Identification of a key intermediate in the asymmetric Appel process: One pot stereoselective synthesis of P-stereogenic phosphines and phosphine boranes from racemic phosphine oxides

Article abstract of DOI:10.1039/c2cc34136k

Sequential treatment of racemic phosphine oxides with oxalyl chloride and chiral non-racemic alcohol generates the same ratios of diastereomeric alkoxyphosphonium salts obtained in the corresponding asymmetric Appel process, strongly implicating the intermediate chlorophosphonium salt in the stereoselecting step. Subsequent reduction allows a novel synthesis of enantioenriched P-stereogenic phosphines-phosphine boranes. The Royal Society of Chemistry 2012.

Full text of DOI:10.1039/c2cc34136k

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2012, 48, 10040–10042
www.rsc.org/chemcomm
COMMUNICATION
Identification of a key intermediate in the asymmetric Appel process: one
pot stereoselective synthesis of P-stereogenic phosphines and phosphine
boranes from racemic phosphine oxidesw
Kamalraj V. Rajendran and Declan G. Gilheany*
Received 8th June 2012, Accepted 15th August 2012
DOI: 10.1039/c2cc34136k
Sequential treatment of racemic phosphine oxides with oxalyl
chloride and chiral non-racemic alcohol generates the same
ratios of diastereomeric alkoxyphosphonium salts obtained in
the corresponding asymmetric Appel process, strongly implicating
the intermediate chlorophosphonium salt in the stereoselecting
step. Subsequent reduction allows a novel synthesis of enantio-
enriched P-stereogenic phosphines–phosphine boranes.
to reach the target phosphines, which can lead to loss of
enantiomeric excess;9 the starting phosphines may be difficult
to prepare requiring storage/manipulation under inert atmos-
phere; the use of HCA is not ideal and its by-product penta-
chloroacetone (PCA) can make purification tedious; and,
finally, the chiral alcohol adjuvant is lost as the chloride.
Furthermore we were initially unsure of the course of the
reaction and the mechanism of stereoselection. We now report
the identification, via independent generation, of the key
intermediate in the process, which, in turn, led to a new,
simpler P-stereogenic methodology.
The use of enantiomerically pure phosphine ligands in asymmetric
catalysis is a popular strategy for asymmetric synthesis1 and much
effort has been directed towards the design, synthesis and testing
of new enantiomerically pure phosphines.2 Several methodologies
have been developed for the synthesis of enantiomerically pure
P-stereogenic phosphines3 and a large number of such ligands
have been reported in the literature.3–5 Some of these methods
can be very effective, but each of them has its own demerits and,
to date, there is no straightforward general way to synthesise
P-stereogenic phosphines.
Our working hypothesis for the course of the asymmetric
Appel process6,7,10 is also shown in Scheme 1. It involves the
transient generation of an intermediate chlorophosphonium
salt (CPS)11 that reacts rapidly with added chiral non-racemic
alcohol (R*OH), giving unequal amounts of diastereomeric
alkoxyphosphonium salts (DAPS route A), which then undergo
slow Arbuzov collapse to form scalemic phosphine oxide.
A complication revealed in our previous work6b is that the
diastereomeric excess (de)12 in the DAPS is the better measure of
the stereoselective step of the reaction because the enantiomeric
excess (ee) observed in the product oxides may be lower due to
non-selective processes than can occur during the Arbusov step.
Therefore to better study the selectivity, we devised a consistent
procedure (see ESIw) to measure the de of DAPS by 31P-NMR
spectroscopy13 and Table 1 (col A) shows how this more reliable
selectivity varies for a range of phosphine–alcohol combinations.
A more significant difficulty in our studies was that it proved
impossible to definitively detect the presence of CPS during the
reaction. This is due to its rapid reaction with the added
alcohol, which must be present from the start to prevent
certain side reactions within the manifold of the Appel condi-
tions.7 Therefore we were most interested in how CPS might
be independently generated as a probe of the reaction. We
were intrigued to discover that it had been known for a long
time that they can be easily obtained from the phosphine oxide
by reaction with oxalyl chloride.14 Recently, this reaction has
been used by us15 and others16,17 for other useful transforma-
tions. Accordingly, we took a series of racemic phosphine
oxides and treated them with oxalyl chloride, which cleanly
generated the corresponding CPSs (as judged by 31P NMR –
see ESIw). When treated with chiral alcohol at À78 1C, these
did indeed yield the same diastereomeric alkoxyphosphonium
We previously developed a successful method for the dynamic
kinetic resolution of racemic arylmethylphenyl phosphines.6 This
was achieved in their oxidation using an asymmetric version
of the Appel reaction conditions7 by treatment (at À78 1C)
with hexachloroacetone (HCA) in the presence of a chiral
non-racemic alcohol (Scheme 1). Although this reaction is an
effective way to make P-stereogenic phosphine oxides,8 it too
has demerits. Subsequent stereospecific reduction is required
Scheme 1 Hypothesis for the course of the asymmetric Appel process
showing the proposed chlorophosphonium salt (CPS) intermediate and
its reaction with alcohol to give diastereomeric alkoxyphosphonium
salts (DAPS route A).
Centre for Synthesis and Chemical Biology, School of Chemistry and
Chemical Biology, University College Dublin, Belfield, Dublin 4,
Ireland. E-mail: declan.gilheany@ucd.ie; Fax: +353-1-7162127;
Tel: +353-1-7162308
w Electronic supplementary information (ESI) available: Full experi-
mental procedure, and full characterization data for phosphine oxides
and boranes; HPLC traces of all scalemic phosphine boranes. See
DOI: 10.1039/c2cc34136k
c
10040 Chem. Commun., 2012, 48, 10040–10042
This journal is The Royal Society of Chemistry 2012

View Article Online
Table 1 Stereoselectivities measured in this work. A/B: Comparison
of diasteromeric excessesa of DAPS prepared from phosphines Ar-
MePhP (route A: Scheme 1)b and oxides ArMePhPQO (route B:
Scheme 2).c C/D: Enantiomeric excessesd of phosphine boranes Ar-
MePhP-BH3 (C) and phosphines (ArMePhP) (D) prepared by treating
DAPS from’ route B with NaBH4 and LiAlH4 respectivelye (Scheme 3)
Scheme 2 DAPS route B by independent generation of CPS from
o-sub.
in Ar group R*OHf (A)
% de % de % ee (C) % eeg (D)
phosphine oxide and subsequent reaction with chiral alcohol.
#
(B)
(config)
(config)
Concurrent with these studies, separate work from
our laboratory18 had shown that DAPS (obtained from the
asymmetric Appel reaction) could be stereospecifically reduced
with LiAlH4 to give phosphine or with NaBH4 to give phosphine
borane directly. In these reactions the de in DAPS corresponded
to the ee of the reduced products, with only small losses of chiral
information. Combining the three ideas of chlorination, dynamic
resolution and reduction then suggested a one-pot stereoselective
synthesis of phosphines and phosphine boranes from racemic
phosphine oxides (Scheme 3).
1
2
3
4
5
6
7
8
9
Me
Me
Me
Me
ent-1
1
2
82
81
62
63
46
50
48
70
60
68
70
68
—
—
81
80
67
65
76
84
83
64
65
46
49
46
74
64
71
71
73
76(R)
À74(S)
À65(S)
À63(S)
46(R)
40(R)
À44(S)
64(R)
À37(S)
À63(S)
71
78(R)
À76(S)
—
—
—
48(R)
À48(S)
67(R)
À30(S)
À54(S)
68
3
Me
5
ent-1
1
4
2
3
ent-1
1
2
3
ent-1
1
ent-1
1
ent-1
OMe
OMe
OMe
OMe
10 OMe
11 CF3h
h
12 CF3h
13 CF3h
14 CF3
15 iPr
À76
À66
i
i
i
—
—
À84j
—
—
In the early experiments, the produced boranes showed
significant loss of stereoselectivity: for example, the borane
derived from methylphenyl(o-tolyl)phosphine oxide (entry 1,
Table 1) initially revealed an ee of only 40%. By monitoring
the reaction closely by 31P NMR, we discovered that this
resulted from reaction of excess oxalyl chloride (used to ensure
complete conversion to CPS) with non-racemic phosphine
oxide formed from some Arbuzov collapse of DAPS. The reformed
CPS reacts with NaBH4 to give racemic phosphine borane
(Scheme 4(i)).15 The reaction protocol was therefore altered;
limiting the amount of oxalyl chloride strictly to one equiva-
lent and employing excess alcohol. An advantage of the
process is that all the alcohol used can be recovered, as it is
regenerated in the reduction.18 A variety of racemic phosphine
oxides was screened, focussing, for proof-of-principle, on
inexpensive menthol as chiral auxiliary with results shown in
Table 1 (columns C/D). The selectivity and the configuration
of the scalemic phosphines and boranes obtained followed
the same general trends observed in the asymmetric Appel
reaction.6 The best result obtained was 84% ee, the highest
reported to date for the dynamic resolution of a phosphine.
It is noticeable that some substrate/alcohol combinations are
i
À76j
82
82
70
68
78
41(R)
À64(S)
51
68(R)
À66(S)
6868i
—
16 iPr
17 Ph
18 Ph
19 Me, p-F
À52
58
—
a
b
Determined by 31P-NMR (see ESI). Reaction conditions: phosphine
(1.1 mmol), alcohol (1.32 mmol.), HCA (1.1 mmol.) at À78 1C, all yields
>95% (as judged by 31P NMR). Reaction conditions: Phosphine
oxide (1.1 mmol.), oxalyl chloride (1.1 mmol) followed by alcohol
c
(1.32 mmol) at À78 1C, all yields are >95% (as judged by 31P NMR).
d
Determined by CSP HPLC (see ESI), negative ee denotes that the
major enantiomer was eluted second; yields >90% except where noted
(as judged by 31P NMR; isolated yields for all (À)-menthol cases),
e
configurations, where given, determined as described in ESI. LiAlH4
(1.1 mmol in PhMe) or NaBH4 (5.5 mmol in diglyme) added to DAPS
f
g
from route B at À78 1C. See Chart 1. Measured by CSP HPLC
(see ESI), after subsequent conversion to the borane with BH3ÁTHF.
h
For route B: oxalyl chloride reaction warmed to 50 1C to ensure full
conversion to chlorophosphonium salt. Unable to measure de due to
i
j
faster Arbuzov collapse of the DAPS. Yielded 60–65% of phosphine
borane and 35–40% of phosphine oxide, also enantioenriched.
Scheme 3 Synthesis of P-stereogenic phosphines and phosphine
boranes.
Chart 1 Chiral alcohols used in Scheme 1–3 and Table 1.
salts as in the asymmetric Appel reaction (DAPS route B,
Scheme 2). A side-by-side comparison of the diastereomeric
excesses (de) in the salts produced by the two different routes
for the same set of phosphine–alcohol combinations is shown
in Table 1 (columns A/B). It can be seen that the selectivities of
both routes are very similar, providing very strong support for
our mechanistic hypothesis and strongly suggesting that the
selectivity of both processes is set in the conversion of CPS to
DAPS. The de is very slightly higher for the oxalyl chloride
route in nearly all cases and we ascribe this to the difference in
counterion (ClÀ vs. PCAÀ).
Scheme 4 Stereoselective formation of phosphine boranes from
racemic oxides showing pathway for ee erosion (i) and competing
alcohol reduction (ii).
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 10040–10042 10041

View Article Online
more prone to erosion of selectivity in the reduction
(e.g. Table 1, entries 1, 6, 11, 17 vs. entry 15). We believe that
this is due to diastereoselection in the small amount of alternative
reduction19 to give oxide and menthane (Scheme 4(ii)).
Soc., 2006, 128, 2788; V. S. Chan, R. G. Bergman and F. D. Toste,
J. Am. Chem. Soc., 2007, 129, 15122; G. Cedric, S. J. Canipa,
P. O’Brien and S. Taylor, J. Am. Chem. Soc., 2006, 128, 9336.
4 For reviews of P-stereogenic compounds, see:(a) A. Grabulosa,
J. Granell and G. Muller, Coord. Chem. Rev., 2007, 251, 25;
(b) D. S. Glueck, Synlett, 2007, 2627; (c) M. J. Johansson and
N. C. Kann, Mini-Rev. Org. Chem., 2004, 1, 233; (e) K. M.
Pietrusiewicz and M. Zablocka, Chem. Rev., 1994, 94, 1375.
