Dihydrogen activation with tBu3P/B(C 6F5)3: A chemically competent indirect mechanism via in situ-generated p -TBu2P - C 6F4 - B(C6F5)2

Article abstract of DOI:10.1021/ja203214f

A chemically competent indirect pathway for the activation of dihydrogen by the nonmetal Lewis acid/Lewis base pair ^tBu₃P/B(C ₆F₅)₃ is described. The reaction between ^tBu₃P and B(C₆F₅)₃ produces [^tBu₃PH]⁺[FB(C₆F ₅)₃]^- and the known phosphinoborane p- ^tBu₂P - C₆F₄ - B(C₆F ₅)₂ (1-^tBu) with elimination of isobutylene. At 1:1 stoichiometry, 1-^tBu is produced rapidly in detectable quantities and can act as a catalyst for the formation of [^tBu ₃PH]⁺[HB(C₆F₅)₃] ^- from ^tBu₃P and B(C₆F ₅)₃ in the presence of H₂. The extent to which this indirect path competes with the direct path is explored.

Full text of DOI:10.1021/ja203214f

COMMUNICATION
pubs.acs.org/JACS
Dihydrogen Activation with tBu3P/B(C6F5)3: A Chemically Competent
Indirect Mechanism via in Situ-Generated p-tBu2PꢀC6F4ꢀB(C6F5)2
Adam J. V. Marwitz, Jason L. Dutton, Lauren G. Mercier, and Warren E. Piers*
Department of Chemistry, University of Calgary, 2500 University Drive NW, Calgary, Alberta, Canada T2N 1N4
S Supporting Information
b
detail available concerning the intimate or even stoichiometric
ABSTRACT: A chemically competent indirect pathway for
the activation of dihydrogen by the nonmetal Lewis acid/
Lewis base pair tBu3P/B(C6F5)3 is described. The reaction
mechanism by which the H2 activation takes place. Qualitatively,
the rates in these R3P/B(C6F5)3 systems appear to be lower11 than
that observed for 1-Mes,7 but no serious comparative kinetic studies
have been reported. What we do know stems mainly from theo-
retical treatments of these important reactions. For compound 1-
Mes, addition of H2 to the BꢀC bondof the PR2-functionalized aryl
group via a side-on H2 adduct of the Lewis acidic boron center,
followed by a series of 1,2-hydrogen shifts, was proposed initially7
and has been shown computationally to be an energetically feasible
pathway.14 However, the computed barriers are too high to account
for the reported rates of this reaction, which occurs rapidly even at
temperatures below 0 °C. In the tBu3P/B(C6F5)3 system, however,
it is thought that an η2-H2 adduct is not involved;15 rather, it has
been proposed that the phosphine and borane form a loosely
associated “encounter complex”16 that is stabilized by several weak
t
between Bu3P and B(C6F5)3 produces [tBu3PH]þ[FB-
(C6F5)3]ꢀ and the known phosphinoborane p-tBu2PꢀC6F4ꢀ
B(C6F5)2 (1-tBu) with elimination of isobutylene. At 1:1
stoichiometry, 1-tBu is produced rapidly in detectable quan-
tities and can act as a catalyst for the formation of [tBu3-
t
PH]þ[HB(C6F5)3]ꢀ from Bu3P and B(C6F5)3 in the pre-
sence of H2. The extent to which this indirect path competes
with the direct path is explored.
esearch into the activation of dihydrogen1,2 and other
R
nonpolar bonds3,4 by non-transition-metal systems has
reached a crescendo of activity after recent discoveries in which
Lewis acid/Lewis base pairs that do not form traditional adducts
were observed to heterolytically cleave dihydrogen under ambi-
ent conditions.5,6 The first such system was the linked phosphi-
noborane 1-Mes7 (Chart 1), which is converted rapidly (<5 min)
to the corresponding phosphonium borate zwitterion upon
exposure to 1 atm H2.
