W(CO)₅(NCMe).

Fig. 8 Solid state structure of compound 19 (50% probability level). Hydrogen atoms are omitted for clarity. Selected bond lengths/Å: C8–P 1.955(17), C1–P 1.850(18), P–W 2.615(5), W–C27 1.983(2), and C27–O1 1.154(3); selected bond angles/°: W–P–C8 122.37(5), W–P–C1 108.98(6), and C1–P–C8 106.50(8).

The (theoretical) energy profile for this reaction was obtained using the very simplified model 4a′′ featuring a methyl substituent at P and 1,3-dimethylimidazol-2-ylidene (here designated as NHC^Me₂) (Fig. 9). The initially formed encounter complex 4a′′·NHC undergoes almost barrierless exergonic chelotropic cycloreversion with the loss of the aldehyde unit and affording the corresponding (model) phosphinidene-NHC adduct 18′′.

Fig. 9 Computed (CPCM_toluene/PWPB95-D3/def2-QZVPP//CPCM_toluene/PBEh-3c) Gibbs energy profile for the reaction of model 4a′′ with NHC^Me₂.

The formation of the encounter complex can be understood in terms of pnictogen bonding, where the approaching nucleophile donates electron density into the “σ-hole”, which conceptually corresponds to the σ*(P–O) orbital. This is visualized by the LUMO having antibonding characteristics for both the P–C and the P–O endocyclic bonds (Fig. 10a). The corresponding bonding interaction between the P and the CO fragments corresponds to the HOMO, thus pointing to singlet phosphinidene characteristics of the R–P fragment, which locates the P lone pair at HOMO−2. By populating the LUMO, the P–C and P–O bonds are weakened, facilitating the detachment of the aldehyde. Therefore, the removal of the carbonyl fragment from oxaphosphiranes upon nucleophilic attack on phosphorus could be potentially used as an important synthetic technique to introduce the P-CPh₃ moiety into a targeted nucleophilic substrate. It is interesting to highlight the resemblance of these FMOs exhibited by the PR fragment in a σ³λ³-oxaphosphirane to those of singlet carbene-type features, showing filled (HOMO–2 in 4a′′) and vacant (HOMO and LUMO in 4a′′) p-type atomic orbitals (Fig. 10a).

Fig. 10 (a) Computed [CPCM_toluene/PWPB95-D3/def2-QZVPP//CPCM_toluene/PBEh-3c] Kohn–Sham frontier molecular orbital isosurfaces (0.06 – isovalue) for oxaphosphirane 4a′′. (b) Variation of the Laplacian of electron density along the pnictogen C⋯P bond path for the ionic-enhanced dative bonding in 4a′′·NHC.

The nature of the NHC–P bond in 4a′′·NHC was analyzed following a well-established protocol³⁴ based on the analysis of the Laplacian of the electron density along the central part of the NHC–P bond path. In the case of 4a′′·NHC, a rather short P⋯C distance (d = 2.130 Å) is observed and the plot along the P⋯C bond path exhibits a small negative value of the Laplacian at the BCP (∇²ρ = −1.098 e Å⁻⁵) with the two VSCC (valence-shell charge concentration) bands located at the two different basins (each side of the BCP) with a small negative relative charge concentration band position parameter, τ_VSCC = −0.016 (Fig. 10b). These features correspond to ionic-enhanced dative bonding, according to a very recent description.³⁵

Conclusions

Herein, a new synthetic protocol to obtain stable and isolable oxaphosphiranes is described using a combination of kinetic protection of the P-center and the ring bond(s) strengthening effect of fluorinated C-bound aryl groups. These first isolated unligated oxaphosphiranes were obtained via decomplexation from molybdenum(0) complexes using (cheap) N-methyl-imidazole under thermal conditions; the newly formed complex [Mo(CO)₃(NMeIm)₃] can be easily separated and, hence, offer the option to be recycled. Reactivity studies on such oxaphosphiranes were performed using one derivative as a case in point. The reaction with tetrachloro-ortho-benzoquinone (TOB) led to oxidation of phosphorus and elimination of the aldehyde unit, finally. A similar reaction outcome was observed in reactions with acidic reagents (H₂O and HCl) as well as with a nucleophilic and basic reagent such as IMe₄, both leading to ring cleavage and the elimination of the aldehyde unit.

