Mechanistic studies of single-step styrene production using a rhodium(I) catalyst

Article abstract of DOI:10.1021/jacs.6b10658

The direct and single-step conversion of benzene, ethylene, and a Cu(II) oxidant to styrene using the Rh(I) catalyst (^FlDAB)Rh(TFA)(η²-C₂H₄) [^FlDAB = N,N′-bis(pentafluorophenyl)-2,3-dimethyl-1,4-diaza-1,3-butadiene; TFA = trifluoroacetate] has been reported to give quantitative yields (with Cu(II) as the limiting reagent) and selectivity combined with turnover numbers >800. This report details mechanistic studies of this catalytic process using a combined experimental and computational approach. Examining catalysis with the complex (^FlDAB)Rh(OAc)(η²-C₂H₄) shows that the reaction rate has a dependence on catalyst concentration between first- and half-order that varies with both temperature and ethylene concentration, a first-order dependence on ethylene concentration with saturation at higher concentrations of ethylene, and a zero-order dependence on the concentration of Cu(II) oxidant. The kinetic isotope effect was found to vary linearly with the order in (^FlDAB)Rh(OAc)(η²-C₂H₄), exhibiting no KIE when [Rh] was in the half-order regime, and a k_H/k_D value of 6.7(6) when [Rh] was in the first-order regime. From these combined experimental and computational studies, competing pathways, which involve all monomeric Rh intermediates and a binuclear Rh intermediate in the other case, are proposed. (Chemical Equation Presented).

Full text of DOI:10.1021/jacs.6b10658

Article
pubs.acs.org/JACS
Mechanistic Studies of Single-Step Styrene Production Using a
Rhodium(I) Catalyst
Benjamin A. Vaughan,† Sarah K. Khani,‡ J. Brannon Gary,‡ James D. Kammert,#
Michael S. Webster-Gardiner,† Bradley A. McKeown,† Robert J. Davis,# Thomas R. Cundari,*,‡
and T. Brent Gunnoe*,†
†Department of Chemistry, University of Virginia, Charlottesville, Virginia 22904, United States
‡Center for Advanced Scientific Computing and Modeling, Department of Chemistry, University of North Texas, Denton, Texas
76203, United States
#Department of Chemical Engineering, University of Virginia, Charlottesville, Virginia 22904, United States
S
* Supporting Information
ABSTRACT: The direct and single-step conversion of benzene, ethylene, and
a Cu(II) oxidant to styrene using the Rh(I) catalyst (FlDAB)Rh(TFA)(η2-
C2H4) [FlDAB = N,N′-bis(pentafluorophenyl)-2,3-dimethyl-1,4-diaza-1,3-buta-
diene; TFA = trifluoroacetate] has been reported to give quantitative yields
(with Cu(II) as the limiting reagent) and selectivity combined with turnover
numbers >800. This report details mechanistic studies of this catalytic process
using a combined experimental and computational approach. Examining
catalysis with the complex (FlDAB)Rh(OAc)(η2-C2H4) shows that the reaction
rate has a dependence on catalyst concentration between first- and half-order
that varies with both temperature and ethylene concentration, a first-order
dependence on ethylene concentration with saturation at higher concen-
trations of ethylene, and a zero-order dependence on the concentration of
Cu(II) oxidant. The kinetic isotope effect was found to vary linearly with the
order in (FlDAB)Rh(OAc)(η2-C2H4), exhibiting no KIE when [Rh] was in the
half-order regime, and a kH/kD value of 6.7(6) when [Rh] was in the first-order regime. From these combined experimental and
computational studies, competing pathways, which involve all monomeric Rh intermediates and a binuclear Rh intermediate in
the other case, are proposed.
the net reaction is the thermodynamically favorable conversion
of benzene, ethylene, and oxygen to styrene and water (eq 1).11
INTRODUCTION
■
Styrene is produced on a scale of ∼18.5 million tons annually
for use in plastics, elastomers, and fine chemicals.1−5 Industrial
synthesis of styrene often involves an acid-catalyzed (i.e.,
Friedel−Crafts or zeolite catalysis) arene alkylation to form
ethylbenzene, trans-alkylation to optimize yield, and ethyl-
benzene dehydrogenation.1−10 Although this method has been
employed by industry for many years, there are some
disadvantages.2 Dehydrogenative addition of an arene C−H
bond across an olefin CC bond (i.e., oxidative arene
vinylation) provides a potential route to directly synthesize
vinyl arenes from arenes, olefins, and oxidants. For styrene
production, one possible route involves metal-mediated
activation of the C−H bond of benzene to yield a M−Ph
bond, ethylene insertion into the resulting M−Ph bond to
produce a M-CH2CH2Ph complex, and β-hydride elimination
from the resulting M-CH2CH2Ph complex to give coordinated
styrene and a M−H bond (Scheme 1). Subsequent styrene
dissociation (regardless of mechanism) and reaction with
oxidant can regenerate the active catalyst, and if the oxidant is
oxygen (either used in situ or used to recycle an in situ oxidant),
While catalysts have been developed for the addition of
unactivated arene C−H bonds (i.e., non-heterofunctionalized
arenes such as benzene, toluene, naphthalene, etc.) across olefin
CC bonds (i.e., catalytic olefin hydroarylation) to afford alkyl
arenes,12−24 few examples of transition metal catalysts for
oxidative arene vinylation using unactivated substrates have
been reported.25−30 Of the reported catalysts, to our knowl-
edge, all suffer from low selectivity and/or low yield/catalytic
turnovers. For example, Hong and co-workers reported a
Received: October 11, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.6b10658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
benzene and ethylene to styrene without the intermediacy of
ethylbenzene, we first sought to confirm that ethylbenzene is
not converted to styrene under catalytic conditions using 1-
TFA (eq 2). Heating a solution of 1-TFA and Cu(OAc)2 in
Scheme 1. Catalytic Cycle for Transition-Metal-Catalyzed
Styrene Production from Benzene, Ethylene, and Oxidant
a
ethylbenzene under 50 psig of ethylene showed no formation of
styrene after 8 h. Isomers of ethyl-vinylbenzene were detected,
but not quantified, by GC/MS.
Apparent Induction Period. Under some conditions,
monitoring the conversion of benzene, ethylene, and Cu(OAc)2
to styrene and CuOAc catalyzed by 1-TFA reveals an apparent
induction period (see Figure 1, for example). We investigated
a
[O] denotes 1 equiv of a two-electron oxidant or 2 equiv of a one-
electron oxidant.
Rh4(CO)12 catalyst that affords 472 turnovers (TO) of styrene;
however, 809 TO of a byproduct (3-pentanone) are also
produced.25 Another example is the work of Sanford and co-
workers, who have reported that (3,5-dichloropyridyl)Pd-
(OAc)2 catalyzes styrene production with 100% selectivity,
but the overall turnover number is low (6.6 TO, 33% overall
yield) and the reaction incorporates the expensive oxidant
t
PhCO3 Bu that cannot be readily regenerated from air or
dioxygen.29 Germane here, Periana and co-workers have
disclosed styrene production using Rh(III) complexes.26
In an initial communication, we reported that the Rh(I)
complex (FlDAB)Rh(TFA)(η2-C2H4) (1-TFA) [FlDAB = N,N′-
bis(pentafluorophenyl)-2,3-dimethyl-1,4-diaza-1,3-butadiene;
TFA = trifluoroacetate] effectively catalyzes the conversion of
benzene and ethylene to styrene in the presence of Cu(II)
oxidants.31 This catalyst affords quantitative yields of styrene
(relative to the Cu(II) limiting reagent) with quantitative
selectivity and gives turnover numbers (TON) > 800. Catalysis
with 1-TFA is also one of the longest-lived catalytic processes
for styrene production using a molecular catalyst, showing
activity for up to 96 h with apparent turnover frequencies
(TOF) on the order of 10−3 s−1. Herein, we report a
mechanistic study of catalysis with 1-TFA.
Figure 1. TO vs time plot for catalysis with 1-TFA. Reaction
conditions: 0.112 mM 1-TFA, 20 mL C6H6, 25 psig ethylene, 13.4
mM Cu(OAc)2 (120 equiv relative to 1-TFA), 150 °C. Each data
point is the average of two independent catalytic reactions, each
analyzed in duplicate by GC/FID. Error bars represent the standard
deviation of all four values.
two possible rationalizations for the observed change in rate of
catalysis: (1) the active catalyst is formed by decomposition of
1-TFA to insoluble Rh nanoparticles that catalyze the
reaction,33 and (2) during the apparent induction period, 1-
TFA converts to (FlDAB)Rh(OAc)(η2-C2H4) (1-OAc), and 1-
OAc catalyzes the reaction at a faster rate than 1-TFA.
Testing for Nanoparticle Formation. A number of tests
for the formation of nanoparticles from the decomposition of
homogeneous complexes are known;34 however, some of these
methods are unsuitable for use with rhodium (e.g., the mercury
drop test, which is unsuitable since Rh does not amalgamate
well).34 One method for the detection of nanoparticles is
transmission electron microscopy (TEM). While it has been
argued that TEM might not provide useful information
regarding the formation of nanoparticles due to its inherent
inability to detect particles below 1 nm in size,34 when coupled
with energy-dispersive X-ray spectroscopy (EDS), which
provides information about elemental composition of samples,
this method can provide evidence for or against the formation
of nanoparticles.
RESULTS AND DISCUSSION
■
Initial studies of arene C−H activation in acidic media by the
cyclooctene variant of 1-TFA revealed fast rates of benzene C−
H activation.32 Since C−H activation is a key step in the
mechanism for oxidative arene vinylation, these results
indicated that 1-TFA might be an effective catalyst for vinyl
arene production. Upon reaction of benzene with ethylene in
the presence of 1-TFA and Cu(II) oxidant, styrene is formed
with high selectivity. As disclosed in a preliminary communi-
cation,31 this catalysis was examined over a range of conditions,
and optimal yields and reaction rates were achieved at 150 °C
using Cu(OAc)2 as the oxidant.
