A selective route to aryl-triphosphiranes and their titanocene-induced fragmentation

Abstract

Triphosphiranes are three-membered phosphorus cycles and their fundamental reactivity has been studied in recent decades. We recently developed a high-yielding, selective synthesis for various aryl-substituted triphosphiranes. Variation of the reaction conditions in combination with theoretical studies helped to rationalize the formation of these homoleptic phosphorus ring systems and highly reactive intermediates could be isolated. In addition we showed that a titanocene synthon [Cp₂Ti(btmsa)] facilitates the selective conversion of these triphosphiranes into titanocene diphosphene complexes. This unexpected reactivity mode was further studied theoretically and experimental evidence is presented for the proposed reaction mechanism.

---

Introduction

Triphosphiranes are three-membered cyclo-phosphines, which are promising synthons in inorganic chemistry (Scheme 1). As early as 1877 the first cyclic oligophosphine was synthesized by Köhler and Michaelis in an attempt to prepare a phosphorus analogue of azobenzene with a PP double bond.¹ Almost 100 years later in 1964 the molecular structure of the product could be identified as P₅Ph₅ by X-ray crystal structure analysis.² Although, Cowley et al. already mentioned the synthesis of P₃(C₂F₅)₃ in 1970,³ it was later discussed that in fact the tetramer and pentamer were formed under the reaction conditions described.⁴ The first stable triphosphirane P₃^tBu₃ was reported by Baudler and co-workers in 1976,^5,6 and various synthetic approaches towards triphosphiranes have since emerged.⁷ Reductive approaches starting from dihalophosphines RPX₂ (X = Cl, Br) result in a mixture of oligophosphines of different ring sizes of P_nR_n (n = 3, 4, 5, 6) and are thus regarded as unspecific.⁸ The ratio of the different oligomers heavily depends on the steric demand of the substituent R.^5,9 Cyclo-condensation reactions, which also allow the preparation of unsymmetrically substituted triphosphiranes, and cyclization by reductive dehalogenation of dihalotriphosphines have emerged as more selective synthetic pathways.¹⁰ Nevertheless, the presence of other cyclic oligophosphines as side products is often observed.

Scheme 1 Selected reactivity modes of differently substituted triphosphiranes.

Jutzi and co-workers have shown that selenium inserts into one P–P bond of P₃Cp*₃ (Cp* = pentamethylcyclopentadienyl), affording a mixture of cyclic selenotriphosphabutanes (Scheme 1, A) and cyclic selenodiphosphapropanes (Scheme 1, B).¹¹ In contrast, thermolysis of P₃Cp*₃ in xylene resulted in the formation of different phosphorus clusters, some of which are structurally related to Hittorf's-phosphorus (Scheme 1, C and D).¹² Ring expansion reactions were reported by Uhl and Benter by the insertion of Ga(I) into a P–P bond of P₃^tBu₃, thus establishing a way to prepare cyclo-galliumtriphosphabutanes (Scheme 1, E, M = Ga).¹³ A similar reactivity is observed when Al(I) compound (AlCp*)₄ reacts with P₃^tBu₃ (Scheme 1, E, M = Al).¹⁴ In addition, the reaction of P₃^tBu₃ with PMe₂Cl or PPh₂Cl in the presence of Me₃SiOTf or GaCl₃, respectively, resulted in the selective ring expansion with insertion of [PMe₂]⁺ into the P–P bond between the two identical P atoms of P₃^tBu₃ to afford [R₂P(P₃^tBu₃)]⁺ (Scheme 1, F; R = Me, Ph).^15,16 More recently, Manners and co-workers showed the addition of P₃^tBu₃ to organic nitriles after activation of the three-membered ring by electrophiles to yield differently substituted 1-aza-2,3,4-triphospholenes in a click-type reaction (Scheme 1, G),^17,18 underlining the value of triphosphiranes as synthons in synthetic inorganic chemistry. Fragmentation of P₃^tBu₃ was observed by Fenske and Ahlrichs in the reaction with Ni(CO)₄, resulting in the formation of [Ni₅(P^tBu)₃(P₃^tBu₃)(CO)₅] with μ₄- and μ₃-bridging P^tBu ligands as well as a P₃^tBu₃ chain, acting as a μ₄(η²,η^1′,η^2′′) ligand to three Ni atoms of the cluster.¹⁹

To the best of our knowledge, only four aryl-substituted triphosphiranes are reported in the literature. P₃Ph₃ was described as early as 1973 as a labile solid with respect to P₅Ph₅,²⁰ and it has been shown that this compound is part of an equilibrium mixture consisting of different oligomers with ring sizes of n = 3, 4, 5, 6.²¹ Tokitoh et al. synthesized (Anth = 9-anthryl, Bbt = 2,6-bis[bis-(trimethylsilyl)methyl]-4-[tris(trimethylsilyl)phenyl]) in good yield by heating a mixture of AnthP=PBbt and ⁿBu₃P=Te.²² P₃Tipp₃ (Tipp = 2,4,6-ⁱPr₃C₆H₂) and P₃Mes₃ (Mes = 2,4,6-Me₃C₆H₂) were described as one of a mixture of products when free phosphinidenes were generated by reductive dechlorination of RPCl₂ (R = Tipp, Mes).^23–25 Moreover, Gaspar and co-workers reported on the photochemical release of the triplet phosphinidene MesP from MesP(C₂H₄) in 1992.²⁶ In the absence of a trapping reagent these triplet phosphinidenes oligomerize to give a mixture containing P₃Mes₃ and P₄Mes₄.

