Chemoselective reduction of the phosphoryl bond of O-alkyl phosphinates and related compounds: An apparently impossible transformation

Article abstract of DOI:10.1039/c5cc06389b

A method is reported for the phosphoryl bond cleavage of O-alkyl phosphinates, phosphinothioates and certain phosphonamidates to furnish the corresponding P(iii) borane adducts. The two-step procedure relies upon initial activation of the phosphoryl bond with an alkyl triflate, followed by reduction of the resulting intermediate using lithium borohydride.

Full text of DOI:10.1039/c5cc06389b

ChemComm
View Article Online
View Journal
COMMUNICATION
Chemoselective reduction of the phosphoryl bond
of O-alkyl phosphinates and related compounds:
an apparently impossible transformation†
Cite this:DOI: 10.1039/c5cc06389b
Received 30th July 2015,
Accepted 18th September 2015
Niall P. Kenny, Kamalraj V. Rajendran and Declan G. Gilheany*
DOI: 10.1039/c5cc06389b
www.rsc.org/chemcomm
A method is reported for the phosphoryl bond cleavage of O-alkyl chloride,12 which is in turn easily reduced with borohydride
phosphinates, phosphinothioates and certain phosphonamidates to the corresponding phosphine borane.8,13,14 Encouraged by
to furnish the corresponding P(III) borane adducts. The two-step this success, we wondered if we could go further and attempt
procedure relies upon initial activation of the phosphoryl bond with the deoxygenation of phosphinates. This would achieve the
an alkyl triflate, followed by reduction of the resulting intermediate apparently impossible task of breaking PQO in the presence of
using lithium borohydride.
P–O. And to make the task even more challenging – there is a
known effect whereby the more P–O bonds present in a P(V)
The development of a general method for the facile and chemo- species, the higher the bond dissociation energy of the asso-
selective reduction of the phosphoryl bond has remained an ciated phosphoryl bond.15 The final products, some of which
issue in chemistry for many decades.1 Traditional methods to could be difficult to synthesise via solely P(III) starting materi-
affect this highly useful transformation have relied upon very als, could find many uses, for instance as ligands in metal
reactive hydride sources, such as silanes2 or hydrides,3 which catalysed transformations.16
often preclude the presence of other susceptible functional
In our initial experiments, we found that neither O-alkyl nor
groups in the substrate, and also have unpredictable stereo- O-aryl phosphinates reacted with oxalyl chloride to give a stable
chemical outcomes.1a,b Progress has accelerated in the inter- chlorophosphonium salt.17 We therefore turned to our second
vening years,4 and, recently, both ourselves5–10 for hydrides and methodology, which is via the reduction of alkoxy phosphonium
Beller and co-workers11 for silanes have each contributed two species,5–7 Scheme 1. The reduction can be done with LAH5,7 or
methodologies such that the chemoselective reduction has preferably L-selectride9 to give the phosphines or with boro-
been solved in certain instances. Now, the phosphoryl bond hydride,5–7,9 preferably lithium borohydride10 to give the phos-
of tertiary and secondary phosphine oxides can be reduced, phine boranes.
with good yield, in the presence of ketones, esters, olefins,
nitriles and aldehydes.8,11b
To our knowledge there is only one report of phosphinate
reduction by such a route, that of Rhomberg and Tavs in
However, there is another type of chemoselectivity that has 1967,18 who used Meerwein’s salt19 for the alkylation followed
seen much less progress. We refer to the reduction of phosphoryl by reduction with magnesium in methanol, but obtained poor
species wherein the phosphorus is formally singly bonded to yields (r35%). Given our own previous success with Meerwein’s
elements other than carbon or hydrogen. Previously, their salt,6 we took that as our starting point but soon found that it was
reduction without concomitant loss of the P–heteroatom bond unsuitable. This is because the higher phosphoryl bond strength
had been an intractable problem. We then reported8 the first necessitates a much higher temperature for the alkylation. For
wide-scope methodology for the reduction of aminophosphine example, O-ethyl diphenylphosphinate 1 reacts to give reasonable
oxides (P(V) compounds with between one and three P–N alkylation yields (480%) only at 160 1C. Since these experiments
bonds). The technique is based on reaction initially with oxalyl had to be conducted in a sealed tube,20 the diethyl ether by-product
chloride, generating the corresponding chlorophosphonium
Centre for Synthesis and Chemical Biology, School of Chemistry,
University College Dublin, Belfield, Dublin 4, Ireland.
E-mail: declan.gilheany@ucd.ie; Fax: +353 1-7162127
† Electronic supplementary information (ESI) available: Syntheses and character-
Scheme 1 Reduction of phosphoryl species via alkoxyphosphonium
isations of compounds 1–8, their analogous boranes and other experimental
intermediates.
details. See DOI: 10.1039/c5cc06389b
This journal is ©The Royal Society of Chemistry 2015
Chem. Commun.

View Article Online
Communication
ChemComm
Scheme 2 Phosphine oxide reduction with Meerwein’s salt/LAH giving
phosphine borane directly (Ar: o-tolyl).
Scheme 4 Competing pathways in borohydride reduction of dialkoxy-
phosphonium triflate.
presents a potential hazard and we were forced to switch to alkyl
triflates as alkylating agents.3b,6,21
fully regenerated. We have previously encountered this phenom-
enon in hydride reductions of pseudo phosphonium salts6,7,9,10 It
is explained by the alternative site of hydride attack23 shown in
Scheme 4 (path A). Fortunately, after some screening (see ESI†), we
found that dilution of the hydride agent suffices to restore the
desired product completely (Scheme 4, path B). Tetrahydrofuran
was identified as the optimum solvent for the reduction step and
dilution of the LiBH4 from 2 M to 0.2 M increased the overall
conversion to product from 0% (starting material fully regener-
ated) to 80% (499% conversion from the intermediate, 73%
isolated yield, for O-ethyl diphenylphosphinate).
Scheme 3 shows the optimised conditions for the two steps
leading to desired P(III) borane product, starting from O-ethyl
diphenylphosphinate 1, showing that indeed a PQO bond can
be removed while leaving a P–O intact.
We then subjected a number of related substrates to the metho-
dology of Scheme 3, with the optimised results given in Table 1.
Digressing briefly, we did make one interesting observation
in these preliminary studies with Meerwein’s salt. When enantio-
enriched tertiary phosphine oxide, (S)-methylphenyl(o-tolyl)-
phosphine oxide (93% ee), was employed, the tetrafluoroborate
salt could be formed by refluxing in chloroform for 2 hours with
1.2 equivalents of Meerwein’s salt. The addition of LAH at room
temperature then gave the corresponding phosphine borane,
with 90% yield, Scheme 2. The reaction is completely stereo-
specific, furnishing the product with inversion of configuration
at phosphorus. Evidently, the excess LAH reduces the tetra-
fluoroborate anion in situ furnishing borane, which then reacts
with the initially formed phosphine.
Returning to the main topic, use of ethyl triflate as alkylating
agent gave the desired dialkoxyphosphonium salt (DiAPS) with
an identical conversion (80%) when reacted with ethyl diphenyl-
phosphinate 1 in chloroform, again only at high temperature
(160 1C). However, remarkably, it was found that switching to
the higher boiling 1,1,2,2-tetrachloroethane (TCE, b.p. 147 1C)
allowed the reaction temperature to be lowered to 100 1C or less
with equivalent results,22 Scheme 3. During these experiments a
peculiar aspect of the alkylation reaction emerged, which is still
not understood: the alkylation proceeds with higher conversion
when activated 4 Å molecular sieves are present in the reaction.
Rather a lot of sieves are necessary to maximise the effect,
approximately 1 g of sieves per 100 mg of phosphorus species.
It is clear that the sieves are not performing their common role of
sequestering water, as all substrates and solutions were confirmed
to be fully dry, by Karl Fischer titration, prior to running of the
reaction, and water is not produced as a by-product.
Table
1
Substrates subjected to two-step reduction according to
Scheme 3 and their optimised conditions
It was found, as expected, that LAH is too harsh for the
reduction of the O,O-dialkyl phosphonium species, cleaving both
of the P–O bonds, and producing the secondary phosphine borane
as the major product. On changing to our preferred lithium
borohydride, a striking dependence on concentration was found. Substrate
Alkylation conditionsa
Timeb
(h)
Temp.
(1C)
Equiv.