We performed TEM/EDS experiments on samples of
reaction mixtures from catalysis with 1-TFA. Catalytic reactions
with 1-TFA and Cu(OAc)2 were allowed to reach completion,
Since our general mechanistic hypothesis for styrene
production (Scheme 1) involves the direct conversion of
B
DOI: 10.1021/jacs.6b10658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
after which samples were deposited on a grid, allowed to
evaporate, and analyzed by TEM and EDS. While many of the
sample regions chosen for TEM imaging appeared to contain
nanoparticles (Figure 2A), EDS results show that these regions
do not contain Rh. Interestingly, some of the regions that
appeared to be amorphous (Figure 2B) showed significant
concentrations of Rh by EDS. We hypothesized that this was
likely a result of remaining molecular Rh complex on the
sample grid. To test this hypothesis, samples of neat reaction
mixture were decanted to leave only solid (mostly Cu) and
remove any soluble materials. The solid was subsequently
sonicated in dioxane (in which 1-TFA is soluble), decanted,
and the remaining solid was deposited on the grid. Results from
EDS of these washed samples showed no Rh present (Figure
2C). As an additional control, the solid material isolated by
filtration was subjected to catalytic conditions, and styrene
production was not observed after 12 h of heating.
In addition to the TEM/EDS analysis, we performed the
Maitlis filtration test (Figure 3).35 Reactions with 1-TFA and
Figure 3. TO vs time plot with the reaction solution filtered at 3 h.
Reaction conditions: 0.112 mM 1-TFA, 20 mL C6H6, 50 psig ethylene,
13.4 mM Cu(OPiv)2 (120 equiv relative to 1-TFA), 150 °C. Each data
point is the average of two independent catalytic reactions, each
analyzed in duplicate by GC/FID. Error bars represent the standard
deviation of all four values.
Cu(OPiv)2 (OPiv = pivalate) were sampled every 30 mins until
3 h, at which time the reaction mixture was filtered through
Celite and the reactors were recharged with the filtrate and
heated to 150 °C (n.b.: Cu(OPiv)2 was used due to its
solubility in benzene). The filtration should remove insoluble
Rh materials, and a second induction period would be expected
if insoluble materials were serving as the active catalyst. The
results from both the TEM and Maitlis test experiments suggest
that insoluble nanoparticles are not serving as the catalyst, and
therefore, that formation of insoluble Rh nanoparticles is likely
not the reason for the apparent induction period.
Comparison of Catalysis with 1-TFA and 1-OAc.
Operating under the assumption that catalysis proceeds by a
pathway similar to that shown in Scheme 1, we hypothesized
that 1-TFA converts to 1-OAc in situ (e.g., after completing one
catalytic loop and releasing HTFA). To test this hypothesis, 1-
OAc was independently synthesized and tested for catalytic
activity. Under identical conditions as catalysis with 1-TFA, the
reaction with 1-OAc using Cu(OPiv)2 shows no apparent
induction period. Cu(OPiv)2 was used in place of Cu(OAc)2
due to the enhanced solubility of Cu(OPiv)2 and the likely
similar reactivity of Rh-OAc and Rh-OPiv intermediates. This
result is consistent with 1-OAc serving as a more active catalyst
compared to 1-TFA. We compared the rate of catalysis of 1-
OAc to 1-TFA after the apparent induction period using a plot
of [styrene] vs time in which data for 1-OAc are time-shifted (t
Figure 2. TEM images of reaction mixtures from catalysis with 1-TFA.
EDS was performed on circled areas. (A) Unwashed sample, region
with nanoparticle-like structures. (B) Unwashed sample, amorphous
region. (C) Solid from catalytic reaction deposited after washing with
dioxane. Reaction conditions for catalysis: 0.112 mM 1-TFA, 20 mL
C6H6, 50 psig ethylene, 13.4 mM Cu(OAc)2, 150 °C, 12 h.
C
DOI: 10.1021/jacs.6b10658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
= 0.5 h shifted to t = 1.5 h) and overlaid with data from
reactions using 1-TFA as the catalyst precursor (Figure 4). The
Figure 5. TO vs time plot for catalysis with 1-TFA using Cu(OAc)2 or
Cu(TFA)2 as the oxidant. Reaction conditions: 0.112 mM 1-TFA, 20
mL C6H6, 13.4 mM CuX2 (120 equiv relative to 1-TFA), 25 psig
C2H4, 150 °C. Each data point is the average of two independent
catalytic reactions, each analyzed in duplicate by GC/FID. Error bars
represent the standard deviation of all four values.
Rh−Ph bond, (c) β-hydride elimination from the Rh-
CH2CH2Ph intermediate, and (d) styrene dissociation. To
display the results of these calculations in an easily viewed
comparison, in Scheme 2 we set the free energy of each of the
complexes 1-OAc or 1-TFA to zero. Our calculations also
included the energetics of conversion of 1-TFA and 1-OAc to
the dimers [(FlDAB)Rh(μ-X)]2 (X = OAc or TFA; 1D-OAc or
1D-TFA, respectively). Using the B3LYP functional, the dimers
were calculated to be more stable than the monomeric species
1-TFA and 1-OAc by 11.2 and 6.3 kcal/mol, respectively. In
contrast to these calculations, experimental results indicate that
the monomeric complexes are more stable than the dimers. For
example, a C6D6 solution of independently synthesized 1D-
TFA completely converts to 1-TFA in the presence of ethylene
at room temperature in <1 min. As a result of the discrepancy
between experimental and computational results, the con-
versions of monomer to dimer were computationally modeled
using the M06 functional, which gave ΔG values (dimer to
monomer) of −1.8 and 1.5 kcal/mol for 1-TFA and 1-OAc at
423 K, respectively, which are more in line with experimental
observations. The M06 functional is known to provide better
estimates of van der Waals interactions, which we believe gives
a more accurate modeling of the energetics for monomer/
dimer interconversion. However, to assess catalysis based on
monomeric catalysts, we retained the B3LYP data. The
calculations predict ΔG⧧ = 29.6 kcal/mol using 1-OAc and
ΔG⧧ = 42.8 kcal/mol using 1-TFA. These barriers are
consistent with experimental rates. Using the Eyring equation,
kobs values (for details on kobs calculations, vide infra) from
reactions using 1-TFA with Cu(TFA)2 as the oxidant (which
results in the Rh-TFA moiety being retained throughout
catalysis) and 1-OAc with Cu(OAc)2 as the oxidant (which
results in the Rh-OAc moiety being retained throughout
catalysis) give ΔG⧧ values which are in general agreement with
ΔG⧧ values from computations (for Rh-TFA, 42.8 kcal/mol;
for Rh-OAc, 29.6 kcal/mol).
Figure 4. (A) [Styrene] vs time plot for catalysis with 1-TFA and 1-
OAc using Cu(OPiv)2 as the oxidant. Reaction conditions: 0.112 mM
1-TFA or 1-OAc, 20 mL C6H6, 13.4 mM Cu(OPiv)2 (120 equiv
relative to 1-TFA or 1-OAc), 50 psig C2H4, 150 °C. Data for 1-OAc
are offset from t = 0.5 h to t = 1.5 h to overlap with data from 1-TFA,
and non-offset times are labeled in red. Each data point is the average
of two independent catalytic reactions, each analyzed in duplicate by
GC/FID. Error bars represent the standard deviation of all four values.
(B) [Styrene] vs time plot for catalysis with 1-OAc from panel A
without time offset, which is consistent with no induction period.
overlap of the plots shows nearly identical [styrene] versus time
profiles for catalysis following the apparent induction period for
1-TFA, which supports the hypothesis that 1-TFA is converting
to 1-OAc in situ.
We performed catalytic experiments with 1-TFA using both
Cu(TFA)2 and Cu(OAc)2 as oxidants. Plots of TO vs time for
these reactions are shown in Figure 5. For catalysis using
Cu(TFA)2, a consistently slower rate is observed. For the
reaction with Cu(OAc)2, the rate of catalysis gradually
increases, which is consistent with conversion of 1-TFA to 1-
OAc. For reactions using 1-TFA and Cu(TFA)2, we do not
observe induction periods.
Computational Studies: Overview. To gain insight into
the difference in rate of catalysis for 1-TFA compared to 1-
OAc, we performed a computational study (B3LYP) of the
fundamental steps involved in the proposed catalytic cycle
(Scheme 2). The catalytic cycle we modeled involves (a)
benzene coordination to (FlDAB)Rh(η2-C2H4)(X) by net
ligand substitution with ethylene and subsequent benzene C−
H activation, (b) ethylene coordination and insertion into the
From the complexes 1-OAc and 1-TFA, multiple possible
pathways for benzene C−H activation were calculated. The
lowest energy pathway for catalysis by both 1-OAc and 1-TFA
involves initial displacement of ethylene to afford the
corresponding η2-benzene complexes (FlDAB)Rh(OAc)(η2-
C6H6) [2-OAc] and (FlDAB)Rh(TFA)(η2-C6H6) [2-TFA]
D
DOI: 10.1021/jacs.6b10658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 2. Calculated Gibbs Free Energies [B3LYP/LANL2DZ+6-311++G(d,p) in kcal/mol] Including Solvent (SMD-benzene)
and Dispersion Corrections for the Lowest Energy Calculated Pathway for Styrene Production Using Complexes 1-OAc (Shown
a
in Black) and 1-TFA (Shown in Red) at 423.15 K
a
The calculated energies for each reaction are relative to the energy of 1-X (X = OAc or TFA), which is set to zero energy for each reaction.
Stationary points without TFA or OAc are degenerate.
with energies of 3.8 and 10.7 kcal/mol relative to 1-OAc and 1-
TFA, respectively. It is interesting to note that ethylene/
benzene exchange is ∼7 kcal/mol less favorable for 1-TFA than
for 1-OAc, accounting for ∼50% of the calculated overall
ΔΔG⧧ of 13.2 kcal/mol for the catalytic process.