Using [W(PMe₃)₆] as a reducing agent the quantitative coupling of RPCl₂ (R = Mes* = 2,4,6-^tBuC₆H₂; 2,4,6-(CF₃)₃C₆H₂) to the respective diphosphenes RP=PR was detected. Starting from TippPCl₂, the initial formation of the diphosphene is detected by ³¹P NMR spectroscopy, however, the reaction continues to produce Tipp₃P₃ as the final product, clearly pointing to the intermediacy of W=PR species.²⁷ Moreover, it was shown that the reductive degradation of P₄ with mesityl-radicals (generated from Mes-Br and Ti(III)-based chlorine atom abstracting reagent [Ti{N(^tBu)(3,5-C₆H₃Me₂)}₃]) yields P₃Mes₃ as the main product in good isolated yields.²⁸

In 1998 Shah and Protasiewicz reported the formation of the triphosphirane P₃Tipp₃ (1a) by treatment of TippPCl₂ with PMe₃ and Zn and subsequent reaction with benzaldehyde (Scheme 2).²⁹ This so-called phospha-Wittig reaction afforded a mixture of P₃Tipp₃ and traces of the desired phosphaalkene Ph(H)C=PTipp.

Scheme 2 Formation of Tipp₃P₃ (1a) and trace amounts of phosphaalkene H(Ph)C=PTipp in a so-called phospha-Wittig protocol.

In this contribution, we report on the synthesis of aryl substituted triphosphiranes using a modified synthesis on the basis of the studies by Protasiewicz et al. Furthermore, we report on the selective degradation of these P₃Ar₃ systems using [Cp₂Ti(btmsa)] (Cp = cyclopentadienyl, btmsa = C₂(SiMe₃)₂) as a Ti(II) synthon.

Results

In an attempt to prepare new variants of pyridinephosphaalkenes,³⁰ we utilized the phospha-Wittig protocol described by Protasiewicz et al. with DippPCl₂ (Dipp = 2,6-ⁱPr₂C₆H₃), PMe₃ and excess of Zn powder in a strict low-temperature regime (−78 °C); after subsequent treatment with pyridine-2-carbaldehyde at that temperature and warming to room temperature the formation of the respective phosphaalkene was not observed. The ³¹P NMR spectrum of the reaction mixture displayed a major product with an A₂B spin system with a doublet at −99.47 ppm and a triplet at −132.90 ppm with a coupling constant of 178.5 Hz, which was identified as P₃Dipp₃ (1b), in line with attempted synthesis of the phospha-Wittig reagent TippPPMe₃ as discussed before.²⁹ X-ray quality crystals of 1b were grown from a saturated n-hexane solution at 5 °C (Fig. 1). 1b crystallises in the monoclinic space group P2₁/c with four molecules in the unit cell. The molecular structure of 1b shows the expected down-down-up orientation of the Dipp groups with respect to the central P₃ plane, with a minimally distorted central P₃-ring [P1–P2 2.1991(4), P2–P3 2.2440(4); P1–P3 2.2124(3) Å] (Fig. 1). These metric parameters are in line with those detected for 1a and 1c (Table S1†),³¹ of which the molecular structures have been reported previously.^32,33

Fig. 1 POV-ray depiction of the molecular structure of 1b. ORTEPs drawn at 30% probability, H atoms are omitted for clarity. Selected bond lengths (Å) and angles (°): P1–P2 2.1991(4), P2–P3 2.2440(4), P1–P3 2.2124(3), P1–C1 1.8526(10), P2–C13 1.8594(10), P3–C25 1.8507(10); P1–P2–P3 59.718(11), P2–P1–P3 61.147(12), P1–P3–P2 59.135(11).

We then utilized the sterically more demanding PEt₃ to better stabilize the reactive phosphanylidenephosphorane intermediate TippP=PEt₃. Phosphanylidenephosphoranes have been identified as a source of the triplet phoshinidenes Ar–P.³⁴

Additionally, we switched to TippPBr₂, as its reduction should be more facile. TippPBr₂, PEt₃ (1.2 equiv.) and Zn (3 equiv.) were combined in THF at −78 °C and the formation of a deep yellow to orange suspension was observed, which again showed P₃Tipp₃ (1a) as the major species in the ³¹P NMR spectrum.

After removal of the solvent and extraction with n-hexane minimal amounts (<0.01 g) of yellow needles suitable for single crystal X-ray analysis were obtained and identified as the elusive diphosphene P₂Tipp₂ (2) (Fig. 2), which has only been observed in solution in the [W(PMe₃)₆] mediated coupling of ArPCl₂ (Ar = Tipp, Mes*, 2,4,6-(CF₃)₃C₆H₂) by ³¹P NMR experiments to date.²⁷ The ³¹P NMR spectrum of isolated 2 showed P₂Tipp₂ (δ(³¹P) = 517.4 ppm) to be the major species, whereas minor amounts of P₃Tipp₃ and P₄Tipp₄ were also detected. Monitoring a C₆D₆ solution of 2 over time at room temperature revealed that P₂Tipp₂ slowly coverts into P₃Tipp₃ and its dimer P₄Tipp₄, vide infra.³¹2 crystallises as its trans-conformer in the triclinic space group P1̄ with one molecule in the unit cell. The P1–P1′ distance [2.0290(5) Å] (cf. d(P=P) P₂Mes*₂ 2.034(2);³⁵ P₂Ter₂ 2.029(1);³⁶ P₂Bbt₂ 2.043(1)³⁷) is in the expected range for a diphosphene (∑r_cov(P=P) = 2.04 Å),³⁸ and rather acute C–P–P′ [99.61(3)°] angles at the dicoordinate P center are detected.

Fig. 2 POV-ray depiction of the molecular structure of 2. ORTEPs drawn at 30% probability. Selected bond lengths (Å) and angles (°): P1–P1′ 2.0290(5), P1–C1 1.8439(10); C1–P1–P1′ 99.61(3); P1′–P1–C1–C6 91.34(8), C1–C2–C3–C4 1.25(16).

