Main group redox catalysis: Reversible PIII/PV redox cycling at a phosphorus platform

Article abstract of DOI:10.1021/ja302963p

A planar, trivalent phosphorus compound is shown to undergo reversible two-electron redox cycling (P^III/P^V) enabling its use as catalyst for a transfer hydrogenation reaction. The trivalent phosphorus compound activates ammonia-borane to furnish a 10-P-5 dihydridophosphorane, which in turn is shown to transfer hydrogen cleanly to azobenzene, yielding diphenylhydrazine and regenerating the initial trivalent phosphorus species. This result constitutes a rare example of two-electron redox catalysis at a main group compound and suggests broader potential for this nonmetal platform to support bond-modifying redox catalysis of the type dominated by transition metal catalysts.

Full text of DOI:10.1021/ja302963p

Communication
pubs.acs.org/JACS
Main Group Redox Catalysis: Reversible PIII/PV Redox Cycling at a
Phosphorus Platform
Nicole L. Dunn, Minji Ha, and Alexander T. Radosevich*
Department of Chemistry, The Pennsylvania State University, University Park, Pennsylvania 16802, United States
S
* Supporting Information
The notion that elements outside the d-block, in particular
ABSTRACT: A planar, trivalent phosphorus compound is
shown to undergo reversible two-electron redox cycling
(PIII/PV) enabling its use as catalyst for a transfer
hydrogenation reaction. The trivalent phosphorus com-
pound activates ammonia-borane to furnish a 10-P-5
dihydridophosphorane, which in turn is shown to transfer
hydrogen cleanly to azobenzene, yielding diphenylhydra-
zine and regenerating the initial trivalent phosphorus
species. This result constitutes a rare example of two-
electron redox catalysis at a main group compound and
suggests broader potential for this nonmetal platform to
support bond-modifying redox catalysis of the type
dominated by transition metal catalysts.
nonmetal p-block elements, may be able to exhibit the redox
reactivity traditionally reserved for the transition metals has
recently been articulated by Bertrand,9 Power,10 and Stephan
and Erker.11 The elementary steps common to redox catalytic
mechanisms (e.g., oxidative addition, reductive elimination)
have been demonstrated in stoichiometric fashion for a large
number of the p-block elements,12−14 and recent breakthroughs
have further expanded the scope of this reactivity, demonstrat-
ing the ability of nonmetals to activate even strong, nonpolar
bonds under mild conditions.15,16 Despite this promising
stoichiometric reactivity, catalytic processes employing redox
catalysts based on p-block elements remain extremely rare.17,18
We now report a nonmetal, phosphorus-based platform that
participates in reversible, two-electron bond modifying
reactivity in analogy to the transition metals. We further
show that this ability to cycle freely between PIII/PV oxidation
states may be harnessed to effect a catalytic transfer
hydrogenation reaction where the central redox element is
not of the d-block.19 This observation extends the reach of
bond activation catalysis beyond the transition metal series,
further cementing the role of certain main group compounds as
viable metal-catalyst surrogates.20
We adopted a structurally unique tricoordinate phosphorus
compound 1, described initially by Arduengo,21−23 as a point of
departure in our study (Figure 1). Whereas three-coordinate
phosphorus commonly assumes a pyramidal geometry,
compound 1 displays a highly unusual T-shaped geometry.
When considered in conjuction with the related 10-P-5
compound 2, this platform suggested a structural analogy to
transition metal pincer complexes,24 where meridionally ligated
metals localize diverse bond activation processes at sites
perpendicular to the ligand plane.
