Reactivity of phosphiranes toward nucleophiles: Theoretical and experimental investigations

Article abstract of DOI:10.1002/hc.20405

The thermodynamic and kinetic parameters for the reaction of 1-methylphosphirane with dimethylphosphide (PMe₂^-), proceeding via attack at the ring phosphorus atom to give Me₂P- PMe^- plus ethylene, were calculated at the G3(MP2)-RAD(+) level. The rate constant for this mechanism (k= 1.1 × 10^-13 L mol ^-1 s^-1 at 25°C) is seven orders of magnitude greater than that for the previously studied mechanism involving attack at carbon. Experimental investigations with 1-phenylphosphirane gave consistent results: Treatment with LiPMePh yielded no detectable polymer, and quenching with MeI gave the known diphosphine (R*, R*)-(±)/(R*,S*)- MePhP-PMePh.

Full text of DOI:10.1002/hc.20405

Heteroatom Chemistry
Volume 19, Number 2, 2008
Reactivity of Phosphiranes Toward
Nucleophiles: Theoretical and Experimental
Investigations
Michelle L. Coote, Elizabeth H. Krenske, Ilham Maulana,
Jo¨rg Steinbach, and S. Bruce Wild
Research School of Chemistry, The Australian National University, Canberra ACT 0200, Australia
Received 3 April 2007; revised 8 July 2007
ture [P(R)CH2CH2]n [1]. We have previously reported
ABSTRACT: The thermodynamic and kinetic param-
computational studies of the radical polymerization
eters for the reaction of 1-methylphosphirane with
of phosphiranes and found that the most likely path-
dimethylphosphide (PMe−2 ), proceeding via attack at
way involved attack by a C-centered radical at a ring
the ring phosphorus atom to give Me2P–PMe− plus
phosphorus atom [2]. In subsequent work, we stud-
ethylene, were calculated at the G3(MP2)-RAD(+)
ied the anionic ring opening of 1 (R = Me), giving
level. The rate constant for this mechanism (k = 1.1
consideration to a mechanism involving attack by
× 1013 L mol−1 s−1 at 25◦C) is seven orders of mag-
PMe−2 at carbon (Scheme 1, step 1), and found that
nitude greater than that for the previously studied
it should be fast enough (k = 6.2 × 105 L mol−1 s−1
mechanism involving attack at carbon. Experimen-
at 25◦C) to enable a practical synthesis of polyphos-
tal investigations with 1-phenylphosphirane gave con-
phines 3 [3]. The borane adduct of 1 (R = Me) had a
sistent results: Treatment with LiPMePh yielded no
similar rate constant (k = 6.8 × 104 L mol−1 s−1).
detectable polymer, and quenching with MeI gave
the known diphosphine (R∗,R∗)-( )/(R∗,S∗)-MePhP–
ꢀ
C
PMePh. 2008 Wiley Periodicals, Inc. Heteroatom Chem
19:178–181, 2008; Published online in Wiley InterScience
(www.interscience.wiley.com). DOI 10.1002/hc.20405
We wished to use these nucleophilic ring-
INTRODUCTION
opening reactions to prepare polyphosphines of the
type 3. However, previous experimental studies have
shown that the phosphirane ring can react with a nu-
cleophile at phosphorus because of the polarization
Small-ring phosphorus heterocycles, such as phos-
phiranes (1), have recently attracted interest as pre-
cursors to polyphosphines. For example, the ring-
opening polymerization of 1 (R = 2,4,6−tBu3-C6H2)
under (cationic) initiation by MeI or MeOTf has been
used to prepare a polymer with the backbone struc-
δ+
δ−
P –C of the P C ring bond [4]. Depending on the
substituents, this can be accompanied by cleavage
of one or both of the P C ring bonds. For example,
it was found that treatment of the coordinated phos-
phirane in 4 with a hard nucleophile such as RO− or
R2N− led to cleavage of one P C bond, whereas soft
nucleophiles such as R2P− and RS− led to cleavage
of both [5]. In the latter case, styrene was liberated
(Scheme 2).
Correspondence to: Michelle L. Coote and S. Bruce Wild; e-mail:
mcoote@rsc.anu.edu.au and sbw@rsc.anu.edu.au.
Contract grant sponsor: Australian Research Council Discovery
Grant.
c
ꢀ
2008 Wiley Periodicals, Inc.
178

Reactivity of Phosphiranes Toward Nucleophiles: Theoretical and Experimental Investigations 179
SCHEME 1
SCHEME 2
We, therefore, set out to determine whether a
calculated thermodynamic and kinetic parameters
[7] for the two modes of attack are given in Table 1.
The calculated equilibrium constants indicate
that attack at carbon and attack at phosphorus are
both thermodynamically very favorable. This pri-
marily reflects the release of ring strain. However,
the rate constant for attack at phosphorus (k = 1.1
× 1013 L mol−1 s−1) is approximately a factor of
107 greater than for attack at carbon (k= 6.2 × 105
L mol−1 s−1). This is presumably due to the lesser
Coulombic repulsion in the transition state when at-
tack occurs at phosphorus.
similar mode of reaction would be found for free
phosphiranes. We have used ab initio molecular or-
bital calculations at the G3(MP2)-RAD(+) level [6]
to investigate the dimethylphosphide-induced dis-
sociation of 1-methylphosphirane into Me2P–PMe−
(5) and ethylene, and then experimentally examined
the reaction of 1-phenylphosphirane (1, R = Ph) with
lithium methylphenylphosphide.
RESULTS AND DISCUSSION
Molecular geometries, obtained at the B3-LYP/6–
31+G(d) level, for the attack of the PMe2−
ion at the ring-carbon and phosphorus in
1-methylphosphirane, are shown in Scheme 3. The
During attack at phosphorus, the PMe−2 ion ap-
proaches roughly in the plane of the ring, leading to
an unsymmetrical transition state in which the two
˚
breaking P C bonds differ in length by 0.331 A and
SCHEME 3
Heteroatom Chemistry DOI 10.1002/hc

180 Coote et al
TABLE 1 Calculated Thermodynamics and Kinetics (25◦C)
of the Ring Opening of 1 (R = Me) by PMe2−
the formation of the known diphosphine (R∗,R∗)-
( )/(R∗,S∗)-7 was evident from the presence of sin-
glets at δ −39.8 and −36.2 for the two diastereomers
Attack at Ca
Attack at P
1
in the 31P{ H} NMR spectrum in C6D6 at 25◦C [9–11].
ꢀH (kJ mol−1
)
−123.5
−145.8
−80.1
−91.0
19.1
The propensity for nucleophiles to attack the
phosphirane ring at phosphorus rather than at car-
bon is predominantly a kinetic effect reflecting the
polarity of the P C bond. Although this reduces the
utility of anionic ring opening as a general route
to poly(phosphiranes), it is possible that successful
polymerization may be achieved if the polarity of the
P C ring bonds were reversed through appropriate
choice of substituents.
ꢀS (J K−1 mol−1
)
ꢀG (kJ mol−1
)
−96.7
K
2.6 × 1015
8.6
8.7 × 1016
−39.3
ꢀH‡ (kJ mol−1
)
ꢀS‡ (J K−1 mol−1
)
−131.7
−153.7
6.6
ꢀG‡ (kJ mol−1
)
47.9
k (L mol−1 s−1
aFrom [3].
)
6.2 × 105
1.1 × 1013
the more detached carbon atom bears an increased
negative charge (natural charge −0.852 in the TS,
compared with −0.645 in the reactant). It was, there-
fore, suspected that the carbanionic species 6 might
be an intermediate en route to 5. However, species
6 was found to lie 14.6 kJ mol−1 above the transition
structure (relative enthalpies at 25◦C at the G3(MP2)-
RAD(+) level), which suggests that the cleavage of
the two P C bonds is a concerted process.
EXPERIMENTAL
All reactions were carried out under a nitrogen or
argon atmosphere using Schlenk techniques. Sol-
vents were purified by conventional methods and
stored under nitrogen or argon [12]. 1H and 31P{ H}
1
NMR spectra were recorded on Varian VXR 300S
and Gemini 300 spectrometers operating at 300.1
MHz (1H) and 121.5 MHz (31P) with chemical shifts
being expressed in ppm downfield from internal
Me4Si (1H) and external 85% aqueous H3PO4 (31P).
Methylphenylphosphine was distilled and kept un-
der argon [13]. n-Butyllithium and iodomethane
were used as supplied by Sigma-Aldrich.
1-Phenylphosphirane was synthesized according
to the published method [8], and kept briefly in
THF solution (1 d) prior to use. For the dilithia-
tion of phenylphosphine, however, a 2:1 mixture of
n-hexane and THF was used as solvent instead of
pure THF, so as to decrease the solubility of the prod-
uct (THF is needed to stabilize the product in the
form of a complex, otherwise the powder obtained
is pyrophoric). The crude 1-phenylphosphirane was
distilled directly from the reaction mixture without
initial separation of the lithium chloride. This im-
proved the yield from 21% to 29%.
The large difference in rate constants for the
two pathways suggests that, although the propaga-
tion step of the desired anionic polymerization of
free phosphiranes is facile [3], the polymerization is
likely to be hampered by attack at phosphorus. We
have found this to be borne out by experiment. When
the phosphirane 1 (R = Ph) [8] was treated with
lithium methylphenylphosphide as initiator in THF,
1
no evidence was found in the 31P{ H} NMR spectrum
for the formation of poly(1-phenylphosphirane), al-
though there were in the spectrum numerous peaks
in the region −43 to −5 ppm. However, when
a solution of 1 (R = Ph) was treated with a sto-
ichiometric amount of lithium methylphenylphos-
phide in THF, and the intermediate phosphinophos-
phide then treated with iodomethane (Scheme 4),
(R∗,R∗)-( )/(R∗,S∗)-1,2-Dimethyl-1,2-
diphenyldiphosphine (R∗,R∗)-( )-/R∗,S∗)-7
In a tube equipped with a septum, a solution of
methylphenylphosphine (0.25 mL, 2.9 mmol) in
SCHEME 4
Heteroatom Chemistry DOI 10.1002/hc

