A Convenient Synthetic Protocol to 1,2-Bis(dialkylphosphino)ethanes.

1,2-Bis(dialkylphosphino)ethanes are readily prepared from the parent phosphine oxides, via a novel sodium aluminium hydride/sodium hydride reduction protocol of intermediate chlorophosphonium chlorides. This approach is amenable to multi-gram syntheses, utilises readily available and inexpensive reagents, and benefits from a facile non-aqueous work-up in the final reductive step.

Organophosphines (R3P) are prime candidates for use as ancillary ligands in many transition metal-catalysed transformations, and are arguably the most important class of ligator in organometallic chemistry.[1] Substitution at the phosphorus atom has a considerable effect on the electronic and steric influence of the ligand over the metal centre, which enables fine tuning of catalyst reactivity and stability.[2] Alkylphosphines are more electron-donating than their arylphosphine counterparts and therefore more effective in promoting key oxidative addition steps common to catalytic cycles (e.g., alkene hydrogenation, hydroformylation).[3] In particular, 1,2-bis(dialkylphosphino)ethanes (R2PCH2CH2PR2) have found myriad applications in a number of important transformations as electron-rich chelating ligands. These include heterolytic H2 activation (Co, Rh, Group 10 triad),[4] C–H bond activation (Pt),[5] CO2 reduction (Fe, Ni),[6] N2 fixation (Fe, Mo),[7] and have facilitated the isolation of structurally interesting molecules incorporating ligand C–H agostic interactions (Ti)[8] and recently a σ-alkane complex (Rh).[9] Despite their utility, the syntheses of these simple yet versatile ligands can be lengthy and involve dangerous (highly toxic and pyrophoric) or costly reagents; subsequently, related arylphosphine ligands that are far less air-sensitive and more readily prepared have hitherto been utilised to a greater extent. The difficulty in preparing alkyl bisphosphines is conveniently illustrated by the synthetic protocols reported for 1,2-bis(dimethylphosphino)ethane (Me2PCH2CH2PMe2, dmpe; Scheme 1). Generation of a phosphide precursor involves either NaNH2/NH3(l) deprotonation of pyrophoric Me2PH (from PH3),[10] or Na reduction of Me2PCl/Me2P–PMe2.[11] Other modifications have utilised Me2P(S)–P(S)Me2,[12, 13] a precursor to Me2P–PMe2, which is prepared via a dangerous reaction of MeMgI with P(=S)Cl3, that has in one case led to a severe accident.[14] Alkylation of Cl2PCH2CH2PCl2 affords R2PCH2CH2PR2 species in moderate yields, yet synthesis of the precursor requires specialised autoclave apparatus (P/PCl3/CH2=CH2; 200 °C, 70 atm)[15] or chlorination of highly pyrophoric H2PCH2CH2PH2.[16] Whilst Cl2PCH2CH2PCl2 is commercially available, its expense limits large-scale syntheses of R2PCH2CH2PR2.

Scheme 1

Previous synthetic routes to Me2PCH2CH2PMe2 (dmpe; 3a).

We required multi-gram quantities of dmpe (and related alkylphosphines), and envisioned a route via reduction of the bisphosphine oxide Me2P(=O)CH2CH2P(=O)Me2 (3a, Scheme 2). However, common strategies used for the deoxygenation of phosphine oxides [for example, electrophilic aluminium hydrides,[17] hydrosilanes][18] require aqueous work-ups that can result in considerable by-product waste, separation and purification issues on scale-up, and accordingly we wished to devise a novel reduction strategy which would avoid an aqueous work-up in the final step.

Scheme 2

Synthesis of Me2PCH2CH2PMe2 (dmpe; 3a): (i) 6 equiv. MeMgCl/THF; (ii) 0.5 equiv. ClCH2CH2Cl, reflux then K2CO3(aq.); (iii) 2.1 equiv. (COCl)2/CH2Cl2.