ROTf
DiAPS
Borane
conv.c (%)
yieldd (%)
Thus, a 2 M solution in THF, did not produce any P(III) products,
protected or otherwise, instead the starting P(V) phosphinate was
1
2
8
8
6
16
2
23
100
100
100
75
80
80
Various
Various
2
2
2
2
1.66
1.66
80
80
100
86
87
58
73
72e
80
60
50
32
0
3
4f
5g
6g
7
o10
8
o10
0
a
With ethyl triflate unless noted otherwise, all substrates dried thoroughly
b
prior to use. Time periods are strict, with deviation in either direction
lowering conversion. By 31P-NMR spectroscopy. Isolated yield, balance
c
d
e
is returned starting material unless noted otherwise. Balance is rear-
ranged starting material. Methyl triflate used, necessitating a lower
reaction temperature. Reduction performed in IPA/THF (4 : 1).
f
Scheme 3 Optimised conditions for the two-step deoxygenation of
O-ethyl diphenylphosphinate 1 to phosphinite borane.
g
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2015

View Article Online
ChemComm
Communication
The thiophosphoryl analogue O-ethyl diphenylphosphinothioate issue arose with phenyl ester 7, echoing its unreactivity with
2,24 on alkylation, has different potential leaving groups, –OEt and oxalyl chloride.17 It is likely that the bond energies for the
–SEt, present in the intermediate. In the event, the reaction phosphoryl systems of these two species are too high to be
proceeds with exclusive loss of the thiolate moiety, forming O-ethyl adequately activated by alkylation under these conditions.15
diphenylphosphinite borane as the product. Unsurprisingly, given
In conclusion, a method for the phosphoryl deoxygenation
the difference in bond energies, the returned P(V) material (20%) is of O-alkyl phosphinates, phosphinothioates and certain phos-
the rearranged compound, with PQO and P–SEt bonds, analogous phonamidates has been described. The method relies upon
to a thiono–thiolo rearrangement.25
activation of the strong phosphoryl bond prior to introduction
The alkylaryl analogue O-ethyl phenyl(isopropyl)phosphinate of the reductant, as well as the creation of a symmetric
(3) behaved well in the reaction and was convenient in that it phosphonium salt intermediate, to enable overall PQO bond
could be fully alkylated, likely due to the stabilising effect of the cleavage using lithium borohydride. The use of such mild
more electron-donating P-alkyl group on the positively charged conditions should also allow the technique to be extended to
DiAPS intermediate. This allowed us to probe the mechanism of substrates featuring other susceptible functional groups.
alkylation, especially whether it is reversible. To do so, a solution
We sincerely thank Science Foundation Ireland (SFI) for
of O,O-diethyl phenyl(isopropyl)phosphonium triflate was formed funding this chemistry under Grants RFP/08/CHE1251 and
and confirmed to be the sole phosphorus species present in 09/IN.1/B2627. We are also grateful to UCD Centre for Synthesis
solution by 31P-NMR spectroscopy. Then, two equivalents of and Chemical Biology (CSCB) and the UCD School of Chemistry
methyl triflate were added and this mixture was heated to 75 1C for access to their extensive analysis facilities.
for 15 hours. Analysis of the resulting mixture showed the
O,O-diethyl phosphonium species as previously observed, but
now accompanied by the O,O-dimethyl and mixed O-methyl,
O-ethyl salts, showing that indeed the reaction between a phos-
Notes and references
´
1 (a) D. Herault, D. H. Nguyen, D. Nuel and G. Buono, Chem. Soc. Rev.,
phinate and alkyl triflate is in dynamic equilibrium.
2015, 44, 2508; (b) K. Pietrusiewicz and M. Zablocka, Chem. Rev.,
1994, 94, 1375; (c) M. J. Gallagher, in The Chemistry of Organo-
phosphorus Compounds, ed. F. R. Hartley, Wiley-Interscience, Chice-
ster, 1992, ch. 2, vol. 2, pp. 53–76; (d) L. Maier, in Organic Phosphorus
Compounds, ed. L. Maier and G. M. Kosolapoff, Wiley-Interscience,
New York, ch. 1, vol. 1, 1972, pp. 1–226.
2 (a) L. Horner and W. D. Balzer, Tetrahedron Lett., 1965, 6, 1157;
(b) R. Tang and K. Mislow, J. Am. Chem. Soc., 1969, 91, 5645;
(c) K. Naumann, G. Zon and K. Mislow, J. Am. Chem. Soc., 1969,
91, 7012.
3 (a) P. D. Henson, S. B. Ockrymiek and R. E. Markham, J. Org. Chem.,
1974, 39, 2296; (b) T. Imamoto, T. Oshiki, T. Onozawa, T. Kusumoto
and K. Sato, J. Am. Chem. Soc., 1990, 112, 5244; (c) G. Keglevich,
T. Chuluunbaatar, K. Ludanyib and L. Tokec, Tetrahedron, 2000,
56, 1; (d) M. Stankevic and K. M. Pietrusiewicz, Synlett, 2003, 1012.