H
F interactions and dispersion forces (I in Chart 1), creating a
3 3 3
pocket in which the δꢀ charge on the phosphine and the δþ charge
on boron generate an electric field capable of polarizing H2 to the
extent that bond cleavage is facile.15 A related mechanism for the
linked system 1-Mes has subsequently been probed and found to
be more consistent with experimentally observed rates.17 Thus,
there is substantial computational support for the encounter-com-
plex mechanism. However, to the best of our knowledge, despite the
Since then, several other systems have been developed,8ꢀ10
including a much studied pairing of tBu3P and B(C6F5)3,11 two
commercially available reagents that split hydrogen to form the
ion pair [tBu3PH]þ[HB(C6F5)3]ꢀ. Trimesitylphosphine is also
an able partner in this chemistry; furthermore, B(C6F5)3 has
been used in combination with a variety of other bulky bases,
notably imines,12 to bring about metal-free imine hydrogenations
via reactions related to the B(C6F5)3-catalyzed hydrosilylation of
carbonyl functions4 and imines.13 The term “frustrated Lewis
pair” has been introduced to describe these bond-activating
systems.5
reasonably substantial computed minima (ꢀ10 to ꢀ13 kcal molꢀ1
)
for such entities, all attempts to date to observe them via NMR
spectroscopy have failed.
t
Although solutions of 1:1 mixtures of Bu3P and B(C6F5)3
have been reported to be colorless,11 we observed that when very
dry solvents (toluene or CH2Cl2) and reagents18 were employed,
these solutions exhibited a pale-yellow color. Postulating that the
color may have arisen from the elusive encounter complex, we
endeavored to study these solutions by UVꢀvis spectroscopy in
both toluene and CH2Cl2. In order to study more concentrated
solutions, a 1 mm path length cuvette was employed. Control
Despite the facility of HꢀH bond cleavage by these systems
and their obvious utility, there is surprisingly little experimental
t
spectra established the absorption profiles of free Bu3P and
B(C6F5)3, which showed that neither absorbs above 365 nm; the
spectrum of B(C6F5)3 agreed with that reported previously (λmax
=
303 nm).19 In mixtures, however, a weak band with λmax = 373 nm
appeared on the shoulder of the main absorption band for B-
(C6F5)3, and we ascribe this band to the species that imbues these
solutions with the pale-yellow color (see Figure S1 in the Supporting
Information). In a scenario in which this species is the encounter
complex, the relative intensity of this band would be expected to
exhibit a monotonic concentration dependence, since higher
Chart 1
Received: April 7, 2011
Published: June 08, 2011
r
2011 American Chemical Society
10026
dx.doi.org/10.1021/ja203214f J. Am. Chem. Soc. 2011, 133, 10026–10029
|

Journal of the American Chemical Society
COMMUNICATION
Scheme 1
Scheme 2
Since the reported UVꢀvis spectrum of 1-tBu is dominated by
an absorption with λmax = 373 nm,20 it is clear that the yellow
t
color observed in dilute (and concentrated) mixtures of Bu3P
and B(C6F5)3 is due to 1-tBu, which likely is formed via the
chemistry depicted in Scheme 1. It is already well-known that
other phosphines react with B(C6F5)3 via attack of the phos-
phine at a p-fluoro group, resulting in the formation of zwitter-
ionic phosphonium fluoroborates. To our knowledge, the precise
species shown here (II) has not been previously reported, but it
may be responsible for the minor unassigned peaks in the 19F
NMR spectra that we recorded for these samples. In the presence
of a large excess of tBu3P and B(C6F5)3, the elements of HꢀF are
rapidly removed from II by deprotonation of a tert-butyl group
and abstraction of a fluoride to produce 2 and 1-tBu, as shown;
this also accounts for the observed isobutylene. We found that
the production of detectable amounts of 1-tBu occurs essentially
upon mixing of tBu3P and B(C6F5)3 in both toluene and CH2Cl2
solutions.
concentrations should favor its formation. This is a phenom-
enon that we did not observe in a reproducible way. Further-
more, using time-dependent density functional theory (DFT)
methods, we computed the UVꢀvis spectrum of the encounter
complex using published coordinates16 and did not find
evidence for a band that is red-shifted relative to the computed
spectrum of free B(C6F5)3.