(n-Bu)2P(=O)CH2CH2P(=O)(n-Bu)2 has previously been prepared from the condensation of (n-Bu)2P(=O)H with XCH2CH2X (X=Cl, Br) using concentrated KOH(aq) in DMSO.[19] Hays et al. documented the synthesis of secondary phosphine oxides [R2P(=O)H; R=Me, Et][20] from the reaction of RMgCl and inexpensive diethyl phosphite [HP(=O)(OEt)2], postulating formation of a solution-phase active intermediate of the form [R2P(=O)MgCl]. This species also acts as a P-centred nucleophilic synthon [R2P(=O)]− in R2P(=O)–C bond formation, yet very few examples of directly producing tertiary phosphine oxides using this method have been documented.[21]

In order to avoid the prior synthesis of Me2P(=O)H, we reacted MeMgCl and HP(=O)(OEt)2 (3:1) in THF, which led to evolution of CH4 and proposed formation of [Me2P(=O)MgCl]. Electrophilic trapping of the latter in situ with 1,2-dichloroethane and subsequent K2CO3(aq) work-up afforded the target Me2P(=O)CH2CH2P(=O)Me2 (1a) in excellent yield (Scheme 2). Hence the entire carbon-phosphorus skeleton of the target phosphine can be assembled in a one-pot protocol.

Direct deoxygenation of 1a was attempted using a variety of standard protocols: LiAlH4 and MeOTf/MeI/Me3SiCl,[22] DIBAL-H,[23] AlH3,[24] Ti(O-i-Pr)4 and PMHS,[25] Cu(OTf)2 and TMDS;18i in each case poor conversions (<10%) and product separation difficulties were encountered. This latter effect is attributed to the strongly electron-donating character of the chelating alkylphosphine. Rajendran and Gilheany recently reduced a variety of phosphorus(V) oxides to their corresponding phosphorus(III) boranes using NaBH4, via their chlorophosphonium(V) chlorides.[26] Accordingly, reaction of 1a with (COCl)2 in CH2Cl2 rapidly affords [Me2P(Cl)CH2CH2P(Cl)Me2][Cl]2 (2a) as a poorly soluble moisture-sensitive solid, in almost quantitative yield (Scheme 2). Whilst the NaBH4 reduction of 2a reaction did cleanly form (dmpe)⋅(BH3)2 (31P NMR: δ=8.52 ppm; THF), dissociation of this strong adduct could neither be achieved by heating in vacuum (150 °C, 10−3 mbar), nor using DABCO deprotection.[27] Other methods for deboronation of electron-rich phosphines using strong acids (e.g., HBF4⋅OMe2, CF3SO3H) and subsequent alkaline hydrolysis[28] were deemed incommensurate with the goals of our synthetic methodology.

Recognising that the soft-soft interaction between BH3 and a tertiary alkylphosphine might be too strong to permit isolation of our target phosphines, our attention turned to the use of NaAlH4 whereby the harder AlH3 by-product should lead to weaker donor-acceptor adducts. Gratifyingly, an NMR-scale reaction of 2a with NaAlH4 in THF afforded uncoordinated dmpe (3a) in quantitative yield (calibrated PPh3 insert), as ascertained by the strongly shielded 31P NMR resonance (δ=−48.3 ppm). 2 equivalents of NaAlH4 are necessary to effect full conversion of 2a to 3a as judged by the disappearance of AlH4− (27Al NMR: δ=98.7 ppm, quintet, 1JAl,H=174 Hz) in solution, thus implying that AlH2Cl is the end-product after hydride transfer under this stoichiometry. Although 3a is hydrocarbon-soluble, it could not be isolated from NaCl and AlH2Cl⋅THF by pentane extraction of the residue obtained upon removal of THF solvent. However, upon re-addition of THF to the solids, free 3a was again observed by 31P NMR spectroscopy. Thus, it appears that the Lewis acidic AlH2Cl preferentially binds the harder O-donor THF when it is in excess, but upon solvent removal it binds the softer phosphine, rendering it a hydrocarbon-insoluble strongly bound adduct that inhibits mechanical separation.