Reactivity of Phosphiranes Toward Nucleophiles: Theoretical and Experimental Investigations 181
THF (6 mL) was cooled to −78◦C (acetone/dry ice).
REFERENCES
A solution of n butyllithium (1.6 M in n-hexane, 1.8
mL, 2.9 mmol) was added via syringe, and the result-
ing orange solution was stirred for a further 10 min
before 1-phenylphosphirane (0.35 mL, 2.9 mmol)
was added. The solution was then allowed to warm
to room temperature, during which time the color
changed to dark yellow. The mixture was quenched
with iodomethane (0.18 mL, 2.9 mmol), whereupon
the color disappeared. After removal of the solvent
and other volatile components in vacuo, a colorless
oil was obtained, which was identified as the diphos-
phine (R∗,R∗)-( )/(R∗,S∗)-MePhPPMePh (7) by com-
[1] Kobayashi, S.; Kadokawa, J.-I. Macromol Rapid
Commun 1994, 15, 567–571.
[2] Hodgson, J. L.; Coote, M. L. Macromolecules 2005,
38, 8902–8910.
[3] Coote, M. L.; Hodgson, J. L.; Krenske, E. H.; Wild, S.
B. Heteroat Chem 2007, 18, 118–128.
[4] See, for example, Mathey, F. Chem Rev 1990, 90, 997–
1025.
[5] Marinetti, A.; Mathey, F. Tetrahedron 1989, 45, 3061–
3070.
[6] The computational methods used were the same
as those used in [3], wherein a full description is
supplied. The G3(MP2)-RAD(+) method is a high-
level procedure that attempts to reproduce coupled-
cluster [CCSD(T)] energies with a large triple zeta
basis set using additivity corrections carried out at
the RMP2 level.
1
parison of its 31P{ H} NMR data, viz. δP −39.8 (s),
−36.2 (s) (C6D6), with the literature values, viz. δP
40.5 (s), 36.9 (s) (C6H6) [11].
[7] Thermodynamic and kinetic parameters were calcu-
lated by the methods described in [3]. The variational
transition state for attack at phosphorus was identi-
fied by a relaxed scan along the forming (P P) bond
length at the B3-LYP/6-31+G(d) level.
SUPPLEMENTARY INFORMATION
[8] Hockless, D. C. R.; Kang, Y. B.; McDonald, M. A.;
Pabel, M.; Willis, A. C.; Wild, S. B. Organometallics
1996, 15, 1301–1306.
B3-LYP/6–31+G(d) geometries, in the form of
Gaussian archive entries, are available from the
author on request (mcoote@rsc.anu.edu.au).
[9] Fluck, E.; Issleib, K. Chem Ber 1965, 98, 2674–2680.
[10] Lambert, J. B.; Jackson, G. F., III; Mueller, D. C. J Am
Chem Soc 1968, 90, 6401–6405.
[11] McFarlane, H. C. E.; McFarlane, W. J Chem Soc,
Chem Commun 1971, 1589–1590.
ACKNOWLEDGMENTS
[12] Armarego, W. L. F.; Chai, C. L. L. Purification of Lab-
oratory Chemicals, 5th ed.; Butterworth-Heinemann:
Amsterdam, 2003; Ch. 4.
[13] Roberts, N. K.; Wild, S. B. J Am Chem Soc 1979, 101,
6254–6260.
We gratefully acknowledge generous allocations of
computing time from the Australian Partnership for
Advanced Computing and the Australian National
University Supercomputer Facility.
Heteroatom Chemistry DOI 10.1002/hc