In a further effort to favor formation of the encounter complex,
30 mg of B(C6F5)3 (0.06 mmol) was dissolved in 200 mg of
t
liquid Bu3P (1.0 mmol). An immediate color change to dark
yellow was observed upon mixing of the two reagents; the borane
dissolved completely in the phosphine, but soon after, colorless
crystals began to deposit from the yellow-orange mixture. After
several hours, these crystals were washed with cold toluene and
collected; an X-ray crystallographic analysis of a single crystal
from this batch showed it to be the phosphonium fluoroborate
[tBu3PH]þ[FB(C6F5)3]ꢀ (2); an ORTEP diagram of the salt
(Figure S2) and crystal and metrical data are given in the
Supporting Information. Multinuclear NMR spectroscopy con-
firmed its identity; in particular, the broad signal at ꢀ188 ppm in
the 19F NMR spectrum that integrated to one fluorine is
consistent with the formation of the fluoroborate anion.
t
The presence of 1-tBu in solutions of Bu3P and B(C6F5)3
raises significant questions concerning the mechanism by which
this particular Lewis pair activates dihydrogen. For example, as
mentioned above, it has been reported that 1-Mes activates H2
t
more rapidly than do solutions of Bu3P and B(C6F5)3. It is
therefore conceivable that 1-tBu is a catalyst for the formation of
[tBu3PH]þ[HB(C6F5)3]ꢀ via the pathway shown in Scheme 2.
t
In such a scenario, H2 addition to 1-tBu generated from Bu3P
The demands of a balanced chemical equation imply that in
addition to 2, isobutylene and the known phosphinoborane 1-tBu20
must also be produced. Thus, the maxiumum yield of 2 in the above
experiment is 50%; we recovered a yield of 48%, with the remainder
presumably lost in the isolation procedure. Accordingly, examina-
tion of a concentrated 1:1 tBu3P/B(C6F5)3 mixture by multinuclear
NMR spectroscopy confirmed the presence of 1-tBu, with a
characteristic doublet of doublets at 25 ppm in the 31P{1H} NMR
spectrum being the most convincing evidence.20 Compounds 2 and
1-tBu were formed in a 1:1 ratio, and with small amounts of other
fluorine-containing species, in addition to the starting materials,
were also present in these samples. Furthermore, when the reaction
was conducted in a sealed NMR tube, the presence of free
and B(C6F5)3 as described above would yield the known linked
phosphonium hydridoborate 3-tBu.20 From there, deprotona-
t
tion of the phosphonium moiety by free Bu3P likely would
precede hydride abstraction, since the phosphonium center in
3-tBu is more acidic than [tBu3PH]þ, while it has been shown
that the boron centers in phosphonium boranes are stronger
Lewis acids than that in B(C6F5)3.20,22,23 Indeed, the stoichio-
t
metric reaction of separately prepared 3-tBu and Bu3P led
rapidly (essentially upon mixing) to ion pair 4, as determined
by NMR spectroscopy (see the Supporting Information). In
particular, a new doublet for the [tBu3PH]þ cation at 5.25
ppm (1JPH = 433 Hz) and two characteristic signals at 59.9 and
20.9 ppm in the 31P NMR spectrum are indicative of formation of
4 from 3-tBu. Conversely, a separate experiment showed that
3-tBu does not surrender hydride to B(C6F5)3. However, treat-
ment of 4 with B(C6F5)3 generated [tBu3PH]þ[HB(C6F5)3]ꢀ
and 1-tBu, again essentially upon mixing.
1
isobutylene was detected in the H NMR spectrum (4.74 and
1.59 ppm). The expulsion of isobutylene from tert-butylphospho-
nium salts has been observed in other recent studies.20,21 Finally,
t
when a yellow Bu3P/B(C6F5)3 mixture was examined by MAL-
DIꢀTOF mass spectrometry, a molecular-ion peak for 1-tBu was
clearly observed in the mass spectrum in addition to peaks
associated with 2 and the Lewis pair components.
These experiments certainly demonstrate the chemical com-
petency of such a phosphinoborane-catalyzed pathway for hy-
t
drogen splitting by the Bu3P/B(C6F5)3 Lewis pair. However,
10027
dx.doi.org/10.1021/ja203214f |J. Am. Chem. Soc. 2011, 133, 10026–10029

Journal of the American Chemical Society
COMMUNICATION
Scheme 3
0.07 M C6D5Br solutions of 1:1 tBu3P/B(C6F5)3 and 1-tBu were
prepared, degassed, exposed to 4 atm H2, and warmed to room
temperature, and the reactions were monitored by 31P{1H} and
1H NMR spectroscopy. In the first sample, production of
[tBu3PH]þ[HB(C6F5)3]ꢀ was complete in less than 10 min
after exposure to H2; traces of 1-tBu were detected in the
baseline of the spectra. This reaction is thus very rapid and does
not require several hours to reach completion as initially
reported.11 In the reaction of pure 1-tBu with H2, the only
species present in the 31P NMR spectra were the starting
material and the expected zwitterionic product of H2 addition,
3-tBu; this process took ∼120 min to reach full conversion.
Thus, it appears that addition of hydrogen to 1-tBu in the
t
absence of Bu3P/B(C6F5)3 is slower than splitting of H2 by
tBu3P/B(C6F5)3 when only a trace of 1-tBu is present (i.e., to a
first approximation, kd > ka).
The comparatively slow addition of H2 to 1-tBu that we
observed occurred in the absence of tBu3P/B(C6F5)3. However,
in the reaction of the unlinked pair with H2, there was always a
small amount of 1-tBu present, which implies that the scenarios
outlined in Scheme 3b,c cannot be ignored. Significantly, when
t
1-tBu was mixed with 1 equiv of Bu3P in the absence of
B(C6F5)3 under the same conditions as described above and
then and exposed to 4 atm H2, ion pair 4 (Scheme 2) was formed
almost immediately (within 5 min) and quantitatively. The
analogous experiment aimed at probing Scheme 3c, wherein
1-tBu was mixed with 1 equiv of B(C6F5)3 and treated with H2,
resulted in addition of H2 to form 3-tBu at a rate consistent with
that observed in the absence of B(C6F5)3.
In an effort to further delineate the relative rates of these
reactions, the rapid processes were probed using the following
procedure. A solution of B(C6F5)3 (0.0488 M in toluene) was
degassed and exposed to H2 (1 atm); a separate solution of tBu3P
(also 0.0488 M in toluene) that had been degassed and saturated
with H2 was added in one portion to the solution of borane under
a strong flow of H2 with vigorous stirring. Precisely 1 min after
addition, the reaction was exposed to vacuum and evaporated to
dryness; the resulting residue was assayed by 1H NMR spectros-
copy using hexamethylbenzene as an internal standard, and it was
found that 38 ( 2% conversion to [tBu3PH]þ[HB(C6F5)3]ꢀ
had occurred. A similar experiment employing a mixture of 1-tBu
and tBu3P, each at 0.0488M, showed a similar extent of conver-
sion to the ion pair (35 ( 2%) after one minute. Thus, the rates of
H2 addition to these two frustrated Lewis pairs appear to be quite
similar (i.e., kd ≈ kb), despite25 the lower Lewis acid strength of
the borane center in 1-tBu in comparison with B(C6F5)3. When
20 mol % 1-tBu was added to the B(C6F5)3 solution, the
conversion to [tBu3PH]þ[HB(C6F5)3]ꢀ after 1 min increased
to 54 ( 2%. This increase was not large enough to indicate the
dominance of a catalyzed path of H2 addition to the Lewis pair
tBu3P and B(C6F5)3, but it does suggest that such a path could be
competitive with direct addition via the encounter complex I
under conditions where significant quantities of 1-tBu are present.
In conclusion, although the Lewis pair consisting of tBu3P and
B(C6F5)3 do not form a classical Lewis acid/base adduct, the
partners do react in a facile process via coordination of the tBu3P
to one of the para carbons of B(C6F5)3. This occurs because the
two essentially degenerate LUMOþ1 orbitals are situated mostly
on the para carbons (see Figure S3). This triggers the formation
of 1-tBu, which can serve as a catalyst for addition of H2 to the
tBu3P/B(C6F5)3 Lewis pair via a chemically competent sequence
of reaction steps. On the basis of semiquantitative kinetic
establishing the kinetic competency of the pathway depicted in
t
Scheme 2, as opposed to direct activation of H2 by Bu3P/
B(C6F5)3, is nontrivial because production of 1-tBu occurs with
great facility upon mixing of the phosphine and borane. In our
t
hands, it was not possible to generate solutions of Bu3P/
B(C6F5)3 that did not contain at least a small amount (1ꢀ5%)
of 1-tBu.
The situation is further complicated by the possibility that
encounter complexes between 1-tBu and tBu3P or B(C6F5)3 may
also be involved in activating hydrogen, providing alternate
pathways for the generation of the observed product [tBu3-
PH]þ[HB(C6F5)3]ꢀ. Thus, four possible H2 activation steps
may be envisioned, as depicted in Scheme 3. Of these possible
paths, three of the four (aꢀc) represent reactions that are
catalytic in 1-tBu, while the fourth (d) is the direct path. In path
a, activation of H2 by 1-tBu via the encounter-complex path
would yield ion pair 5 as the kinetic product; in the absence of
tBu3P/B(C6F5)3, this would likely disproportionate to 1-tBu/
3-tBu,24 but in the presence of an excess of the Lewis pair, it
would rapidly be converted to product based on the results
shown in Scheme 2. The kinetic product of path b would in fact
be ion pair 4, which would be rapidly converted to product in the
presence of borane (see above), and it would also be expected
that the ion pair [p-(F5C6)2BꢀC6F4ꢀP(H)tBu2]þ[HB(C6-
F5)3]ꢀ (40) formed in path c would also be rapidly converted
to product and/or 3-tBu once formed.
We attempted to probe the relative importance of the
catalyzed paths versus the direct path by a semiquantitative
investigation of the reactions of solutions of 1-tBu (with and
without added phosphine or borane) and the unlinked pair
tBu3P/B(C6F5)3 (prepared in such a way as to minimize produc-
tion of 1-tBu) with H2 under the same conditions. Thus, separate
10028
dx.doi.org/10.1021/ja203214f |J. Am. Chem. Soc. 2011, 133, 10026–10029

Journal of the American Chemical Society
COMMUNICATION
experiments, it appears that direct addition is likely still the
dominant reaction coordinate, but under certain conditions, the
1-tBu catalyzed path may constitute a significant fraction of the
path by which [tBu3PH]þ[HB(C6F5)3]ꢀ is formed. Whether or
not this mechanistic scenario is specific to the tBu3P/B(C6F5)3
Lewis pair remains an open question, and we are currently
investigating the feasibility of this alternate path in other Lewis
base/B(C6F5)3 Lewis pairs.10,26,27 Irrespective of the generality
of this indirect mechanism, the results described here show that
despite the seeming simplicity of the reactions observed, the
mechanistic picture is complex and requires further study.
(19) Chojnowski, J.; Rubinsztajn, S.; Cella, J. A.; Fortuniak, W.;
Cypryk, M.; Kurjata, J.; Kazmierski, K. Organometallics 2005, 24, 6077–
6084.
(20) Welch, G. C.; Cabrera, L.; Chase, P. A.; Hollink, E.; Masuda,
J. D.; Wei, P.; Stephan, D. W. Dalton Trans. 2007, 3407–3414.
(21) Jiang, C.; Blacque, O.; Fox, T.; Berke, H. Dalton Trans. 2011,
40, 1091–1097.
€
(22) Chiu, C.-W.; Kim, Y.; Gabbai, F. P. J. Am. Chem. Soc. 2009, 131,
60–61.
(23) Kim, Y.; Gabbai, F. P. J. Am. Chem. Soc. 2009, 131, 3363–3369.
(24) Indeed, separate syntheses of the linked hydridoborate and
phosphonium ions as their [tBu3PH] and [B(C6F5)4] salts, respectively,
showed that their combination results in immediate disproportionation.