Theoretical investigations at the M062X/TZVP level of density functional theory were carried out, assuming that transient phosphinidenes are formed. The gas-phase trimerization of Dipp-P with a triplet ground state (the corresponding singlet state is less stable by 26.01 kcal mol⁻¹) is exergonic (−91.39 kcal mol⁻¹). In addition, we computed the transfer reaction of a Dipp-P fragment (which may be formed intermediately at low temperatures) via DippPPMe₃ to P₂Dipp₂ and found this reaction to be exergonic by −15.74 kcal mol⁻¹ (energy barriers were not calculated). This is in line with the isolation of 2.

Since there are only few high-yielding, selective methods for the preparation of aryl-substituted triphosphiranes outlined in the literature, we decided to take a closer look at this synthetic approach. We therefore tested different aryl(dichloro)phosphines ArPCl₂ (Ar = Mes, Dipp, Tipp) to elucidate whether treatment with PR₃ (R = Me, Et) and Zn gives general access to aryl-substituted triphosphiranes (Scheme 3).

Scheme 3 (i and ii) General procedure for the preparation of 1a–c; (iii) identification of PMe₃ as the active reductant; (iv) Zn can be excluded as active reductant.

The reaction of ArPCl₂ with PMe₃ (2.5 equiv.) and an excess of Zn (5 equiv.) in anhydrous THF afforded P₃Ar₃ (Ar = Tipp (1a), Dipp (1b), Mes (1c)) as expected (Scheme 3, reaction (i)).

Purification by recrystallisation from a saturated n-hexane solution at 5 °C yielded 1a–c as colourless crystalline solids in 47, 50 and 10% isolated yield, respectively.

Starting from the easily accessible mixed dihalophosphines ArPX₂ (Ar = Tipp, Dipp, Mes; X = Cl, Br; obtained through treatment of ArMgBr with PCl₃),³⁹ with PMe₃ and Zn in a 1/2/2.5 molar ratio in THF at room temperature (Scheme 3, reaction (ii)), 1a, 1b and 1c could be obtained in up to 72%, 75% and 52% isolated yield, respectively, after extraction with benzene or Et₂O in case of 1c. 1a–c show good thermal stability with melting points of higher than 167 °C.³¹ Heating a solution of 1a in C₆D₆ for 36 h at 80 °C showed no decomposition or rearrangement products in the ³¹P NMR spectrum.

Since either PMe₃ or Zn can act as reducing agents, we reduced TippPCl₂ with each reductant separately (Scheme 3(iii) and (iv)). While there is no reaction observed, when TippPCl₂ or TippPBr₂ are stirred with an excess of Zn in THF over a period of 24 h, treatment of ArPCl₂ with a fivefold excess of PMe₃ afforded 1a–c in 43, 66 and 18% isolated yield, respectively. The potential of PMe₃ to act as a chlorine abstracting reagent is documented in the literature and results in oxidation to the respective dichlorophosphorane,^40,41 or the homoleptic dication salt [Me₃PPMe₃]₂Cl₂. This concept has been used to access cyclo-tetra(stibinophosphonium) triflate salts of the type [Sb₄(PR₃)₄][OTf]₄ (R = Me, Et, Pr, Bu), cationic antimony compounds related to the cyclic oligophosphines.⁴² To shed light on this proposition, we independently synthesized PMe₃Cl₂ and treated it with an excess of zinc dust in the presence of TippPCl₂ in a mixture of MeCN/THF (3 : 1) over 24 h. A ³¹P NMR spectrum of the reaction mixture indeed showed 1a to be the main product of this reaction.³¹ It can thus be concluded that PMe₃Cl₂ is a plausible by-product of the reduction with PMe₃ and zinc can reduce it back to PMe₃, vide infra. This opens the pathway for potential catalytic reduction of ArPCl₂ with PMe₃ and Zn as a sacrificial reductant. In another experiment DippPCl₂ was reduced with an excess of PMe₃ and the white precipitate was carefully washed with benzene and n-hexane. Subsequently, the precipitate was treated with AgOTf in CH₂Cl₂. After filtration a colourless solid was obtained, which was dissolved in CD₃CN, allowing to unambiguously identify [Me₃PCl]OTf (δ³¹P{¹H} = 93.6 ppm),⁴³ and [Me₃P–PMe₃][OTf]₂ (δ³¹P{¹H} = 28.4 ppm)⁴⁴ among three unidentified PMe₃ containing species (Scheme 3(iii)).³¹

The synthetic approach using Zn/PMe₃ showed a high selectivity towards the respective triphosphiranes. In the case of 1a and 1b just little amounts of the corresponding cyclic tetraphosphines P₄Ar₄ were detected as side products by ³¹P NMR spectroscopy of the reaction mixture. When MesPCl₂ is applied in our approach, the selectivity decreases and the formation of little amounts of the cyclic tetraphosphine P₄Mes₄, and the cyclic pentaphosphine P₅Mes₅ species can be detected. We conclude that this is due to lesser steric bulk imposed by the mesityl substituent. The sterically more demanding substituents Tipp and Dipp promote the formation of the three-membered phosphorus ring more effectively.⁷

Having prepared 1a–c we wanted to explore their reactivity with the titanocene synthon [Cp₂Ti(btmsa)] in order to access titanium phosphinidene complexes.