Following the formation of a benzene adduct, the lowest
energy C−H activation pathway for both complexes was found
to involve concerted metalation−deprotonation (CMD) of
benzene using an acetate ligand,36−42 with activation barriers of
21.7 kcal/mol for the Rh-OAc complex and 22.9 kcal/mol for
the Rh-TFA complex relative to 2-OAc and 2-TFA,
respectively. After benzene C−H activation, the coordinated
HX is displaced by ethylene to generate (FlDAB)Rh(Ph)(η2-
C2H4) (4). Comparing the optimized transition-state geo-
metries for CMD of benzene from 2-OAc and 2-TFA (Figure
6), it can be observed that the Rh−C and C−O bond lengths
(2.19 and 1.25 Å, respectively, for the Rh-TFA transition state)
elongate by 0.02 Å for the Rh-OAc transition state. Conversely,
the Rh−O bond length shrinks by 0.02 Å. More noticeable is
the change in the C−H bond distance in the Rh-TFA transition
state, which is 0.06 Å longer than the corresponding bond in
the Rh-OAc transition state; the O−H bond in the Rh-OAc
transition state is elongated by 0.06 Å versus the same bond in
the Rh-TFA transition state. Hence, the transition state for the
Rh-TFA complex appears to be later than the corresponding
transition state for Rh-OAc, which is consistent with the
Hammond postulate, as the ΔH for benzene C−H activation
by 2-TFA is calculated to be more endothermic (ΔH = 17.3
kcal/mol) and endergonic (ΔG = 18.2 kcal/mol) than that of
2-OAc (ΔH = 12.9 kcal/mol, ΔG = 12.0 kcal/mol).
Previous studies of ethylene insertion into Pt(II) hydrocarbyl
and aryl bonds demonstrate the viability of such reactions for
cationic complexes, but the results suggest that the activation
barrier for olefin insertion might be higher for overall charge-
Figure 6. Optimized calculated geometries for the transition states for
benzene C−H activation by (A) (FlDAB)Rh(η2-C6H6)(TFA) (2-
TFA), and (B) (FlDAB)Rh(η2-C6H6)(OAc) (2-OAc). Bond lengths in
Å.
neutral complexes than for cationic complexes.22 Thus, we
anticipated that activation barriers for ethylene insertion into
Rh−Ph bonds of charge-neutral Rh(I) complexes might be
higher than those of related Pt(II) complexes. Previously, we
have shown that the conversion of [(tbpy)Pt(Ph)(THF)]+
(tbpy =4,4′-di-tert-butyl-2,2′-bipyridyl) and ethylene to [(tbpy)-
Pt(CH2CH2Ph)(η2-C2H4)]+ at 23 °C occurs with an
experimentally determined ΔG⧧ of 21.0(1) kcal/mol.17 This
E
DOI: 10.1021/jacs.6b10658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 3. Comparison of Calculated Free Energies (in kcal/mol) for β-Hydride Elimination To Form Styrene and Benzene C−
H Activation To Form Ethylbenzene from (FlDAB)Rh(CH2CH2Ph) (Shown in Black) and (tbpy)Pt(CH2CH2Ph) (Shown in
Red), with the Important Energy Differences between the Two Pathways Highlighted
activation energy is similar to the reported ΔG⧧ of 19.2 kcal/
mol for ethylene insertion into a Pt(II)−hydride bond.43
Through the use of a diimine ligand with electron-withdrawing
perfluorophenyl groups (i.e., FlDAB) we sought to generate a
Rh center with similar properties to cationic Pt(II) such as
[(Xbpy)Pt(Ph)(THF)]+ (Xbpy = 4,4′-X-2,2′-bipyrid-
yl).14,17,19−22 Insertion of ethylene into the Rh−Ph bond of
(FlDAB)Rh(Ph)(η2-C2H4) (4) is calculated to have a free
energy of activation of 20.2 kcal/mol from complex 4, which is
commensurate with the cationic Pt(II) complex (tbpy)Pt(Ph)-
(η2-C2H4), for which the free energy of activation for ethylene
insertion into the Pt−Ph bond was calculated to be 21.5 kcal/
mol. (n.b.: to provide an accurate comparison, the energy
values for Pt complexes17 have been recalculated using the
same computational parameters as have been used for the Rh
complexes presented herein).
the transition state for ethylene insertion is the calculated
highest energy species for the Rh catalysis. Notably, for the Pt
catalysis an inverse dependence on ethylene concentration was
experimentally demonstrated, while for the Rh catalysis a first-
order dependence is observed (vide infra).17
Computational Studies: Selectivity for Styrene. A
challenge for oxidative arene vinylation has been achieving
selectivity for the vinyl arene product. In fact, [(tbpy)Pt(THF)-
(Ph)]+ has been reported to selectively yield ethylbenzene,
even in the presence of oxidants including various Cu(II) salts
(unpublished results).17,20 As noted below, modification of the
donor ability of the 2,2′-bipyridyl ligand coordinated to Pt(II)
can result in the production of styrene, and in one case a few
turnovers are observed with ethylene as the oxidant.20 From a
M-CH2CH2Ph intermediate, ethylbenzene is formed from
benzene C−H activation, whereas styrene is formed from β-
hydride elimination and net styrene dissociation. In order to
understand why the [(tbpy)Pt(Ph)]+ catalyst intermediate is
selective for ethylbenzene production while the (FlDAB)Rh-
(Ph) intermediate is selective for styrene production, we
modeled the likely reaction steps for each product, which are
shown in Scheme 3. The two pathways to form either styrene
or ethylbenzene diverge from the M-CH2CH2Ph intermediate.
In the pathway to form styrene, β-hydride elimination occurs to
give a M(H)(η2-styrene) intermediate that subsequently
liberates styrene. In the pathway to form ethylbenzene, a
second equivalent of benzene coordinates to the coordinatively
unsaturated M−CH2CH2Ph complex, and subsequent benzene
C−H activation affords free ethylbenzene and a new M-Ph
complex.
For catalysis with 1-OAc, the ΔΔG⧧ between C−H
activation and olefin insertion is 4.1 kcal/mol, suggesting that
ethylene insertion is the rate-determining step. For catalysis
with 1-TFA, the ΔΔG⧧ for the same two steps is 9.2 kcal/mol,
and also predicts that ethylene insertion is the rate-determining
step. This is in contrast to Pt(II)- and Ru(II)-based catalysts for
ethylbenzene formation for which C−H activation is calculated
to be rate-determining.13,17 We note that some calculations for
the Ru(II) catalyst indicated ethylene insertion as the rate-
determining step.44 The calculations are consistent with a
kinetic advantage for 1-OAc over 1-TFA. The calculated ΔΔG⧧
for catalysis using 1-OAc vs 1-TFA is 13.2 kcal/mol, and
computational modeling suggests that the difference in ground-
state energies of 1-OAc and 1-TFA greatly influences the
difference in rate of catalysis. The penultimate step in the
pathway for styrene production is β-hydride elimination, which
was calculated to have an activation barrier of only 0.3 kcal/mol
relative to (FlDAB)Rh(CH2CH2Ph) (5).
For the [(tbpy)Pt]+ catalyst, the difference in energy between
these two pathways was calculated to be 4.5 kcal/mol. For the
Rh complex, the energy difference between the two pathways is
12.6 kcal/mol. Thus, the calculations predict a greater
predilection toward styrene formation for (FlDAB)Rh vs
[(tbpy)Pt]+, which is consistent with experimental results.17
Interestingly, the calculations reveal that energy differences for
benzene C−H activation (ΔΔG⧧ = 2 kcal/mol) and β-hydride
elimination (ΔΔG⧧ = 0.2 kcal/mol), the two key steps for the
The calculations predict different rate-determining steps for
ethylene hydrophenylation by [(tbpy)Pt(Ph)(THF)]+ and
oxidative hydrophenylation of ethylene by 1-TFA/1-OAc.
For the Pt catalyst, the transition state for benzene C−H
activation is calculated to be the highest energy species, while
F
DOI: 10.1021/jacs.6b10658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
formation of ethylbenzene and styrene, respectively, are small
and likely cannot be used to rationalize the switch in selectivity
between (FlDAB)Rh and [(tbpy)Pt]+. Rather, it appears that the
propensity toward styrene dissociation is the key parameter,
with calculated ΔG values for styrene dissociation of 11.8 kcal/
mol for (FlDAB)Rh and a surprisingly much larger 27.4 kcal/
mol for [(tbpy)Pt]+.
Thus, we propose that both (FlDAB)Rh and [(tbpy)Pt]+
likely undergo β-hydride elimination from M−CH2CH2Ph
intermediates, but for [(tbpy)Pt]+ this process is reversible
because the energetics for styrene dissociation are too
unfavorable. We have reported evidence for reversible β-
hydride elimination for ethylbenzene formation by TpRu-
(CO)(NCMe)(Ph) (Tp = hydridotris(pyrazolyl)borate).13 We
note that the formation of diethylbenzenes by [(tbpy)Pt]+ is
explained by a similar rationale.17 That is, dissociation of
ethylbenzene from [(tbpy)Pt(Ph)(ethylbenzene)]+ is slow
(relatively), which allows a second arene C−H activation
(ultimately leading to diethylbenzenes) to compete with the
dissociation of ethylbenzene. More difficult to explain are the
calculations for [(tbpy)Pt]+, which predict styrene formation
over ethylbenzene formation with a ΔΔG⧧ of 4.5 kcal/mol,
which can arise from the functional and basis sets chosen.
Previous calculations and experiments on dipyridyl-supported
Pt(II) catalysts17,19−21 show that a subtle balance of factors
discriminate ethylbenzene from styrene formation. Nonethe-
less, the present calculations are consistent with the
experimental observation of much greater propensity for the
latter for Rh(I) versus Pt(II).
Kinetic and Mechanistic Studies. We studied the kinetics
of catalysis using 1-OAc and copper(II) pivalate [Cu(OPiv)2].
The oxidant Cu(OPiv)2 was used due to its solubility in
benzene. The concentration of ethylene dissolved in benzene
was determined according to the method described by Holder
and Macauley.45 To determine the order of the reaction in 1-
OAc, ethylene, and Cu(OPiv)2, three sets of kinetic experi-
ments were conducted in which the concentrations of two of
the compounds were held constant while one was varied.