omogeneous transition metal catalysis has profoundly
1
Himpacted chemical synthesis, enabling the construction
of covalent bonds with exquisite levels of efficiency and
selectivity.2−5 Methods based on late transition metal catalysts
(Rh, Ir, Pd, Pt) are particularly prominent. The facility with
which these complexes mediate the making and breaking of
covalent bonds stems in part from the ability of d-block metals
to readily cycle between n and n + 2 oxidation states with
concomitant expansion and contraction of their primary
coordination sphere (Figure 1).6 This reversible coordinative
and oxidative reactivity is intimately linked to the energetic and
spatial disposition of the valence d-orbitals about the metal
center.7 Consequently, catalytic two-electron bond modifying
redox reactivity has traditionally been considered the exclusive
purview of the transition metals.8
Motivated by the perceived analogy to pincer chemistry, we
evaluated the capacity of 1 to affect dehydrogenation of small
molecule substrates.25 Three-coordinate 1 is stable in alkane
solutions; however, reaction with the alkane isostere ammonia-
borane26 in acetonitrile solution results in conversion to a single
new phosphorus-containing compound with a 31P{1H} NMR
resonance at δ −43.7 ppm. This signal evolves in the proton-
coupled 31P NMR spectrum to a triplet of triplets with J = 670
and 34 Hz (Figure 2, bottom). In situ monitoring of the
reaction progress indicates that no other phosphorus-
containing intermediates or products are formed. Spectroscopi-
cally pure material may be isolated in 75% yield following
Figure 1. (Left) Schematic representation of redox cycling at a metal
complex during transfer hydrogenation. (Right) Phosphorus com-
pounds 1−3 employed in this study.
Received: March 27, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja302963p | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Figure 3. (Left) Thermal ellipsoid plot (50%) of 3. Hydrogen atoms
on carbon omitted for clarity. Selected bond distances (Å) and angles
(deg): P(1)−N(1) = 1.676(2); P(1)−O(1) = 1.721(1); P(1)−O(2) =
1.699(1); P(1)−H(1) = 1.29(2); P(1)−H(2) = 1.31(2); O(1)−
P(1)−O(2) = 170.5(1); N(1)−P(1)−H(1) = 134.8(1); N(1)−P(1)−
H(2) = 116.1(1); H(1)−P(1)−H(2) = 109.1(2). (Right) Schematic
illustrating bond angles.
1
Figure 2. (Top) Abridged H NMR spectrum for 3 in CDCl3. An
1
additional H signal for −C(CH3)3 is found at δ 1.06 ppm. (Bottom)
Annotated 31P NMR spectrum for 3 in CDCl3. Units are ppm relative
to Me4Si (1H) and H3PO4 (31P).
pentane trituration of the crude reaction residue. Alternatively,
the action of sodium cyanoborohydride on dichlorophosphor-
ane 2 gives the same species in 59% isolated yield, which we
assign as dihydridophosphorane 3 (Scheme 1).27,28 The
azobenzene in acetonitrile-d3 solution at 40 °C for 19 h results
in 72 5% conversion of 3 to 1 with concomitant formation of
1,2-diphenylhydrazine in 63 5% yield (eq 1). In the presence
Scheme 1. Preparation of Dihydridophosphorane 3
of excess azobenzene (which mimics catalytic conditions, vide
infra), the conversion of 3 to 1 can be driven cleanly to
completion. When monitored by NMR, no other intermediates
or products are evident in this hydrogen transfer. The
consumption of 3 in acetonitrile-d3 solution in the presence
of excess azobenzene follows pseudo-first order kinetics, and an
Eyring analysis over the temperature range 30−60 °C furnishes
magnitude and multiplicity of the 31P NMR resonance are
indicative of 1JPH coupling of the phosphorus center with bound
hydrogen nuclei of an apparently equivalent pair (i.e., a −P(H)2
3
unit, vide infra) and longer range JPH coupling to the remote
vinylic hydrogen nuclei. The complementary coupling securing
1
these assignments is evident in the H NMR spectrum (Figure
ΔH⧧ = 12.4
0.7 kcal/mol and ΔS⧧ = −36
7 eu (see
2, top); phosphorus-coupled doublets are observed at δ 7.95
Supporting Information (SI)). Although a more detailed
mechanistic description awaits further study, the sign and
magnitude of the entropy of activation are consistent with a
rate-limiting step involving a transition structure from
bimolecular reaction.
ppm (1JPH = 670 Hz) and δ 5.83 ppm (3JPH = 34 Hz).