This is a feature of the chemistry of hydrogen addition to compounds 1
that was not anticipated by the theoretical treatment of Guo and Li17 and
may account for the complex kinetic behavior observed for this
reaction.5,7
(25) Rokob, T. A.; Hamza, A.; Pꢀapai, I. J. Am. Chem. Soc. 2009, 131,
10701–10710.
(26) Sumerin, V.; Schulz, F.; Nieger, M.; Leskel€a, M.; Repo, T.;
Rieger, B. Angew. Chem., Int. Ed. 2008, 47, 6001–6003.
(27) Chase, P. A.; Stephan, D. W. Angew. Chem., Int. Ed. 2008, 47,
7433–7437.
’ ASSOCIATED CONTENT
S
Supporting Information. Full experimental details,
b
UVꢀvis spectra, X-ray crystallographic data for 2 (CIF), and
DFT computational output. This material is available free of
charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
wpiers@ucalgary.ca
’ ACKNOWLEDGMENT
This work was supported by the Natural Sciences and
Engineering Research Council (NSERC) of Canada, an NSERC
Graduate Scholarship (L.G.M.), and an Alberta Innovates Stu-
dentship (L.G.M.). The authors thank the reviewers for thought-
ful and constructive comments.
’ REFERENCES
(1) Frey, G. D.; Lavallo, V.; Donnadieu, B.; Schoeller, W. W.;
Bertrand, G. Science 2007, 316, 439–441.
(2) Spikes, G. H.; Fettinger, J. C.; Power, P. P. J. Am. Chem. Soc.
2005, 127, 12232–12233.
(3) Parks, D. J.; Piers, W. E. J. Am. Chem. Soc. 1996, 118, 9440–9441.
(4) Parks, D. J.; Blackwell, J. M.; Piers, W. E. J. Org. Chem. 2000,
65, 3090–3098.
(5) Stephan, D. W.; Erker, G. Angew. Chem., Int. Ed. 2010, 49, 46–76.
(6) Stephan, D. W. Org. Biomol. Chem. 2008, 6, 1535–1539.
(7) Welch, G. C.; Juan, R. R. S.; Masuda, J. D.; Stephan, D. W. Science
2006, 314, 1124–1126.
(8) Spies, P.; Erker, G.; Kehr, G.; Bergander, K.; Fr€ohlich, R.;
Grimme, S.; Stephan, D. W. Chem. Commun. 2007, 5072–5074.
(9) Sumerin, V.; Schulz, F.; Atsumi, M.; Wang, C.; Nieger, M.;
Leskel€a, M.; Repo, T.; Pyykk€o, P.; Rieger, B. J. Am. Chem. Soc. 2008,
130, 14117–14119.
(10) Holschumacher, D.; Bannenberg, T.; Hrib, C. G.; Jones, P. G.;
Tamm, M. Angew. Chem., Int. Ed. 2008, 47, 7428–7432.
(11) Welch, G. C.; Stephan, D. W. J. Am. Chem. Soc. 2007, 129,
1880–1881.
(12) Chase, P. A.; Jurca, T.; Stephan, D. W. Chem. Commun. 2008,
1701–1703.
(13) Blackwell, J. M.; Sonmor, E. R.; Scoccitti, T.; Piers, W. E. Org.
Lett. 2000, 2, 3921–3923.
(14) Rajeev, R.; Sunoj, R. B. Chem.—Eur. J. 2009, 15, 12846–12855.
(15) Grimme, S.; Kruse, H.; Goerigk, L.; Erker, G. Angew. Chem., Int.
Ed. 2010, 49, 1402–1405.
(16) Rokob, T. A.; Hamza, A.; Stirling, A.; Soꢀos, T.; Pꢀapai, I. Angew.
Chem., Int. Ed. 2008, 47, 2435–2438.
(17) Guo, Y.; Li, S. Inorg. Chem. 2008, 47, 6212–6219.
(18) Beringhelli, T.; Maggioni, D.; D’Alfonso, G. Organometallics
2001, 20, 4927–4938.
10029
dx.doi.org/10.1021/ja203214f |J. Am. Chem. Soc. 2011, 133, 10026–10029