Observed rate constants (kobs) were extracted from linear fits of
the initial rate regime (linear region where likely no catalyst
deactivation was occurring) of [styrene] vs time plots. Due to
the complex nature of Rh in solution, the slopes of linear fits
were used as kobs without extracting the [Rh]. Log−log plots
were used to determine the order of the reaction in each of the
compounds examined (Figure 7).
A first-order dependence on ethylene concentration is
observed over a concentration range from 79 mM to 237
mM (Figure 7B). While the first-order dependence on ethylene
is consistent with the previously reported dependence of
apparent TOF on ethylene pressure for catalysis with 1-TFA,31
it is in contrast to previously reported Pt(II)- and Ru(II)-based
hydroarylation catalysts. The rate of catalysis with these
complexes shows an inverse dependence on the concentration
of ethylene, which is due to the formation of M(CH2CH2Ph)-
(η2-C2H4) intermediates (M = TpRu(CO) or [(tbpy)Pt]+),
which were identified as off-cycle resting states.13,15,17 An Ir-
based olefin hydroarylation catalyst shows a bell-curve
dependence on the concentration of ethylene, exhibiting first-
order kinetics at low concentrations and inverse first-order
kinetics at higher concentrations.24 The observed first-order
dependence on [C2H4] for catalysis with 1-OAc is consistent
with a different resting state from our Pt(II) and Ru(II)
catalysts. In fact, for catalysis with 1-OAc, the energy of the
Figure 7. Log−log plots of observed rate constants as a function of
(A) concentration of 1-OAc (slope = 0.66(2), R2 = 0.99). Reaction
conditions: 20 mL C6H6, 26.9 mM Cu(OPiv)2, 50 psig C2H4, 150 °C.
(B) Concentration of C2H4 (slope = 1.02(8), R2 = 0.99). Reaction
conditions: 0.11 mM 1-OAc, 20 mL C6H6, 13.4 mM Cu(OPiv)2, 150
°C. (C) Concentration of Cu(OPiv)2 (slope = −0.17(2), R2 = 0.99).
Reaction conditions: 0.11 mM 1-OAc, 20 mL C6H6, 50 psig C2H4, 150
°C. Each data point represents the average of two independent
catalytic reactions, each analyzed in duplicate by GC/FID. Error bars
represent the standard deviation of all four values.
(FlDAB)Rh(CH2CH2Ph)(η2-C2H4) complex, which is the
analog of the proposed resting states for the Pt(II) and Ru(II)
catalysts,13,17 is calculated to be higher than that of 1-OAc by
7.6 kcal/mol.
G
DOI: 10.1021/jacs.6b10658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
The log−log plot for Cu(OPiv)2 (Figure 7C) shows a near
zero-order dependence {slope = −0.17(2)} on [Cu(OPiv)2]
over a concentration range from 13 mM to 54 mM. Given that
the involvement of Cu(II) occurs after the proposed rate-
limiting step (vide infra), a zero-order dependence is expected.
This type of mechanism has been reported for similar oxidative
processes, and has been shown to exhibit a zero-order
dependence on oxidant.46 The slight negative slope indicates
that Cu(II) might suppress catalytic activity. The need for
Cu(II) as an oxidant, but the apparent need of access to low
valent Rh(I) for C−H activation sets up a difficult balance.
Thus, one possible explanation of the slight inverse dependence
on Cu(II) concentration is that Cu(II) oxidizes a Rh(I)
intermediate, pulling it out of the catalytic cycle, but that the
equilibrium favors Rh(I), but the slight magnitude of the
apparent inhibition makes it difficult to draw a definitive
conclusion.
per-protio styrene (m/z = 104) and styrene-d5 (m/z = 109) by
mass spectrometry. In a separate series of reactions, catalysis
was performed in C6H6 and C6D6 independently (Scheme 5). A
Scheme 5. Kinetic Isotope Experiment Using Independent
a
Reactions in C6H6 and C6D6
The log−log plot for 1-OAc reveals a slope of 0.66(2), which
is between first-order and half-order (Figure 7A). Further
studies showed that the dependence of the reaction rate on 1-
OAc is complicated, varying between first- and half-order as a
function of both the concentration of C2H4 and temperature
(see Supporting Information for plots). The order in Rh was
determined over a concentration range from 0.056 to 0.23 mM
at three different concentrations of C2H4, and, in a separate series
of experiments, the order in 1-OAc was determined at three
different temperatures. With [C2H4] variation, the order in Rh
varied exhibiting close to half-order kinetics at lower [C2H4] {at
7.9 mM C2H4, reaction order in 1-OAc = 0.58(3)} and close to
first order kinetics at higher [C2H4] {at 17.5 mM C2H4,
reaction order in 1-OAc = 0.96(1)}. The order in Rh was found
to vary inversely with temperature, exhibiting closer to first-
order kinetics at lower temperatures {at 130 °C, order in 1-
OAc = 0.83(1)} and closer to half-order kinetics at higher
temperatures {at 160 °C, order in 1-OAc = 0.64(1)}.
Limitations of the catalytic process make kinetic studies outside
of this temperature range challenging. These results indicate the
possibility of competing catalytic pathways (vide infra).
a
kH and kD values were determined using the method of initial rates for
two independent catalytic reactions each, with all samples analyzed in
duplicate by GC/FID. The reported error represents the propagated
standard deviation of all values and linear regressions.
kH/kD of 3.3(2) was determined for these reactions using kobs
values calculated after 1 h of reaction. The two kH/kD values are
statistically identical (for more discussion of KIEs for catalysis,
vide infra). The KIEs are consistent with reported values for
C−H activation by d8 complexes, which often exhibit KIE
values of ≥2.5.47,48 It is also important to note that d6−8
products were not observed except those predicted by the
natural abundance of deuterium in ethylene.
We also sought to probe the reversibility of the benzene C−
H activation step through benzene H/D exchange experiments.
A solution of 1-OAc (0.001 mol %) and Cu(OPiv)2 (120
equiv) in a 1:1 molar mixture of C6H6 and C6D6 was heated to
150 °C. The isotopic distribution of benzene was determined
by GC/MS for the initial solution prior to heating, after 4 h of
heating, and after 24 h of heating. No change was observed in
the isotopic distribution. To determine whether the acetic acid
generated in the catalytic cycle could contribute to the
reversibility of benzene C−H activation, a solution of 1-OAc
(0.001 mol %), Cu(OPiv)2 (120 equiv), and CD3CO2D (500
equiv) in C6H6 was heated to 150 °C. The isotopic distribution
of benzene was determined by GC/MS for the initial solution
prior to heating, after 4 h of heating, and after 24 h of heating
with no change observed.
The kinetic isotope effect (KIE) for catalysis using C6H6 vs
C6D6 was determined using two methods. In one experiment,
catalysis was performed using an equimolar mixture of per-
protio and per-deutero benzene (Scheme 4). After 30 min, a
kH/kD of 3.0(1) was observed upon comparison of the peaks for
Scheme 4. Kinetic Isotope Experiment Using a 1:1 Molar
a
Mixture of C6H6:C6D6
To determine if the equilibrium between monomer and
dimer plays a role in benzene C−H activation, ethylene was
added to H/D exchange reactions as experiments show that
only monomer exists in the presence of ethylene. A solution of
1-OAc (0.001 mol %), Cu(OPiv)2 (120 equiv), ethylene (50
psig), and CD3CO2D (500 equiv) in C6H6 was heated to 150
°C. The isotopic distribution of benzene was determined by
GC/MS for the initial solution prior to heating, after 4 h of
heating, and after 24 h of heating, and no change was observed
over time. The isotopic distribution of the reaction product,
styrene was also determined by GC/MS. Comparing the
isotopic distribution in the reaction after 4 and 24 h to the MS
of an authentic standard, the amount of styrene-d1 increased by
∼5% over the course of the reaction. This indicates that the
acid could be catalyzing H/D exchange with styrene, but the
a
kH/kD value represents the average of three independent catalytic
reactions, each analyzed in triplicate by GC/MS. Reported error
represents the standard deviation of all nine values.
H
DOI: 10.1021/jacs.6b10658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
lack of deuterium incorporation into the benzene indicates that
the C−H activation step is likely irreversible.
would give rise to half-order kinetics with respect to Rh, and
the pathway that does not would give rise to first-order kinetics
with respect to Rh.
Wang and co-workers proposed that a Rh(I) catalyst
precursor was rapidly converted to catalytically active Rh(III)
in the presence of a Cu(II) oxidant and O2.49 To probe
whether a similar reaction was occurring under our catalytic
conditions, the complex (FlDAB)RhCl3(η2-C2H4) was synthe-
sized in situ and tested for catalytic activity. Heating a solution
of FlDAB (0.001 mol %), RhCl3 (0.001 mol %), Cu(OPiv)2
(120 equiv), and ethylene (50 psig) in benzene at 150 °C
afforded no styrene production after 24 h, suggesting that
Rh(III) cannot catalyze styrene production under our
conditions.
A rate law for the cycle shown in Scheme 6 was derived using
the steady-state approximation (full derivation given in the
Supporting Information). The two divergent pathways give rise
to two terms in the rate law, one of which describes first-order
behavior with respect to Rh and the other of which describes
half-order behavior. Considering the limiting forms of the rate
law where [C6H6] is substantial predicts two kinetic regimes,
one that is operative at low [C2H4] in which the first-order
term cancels, predicting a half-order dependence on [Rh], first-
order dependence on ethylene, zero-order dependence on Cu,
and an inverse dependence on HX. The other that is operative
at high [C2H4], in which the half-order term cancels, predicts a
first-order dependence on [Rh] and saturation in all other
components except benzene. Under catalytic conditions, the
observed rate is likely a convolution of both the half- and first-
order terms, as observed in the experiment where a change in
the order in [Rh] is observed at different [C2H4] (vide supra).
Also, at low [C2H4], an assumption is made that [Rh]1/2
dominates the [Rh]Tot term, which gives rise to the half-order
dependence on Rh and is consistent with experimental
observations (see above).
Proposed Mechanism for the Catalytic Cycle. Based on
the experimental and computational results, a cycle for styrene
production using 1-OAc is proposed in Scheme 6. Benzene
Scheme 6. Proposed Mechanism and Rate Law for Catalysis
with 1-OAc
Based on this derivation, the rate constant for C−H
activation (k3) is not included in the low [C2H4] limiting
form of the rate law, which indicates that catalysis in this regime
should not exhibit a KIE. Thus, the KIE for the overall reaction
would be expected to decrease as the order in [Rh] approaches
half-order. In contrast, at higher concentrations of C2H4 a KIE
is expected as the limiting form of the rate law contains k3. To
probe this, catalysis was performed in an equimolar mixture of
per-protio and per-deutero benzene at 35 and 150 psig
ethylene, and the resulting kH/kD values were compared to
the data from the initial KIE experiments at 50 psig ethylene
(vide supra). A plot of observed KIE vs order in [Rh] (Figure 8,
vide supra for discussion of variable order of the reaction in
[Rh]) shows a linear correlation that predicts a KIE of 1.2(6) in
the half-order regime (extrapolated from the linear fit) and a
coordinates to 1-OAc, displacing ethylene to form 2-OAc. The
benzene C−H bond is cleaved in an irreversible reaction,
consistent with lack of H/D exchange into benzene (vide
supra), to afford (FlDAB)Rh(Ph)(HOAc) (3-OAc). Acetic acid
is displaced by ethylene to form (FlDAB)Rh(Ph)(η2-C2H4) (4),
which inserts ethylene to afford (FlDAB)Rh(CH2CH2Ph) (5).
Rapid, irreversible β-hydride elimination occurs to give
(FlDAB)Rh(H)(η2-styrene) (6). Two divergent pathways
from 6 are proposed, one in which styrene dissociates and
the Rh complex reacts with the Cu(II) oxidant to regenerate
the dimer 1D-OAc (favored at low [C2H4]), and another in
which styrene is displaced by ethylene (likely via associative
ligand exchange) before reaction with the oxidant (favored at
high [C2H4]). The pathway that proceeds through 1D-OAc
Figure 8. Plot of kH/kD vs order in [Rh] (R2 = 0.99). Reaction
conditions: 0.11 mM 1-OAc, 10 mL 1:1 C6H6 and C6D6, 26.8 mM
Cu(OPiv)2, 35−150 psig C2H4, 150 °C. Order in [Rh] (and the
corresponding horizontal error bars) was determined using the data
shown in Figure S13. Each data point represents the average of three
independent catalytic reactions, each analyzed in triplicate by GC/MS.
Vertical error bars represent the standard deviation of all nine values.
I
DOI: 10.1021/jacs.6b10658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
KIE of 6.7(6) in the first-order regime. Within deviation of the
data, these results are consistent with the proposed reaction
pathways.
To further probe our proposed mechanism, experiments
were run at high [C2H4] in an attempt to observe saturation
kinetics (Figure 9). At concentrations less than 237 mM, a first-
not observed under the catalytic conditions we tested. This is
likely due to the low concentration of acid in solution under the
conditions tested. The limiting form for the high [C2H4]
regime predicts a zero-order dependence on acid (HX), so
catalysis was performed at 400 psig C2H4 with 0, 500, and 1000
equiv of AcOH added (relative to Rh). At this high
concentration of ethylene, the observed rate of reaction
without added AcOH is statistically identical to that of catalytic
experiments with added acid, which is consistent with the
proposed reaction pathway. Rationalization of the change in the
dependence on Rh as temperature is variation is more difficult
since the route that is first-order in Rh involves a series of ill-
defined reactions with Cu(II). But, it is not surprising that the
contribution of the two competing pathways varies with
temperature.
SUMMARY AND CONCLUSIONS
■
The complex (FlDAB)Rh(OAc)(η2-C2H4) is an effective
catalyst for the direct conversion of benzene, ethylene and
Cu(II) oxidant to styrene. This Rh catalyst precursor is unique
for its selectivity for styrene formation as well as catalyst
longevity. The mechanism of this reaction has been studied and
compared to both electron-rich Ru(II) and related electrophilic
cationic Pt(II) catalysts. Based on these results, the following
conclusions can be drawn:
Figure 9. Plot of kobs vs [C2H4]. Reaction conditions: 0.11 mM 1-
OAc, 20 mL C6H6, 13.4 mM Cu(OPiv)2, 150 °C. Each data point
represents the average of two independent catalytic reactions, each
analyzed in duplicate by GC/FID. Error bars represent the standard
deviation of all four values.
(1) Combined experimental and computational studies are
consistent with catalysis that proceeds through benzene
C−H activation, displacement of coordinated acetic acid
by ethylene, rate-limiting insertion of ethylene into the
Rh−Ph bond, and β-hydride elimination followed by
liberation of styrene and reaction with Cu(II) (vide infra,
comment (3), for additional comments).
(2) The apparent induction period observed for catalysis
with (FlDAB)Rh(TFA)(η2-C2H4) is not likely the result
of in situ formation of insoluble Rh nanoparticles as the
active catalyst. Rather, (FlDAB)Rh(TFA)(η2-C2H4) con-
verts to (FlDAB)Rh(OAc)(η2-C2H4) in situ, which
catalyzes styrene production at a faster rate than the
TFA analog. The difference in rate of catalysis for 1-OAc
and 1-TFA is likely a result of the difference in ground
state energies rather than OAc/TFA influence on
transition states, since the proposed rate-determining
step occurs once the carboxylate ligands are completely
dissociated.
order dependence on ethylene is observed, but above 237 mM,
saturation is observed. In an attempt to observe the inverse
dependence on acid (HX) predicted by the proposed rate law,
catalysis with 1-OAc was performed in the presence of added
AcOH (500 and 1000 equiv relative to 1-OAc). Using initial
rates, we found that the rate of catalysis is suppressed by AcOH
(Figure 10). This is consistent with the limiting form for the
low [C2H4] regime.
It is important to note that while the inverse dependence on
acid would predict inhibition as the reaction proceeds, this is
(3) A mechanism for catalysis with (FlDAB)Rh(X)(η2-C2H4)
(X = OAc or TFA) has been proposed that involves two
pathways whose contributions vary with [C2H4] and
temperature. Accordingly, the derived rate law predicts
different behavior at low and high [C2H4]. At low
[C2H4], the limiting form predicts half-order in [Rh],
first-order in ethylene, zero-order in Cu, and an inverse
dependence on acid concentration. At high [C2H4], the
limiting form predicts first-order in [Rh], and zero-order
dependence on all others. Saturation behavior has been
observed for ethylene, and the rate of reaction is
suppressed in the presence of added acid at low
[C2H4] but not at high [C2H4], both results are
consistent with the proposed mechanism and rate law.
The kinetics are also consistent with computational
modeling, which predicts rate-limiting ethylene insertion
into the Rh−Ph bond. However, alternative explanations
of the kinetic data are possible, such as inhibition of
Figure 10. [Styrene] vs time plot for the initial rate regime of catalysis
with 1-OAc upon the addition of 0 equiv (slope = 0.005, R2 = 0.99),
500 equiv (slope = 0.0016, R2 = 0.99), and 1000 equiv (slope =
0.0008, R2 = 0.99) of AcOH relative to the concentration of 1-OAc.
Reaction conditions: 20 mL C6H6, 13.4 mM Cu(OPiv)2, 50 psig C2H4,
AcOH (0, 500, or 1000 equiv relative to 1-OAc), 150 °C. Each data
point represents the average of two independent catalytic reactions,
each analyzed in duplicate by GC/FID. Error bars represent the
standard deviation of all four values.
J
DOI: 10.1021/jacs.6b10658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
along with the LANL2DZ pseudopotential basis set for Rh and the 6-
31G(d) basis set for main group elements; the 6-311++G(d,p)
pseudopotential basis set was used on main group elements for single
point calculations at the B3LYP/LANL2DZ+6-31G(d) stationary
points. Free energies are reported in kcal/mol at 423.15 K assuming a
pressure of 1 atm. Unscaled B3LYP/LANL2DZ+6-31G(d) vibrational
frequencies were used for the enthalpic and entropic corrections. The
GD3BJ dispersion correction58 and the SMD solvation model59 were
utilized in the presence of benzene as the continuum solvent on the
single-point calculations. Stationary points were differentiated as
minima or transition states based on the number of imaginary
frequencies (zero or one, respectively) calculated based on the energy
Hessian.
catalysis upon reaction of the catalyst with acid (to
remove Rh from the catalytic cycle) that is influenced by
concentration of ethylene. Despite other possible
explanations for this complex catalytic process, our
mechanistic proposal is consistent with both experimen-
tal and computational data and, we believe, is the
mechanistic model that is most consistent with the data.
(4) The selectivity for styrene production (versus ethyl-
benzene) using (FlDAB)Rh(X)(η2-C2H4) appears to
result from more favorable styrene dissociation compared
to related cationic Pt(II) catalysts.
The monomer/dimer equilibrium of both 1-TFA and 1-OAc was
also examined using DFT. All stationary points were obtained using
the M0660 functional along with the LANL2DZ pseudopotential basis
set for Rh and the 6-31G(d) basis set for main group elements; the 6-
311++G(d,p) pseudopotential basis set was used on main group
elements for single-point calculations at the M06/LANL2DZ+6-
31G(d) stationary points. Free energies are reported in kcal/mol at
423.15 K, assuming a pressure of 1 atm. Unscaled M06/LANL2DZ+6-
31G(d) vibrational frequencies were used for the enthalpic and
entropic corrections. The SMD solvation model59 was utilized in the
presence of benzene as the continuum solvent on the single-point
calculations.
