Convenient and Scalable Synthesis of Aryldichlorophosphines and Primary Arylphosphines via Perthiophosphonic Anhydrides

A scalable synthetic route to both primary arylphosphines ArPH 2 and aryldichlorophosphines ArPCl 2 is reported. The C–P bond for- mation was performed in a highly regiospecific manner through elec-trophilic substitution of selected aromatic hydrocarbons (ArH) with phosphorus pentasulfide. The resultant perthiophosphonic anhydrides Ar 2 P 2 S 4 were then reacted with LiAlH 4 to give primary phosphines ArPH 2 . Subsequent reaction of ArPH 2 with phosgene solution gives di- chlorophosphines ArPCl 2 . Each reaction step requires minimum purifi- cation and uses commercially available and economical precursors. The scope of the reaction was shown to include alkoxy and phenoxy substi- tuted benzenes as well as naphthalene and fluorene as starting materials.

Both primary phosphines RPH 2 and dichlorophosphines RPCl 2 have a rich chemistry and are valuable reagents in reactions such as hydrophosphinations, dehydrocoupling and P-C bond formation. [1][2][3] This is facilitated by the reactive nature of P-H and P-Cl bonds, respectively.
However, this also means that primary phosphines, particularly those with a low molecular weight and minimal steric bulk protecting the -PH 2 moiety, are extremely pyrophoric, rendering them difficult to synthesise and manipulate. Typically, primary phosphines are synthesised by the reaction between either alkyl/aryl dichlorophosphines (Scheme 1, route (i)) or phosphonates (Scheme 1, route (ii)) and a strong reducing agent. [4][5][6] Dichlorophosphines appear to be ubiquitous within synthetic organophosphorus chemistry; [7][8][9][10][11][12] despite this, the commercial availability of aryldichlorophosphines is limited and so laboratory-scale syntheses have to be employed where access to a wider range of these compounds is required. A major challenge to overcome in these syntheses is the need for a highly selective formation of the desired species, which is crucial due to the limited options for postsynthetic purification, generally limited to fractional distillation (for sufficiently volatile species). Other options (in particular chromatography) are generally not accessible due to the very reactive nature of dichlorophosphines towards both oxygen and moisture. P-C bond formation is a crucial step in the synthesis of aryldichlorophosphines. Early examples involved reacting aromatic hydrocarbons or anisol derivatives with PCl 3 in the presence of a Lewis acid catalyst such as AlCl 3 or SnCl 4 (Scheme 2, route (i)). 13,14 Later examples used the reaction of PCl 3 with respective Grignard (or organolithium) reagent (Scheme 2, route (ii)) and this method is still routinely employed. 15,16 The disadvantage of this synthetic route is that due to the high reactivity of the Grignard (or organolithium) reagent, often the products of multiple substitution (Ar 2 PCl and Ar 3 P) are formed in significant amounts, impacting the yield of the dichlorophosphine. This synthetic route also suffers from poor functional group tolerance. To prevent the formation of byproducts due to multiple substitution, a protective group strategy was devised in which PCl 3 was replaced with ClP(NR 2 ) 2 (Scheme 2, route (iii)). The intermediate aminophosphinyl species ArP(NR 2 ) 2 is formed, which is isolated and then reacted with excess HCl gas to form the dichlorophosphine. 4,17 While this synthetic route does prevent multiple substitution from occurring, a new challenge is presented through the use of HCl gas and separation of the dichlorophosphine from the coformed dialkylammonium chloride.

Scheme 1 Examples of primary phosphine syntheses
More recently a systematic investigation of the synthesis of aryl-and heteroaryldichlorophosphines was reported by Karaghiosoff 18 in which organozinc reagents were employed in the place of Grignard or organolithium reagents (Scheme 2, route (iv)). Lower polarity of C-Zn vs. C-Mg and C-Li bonds resulted in more controlled reactivity of organozinc species compare to Grignard reagents and organolithiums. Whilst this route presents improvement over the more conventional routes as it does avoid multiple substitution products and offers good functional group tolerance, it requires observing the right stoichiometry carefully, and distillation (or recrystallisation for solids) of the highly reactive product as a final purification step.
As outlined above, the synthesis of dichlorophosphines is not straightforward, with multiple issues presented for each synthetic method. Our recent investigations required a synthesis of a series of aryldichlorophosphines to allow for fine tuning of electronics of a target phosphorus-containing molecule. This prompted the synthesis of a series of aryldichlorophosphines, each in multigram quantities, with differing aryl groups. To synthesise these we have expanded on our previously reported 'niche' synthetic route, originally used to form dichloroferrocenylphosphine FcPCl 2 , via the perthiophosphonic anhydride, Fc 2 P 2 S 4 (Scheme 3). 19 This unique method offers a route to a previously difficult to access dichlorophosphine FcPCl 2 in high yield, is easily scalable, uses economical commercially available precursors, and does not require complex purification. The P-C bond-forming step, in which the perthiophosphonic anhydride Fc 2 P 2 S 4 is formed from ferrocene and P 4 S 10 , proceeds fully regioselectively with high yield (>80%). 19 Herein, we report the expansion of this synthetic route to make a selection of primary and dichloroaryl phosphines.

Perthiophosphonic Anhydrides Synthesis
The synthetic route towards primary and dichloroaryl phosphines (Scheme 4) begins with the synthesis of the perthiophosphonic anhydride compounds 1A-F. These six compounds were selected due to the ease with which they can be synthesised; this is achieved by a simple heating of P 4 S 10 with the related hydrocarbon precursor -anisole, phenetole, 2-tert-butylanisole, diphenyl ether, fluorene and naphthalene, respectively, at high temperature (150-190°C , reaction time 1-5 hours). All of these compounds have previously been reported in the literature, [20][21][22] with the exception of the fluorene derivative 1F. The arene reactants served also as solvents for the reaction; hence, the molar ratio of reactants 1:2 to 1:2.5 was used (P/ArH). The reactions were performed under a gentle stream of nitrogen and the exhaust gasses were bubbled through an aqueous NaOH solution to remove H 2 S formed as a by-product. The liquid mixtures obtained after heating were left to cool to room temperature and the formed solid products were filtered off, washed with dichloromethane or diethyl ether and dried in vacuo. Compounds 1A-F form as yellow or white solids in yields in the range of 31-65% (see Table 1). They can be manipulated in air; however, they hydrolyse slowly, hence long-term storage requires well-sealed vials.
While other routes exist for the synthesis of perthiophosphonic anhydrides, these are generally more complex and require several steps, hence are less suitable to prepare starting materials for the synthetic sequence in this work. 23,24 An important point to make is that the arenes selected in this work contain no functional groups that will provide additional reactivity toward P 4 S 10 , such as ketones, esters and alcohols. 25,26 On the other hand, the presence of electron-donating groups (such as OMe) on the aromatic ring results in improved yields and shorter reaction times. Due to the inherent insolubility of perthiophosphonic anhydrides, no NMR data could be collected for 1A-F. Nevertheless, the follow-on reactivity (see below) indicates the P-C bond formation is fully regiospecific in the reactions of

Special Topic Synthesis
P 4 S 10 with both activated and non-activated arenes used in this study, with no other regioisomers detected by 31 P{ 1 H} NMR analysis in the phosphines 2A-F formed from 1A-F in the next step.

Reduction to Primary Phosphines
With the desired Ar 2 P 2 S 4 compounds in hand, the next step was to reduce these with LiAlH 4 (4 equiv of LiAlH 4 per Ar 2 P 2 S 4 were used) to the corresponding primary phosphines ArPH 2 (2A-F). Both Ar 2 P 2 S 4 and LiAlH 4 were suspended in Et 2 O, and the two suspensions were added together slowly at 0 °C. The resulting suspension was filtered, degassed water was added, and the mixture was filtered a second time. In both filtrations, efficient washing of the solid on the filter was essential to achieve good yields. The filtrate and washings were collected, and the volatiles were removed in vacuo to yield the desired primary phosphines. No further purification was performed, and phosphines 2A-D were obtained in good purity and reasonable yields (33-52%). The naphthalene species 2F was obtained in 8% yield only and small amount of naphthalene (formed in the reduction step rather than carried over from previous step) was present, as shown by 1 H NMR analysis. Also, the reduc-tion of 1E led to partial cleavage of the C-P bond, with small amounts of fluorene being detected by 1 H NMR analysis alongside the major product 2E. Despite this, both 2E and 2F were of sufficient purity for further synthetic use and were used as obtained in the preparations of respective dichlorophosphines as described below.
Other reducing reagents (NaH and NaBH 4 ) were tested for the reduction of 2A-F; however, no phosphine was produced even at elevated temperatures in ethereal solvents.
Interestingly, of the six primary phosphine compounds synthesised in this work, only two have been previously reported in the literature (2A and 2F), 27,28 demonstrating the ability of this synthetic route to provide access to a wider range of primary phosphines. Despite the lack of steric bulk protection, phosphines 2E and 2F showed remarkable stability in air in both the solid state and in solution, with minimal oxidation observed. This enhanced stability to oxidation could be due to the conjugated aromatic system of naphthalene and fluorene, as the additional conjugation has been shown to stabilise primary phosphines against oxidation. 27