5 P-Stereogenic Ligands in Enantioselective Catalysis, ed.
A. Grabulosa, Royal Society of Chemistry, Cambridge, UK, 2011.
6 (a) E. Bergin, C. T. O’Connor, S. B. Robinson, E. M. McGarrigle,
C. P. O’Mahony and D. G. Gilheany, J. Am. Chem. Soc., 2007,
129, 9566; (b) K. V. Rajendran, L. Kennedy and D. G. Gilheany,
Eur. J. Org. Chem., 2010, 5642.
7 R. Appel and M. Halstenberg, Tertiary Phosphane Halogenoalkane
Reagents, in Organophosphorus Regents in Organic Synthesis, ed. J. I. G.
Cadogan, Academic Press, London, 1979, ch. 9, p. 387; I. M. Downie,
J. B. Holmes and J. B. Lee, Chem. Ind. (London, U. K.), 1966, 900.
8 D. G. Gilheany, S. B. Robinson, C. P. O’Mahony, C. T.
O’Connor, E. Bergin, D. M. Walsh, E. F. Clarke, B. G. Kelly
and E. M. McGarrigle, Int. Patent, WO 118603, 2005. This method
has been successfully run on a multikilo scale in an industrial
process.
9 K. Naumann, G. Zon and K. Mislow, J. Am. Chem. Soc., 1969,
91, 7012; K. Naumann, G. Zon and K. Mislow, J. Am. Chem. Soc.,
1969, 91, 2788; G. Zon, K. E. DeBruin, K. Naumann and K. Mislow,
J. Am. Chem. Soc., 1969, 91, 7023; D. Valentine, Jr., J. F. Blount and
K. Toth, J. Org. Chem., 1980, 45, 3691; L. D. Quin, K. C. Caster,
J. C. Kislaus and K. A. Masch, J. Am. Chem. Soc., 1984, 106, 7021.
10 A fuller description of our working hypothesis is given in the ESIw.
11 For studies of such salts see: S. M. Godfrey, C. A. McAuliffe,
R. G. Pritchard and J. M. Sheffield, Chem. Commun., 1998, 921;
N. C. Gonnella, C. Busacca, S. Campbell, M. Eriksson, N. Grinberg,
T. Bartholomeyzik, S. Ma and D. L. Norwood, Magn. Reson. Chem.,
2009, 47, 461.
In summary, we have adduced convincing evidence for our
proposed course for the asymmetric Appel process. This will
enable us to work to improve its selectivity. Also during the
study, we discovered an unprecedented alternative method for
the creation of P-stereogenicity. The one-pot method starts
from the more convenient oxides, has more easily removed
by-products (CO, CO2, HCl) and yields the protected phosphine
directly. To be sure, much development work is needed, both to
raise the selectivity and minimise its erosion. However, in that
regard, we now have greater scope in our choice of chiral alcohol
auxiliary because it can be recovered at the end of the reaction.
We thank sincerely Science Foundation Ireland (SFI) for
funding this chemistry under Grants RFP/08/CHE1251 and
09/IN.1/B2627. We are also grateful to UCD Centre for
Synthesis and Chemical Biology (CSCB) and the UCD School
of Chemistry and Chemical Biology for access to their extensive
analysis facilities. DGG thanks sincerely University College
Dublin for a President’s Research Fellowship during which
the conception of this work took place. The Fellowship was
held partly in Stanford University in the laboratory of Professor
James Collman, to whom DGG is warmly appreciative for both
his hospitality and stimulating intellectual discussions.
Notes and references
12 Diastereomeric excess (de) is used rather than diastereomeric ratio
(dr) to facilitate comparison with the enantiomeric excess (ee) of
the ultimate products of the reactions (oxides or boranes).
13 E.g.: the reaction mixture derived from oxalyl chloride treatment
of methylphenyl(o-tolyl)phosphine oxide directly after addition of
(À)-menthol shows two narrow signals for DAPS of unequal
heights (92 : 8) at d 67.8 and d 67.3 ppm, replacing the CPS signal
at d 71.0. An acquisition period of 3 s was set for all 31P spectra to
allow full relaxation, extensive precautions taken to ensure that the
measured de truly reflected that produced in the reactions: see ESIw
for details.
1 Catalytic Asymmetric Synthesis, ed. I. Ojima, Wiley-VCH,
New York, 3rd edn, 2010; Tetrahedron: Asymmetry, ed.
X. Zhang, 2004, vol. 15, pp. 2099–2311, special issue; K. V. L.
Crepy and T. Imamoto, Top. Curr. Chem., 2003, 229, 1;
S. J. Connon, Angew. Chem., Int. Ed., 2006, 45, 3909; J. L.
Methot and W. R. Roush, Adv. Synth. Catal., 2004, 346, 1035.
2 Phosphorus Ligands in Asymmetric Catalysis, ed. A. Borner,
¨
¨
Wiley-VCH, Weinheim, 2008, vol. I–III; S. Luhr, J. Holz and
A. Borner, ChemCatChem, 2011, 3, 1708.
¨
3 Leading references to the major methods: Methyl phosphinate route:
O. Korpium and K. Mislow, J. Am. Chem. Soc., 1967, 89, 4784;
B. D. Gatineau, L. Laurent Giordano and G. Buono, J. Am. Chem.
Soc., 2011, 133, 10728; Q. Xu, C.-Q. Zhao and L.-B. Han, J. Am.
14 M. Masaki and K. Fukui, Chem. Lett., 1977, 151; K. Fukui and
N. Kakeya, U. S. Patent, 4,301,301 Nov. 17, 1981.
15 Direct (achiral) conversion of phosphine oxides to boranes:
K. V. Rajendran and D. G. Gilheany, Chem. Commun., 2011,
48, 817.
Chem. Soc., 2008, 130, 12648. Cyclic phosphoramidate route: S. Juge
and J. P. Genet, Tetrahedron Lett., 1989, 30, 2783; C. Darcel, J. Uziel
and S. Juge, in ref. 2, vol. 3, pp. 1211–1233; T. Leon, A. Riera and
´
´
16 Electroreduction of phosphine oxides: M. Kuroboshi, T. Yano,
S. Kamenoue, H. Kawakubo and H. Tanaka, Tetrahedron, 2011,
67, 5825.
17 Catalytic Appel reaction: R. M. Denton, J. An and B. Adeniran,
Chem. Commun., 2010, 46, 3025; R. M. Denton, J. An, B. Adeniran,
A. J. Blake, W. Lewis and A. M. Poulton, J. Org. Chem., 2011,
76, 6749.
18 K. V. Rajendran, J. S. Kudavalli, K. S. Dunne and D. G. Gilheany,
Eur. J. Org. Chem., 2012, 2720–2723.
19 K. E. Elson, I. D. Jenkins and W. A. Loughlin, Org. Biomol. Chem.,
2003, 1, 2958; J. B. Hendrickson, M. Singer and Md. S. Hussoin,
J. Org. Chem., 1993, 58, 6913. Also in LiAlH4 reductions any
produced oxide will be reduced racemically: see ref. 18.
X. Verdaguer, J. Am. Chem. Soc., 2011, 133, 5740; Desymmetrisation:
A. R. Muci, K. R. Campos and D. A. Evans, J. Am. Chem. Soc.,
1995, 117, 9075; A. Ohashi, S. I. Kikuchi, M. Yasutake and
T. Imamoto, Eur. J. Org. Chem., 2002, 2535; J. J. Gammon,
V. H. Gessner, G. R. Barker, J. Granander, A. C. Whitwood,
C. Strohmann, P. O’Brien and B. Kelly, J. Am. Chem. Soc., 2010,
132, 13922; J. Granander, F. Secci, S. J. Canipa, P. O’Brien and
B. Kelly, J. Org. Chem., 2011, 76, 4794; Enzymatic resolution:
P. Kielbasinski, J. Omelanczuk and M. Mikolajczyk, Tetrahedron:
Asymmetry, 1998, 9, 3283; Dynamic resolution: H. Heath, B. Wolfe,
T. Livinghouse and S. K. Bae, Synthesis, 2001, 2341; C. E. Headley
and S. P. Marsden, J. Org. Chem., 2007, 72, 7185; Catalytic
asymmetric synthesis: C. Scriban and D. S. Glueck, J. Am. Chem.
c
10042 Chem. Commun., 2012, 48, 10040–10042
This journal is The Royal Society of Chemistry 2012