EXPERIMENTAL SECTION
■
General Considerations. All manipulations were performed
under an atmosphere of dry nitrogen using standard Schlenk or
high-vacuum techniques and/or in a Vac Atmospheres Dri-Lab
glovebox equipped with a Dri-Train MO-41 purifier. Dry, oxygen-
free solvents were employed throughout. Benzene was dried by
passage through columns of activated alumina. THF was dried by
passage through columns of activated alumina, followed by distillation
from sodium benzophenone ketyl. Pentane was dried over sodium
benzophenone ketyl. Deuterated solvents were purchased from
Cambridge Isotope Labs, degassed, and dried over molecular sieves.
NMR spectra were recorded on a Varian NMRS 600 MHz NMR
spectrometer (19F, 564.33 MHz operating frequency), Bruker Avance
III 600 MHz NMR spectrometer (1H, 600.13 MHz operating
frequency), or a Bruker Avance III 800 MHz NMR spectrometer
(13C, 201.27 MHz operating frequency), and are reported with
reference to residual solvent resonances. GC/MS was performed using
a Shimadzu GCMS-QP2010 Plus system with a 30 m × 0.25 mm
SHRXI-5MS column with 0.25 μm film thickness using electron
impact (EI) ionization. GC/FID was performed using a Shimadzu
GC-2014 system with a 30 m × 0.25 mm HP5 column with 0.25 μm
film thickness. Infrared spectra were collected on a Shimadzu
IRAffinity-1 FT-IR instrument using KBr pellets.
Styrene production was quantified using linear regression analysis of
gas chromatograms of standard samples of authentic product. A plot of
peak area ratios versus molar ratios gave a regression line. For the GC/
FID system, the slope and correlation coefficient of the regression line
were 1.34 and 0.99, respectively. Ethylene was purchased in gas
cylinders from GTS-Welco and used as received. All other reagents
were purchased from commercial sources and used as received.
[Rh(η2-C2H4)2(μ-Cl)]2 was prepared according to literature proce-
dures.50 [Rh(η2-C2H4)2(μ-TFA)]2 was prepared using an adaptation of
literature procedures, substituting AgTFA for TlTFA.51 While the
synthesis of [Rh(η2-C2H4)2(μ-OAc)]2 has been reported previously,52
higher yields were obtained by using the same method as for the
synthesis of [Rh(η2-C2H4)2(μ-TFA)]2,51 substituting TlOAc for
TlTFA. FlDAB was synthesized according to literature procedures.53
Copper(II) pivalate [Cu(OPiv)2] was synthesized according to
literature procedures.54 Procedures for catalytic reactions using 1-
TFA have been reported previously.31 The Maitlis filtration test was
performed according to the previously reported procedure.35
Synthesis of (FlDAB)Rh(OAc)(η2-C2H4) [1-OAc]. To a stirring
solution of [Rh(η2-C2H4)2(μ-OAc)]2 (200 mg, 0.459 mmol) in THF
(20 mL), FlDAB (382 mg, 0.918 mmol) was added, and the mixture
was stirred for 30 min. The solvent was removed in vacuo, and the
purple solid was washed with pentane (ca. 80 mL) and dried to afford
1 as a purple powder (218 mg, 0.360 mmol, 40%). Upon prolonged
drying in vacuo, the ethylene is removed to form [(FlDAB)Rh(μ-
OAc)]2 (1D-OAc), the characterization data for which are as follows.
1H NMR (600.13 MHz, C6D6): δ 1.75 (s, 6H, COOCH3), −0.95 (s,
12H, DAB-CH3). 13C NMR (201.27 MHz, C6D6): δ 186.0 (s,
COOCH3), 163.9 (s, CN), 141.2 (m, C6F5-meta C), 140.0 (m,
C6F5-meta C), 138.9 (m, C6F5-ortho C), 137.7 (m, C6F5-ortho C),
131.0 (m, C6F5-para C), 130.1 (s, C6F5-ipso C), 22.8 (s, COOCH3),
19.7 (s, DAB-CH3). 19F NMR (564.33 MHz, C6D6): δ −144.0 (d, 3JFF
3
= 23 Hz, C6F5-ortho F), −151.3 (d, JFF = 23 Hz, C6F5-ortho F),
−156.6 (t, 3JFF = 23 Hz, C6F5-meta F), −163.1 (t, 3JFF = 23 Hz, C6F5-
meta F), −165.9 (t, 3JFF = 23 Hz, C6F5-para F). IR (KBr): 1530, 1497
cm−1 (O2CMe sym and asym). Upon pressurizing with ethylene, 1-
OAc can be regenerated, and its in situ characterization data are as
1
follows: H NMR (600.13 MHz, C6D6): δ 2.5 (broad s, 4H, C2H4),
1.68 (s, 3H, COOCH3), − 1.79 (s, 6H, DAB-CH3).
Reaction of 1-TFA with Ethylbenzene. A stock solution
containing 1-TFA (0.112 mM), decane (10 equiv relative to 1-
TFA), and ethylbenzene (50 mL) was prepared in a volumetric flask.
Fisher-Porter reactors were charged with stock solution (20 mL) and
copper acetate (13.4 mM). The vessels were sealed, pressurized with
ethylene (50 psig), and stirred while heated to 150 °C. The reactions
were sampled at 4, 8, 12, and 24 h. At each time point the reactors
were cooled to room temperature, sampled, recharged with ethylene,
and reheated. Aliquots of the reaction mixture were analyzed by GC/
MS, and products were identified using through a search of the NIST
mass spectral database.
Procedure for TEM/EDS Experiments. Transmission electron
microscopy images were obtained using a JEOL 2000FX-II electron
microscope equipped with a Gresham high-angle X-ray detector. An
accelerating voltage of 200 kV was used with a resolution of 1.4 Å.
Energy-dispersive X-ray spectroscopy (EDS) data were used to detect
the presence of Rh and Cu, and were collected using an incident beam
diameter of approximately 25 nm. Samples for electron microscopy
were prepared by ultrasonic dispersion in 1,4-dioxane followed by
deposition onto a holey carbon film supported on a 400 mesh Cu grid.
Samples were allowed to dry in air for 24 h before insertion into the
microscope vacuum chamber.
Attempted Oxidative Hydrophenylation of Ethylene Using
Washed Materials Recovered from Catalysis with 1-TFA and
Cu(OAc)2. Two duplicate catalytic reactions with 1-TFA were run to
completion. The resulting mixtures were combined and dried in vacuo.
The solid was sonicated in 1,4-dioxane for 30 min, after which the
solution was decanted leaving only insoluble materials, which were
dried in vacuo. Two Fisher-Porter reactors were each charged with half
of the recovered material, decane (4 μL, 0.022 mmol), benzene (20
mL), and Cu(OAc)2 (49 mg, 0.27 mmol). The vessels were sealed,
pressurized with ethylene (50 psig), and stirred while heated to 150
°C. The reactions were sampled at 4 and 24 h. At each time point the
reactors were cooled to room temperature, sampled, recharged with
ethylene, and reheated. Aliquots of the reaction mixture were analyzed
Computational Methods. Density functional theory (DFT) was
applied to the study of styrene formation using both 1-TFA and 1-
OAc. The Gaussian 09 package was used to perform all simulations.55
All stationary points were obtained using the B3LYP56,57 functional
K
DOI: 10.1021/jacs.6b10658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
by GC/FID using relative peak area vs an internal standard (decane).
No styrene was observed after 24 h of heating.
different pressures were used, all of which correspond to 114 mM
C2H4: 30 psig for 130 °C reactions, 50 psig for 150 °C reactions, or 60
psig for 160 °C reactions; 2 reactors per pressure at each Rh
concentration), and stirred while heated to 130 °C, 150 °C, or 160 °C.
The reactions were sampled every 30 min until complete. At each time
point the reactors were cooled to room temperature, sampled,
recharged with ethylene, and reheated. Aliquots of the reaction
mixture were analyzed by GC/FID using relative peak area vs an
internal standard (decane).
H/D Exchange Experiments1:1 Ratio of C6H6 to C6H6. A
stock solution containing 1-OAc (0.112 mM), Cu(OPiv)2 (13.4 mM),
and a 1:1 molar mixture of C6H6 and C6D6 (25 mL) was prepared in a
volumetric flask. PTFE-valved reaction tubes were charged with stock
solution (10 mL), sealed, and heated to 150 °C. The reactions were
sampled at 4 and 24 h. At each time point the reactors were cooled to
room temperature, sampled, and reheated. Aliquots of the reaction
mixture were analyzed by GC/MS. The isotopic distribution of
benzene-dn (n = 0−6) was measured for the initial stock solution and
at each time point.
H/D Exchange Experiments500 equiv CD3COOD. A stock
solution containing 1-OAc (0.112 mM), Cu(OPiv)2 (13.4 mM),
CD3COOD (56 mM), and C6H6 (25 mL) was prepared in a
volumetric flask. PTFE-valved reaction tubes were charged with stock
solution (10 mL), sealed, and heated to 150 °C. The reactions were
sampled at 4 and 24 h. At each time point the reactors were cooled to
room temperature, sampled, and reheated. Aliquots of the reaction
mixture were analyzed by GC/MS. The isotopic distribution of
benzene-dn (n = 0−6) was measured for the initial stock solution and
at each time point.
H/D Exchange Experiments500 equiv CD3COOD with
Added Ethylene. A stock solution containing 1-OAc (0.112 mM)
and C6H6 (25 mL) was prepared in a volumetric flask. Fisher-Porter
reactors were charged with stock solution (10 mL), copper pivalate
(13.4 mM), and CD3COOD (56 mM). The vessels were sealed,
pressurized with ethylene (50 psig), and stirred while heated to 150
°C. The reactions were sampled at 4 and 24 h. At each time point the
reactors were cooled to room temperature, sampled, and reheated.
Aliquots of the reaction mixture were analyzed by GC/MS. The
isotopic distribution of benzene-dn (n = 0−6) was measured for the
initial stock solution and at each time point. The isotopic distribution
of styrene-dn (n = 0−8) was also measured at each time point.