In order to solve this predicament, we adapted the known reaction of AlCl3 and MH (M=Li, Na, K) to form MCl and MAlH4 salts,[29] recognising that incipiently formed AlH2Cl or AlH3 could regenerate AlH4−, thus blocking Al Lewis acids from binding 3a. Satisfyingly, addition of activated NaH30a,b to the reaction mixture of 3a formed from 2 NaAlH4/2a resulted in immediate appearance of the AlH4− resonance in the 27Al NMR spectrum, demonstrating that regeneration of NaAlH4 by NaH is facile under these conditions. Additionally, 2a could be reacted on a 5–20-gram scale with NaAlH4/activated NaH (1:2:4 ratio) to afford 3a in high yield (Scheme 2), which was purified by vacuum distillation. Work-up is facile and consists of simple filtration from NaCl, pentane extraction, and removal of volatiles. NaAlH4 was used in place of LiAlH4 since it is considerably less hazardous[31] due to its greater thermal stability and furthermore, the poorly soluble NaCl by-product precipitates from THF solution, thus simplifying the extraction process. Advantageously, NaAlH4 was readily recovered in near quantitative yield (>95%), and could be used in further reactions with no impaired reactivity.[32]

In order to test the scope of our reaction sequence, we synthesised the corresponding R2P(=O)CH2CH2P(=O)R2 (1b–e) from RMgCl (R=Et, i-Pr, i-Bu) or RLi (R=t-Bu),[33] and reduced them to R2PCH2CH2PR2 (3b–e) via­ 2b–e respectively; comparable yields are obtained in each case (Table 1). Particularly noteworthy is that all reactions scale-up linearly with no ensuing complications, even for sterically bulky substrates (R=i-Pr, t-Bu), and the syntheses have frequently been performed on a 5-30-gram manipulation.

Table 1

31P NMR spectral data and yields for compounds

Compound		R	31P (δ [ppm])[a]	Yield [%][c]
	1a	Me	42.05	84
	1b	Et	51.21	78
	1c	i-Pr	56.07	76
	1d	i-Bu	46.43	92
	1e	t-Bu	60.09	85
	2a	Me	–[b]	96
	2b	Et	107.21	97
	2c	i-Pr	114.47	94
	2d	i-Bu	99.03	96
	2e	t-Bu	120.01	96
	3a	Me	−48.79	73
	3 b	Et	−18.77	84
	3c	i-Pr	9.12	85
	3d	i-Bu	−35.96	73
	3e	t-Bu	35.72	85

[a] NMR spectra for compounds 1a–e, 2b–e were recorded in CDCl3; those of 3a–e were recorded in C6D6.

[b] Compound insoluble in all common solvents attempted.

[c] Purified yields (distillation/sublimation).

Subsequently we sought to investigate the possibility of extending the protocol to bisphosphines with greater functionality. The tridentate ‘pincer’ ligand 1,2-bis(di-tert-butylphosphinomethyl)pyridine (3f) has shown itself to be a strongly electron-donating framework which forms structurally interesting and catalytically active complexes with an extensive range of transition metals.[34] Pyridine 3f is currently prepared from expensive HP(t-Bu)2 or ClP(t-Bu)2 using the methods of Kawatsura and Hartwig34a or Milstein.34c In view of the ability of LiAlH4 to attack pyridine and form lithium tetrakis(N-dihydropyridyl)aluminate isomers,[34] we thought that 3f would be a viable target to test the selectivity of our reductive step. Gratifyingly, 2,6-bis(chloromethyl)pyridine was phosphorylated with the [(t-Bu)2P(=O)Li] reagent to provide 1,2-bis(di-tert-butylphosphinomethyl)pyridine P,P′-dioxide (1f; 72% yield from diethyl phosphite), which was successfully converted (via­ 2f) to 3f in 79% yield on a multi-gram scale (Scheme 3).

Scheme 3

Novel synthesis of 1,2-bis(di-tert-butylphosphinomethyl)pyridine (3f); (i) 6 equiv. t-BuLi/Et2O/2 equiv. HP(=O)(OEt)2; (ii) excess (COCl)2/CH2Cl2; (iii) 2 equiv. NaAlH4/4 equiv. NaH/THF.