Computational analysis shows that full oxidation with TOB takes place in three exergonic low-barrier steps. The hydrolysis proceeds by preferential P- over C-attack of water on the ring in model compounds with only moderate steric bulk at phosphorus, although it could be expected to be reversed in P-trityl substituted derivatives. The initially formed phosphinic acid rearranges to yield the final phosphane oxide. However, it can also eliminate an aldehyde unit, affording a new phosphinic acid as a precursor of a primary phosphane oxide R-PH₂(O). A reaction with HCl or an NHC derivative furnishes P-addition products as a result of the remarkable reactivity feature of a formal singlet phosphinidene being (still) part of the oxaphosphirane. To the best of our knowledge, the latter bonding feature of a σ³λ³-P center in a heterocycle is of particular importance, which has not been reported for small rings, so far. It should also be noted that geometrically distorted, T-shaped cyclic phosphanes show a similar reactivity.

Author contributions

Conceptualization: R. S. and A. E. F.; methodology: N. V., S. B. and A. G. A.; validation: A. E. F. and R. S.; formal analysis and investigation: N. V., J. H., S. B. and A. G. A.; resources: R. S. and A. E. F.; data curation: A. E. F. and R. S.; writing—review and editing: A. E. F. and R. S.; and supervision: A. E. F. and R. S. All authors have read and agreed to the published version of the manuscript.

Data availability

Electronic supplementary information available: experimental protocols, NMR spectra, X-ray diffraction data and theoretical results.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was conducted in memory of Prof. Dr E. Niecke and Prof. Dr F. Mathey and to honor their ground-breaking discoveries in the chemistry of small P-heterocycles.

The authors are grateful to the Deutsche Forschungsgemeinschaft (STR 411/29-3), the University of Bonn, for financial support (R. S.), to the Servicio de Cálculo Científico at the University of Murcia for technical support and computational resources (A. E. F. and A. G. A.), and to the University of Murcia for a pre-doctoral fellowship (A. G. A).

References

1. (a) A. Wurtz, Liebigs Ann. Chem., 1859, 110, 125; (b) S. Rebsdat and D. Mayer, in Ullmann's encyclopedia of industrial chemistry, Wiley-VCH, Wiley online library, Weinheim, 7th edn, 2010.
2. R. E. Parker and N. S. Isaacs, Chem. Rev., 1959, 59, 737.
3. H. Q. Pham and M. J. Marks, in Ullmann's encyclopedia of industrial chemistry, Wiley-VCH, Wiley online library, Weinheim, 7th edn, 2010.
4. G. He, O. Shynkaruk, M. W. Lui and E. Rivard, Chem. Rev., 2014, 114, 7815.
5. S. Ishida, T. Iwamoto and M. Kira, Organometallics, 2010, 29, 5526.
6. A. Rey Planells and A. Espinosa Ferao, Inorg. Chem., 2022, 61, 13846.
7. A. Espinosa Ferao and A. Rey Planells, Chem. – Eur. J., 2023, 29, e202302243.
8. A. Chiaroni, G. Hanquet, M. Lusinchi and C. Riche, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1995, 51, 2047.
9. W. Sander, K. Schroeder, S. Muthusamy, A. Kirschfeld, W. Kappert, R. Boese, E. Kraka, C. Sosa and D. Cremer, J. Am. Chem. Soc., 1997, 119, 7265.
10. (a) P. R. Schreiner, H. P. Reisenauer, J. Romanski and G. Mloston, J. Am. Chem. Soc., 2010, 132, 7240; (b) H. P. Reisenauer, G. Mloston, J. Romanski and P. R. Schreiner, Eur. J. Org. Chem., 2011, 6269; (c) T. Akasaka and W. Ando, Phosphorus, Sulfur Silicon Relat. Elem., 1994, 95, 437.
11. S. Bauer, A. Marinetti, L. Ricard and F. Mathey, Angew. Chem., Int. Ed. Engl., 1990, 29, 1166 ( Angew. Chem. , 1990 , 102 , 1188 ).
12. R. Streubel, A. Kusenberg, J. Jeske and P. G. Jones, Angew. Chem., Int. Ed. Engl., 1995, 33, 2427 ( Angew. Chem. , 1994 , 106 , 2564 ).
13. A. Özbolat, G. von Frantzius, J. M. Pérez, M. Nieger and R. Streubel, Angew. Chem., Int. Ed., 2007, 46, 9327 ( Angew. Chem. , 2007 , 119 , 9488 ).
14. (a) C. Albrecht, M. Bode, J. M. Pérez, J. Daniels, G. Schnakenburg and R. Streubel, Dalton Trans., 2011, 40, 2654; (b) V. Nesterov, G. Schnakenburg, A. Espinosa and R. Streubel, Inorg. Chem., 2012, 51, 12343; (c) J. Faßbender, N. Künemund, A. Espinosa Ferao, G. Schnakenburg and R. Streubel, Organometallics, 2018, 37, 1331.
15. N. Volk, P. Malik, A. García Alcaraz, A. Espinosa Ferao and R. Streubel, Coord. Chem. Rev., 2021, 437, 213818.
16. F. Ramirez, A. S. Gulati and C. P. Smith, J. Org. Chem., 1968, 33, 13.
17. A. Espinosa Ferao, Inorg. Chem., 2018, 57, 8058.
18. G.-V. Röschenthaler, K. Sauerbrey and R. Schmutzler, Eur. J. Inorg. Chem., 1978, 111, 3105.
19. J. Faßbender, N. Volk, A. García Alcaraz, S. Balasubramaniam, A. Espinosa Ferao and R. Streubel, Chem. Commun., 2023, 59, 1285.
20. A. Espinosa Ferao and R. Streubel, Eur. J. Inorg. Chem., 2017, 2707.
21. C. Murcia-García, A. Bauzá, G. Schnakenburg, A. Frontera and R. Streubel, CrystEngComm, 2015, 17, 1769.
22. A. Schmer, P. Junker, A. Espinosa Ferao and R. Streubel, Acc. Chem. Res., 2021, 54, 1754.
23. C. Murcia García, PhD thesis, Rheinische Friedrich-Wilhelms-Universität Bonn, 2017.
24. A. Marinetti, F. Mathey, J. Fischer and A. Mitschler, J. Chem. Soc., Chem. Commun., 1984, 45.
25. H. D. Flack and G. Bernardinelli, Chirality, 2008, 20, 681.
26. (a) M. Schnebel, I. Weidner, R. Wartchow and H. Butenschön, Eur. J. Org. Chem., 2003, 2003, 4363; (b) J.-Y. Shie, J.-L. Zhu and Y.-P. Wu, J. Chin. Chem. Soc., 2015, 62, 665.
27. S. Sase, N. Kano and T. Kawashima, J. Org. Chem., 2006, 71, 5448.
28. V. Plack, J. R. Goerlich, H. Thönnessen, P. G. Jones and R. Schmutzler, Heteroat. Chem., 1999, 10, 277.
29. M. Nieger and M. W. M. Boras, CCDC 1843600:†Experimental Crystal Structure Determination, 2018.
30. S. Priya, M. S. Balakrishna, J. T. Mague and S. M. Mobin, Inorg. Chem., 2003, 42, 1272.
31. A. Espinosa Ferao, A. Rey Planells and R. Streubel, Eur. J. Inorg. Chem., 2021, 348.
32. (a) M. Schröder, PhD thesis, Universität Kaiserslautern, 1999; (b) J. M. Perez, PhD thesis, Rheinische Friedrich-Wilhelms-Universität Bonn, 2010; (c) J. Marinas Pérez, C. Albrecht, H. Helten, G. Schnakenburg and R. Streubel, Chem. Commun., 2010, 46, 7244; (d) J. Marinas Pérez, H. Helten, B. Donnadieu, C. A. Reed and R. Streubel, Angew. Chem., Int. Ed., 2010, 49, 2615 ( Angew. Chem. , 2010 , 122 , 2670 ); (e) J. Faßbender, PhD thesis, 2017; (f) J. Fassbender, G. Schnakenburg, D. P. Gates, A. Espinosa Ferao and R. Streubel, Chem. Commun., 2018, 54, 14013.
33. V. Plack, Diploma Thesis, Universität Braunschweig, 1993.
34. (a) V. Nesterov, R. Baierl, F. Hanusch, A. Espinosa Ferao and S. Inoue, J. Am. Chem. Soc., 2019, 141, 14576; (b) M. W. Stanford, J. I. Schweizer, M. Menche, G. S. Nichol, M. C. Holthausen and M. J. Cowley, Angew. Chem., Int. Ed., 2019, 58, 1329 ( Angew. Chem. , 2019 , 131 , 1343 ); (c) A. Espinosa Ferao, A. García Alcaraz, S. Zaragoza Noguera and R. Streubel, Inorg. Chem., 2020, 59, 12829.
35. D. Biskup, G. Schnakenburg, R. T. Boeré, A. Espinosa Ferao and R. Streubel, Dalton Trans., 2023, 52, 13781.

---

Footnotes

† Electronic supplementary information (ESI) available: Experimental protocols, NMR spectra, X-ray data and computational results. CCDC 2382051–2382055. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d4dt02915a

‡ These authors contributed equally to this work.

---