X-ray diffraction of a colorless single crystal obtained by
sublimation of 3 in vacuo at 35 °C affords additional structural
detail. The structure of 3 is found to deviate, albeit modestly,
from an idealized 10-P-5 trigonal bipyramid toward a distorted
square pyramid about phosphorus (Figure 3). The phosphorus
center resides in a plane defined by the nitrogen atom and the
two P−H hydrogen atoms. The positions of the P−H
hydrogens, both of which were located on the difference
Fourier map, are splayed obliquely with respect to the O,N,O-
supporting structure and are crystallographically inequivalent
(∠N−P−H1 = 134.8(1)°; ∠N−P−H2 = 116.1(1)°).29
Tridentate binding of the O,N,O-motif is evident with oxygen
atoms spanning the phosphorus center at an angle of 170.5(1)°.
The ring systems experience only a minor puckering with
nitrogen displaying slight pyramidalization (φ = 6.7°).30
Reaction of dihydridophosphorane 3 with an unsaturated
organic substrate results in transfer of hydrogen from
phosphorus, regenerating reduced three-coordinate 1.31 In a
stoichiometric experiment, exposure of 3 to 1 equiv of
The combined reactivities of 1 as hydrogen acceptor from
ammonia-borane and 3 as hydrogen donor to an organic
substrate suggested the possibility of employing this
phosphorus platform as a redox catalyst for transfer hydro-
genation. The direct reaction of azobenzene with a 4-fold excess
of ammonia-borane in acetonitrile solution at 80 °C proceeds
only to minimal conversion (ca. 5%) after 24 h (Table 1, entry
1). However, the addition of 10 mol % of 1 yields
diphenylhydrazine in 80% yield in the same time frame
(entry 2), even under less forcing conditions (40 °C). Longer
reactions times lead to modestly increased yields (entry 3).
Presumably, under these conditions, three-coordinate 1 accepts
hydrogen from ammonia-borane forming dihydridophosphor-
ane 3, then subsequent hydrogen transfer to azobenzene
regenerates 1 and closes the catalytic cycle. Consistent with this
scheme, the use of a catalytic amount of 3 also promotes
B
dx.doi.org/10.1021/ja302963p | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
These results demonstrate the capacity for a three-coordinate
phosphorus compound to support two electron redox
chemistry of the type commonly observed in the transition
metals. This constitutes a rare example of two-electron bond
modifying redox catalysis outside of the d-block. Moreover,
whereas phosphines are well-known nucleophilic32 and
frustrated Lewis pair (FLP) catalysts,33 the behavior of the
phosphorus center described above is of a different type. In
both nucleophilic and FLP paradigms, the phosphorus behaves
exclusively as an electron donor.34 By contrast, catalysis with 1
results from cycling between discrete tricoordinate PIII and
pentacoordinate PV states, capitalizing on the colocalization of
electron-donating and electron-accepting function at a single
atomic center. It is precisely this coordinative and oxidative
reactivity that enables the valuable two-electron redox trans-
formations observed in transition metal catalysis. The
identification of this reactivity in 1 suggests broader potential
for this nonmetal platform to support additional bond-
modifying redox catalysis of the type currently dominated by
transition metal catalysts. Research along these lines is currently
underway in our laboratory.
Table 1. Catalytic Transfer Hydrogenation with 1/3
a
b
Loading relative to azobenzene substrate. Determined by1H NMR
of crude reaction mixtures. Error estimate 5%.
transfer hydrogenation (entries 4 and 5). Interrogation of the
phosphorus speciation by 31P NMR under the catalytic
conditions supports this mechanistic outline; the signal
corresponding to 1 (δ 186.7 ppm) is replaced by
dihydridophosphorane 3 (δ −43.7 ppm) at early time points,
after which compound 3 persists in solution as the sole
observable phosphorus species in the range +250 to −150 ppm
until the catalytic reaction is terminated at 48 h (Figure 4).