In conclusion, we have developed a new, high yielding and facile synthetic route to 1,2-bis(dialkylphosphino)ethanes which uses inexpensive reagents,[35] mild conditions, and demonstrates versatility for modulation of the alkyl groups and ligand backbone. Advantageously, the phosphine oxide and chlorophosphonium chloride precursors are readily handled solids, and amenable to large-scale syntheses. The novel NaAlH4/NaH reductive protocol is highly economical, and avoids issues encountered with other common reduction methods (boranes and silanes), for a convenient extraction process that obviates the need for an aqueous work-up. We anticipate the improved availability of these desirable alkylated bisphosphines will stimulate an enhanced interest in the study of their diverse and effectual coordination chemistry, which can be extended to other substrates incorporating dialkylphosphine moieties.

Experimental Section

General Remarks

All chemical manipulations were performed under an N2 atmosphere either using standard Schlenk-line techniques or in a MBraun Labmaster DP glovebox, unless stated otherwise; in particular, the phosphine compounds 3a–f are highly oxygen sensitive. Solvents were purchased from VWR: pentane and CH2Cl2 were dried using an Innovative Technology Pure Solv SPS-400; THF and Et2O were distilled from purple Na/benzophenone indicator. Solvents were degassed by thorough sparging with N2 gas followed by storage in gas-tight ampoules over suitable drying agents: CH2Cl2 (4 Å molecular sieves); pentane, Et2O (K mirror). Deuterated solvents were freeze-thaw degassed, dried, and stored under N2 in gas-tight ampoules: C6D6 (Sigma–Aldrich, 99.6% D; K mirror); CDCl3 (Merck, 99.8% D; 4 Å molecular sieves). Diethyl phosphite (Sigma–Aldrich, 98%) and 1,2-dichloroethane (Sigma–Aldrich, ≥99.0%) were degassed by thorough sparging with N2 and stored over 4 Å molecular sieves. Oxalyl chloride (Sigma–Aldrich, 98%), NaAlH4 (Sigma–Aldrich, hydrogen-storage grade) and NaH (60 wt% dispersion in oil) were used as supplied. MeMgCl (3.0 M in THF), EtMgCl (2.0 M in THF), (i-Pr)MgCl (2.0 M in THF), (i-Bu)MgCl (2.0 M in THF), and t-BuLi (1.7 M in pentane) were purchased from Sigma–Aldrich and freshly titrated against sec-butanol/1,10-phenanthroline indicator prior to use. 2,6-Bis(chloromethyl)pyridine was prepared according to a literature procedure.[36]

NMR, HR-MS, IR and elemental analysis data for the compounds are presented in the Supporting Information. NMR spectra were recorded using Bruker AV-400 (400 MHz) spectrometers. Chemical shifts, δ, are reported in parts per million (ppm). 1H and 13C{1H} chemical shifts are given relative to Me4Si and referenced internally to the residual proton shift of the deuterated solvent employed. 27Al and 31P{1H} NMR chemical shifts were referenced (δ=0) externally to 1 M AlCl3 (aq.) and 85% H3PO4 (aq.). Air or moisture sensitive samples were prepared inside the glovebox using NMR tubes fitted with J. Young valves. High resolution mass spectrometry samples (HR-MS; EI and ESI) were recorded using either a Micromass Autospec Premier or a Micromass LCT Premier spectrometer. Infrared (IR) spectra were recorded as Nujol mulls using a Perkin–Elmer FT-IR Spectrum GX spectrometer. Elemental analysis was performed by Mr. S. Boyer of the London Metropolitan University.

General Procedure for the Synthesis of 1,2-Bis(dialkylphosphoryl)ethanes (1a–e)