4 (a) S. C. Berk and S. L. Buchwald, J. Org. Chem., 1992, 57, 3751;
(b) N. J. Lawrence, M. D. Drew and S. M. Bushell, J. Chem. Soc., Perkin
Trans. 1, 1999, 3381; (c) H.-C. Wu, J.-Q. Yu and J. B. Spencer, Org.
Variation of the ester group was explored with O-methyl
species 4. In this case, methyl triflate (b.p. 97–99 1C) was used
to ensure formation of a symmetric triflate salt. A similar
alkylation yield was obtained but a lower yield of the O-methyl
diphenylphosphinite borane was obtained (60%). Once again,
we explain this on the availability of the two different reaction
pathways of reduction (Scheme 4). The lower steric requirement
about the alkoxy a-carbon in the case of a methyl group makes
substitution at that position more competitive with that at the
phosphorus centre, relative to the P–OEt examples, causing a
greater proportion of starting material to be reformed.
Next we sought to expand the scope to phosphonamidates:
analogous P(V) compounds featuring a P–N bond, as well as the
P–O bond (substrates 5 and 6). Under the same conditions as
for phosphinate alkylation, these compounds were found to
undergo significant decomposition via cleavage of the P–N
bond. However, by lowering both the reaction temperature for
the first step, to 80 1C, and the excess of ethyl triflate, to 1.66,
more favourable conversions to the intermediate could be
obtained. The subsequent treatment with lithium borohydride
(0.2 M in THF) once again fully returned the starting phospho-
namidate. However a solvent screen showed that a mixture of
IPA/THF (4 : 1, overall concentration of reductant: 0.4 M) gave
reasonable yields of 50% and 32% for the derived boranes of
5 and 6 respectively. Again, the mass balance mostly consisted
of returned starting material. This also can be explained on the
basis of Scheme 4, in this case a consequence of the participa-
tion of the nitrogen lone pair of electrons in the P–N bonding.26
This would reduce the positive charge on phosphorus in the
intermediate, making it less electrophilic, thereby discouraging
pathway B.
´
Lett., 2004, 6, 4675; (d) C. Petit, A. Favre-Reguillon, B. Albela,
L. Bonneviot, G. Mignani and M. Lemaire, Organometallics, 2009,
28, 6379; (e) C. J. O’Brien, J. L. Tellez, Z. S. Nixon, L. J. Kang,
A. L. Carter, S. R. Kunkel, K. C. Przeworski and G. A. Chass, Angew.
Chem., Int. Ed., 2009, 48, 6836; ( f ) J. Gatignol, C. Alayrac, J.-F.
Lohier, J. Ballester, M. Taillefer and A.-C. Gaumont, Adv. Synth.
Catal., 2013, 355, 2822–2826; (g) L. R. Doyle, A. Heath, C. H. Low and
A. E. Ashley, Adv. Synth. Catal., 2014, 356, 603–608; (h) S. Sowa,
´
M. Stankevic, A. Szmigielska, H. Małuszynska, A. E. Kozioł and
K. M. Pietrusiewicz, J. Org. Chem., 2015, 80, 1672–1688.
5 K. V. Rajendran, J. S. Kudavalli, K. S. Dunne and D. G. Gilheany,
Eur. J. Org. Chem., 2012, 2720–2723.
6 K. V. Rajendran and D. G. Gilheany, Chem. Commun., 2012, 48, 817.
7 K. V. Rajendran and D. G. Gilheany, Chem. Commun., 2012,
48, 10040.
8 N. P. Kenny, K. V. Rajendran, E. V. Jennings and D. G. Gilheany,
Chem. – Eur. J., 2013, 19, 14210.
9 K. V. Rajendran, K. Nikitin and D. G. Gilheany, J. Am. Chem. Soc.,
2015, 137, 9375.
10 S. S. Al Sulaimi, K. V. Rajendran and D. G. Gilheany, Eur. J. Org.
Chem., 2015, 5959.
11 (a) Y. Li, S. Das, S. Zhou, K. Junge and M. Beller, J. Am. Chem. Soc.,
2012, 134, 9727; (b) Y. Li, L.-Q. Lu, S. Das, S. Pisiewicz, K. Junge and
M. Beller, J. Am. Chem. Soc., 2012, 134, 18325.
12 M. Masaki and K. Fukui, Chem. Lett., 1977, 151; T. Yano,
M. Kuroboshi and H. Tanaka, Tetrahedron Lett., 2010, 51, 698; R. M.
Denton, J. An, B. Adeniran, A. J. Blake, W. Lewis and A. M. Poulton,
Finally extension to phosphonate 8 was unsuccessful, being
hindered by low conversion in the alkylation step. The same
This journal is ©The Royal Society of Chemistry 2015
Chem. Commun.

View Article Online
Communication
ChemComm
J. Org. Chem., 2011, 76, 6749; K. Nikitin, H. Muller-Bunz and 17 In the case of the O-alkyl compounds, the products immediately
D. G. Gilheany, Chem. Commun., 2013, 49, 1434.
13 The boranes adducts are useful, having enhanced air and moisture
react further by an Arbusov-type process to give the phosphinic
chloride, while in the O-aryl case, no reaction occurs.
stability, compared to their free P(III) equivalents, and they can be 18 A. Rhomberg and P. Tavs, Monatsh. Chem., 1967, 98, 105.
deprotected in a separate reaction: (a) A. Staubitz, A. P. M. 19 H. Meerwein, G. Hinz, P. Hofmann, E. Kroning and E. Pfeil, J. Prakt.
Robertson, M. E. Sloan and I. Manners, Chem. Rev., 2010, 110,
Chem., 1937, 147, 257; H. Meerwein, Org. Synth., 1973, 5, 1080.
4023; (b) M. V. Overschelde, E. Vervecken, S. G. Modha, S. Cogen, 20 The reactions were performed with 2 equiv. Meerwein’s salt in
E. V. D. Eycken and J. V. D. Eycken, Tetrahedron, 2009, 65, 6410;
(c) M. Stankevic and K. M. Pietrusiewicz, Synthesis, 2005, 1279;
(d) J. Uziel, C. Darcel, D. Moulin, C. Baudin and S. Juge, Tetrahedron:
chloroform (2 mL) at 160 1C on small scale (200 mg phosphinate)
in thick-walled 100 mL Young’s tubes that could be screwed tightly
shut at the top. For diagram and detailed procedure, see ESI†.
Asymmetry, 2001, 12, 1441; (e) H. Yang, N. Lugan and R. Mathieu, 21 T. Imamoto, S. Kikuchi, T. Miura and Y. Wada, Org. Lett., 2000, 3, 87;
Organometallics, 1997, 16, 2089; ( f ) L. McKinstry and T. Livinghouse,
Tetrahedron Lett., 1994, 35, 9319.
14 The new reaction lifted a major limitation in organophosphorus
J. Omelanczuk and M. Mikolajczyk, J. Am. Chem. Soc., 1979, 101, 7292.
22 This is because, in this solvent, at 100 1C, ethyl triflate remains in
the (now liquid) reaction mixture: for detailed explanation, see ESI†.
synthesis methodology: until then it had not been possible to 23 K. E. Elson, I. D. Jenkins and W. A. Loughlin, Org. Biomol. Chem.,
construct a P(III) aminophosphine by a route that involved attachment 2003, 1, 2958.
of the nitrogen substituent to a P(V) centre. Since many important 24 Tertiary phosphine sulfides and selenides have been reduced stereo-
organophosphorus methods necessarily generate P(V) centres, this was
significant.
15 C. L. Chernick and H. A. Skinner, J. Chem. Soc., 1956, 1401; W.-R.
specifically with hydride: W. J. Stec, A. Okruszek and J. Michalski,
Angew. Chem., Int. Ed. Engl., 1971, 10, 494; R. Luckenbach, Tetrahedron
Lett., 1971, 24, 2177.
Zheng, J.-L. Xu, T. Huang, Z.-C. Chen and Q. Yang, Comput. Theor. 25 (a) W. J. Stec, B. Uznanski, B. Bruzik and M. J. Karol, J. Org. Chem., 1976,
Chem., 2011, 968, 1. 41, 1291; (b) K. Bruzik and W. J. Stec, J. Org. Chem., 1981, 46, 1618.
16 T. V. Rajanbabu, in Phosphinite and Phosphonite Ligands, ed. 26 N. Burford, A. I. Dipchand, B. W. Royan and P. S. White, Inorg.
P. C. J. Kamer and P. W. N. M. van Leeuwen, John Wiley & Sons,
2012.
Chem., 1990, 29, 4938; N. Burford, P. Losier, A. D. Phillips, P. J.
Ragogna and T. S. Cameron, Inorg. Chem., 2003, 42, 1087.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2015