Frustrated lewis pair route to hydrodesilylation of silylphosphines

Article abstract of DOI:10.1021/om4007278

A 1:1 mixture of P(SiMe₃)₃ and B(p-C ₆F₄H)₃ reacts with 3 equiv of 4-heptanone to afford a 1:2 mixture of [(Me₃SiO)(n-Pr)₂C]H ₂P-B(p-C₆F₄H)₃ and the silyl enol ether, 4-trimethylsiloxy-3-heptene. Subsequent thermolysis of the adduct produces H₃P-B(p-C₆F₄H)₃ and an additional equiv of silyl enol ether. In the presence of a catalytic quantity of B(p-C₆F₄H)₃, however, P(SiMe₃) ₃ reacts with 4-heptanone to produce a 1:1 mixture of [(Me ₃SiO)(n-Pr)₂C]₂PH and silyl enol ether. Heating this mixture further produces [(Me₃SiO)(n-Pr)₂C]PH ₂, which is eventually converted to elemental phosphorus.

Full text of DOI:10.1021/om4007278

Communication
pubs.acs.org/Organometallics
Frustrated Lewis Pair Route to Hydrodesilylation of Silylphosphines
Katsuhiko Takeuchi, Lindsay J. Hounjet, and Douglas W. Stephan*
Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario M5S 3H6, Canada
S
* Supporting Information
ABSTRACT: A 1:1 mixture of P(SiMe3)3 and B(p-C6F4H)3 reacts with
3 equiv of 4-heptanone to afford a 1:2 mixture of [(Me3SiO)(n-
Pr)2C]H2P−B(p-C6F4H)3 and the silyl enol ether, 4-trimethylsiloxy-3-
heptene. Subsequent thermolysis of the adduct produces H3P−B(p-
C6F4H)3 and an additional equiv of silyl enol ether. In the presence of a
catalytic quantity of B(p-C6F4H)3, however, P(SiMe3)3 reacts with 4-
heptanone to produce a 1:1 mixture of [(Me3SiO)(n-Pr)2C]2PH and
silyl enol ether. Heating this mixture further produces [(Me3SiO)(n-
Pr)2C]PH2, which is eventually converted to elemental phosphorus.
hosphine ligands are ubiquitous in transition metal
B(p-C6F4H)3 1 and silyl enol ether CH3CH2CH2C(OSiMe3)
Pchemistry and indeed an enormous variety of mono- and C(H)CH2CH3 29 (E/Z = 1:4) in a 1:2 ratio (Scheme 1), as
multidentate ligands have been prepared and utilized.1 While a
variety of strategies have been developed to achieve the
synthesis of phosphine ligands, principal among the synthons
used are secondary and primary phosphines. These precursors
are often challenging to handle as a result of their volatility, air-
sensitivity, and disagreeable odor. Typically, primary and
secondary phosphines are made by treatment of phosphorus-
halide or alkoxide species with hydride reagents. Alternatively,
alkylation of phosphide such as NaPH2 and reduction of
hydroxyl-methylphosphines have been exploited to prepare
either primary2 or secondary3 phosphines. Nonetheless, the
number of synthetic routes to these useful synthons is limited.
The advent of frustrated Lewis pair (FLP) chemistry4 has
seen the unquenched reactivity of sterically encumbered
phosphines and boranes applied to activations of various
small molecules. FLP chemistry has also been exploited for
hydrogenation catalysis.5 Among the various reductions that
have been demonstrated, we have also shown that P−P bonds
can be hydrogenated to provide an FLP-mediated approach to
secondary or primary phosphines.6 In more recent work, we
reported that the FLP derived from P(SiMe3)3 and B(p-
C6F4H)3 participates in the sequential activation of H2 and
CO2, with migrations of one or more Me3Si groups to oxygen.7
Herein, we report that FLPs derived from silylphosphines and
B(p-C6F4H)38 react with aliphatic ketone in stoichiometric and
catalytic reactions with the elimination of silyl enol ethers. This
hydrodesilylation reaction is shown to provide a new synthetic
route to secondary and primary phosphines, PH3 and elemental
phosphorus, respectively.