Examplary synthesis of 1,2-bis(dimethylphosphoryl)ethane ( diethyl phosphite (51.5 mL, 0.40 mol) was added dropwise via a dropping funnel to a 3.01 M THF solution of MeMgCl (400 mL, 1.20 mol) at 0 °C. CH4 gas was evolved during addition and the excess pressure vented via a paraffin oil bubbler. The mixture was stirred at 0 °C for 30 min followed by 6 h at room temperature. The resulting grey suspension was cooled to 0 °C and 1,2-dichloroethane (15.8 mL, 0.201 mol) was added slowly. The mixture was heated to reflux for 12 h yielding a viscous grey suspension which was poured into aqueous K2CO3 (166 g, 1.20 mol, 200 mL) and the THF/H2O mixture removed by decanting. The remaining white precipitate was washed with hot MeOH (6×300 mL) and the combined MeOH washings were concentrated under vacuum. The oily solid obtained was dissolved in CHCl3 (300 mL) and the solution dried over Na2SO4, filtered and evaporated. The resulting solids were dried under vacuum at 80 °C to yield 1a as a hygroscopic white powder; yield: 30.2 g (84%). 1H NMR (400 MHz, CDCl3): δ=1.98 (d, J=2.6 Hz, 4 H, CH2), 1.53 (m, 12 H, CH3); 13C{1H} NMR (100 MHz, CDCl3): δ=24.3–23.1 (m), 16.8–15.7 (m); 31P{1H} NMR (162 MHz, CDCl3): δ=42.1 (s). HR-MS (ESI): m/z=183.0704, calcd. for C6H17O2P2: 183.0704; IR (nujol): ν=115 cm−1:6 (s, νPO); anal. calcd. for C6H16O2P2: C 39.57, H 8.85; found: C 39.66, H 8.78.

General Procedure for the Synthesis of Ethylenebis(dialkylchlorophosphonium) Dichlorides (2a–e)

Examplary synthesis of ethylenebis(dimethylchlorophosphonium) dichloride ( to a solution of 1a (20 g, 110 mmol) in CH2Cl2, was added dropwise (COCl)2 (19.5 mL, 230 mmol) with stirring (caution: gas evolution) resulting in precipitation. The reaction was allowed to stir for 1 h at room temperature after which the suspension was filtered, washed with Et2O, and dried under vacuum to afford 2a as a white powder; yield: 30.7 g (96%); anal. calcd. for C6H16Cl4P2: C 24.68, H 5.52; found: C 24.76, H 5.46.

General Procedure for the Synthesis of 1,2-Bis(dialkylphosphino)ethanes (3a–e)

Examplary synthesis of 1,2-bis(dimethylphosphino)ethane ( NaH (60 wt% in mineral oil, 11.52 g, 288 mmol) was placed in a Schlenk tube and washed with THF (2×50 mL) to remove mineral oil. Subsequent activation[30] of NaH was achieved by stirring with a 1 M solution of LiAlH4 in THF (30 mL, 30 mmol) for 2 h, which was then removed by cannula filtration and the solid rinsed with a further 2×20 mL portions of THF. 100 mL of fresh THF were then added and the suspension cooled to −78 °C before addition of NaAlH4 (8.14 g, 151 mmol) as a solid under a flush of N2. 2a (20 g, 68.5 mmol) was then added as a suspension in 150 mL of THF at −78 °C via a wide Teflon cannula. Reaction proceeded immediately with accompanying gas evolution (caution). The mixture was allowed to warm to room temperature, stirred for a further 1 h, and then filtered through a glass frit with a pad of Celite® and washed with THF (2×50 mL). THF was removed under reduced pressure (20 mmHg) and the resulting white solids extracted with pentane (4×150 mL); remaining solid NaAlH4 was washed with cold Et2O and reused in further reactions (recovery: 7.7 g; 95%). The pentane was then removed under reduced pressure (20 mmHg), leaving behind a colourless oil which was vacuum distilled at 58 °C (5 mmHg) to afford 3a as a colourless oil; yield: 7.44 g (72 %). 1H NMR (400 MHz, C6D6): δ=1.30 (m, 4 H, CH2), 0.81 (t, J=1.5 Hz, 12 H, CH3); 13C{1H} NMR (100 MHz, C6D6): δ=28.2 (s), 13.9–14.1 (m); 31P{1H} NMR (162 MHz, C6D6): δ=−48.8 (s); HR-MS (EI): m/z=150.0726, calcd. for C6H16P2: 150.0727; anal. calcd. for C6H16P2: C 48.00, H 10.74; found: C 48.23, H 10.64.