Special Topic Synthesis
of 199-202 Hz as expected. The purity of the products was further assessed by 1 H and 13 C{ 1 H} NMR analysis. In addition to multinuclear NMR spectroscopy, the novel compounds 2B-E were characterised by MS analysis. The spectroscopic data obtained by us for the previously reported species (2A and 2F) were in agreement with the literature (see Experimental Section).
For 2B, 2D and 2F, minor impurities were observed in the 31 P{ 1 H} spectra as two singlets at approximately  P -70 ppm (2-4% of integral intensity). These were assigned as the respective diphosphines ArP(H)-P(H)Ar, which, due to the presence of two chiral P atoms, exist in two diastereomeric forms (meso and rac). 30 In the 31 P NMR spectra, these minor signals split into symmetrical multiplets with pattern consistent with a AA′XX′ spin system (A, A′ = P, X, X′ = H). Spin system simulations were carried out to replicate the observed splitting pattern for selected examples ( Figure  1). 31 These simulations yielded 1 J PP =150 Hz, 1 J PH = 150 Hz and 2 J PH = 10 Hz for one of the diastereomers of ArP(H)-P(H)Ar (Ar = p-EtOC 6 H 4 ), which is fully consistent with the suggested diphosphine structure. Note the contribution of the ortho protons from the aryl groups has been omitted from the simulated spectrum due to the extra complexity this presents.
The diphosphine impurities presented no issues for the subsequent chlorination step as they were present in very small amounts and were converted into the same end product (ArPCl 2 ) on chlorination. Hence, no attempt was made to remove these through further purification.

Chlorination to Dichlorophosphines
In the last step of the synthetic sequence shown in Scheme 4, the primary phosphines 2A-F were chlorinated to the aryldichlorophosphines 3A-F. A commercially available solution of phosgene in toluene (slight excess, 2.1 equiv) was added slowly, at -10 °C, to the solution of primary phosphine. The reaction mixture was left to stir overnight and subsequent removal of the volatiles in vacuo afforded 3A-F. The dichlorophosphines were isolated in yields of 74-90% and were obtained as oils, except for 3F, which was isolated in 46% yield as an off-white solid. The 31 P{ 1 H} spectra of 3A-F were as expected (singlets within a narrow range of  P 160-164 ppm) and showed that all the primary phosphine starting material had been consumed during the chlorination step, with no other phosphoruscontaining species present. The purity of the products was further confirmed by 1 H and 13 C{ 1 H} NMR analysis, which indicated some fluorene was present in the sample of 3E, whilst the purity of the other samples was very good. This represents a marked improvement on previously reported methods, where vacuum distillation was required as a final purification step. Stoichiometric amount of triphosgene was used as an alternative chlorinating reagent at room temperature for selected examples of primary phosphines, giving full conversion into the respective dichlorophosphines as judged by 31 P{ 1 H} NMR analysis.
Syntheses of 3B-E have not been reported previously. In addition to multinuclear NMR spectroscopy (as discussed above) the compounds were also characterised by MS analysis. The spectroscopic data obtained by us for previously reported species (3A and 3F) agreed with the literature (see experimental section).
In summary, a multigram synthesis of a series of primary arylphosphines and aryldichlorophosphines has been reported. All steps use convenient, commercially available, and economical reagents. Each step requires minimum purification, which is of importance due to the highly reactive

Special Topic Synthesis
nature of both compound types. The initial step (formation of perthiophosphonic anhydrides) proceeds highly regiospecifically, and the nature of subsequent steps means multiple substitution is avoided, hence employing protection and subsequent deprotection strategies is not required. Future work will look to further extend the scope of this reaction to other R 2 P 2 S 4 compounds.
All manipulations (unless indicated otherwise) were performed under an atmosphere of nitrogen using standard Schlenk line techniques or under an atmosphere of argon in a Saffron glove box. Diethyl ether and dichloromethane (DCM) were collected from an MBraun solvent purification system and stored over activated 4Å molecular sieves. 2tert-Butylanisole was prepared according to a literature method; 32 all other reagents were commercially available. All new compounds were characterised via 1 H, 31 P{ 1 H} and 13 C{ 1 H} NMR spectroscopy including the measurement of H-H COSY, H-C HSQC, H-C HMBC, and H-P HMBC. 13 C NMR spectra were recorded using the DEPT-Q pulse sequence. All spectra were recorded at 25 °C with either a Bruker Avance III (500 MHz) spectrometer or a Bruker Avance II (400 MHz) spectrometer. In vacuo refers to pressure of ca. 0.01-0.1 mbar. MS were acquired with a Micromass LCT (electrospray ionisation) from solutions of the analyte in methanol.

Synthesis of Perthiophosphonic Anhydrides 1A-F
All syntheses in this section were performed under a gentle stream of nitrogen, and the exhaust gasses were bubbled through aqueous NaOH solution to remove H 2 S formed as a by-product. The subsequent workup was performed in air (in a well ventilated fumehood).

Lawesson's Reagent, (4-MeO-C 6 H 4 ) 2 P 2 S 4 (1A)
1A was synthesised according to a reported literature procedure. 20 Anisole (48.7 g, 450 mmol) and P 4 S 10 (20.0 g, 45.0 mmol) were heated under reflux at 150 °C for 5 hours. The liquid mixture was allowed to cool to r.t. and a pale-yellow crystalline solid precipitated out, which was isolated by vacuum filtration. The solid was washed with ether (2 × 25 mL) and dried in vacuo.

(3-tBu-4-MeOC 6 H 3 ) 2 P 2 S 4 (1C)
1C was synthesised according to a reported literature procedure. 22 2tert-Butylanisole (27.0 g, 164 mmol) and P 4 S 10 (10.0 g, 22.5 mmol) were combined and heated to 180 °C with stirring for 1 hour. The solution was then allowed to cool to r.t. A crystalline pale-yellow solid precipitated out of solution, which was collected by vacuum filtration, washed with ether (50 mL) and dried in vacuo.

General Procedure for the Synthesis of Primary Phosphines (2A-F)
The respective perthiophosphonic anhydride 1A-F was suspended in diethyl ether (150 mL) and cooled to 0 °C. A suspension of LiAlH 4 (4 equiv) in ether (50 mL) was added in small portions with vigorous stirring. The resulting mixture was stirred for 1 hour at 0 °C. The mixture was then filtered to remove the insoluble solids, which were washed with ether (2 × 10 mL). The filtrate and washings were collected, cooled to 0 °C and degassed water (2 mL) was added cautiously dropwise. The resulting suspension was filtered again to remove the insoluble solids that had formed, and the solid on the filter was washed with DCM (2 × 20 mL). The filtrate and washings were collected, and the volatiles were removed in vacuo to yield the desired primary phosphines 2A-F.

9H-Fluoren-2-ylphosphine (2E)
Starting from 1E (6.00 g, 11.5 mmol), 2E was isolated as a white solid (1.50 g, 7.56 mmol, 33%). The respective primary phosphine (2A-F) was dissolved in DCM (150 mL) and cooled to -10 °C. A solution of phosgene (20% solution in toluene, 2.1 equiv) was added dropwise over 30 minutes. The resulting solution was allowed to warm to r.t. and stirred for a further 5 hours. The volatiles were removed in vacuo to yield the desired dichlorophosphine 3A-F. Safety note: Phosgene and carbon monoxide (evolved in the chlorination reaction) are highly toxic, use of a well-ventilated fumehood is essential for this step.