Scheme 1. Stoichiometric Synthesis of 1 and 2
made evident by 1H nuclear magnetic resonance (NMR)
spectroscopy. Following evaporation of the solvent and volatile
2 from the reaction mixture, recrystallization of the residual
solid with hexanes produced crystals of 1 in 78% yield. The 1H
NMR spectrum of 1 shows a doublet at δ 5.19 with a large 1JPH
= 385 Hz, and the 31P NMR spectrum shows a triplet with the
same coupling constant at δ −37.1, consistent with that of the
RPH2 moiety. The 11B{1H} NMR spectrum of 1 shows a broad
signal at δ −15.4, indicating a 4-coordinate B center. These
spectral data are consistent with those reported for the adduct
1
(Cy)H2P−B(C6F5)3 (δH 4.65, δB −17.5, and δP −30.0; JPH
=
393 Hz).10 The X-ray structure of 1 (Figure 1) shows the
pseudo-tetrahedral geometries at P and B, with a P−B bond
length of 2.042(2) Å and P−H bond lengths of 1.28(2) and
1.29(2) Å,11 similar to those reported for (t-Bu)H2P−B(C6F5)3
[P−B = 2.015(3) Å, P−H = 1.28(2) and 1.30(3) Å].12
The silylphosphine, P(SiMe3)3, also reacted with 3
equivalents of 4-heptanone in the presence of a catalytic
amount of B(p-C6F4H)3 (5 mol %) at ambient temperature. In
this case, however, NMR data were consistent with a product
that was not a primary phosphine but rather the secondary
While the combination of P(SiMe3)3 and 4-heptanone shows
no evidence of reaction, the stoichiometric combination of a
1:1 ratio of P(SiMe3)3 and B(p-C6F4H)3 with 3 equivalents of
4-heptanone in toluene at ambient temperature produced the
phosphine−borane adduct [(CH3CH2CH2)2C(OSiMe3)]PH2-
Received: July 23, 2013
Published: August 12, 2013
© 2013 American Chemical Society
4469
dx.doi.org/10.1021/om4007278 | Organometallics 2013, 32, 4469−4472

Organometallics
Communication
and 2 to 100 °C in the presence of 5 mol % B(p-C6F4H)3 for an
additional 24 h eventually gave rise to an insoluble orange
precipitate. At this point, the 31P NMR spectrum of the mother
1
liquor showed a new broad quartet at δ −238.9 with JPH = ca.
180 Hz, corresponding to PH3. However, monitoring the
thermal degradation of adduct 1 in solution upon heating to
100 °C for 24 h by NMR spectroscopy again showed the
gradual production of a new adduct 5 accompanied by the loss
of 2. Removal of the solvent and 2 in vacuo, followed by
recrystallization of the residues in 10:1 n-pentane/CH2Cl2,
1
afforded colorless crystals of 5 in 35% yield. The H NMR
spectrum of 5 shows the characteristic doublet of the PH3 unit
at δ 5.19 (1JPH = 407 Hz). This signal is similar but shifted
downfield from that reported for the adduct (H3P)B(C6F5)3 (δ
3.34, JPH = 410 Hz).6 The 31P NMR spectrum of 5 shows a
1
quartet with the same coupling constant at δ −93.5, consistent
with that of the PH3 moiety, while the 11B{1H} NMR spectrum
shows a broad signal at δ −16.1, indicating a 4-coordinate B
atom. Collectively, these data are consistent with the
formulation of 5, as (H3P)B(p-C6F4H)3. This assignment was
confirmed by an X-ray diffraction analysis (Figure 2),11 which
Figure 1. POV−Ray depiction of 1. P: orange; Si: sky blue; O: red; F:
pink; C: black; B: green.
phosphine [(CH3CH2CH2)2C(OSiMe3)]2PH (3), accompa-
nied by 2 (E/Z = 1:4) (see the Supporting Information). The
1
1H NMR spectrum of 3 shows a doublet at δ 3.44 with JPH
=
207 Hz, while the 31P NMR spectrum exhibits a doublet with
the same coupling constant at δ −18.8, corresponding to a
secondary phosphine. Heating the toluene solution of 3 to 50
°C for 72 h in the presence of 5 mol % B(p-C6F4H)3 results in
gradual conversion of the resonances at δH 3.44 and δP −18.8,
to a new doublet at δH 3.00 and a triplet signal at δP −98.4 with
1JPH = 192 Hz, again accompanied by the production of 2 (E/Z
= 1:4). The change in chemical shift is indeed consistent with
the production of the primary phosphine (CH3CH2CH2)2C-
(OSiMe3)PH2 4, which although not isolated, could be
converted to adduct 1 upon the addition of B(p-C6F4H)3,
enabling its facile spectroscopic identification (Scheme 2). It is
also interesting to note that addition of 1 equivalent of B(p-
C6F4H)3 to 3 prompted the immediate formation of adduct 1
and silyl enol ether 2 at room temperature.
Figure 2. POV−Ray depiction of 5. P: orange; F: pink; C: black; B:
green.
Compound 3 was stable on heating to 50 °C in the absence
of B(p-C6F4H)3, consistent with a borane-catalyzed elimination
of silyl enol ether 2 on formation of 1. Heating the mixture of 4
showed the expected pseudo-tetrahedral geometries of the P
and B centers, with a P−B bond length of 2.042(2) Å and P−H
bond lengths of 1.29(2), 1.30(2), and 1.33(2) Å, similar to
those of H3P−B(C6F5)3 [P−B = 2.044(8) Å, P−H = 1.24(3) to
1.36(3) Å].13
Scheme 2. Syntheses of 3-5
Further heating of 5 at 100 °C resulted in its gradual
decomposition with increased production of the isolable orange
precipitate. The solid was identified as elemental phosphorus
(>90 wt %) by elemental analysis, ICP-AES, and DART-MS
data. The air-stable nature of this red-orange solid is also
consistent with its characterization as elemental (red)
phosphorus.
In consideration of the mechanisms of the above reactions, it
is important to note that the use of B(p-C6F4H)3 is critical
because this borane does not undergo para-attack by
silylphosphine, as is seen with B(C6F5)3.14 This combination
therefore represents a stable FLP. As shown in Scheme 3, a
plausible mechanism for the catalytic reaction of ketone and
P(SiMe3)3 in the presence of B(p-C6F4H)3 is thought to be
initiated by activation of the carbonyl group by coordination of
O to B. This prompts nucleophilic attack of P(SiMe3)3 at C to
give the proposed zwitterionic intermediate Int1. Migration of
4470
dx.doi.org/10.1021/om4007278 | Organometallics 2013, 32, 4469−4472

Organometallics
Communication
Scheme 3. Proposed Reaction Mechanism for Desilylation of
P(SiMe3)3
Scheme 4. Reactions of Ph2P(SiMe3) with Ketone and
Catalytic Borane
adducts are isolated, whereas in catalytic reactions, the free
phosphines are generated. These reactions result from the
generation of a phosphine/borane FLP, allowing activation of
the ketone by borane, which then undergoes nucleophilic attack
by phosphine. Subsequent silyl-migration to O and β-
elimination generates the P−H bond. To the best of our
knowledge, this report describes the first reactions of an
unactivated dialkylketone with an FLP. Nonetheless, it is
noteworthy that Erker et al. have recently reported FLP
activations of conjugated ketones, such as ynones and
enediones, affording addition products.15 The utility of the
current FLP-catalyzed reactivity in the synthesis of unique
phosphine products and the mechanism of Lewis acid-catalyzed
P−P dehydrocoupling continues to be the subject of current
investigations.
a silyl group from phosphorus to oxygen is concurrent with
liberation of the borane catalyst, affording the intermediate
phosphine, (Me3Si)2PC(CH2CH2CH3)2(OSiMe3) Int2. Steric
crowding and coordination of the borane, B(p-C6F4H)3, is
thought to prompt 1,3-migration of proton from C to P,
producing 3 with concurrent β-elimination of silyl enol ether 2.
Conversely, one can envision a mechanism whereby each
silaphosphination step is followed by an elimination step
(rather than three silylphosphination steps followed by three
elimination steps). To the best of our knowledge, reactions
between P(SiMe3)3 and ketones in the absence of a Lewis acid
catalyst have not been reported. It is noteworthy that the
combination of P(SiMe3)3/B(p-C6F4H)3 did not react with the
nonenolizable ketones such as benzophenone or di-tert-
butylketone, even upon heating to 100 °C for 24 h, suggesting
that the elimination of silyl enol ether drives these reactions to
completion. Nonetheless, in a related sense, the silylphosphine,
Me3SiPEt2, has been previously reported to react with α-
diketones to give Me3SiOC(Me)PR2C(O)Me.14 Products 4
and PH3 are thought to be produced by successive borane-
catalyzed β-eliminations from 3 and 4. In these catalytic
reactions, formation of the secondary phosphine 3 at ambient
conditions presumably occurs because this species retains
significant steric congestion about phosphorus and thus
requires additional energy to prompt further borane-catalyzed
β-elimination to give 4 and PH3. Binding the toxic PH3 gas to
B(p-C6F4H)3 affords the isolable adduct 5.
In order to probe the generality of this FLP-mediated
hydrodesilylation reaction, the monosilylphosphine,
Ph2PSiMe3, was used as a precursor. Heating a 0.06 M
C6D5Br solution of 1:1 Ph2PSiMe3/4-heptanone with 5 mol %
B(p-C6F4H)3 to 130 °C resulted in the complete consumption
of the starting material Ph2PSiMe3 and the generation of
Ph2PH (δP −40.3) as the major product (Scheme 4). The yield
of the secondary phosphine was 80%, as evidenced by 31P{1H}
NMR spectroscopy. Interestingly, a minor product that was
formed in this reaction in 12% yield was identified as the
biphosphine product of homodehydrocoupling Ph2PPPh2 (δP
−15.2). Increasing the reaction concentration under otherwise
identical reaction conditions resulted in >95% conversion to
Ph2PPPh2.
ASSOCIATED CONTENT
* Supporting Information
Preparations of compounds, reaction procedures, spectra, and
experimental crystallographic details. This material is available
free of charge via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
*E-mail: dstephan@chem.utoronto.ca.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
D.W.S. gratefully acknowledges the financial support of the
NSERC of Canada and the award of a Canada Research Chair,
and K.T. is grateful for the award of a JSPS Postdoctoral
Fellowship for Research Abroad.
REFERENCES
■
(1) Adrio, L. A.; Hii, K. K. Organomet. Chem. 2009, 35, 62−92.
(2) (a) Hiney, R. M.; Ficks, A.; Muller-Bunz, H.; Gilheany, D. G.;
Higham, L. J. Organomet. Chem. 2011, 37, 27−45. (b) Higham, L. J.
Catal. Met. Complexes 2011, 37, 1−19. (c) Marinetti, A.; Genet, J.-P.
C.R. Chim. 2003, 6, 507−514. (d) Katti, K. V.; Pillarsetty, N.;
Raghuraman, K. Top. Curr. Chem. 2003, 229, 121−141. (e) Malisch,
W.; Klupfel, B.; Schumacher, D.; Nieger, M. J. Organomet. Chem. 2002,
661, 95−110. (f) Prabhu, K. R.; Kishore, P. N.; Gali, H.; Katti, K. V.
Curr. Sci. 2000, 78, 431−439.
(3) (a) Zhao, D.; Wang, R. Chem. Soc. Rev. 2012, 41, 2095−2108.
(b) Shaikh, T. M.; Weng, C.-M.; Hong, F.-E. Coord. Chem. Rev. 2012,
256, 771−803. (c) Glueck, D. S. Chem.−Eur. J. 2008, 14, 7108−7117.
(d) Ackermann, L. In Phosphorus Ligands in Asymmetric Catalysis;
In summary, the stoichiometric and catalytic reactions of
silylphosphines in the presence of the borane, B(p-C6F4H)3,
and the ketone, 4-heptanone, provide a synthetic route to
secondary and primary phosphines with liberation of the silyl
enol ether 2 via SiMe3 group migration from phosphorus to
oxygen. In stoichiometric reactions the phosphine−borane
Borner, A., Ed.; 2008; Vol. 2, pp 831−847; (e) Honaker, M. T.;
̈
Hovland, J. M.; Salvatore, R. N. Curr. Org. Synth. 2007, 4, 31−45.
(4) Stephan, D. W.; Erker, G. Angew. Chem., Int. Ed. 2010, 49, 46−76.
4471
dx.doi.org/10.1021/om4007278 | Organometallics 2013, 32, 4469−4472

Organometallics
Communication
(5) Stephan, D. W. Org. Biomol. Chem. 2012, 10, 5740−5746.
(6) Geier, S. J.; Stephan, D. W. Chem. Commun. 2010, 46, 1026−
1028.
(7) Takeuchi, K.; Stephan, D. W. Chem. Commun. 2012, 48, 11304−
6.
(8) (a) Ullrich, M.; Lough, A. J.; Stephan, D. W. Organometallics
2010, 29, 3647−3654. (b) Ullrich, M.; Lough, A. J.; Stephan, D. W. J.
Am. Chem. Soc. 2009, 131, 52−53.
(9) The presence of silylenol ether 2 was verified by GC/MS, and the
resulting data are consistent with those provided by Wiley
Subscription Services, Inc. (US). 1H and 13C{1H} NMR data are
consistent with those provided within the previous report: Davis, F. A.;
Yang, B. J. Am. Chem. Soc. 2005, 127, 8398.
(10) Welch, G. C.; Holtrichter-Roessmann, T.; Stephan, D. W. Inorg.
Chem. 2008, 47, 1904−1906.
(11) X-ray crystallographic data for 1: triclinic, P1, a = 9.9406(8) Å, b
̅
= 11.7049(8) Å, c = 13.254(1) Å, α = 95.840(4)°, β = 97.626(4)°, γ =
102.456(4)°, V = 1478.89(19) Å3, Z = 2, data (>2σ) = 26289, variables
= 7239, R1(>2σ) = 0.0348, wR2(all) = 0.1270, GOF = 0.923. 5:
Orthorhombic, Pbca, a = 10.1556(3) Å, b = 17.4865(5) Å, c =
20.2840(5) Å, V = 3602.2(2) Å3, Z = 8, data (>2σ) = 16575, variables
= 4140, R1(>2σ) = 0.0328, wR2(all) = 0.0819, GOF = 1.017.
(12) Bradley, D. C.; Harding, I. S.; Keefe, A. D.; Motevalli, M.;
Zheng, D. H. J. Chem. Soc., Dalton Trans. 1996, 3931.
(13) Bradley, D. C.; Hursthouse, M. B.; Motevalli, M.; Zheng, D. H.
J. Chem. Soc., Chem. Commun. 1991, 7−8.
(14) Schulz, F.; Sumerin, V.; Leskela, M.; Repo, T.; Rieger, B. Dalton
̈
Trans. 2010, 39, 1920−2.
(15) Xu, B.; Yanez, R.; Nakatsuka, H.; Kitamura, M.; Frohlich, R.;
Kehr, G.; Erker, G. Chem. Asian J. 2012, 7, 1347−1356.
̈
4472
dx.doi.org/10.1021/om4007278 | Organometallics 2013, 32, 4469−4472