Kinetic Isotope Effect Experiments using a 1:1 Molar
Mixture of C6H6 and C6D6. A stock solution containing 1-OAc
(0.112 mM) and a 1:1 molar mixture of C6H6 and C6D6 (100 mL) was
prepared in a volumetric flask. For reactions at 35 and 50 psig, Fisher-
Porter reactors were charged with stock solution (10 mL) and
Cu(OPiv)2 (240 equiv relative to 1-OAc). The vessels were sealed,
pressurized with ethylene (50 psig), and stirred while heated to 150
°C. The reactions were sampled at 30 min and at 1, 2, and 3 h. At each
time point the reactors were cooled to room temperature, sampled,
recharged with ethylene, and reheated. For reactions at 150 psig,
stainless steel high-pressure reactors equipped with glass liners were
charged with stock solution (10 mL) and Cu(OPiv)2 (240 equiv
relative to 1-OAc). The vessels were sealed, pressurized with ethylene
(150 psig), and stirred while heated to 150 °C. The reactions were
sampled at 30 min, 1, 2, and 3 h. At each time point the reactors were
sampled at temperature using a narrow bore dip tube. Aliquots of the
reaction mixture were analyzed by GC/MS. KIE was determined by
examining the ratio of styrene (m/z = 104) to styrene-d5 (m/z = 109)
in the mass spectrum, accounting for the initial isotopic distribution
and natural abundance. No change in the isotopic distribution for
benzene was observed over the course of the reaction, and the
observed isotopic distribution of product was consistent with the initial
distribution. No d6−8 products were observed, except those predicted
by the natural abundance of deuterium in ethylene.
Oxidative Hydrophenylation of Ethylene Using 1-OAc. A
representative catalytic reaction is described. A stock solution
containing 1-OAc (3.4 mg, 0.0056 mmol), decane (11 μL, 0.056
mmol), and benzene (50 mL) was prepared in a volumetric flask.
Fisher-Porter reactors were charged with stock solution (20 mL) and
Cu(OAc)2 (49 mg, 0.27 mmol). The vessels were sealed, pressurized
with ethylene (25 psig), and stirred while heated to 150 °C. The
reactions were sampled every hour until complete. At each time point
the reactors were cooled to room temperature, sampled, recharged
with ethylene, and reheated. Aliquots of the reaction mixture were
analyzed by GC/FID using relative peak area vs an internal standard
(decane).
Oxidative Hydrophenylation of Ethylene as a Function of
[C2H4]. A stock solution containing 1-OAc (0.112 mM), decane (10
equiv relative to 1-OAc), and benzene (250 mL) was prepared in a
volumetric flask. Fisher-Porter reactors (2 reactors per concentration
level) were charged with stock solution (20 mL) and copper pivalate
(13.4 mM). The vessels were sealed, pressurized with ethylene (35, 50,
75, or 100 psig), and stirred while heated to 150 °C. The reactions
were sampled every 30 min until complete. At each time point the
reactors were cooled to room temperature, sampled, recharged with
ethylene, and reheated. Aliquots of the reaction mixture were analyzed
by GC/FID using relative peak area vs an internal standard (decane).
Oxidative Hydrophenylation of Ethylene as a Function of
[Cu(OPiv)2]. A stock solution containing 1-OAc (0.112 mM), decane
(10 equiv relative to 1-OAc), and benzene (250 mL) was prepared in a
volumetric flask. Fisher-Porter reactors (2 reactors per concentration
level) were charged with stock solution (20 mL) and copper pivalate
(13.4, 19.0, 26.9, 38.1, or 53.8 mM). The vessels were sealed,
pressurized with ethylene (50 psig), and stirred while heated to 150
°C. The reactions were sampled every 30 min until complete. At each
time point the reactors were cooled to room temperature, sampled,
recharged with ethylene, and reheated. Aliquots of the reaction
mixture were analyzed by GC/FID using relative peak area vs an
internal standard (decane).
Oxidative Hydrophenylation of Ethylene as a Function of
[1-OAc]. Five separate stock solutions were prepared in 50 mL
volumetric flasks, each containing 1-OAc (0.225, 0.168, 0.112, 0.079,
or 0.056 mM), decane (10 equiv relative to 1-OAc), and benzene (50
mL). Fisher-Porter reactors (2 reactors per concentration level) were
charged with stock solution (20 mL) and copper pivalate (26.9 mM).
The vessels were sealed, pressurized with ethylene (50 psig), and
stirred while heated to 150 °C. The reactions were sampled every 30
min until complete. At each time point the reactors were cooled to
room temperature, sampled, recharged with ethylene, and reheated.
Aliquots of the reaction mixture were analyzed by GC/FID using
relative peak area vs an internal standard (decane).
Dependence of Order in [1-OAc] on [C2H4]. Three separate
stock solutions were prepared in 50 mL volumetric flasks, each
containing 1-OAc (0.225 mM, 0.112 mM, or 0.056 mM), decane (10
equiv relative to 1-OAc), and benzene (50 mL). Fisher-Porter reactors
(6 reactors per concentration of Rh) were charged with stock solution
(20 mL) and copper pivalate (26.9 mM). The vessels were sealed,
pressurized with ethylene (35 psig, 50 psig, or 75 psig; 2 reactors per
pressure at each Rh concentration), and stirred while heated to 150
°C. The reactions were sampled every 30 min until complete. At each
time point the reactors were cooled to room temperature, sampled,
recharged with ethylene, and reheated. Aliquots of the reaction
mixture were analyzed by GC/FID using relative peak area vs an
internal standard (decane).
Dependence of Order in [1-OAc] on Temperature. Three
separate stock solutions were prepared in 50 mL volumetric flasks,
each containing 1-OAc (0.225 mM, 0.112 mM, or 0.056 mM), decane
(10 equiv relative to 1-OAc), and benzene (50 mL). Fisher-Porter
reactors (6 reactors per concentration of Rh) were charged with stock
solution (20 mL) and copper pivalate (26.9 mM). The vessels were
sealed, pressurized with ethylene (to keep the concentration of
ethylene dissolved in solution constant across all temperatures,
Oxidative Hydrophenylation of Ethylene Using 1-OAc in
C6D6. A stock solution containing 1-OAc (0.112 mM), decane (10
equiv relative to 1-OAc), and C6D6 (50 mL) was prepared in a
volumetric flask. Fisher-Porter reactors were charged with stock
solution (20 mL) and copper pivalate (26.9 mM). The vessels were
L
DOI: 10.1021/jacs.6b10658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
sealed, pressurized with ethylene (50 psig), and stirred while heated to
150 °C. The reactions were sampled every 30 min for 3 h. At each
time point the reactors were cooled to room temperature, sampled,
recharged with ethylene, and reheated. Aliquots of the reaction
mixture were analyzed by GC/FID using relative peak area vs an
internal standard (decane).
ORCID
Benjamin A. Vaughan: 0000-0001-9607-5696
T. Brent Gunnoe: 0000-0001-5714-3887
Notes
The authors declare no competing financial interest.
Oxidative Hydrophenylation of Ethylene Using in Situ
Generated (FlDAB)RhCl3(η2-C2H4). A stock solution containing
RhCl3 (0.112 mM), FlDAB (0.112 mM), decane (1.12 mM), and
benzene (50 mL) was prepared in a volumetric flask. Fisher-Porter
reactors were charged with stock solution (20 mL) and Cu(OPiv)2 (27
mM). The vessels were sealed, pressurized with ethylene (50 psig),
and stirred while heated to 150 °C. The reactions were sampled at 4
and 24 h. At each time point the reactors were cooled to room
temperature, sampled, recharged with ethylene, and reheated. Aliquots
of the reaction mixture were analyzed by GC/FID using relative peak
area vs an internal standard (decane).
ACKNOWLEDGMENTS
■
This work was supported by the U.S. Department of Energy,
Office of Basic Energy Sciences [DE-SC0000776 (T.B.G.) and
DE-FG02-03ER15387 (T.R.C.)], an AES Graduate Fellowship
in Energy Research (M.S.W.-G.), and a University of Virginia
Jefferson Arts & Sciences Dissertation Year Fellowship
(M.S.W.-G.). The authors also thank Prof. Mahdi Abu-Omar
and Dr. Mark Pouy for useful discussions.
Oxidative Hydrophenylation of Ethylene Using 1-OAc at
Pressures >100 psig. A representative catalytic reaction is described.
A stock solution containing 1-OAc (0.112 mM), decane (1.12 mM),
and benzene (50 mL) was prepared in a volumetric flask. Stainless
steel high pressure reactors equipped with glass liners were charged
with stock solution (10 mL) and Cu(OPiv)2 (27 mM). The vessels
were sealed, pressurized with ethylene (400 psig), and stirred while
heated to 150 °C. The reactions were sampled every hour until
complete. At each time point the reactors were sampled at temperature
using a narrow bore dip tube. Aliquots of the reaction mixture were
analyzed by GC/FID using relative peak area vs an internal standard
(decane).
Oxidative Hydrophenylation of Ethylene Using 1-OAc with
Added AcOH at Low Pressures <100 psig. A representative
catalytic reaction is described. A stock solution containing 1-OAc
(0.112 mM), decane (1.12 mM), and benzene (50 mL) was prepared
in a volumetric flask. Fisher-Porter reactors were charged with stock
solution (20 mL), AcOH (56 mM), and Cu(OPiv)2 (27 mM). The
vessels were sealed, pressurized with ethylene (50 psig), and stirred
while heated to 150 °C. The reactions were sampled every 30 min
until complete. At each time point the reactors were cooled to room
temperature, sampled, recharged with ethylene, and reheated. Aliquots
of the reaction mixture were analyzed by GC/FID using relative peak
area vs an internal standard (decane).
REFERENCES
(1) Olah, G. A.; Molnar
Hoboken, NJ, 2003.
■
́
́
, A. Hydrocarbon Chemistry, 2 ed.; Wiley:
(2) Perego, C.; Pollesel, P. Advances in Aromatics Processing Using
Zeolite Catalysts. In Advances in Nanoporous Materials; Stefan, E., Ed.;
Elsevier: Oxford, 2010; Vol. 1, pp 97.
(3) Chen, S.-S. Styrene. Kirk-Othmer Encyclopedia of Chemical
Technology; Wiley: Hoboken, NJ, 2000.
(4) Wittcoff, H. A.; Reuben, B. G.; Plotkin, J. S. Chemicals and
Polymers from Ethylene. Industrial Organic Chemicals; Wiley:
Hoboken, NJ, 2004; pp 100.
(5) Chem Systems, I. Styrene/Ethylbenzene; Process Evaluation/
Research Planning (PERP) Program Report 91-9;Chem Systems,
Inc.: New York, 1992.
(6) Lucchini, M.; Galeotti, A. International Patent
WO2007073918A1, 2007.
(7) Perego, C.; Ingallina, P. Catal. Today 2002, 73, 3.
(8) Perego, C.; Ingallina, P. Green Chem. 2004, 6, 274.
̌
́
(9) Cejka, J.; Wichterlova, B. Catal. Rev.: Sci. Eng. 2002, 44, 375.
(10) Gerzeliev, I. M.; Khadzhiev, S. N.; Sakharova, I. E. Pet. Chem.
2011, 51, 39.
(11) CRC handbook of chemistry and physics; Weast, R. C., Ed.; CRC
Oxidative Hydrophenylation of Ethylene Using 1-OAc with
Added AcOH at High Pressures >100 psig. A representative
catalytic reaction is described. A stock solution containing 1-OAc
(0.112 mM), decane (1.12 mM), and benzene (50 mL) was prepared
in a volumetric flask. Stainless steel high pressure reactors equipped
with glass liners were charged with stock solution (10 mL), AcOH (56
mM), and Cu(OPiv)2 (27 mM). The vessels were sealed, pressurized
with ethylene (400 psig), and stirred while heated to 150 °C. The
reactions were sampled every 30 min until complete. At each time
point the reactors were sampled at temperature using a narrow-bore
dip tube. Aliquots of the reaction mixture were analyzed by GC/FID
using relative peak area vs an internal standard (decane).
Press/Taylor & Francis: Boca Raton, FL, 1977.
(12) Lail, M.; Arrowood, B. N.; Gunnoe, T. B. J. Am. Chem. Soc.
2003, 125, 7506.
(13) Lail, M.; Bell, C. M.; Conner, D.; Cundari, T. R.; Gunnoe, T. B.;
Petersen, J. L. Organometallics 2004, 23, 5007.
(14) McKeown, B. A.; Foley, N. A.; Lee, J. P.; Gunnoe, T. B.
Organometallics 2008, 27, 4031.
(15) Foley, N. A.; Lee, J. P.; Ke, Z.; Gunnoe, T. B.; Cundari, T. R.
Acc. Chem. Res. 2009, 42, 585.
(16) Andreatta, J. R.; McKeown, B. A.; Gunnoe, T. B. J. Organomet.
Chem. 2011, 696, 305.
(17) McKeown, B. A.; Gonzalez, H. E.; Friedfeld, M. R.; Gunnoe, T.
B.; Cundari, T. R.; Sabat, M. J. Am. Chem. Soc. 2011, 133, 19131.
(18) Joslin, E. E.; McMullin, C. L.; Gunnoe, T. B.; Cundari, T. R.;
Sabat, M.; Myers, W. H. Organometallics 2012, 31, 6851.
(19) McKeown, B. A.; Gonzalez, H. E.; Michaelos, T.; Gunnoe, T. B.;
Cundari, T. R.; Crabtree, R. H.; Sabat, M. Organometallics 2013, 32,
3903.
(20) McKeown, B. A.; Gonzalez, H. E.; Friedfeld, M. R.; Brosnahan,
A. M.; Gunnoe, T. B.; Cundari, T. R.; Sabat, M. Organometallics 2013,
32, 2857.
(21) McKeown, B. A.; Gonzalez, H. E.; Gunnoe, T. B.; Cundari, T.
R.; Sabat, M. ACS Catal. 2013, 3, 1165.
(22) McKeown, B. A.; Prince, B. M.; Ramiro, Z.; Gunnoe, T. B.;
Cundari, T. R. ACS Catal. 2014, 4, 1607.
(23) Burgess, S. A.; Joslin, E. E.; Gunnoe, T. B.; Cundari, T. R.;
Sabat, M.; Myers, W. H. Chem. Sci. 2014, 5, 4355.
(24) Bhalla, G.; Bischof, S. M.; Ganesh, S. K.; Liu, X. Y.; Jones, C. J.;
Borzenko, A.; Tenn, W. J., III; Ess, D. H.; Hashiguchi, B. G.; Lokare,
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.6b10658.
■
S
Initial rate plots for catalytic reactions, kinetic deriva-
tions, mass spectra, and TEM images, including Figures
S1−S17 (PDF)
Full computational results for B3LYP (XYZ)
Full computational results for M06 (XYZ)
AUTHOR INFORMATION
■
Corresponding Authors
*t@unt.edu
*tbg7h@virginia.edu
M
DOI: 10.1021/jacs.6b10658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
̈
K. S.; Leung, C. H.; Oxgaard, J.; Goddard, W. A., III; Periana, R. A.
Green Chem. 2011, 13, 69.
(25) Hong, P.; Yamazaki, H. J. Mol. Catal. 1984, 26, 297.
(26) Taube, D.; Periana, R.; Matsumoto, T. U.S. Patent 6127590A,
2000.
(27) Fujiwara, Y.; Moritani, I.; Danno, S.; Asano, R.; Teranishi, S. J.
Am. Chem. Soc. 1969, 91, 7166.
(28) Yamada, T.; Sakakura, A.; Sakaguchi, S.; Obora, Y.; Ishii, Y. New
J. Chem. 2008, 32, 738.
(29) Kubota, A.; Emmert, M. H.; Sanford, M. S. Org. Lett. 2012, 14,
D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J.
Gaussian 09; Gaussian, Inc.: Wallingford, CT, 2009.
(56) Becke, A. D. Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38, 3098.
(57) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B: Condens. Matter
Mater. Phys. 1988, 37, 785.
(58) Grimme, S.; Ehrlich, S.; Goerigk, L. J. Comput. Chem. 2011, 32,
1456.
(59) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B
2009, 113, 6378.
(60) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215.
1760.
(30) Sasaki, K.; Sakakura, T.; Tokunaga, Y.; Wada, K.; Tanaka, M.
Chem. Lett. 1988, 17, 685.
(31) Vaughan, B. A.; Webster-Gardiner, M. S.; Cundari, T. R.;
Gunnoe, T. B. Science 2015, 348, 421.
(32) Webster-Gardiner, M. S.; Fu, R.; Fortman, G. C.; Nielsen, R. J.;
Gunnoe, T. B.; Goddard Iii, W. A. Catal. Sci. Technol. 2015, 5, 96.
(33) Webb, J. R.; Pierpont, A. W.; Munro-Leighton, C.; Gunnoe, T.
B.; Cundari, T. R.; Boyle, P. D. J. Am. Chem. Soc. 2010, 132, 4520.
(34) Widegren, J. A.; Finke, R. G. J. Mol. Catal. A: Chem. 2003, 198,
317.
(35) Hamlin, J. E.; Hirai, K.; Gibson, V. C.; Maitlis, P. M. J. Mol.
Catal. 1982, 15, 337.
(36) Gorelsky, S. I.; Lapointe, D.; Fagnou, K. J. Am. Chem. Soc. 2008,
130, 10848.
(37) Sperger, T.; Sanhueza, I. A.; Kalvet, I.; Schoenebeck, F. Chem.
Rev. 2015, 115, 9532.
(38) Balcells, D.; Clot, E.; Eisenstein, O. Chem. Rev. 2010, 110, 749.
(39) Ackermann, L. Acc. Chem. Res. 2014, 47, 281.
(40) Boutadla, Y.; Davies, D. L.; Macgregor, S. A.; Poblador-
Bahamonde, A. I. Dalton Trans. 2009, 5820.
(41) Webb, J. R.; Burgess, S. A.; Cundari, T. R.; Gunnoe, T. B. Dalton
Trans. 2013, 42, 16646.
(42) Gunnoe, T. B. Eur. J. Inorg. Chem. 2007, 2007, 1185.
(43) Shiotsuki, M.; White, P. S.; Brookhart, M.; Templeton, J. L. J.
Am. Chem. Soc. 2007, 129, 4058.
(44) Oxgaard, J.; Periana, R. A.; Goddard, W. A. J. Am. Chem. Soc.
2004, 126, 11658.
(45) Holder, G. A.; Macauley, D. J. Chem. Eng. Data 1992, 37, 100.
(46) Baxter, R. D.; Sale, D.; Engle, K. M.; Yu, J.-Q.; Blackmond, D. G.
J. Am. Chem. Soc. 2012, 134, 4600.
(47) Northcutt, T. O.; Wick, D. D.; Vetter, A. J.; Jones, W. D. J. Am.
Chem. Soc. 2001, 123, 7257.
(48) Vetter, A. J.; Flaschenriem, C.; Jones, W. D. J. Am. Chem. Soc.
2005, 127, 12315.
(49) Zheng, L.; Wang, J. Chem. - Eur. J. 2012, 18, 9699.
(50) Cramer, R. Inorg. Synth. 1990, 28, 86.
(51) Bianchi, F.; Gallazzi, M. C.; Porri, L.; Diversi, P. J. Organomet.
Chem. 1980, 202, 99.
(52) Werner, H.; Poelsma, S.; Schneider, M. E.; Windmuller, B.;
̈
Barth, D. Chem. Ber. 1996, 129, 647.
(53) Khusniyarov, M. M.; Harms, K.; Burghaus, O.; Sundermeyer, J.
Eur. J. Inorg. Chem. 2006, 2006, 2985.
(54) Xie, L.-H.; Suh, M. P. Chem. - Eur. J. 2011, 17, 13653.
(55) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M. J.; Heyd, J.; Brothers, E. N.;
Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.;
Raghavachari, K.; Rendell, A. P.; Burant, J. C.; Iyengar, S. S.;
Tomasi, J.; Cossi, M.; Rega, N.; Millam, N. J.; Klene, M.; Knox, J. E.;
Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.;
Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.;
Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.;
Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A.
N
DOI: 10.1021/jacs.6b10658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX