Hydrogenolysis of palladium(II) hydroxide, phenoxide, and alkoxide complexes

Article abstract of DOI:10.1021/ja205824q

A series of pincer (^tBuPCP)Pd(II)-OR complexes ( ^tBuPCP = 2,6-bis(CH₂P^tBu₂)C ₆H₃, R = H, CH₃, C₆H₅, CH₂C(CH₃)₃, CH₂CH₂F, CH₂CHF₂, CH₂CF₃) were synthesized to explore the generality of hydrogenolysis reactions of palladium-oxygen bonds. Hydrogenolysis of the Pd hydroxide complex to generate the Pd hydride complex and water was shown to be inhibited by formation of a water-bridged, hydrogen-bonded Pd(II) hydroxide dimer. The Pd alkoxide and aryloxide complexes exhibited more diverse reactivity. Depending on the characteristics of the -OR ligand (steric bulk, electron-donating ability, and/or the presence of β-hydrogen atoms), hydrogenolysis was complicated by hydrolysis by adventitious water, a lack of reactivity with hydrogen, or a competing dissociative β-hydride abstraction reaction pathway. Full selectivity for hydrogenolysis was observed with the partially fluorinated Pd(II) 2-fluoroethoxide complex. The wide range of Pd-OR substrates examined helps to clarify the variety of reaction pathways available to late-transition-metal alkoxides as well as the conditions necessary to tune the reactivity to hydrogenolysis, hydrolysis, or dissociative β-hydride abstraction.

Full text of DOI:10.1021/ja205824q

ARTICLE
pubs.acs.org/JACS
Hydrogenolysis of Palladium(II) Hydroxide, Phenoxide, and Alkoxide
Complexes
Gregory R. Fulmer,† Alexandra N. Herndon,† Werner Kaminsky,† Richard A. Kemp,*,‡,§ and
Karen I. Goldberg*,†
†Department of Chemistry, University of Washington, Box 351700, Seattle, Washington 98195-1700, United States
‡Department of Chemistry and Chemical Biology, University of New Mexico, Albuquerque, New Mexico 87131, United States
§Advanced Materials Laboratory, Sandia National Laboratories, Albuquerque, New Mexico 87106, United States
S Supporting Information
b
ABSTRACT: A series of pincer (tBuPCP)Pd(II)ꢀOR complexes (tBuPCP = 2,6-
bis(CH2PtBu2)C6H3, R = H, CH3, C6H5, CH2C(CH3)3, CH2CH2F, CH2CHF2,
CH2CF3) were synthesized to explore the generality of hydrogenolysis reactions
of palladiumꢀoxygen bonds. Hydrogenolysis of the Pd hydroxide complex to
generate the Pd hydride complex and water was shown to be inhibited by
formation of a water-bridged, hydrogen-bonded Pd(II) hydroxide dimer. The Pd
alkoxide and aryloxide complexes exhibited more diverse reactivity. Depending
on the characteristics of the ꢀOR ligand (steric bulk, electron-donating ability, and/or the presence of β-hydrogen atoms),
hydrogenolysis was complicated by hydrolysis by adventitious water, a lack of reactivity with hydrogen, or a competing dissociative
β-hydride abstraction reaction pathway. Full selectivity for hydrogenolysis was observed with the partially fluorinated Pd(II)
2-fluoroethoxide complex. The wide range of PdꢀOR substrates examined helps to clarify the variety of reaction pathways available
to late-transition-metal alkoxides as well as the conditions necessary to tune the reactivity to hydrogenolysis, hydrolysis, or
dissociative β-hydride abstraction.
Scheme 1
’ INTRODUCTION
Hydrogenation reactions play a very large role in homoge-
neous and heterogeneous metal catalysis, aiding in the produc-
tion of a host of chemicals and pharmaceuticals.1ꢀ3 A key
reaction step in metal-catalyzed hydrogenation cycles is the
formation of a metal hydride complex. This fundamental orga-
nometallic reaction has received extensive study, and as illu-
strated in Scheme 1, this hydride formation can proceed through
a wide variety of mechanisms. Initial interaction of H2 with the
metal center typically proceeds via coordination and formation of
an η2-dihydrogen species.4 Next, the reaction pathway could
proceed via (a) oxidative addition (OA) to generate a metal
dihydride species.1,2 Notably, migratory insertion of a hydride to
a bound olefin, followed by CꢀH reductive elimination of a
hydrocarbon from this higher oxidation state species, would be a
typical pathway for an olefin hydrogenation catalyst. An alter-
native reaction for the η2-dihydrogen species would be (b) σ-
bond metathesis (SBM), resulting in hydrogenolysis of the MꢀY
bond to generate a metal hydride complex along with the
concurrent release of a hydrogenated ligand (HꢀY).5 Deproto-
nation of the dihydrogen moiety by (c) an external base or by (d)
an internal base also results in formation of metal hydride species.
The most prominent examples of pathway d are the well-known
Noyori and Shvo-type catalytic systems for the hydrogenation of
CdO and CdN bonds.6,7 In these systems, the internal base is
part of a chelating or multidentate ligand which prevents the
dissociation of the protonated base.
What happens if the heteroatom ligand (Y) receiving one of
the hydrogens is a monodentate ligand such as a hydroxide,
alkoxide, or amide group? While similar to SBM shown as path b
in Scheme 1, reactions involving these ligand types would be
mechanistically distinct since the lone pair on the ligand could be
involved in the dihydrogen cleavage step, as in path d.8,9
However, the reaction could also proceed via path a with HꢀY
reductive elimination from the dihydride species to yield the
Received: July 3, 2011
Published: September 21, 2011
r
2011 American Chemical Society
17713
dx.doi.org/10.1021/ja205824q J. Am. Chem. Soc. 2011, 133, 17713–17726
|

Journal of the American Chemical Society
ARTICLE
same products. In fact, very little is known about this type of
hydrogenation reaction, which results in hydrogenolysis of a
metalꢀheteroatom bond. Contributing to the difficulty in study-
ing this reaction for metalꢀhydroxide, ꢀalkoxide, and ꢀamide
bonds are the variety of competing reactions available to metal
complexes of these ligands. Hydrolysis and dissociative β-hy-
dride abstraction (DBHA) are two relatively common competi-
tive reactions noted for metal alkoxides and amides.10,11 It is
notable that DBHA leads to generation of the identical metal
hydride formed in hydrogenolysis, albeit without production of
the alcohol or amine. Instead, an unsaturated organic with a
CdO or CdN bond is generated.
Despite the paucity of direct observation and mechanistic
study of hydrogenolysis of late-transition-metalꢀoxygen
bonds, the addition of hydrogen across MꢀOR bonds has
been recognized to play a vital role in a number of important
stoichiometric and catalytic reactions. For example, hetero-
geneous complexes like Pearlman’s and Adams’ catalysts,12,13
heavily used in industrial-scale hydrogenation and hydroge-
nolysis processes, are thought to exhibit exceptional catalytic
activity due to the formation of highly reactive MꢀH bonds
on their surface after treatment of these solid state hydroxides
and “hydrated oxides” with H2. Stryker’s reagent, a copper
hydride complex used in the homogeneous catalysis of con-
jugate reduction reactions, is typically generated in situ by
the hydrogenolysis of CuꢀOtBu bonds.14 Hydrogenation
of metalꢀformate oxygen bonds is also proposed as the pro-
duct release step in carbon dioxide reduction to produce
formic acid.15
It is evident from the examples above that hydrogenolysis of
late-transition-metalꢀoxygen bonds is a fundamental reaction
with important consequences in catalysis. This chemical reaction,
releasing water or alcohol while producing a metalꢀhydride,
constitutes a powerful combination product release and catalyst
regeneration step in a catalytic cycle.14a,16,17 Many applications
of this reaction can be envisioned in the design of new catalytic
transformations. Hydrogenation of metal hydroxides generated
after oxygen transfer to substrates from metal hydroperoxide
species (formed by oxygen insertion into metalꢀhydride bonds)
would allow for oxygenase reactions using molecular oxygen as
the oxidant. Catalytic hydrogenation of aldehydes and ketones
could similarly involve hydrogenolysis of metalꢀalkoxide bonds.
To fully exploit the unique reactivity of metalꢀoxygen bonds,
however, investigations of mechanism and the scope of the
hydrogenolysis reaction are needed. In particular, a mechanistic
understanding is essential to developing the ability to select and
tune between the competitive reaction pathways (e.g., hydro-
genolysis, DBHA) available to metal alkoxides for the effective
use of these complexes in catalysis.
and electronic considerations of the alkoxide groups are deli-
neated. Kinetic and thermodynamic factors are also addressed.
Similar issues of competitive, alternative pathways are likely to be
encountered for hydrogenolysis reactions of other metalꢀ
heteroatom bonds (e.g., MꢀN bonds). Thus, results from this
study provide broad insight into the application of this reaction step
in the design of new transition-metal-catalyzed processes.
’ RESULTS AND DISCUSSION
(
tBuPCP)PdOH. When C6D6 solutions of the palladium(II)
hydroxide complex (tBuPCP)PdOH (1)18 were exposed to
pressures of hydrogen gas (7.0 atm), room temperature conver-
sion to (tBuPCP)PdH (2) and water was observed over a period
of days.9 The reaction progress was easily monitored by 1H and
31P NMR spectroscopy, and rates were determined by following
the disappearance of the methylene signal of 1 as compared to
the methyl signal of the internal standard hexamethylbenzene
(s, δ = 2.1319).
Notably, a few specific resonances in the 1H NMR spectrum
for complex 1 shifted upfield during the reaction. This shift can
be attributed to the increase in the water concentration due to
the formation of water as a product during the reaction. The
same upfield shift for 1 was observed in wet C6D6 solutions of
1 and was attributed to the interaction between complex 1 and
H2O.20 Kinetic experiments with different pressures of hydro-
gen (3.5ꢀ7.0 atm) were performed, and in all cases, it was found
that the reaction did not follow first-order kinetics in [1]. As
shown by the example in Figure 1a, plots of ln[1] against time do
not exhibit linear behavior. In contrast, when the data are plotted
Preliminary results on the hydrogenolysis of the pincer
palladium(II) hydroxide complex (tBuPCP)PdOH (1, tBuPCP =
2,6-bis(CH2PtBu2)C6H3) to generate the palladium(II) hydride
complex and water were recently reported by us.9 In this con-
tribution, a broad study investigating the hydrogenolysis of
palladiumꢀoxygen bonds is presented using the tBuPCP pincer
framework. The scope of the reaction with respect to hydro-
genolysis of various palladium alkoxide and aryloxide linkages is
explored. Attemped hydrogenolysis of these compounds to
generate alcohol and the corresponding palladium(II) hydride
complex (tBuPCP)PdH (2) are described, along with our ob-
servations of competitive alternative reaction pathways. The
limitations of the hydrogenolysis reaction with respect to steric
Figure 1. (a) Pseudo-first-order and (b) pseudo-half-order rate plots
for the reaction of 1 and H2 (7.0 atm) in anhydrous C6D6 at 25 °C.
17714
dx.doi.org/10.1021/ja205824q |J. Am. Chem. Soc. 2011, 133, 17713–17726

Journal of the American Chemical Society
ARTICLE
Scheme 2
Figure 2. Linear relationship between [1]1/2 and time. Kinetic plots for
reactions of 1 with 3.5 atm (2) and 7.0 atm (9) H2 in wet C6D6 at 25 °C
are shown ([1]0 = 4.2 mM, [H2O]0 = 38 mM).
Figure 4. ORTEP of complex [(tBuPCP)PdOH]2•4H2O. Ellipsoids are
shown at 50% probability, and hydrogen atoms are omitted for clarity.
dependence of the rate on [H2].22 The empirical rate law for the
reaction can then be written as ꢀd[1]/dt = kobs[H2][1]1/2
.
A mechanism involving a dimerꢀmonomer pre-equilibrium prior
to rate-determining hydrogenolysis of the monomer (Scheme 2)
would be consistent with the unusual half-order dependence
on 1.23 X-ray structural determinations of the hydroxide complex
1 provide further support. While crystals of 1 grown under rig-
orously anhydrous conditions afford the monomeric structure
shown in Figure 3, crystals of 1 grown in the presence of H2O
revealed a water-bridged dimeric complex [(tBuPCP)PdOH]2•
4H2O (Figure 4).9 Other PCP-type palladium(II) hydroxides
have also been observed to dimerize through bridging water mol-
ecules in the solid state.20,24 Selected bond lengths and angles for both
the monomer and dimer are detailed in Table 1. The structural
parameters are similar for both forms, with notable elongation of the
PdꢀO bond from 2.066(3) to 2.094(3) Å when the hydroxide ligand
hydrogen-bonds with the bridging water molecules.
The empirical rate law can then be explained by the mechan-
ism shown in eqs 1ꢀ2. Hydrogenolysis results from the reaction
of the monomeric form of 1 with hydrogen. Formation of the
dimeric species inhibits the reaction by removing the reactive
monomer from solution. The participation of water in forming
the dimer explains the slower rate of reaction observed in the
presence of excess water. In addition, this would be consistent
with the irreproducibility of the rate constants when excess water
was not deliberately added. In this case, differing amounts of
adventitious water yield different rates of reaction.
Figure 3. ORTEP of complex (tBuPCP)PdOH (1). Ellipsoids are
shown at 50% probability, and hydrogen atoms are omitted for clarity.
of [1]1/2 versus time, excellent correlation to a linear fit is
observed (Figure 1b).
Each individual reaction of 1 with H2 was found to obey half-
order kinetics with respect to [1] (similar to the example data set
shown in Figure 1b). Initially though, the rate constants calcu-
lated from the slopes of different kinetic experiments conducted
at a given pressure of H2 were not consistent. However, when
excess water (9 equiv)21 was deliberately added to the reaction
solutions, the observed rate constants were highly reproducible,
albeit smaller than in the case without added water. At ambient tem-
perature in the presence of excess H2O, doubling the pressure of H2
from 3.5 to 7.0 atm effected a doubling of the observed rate
k1
2PdOH þ nH2O h ½PdOHꢂ2 nH2O
ð1Þ
ð2Þ
3
k
ꢀ1
k2
f
PdOH þ H2
s
PdH þ H O
2
constant (Figure 2; kobs = 2.0(1) ꢁ 10ꢀ7 Mꢀ1/2 sꢀ1 to kobs
=
Assuming that the palladium(II) hydroxide signals observed
4.3(3) ꢁ 10ꢀ7 Mꢀ1/2 sꢀ1, respectively). This indicates a first-order
in the 1H NMR spectra were an average of the signals for the
17715
dx.doi.org/10.1021/ja205824q |J. Am. Chem. Soc. 2011, 133, 17713–17726

Journal of the American Chemical Society
ARTICLE
Scheme 3
as intermediates in PdII mediated reactions.27 The second
proposed mechanism (Scheme 3, path b) involves a four-center
transition state wherein a proton is transferred intramolecularly
from a coordinated dihydrogen to the oxygen.28,29 Both of these
pathways could proceed via the initial coordination of H2 to the
16 electron palladium center. Notably, as a cationic dihydrogen
complex of the tBuPCP platinum system is known,30 a third
mechanism (Scheme 3, path c) would be the dissociation of the
OHꢀ ligand, H2 coordination, and deprotonation by the result-
ing OHꢀ to form water and the PdIIꢀH. However, if hydroxide
dissociation were involved as a preliminary step, the addition of
excess water to the reaction would be expected to increase,31
rather than decrease, the rate of reaction as was confirmed
experimentally.
In the absence of observable intermediates, it is challenging to
experimentally distinguish between the remaining two pathways.
Computational methods (B3LYP density functional theory32)
were then employed to compare the energetics of the pathways.
As illustrated in Figure 5 for a model PCP complex, the four-
center transition state (B) was found to be significantly lower in
energy than the oxidative addition intermediate (A). Notably the
involvement of the oxygen lone pair in the four-center transition
state distinguishes path B from a true σ-bond metathesis, and this
type of intramolecular proton transfer has been termed internal
electrophilic substitution (IES).29
Table 1. Selected Bond Lengths and Angles for Complexes 1
and [(tBuPCP)PdOH]2•4H2O
1
[(tBuPCP)PdOH]2•4H2O
bond length (Å)
Pd1ꢀP1
2.2864(11)
2.2814(11)
2.022(4)
2.3027(13)
2.2991(13)
2.024(4)
2.094(3)
2.813(4)
3.041(4)
Pd1ꢀP2
Pd1ꢀC1
Pd1ꢀO1
2.066(3)
O1•••O2
O1•••O20
bond angle (deg)
P1ꢀPd1ꢀP2
C1ꢀPd1ꢀO1
C1ꢀPd1ꢀP1
C1ꢀPd1ꢀP2
O1ꢀO2ꢀO10
O1ꢀO20ꢀO10
166.57(4)
177.30(16)
83.42(12)
83.16(12)
167.06(5)
176.68(17)
83.73(14)
83.63(14)
100.01(22)
100.01(22)
monomer and for the dimer in a fast equilibrium, the total
palladium(II) hydroxide concentration can be described as
[PdOH]tot = 2[PdOH]2 + [PdOH]. The rate expression for
the hydrogenolysis reaction depicted in Scheme 2 and eqs 1ꢀ2
was derived in terms of [PdOH]tot (eq 3), with the composite
Our experimental and computational studies of the hydro-
genolysis reaction of 1 are consistent with the concerted reaction
of hydrogen with a monomeric d8 metal center and the lone pair
of the hydroxide oxygen. Similar reactivity should be expected for
monomeric d8 metal alkoxides and aryloxides with hydrogen. In
fact, one might anticipate that hydrogenolysis of metal alkoxide
or aryloxide species should be more straightforward as the
tendency to dimerize through bridging water and/or alcohol
molecules would be less significant. Both the sterics of the
alkoxide or aryloxide substituent and the lack of a hydrogen
bound to oxygen should inhibit hydrogen-bonded dimerization.
In a broader context, the study of hydrogenolysis of the alkoxide
and aryloxide analogues enables us to investigate the generality of
hydrogenolysis reactions with respect to a range of MꢀO bonds.
A greater understanding of the scope of this reaction will allow
for its successful application in catalysis.
constants k = k2/4K0eq and K0eq = k1[H2O]n/k
.
ꢀ1
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Rate ¼ k½H2ꢂð 1 þ 8K0eq½PdOHꢂtot ꢀ 1Þ
ð3Þ
Since eq 3 does not have an easily handled analytical solution, the
kinetic data were fitted by numerical integration using the
RungeꢀKutta fourth-order method.25 By this method, K0eq for
dimer formation (eq 1) was determined to be 1.0 ((0.1) ꢁ 103
M
ꢀ1 and k2, the rate constant for the reaction of the monomeric
Pd hydroxide with H2 (eq 2), was 1.5 ((0.3) ꢁ 10ꢀ3 Mꢀ1 sꢀ1 26
.
Several mechanisms for the key hydrogenolysis step(s) invol-
ving the monomeric palladium complex 1 to form 2 and H2O can
be considered. One mechanism (Scheme 3, path a) is oxida-
tive addition of H2 to the PdIIꢀOH complex to form an
octahedral PdIV intermediate, followed by reductive elimina-
tion of H2O to form 2. Notably, PdIV complexes have been
reported and significant evidence provided for their involvement
(
tBuPCP)PdOCH3. The first palladium(II) alkoxide species
targeted was the methoxide complex (tBuPCP)PdOCH3 (3).9
It was quickly discovered, however, that complex 3 is extremely
17716
dx.doi.org/10.1021/ja205824q |J. Am. Chem. Soc. 2011, 133, 17713–17726

Journal of the American Chemical Society
ARTICLE
Scheme 5. Two Possible Paths for the Conversion of
(
tBuPCP)PdOR and H2 to (tBuPCP)PdH and ROH (R ¼ H)
Figure 5. Energies (ΔH [ΔG]) are in kcal molꢀ1 for gas phase species.
A corresponds to the PdIV intermediate for the oxidative addition/
reductive elimination pathway, and B to the transition state for the four-
center intramolecular proton transfer. PCP = MePCP (2,6-bis(CH2P-
(CH3)2)C6H3).
maintain anhydrous conditions, a small amount of the hydroxide
complex 1 was always observed in the hydrogenolysis reactions
of palladium methoxide complex 3, indicative of the presence of
adventitious water. Furthermore, the hydroxide complex 1, the
hydrolysis product of 3, will also undergo hydrogenolysis to form
the palladium hydride complex 2.
Scheme 4
Avoiding the hydrolysis reaction is critical to achieving repro-
ducible kinetics for the hydrogenolysis of complex 3. As illu-
strated in Scheme 5, the presence of H2O during a hydro-
genolysis reaction actually leads to the production of the same
products through two competing pathways a and b. Acting as a
catalyst for path b, H2O converts the alkoxide complex to the
hydroxide complex, which then reacts with hydrogen to regen-
erate water. Both pathways have indistinguishable products, as
both yield alcohol and complex 2. Yet, the order of the reaction
with respect to [Pd] would be different for the two mechanisms
(i.e., first- and half-order in [Pd] for path a and b, respectively). In
addition, the rate constant for the hydrogenolysis of the mono-
meric palladium hydroxide and the palladium methoxide is also
likely to be different. Thus, if path b is active to different extents
based on adventitious water, obtaining reproducible kinetics
would be challenging.
sensitive to adventitious water. Notably, dimerization was not the
culprit but rather hydrolysis. Upon deliberate exposure of 3
(19 mM) in benzene-d6 to water (8.0 mM), rapid conversion to
the hydroxide complex 1 and the release of methanol were clearly
1
observed by H and 31P NMR spectroscopy. To probe the
equilibrium between these two palladium species, a benzene-d6
solution of the methoxide complex 3 was treated with 0.5 equiv of
H2O, and the change in concentration between complexes 3 and
1
1 was monitored by H NMR spectroscopy over the course of
Adding excess methanol to the methoxide system should shift
the equilibrium in Scheme 4 in favor of the methoxide complex 3.
However, the addition of methanol resulted in promotion of a
different type of reaction. As illustrated in Scheme 6, the presence
of CH3OH assists in the conversion of complex 3 to the hydride
complex 2 through an alcohol-promoted dissociative β-hydride
abstraction (DBHA) mechanism. When 10 equiv methanol
were added to C6D6 solutions of 3 in the absence of H2, complete
conversion to the hydride complex 2 at room temperature was
observed over 3 days. DBHA has been reported for other late-
transition-metal systems11,35,36 and is distinct from the tradi-
tional β-hydride elimination of metal alkyls. The DBHA pathway
proceeds through the heterolytic dissociation of the alkoxide
anion, producing a (tBuPCP)Pd cation. Hydride abstraction from
the alkoxide anion leads to 2 and formaldehyde. The addition of
methanol accelerates this reaction path by stabilizing the alkoxide
anion through hydrogen bonding and thus assisting in the
alkoxide dissociation.
5 h. Using eq 4 and the concentrations of 1, 3, methanol, and
water33 calculated against an internal hexamethylbenzene stan-
dard, an equilibrium constant (Keq) of 9.0 ((0.3) was obtained
(Scheme 4). Thus, when 1 equiv of methanol was added to a
benzene-d6 solution of the hydroxide complex 1, only a small
percentage of complex 3 (<10%) was observed by NMR
spectroscopy. Clearly, the reaction of 3 with H2O favors the
production of 1 and methanol. This result is consistent with
other examples in the literature where a late-transition-metal
alkoxide complex undergoes hydrolysis to generate a hydroxide
ligand and the corresponding alcohol.10,34,35
Keq ¼ ½PdOHꢂ½CH3OHꢂ=½PdOCH3ꢂ½H2Oꢂ
ð4Þ
Analogous to the hydrogenolysis reaction observed with the
hydroxide complex 1, exposure of the methoxide complex 3 to
dihydrogen (7.0 atm) in C6D6 yields the palladium(II) hydride
complex 2 and methanol over a period of 80 h at room
Methanol-free solutions of 3 in benzene-d6 are relatively stable
at ambient temperatures in the absence of H2. Only a small
percentage (<10%) of 3 was observed to decompose over the
course of 1 week at room temperature. A similar amount of the
hydride complex 2 was formed during this period, yet no
1
temperature in 96% yield, as determined by H and 31P{1H}
NMR spectroscopy.9 However, the rate for hydrogenolysis of 3
was not reproducible. Given the results described above for the
hydrogenolysis reaction of the Pd hydroxide complex 1, it is
reasonable to suggest that adventitious H2O is responsible for the
irreproducibility. Notably, even when extreme care was taken to
1
methanol production was detected by H NMR spectroscopy.
At 65 °C, however, complete conversion of 3 to 2 (80%) and a
17717
dx.doi.org/10.1021/ja205824q |J. Am. Chem. Soc. 2011, 133, 17713–17726

Journal of the American Chemical Society
ARTICLE
Scheme 6
Scheme 7
Scheme 8
new minor species (20%) was observed in 15 h. The minor
species has a very similar 1H NMR spectrum to complex 3 with
an additional resonance at 5.72 ppm that integrates to two
protons. In analogy to reactivity observed in a Pt(IV)ꢀOCH3
system35 and consistent with formaldehyde as a product in the
DBHA reaction of complex 3, this new species is proposed to be
the OCH2 insertion product (tBuPCP)PdOCH2OCH3 (4). To
confirm the identity of complex 4 as well as to establish the
probable pathway for its production, the palladium(II) methox-
ide complex 3 was allowed to react with para-formaldehyde (10
equiv) to provide an independent synthesis of 4 (Scheme 7).
Within minutes of dissolving the two reagents together in C6D6
at room temperature, partial conversion of 3 to 4 was observed by
NMR spectroscopy. The complete disappearance of 3 was
observed after 30 min, and the presence of complex 4 along
with other Pd species was detected by NMR spectroscopy.
Similar to the Pt system,35 the additional products are most
likely the result of multiple OCH2 insertions.
complex 3 in an attempt to suppress hydrolysis instead resulted
in the promotion of the DBHA reaction pathway (Scheme 8,
path b). While the hydrogenolysis of complex 3 was observed to
occur (Scheme 8, path c), the competing hydrolysis and DBHA
reactions, which ultimately also result in the formation of the same
hydride complex 2, made kinetic studies and a valid comparison
to the hydroxide system impractical. To investigate hydrogeno-
lysis reactions of palladium alkoxides in greater detail, PdOR
complexes resistant to hydrolysis and to DBHA reactivity are
needed. Alkoxide ligands that either lack β-hydrogens or are too
bulky to undergo the rearrangement necessary for the DBHA
mechanism should be considered as viable candidates.
(
tBuPCP)PdOC6H5. As a phenoxide group lacks β-hydrogens
necessary for a competitive DBHA pathway, the palladium(II)
phenoxide complex (tBuPCP)PdOC6H5 (5) was prepared. Complex
5 was synthesized by a procedure analogous to that of 3, in
which the nitrate complex, (tBuPCP)PdONO2, was mixed with
potassium phenoxide in THF (Scheme 9). The phenoxide
complex was characterized by NMR spectroscopy, elemental
analysis, and X-ray crystallography. The ORTEP of 5 is shown
in Figure 6, exhibiting the square planar geometry expected for
a palladium(II) species, and selected bond lengths and angles
are presented in Table 2. Of note, the PdꢀO bond for 5
(2.0900(12) Å) is just slightly longer than that in the methoxide
complex 3 (2.084(3) Å).9
The seemingly subtle modification from PdꢀOH to ꢀOCH3
effects a significant change in reactivity between the two metal
complexes. While the addition of a methoxide moiety adds steric
encumbrance that can be expected to inhibit dimerization, the
new palladium(II) species 3 is also extremely sensitive to
hydrolysis, leading to formation of the hydroxide complex 1
(Scheme 8, path a). The addition of methanol to solutions of
17718
dx.doi.org/10.1021/ja205824q |J. Am. Chem. Soc. 2011, 133, 17713–17726

Journal of the American Chemical Society
ARTICLE
Scheme 9
Figure 7. ORTEP of dimeric complex [(μ-tBuPCHP)Pd]2 (6). Ellip-
soids are shown at 50% probability, and hydrogen atoms are omitted for
clarity. Selected bond lengths (Å) and angles (deg): Pd1ꢀP1, 2.271(2);
Pd1ꢀP2, 2.265(3); P1ꢀPd1ꢀP2, 171.78(10).
Scheme 10
Figure 6. ORTEP of complex (tBuPCP)PdOC6H5 (5). Ellipsoids are
shown at 50% probability, and hydrogen atoms are omitted for clarity.
Table 2. Selected Bond Lengths and Angles for Complex 5
bond length
Å
bond angle
deg
hydride complex 2 with phenol to liberate hydrogen and generate
5 is thermodynamically favored. Given that phenols are con-
siderably more acidic (pKa = 7ꢀ10) than alcohols (pKa =
16ꢀ20),38 the thermodynamics of hydrogenolysis may be more
favorable for alcohols. A palladium neopentoxide complex could
provide for a favorable hydrogenolysis, and although the neo-
pentoxide ligand contains β-hydrogens, its steric bulk should
make it difficult to align its β-hydrogens for DBHA. Therefore, as
shown in Scheme 10, the palladium(II) neopentoxide complex
Pd1ꢀP1
Pd1ꢀP2
Pd1ꢀC1
Pd1ꢀO1
O1ꢀC25
2.3161(4)
2.2931(4)
2.0061(15)
2.0900(12)
1.3110(19)
P1ꢀPd1ꢀP2
C1ꢀPd1ꢀO1
C1ꢀPd1ꢀP1
C1ꢀPd1ꢀP2
Pd1ꢀO1ꢀC25
164.605(15)
163.94(6)
83.66(5)
97.82(4)
137.20(11)
Notably when an excess of H2O (280 mM) was added to a
benzene-d6 solution of the phenoxide complex 5 (6.5 mM) and
heated to 125 °C for 1 day, no reaction was detected by NMR
spectroscopy. The lack of reactivity with water indicates that
adventitious water is unlikely to affect the hydrogenolysis reac-
tion. However, when 5 was treated with H2 gas (7.0 atm) at
room temperature, the conversion to the hydride complex 2
and free phenol was observed to only reach 20% completion
after 2 weeks. The reaction was monitored for an additional
week, but the conversion of 5 to 2 appeared to halt at 20%. It
was suspected that the reaction had reached equilibrium.
Consistent with this proposal, when a C6D6 solution of the
palladium(II) hydride complex 2 (20 mM) was exposed to
phenol (40 mM), rapid bubbling occurred and the conversion
(
tBuPCP)PdOCH2C(CH3)3 (7) was prepared from the Pd(II)
nitrate and potassium neopentoxide. Complex 7 was fully
characterized by NMR spectroscopy, elemental analysis, and
X-ray crystallography. The ORTEP is shown in Figure 8, and
selected bond distances and angles are reported in Table 3. Like
complex 5, a typical square planar geometry is observed in the
solid state for the Pd(II) complex 7. Additionally, the PdꢀO
bond length for 7 (2.081(2) Å) is similar to that in the methoxide
complex 3(2.084(3) Å)9 and phenoxide complex 5(2.0900(12) Å)
and slightly longer than that in the hydroxide complex 1
(2.066(3) Å)
The palladium(II) neopentoxide complex did not undergo
DBHA even when solutions of 7 were heated to 150 °C in the
presence of excess neopentyl alcohol. Surprisingly, however, there
was no reaction of complex 7 with dihydrogen (7.0 atm), even at
temperatures exceeding 100 °C. This result was unexpected,
especially since the PdꢀO bond distances for 7 and the methoxide
complex 3 are so similar. As the electronics of the two alkoxides
are likely to be similar, the lack of reactivity may be attributed to
the increase in steric bulk of the neopentoxide group. Examina-
tion of the space-filling diagram39 reveals limited access to the
palladium and oxygen atoms due to the large size of the tert-
butylphosphino groups combined with that of the neopentoxide
of 2 to 5 was observed by H and 31P NMR spectroscopies.
1
Over time, faint-yellow crystals grew out of solution, which were
characterized by X-ray crystallography as the dimeric Pd(0)
complex [(μ-tBuPCHP)Pd]2 (6, Figure 7). Complex 6 contains
the bridging ligand α,α0-bis(di-tert-butylphosphino)-m-xylene
(
tBuPCHP), and precedent for this type of alcohol-promoted
reduction of Pd(II) complexes has been observed in the analo-
gous iPrPCP system.37
(
tBuPCP)PdOCH2C(CH3)3. As demonstrated above, the phen-
oxide complex (tBuPCP)PdOC6H5 (5) does not undergo hydro-
genolysis but instead the reverse reaction. Protonation of the
17719
dx.doi.org/10.1021/ja205824q |J. Am. Chem. Soc. 2011, 133, 17713–17726

Journal of the American Chemical Society
ARTICLE
Scheme 11
Scheme 12
Figure 8. ORTEP of complex (tBuPCP)PdOCH2C(CH3)3 (7). Ellip-
soids are shown at 50% probability, and hydrogen atoms are omitted for
clarity.
as a quartet at 4.65 ppm (3JHF = 9.9 Hz), and the 19F NMR
spectrum contains a triplet at ꢀ73.3 ppm (3JFH = 9.9 Hz).
Similarly, treatment of a C6D6 solution of 2 (15.0 mM) with DFE
(198 mM) resulted in the formation of 9 and hydrogen. The
methyl proton on the 2,2-difluoroethoxide ligand (CHF2) of 9
Table 3. Selected Bond Lengths and Angles for Complex 7
bond length
Å
bond angle
deg
Pd1ꢀP1
Pd1ꢀP2
Pd1ꢀC1
Pd1ꢀO1
O1ꢀC25
2.3247(8)
2.2853(8)
2.023(3)
2.081(2)
1.383(3)
P1ꢀPd1ꢀP2
C1ꢀPd1ꢀO1
C1ꢀPd1ꢀP1
C1ꢀPd1ꢀP2
Pd1ꢀO1ꢀC25
165.17(3)
171.99(10)
83.02(9)
appears as a characteristic triplet of triplets at 6.07 ppm (2JHF
=
3
58.0 Hz, JHH = 4.2 Hz), and the methylene protons
(CH2CHF2) are observed as a triplet of doublets at 4.51 ppm -
(3JHF = 15.2 Hz, 3JHH = 4.2 Hz) in the 1H NMR spectrum. In the
19F NMR spectrum, complex 9 displays a doublet of triplets at
82.20(9)
123.99(18)
ligand. As discussed above, involvement of both the Pd center
and the lone pair on the oxygen was implicated for the hydro-
genolysis reaction of the palladium(II) hydroxide complex 1.
ꢀ124.5 ppm (2JFH = 58.0 Hz, JFH = 15.2 Hz). Of note, the
3
conversion of 2 to 9 with DFE was significantly slower than the
reaction of 2 and TFE, requiring several hours to reach comple-
tion as compared to seconds for formation of 8 (Scheme 12).
The complete conversion of these reactions of 2 with TFE or
DFEtoproducethecorrespondingpalladiumalkoxideindicates the
thermodynamics of hydrogenolysis of the palladium alkoxide are
unfavorable, similar to the situation observed with the palladium
phenoxide complex.
(
tBuPCP)PdOCH2CF3 and (tBuPCP)PdOCH2CHF2. Palladium
complexes with fluorinated ethoxide ligands should have less
steric bulk than the neopentoxide analog, and the terminal
electron-withdrawing groups (EWG) should deter a potential
DBHA decomposition pathway. The DBHA pathway would be
disfavored from the metal alkoxide complex with analogy to the
equilibrium behavior between aldehydes and their corresponding
hydrates (Scheme 11a). The addition of terminal EWG to alde-
hydes causes a shift in equilibrium toward sp3 hybridized hy-
drates over sp2 hybridized aldehydes.40 The same trend should
apply to sp3 hybridized alkoxide ligands and sp2 hybridized
aldehydes (Scheme 11b). An alkoxide ligand containing term-
inal electron-withdrawing fluorine atoms should therefore be
more resistant to DBHA (i.e., conversion to sp2 hybridized
aldehyde products).
(
tBuPCP)PdOCH2CH2F. The trend of decreasing acidity with
the decrease in the number of terminal fluorine atoms on the
series of fluoroethanols prompted the investigation of the re-
activity of 2-fluoroethanol (FE, pKa = 14.4)42 with the palladium
hydride complex. When an excess of FE (213 mM) was exposed
to a C6D6 solution of complex 2 (15.0 mM) at room tempera-
ture, no conversion to (tBuPCP)PdOCH2CH2F (10) was ob-
served over the course of 2 weeks.
The palladium(II) 2,2,2-trifluoroethoxide complex (tBuPCP)-
PdOCH2CF3 (8) and the palladium(II) 2,2-difluoroethoxide
complex (tBuPCP)PdOCH2CHF2 (9) were prepared. Notably,
however, 2,2,2-trifluoroethanol (TFE, pKa = 12.4) and 2,2-
difluoroethanol (DFE, pKa = 13.3) are fairly acidic.41 Thus,
reaction of the hydride complex 2 with TFE or DFE to release H2
and produce the palladium alkoxide, analogous to the reaction
observed for 2 with phenol, might be expected. Indeed, the
reaction of a C6D6 solution of 2 (15.0 mM) with an excess of TFE
(175 mM) resulted in the complete conversion to 8 and
hydrogen within seconds (Scheme 12), as observed by NMR
spectroscopy. The methylene resonance for the 2,2,2-trifluoro-
ethoxide ligand in complex 8 appears in the 1H NMR spectrum
An independent synthesis of complex 10 was next attempted
following the analogous preparations for complexes 3, 5, and 7.
Unfortunately, the potassium reagent KOCH2CH2F could not
be prepared through the reaction of FE and potassium metal.
Instead, a highly viscous mixture of several products was
obtained. Therefore, an alternate route involving the displace-
ment of the neopentoxide ligand of complex 7 with FE was used
(Scheme 13). This reaction yielded complex 10 and an equiva-
lent of neopentyl alcohol (NpOH). The NpOH was then
removed under reduced pressure, and complex 10 was isolated
and characterized by NMR spectroscopy. The 19F NMR spec-
trum for complex 10 contains a triplet of triplets at ꢀ218.6
3
1
ppm (2JFH = 48.7 Hz, JFH = 21.3 Hz), and the H NMR
17720
dx.doi.org/10.1021/ja205824q |J. Am. Chem. Soc. 2011, 133, 17713–17726

Journal of the American Chemical Society
ARTICLE
Scheme 13
Scheme 14
Figure 9. ORTEP of complex (tBuPCP)PdOH•HOCH2CH2F (1•FE).
Ellipsoids are shown at 50% probability, and hydrogen atoms are
omitted for clarity.
Figure 10. ORTEP of complex (tBuPCP)PdOFE (10). Ellipsoids are
shown at 50% probability, and hydrogen atoms are omitted for clarity.
Table 4. Selected Bond Lengths and Angles for Complex
1•FE
Table 5. Selected Bond Lengths and Angles for Complex 10
bond length
Å
bond angle
deg
bond length
Å
bond angle
deg
Pd1ꢀP1
Pd1ꢀP2
Pd1ꢀC1
Pd1ꢀO1
O1ꢀC25
C26ꢀF1
2.2996(8)
2.2965(8)
2.021(3)
2.082(2)
1.361(4)
1.375(4)
P1ꢀPd1ꢀP2
C1ꢀPd1ꢀO1
C1ꢀPd1ꢀP1
C1ꢀPd1ꢀP2
Pd1ꢀO1ꢀC25
C25ꢀC26ꢀF1
166.45(3)
172.38(11)
83.01(8)
83.44(8)
129.2(2)
113.3(3)
Pd1ꢀP1
Pd1ꢀP2
Pd1ꢀC1
Pd1ꢀO1
O1•••O2
O1•••F1
2.285(2)
2.280(2)
1.996(8)
2.090(6)
2.602(11)
3.011(11)
P1ꢀPd1ꢀP2
C1ꢀPd1ꢀO1
C1ꢀPd1ꢀP1
C1ꢀPd1ꢀP2
Pd1ꢀO1•••O2
Pd1ꢀO1•••F1
166.61(9)
179.1(4)
84.0(2)
83.6(2)
157.89(13)
141.76(13)
illustrated in Figure 10, the ORTEP of complex 10 exhibits the
same square planar geometry observed for its palladium alkoxide
analogues, and the Pd1ꢀO1 bond length is also comparable at
2.082(2) Å (Table 5).
resonance for the ꢀCH2F protons appears at 4.78 ppm as a
doublet of triplets (2JHF = 48.7 Hz, 3JHH = 5.7 Hz).
Highlighting the hygroscopic nature of these alkoxide com-
plexes, an attempt to grow X-ray quality crystals of (tBuPCP)-
PdOCH2CH2F by slow diffusion of pentane into a concentrated
benzene solution of 10 instead resulted in the hydroxide complex
1 with an associated fluoroethanol (tBuPCP)PdOH•HOCH2CH2F
(1•FE, Figure 9). Examination of the crystallographic data in
Table 4 reveals that the PdꢀO bond length in complex 1•FE is
significantly elongated (2.090(6) Å) as compared to complex 1
(2.066(3) Å), indicative of electron donation from the hydroxide
ligand to FE through hydrogen bonding. Presumably, an equiva-
lent of H2O was introduced during crystallization, displacing the
2-fluoroethoxide ligand of complex 10 (Scheme 14), thus con-
verting the alkoxide complex to 1 and FE. This chemistry is very
similar to that described above for the methoxide complex 3 in
the presence of water where conversion to the hydroxide complex
1 and methanol was favored.
In the absence of water, solutions of complex 10 in C6D6 are
extremely stable, even in the presence of excess FE. Notably, the
1H NMR resonances for 10 do not shift in varying concentrations
of FE, suggesting that no significant hydrogen bonding between
complex 10 and FE is occurring in solution. As a control, an
NMR tube was prepared with a C6D6 solution of 10 and 9 equiv
of FE. In the absence of H2, the sample was heated to 60 °C in a
1
temperature-controlled oil bath for 7 days. By H NMR spec-
troscopy, < 5% conversion of complex 10 to the hydride complex
2 through DBHA was observed. Notably, virtually complete
conversion of the analogous methoxide complex 3 to 2 by DBHA
occurs in 15 h under similar conditions. Thus 10 shows a
significantly reduced proclivity for the DBHA pathway as com-
pared with complex 3. The stability of 10 toward DBHA and the
lack of interaction with FE allows for kinetic studies of the
hydrogenolysis reaction in the presence of excess FE, which is
needed to prevent hydrolysis by adventitious water (see Scheme 5).
In a separate experiment being extremely careful to avoid H2O
contamination, complex 10 was successfully isolated and char-
acterized by elemental analysis and X-ray crystallography. As
17721
dx.doi.org/10.1021/ja205824q |J. Am. Chem. Soc. 2011, 133, 17713–17726

Journal of the American Chemical Society
ARTICLE
Scheme 15
Figure 11. Partial 1H NMR spectra for the reaction of 10 (3.6 μmol)
and H2 (7.0 atm) in C6D6 with excess FE (32 μmol) at 60 °C over a
period of 24 h. The disappearance of the ꢀCH2CH2F signal for 10
(doublet of triplets, ca. 4.78 ppm) is observed along with the appearance
of the PdꢀH signal for 2 (virtual triplet, ca. ꢀ3.72 ppm).
The addition of H2 (7.0 atm) to a benzene-d6 solution of the
palladium 2-fluoroethoxide complex 10 at room temperature in
the presence of excess FE (9 equiv) resulted in the quantitative
conversion to the palladium hydride complex 2 and FE over the
course of nearly 1 month. At 60 °C, however, the reaction reaches
completion within 24 h (Scheme 15).
Figure 12. (a) Pseudo-first-order and (b) pseudo-half-order rate plots
for the reaction of 10 (3.6 μmol) and H2 (7.0 atm) in C6D6 with excess
FE (32 μmol) at 60 °C.
The reaction progress at 60 °C was monitored by NMR
spectroscopy; in the 1H NMR spectrum, the disappearance of the
ꢀCH2F resonance for 10 was observed along with the appear-
ance of the PdꢀH resonance for 2 (Figure 11). It was found that
each individual reaction of 10 with excess H2 (in the presence of
excess FE) at 60 °C obeyed a pseudo-first-order rate dependency
with respect to [10], exhibiting an excellent linear fit to the data
(Figure 12a). By contrast, plots of [10]1/2 against time are clearly
not linear (Figure 12b).43 Doubling the partial pressure of H2
from 3.5 to 7.0 atm resulted in a doubling of the observed rate
constant (1.5 ((0.4) ꢁ 10ꢀ5 sꢀ1 to 3.4 ((0.2) ꢁ 10ꢀ5 sꢀ1
,
respectively), indicating a first-order rate dependence in [H2]
(Figure 13). Thus the empirical rate law for the hydrogenolysis of
10 is ꢀd[10]/dt = kobs[H2][10].
Figure 13. Linear relationship between ln[10] and time. Kinetic plots
for reactions of 10 (3.6 μmol) with 3.5 atm (2) and 7.0 atm (9) H2 in
C6D6 with excess FE (32 μmol) at 60 °C are shown.
Notably, when hydrogenolysis reactions of 10 were performed
in the absence of excess FE, small amounts of the hydroxide
complex 1 were consistently observed by NMR spectroscopy as a
contaminant. The presence of 1 is an indication of the sensitivity
of complex 10 to hydrolysis by adventitious H2O. It was generally
observed that the experiments in which higher concentrations of
1 were detected were also noted to yield faster reaction times.
These reactions also exhibited deviations from the first-order
reaction dependency in [10]; however, the data did not fit half-
order kinetics either. This kinetic behavior would be consistent
with the simultaneous operations of paths a and b (Scheme 5),
with the contribution of path b depending on the concentration
of adventitious water available to generate 1. In this situation,
the palladium hydride complex 2 can be formed from the
hydrogenolyses of both complexes 10 and 1. In path b, H2O
acts as a catalyst, hydrolyzing the alkoxide complex to the
hydroxide complex 1, which then reacts with hydrogen to
regenerate water. The perturbation from first-order kinetics is
attributed to the dimer formation that takes place for 1 as
discussed above. In contrast, when an excess of FE is present
during hydrogenolysis reactions of the 2-fluoroethoxide complex
17722
dx.doi.org/10.1021/ja205824q |J. Am. Chem. Soc. 2011, 133, 17713–17726

Journal of the American Chemical Society
ARTICLE
10, the hydrolysis of 10 is inhibited, and the hydrogenolysis of 10
primarily proceeds via path a in Scheme 5.44
kinetics are consistent with a direct IES reaction of 10 with dihy-
drogen, similar, albeit slower, to that observed for the monomeric
hydroxide 1.
The empirical rate law for the hydrogenolysis of 10, deter-
mined with excess FE present, shows a first-order dependence on
[10] and a first-order dependence on [H2], and is thus consistent
with a mechanism analogous to that proposed for the PdꢀOH
monomer (eq 5). Interaction of H2 with the Pd center and the
lone pair of the oxygen leads directly to the PdꢀH and OꢀH
bond formations. Thus, once the competing reactions of dimer-
ization, hydrolysis, and DBHA are inhibited, hydrogenolysis of
PdꢀO bonds of hydroxide and alkoxide complexes was found to
occur via a common mechanism.
It is evident from this study that hydrogenolysis of PdꢀOR
bonds can occur in a directly analogous fashion to the PdꢀOH
reaction with the caveat that inhibition by dimer formation is not
generally observed for the alkoxides. In addition, in the case of
the production of acidic alcohols, hydrogenolysis is thermody-
namically unfavorable. Finally, there are significant competing
side reactions evident for alkoxide complexes. DBHA generates
the same palladium product as hydrogenolysis, but instead of the
alchohol, unsaturated organics are produced. DBHA can be
promoted with the addition of the alcohol, and it can be
minimized through the use of sterically bulky alkoxides or
with electron-withdrawing substituents on the alkoxide. Finally,
hydrolysis followed by hydrogenolysis of the PdꢀOH will
actually generate the identical products as hydrogenolysis of
the PdꢀOR complex. Thus, hydrogenolysis of metal hydroxides
and metal alkoxides can both lead to regeneration of a metal
hydride in a catalytic cycle. The wide range of substrates PdꢀOR
(R = H, CH3, C6H5, CH2C(CH3)3, CH2CF3, CH2CHF2,
CH2CH2F) examined in this study helps to clarify the conditions
and the considerations necessary for the successful utilization of
this hydrogenolysis reaction in a catalytic system.
k
PdOCH2CH2F þ H2
’ SUMMARY
s
PdH þ HOCH CH F
ð5Þ
f
2
2
Comparison of the reactivity of palladium(II) hydroxide,
alkoxide, and phenoxide complexes with hydrogen has revealed
a range of reactions. Whether hydrogenolysis, β-hydride abstrac-
tion, hydrolysis (by adventitious water), or even a lack of reaction
was observed, the reactivity was found to be controlled by the
basicity of the ꢀOR group, its steric bulk, and its susceptibility to
a dissociative β-hydride abstraction pathway. Hydrogen reacts
directly with the hydroxide complex 1 to generate water and the
palladium(II) hydride complex 2. This hydrogenolysis reaction is
retarded by dimer formation involving hydrogen-bonded water
molecules. Hydrogenolysis occurs with the monomeric d8 metal
species through an internal electrophilic substitution (IES)
mechanism (Scheme 3). A similar direct hydrogenolysis occurs
with the palladium(II) methoxide complex 3 to generate the
hydride complex 2 and methanol. However detailed kinetic and
mechanistic studies of the hydrogenolysis of 3 could not be
performed due to a competing hydrolysis by adventitious water
which forms the hydroxide complex 1. Addition of excess
methanol to thwart hydrolysis results in a rapid dissociative
β-hydride abstraction (DBHA) reaction that also produces the
hydride complex 2.
The palladium(II) neopentoxide complex 7 was also synthe-
sized. Notably thiscomplex was found to be stableto DBHA, even
in the presence of added neopentanol at elevated temperatures.
The sterics of the neopentyl group apparently hinder the DBHA
reaction. However, the sterics of the neopentoxide ligand, combi-
ned with the sterics of the tBu groups on the phosphine, also inhi-
bit reaction with hydrogen, and no hydrogenolysis was observed.
The reactivity of the palladium(II) phenoxide complex 5,
2,2,2-trifluoroethoxide complex 8, and 2,2-difluoroethoxide
complex 9 with hydrogen was limited. In these cases, thermo-
dynamics are the main consideration as the palladium alkoxide
complexes and H2 are favored over the palladium(II) hydride
complex 2 and alcohol products. This preference was convincingly
demonstrated by the reaction of the hydride complex 2 with the
alcohol. Notably the alcohols in question (i.e., phenol, TFE, and
DFE) are all quite acidic, with pKa values of 9.95, 12.4, and 13.3,
respectively.38,41 FE is less acidic with a pKa of 14.4, and the
equilibrium in this case favored the hydride. Thus similar to the
palladium hydroxide 1 and the palladium methoxide 3, the
palladium(II) 2-fluoroethoxide complex 10 undergoes hydrogeno-
lysis. The reaction of 10 with hydrogen produces the palladium-
(II) hydride and FE. The fluorine substitution disfavors a DBHA
reaction, and kinetic studies of the hydrogenolysis could be carried
out in the presence of excess FE to prevent hydrolysis reactions. The
’ EXPERIMENTAL SECTION
Unless specified otherwise, all manipulations were carried out under
nitrogen using conventional vacuum line techniques or a glovebox
equipped with a ꢀ35 °C freezer. Solvents were purified before use.
THF, benzene, and pentane were purified by passage through columns
of activated alumina and molecular sieves. C6D6 was dried over sodium
metal/benzophenone. Fluoroethanols were dried over activated molec-
ular sieves (3 Å) and distilled under a nitrogen atmosphere. Unless
otherwise noted, all other reagents were used as obtained from
commercial suppliers. NMR spectra were obtained at room temperature
(25 °C) on Bruker AV300 and AV500 MHz spectrometers, with
chemical shifts (δ) reported in ppm. All 1H NMR spectra were
referenced to the residual protiated solvent signal; 31P NMR and 19F
NMR spectra were referenced externally to 85% H3PO4 (0 ppm) and
CF3COOH (ꢀ76.55 ppm), respectively. All 13C and 31P NMR data
were collected proton-decoupled (13C{1H} and 31P{1H}). All 1H and
13C resonances were assigned using gradient-selected heteronuclear
multiple-quantum coherence (gsHMQC) NMR spectroscopy and
gradient-selected correlation spectroscopy (gsCOSY). Multiplicity is
reported as follows: s, singlet; d, doublet; t, triplet; q, quartet; vt, virtual
triplet; br, broad. Figure 14 illustrates the aromatic carbon-numbering
scheme for all reported NMR data. Elemental analyses were carried out
by Atlantic Microlab, Inc. of Norcross, GA. All crystallographic data was
collected at the University of Washington X-ray crystallography lab. H2-
pressurized NMR-scale reactions were achieved using a gas pressuriza-
tion apparatus.45 The following complexes were prepared according to
Figure 14. Aromatic carbon-numbering scheme for (tBuPCP)Pd
complexes.
17723
dx.doi.org/10.1021/ja205824q |J. Am. Chem. Soc. 2011, 133, 17713–17726

Journal of the American Chemical Society
ARTICLE
published procedures: (tBuPCP)PdOH (1),18
tBuPCP)PdOCH3 (3),9 (tBuPCP)PdONO2.20
Safety Note: Great caution must be taken when handling and heating
pressurized medium-walled NMR tubes. At all times, pressurized NMR
tubes were heated inside stainless steel jackets and transported and
stored in a protective plastic jacket.
Kinetic Studies: Reaction of 1 and H2. In a typical experiment, a
C6D6 solution of 1 (4.2 mM), H2O (37.8 mM), and hexamethylbenzene
(internal standard) was transferred to a medium-walled NMR tube fitted
with a resealable Telflon valve. The sample was degassed by three
freezeꢀpumpꢀthaw cycles and then placed under hydrogen pressure
(7.0 or 3.5 atm) using a gas pressurization apparatus.45 Samples were
placed in a 25 °C temperature-controlled oil bath during the reaction.
Rates were determined by following the disappearance of the methylene
signal (CH2P) of 1 through three half-lives, as compared to the methyl
signal of hexamethylbenzene.
(
tBuPCP)PdH (2),18
for 4 h. The volatiles were removed under reduced pressure, and the
peach-colored solid was extracted with benzene (20 mL) and filtered
through a Teflon filter (0.2 μm). The solvent was removed from the
filtrate under vacuum, leaving a peach solid which was then dissolved in a
minimum of benzene (0.5 mL). The solution was layered with pentane
(4 mL) and cooled to ꢀ35 °C for 3 days. During this time, faint-yellow
crystals formed. The solvent was decanted, and the crystals were dried in
vacuo. Yield: 38.2 mg (48.1%). 1H NMR (C6D6, 500 MHz) δ = 7.38 (t,
2H, 3JHH = 7.1 Hz, m-Ph-H), 7.08 (d, 2H, 3JHH = 7.5 Hz, o-Ph-H), 7.03
(t, 1H, 3JHH = 7.5 Hz, H4), 6.92 (d, 2H, 3JHH = 7.5 Hz, H3,5), 6.73 (t,
2H, 3JHH = 7.5 Hz, p-Ph-H), 2.94 (vt, 4H, 2JHP = 3.7 Hz, CH2P), 1.21 (vt,
36H, 3JHP = 6.8 Hz, C(CH3)3). 13C{1H} NMR (C6D6, 126 MHz) δ =
171.3 (s, i-Ph-C), 155.7 (s, C1), 151.7 (vt, 3JCP = 10.3 Hz, C2,6), 129.1
(s, m-Ph-C), 125.0 (s, C4), 122.5 (vt, 3JCP = 10.0 Hz, C3,5), 121.1 (s, o-
Ph-C), 112.1 (s, p-Ph-C), 34.7 (vt, 1JCP = 7.0 Hz, C(CH3)3), 34.1 (vt,
1JCP = 10.2 Hz, CH2P), 29.4 (s br, C(CH3)3). 31P{1H} NMR (C6D6,
203 MHz) δ = 70.90. Anal. Calcd for C30H48OP2Pd: C, 60.76; H, 8.16.
Found: C, 60.82; H, 8.20.
(
Reaction of 3 with H2. Formation of (tBuPCP)PdH (2). A
medium-walled NMR tube fitted with a resealable Teflon valve was
charged with complex 3 (2.0 mg, 3.9 μmol) and hexamethylbenzene
(internal standard). The solids were then dissolved in C6D6 (0.4 mL).
Thesolutionwas degassed, andthedesired hydrogen pressure wasadded.
The reaction was monitored by 1H NMR spectroscopy, and the disap-
pearance of3, along withthe appearanceof 2 and methanol, was observed.
Reaction of 3 with (p-OCH2)n. Formation of (tBuPCP)Pd-
OCH2OCH3 (4). An NMR tube fitted with a resealable Teflon valve was
charged with complex 3 (2.2 mg, 4.1 μmol), para-formaldehyde (1.2 mg,
40 μmol), and the internal standard hexamethylbenzene (1.0 mg,
6.2 μmol). Benzene-d6 (0.4 mL) was vacuum transferred into the tube,
and the solution was immediately analyzed by NMR spectroscopy. The
partial conversion of 3 and formation of (tBuPCP)PdOCH2OCH3 (4)
was observed: 1H NMR (C6D6, 500 MHz) δ = 7.01 (t, 1H, JHH = 7.3 Hz,
H4), 6.92 (d, 2H, JHH = 7.3 Hz, H3,5), 5.72 (s, 2H, OCH2OCH3), 3.65
(s, 3H, OCH2OCH3), 2.93 (vt, 4H, JHP = 3.4 Hz, CH2P), 1.29 (vt, 36H,
Reaction of 2 with HOPh. Formation of [(μ-tBuPCHP)Pd]2
(6). Complex 2 (6.0 mg, 12 μmol), phenol (2.3 mg, 24 μmol), and
hexamethylbenzene (1ꢀ2 flakes, internal standard) were weighed into a
medium-walled NMR tube fitted with a resealable Teflon valve. C6D6
(0.4 mL) was vacuum transferred into the tube, and once the solids dissolved,
immediate formation of bubbles was observed. 1H NMR spectroscopy
revealed the initial formation of the phenoxide complex 5 and free H2, and
overthecourseof2h, thesolutionchanged from colorless to pale-yellow. The
formation of faint-yellow crystals occurred over a period of 2 days, which
were identified by X-ray crystallography as [(μ-tBuPCHP)Pd]2 (6).
(
tBuPCP)PdOCH2C(CH3)3 (7). The nitrato complex (tBuPCP)PdO-
NO2 (171.3 mg, 0.305 mmol) was added to a round-bottom flask along
with an excess of KOCH2C(CH3)3 (58.7 mg, 0.465 mmol). The solids
were dissolved in THF (20 mL), and the pale orange-colored solution
was stirred for 12 h. The volatiles were removed under reduced pressure,
and the peach-colored solid was extracted with benzene (25 mL) and
filtered through a Teflon filter. The solvent was removed from the filtrate
under vacuum, leaving a peach-colored solid. Complex 7 was redissolved
in pentane and was recrystallized from a slow evaporation of the
J
HP = 6.5 Hz, C(CH3)3). 31P{1H} NMR (C6D6, 203 MHz) δ = 70.9.
After an additional 30 min, the complete conversion of 3 had occurred.
In the 1H NMR spectrum, complex 4 was observed along with numerous
peaks between 3ꢀ4 ppm and 4.5ꢀ6 ppm, corresponding to ꢀOCH2ꢀ
and ꢀOCH3 protons of (tBuPCP)Pd(OCH2)nOCH3 complexes,
respectively, resulting from an additional n number of insertions of
OCH2 into the PdꢀO bond of 3. In the 31P{1H} NMR spectrum,
several overlapping singlets resonated at 72ꢀ70 ppm.
1
solution. Yield: 88.7 mg (49.6%). H NMR (C6D6, 500 MHz) δ =
7.04 (t, 1H, 3JHH = 7.4 Hz, H4), 6.96 (d, 2H, 3JHH = 7.4 Hz, H3,5), 4.06
(s, 2H, OCH2C(CH3)3), 2.98 (vt, 4H, 2JHP = 3.8 Hz, CH2P), 1.31 (vt,
36H, 3JHP = 6.6 Hz, C(CH3)3), 1.31 (s, 9H, OCH2C(CH3)3). 13C{1H}
NMR (C6D6, 126 MHz) δ = 158.4 (s, C1), 151.3 (vt, 2JCP = 10.0 Hz,
C2,6), 124.5 (s, C4), 121.9 (vt, 3JCP = 9.6 Hz, C3,5), 85.9 (s, OCH2C-
(CH3)3), 36.6 (OCH2C(CH3)3), 35.2 (vt, 1JCP = 10.2 Hz, CH2P), 34.8
(vt, 1JCP = 6.7 Hz, C(CH3)3), 29.4 (vt, 2JCP = 3.3 Hz, C(CH3)3), 28.4 (s,
OCH2C(CH3)3). 31P{1H} NMR (C6D6, 203 MHz) δ = 68.87. Anal.
Calcd for C29H54OP2Pd: C, 59.33; H, 9.27. Found: C, 59.28; H, 9.33.
Reaction of 2 with HOCH2CF3. Formation of (tBuPCP)Pd-
OCH2CF3 (8). Complex 2 (3.0 mg, 6.0 μmol) was weighed into a
medium-walled NMR tube fitted with a resealable Teflon valve. The
complex was dissolved in C6D6 (0.40 mL). An excess of deoxygenated
2,2,2-trifluoroethanol (5.0 μL, 70 μmol) was added by syringe. Im-
mediate formation of bubbles was observed, and upon closer inspection
by NMR spectroscopy, complete conversion of 2 to the 2,2,2-trifluor-
oethoxide complex 8 and H2 had occurred. 1H NMR (C6D6, 500 MHz)
δ = 7.00 (t, 1H, 3JHH = 7.3 Hz, H4), 6.89 (d, 1H, 3JHH = 7.3 Hz, H3,5),
Reaction of 3 with H2O. A standard solution of complex 3
(25 mM) and the internal standard hexamethylbenzene (13 mM) was
made in benzene-d6. Performed in triplicate, 0.380 mL of the stock solu-
tion of complex 3(5.0 mg, 9.4 μmol) was transferred by syringe into an NMR
tube with a resealable Teflon valve. After initial 1H and 31P{1H} NMR
spectra were acquired, each NMR tube was treated with nearly 0.5 equiv
of water (120 μL of a 33 mM H2O solution in C6D6, 4.0 μmol).46 The total
volume of the solution was 0.500 mL, [3]0 = 19 mM, and [H2O]0 =
8.0 mM. The partial conversion of complex 3 and water to 1 and methanol
was observed by NMR spectroscopy. The concentrations of 1, 3, methanol,
and water33 were measured by integration against the methyl signal of hexa-
methylbenzene over the course of 5 h until concentrations equilibrated.
Reaction of 1 with CH3OH. A benzene-d6 solution (0.400 mL) of
complex 1 (26.5 mM) and hexamethylbenzene (13 mM) was added to
1
an NMR tube with a resealable Teflon valve. Initial H and 31P{1H}
3
2
4.65 (q, 2H, JHF = 9.9 Hz, CH2CF3), 2.89 (vt, 4H, JHP = 3.8 Hz,
NMR spectra were collected. The solution was then treated with 1 equiv
of methanol (0.43 μL, 10.6 μmol), and the reaction mixture was
monitored by 1H NMR spectroscopy. Over the course of 2 days, only
a small amount of complex 1 (<10%) had converted to complex 3.
CH2P), 1.23 (vt, 36H, 3JHP = 6.7 Hz, C(CH3)3). 13C{1H} NMR (C6D6,
126 MHz) δ = 156.3 (s, C1), 151.5 (vt, JCP = 10.2 Hz, C2,6), 124.8
2
3
1
(C4), 122.2 (vt, JCP = 9.9 Hz, C3,5), 128.5 (q, JCF = 284.2 Hz,
CH2CF3), 72.0 (q, 2JCF = 29.6 Hz, CH2CF3), 34.7 (vt, 1JCP = 7.0 Hz,
(
tBuPCP)PdOC6H5 (5). The nitrato complex (tBuPCP)PdONO2
1
2
C(CH3)3), 34.4 (vt, JCP = 10.2 Hz, CH2P), 29.2 (vt, JCP = 3.2 Hz,
C(CH3)3). 31P{1H} NMR (C6D6, 203 MHz) δ = 70.91. 19F NMR
(C6D6, 282 MHz) δ = ꢀ73.3 (t, 3JFH = 9.9 Hz).
(75.5 mg, 0.134 mmol) was added to a round-bottom flask along with an
excess of KOC6H5 (26.0 mg, 0.197 mmol). The solids were dissolved in
THF (15 mL), and the light orangish-yellow colored solution was stirred
17724
dx.doi.org/10.1021/ja205824q |J. Am. Chem. Soc. 2011, 133, 17713–17726

Journal of the American Chemical Society
ARTICLE
Reaction of 2 with HOCH2CHF2. Formation of (tBuPCP)Pd-
OCH2CHF2 (9). Complex 2 (3.0 mg, 6.0 μmol) was weighed into a
medium-walled NMR tube fitted with a resealable Teflon valve. The
complex was dissolved in C6D6 (0.40 mL). An excess of deoxygenated
2,2-difluoroethanol (5.0 μL, 79 μmol) was added by syringe. The
complete conversion of complex 2 to the 2,2-difluoroethoxide complex
9 and H2 was observed by NMR spectroscopy over the course of 4 h. 1H
NMR (C6D6, 500 MHz) δ = 7.01 (t, 1H, 3JHH = 7.2 Hz, H4), 6.91 (d,
1H, 3JHH = 7.2 Hz, H3,5), 6.07 (tt, 1H, 2JHF = 58.0 Hz, CHF2, 3JHH = 4.2
Hz) 4.51 (td, 2H, 3JHF = 15.2 Hz, 3JHH = 4.2 Hz, CH2CHF2), 2.91 (vt,
CH2CF3), 3.29 (s, 1H, OH). 19F NMR (C6D6, 282 MHz) δ = ꢀ76.54
3
1
(t, JFH = 8.9 Hz). 2,2-Difluoroethanol: H NMR (C6D6, 500 MHz)
δ = 5.26 (tt, 1H, 2JHF = 55.8 Hz, 3JHH = 3.8 Hz, CHF2), 3.18 (td, 2H,
3JHH = 3.8 Hz, 3JHF = 14.7 Hz, CH2CHF2), 2.04 (s, 1H, OH). 19F NMR
2
3
(C6D6, 282 MHz) δ = ꢀ126.2 (dt, JFH = 55.8 Hz, JFH = 14.7 Hz).
2-Fluoroethanol: 1H NMR (C6D6, 500 MHz) δ = 4.21 (dt, 2H, 2JHF
=
=
48.0 Hz, 3JHH = 4.2 Hz, CH2F), 4.00 (s, 1H, OH), 3.51 (dt, 2H, 3JHF
3
29.8 Hz, JHH = 4.2 Hz, CH2CH2F). 19F NMR (C6D6, 282 MHz)
δ = ꢀ224.4 (tt, 2JFH = 48.0 Hz, 3JFH = 29.8 Hz).48
2
3
4H, JHP = 3.8 Hz, CH2P), 1.24 (vt, 36H, JHP = 6.7 Hz, C(CH3)3).
’ ASSOCIATED CONTENT
13C{1H} NMR (C6D6, 126 MHz) δ = 156.3 (s, C1), 151.4 (vt, 2JCP
10.0 Hz, C2,6), 124.9 (C4), 122.2 (vt, 3JCP = 9.8 Hz, C3,5), 120.3 (t, 1JCF
244.1 Hz, CH2CHF2), 73.7 (t, 2JCF = 24.1 Hz, CH2CHF2), 34.8 (vt, 1JCP
=
=
=
=
S
Supporting Information. X-ray crystallographic data in-
b
cluding files in CIF format for 1•FE, 5, 6, 7, and 10, as well as
additional kinetic plots, computational methodology and XYZ
coordinates, and apparatus diagrams. This material is available
free of charge via the Internet at http://pubs.acs.org.
1
2
6.7 Hz, C(CH3)3), 34.4 (vt, JCP = 10.3 Hz, CH2P), 29.3 (vt, JCP
3.2 Hz, C(CH3)3). 31P{1H} NMR (C6D6, 203 MHz) δ = 70.42. 19F NMR
(C6D6, 282 MHz) δ = ꢀ124.5 (dt, 2JFH = 58.0 Hz, 3JFH = 15.2 Hz).
(
tBuPCP)PdOCH2CH2F (10). Complex 7 (256.0 mg, 436 μmol)
was added to a modified H-flask47 fitted with a resealable Teflon valve
and dissolved in benzene (5 mL). To the solution, an excess of
2-fluoroethanol (50 μL, 850 μmol) was added by syringe. The reaction
was stirred for 2 h, whereupon the volatiles were removed under
vacuum. The resulting brown solid was dried under reduced pressure
using a vacuum line. Benzene (5 mL) was vacuum transferred into the
H-flask to dissolve the solids. With the solution under static vacuum, a
cold bath (evaporating acetone, 19 °C) was used to slowly vacuum
transfer the benzene from one well of the H-flask to the other. Over the
course of two days, tan-colored crystals of 10 formed as the solution
gradually concentrated. The supernatant was decanted, and the crystals
of 10 were isolated and pumped to dryness on the vacuum line. Yield:
’ AUTHOR INFORMATION
Corresponding Author
goldberg@chem.washington.edu; rakemp@unm.edu
’ ACKNOWLEDGMENT
We thank Richard P. Muller for computations, Alexander J. M.
Miller for insightful discussions, and the Department of Energy
(DE-FG02-06ER15765) for support.
138.7 mg (56.5%). 1H NMR (C6D6, 500 MHz) δ = 7.03 (t, 1H, 3JHH
=
’ REFERENCES
7.4 Hz, H4), 6.93 (d, 1H, 3JHH = 7.4 Hz, H3,5), 4.78 (dt, 2H, 2JHF = 48.7
Hz, 3JHH = 5.7 Hz, CH2F), 4.53 (dt, 2H, 3JHF = 21.3 Hz, 3JHH = 5.7 Hz,
(1) Hartwig, J. F. Organotransition Metal Chemistry: From Bonding to
Catalysis; University Science Books: Sausalito, CA, 2010.
(2) Crabtree, R. H. The Organometallic Chemistry of the Transition
Metals, 4th ed.; John Wiley & Sons: New York, 2005.
(3) Handbook of Homogeneous Hydrogenation; De Vries, J. G., Elsevier,
C. J., Eds.; Wiley-VCH: Weinheim, Germany, 2007.
(4) Esteruelas, M. A.; Oro, L. A. Chem. Rev. 1998, 98, 577.
(5) (a) Maron, L.; Eisenstein, O. J. Am. Chem. Soc. 2001, 123, 1036.
(b) Lee, J. C., Jr.; Peris, E.; Rheingold, A. L.; Crabtree, R. H. J. Am. Chem.
Soc. 1994, 116, 11014. (c) Joshi, A. M.; James, B. R. Organometallics
1990, 9, 199. (d) Watson, P. L.; Parshall, G. W. Acc. Chem. Res. 1985,
18, 51.
CH2CH2F), 2.94 (vt, 4H, 2JHP = 3.6 Hz, CH2P), 1.28 (vt, 36H, 3JHP
=
6.7 Hz, C(CH3)3). 13C{1H} NMR (C6D6, 126 MHz) δ = 157.9 (s, C1),
151.5 (vt, 2JCP = 10.1 Hz, C2,6), 124.6 (C4), 122.0 (vt, 3JCP = 9.8 Hz,
1
2
C3,5), 88.9 (d, JCF = 171.0 Hz, CH2CH2F), 72.1 (t, JCF = 20.2 Hz,
CH2CH2F), 32.7 (vt, 1JCP = 10.1 Hz, CH2P), 32.7 (vt, 1JCP = 6.8 Hz,
C(CH3)3), 29.3 (vt, 2JCP = 3.4 Hz, C(CH3)3). 31P{1H} NMR (C6D6,
203 MHz) δ = 69.86. 19F NMR (C6D6, 282 MHz) δ = ꢀ218.6 (tt, 2JFH
=
48.7 Hz, 3JFH = 21.3 Hz). Anal. Calcd for C26H47FOP2Pd: C, 55.47; H,
8.41. Found: C, 55.35; H, 8.64.
Control Experiment: Complex 10 in C6D6 with Excess FE
in Absence of H2. Complex 10 (2.0 mg, 3.6 μmol), 2-fluoroethanol
(1.9 μL, 32 μmol), and hexamethylbenzene (1ꢀ2 flakes, internal
standard) were added to a medium-walled NMR tube fitted with a
resealable Teflon valve. The NMR tube was sealed under an atmosphere
of N2 and heated to 60 °C in a temperature-controlled oil bath. The
sample was analyzed by 1H NMR spectroscopy over the course of 7 days,
during which <5% conversion of 10 to 2 was observed.
Kinetic Studies: Reaction of 10 and H2. In a typical experiment,
a medium-walled NMR tube fitted with a resealable Teflon valve was
charged with complex 10 (2.0 mg, 3.6 μmol), 2-fluoroethanol (1.9 μL,
32 μmol), and hexamethylbenzene (1ꢀ2 flakes, internal standard).
Using a vacuum line, 0.40 mL of the C6D6 was vacuum transferred from
a sodiumꢀpotassium alloy into the NMR tube. The degassed sample
was then placed under hydrogen pressure (7.0ꢀ3.5 atm) using a gas
pressurization apparatus.45 Initial 1H NMR spectra were collected before
heating the sample to 60 °C in a temperature-controlled oil bath. Rates
were determined by following the disappearance of complex 10’s
ꢀCH2F resonance through three half-lives, as compared to the methyl
signal of hexamethylbenzene.
(6) (a) Samec, J. S. M.; B€ackvall, J. E.; Andersson, P. G.; Brandt, P.
Chem. Soc. Rev. 2006, 35, 237. (b) Noyori, R.; Ohkuma, T. Angew. Chem.,
Int. Ed. 2001, 40, 40.
(7) (a) Conley, B. L.; Pennington-Boggio, M. K.; Boz, E.; Williams,
T. J. Chem. Rev. 2010, 110, 2294. (b) Blum, Y.; Czarkie, D.; Rahamim, Y.;
Shvo, Y. Organometallics 1985, 4, 1459.
(8) This type of intramolecular proton transfer has been termed
internal electrophilic substitution (IES). (a) Oxgaard, J.; Tenn, W. J., III;
Nielsen, R. J.; Periana, R. A.; Goddard, W. A., III. Organometallics 2007,
26, 1565. (b) Cundari, T. R.; Grimes, T. V.; Gunnoe, T. B. J. Am. Chem.
Soc. 2007, 129, 13172.
(9) For a preliminary report on the hydrogenolysis of palladium(II)
hydroxide and methoxide pincer complexes, see: Fulmer, G. R.; Muller,
R. P.; Kemp, R. A.; Goldberg, K. I. J. Am. Chem. Soc. 2009, 131, 1346.
(10) Bryndza, H. E.; Fong, L. K.; Paciello, R. A.; Tam, W.; Bercaw,
J. E. J. Am. Chem. Soc. 1987, 109, 1444.
(11) (a) Blum, O.; Milstein, D. J. Organomet. Chem. 2000, 479, 593.
(b) Ritter, J. C. M.; Bergman, R. G. J. Am. Chem. Soc. 1998, 120, 6826.
(c) Blum, O.; Milstein, D. J. Am. Chem. Soc. 1995, 117, 4582.
(12) Pearlman, W. M. Tetrahedron Lett. 1967, 1663.
(13) (a) Adams, R.; Shriner, R. L. J. Am. Chem. Soc. 1923, 45, 2171.
(b) Carothers, W. H.; Adams, R. J. Am. Chem. Soc. 1923, 45, 1071.
(c) Voorhees, V.; Adams, R. J. Am. Chem. Soc. 1922, 44, 1397.
Fluoroethanol NMR Spectroscopic Data. 2,2,2-Trifluoroetha-
1
3
nol: H NMR (C6D6, 500 MHz) δ = 3.41 (q, 2H, JHF = 8.8 Hz,
17725
dx.doi.org/10.1021/ja205824q |J. Am. Chem. Soc. 2011, 133, 17713–17726

Journal of the American Chemical Society
ARTICLE
(14) (a) Deutsch, C.; Krause, N.; Lipshutz, B. H. Chem. Rev. 2008,
108, 2916. (b) Mahoney, W. S.; Brestensky, D. M.; Stryker, J. M. J. Am.
Chem. Soc. 1988, 110, 291. (c) Goeden, G. V.; Caulton, K. G. J. Am.
Chem. Soc. 1981, 103, 7354.
(39) For space-filling diagrams of 1, 3, 5, 7, and 10, see Figure S6 in
the Supporting Information.
(40) Smith, M. B.; March, J. March’s Advanced Organic Chemistry:
Reactions, Mechanisms, and Structure, 5th ed.; John Wiley & Sons:
New York, 2001; p 1176.
(41) Ballinger, P.; Long, F. A. J. Am. Chem. Soc. 1960, 82, 795.
(42) Mock, W. L.; Zhang, J. Z. Tetrahedron Lett. 1990, 5687.
(43) For additional kinetic plots, see Figure S1 in the Supporting
Information.
(44) It was considered that the hydrogenolysis of 10 could proceed
exclusively via path b (Scheme 5) in the presence of excess FE but
follows a well-behaved first-order dependency on Pd if [H2O] were too
low for 1 to dimerize. However, when hydrogenolysis reactions were
performed on 1 in the presence of a drying agent (MgSO4), the reaction
kinetics were still observed to be half-order in [1], indicating that
dimerization still occurs when H2O concentrations are below the
detection limit of 1H NMR spectroscopy.
(15) Jessop, P. G.; Ikariya, T.; Noyori, R. Chem. Rev. 1995, 95, 259.
(16) Wenzel, T. T. Dioxygen Activation and Homogeneous Catalytic
Oxidation; Simꢀandi, L. I., Ed.; Studies in Surface Science and Catalysis
66; Elsevier Science Publishers B. V.: Amsterdam, 1991; p 545.
(17) (a) Joanna, R.; Webb, J. R.; Munro-Leighton, C.; Pierpont,
A. W.; Gurkin, J. T.; Gunnoe, T. B.; Cundari, T. R.; Sabat, M.; Petersen,
J. L.; Boyle, P. D. Inorg. Chem. 2011, 50, 4195. (b) Webb, J. R.; Pierpont,
A. W.; Colleen Munro-Leighton, C.; T. Brent Gunnoe, T. B.; Cundari,
T. R.; Boyle, P. D. J. Am. Chem. Soc. 2010, 132, 4520. (c) B€ohler, C.;
Avarvari, N.; Sch€onberg, H.; W€orle, M.; R€uegger, H.; Gr€utzmacher, H.
Helv. Chim. Acta 2001, 84, 3127. (d) Thompson, J. S.; Randall, S. L.;
Atwood, J. D. Organometallics 1991, 10, 3906. (e) Thompson, J. S.;
Bernard, K. A.; Rappoli, B. J.; Atwood, J. D. Organometallics 1990,
9, 2727.
(45) For diagrams of gas pressurization apparatus, see Figures
(18) Denney, M. C.; Smythe, N. A.; Cetto, K. L.; Kemp, R. A.;
Goldberg, K. I. J. Am. Chem. Soc. 2006, 128, 2508.
S2ꢀS3 in the Supporting Information.
(46) Notably, the reported solubility of water in benzene at 25 °C is
36 mM. See ref 21.
(19) Fulmer, G. R.; Miller, A. J. M.; Sherden, N. H.; Gottlieb, H. E.;
Nudelman, A.; Stoltz, B. M.; Bercaw, J. E.; Goldberg, K. I. Organome-
tallics 2010, 29, 2176.
(47) For a diagram of modified H-flask, see Figure S4 in the
Supporting Information.
(48) For more information on fluoroethanols, see: Elliott, A. J.
Fluoroethanols: Kirk-Othmer Encyclopedia of Chemical Terminology; John
Wiley & Sons: 2000.
€
(20) Johansson, R.; Ohrstr€om, L.; Wendt, O. F. Cryst. Growth Des.
2007, 7, 1974.
(21) 9 equiv corresponds to a water concentration of 38 mM, very
close to the reported solubility of water in benzene at 25 °C (36 mM).
Karlsson, R. J. Chem. Eng. Data 1973, 18, 290.
(22) Derived from an Ostwald coefficient of 0.0718, hydrogen
concentrations were calculated to be 21 mM (8.2 μmol) and 10 mM
(4.1 μmol) for 7.0 and 3.5 atm, respectively. Young, C. L. Hydrogen and
Deuterium, Solubility Data Series; Pergamon Press: Oxford, 1981; Vol. 5/6.
(23) Espenson, J. H. Chemical Kinetics and Reaction Mechanisms, 2nd
ed.; McGraw-Hill: New York, 1995; p 82.
ꢀ
(24) Cꢀampora, J.; Palma, P.; del Río, D.; Alvarez, E. Organometallics
2004, 23, 1652.
(25) Billo, E. J. Excel for Chemists, 2nd ed.; John Wiley & Sons, Inc.:
Canada, 2001; p 184.
(26) Additional data concerning the RuggaꢀKutta Fourth Order
method can be found in the Supporting Information.
(27) (a) Hull, K. L.; Lanni, E. L.; Sanford, M. S. J. Am. Chem. Soc.
2006, 128, 14047. (b) Canty, A. Acc. Chem. Res. 1992, 25, 83.
(28) (a) Milet, A.; Dedieu, A.; Kapteijn, G. M.; van Koten, G. Inorg.
Chem. 1997, 36, 3223. (b) Hutschka, F.; Dedieu, A.; Eichberger, M.;
Fornika, R.; Leitner, W. J. Am. Chem. Soc. 1997, 119, 4432. (c) Hutschka,
F.; Dedieu, A.; Leitner, W. Angew. Chem., Int. Ed. Engl. 1995, 34, 1742.
(29) (a) Cundari, T. R.; Grimes, T. V.; Gunnoe, T. B. J. Am. Chem.
Soc. 2007, 129, 13172. (b) Oxgaard, J.; Tenn, W. J., III; Nielsen, R. J.;
Periana, R. A.; Goddard, W. A., III. Organometallics 2007, 26, 1565.
(30) Kimmich, B. F. M.; Bullock, R. M. Organometallics 2002,
21, 1504.
(31) (a) Pawlikowski, A. V.; Getty, A. D.; Goldberg, K. I. J. Am.
Chem. Soc. 2007, 129, 10382. (b) Williams, B. S.; Goldberg, K. I. J. Am.
Chem. Soc. 2001, 123, 2576.
(32) See the Supporting Information for computational methodol-
ogy, XYZ coordinates, and additional energy data.
1
(33) Integrations of the H NMR resonance for water were not
accurate due to exchange with the hydroxide complex 1; therefore, the
concentration of water was determined by subtracting [CH3OH] from
the initial concentration of water ([H2O]0).
(34) Bryndza, H. E.; Tam, W. Chem. Rev. 1988, 88, 1163.
(35) Smythe, N. A.; Grice, K. A.; Williams, B. S.; Goldberg, K. I.
Organometallics 2009, 28, 277.
(36) Fafard, C. M.; Ozerov, O. V. Inorg. Chim. Acta 2007, 360, 286.
(37) Melero, C.; Martínez-Prieto, L. M.; Plama, P.; del Rio, D.;
Alvarez, E.; Cꢀampora, J. Chem. Commun. 2010, 46, 8851.
ꢀ
(38) Bruice, P. Y. Organic Chemistry, 3rd ed.; Prentice Hall: NJ,
2001; p 645.
17726
dx.doi.org/10.1021/ja205824q |J. Am. Chem. Soc. 2011, 133, 17713–17726