Titanocene-induced degradation of R₃P₃

Stephan and co-workers have shown the phospha-Wittig-type phosphinidene transfer for [Cp₂Zr=PMes*(PMe₃)] resulting in the formation of phosphaalkenes in the reaction with aldehydes along with the formation of [Cp₂ZrO]_n.⁴⁵ Similar reactivity was observed by Cummins and Schrock for the terminal tantalum phosphinidene complexes, [(N₃N)Ta=PR] (N₃N = (Me₃Si–NCH₂CH₂)₃N).⁴⁶

With the series of triphosphiranes 1a–c synthesized, we wanted to investigate the propensity to access monomeric, terminal Cp₂Ti=PR complexes, by reaction of 1 with the titanocene synthon [Cp₂Ti(btmsa)]. Cp₂Ti=PR has not been described in the literature. There are reports of neutral and zwitterionic terminal titanium phosphinidene complexes of the type [(^ArNacnac)Ti=PAr′(R)] (Ar′ = Tipp, Mes*; R = CH₂^tBu, CH₃, CH₃[B(C₆F₅)₃]) by Mindiola and co-workers with a bulky β-diketiminate ligand (^ArNacnac=[Ar]NC(Me)CHC(Me)N[Ar], Ar = Dipp) on titanium.^47,48 [Cp₂Ti(btmsa)] is obtained by reduction of Cp₂TiCl₂ in the presence of btmsa. In these complexes btmsa acts as a spectator ligand and its facile release under the respective reaction conditions generates the highly reactive 14-electron [Cp₂Ti] fragment in situ.⁴⁹ Combination of three equivalents [Cp₂Ti(btmsa)] with 1b in C₆D₆ at room temperature and monitoring by ³¹P NMR spectroscopy revealed slow, but selective, conversion into a phosphorus-containing species with a singlet resonance at 283.8 ppm. Heating this reaction mixture to 80 °C over a period of 16 h in a sealed NMR tube resulted in consumption of [Cp₂Ti(btmsa)] according to ¹H NMR spectroscopy. However, unreacted P₃Ar₃ remained in the reaction mixture and thus, more [Cp₂Ti(btmsa)] was added to the reaction mixture and heating to 80 °C was continued. Fractional crystallisation from C₆D₆ and determination of the molecular structure by single crystal X-ray analysis revealed the formation of the η²-diphosphene complex [Cp₂Ti(P₂Dipp₂)] (3b) (Fig. 3, right). Consequently, the reaction was repeated in the correct stoichiometry with [Cp₂Ti(btmsa)] and 1b in a 3 : 2 molar ratio in benzene, which allowed for full conversion into 3b after stirring at 80 °C over a period of 16 h. In analogy, 1a and 1c were converted into the respective titanocene diphosphene complexes [Cp₂Ti(P₂Tipp₂)] (3a, Fig. 3, left) and [Cp₂Ti(P₂Mes₂)] (3c) (Scheme 4). Filtration and subsequent concentration of the reaction mixtures and standing overnight at 5 °C resulted in the formation of deep yellow crystals of 3a suitable for X-ray analysis, whereas formation of 3c was authenticated by NMR spectroscopy, elemental analysis and HR-MS studies.³¹ Interestingly, in the ¹H NMR spectrum three or two independent septets are detected for 3a and 3b, respectively. This indicates hindered rotation about the P–C_Ar bond and the Me group of the isopropyl moiety in close proximity to the Cp₂Ti-fragment is significantly upfield-shifted, resonating at −0.99 ppm in 3a and 3b. This hindered rotation is also evident in 3c, in which three ¹H NMR signals are detected for the Me groups of the Mes moiety.

Fig. 3 POV-ray depiction of the molecular structure of 3a and 3b. ORTEPs drawn at 30% probability, all H-atoms are omitted for clarity. Selected bond lengths (Å) and angles (°) of 3a: P1–P1′ 2.1826(7), P1–C1 1.8548(13), P1–Ti1 2.5329(5); C1–P1–P1′ 108.39(5), P1–Ti1–P1′ 51.042(17). 3b: P1–P2 2.1699(5), P1–Ti1 2.5425(5), P2–Ti1 2.5230(5), P1–C11 1.8548(13), P2–C23 1.8495(13); C11–P1–P2 108.88(4), C23–P2–P1 112.53(4), P1–Ti1–P2 50.725(12).

Scheme 4 Selective degradation of P₃Ar₃ (1a–c) into [Cp₂Ti(P₂Ar₂)] (3a–c) complexes using [Cp₂Ti(btmsa)] as a synthon for [Cp₂Ti].

3a crystallises in the monoclinic space group C2/c with four molecules in the unit cell as a benzene solvate. 3b crystallises in the monoclinic space group P2₁/c with four molecules of 3b and four C₆D₆ molecules in the unit cell. 3a is located on a special position and thus shows C₂ symmetry in the solid state. The P–P distances in 3a [2.1826(7) Å] and 3b [2.1699(5) Å] are intermediate between a P–P single and double bond (∑r_cov(P=P) = 2.04 Å; (P–P) 2.22 Å)³⁸ and are in line with the P–P distance [2.173(4) Å] in [rac-(EBTHI)Ti(P₂Ph₂)] (ETBHI = ethylene-1,2-bis(5-4,5,6,7-tetrahydro-1-indenyl)), the only titanium diphosphene complex known to date.⁵⁰ It is worth noting that green [rac-(EBTHI)Ti(P₂Ph₂)] is insoluble in common non-halogenated organic solvents and thus, NMR data was not obtained. It is formed through the dehydrocoupling of PhPH₂ in the presence of the Ti(III)-hydride dimer [rac-(EBTHI)-TiH]₂.⁵¹ η²-Diphosphene complexes of various transition metals have been known and were thoroughly reviewed by Weber.⁵² Noteworthy, is the formation of [(Ph₃P)₂M(P₂{C₆F₅}₂)] with an E-configured diphosphene ligand by the degradation of cyclic tetraphosphine P₄(C₆F₅)₄ in the presence of M(PPh₃)₄ (M = Pt,⁵³ Pd⁵⁴). Other known diphosphene complexes of group 4 include the anionic species [Cp₂Zr(PPh)₂Br]⁻ with a P–P distance [2.145(3) Å] shorter than in 3a and 3b,⁵⁵ and the related Mes-substituted complex [Cp₂Zr(P₂Mes₂)] with a similar P–P distance [2.188(3) Å].⁵⁶ The Ti–P distances in 3a [2.5425(5), 2.5230(5) Å] and 3b [2.5329(5)) Å], as well as the P–Ti–P angles (3a 50.725(12)°; 3b 51.042(17)°), are similar to that in [Cp₂Zr(PPh)₂Br]⁻ [d(Ti–P) 2.525(2) Å; <(P–Ti–P) 51.00(6)°] and point to a Ti(IV) center and an overall titana-cyclo-propane, rather than a titana-cyclo-propene type structure.

The surprising selective formation of the titanocene diphosphene species 3, prompted us to study the reactivity by DFT calculations on the M062X/TZVP level of theory. The calculated gas phase structure of 1b and 3b and the metric parameters derived from X-ray crystallography are in good agreement. In a next step the reaction of [Cp₂Ti(btmsa)] with 1b in a 3 : 2 ratio was investigated. It is found that the gas phase reaction is exergonic by −15.93 kcal mol⁻¹, indicating that the reaction is accessible thermodynamically, even though energy barriers for this transformation could not be determined (Scheme 5(i)). Using the truncated model compound P₃Ph₃ (1Ph) the same exergonic character was calculated (ΔG = −18.32 kcal mol⁻¹) for this transformation. Additionally, we were interested to determine whether the free trans-diphosphenes P₂Dipp₂ and P₂Ph₂ can displace the btmsa molecule in [Cp₂Ti(btmsa)] to afford complexes 3b and [Cp₂Ti(P₂Ph₂)] (3Ph), respectively (Scheme 7, bottom). Interestingly, this reaction is also exergonic for P₂Dipp₂ and P₂Ph₂ by −10.69 and −20.87 kcal mol⁻¹, respectively, illustrating that diphosphenes are potential intermediates along the reaction pathway (Scheme 5(iv)).

Scheme 5 M062X/TZVP (i and iv) and BP86/TZVP ((ii and iii) (M062X) for R = ^tBu) computed reaction free energies for possible paths of formation of [Cp₂Ti(P₂R₂)] in the gas phase.

With minimal amounts of the free diphosphene Tipp₂P₂ (2) in hand, we treated 2 with [Cp₂Ti(btmsa)] in a 1 : 1 ratio at room temperature in C₆D₆. Having shown that the reaction of 1a with [Cp₂Ti(btmsa)] is slow at room temperature and full conversion is only achieved at 80 °C, we were delighted to see the disappearance of the diagnostic diphosphene signal at 517.4 ppm and formation of 3a with a characteristic ³¹P NMR shift of 290.7 ppm. This clearly shows, that diphosphenes are potential intermediates in the reaction of 1 with [Cp₂Ti(btmsa)]. Furthermore, this shows the drastic influences of the sterically demanding groups attached to phosphorus, as the diphosphene P₂Mes*₂ was shown to not afford the respective diphosphene complex in the reaction with [Cp₂Ti(btmsa)].⁵⁷

To compare the reactivity of the aryl-substituted triphosphiranes with alkyl-substituted derivatives we treated [Cp₂Ti(btmsa)] with the known triphosphiranes P₃^tBu₃ (1d) and P₃Ad₃ (Ad = adamantyl),^58,59 in a 1 : 1 ratio in benzene at 80 °C in C₆D₆ (Scheme 6). Interestingly, in the case of 1d full consumption of both starting materials was noted, with a new characteristic A₂B spin system in the ³¹P NMR spectrum. 1e also cleanly reacted in similar fashion, however full consumption was not achieved due to the poor solubility of 1e. Compared to 1d and 1e the A₂-part of the ³¹P NMR signal is downfield-shifted, thus indicating selective insertion into the P–P bond with the two identical P atoms and the formation of the triphosphanato-complexes [Cp₂Ti(P₃^tBu₃)] (4a) and [Cp₂Ti(P₃Ad₃)] (4b).³¹ Complex 4a among other [Cp₂Ti(P₃R₃)] species has been described before by Köpf and co-workers in the reaction of Cp₂TiCl₂ with the salt K₂[P₄^tBu₄] in a salt elimination reaction on the basis of NMR experiments.⁶⁰ Extraction of the reaction mixture with Et₂O, concentration to incipient crystallisation and standing at 5 °C overnight, afforded deeply coloured brown crystals of 4a suitable for X-ray analysis (Fig. 4) in 64% yield. To the best of our knowledge this is the first structural characterization of a cyclo-titanatriphosphine.

Scheme 6 Formation of the cyclo-titanatriphosphabutanes [Cp₂Ti(P₃R₃)] (R = ^tBu (4a), Ad (4b)) starting from [Cp₂Ti(btmsa)] and triphosphiranes 1d and 1e.

Fig. 4 POV-ray depiction of the molecular structure of 4a. ORTEPs drawn at 30% probability, H atoms are omitted for clarity. Selected bond lengths (Å) and angles (°): P1–P2 2.1953(8), P2–P3 2.1840(8), Ti1–P1 2.5354(6), Ti1–P2 3.0348(7), Ti1–P3 2.5480(7); P1–Ti1–P3 90.34(2).

4a crystallises in the orthorhombic space group P2₁2₁2₁ with four molecules in the unit cell. The P–P distances [P1–P2 2.1953(8), P2–P3 2.1840(8)] are shorter than a P–P single bond (∑r_cov(P–P) = 2.22 Å)³⁸ and the P–Ti–P angle [90.34(2)°] is wider than in 3a and 3b and compares nicely with the P–Zr–P angle [89.8(2)°] found in the related compound [Cp₂Zr(P₃Ph₃)].⁵⁵

To rationalize the contrasting reactivity of alkyl- and aryl-substituted triphosphiranes noted in this study, we calculated the free enthalpies for the gas phase reaction of [Cp₂Ti(btmsa)] with Dipp₃P₃ to afford the insertion product [Cp₂Ti(P₃Dipp₃)] under liberation of btmsa at the BP86//TZVP/LANL2DZ level of theory.³¹ This transformation was found to be endergonic by 11.41 kcal mol⁻¹, whereas this insertion process was computed to be almost thermo-neutral for P₃^tBu₃ (+1.64 (+4.20 M062X) kcal mol⁻¹) to give 4a (Scheme 5(ii)). The selective degradation of [Cp₂Ti(P₃Dipp₃)] to yield 3b and half an equivalent of P₂Dipp₂ was also considered and is shown to be exergonic by −23.27 kcal mol⁻¹, whereas the same process is endergonic by +0.21 (+9.97 M062X) kcal mol⁻¹ for 4a (Scheme 5(iii)). These results are in line with the observed difference in reactivity of alkyl- and aryl-substituted triphosphiranes and that the reactions only take place at elevated temperatures. We then wanted to determine whether single electron transfer (SET) is preferred over reduction of the cyclo-P₃R₃ in two electron steps by comparison of the free energies of the reduction products. It is noted from successive theoretical one-electron addition to triphosphiranes P₃R₃ that the single-electron transfer step is exergonic and favoured thermodynamically, while the two-electron transfer process is endergonic and thermodynamically not favored.³¹

On the basis of these results, one can expect a stepwise reaction mechanism for the electron transfer reactions. Furthermore one of the P–P bonds in the radical anion species [P₃R₃]˙⁻ is considerably elongated [2.814 (R = Dipp), 2.973 Å (R = Ph)], which would allow for the liberation of a phosphinidene fragment or the recombination of two radical anions, under formal exchange of P–R groups. If arylphosphinidenes were formed in this transformation these would be triplet species, with the triplet state being thermodynamically favored by −26.01 (R = Dipp) and −33.71 kcal mol⁻¹ (R = Ph), respectively. With these insights we set out to generate experimental evidence for these assumptions.

On the basis of these results, one can expect a stepwise reaction mechanism for the electron transfer reactions and the possible intermediary formation of a titanocene phosphinidene species. Electrochemical studies revealed an electrochemically irreversible reduction of 1b in THF at a potential of −3.09 V (vs. F_c/F_c⁺), which is in line with degradation of the aryl-substituted triphosphiranes into diphosphene fragments upon treatment with [Cp₂Ti(btmsa)]. Investigation of the reaction mixture of [Cp₂Ti(btmsa)] and 1a (3 : 2 ratio, after heating to 80 °C for 1 h) at room temperature by electron paramagnetic resonance (EPR) spectroscopy revealed the occurrence of an EPR-active intermediate (Fig. 5) with an isotropic g-factor of 1.978. This doublet signal shows strong coupling to one ³¹P nucleus with a(³¹P) = 72 MHz and hyperfine coupling to titanium a(^49/47Ti) = 22 MHz. The rather large g-value and small hyperfine coupling to Ti indicates a species with a high spin density on phosphorus, in which only one phosphorus is attached to titanium, as a more complex EPR-signal would be expected otherwise.⁶¹ In addition, there is an underlying signal stemming from [Cp₂Ti(btmsa)], which could be fitted to a species with g_iso = 1.973 and a(¹H) = 32 MHz.⁶² This could indicate a hydridic species such as [Cp₂Ti(III)–H], which has been discussed as resting state of [Cp₂Ti] in solution. In this case hydrogen release would generate the free titanocene and subsequent addition of H₂ regenerates the [Cp₂TiH] species.⁶³

Fig. 5 Experimental (black) and simulated (red) X-band EPR spectra of the intermediate formed in the reaction of [Cp₂Ti(btmsa)] with Tipp₃P₃ to yield 3a in benzene solution at room temperature. The simulation includes an impurity (g_iso = 1.973, a(¹H) = 32 MHz) which is present in the titanium precursor [Cp₂Ti(btmsa)]. Simulation parameters: g_iso = 1.978, a(³¹P) = 72 MHz, and a(^47,49Ti) = 22 MHz.

We then wanted to generate more evidence for the end group liberation and formation of free phosphinidenes during the reaction. If this is the case, starting from a 1 : 1 mixture of differently substituted triphosphiranes P₃Ar₃ and P₃Ar′₃ should result in the formation of the mixed diphosphene complex [Cp₂TiP₂ArAr′] (from recombination of differently substituted phosphinidenes) along with [Cp₂TiP₂Ar₂] and [Cp₂TiP₂Ar′₂]. Therefore, a 1 : 1 mixture of 1a and 1b (1 equiv.) was mixed with 1.5 equiv. of [Cp₂Ti(btmsa)] in C₆D₆ in an NMR scale reaction.

The ³¹P NMR spectrum of the resulting product solution is shown in Fig. 6. For comparison the spectra of the pure compounds 3a and 3b are depicted as well. In the spectrum of the product mixture the singlet signals of the symmetric compounds 3a and 3b can be seen clearly at 283.8 and 290.7 ppm, respectively. Additionally, there are two doublets, indicating the formation of the mixed diphosphene complex [Cp₂Ti(P₂DippTipp)] (3ab). We conclude from this experiment that an exchange of P–R end groups or the intermediacy of phospinidenes P–R are likely in the course of the reaction.

Fig. 6 Formation of the mixed diphosphene complex 3ab in a scrambling experiment utilizing a 1 : 1 mixture of 3a and 3b in the presence of 1.5 equiv. [Cp₂Ti(btmsa)].

Moreover, titana- and zirconacycles are regularly applied in the formation of main group element substituted heterocycles.^64,65 We wanted to probe this reactivity by treating isolated 3a with TippPCl₂ and found 1a as the product along with the formation of Cp₂TiCl₂ (Scheme 7),³¹ which clearly shows the potential of complexes 3 for the formation of small inorganic ring systems.

Scheme 7 Transmetalation of 3a with Tipp–PCl₂, resulting in the formation of 1a.

Conclusions

We have shown in here a simple and selective synthetic protocol for the formation of aryl-substituted triphosphiranes 1 of the type P₃Ar₃ and identified PMe₃ as the active reductant. These findings open the way for future studies to render these transformations catalytic with respect to PMe₃. Moreover, we have shown that the Ti(II) synthon [Cp₂Ti(btmsa)] reacts with 1 to yield the respective titanocene diphosphene complexes 3 in straightforward fashion. Combined theoretical and experimental studies suggest the intermediate formation of a paramagnetic titanium phosphorus species, indicating single electron transfer steps. Moreover, experimental evidence is presented for the intermediacy of free diphosphenes, authenticated by reaction of the elusive diphosphene P₂Tipp₂ (2) with [Cp₂Ti(btmsa)]. In first reactivity studies we have shown that 3 can be utilized as a P₂R₂-transfer reagent in transmetalation protocols using TippPCl₂. This opens the pathway to generate new P₂R₂-containing heterocycles.

Studies to further elucidate the reaction mechanism of the P₃Ar₃ degradation reaction are ongoing, to further investigate the nature of the paramagnetic intermediate. Additionally, application of the P₃Ar₃ systems in phosphinidene transfer reactions will be investigated.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

C. H.-J. thanks Prof. M. Beller for his support, the European Union for funding (H2020-MSCA-IF-2017 792177) and the Max Buchner-Foundation for a Scientific Fellowship. The CV studies were co-funded through the Leibniz Science Campus Phosphorous Research Rostock and the FCI (SK 202/22). We thank our technical and analytical staff for assistance, especially Dr Anke Spannenberg for her support regarding X-ray analysis. Dr Jonas Bresien is kindly acknowledged for help with vibrational spectroscopy.

Notes and references

1. H. Köhler and A. Michaelis, Ber. Dtsch. Chem. Ges., 1877, 10, 807.
2. J. J. Daly and L. Maier, Nature, 1964, 203, 1167.
3. A. H. Cowley, T. A. Furtsch and D. S. Dierdorf, J. Chem. Soc., Chem. Commun., 1970, 523.
4. P. S. Elmes, M. E. Redwood and B. O. West, J. Chem. Soc. D, 1970, 1120.
5. M. Baudler, B. Carlsohn, W. Böhm and G. Reuschenbach, Z. Naturforsch., B: Anorg. Chem., Org. Chem., 1976, 31, 558.
6. M. Baudler, B. Carlsohn, B. Kloth and D. Koch, Z. Anorg. Allg. Chem., 1977, 432, 67.
7. M. Baudler and K. Glinka, Chem. Rev., 1993, 93, 1623.
8. W. Mahler and A. B. Burg, J. Am. Chem. Soc., 1957, 79, 251.
9. M. Baudler, J. Hahn and E. Clef, Z. Naturforsch., B: Anorg. Chem., Org. Chem., 1984, 39, 438.
10. M. Baudler and J. Hellmann, Z. Naturforsch., B: Anorg. Chem., Org. Chem., 1981, 36, 266.
11. P. Jutzi, N. Brusdeilins, H.-G. Stammler and B. Neumann, Chem. Ber., 1994, 127, 997.
12. P. Jutzi and N. Brusdeilins, Z. Anorg. Allg. Chem., 1994, 620, 1375.
13. W. Uhl and M. Benter, J. Chem. Soc., Dalton Trans., 2000, 3133.
14. C. Üffing, C. v. Hänisch and H. Schnöckel, Z. Anorg. Allg. Chem., 2000, 626, 1557.
15. C. A. Dyker, N. Burford, G. Menard, M. D. Lumsden and A. Decken, Inorg. Chem., 2007, 46, 4277.
16. M. H. Holthausen, D. Knackstedt, N. Burford and J. J. Weigand, Aust. J. Chem., 2013, 66, 1155.
17. S. S. Chitnis, R. A. Musgrave, H. A. Sparkes, N. E. Pridmore, V. T. Annibale and I. Manners, Inorg. Chem., 2017, 56, 4521.
18. S. S. Chitnis, H. A. Sparkes, V. T. Annibale, N. E. Pridmore, A. M. Oliver and I. Manners, Angew. Chem., Int. Ed., 2017, 56, 9536.
19. R. Ahlrichs, D. Fenske, H. Oesen and U. Schneider, Angew. Chem., Int. Ed. Engl., 1992, 31, 323–326.
20. M. Baudler and M. Bock, Z. Anorg. Allg. Chem., 1973, 395, 37.
21. K. Schwedtmann, R. Schoemaker, F. Hennersdorf, A. Bauzá, A. Frontera, R. Weiss and J. J. Weigand, Dalton Trans., 2016, 45, 11384.
22. N. Tokitoh, A. Tsurusaki and T. Sasamori, Phosphorus, Sulfur Silicon Relat. Elem., 2009, 184, 979.
23. C. N. Smit, T. A. van der Knaap and F. Bickelhaupt, Tetrahedron Lett., 1983, 24, 2031.
24. L. Weber, D. Bungardt, R. Boese and D. Bläser, Chem. Ber., 1988, 121, 1033.
25. D. G. Yakhvarov, E. Hey-Hawkins, R. M. Kagirov, Y. H. Budnikova, Y. S. Ganushevich and O. G. Sinyashin, Russ. Chem. Bull., 2007, 56, 935.
26. X. Li, D. Lei, M. Y. Chiang and P. P. Gaspar, J. Am. Chem. Soc., 1992, 114, 8526.
27. K. B. Dillon, V. C. Gibson and L. J. Sequeira, J. Chem. Soc., Chem. Commun., 1995, 2429.
28. B. M. Cossairt and C. C. Cummins, New J. Chem., 2010, 34, 1533.
29. S. Shah and J. D. Protasiewicz, Chem. Commun., 1998, 1585.
30. P. Le Floch, Coord. Chem. Rev., 2006, 250, 627.
31. Experimental and computational details, and details on the X-ray diffraction studies are included in the ESI.†.
32. K. S. A. Motolko, D. J. H. Emslie, H. A. Jenkins and J. F. Britten, Eur. J. Inorg. Chem., 2017, 2920.
33. C. Frenzel and E. Hey-Hawkins, Phosphorus, Sulfur Silicon Relat. Elem., 1998, 143, 1.
34. J. D. Protasiewicz, Eur. J. Inorg. Chem., 2012, 4539.
35. M. Yoshifuji, I. Shima, N. Inamoto, K. Hirotsu and T. Higuchi, J. Am. Chem. Soc., 1981, 103, 4587.
36. J. D. Protasiewicz, M. P. Washington, V. B. Gudimetla, J. L. Payton and M. C. Simpson, Inorg. Chim. Acta, 2010, 364, 39.
37. T. Sasamori, N. Takeda and N. Tokitoh, J. Phys. Org. Chem., 2003, 16, 450.
38. P. Pyykkö and M. Atsumi, Chem.–Eur. J., 2009, 15, 12770.
39. D. Dhara, D. Mandal, A. Maiti, C. B. Yildiz, P. Kalita, N. Chrysochos, C. Schulzke, V. Chandrasekhar and A. Jana, Dalton Trans., 2016, 45, 19290.
40. A. Aistars, R. J. Doedens and N. M. Doherty, Inorg. Chem., 1994, 33, 4360.
41. R. Appel and H. Schöler, Chem. Ber., 1977, 110, 2382.
42. S. S. Chitnis, A. P. M. Robertson, N. Burford, J. J. Weigand and R. Fischer, Chem. Sci., 2015, 6, 2559.
43. J. J. Weigand, N. Burford, A. Decken and A. Schulz, Eur. J. Inorg. Chem., 2007, 4868.
44. A. P. M. Robertson, S. S. Chitnis, H. A. Jenkins, R. McDonal, M. J. Ferguson and N. Burford, Chem.–Eur. J., 2015, 21, 7902.
45. T. L. Breen and D. W. Stephan, J. Am. Chem. Soc., 1995, 117, 11914.
46. C. C. Cummins, R. R. Schrock and W. M. Davis, Angew. Chem., Int. Ed. Engl., 1993, 32, 756.
47. F. Basuli, J. Tomaszewski, J. C. Huffman and D. J. Mindiola, J. Am. Chem. Soc., 2003, 125, 10170.
48. G. Zhao, F. Basuli, U. J. Kilgore, H. Fan, H. Aneetha, J. C. Huffman, G. Wu and D. J. Mindiola, J. Am. Chem. Soc., 2006, 128, 13575.
49. T. Beweries, M. Haehnel and U. Rosenthal, Catal. Sci. Technol., 2013, 3, 18.
50. S. Xin, H. G. Woo, J. F. Harrod, E. Samuel and A.-M. Lebuis, J. Am. Chem. Soc., 1997, 119, 5307.
51. S. Xin, J. F. Harrod and E. Samuel, J. Am. Chem. Soc., 1994, 116, 11562.
52. L. Weber, Chem. Rev., 1992, 92, 1839.
53. P. S. Elmes, M. L. Scudder and B. O. West, J. Organomet. Chem., 1976, 122, 281.
54. J. Chatt, P. B. Hitchcock, A. Pidcock, C. P. Warrens and K. R. Dixon, J. Chem. Soc., Dalton Trans., 1984, 2237.
55. J. Ho, T. L. Breen, A. Ozarowski and D. W. Stephan, Inorg. Chem., 1994, 33, 865.
56. S. Kurz and E. Hey-Hawkins, J. Organomet. Chem., 1993, 462, 203.
57. M. Schaffrath, A. Villinger, D. Michalik, U. Rosenthal and A. Schulz, Organometallics, 2008, 27, 1393.
58. M. Baudler, K. Glinka, A. H. Cowley and M. Pakulski, Organocyclophosphanes, in Inorg. Synth., ed. H. R. Allcock, 2007.
59. J. R. Goehrlich and R. Schmutzler, Z. Anorg. Allg. Chem., 1994, 620, 173.
60. H. Köpf and R. Voigtländer, Chem. Ber., 1981, 114, 2731.
61. A. T. Normand, Q. Bonnin, S. Brandès, P. Richard, P. Fleurat-Lessard, C. H. Devillers, C. Balan, P. Le Gendre, G. Kehr and G. Erker, Chem.–Eur. J., 2019, 25, 2803.
62. J. Pinkas, R. Gyepes, I. Císařová, J. Kubišta, M. Horáček, N. Žilkováa and K. Mach, Dalton Trans., 2018, 47, 8921.
63. M. Polášek and J. Kubišta, J. Organomet. Chem., 2007, 692, 4073.
64. X. Yan and C. Xi, Acc. Chem. Res., 2015, 48, 935.
65. T. Baumgartner and R. Réau, Chem. Rev., 2006, 106, 4681.

---

Footnote

† Electronic supplementary information (ESI) available: Synthesis and characterization of compounds, NMR spectra, crystallographic, EPR and computational details. CCDC 1915056–1915060. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c9sc02322d

---