ASSOCIATED CONTENT
* Supporting Information
Synthetic procedures and spectra, kinetic data, crystallographic
and computational details. This material is available free of
charge via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
radosevich@psu.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We gratefully acknowledge the Pennsylvania State University
for financial support of this work. We thank Dr. Hemant
Yennawar for crystallographic structure determination of 3.
REFERENCES
■
(1) Parshall, G. W.; Ittel, S. D. Homogeneous Catalysis; Wiley: New
York, 1992.
(2) Beller, M.; Bolm, C.; Eds. Transition Metals for Organic Synthesis,
2nd ed.; Wiley: Weinheim, 2004.
(3) Diederich, F.; Stang, P. J. Metal-Catalyzed Cross Coupling
Reactions; Wiley-VCH: Weinheim, 1998.
(4) De Vries, J. G.; Elsevier, C. J. Handbook of Homogeneous
Hydrogenation; Wiley-VCH, Weinheim, 2007.
(5) Ojima, I, Ed. Catalytic Asymmetric Synthesis, 3rd ed.; Wiley: New
York, 2010.
(6) (a) Collman, J. P. Acc. Chem. Res. 1968, 1, 136. (b) Halpern, J.
Acc. Chem. Res. 1970, 3, 386.
(7) Niu, S.; Hall, M. B. Chem. Rev. 2000, 100, 353.
(8) Hartwig, J. F. Organotransition Metal Chemistry; University
Science Books: Sausalito, CA, 2010; p 261.
Figure 4. Evolution of phosphorus species during transfer hydro-
genation catalysis with 1 (Table 1, entry 3). Time-stacked 31P NMR
spectra at t = 0, 3, 8, and 48 h. See SI for full, unabridged spectra.
(9) Martin, D.; Soleilhavoup, M.; Bertrand, G. Chem. Sci. 2011, 2,
389.
(10) Power, P. P. Nature 2010, 463, 171.
(11) Stephan, D. W.; Erker, G. Angew. Chem., Int. Ed. 2010, 49, 46.
(12) Oae, S.; Uchida, Y. Acc. Chem. Res. 1991, 24, 202.
(13) Akiba, K. Y. Chemistry of Hypervalent Compounds; Wiley VCH:
New York, 1999.
These results implicate 3 as the resting state of the catalytic
cycle. The redox platform 1/3 is unique among phosphorus
species surveyed in enabling this catalytic transfer hydro-
genation reactivity; none of the other trivalent phosphorus
compounds examined promote the transfer hydrogenation at a
comparable level, even at higher loadings and temperatures
(entries 4−6).
(14) Finet, J.-P. Ligand Coupling Reactions with Heteroatomic
Compounds; Elsevier: Oxford; 1998.
C
dx.doi.org/10.1021/ja302963p | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
(15) Frey, G. D.; Lavallo, V.; Donnadieu, B.; Schoeller, W. W.;
Bertrand, G. Science 2007, 316, 439.
(16) For C−H activation at a redox active Al(III)−oxo complex, see:
Myers, T. W.; Berben, L. A. J. Am. Chem. Soc. 2011, 133, 11865.
(17) For phosphine-catalyzed redox transformations, see: (a) McGo-
nagle, A. E.; Marsden, S. P.; McKever-Abbas, B. Org. Lett. 2008, 10,
2589. (b) O’Brien, C. J.; Tellez, J. L.; Nixon, Z. S.; Kang, L. J.; Carter,
A. L.; Kunkel, S. R.; Przeworski, K. C.; Chass, G. C. Angew. Chem., Int.
Ed. 2009, 48, 6836. (c) Cao, J.-J.; Zhou, F.; Zhou, J. Angew. Chem., Int.
Ed. 2010, 49, 4976. (d) Harris, J. R.; Haynes, M. T., II; Thomas, A. M.;
Woerpel, K. A. J. Org. Chem. 2010, 75, 5083. (e) Denton, R. M.; Jie,
A.; Adeniran, B. Chem. Commun. 2010, 46, 3025.
(18) For a review of organoiodine-catalyzed redox transformations,
see: Dohi, T.; Kita, Y. Chem. Commun. 2009, 2073.
(19) Bullock, R. M. Catalysis without Precious Metals; Wiley VCH:
Weinheim, 2010.
(20) For instance, see organocatalysis: (a) MacMillan, D. W. C.
Nature 2008, 455, 304. (b) List, B. Chem. Rev. 2007, 107, 5413.
(21) Culley, S. A.; Arduengo, A. J. J. Am. Chem. Soc. 1984, 106, 1164.
(22) (a) Arduengo, A. J.; Stewart, C. A.; Davidson, F.; Dixon, D. A.;
Becker, J. Y.; Culley, S. A.; Mizen, M. B. J. Am. Chem. Soc. 1987, 109,
627. (b) Arduengo, A. J.; Stewart, C. A. Chem. Rev. 1994, 94, 1215.
(23) Driess, M.; Muresan, N.; Merz, K.; Pach, M. Angew. Chem., Int.
Ed. 2005, 44, 6734.
(24) Morales-Morales, D.; Jensen, C. M. The Chemistry of Pincer
Compounds; Elsevier: Amsterdam, 2007.
(25) Choi, J.; MacArthur, A. H. R.; Brookhart, M.; Goldman, A. S.
Chem. Rev. 2011, 111, 1761.
(26) Staubitz, A; Robertson, A. P. M.; Sloan, M. E.; Manners, I.
Chem. Rev. 2010, 110, 4023.
(27) Dihydridophosphorane 3 has been mentioned in the literature
(see ref 22b), but no procedure for its preparation has been reported
in either the peer-review or patent literature.
(28) For a related dihydridophosphorane, see: Murillo, A.; Chiquete,
M. L.; Joseph-Nathan, P.; Contreras, R. Phosphorus, Sulfur Silicon Relat.
Elem. 1990, 53, 87.
(29) Calculations (B3LYP/6-311+G*) reproduce this nonsymmetric
ground state structure for 3. A higher symmetry C2v-structure resides
at first-order saddle point (+0.7 kcal/mol). See Supporting
Information for full details.
(30) Pyramidalization angle φ is defined as: cos φ = −cos(Cα-N-P)/
[cos 1/2(Cα-N-Cα′)]. See: Borden, W. T. Chem. Rev. 1989, 89, 1095.
(31) For related hydridic P−H reactivity, see: (a) Gudat, D.;
Haghverdi, A.; Nieger, M. Angew. Chem., Int. Ed. 2000, 39, 3084.
(b) Burck, S.; Gudat, D.; Nieger, M.; Du Mont, W. W. J. Am. Chem.
Soc. 2006, 128, 3946. (c) Miyake, H.; Kano, N.; Kawashima, T. J. Am.
Chem. Soc. 2009, 131, 16622. (d) Miyake, H.; Kano, N.; Kawashima,
T. Inorg. Chem. 2011, 50, 9083.
(32) For reviews of phosphine-based nucleophilic catalysis, see:
(a) Methot, J. L.; Roush, W. R. Adv. Synth. Catal. 2004, 346, 1035.
(b) Ye, L.-W.; Zhou, J.; Tang, Y. Chem. Soc. Rev. 2008, 37, 1140.
(c) Lu, X.; Zhang, C.; Xu, Z. Acc. Chem. Res. 2001, 34, 535.
(33) For reviews of frustrated Lewis pair reactivity and catalysis, see:
(a) Stephan, D. W. Org. Biomol. Chem. 2008, 6, 1535. (b) Kenward, A.
L.; Piers, W. E. Angew. Chem., Int. Ed. 2007, 47, 38.
(34) For Lewis base catalysis, see: Denmark, S. E.; Beutner, G. L.
Angew. Chem., Int. Ed. 2008, 47, 1560.
D
dx.doi.org/10.1021/ja302963p | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX