Reaction pathways and energetics of etheric C-O bond cleavage catalyzed by lanthanide triflates

Article abstract of DOI:10.1021/cs400483q

Efficient and selective cleavage of etheric C-O bonds is crucial for converting biomass into platform chemicals and liquid transportation fuels. In this contribution, computational methods at the DFT B3LYP level of theory are employed to understand the efficacy of lanthanide triflate catalysts (Ln(OTf)₃, Ln = La, Ce, Sm, Gd, Yb, and Lu) in cleaving etheric C-O bonds. In agreement with experiment, the calculations indicate that the reaction pathway for C-O cleavage occurs via a C-H → O-H proton transfer in concert with weakening of the C-O bond of the coordinated ether substrate to ultimately yield a coordinated alkenol. The activation energy for this process falls as the lanthanide ionic radius decreases, reflecting enhanced metal ion electrophilicity. Details of the reaction mechanism for Yb(OTf) ₃-catalyzed ring opening are explored in depth, and for 1-methyl-d₃-butyl phenyl ether, the computed primary kinetic isotope effect of 2.4 is in excellent agreement with experiment (2.7), confirming that etheric ring-opening pathway involves proton transfer from the methyl group alpha to the etheric oxygen atom, which is activated by the electrophilic lanthanide ion. Calculations of the catalytic pathway using eight different ether substrates indicate that the more rapid cleavage of acyclic versus cyclic ethers is largely due to entropic effects, with the former C-O bond scission processes increasing the degrees of freedom/particles as the transition state is approached.

Full text of DOI:10.1021/cs400483q

Research Article
pubs.acs.org/acscatalysis
Reaction Pathways and Energetics of Etheric C−O Bond Cleavage
Catalyzed by Lanthanide Triflates
Rajeev S. Assary,*,† Abdurrahman C. Atesin,‡ Zhi Li,‡ Larry A. Curtiss,*,†,§ and Tobin J. Marks*,‡
§
†Materials Science Division and Center for Nanoscale Materials, Argonne National Laboratory, Argonne, Illinois 60439, United
States
‡Department of Chemistry, Northwestern University, Evanston, Illinois 60208, United States
S
* Supporting Information
ABSTRACT: Efficient and selective cleavage of etheric C−O
bonds is crucial for converting biomass into platform chemicals
and liquid transportation fuels. In this contribution, computa-
tional methods at the DFT B3LYP level of theory are
employed to understand the efficacy of lanthanide triflate
catalysts (Ln(OTf)3, Ln = La, Ce, Sm, Gd, Yb, and Lu) in
cleaving etheric C−O bonds. In agreement with experiment,
the calculations indicate that the reaction pathway for C−O
cleavage occurs via a C−H → O−H proton transfer in concert
with weakening of the C−O bond of the coordinated ether
substrate to ultimately yield a coordinated alkenol. The
activation energy for this process falls as the lanthanide ionic
radius decreases, reflecting enhanced metal ion electrophilicity. Details of the reaction mechanism for Yb(OTf)3-catalyzed ring
opening are explored in depth, and for 1-methyl-d3-butyl phenyl ether, the computed primary kinetic isotope effect of 2.4 is in
excellent agreement with experiment (2.7), confirming that etheric ring-opening pathway involves proton transfer from the
methyl group alpha to the etheric oxygen atom, which is activated by the electrophilic lanthanide ion. Calculations of the catalytic
pathway using eight different ether substrates indicate that the more rapid cleavage of acyclic versus cyclic ethers is largely due to
entropic effects, with the former C−O bond scission processes increasing the degrees of freedom/particles as the transition state
is approached.
KEYWORDS: biomass conversion, computational ether C−O hydrogenolysis, ionic liquids, lanthanide triflate catalysts,
density functional theory, activation energy, kinetic isotopic effect
which is found as ether linkages between aliphatic, aromatic,
and keto/aldehyde groups.2 Efficient and selective cleavage of
these etheric linkages would facilitate the liquification of
biomass and coal into processable fuel and commodity
chemical precursors.20,21 Recently, Marks and co-workers22
reported on a selective Ln(OTf)3/Pd nanoparticle-mediated
tandem catalytic etheric C−O bond hydrogenolysis process in
“green” ionic liquid media.
Based on the microscopic reverse of the well-documented
alkene hydroalkoxylation reaction,23,24 this atom-economical
route consists of a tandem system in which the Ln(OTf)3
cleaves the etheric C−O bond through an endothermic
homogeneous bond scission process (Scheme 1, Cycle A) to
yield an alkenol intermediate, which then undergoes
exothermic heterogeneous hydrogenation mediated by the Pd
nanoparticles (Scheme 1, Cycle B). Due to the high reaction
temperatures, significant deactivation of the Pd nanoparticle
catalyst is observed, presumably due to agglomeration, when
commercial catalysts such as Pd/C and Pd/BaSO4 are used.
1. INTRODUCTION
Due to the rapid increase in demand for transportation fuels
and industrial chemicals around the globe, efficient utilization
of renewable energy resources such as biomass is crucial for
environmental sustainability and continued economic
growth.1−3 Ongoing research in this area during the past
decade has led to significant progress in developing efficient
routes for the conversion of cellulosic biomass into liquid
transportation fuels and industrial chemicals via a variety of
processes.4−9 These processes, however, must be scalable and
cost-effective in order for the chemical transformations of
lignocellulosic materials into useful chemicals to have significant
priority in our energy future.10−16 Although thermochemical
processes such as acid treatment and pyrolysis are currently
available for the conversion of biomass into lower molecular
weight products, they are hindered by low efficiency and poor
selectivity toward desired products and are ineffective in terms
of industrial level conversion of biomass to platform
chemicals.9,17−19
Lignocellulosic biomass and the products obtained from acid-
catalyzed reactions (furfurals) and fast pyrolysis (pyrans,
furans) have a high oxygen content, a substantial amount of
Received: April 24, 2013
Published: July 15, 2013
© 2013 American Chemical Society
1908
dx.doi.org/10.1021/cs400483q | ACS Catal. 2013, 3, 1908−1914

ACS Catalysis
Research Article
a
Scheme 1. Proposed Pathway for Tandem Lanthanide Triflate/Pd Nanoparticle-Mediated Etheric C−O Bond Hydrogenolysis
a
Ln = lanthanide; R1, R2 = organic functional group. Cycle A is found to be turnover-limiting.
However, sinter-resistant Pd nanoparticles deposited on Al2O3
via atomic layer deposition are significantly more robust, and
the hydrogenation reactions proceed to completion in this
case.25,26 In this catalytic process, it is found that the C−O
bonds of acyclic ethers cleaved more rapidly than those of cyclic
ethers, possibly reflecting entropic and/or steric factors. Far
slower or negligible C−O cleavage is observed with substrates
lacking a methyl group alpha to the etheric O atom, such as di-
n-octylether and tetrahydropyran. This observation is in good
agreement with the regiochemistry of the reverse hydro-
alkoxylation process (reverse of Scheme 1, Cycle A). Kinetic
studies reveal that C−O bond scission is the turnover-limiting
step here, with the rate law being first-order with respect to
substrate and Ln(OTf)3 concentration, and zero-order in
hydrogen pressure.22
As a consequence of the endothermicity of the ring-opening
process, the kinetics and thermodynamics that control rate and
selectivity in the Ln(OTf)3-catalyzed dehydroalkoxylation
process are key features that require molecular level under-
standing if ever more efficient and selective catalysts are to be
invented. In this contribution, we employ density functional
theory at the B3LYP level to analyze mechanistic and reaction
energetic trends for a series of ethers and lanthanide ions,
including analysis of kinetic isotope effects.
solution computed here include the solvation energies.31 In
addition to the B3LYP/6-31+G(d) calculations, MP2/6-
311+G(2df,p)//B3LYP/6-31+G(d) level calculations were
also performed for the uncatalyzed reactions to check the
accuracy of B3LYP method. All calculations were performed
using the G09 software.32
3. EXPERIMENTAL DETAILS
3.1. Materials and Methods. All manipulations of
reagents were carried out under an N2 atmosphere. CDCl3
was obtained from commercial sources and stored over
molecular sieves. Heptanal, CD3MgI, phenol, triphenylphos-
phine, and diethyl azodicarboxylate were obtained from Aldrich
Chemical Co. and used without further purification. Yb-
33
(OTf)3 and Pd nanoparticles on alumina (Pd@ALD)25
containing 2% Pd by weight were synthesized according to
previously published procedures. The labeled substrates 1-d3-2-
octanol34 and 1-methyl-d3-heptyl phenyl ether35 were synthe-
sized according to modified literature procedures as outlined
1
below. The H NMR integration internal standard, 1,1,2,2-
tetrachloroethane, was purchased from Aldrich Chemical Co.,
distilled under reduced pressure, and stored under N2. The
ionic liquid [EMIM][OTf] was prepared according to a
published procedure36 and stored under an inert atmosphere.
3.2. Physical and Analytical Measurements. NMR
2. COMPUTATIONAL DETAILS
1
spectra were recorded on a Varian Inova-500 (500 MHz, H,
All calculations described in this investigation were performed
using the B3LYP27 density functional. A 6-31+G(d) basis set
was used for atoms such as C, H, O, N, S, and F. The Stuttgart
RSC 1997 ECP basis set was used for Ce, Sm, Gd, Yb, and Lu
atoms. The LANL2TZ(f) basis set was used for La. The basis
sets for the lanthanides (Ln) were obtained from the EMSL
database.28,29 A detailed description regarding the modeling of
the catalyst (Figure S1 of the Supporting Information) and
reactions is presented in the Supporting Information. All
structures presented in the paper are fully relaxed, and normal
coordinate analysis was performed for all stationary points to
characterize the transition states (one imaginary frequency) and
the equilibrium structures (no imaginary frequencies) and to
calculate the zero point energy correction (ZPE). To include
the contributions from solvent medium in the reaction
energetics (enthalpies and free energies at 298 K and 1 atm
pressure), the solvation energies were computed using the
SMD solvation model30 using a dielectric medium of 2-butanol
(ε = 15.9) at 298 K and using the gas phase geometry at the
B3LYP level of theory. Thus, the enthalpies of activation in
1
100 MHz, 13C) spectrometer. Chemical shifts (δ) for H and
13C are referenced to solvent peaks. A Varian 1200 Quadrupole
mass spectrometer was used to obtain mass spectral data. C−O
cleavage reactions were performed in a Parr reactor (model
4590, 100 mL, Parr Co., Moline, IL), capable of reaching 300
°C, 1000 psi, and a 3000 rpm stirring rate.
3.3. Preparation of 1-d3-2-octanol. To a 3-neck round-
bottom flask equipped with a condenser and magnetic stirrer
was charged 50 mL of 1 M CD3MgI in Et2O. Heptanal (45
mmol) was added dropwise with a syringe to allow a gentle
refluxing of the ether solution, which was then refluxed for 3 h
prior to quenching with 100 mL of 1 M aqueous HCl. The
mixture was then separated and extracted with 3 × 50 mL of
Et2O. The combined organic phase was washed with 2 × 50 mL
of brine, dried with Na2SO4, and evaporated in vacuo. The
resulting liquid was passed through a short silica gel column
with 3:1 hexane/ethyl acetate mixture as the eluent.
Evaporation of the solvents resulted in 1-d3-2-octanol (5.6g,
1
93%). H NMR (CDCl3): δ 3.78 (m, 1 H), 1.22−1.52 (m, 10
1909
dx.doi.org/10.1021/cs400483q | ACS Catal. 2013, 3, 1908−1914

ACS Catalysis
Research Article
H), 0.89 (t, J = 7.1 Hz, 3 H). 13C {1H} NMR (CDCl3): δ 68.2,
39.5, 32.0, 29.6, 25.7, 22.8, 14.2.
Scheme 2. Schematic Representation of the Catalytic
Conversion of Ether to Alkenol to Alcohol through Ring
a
3.4. Preparation of 1-Methyl-d3-Heptyl Phenyl Ether.
To a round-bottom flask were charged 1-d3-2-octanol (42
mmol), phenol (84 mmol), triphenylphosphine (88 mmol),
and THF (170 mL). The mixture was stirred for 20 min to let
all solids dissolve prior to cooling to 0 °C with an ice bath.
Diethyl azodicarboxylate (88 mmol) was then added very
slowly at 0 °C over 20 min. The mixture was warmed to room
temperature and stirred overnight. The solvent was then
evaporated, and residue was redissolved in 100 mL of CH2Cl2.
The solution was passed through a silica gel pad (11 cm
diameter, 7 cm height) in a fritted funnel. The silica pad was
washed with 500 mL of CH2Cl2, and all organics were collected
and evaporated to 300 mL total volume. The solution was then
washed sequentially with 300 mL each of 15% H2O2, saturated
Na2SO3, and H2O. The last H2O wash was back extracted with
200 mL of CH2Cl2, and all organics were collected and dried on
Na2SO4. The solution was then filtered through another 11 × 5
silica pad. The silica pad was washed with 5 × 200 mL of
hexanes. Evaporation of all collected fractions provided the title
compound (7.9g, 90%). 1H NMR (CDCl3): δ 7.17−7.28 (m, 2
H), 6.80−6.93 (m, 3 H), 4.25−4.35 (m, 1 H), 1.63−1.80 (m, 1
H),1.18−1.59 (m, 9 H), 0.78−0.93 (m, 3 H). 13C {1H} NMR
(CDCl3): δ 158.4, 129.5, 120.5, 116.0, 73.8, 36.6, 32.0, 29.5,
Opening and Subsequent Hydrogenation
a
The superscript “s” indicates that the ether is in solution and the rest
of the species are bound to the catalyst in solution. Also shown are the
various energy terms (ΔE) with respect to reaction coordinate.
etheric C−O bond. According to this mechanistic picture, the
ether molecule binds to the metal center (Ln3+) through the
electron-rich oxygen, followed by a proton transfer from the α-
methyl group to the etheric oxygen to form an alkenol
intermediate. Subsequent hydrogenation results in the for-
mation of saturated alcohols.
Experimentally, it is observed that the catalytic activity of the
lanthanide triflates increases as the ionic radius of the Ln3+ ion
decreases (Lewis acidity increases37). To acquire a better
theoretical understanding of this observation, we chose
La(OTf)3, Ce(OTf)3, Sm(OTf)3, Gd(OTf)3, Yb(OTf)3, and
Lu(OTf)3 as catalyst models and computed their energetics and
reaction barriers for C−O bond cleavage. Optimized structures
of the metal triflates, complexes of metal triflates with ether,
alkenol, and alcohol, and the transition state structure for the
C−O bond cleavage are presented in the Supporting
Information. The ether substrate, 2,3-dihydro-2-methylbenzo-
furan, 1A, was chosen to understand details of the energy
profile along the reaction coordinate to alkenol (1B) and
subsequent hydrogenation to afford n-propyl phenol (1C) (eq
1).
2
25.7, 22.8, 19.1 (m, CD3), 14.2. D {1H} NMR (CDCl3): δ
1.30.
3.5. Yb(OTf)3-Mediated Reaction of 1-Methyl-d3-
heptyl Phenyl Ether in [EMIM][OTf]. Yb(OTf)3 (10 mol
%, 6.5 × 10−5 mol), Pd@ALD (1.3 × 10−5 mol Pd),
[EMIM][OTf] (2.6 × 10−2 mol), and substrate (6.5 × 10−4
mol, 10 equiv with respect to Yb(OTf)3) were added to a 100
mL Teflon sleeve and then put into the Parr reactor in a
glovebox and sealed. The reactor was purged with 100 psi
hydrogen 10 times, after which the reactor was heated,
pressurized, and stirred at 1000 rpm for 18 h. The progress
of the reaction was followed by taking aliquots during the
reaction through a needle valve on the Parr reactor after cooling
the reaction mixture to 80 °C and decreasing the pressure to 30
psi. Upon the completion of the reaction, the reactor was
cooled down to room temperature and depressurized. The final
products phenol-OD and octane-1,1-d2 were separated from the
ionic liquid by vacuum distillation. The kinetic isotope effect
was calculated by measuring the initial rates of the reaction and
comparing it to the protio substrate under identical reaction
conditions.22 Spectral data of phenol-OD matched those of
1
commercially available phenol-OD. Octane-1,1-d2: H NMR
(CDCl3): δ 1.20−1.35 (m, 12 H), 0.84−0.96 (m, 4 H). 13C
2
{1H} NMR (CDCl3): δ 32.8, 30.5, 23.4, 22.0, 21.2, 14.1. D
{1H} NMR (CDCl3): δ 1.22.
Since Ln(III) triflates have differing solvation energies, we
have also incorporated the solvation energetics in this study.
Two important parameters were considered and computed: the
intrinsic enthalpy of activation (shown in Scheme 2, ΔHint) and
the free energy of reaction (ΔGb‑c) from the Ln(OTf)3
coordinated ether to alkenol. The catalytic reactions that
require a relatively low activation enthalpy and have a negative
free energy for the ether to alkenol conversion are considered
to be most likely. Following the aforementioned criterion, the
enthalpy of activation for the etheric C−O bond cleavage of 1A
using various Ln(OTf)3 catalysts and the free energy change for
the ether-alkenol reaction were computed. The results of these
calculations are shown in Figure 1. Details of the computed
enthalpy parameters in gas phase and in solution can be found
4. RESULTS AND DISCUSSION
A schematic representation of the reaction coordinate and
various energy terms associated with the conversion of ether to
alcohol via an alkenol is shown in Scheme 2. The changes in
energies of the ether, alkenol intermediate, and alcohol product
when bound to the catalyst are given as ΔE in the scheme but
are reported as free energies (ΔG°(298 K)) and enthalpies
(ΔH°(298 K)) as appropriate in this section. The labels ΔEapp
and ΔEint stand for apparent and intrinsic activation energies.
On the basis of the experimental kinetic results, Marks and co-
workers proposed that the turnover-limiting step for the
tandem ether to alcohol conversion is the cleavage of the
1910
dx.doi.org/10.1021/cs400483q | ACS Catal. 2013, 3, 1908−1914

ACS Catalysis
Research Article
Table 1. Comparison of Ionic Radii and Computed Ln3+−
OEther Bond Lengths (Å) for Various Ln3+ Ions and
Ln(OTf)3 Complex with the Ethers
La
Ce
Sm
Gd
Yb
Lu
ionic radius (Å)
1.03
2.68
2.60
2.62
1.01
2.69
2.66
2.58
0.96
2.61
2.72
2.47
0.94
2.59
2.77
2.44
0.87
2.44
2.81
2.38
0.86
2.41
2.80
2.35
a
APT charge (e)
charge density
Ln3+−Oether (Å)
a
Atomic polar tensor (APT) charges are obtained from DFT
frequency calculations.
activation of etheric bond and subsequent proton transfer to
cleave the etheric C−O bond. This is partially supported by the
increase in computed charge density (Table 1) of various Ln3+
ions in Ln(OTf)3 with the decrease of effective ionic radii.
To probe the proposed mechanistic scenario outlined in
Schemes 1 and 2, kinetic isotope effect data were obtained for
the cleavage of the etheric C−O bond in 1-methyl-d3-heptyl
1
phenyl ether by H NMR spectroscopy (Ph-O-C(H)(R)-CH3
Figure 1. Computed enthalpy barriers (top) and free energies for the
conversion of ether to enol catalyzed by various Ln(OTf)3 (Ln = La,
Ce, Sm, Gd, Yb, and Lu) catalysts in solution. All energies were
computed at the B3LYP level of theory. Under identical reaction
conditions, the experimental turnover frequencies of catalysts
La(OTf)3, Sm(OTf)3, and Yb(OTf)3 are 0.12, 0.22, and 0.55 h−1,
respectively.22
vs Ph-O-C(H)(R)-CD3). These experiments yielded a kH/kD of
2.7, which is suggestive of a primary KIE. Next, 1-methylbutyl
phenyl ether, 8A, was used in the calculations to compare with
the experimentally obtained value. The detailed enthalpy
landscape in the gas phase for the conversion of this ether to
alcohol is shown in Figure 2. The computed intrinsic activation
in Table S1 of the Supporting Information. The Ln(OTf)3
cases for which there is experimental data are highlighted with a
rectangular box, (Yb(OTf)3, Sm(OTf)3, and La(OTf)3).
Optimized geometries and gas phase enthalpy profiles for the
lanthanide triflates can be found in Figures S3−S6 of the
Supporting Information. Based on the computed activation
enthalpies and the reaction free energies associated with the 1A
→ 1B conversion, Yb(OTf)3 is found to be the most active
catalyst among the various Ln(OTf)3 catalytic models examined
in this comparison. The computed activation enthalpy for the
1A → 1B cleavage by Yb3+ is 24.3 kcal/mol, in contrast to the
uncatalyzed reaction that has an activation barrier of 56 kcal/
mol (see Figure S2, Supporting Information), and the reaction
is calculated to be endergonic by 6 kcal/mol. In terms of the
reaction barriers, Ce(OTf)3, Sm(OTf)3, and Gd(OTf)3 show
similar intrinsic enthalpic barriers (∼26 kcal/mol) for the
reaction, while both the La(OTf)3- and Lu(OTf)3-catalyzed
reactions require relatively higher barriers of 32 and 28 kcal/
mol, respectively. In terms of free energies of reaction, the
La(OTf)3-, Lu(OTf)3-, and Ce(OTf)3-catalyzed reactions are
slightly more endergonic versus the Yb(OTf)3-, Sm(OTf)3-,
and Gd(OTf)3-catalyzed reactions. The activity of the catalysts
increases with the increase of the charge density and the
decrease of the ionic radius of the central metal ions. Note that
a direct correlation between the Ln3+−Oether bond distances
(optimized geometry) and the ionic radii/charge density of the
Ln3+ ions is observed (Table 1). This computed trend in
lanthanide catalytic activities is consistent with the experimental
results of Marks and co-workers, where they investigated
Yb(OTf)3, Sm(OTf)3, and La(OTf)3 for the cleavage of the 1A
etheric bond.22 A trend of decreasing activation barrier with
decreasing Ln3+ ionic radius and Ln3+−Oether bond length is
observed in the lanthanide-catalyzed ether ring opening.22 We
speculate that the increase in charge density of the central
cation (upon lanthanide contraction) results in an increased
Lewis acidity, which could be the reason for the effective
Figure 2. Computed gas phase enthalpy profile for the conversion of
8A to pentane and phenol catalyzed by Yb(OTf)3 at the B3LYP level
of theory.
enthalpy required for the C−O bond cleavage of 8A is 20.9
kcal/mol. The ring cleavage process yields 1-pentene and a
Yb(OTf)3-phenol complex. This reaction is computed to be
endothermic by 5.8 kcal/mol, while the subsequent hydro-
genation of 1-pentene is exothermic by 19.9 kcal/mol.
Experimentally, Pd nanoparticles are reported to catalyze the
hydrogenation reaction, as noted above. For computational
simplicity, the decomplexation kinetics of alkene and phenol
from the Yb3+ center and the subsequent energetics of
hydrogenation by the Pd nanoparticles were not computed.
The computed activation energy differences (ΔΔETS(H/D))
between the α-CH3 and the α-CD3 derivatives of phenyl-2-
pentyl ether is 0.51 kcal/mol; the details of the KIE evaluation
are presented in Scheme S1 and Table S2 of the Supporting
Information. This small change in energy corresponds to a KIE
for proton transfer of 2.4, which is in excellent agreement with
the experimental value of 2.7.
1911
dx.doi.org/10.1021/cs400483q | ACS Catal. 2013, 3, 1908−1914

ACS Catalysis
Research Article
Figure 3 shows the gas phase enthalpy profile associated with
the ring opening and subsequent hydrogenation of 1A to 1C. In
Figure 3. Computed gas phase enthalpic profile for the conversion of
2,3-dihydro-2-methylbenzofuran (1A) to propyl phenol catalyzed by
Yb(OTf)3 at the B3LYP level of theory.
the gas phase, binding of the ether through its oxygen atom to
Yb(OTf)3 is an exothermic process (−14 kcal/mol, not
shown), and the intrinsic activation enthalpy for ring opening
(ether to TS) is +31.8 kcal/mol. Note that conversion of the
ether to alkenol is an endothermic process (+10.7 kcal/mol),
while the subsequent hydrogenation (alkenol to alkanol) is
exothermic (−24.5 kcal/mol). The computed activation energy
differences (ΔΔETS(H/D)) between the α-CH3 and the α-CD3
derivatives of 1A is 0.67 kcal/mol, with the details of the KIE
evaluation presented in the Table S2 of the Supporting
Information. This small change in energy corresponds to a KIE
for proton transfer of 3.2 (see Table S2 of the Supporting
Information). Although we have not measured the experimental
KIE for this substrate, the computed KIE suggests a proton
transfer pathway similar to that of 8A.
Comparing the ring opening of 8A to that of 1A, it can be
seen that the linear ether incurs a smaller barrier (20.9 kcal/
mol) and that the C−O bond cleavage reaction is less
endergonic (+5.8 kcal/mol) when compared to the corre-
sponding parameters for 1A, 31.8 and 10.7 kcal/mol,
respectively. These results indicate that the catalytic reaction
of linear ethers is kinetically and thermodynamically more
favorable than for the cyclic ethers. Experimentally, it is found
that linear ethers are more rapidly cleaved than cyclic ethers,
requiring lower temperatures for the corresponding reactions to
proceed to completion (110 vs 180 °C).22
A total of eight etheric substrates that have been studied by
Marks and co-workers were next selected in order to compute
the energetics of the ring opening and subsequent hydro-
genation processes as shown in Figure 4. All the substrates have
a methyl substituent at the α-position to the etheric oxygen
atom. Detailed gas phase thermochemistry including reaction
barriers were computed at the B3LYP/6-31+G(d) level of
theory (gas phase) for the uncatalyzed C−O bond cleavage and
subsequent hydrogenation reactions. This level of theory
provides reasonably good agreement compared to a MP2/6-
311+G(3df,2p)//B3LYP/6-31+G(d) level of calculation in
predicting most of the reaction energetics and reaction barriers
(see Figure S7 in the Supporting Information). Exceptions are
the barriers for the C−O bond-breaking reactions in ethers 1A
and 5A (Figure 4). These substrates have a phenyl group
Figure 4. Computed energetics of uncatalyzed conversion of ether to
alcohol in the gas phase at the B3LYP/6-31+G(d) level of theory.
Note that for substrate 4 the computed reaction barrier is for breaking
one C−O bond, but reaction energies are for two bonds.
adjacent to the etheric oxygen, and density functional theory
underestimates the reaction barriers by ∼10 kcal/mol.
These calculations clearly indicate that the ring opening is
endothermic for all substrates (12−21 kcal/mol), while
hydrogenation of the corresponding alkenols is exothermic
((−24 to −32 kcal/mol; Table S1 of the Supporting
Information). In terms of the reaction barriers for the
uncatalyzed reactions, the formation of the alkenol inter-
mediates from the ether species invariably incurs an activation
barrier (ΔHint = ΔHapp) of about 60 kcal/mol. The formation
of alkenol in the gas phase is found to take place via proton
transfer (Figure S2 of the Supporting Information), and the
computed barrier is extremely high, indicating that these
reactions are kinetically not feasible. Another essential
parameter that was computed is the proton affinities of oxygen
atoms (PA) of the substrates, for which the computed values
range from 190 to 206 kcal/mol, depending on the substituents
around the etheric oxygen atom. Note that the PAs of the
ethers increase in the presence of electron-donating sub-
stituents (methyl or alkyl groups) adjacent to the ether oxygen.
It is therefore likely that substrates with higher proton affinities
would interact with the positively charged lanthanide ions
(Ln3+) more strongly than the substrates with lower proton
1912
dx.doi.org/10.1021/cs400483q | ACS Catal. 2013, 3, 1908−1914

ACS Catalysis
Research Article
affinities. A stronger interaction suggests an effective Lewis
“adduct” between the metal ion (Lewis acid) and the etheric
oxygen (Lewis base) and subsequent activation of the C−O
bond is likely to be determined by the Lewis basicity of various
metal ions (Lewis acid).
The computed intrinsic activation enthalpies, free energy
changes for the catalyzed conversion of the ethers to alkenols,
and the enthalpies of binding the ether substrates to Yb(OTf)3
are summarized in Figure 5. Detailed relative free energies of
central metal ion decreases. Yb(OTf)3 was found to be the
most effective catalyst among the various lanthanide triflates
considered in this investigation. The detailed reaction
mechanism of the Yb(OTf)3-catalyzed ether ring opening was
explored by measurement of the KIE using both experiment
and density functional theory. The measured KIE for phenyl-2-
octyl ether (2.7) and computed KIE for phenyl-2-pentyl-ether
(2.4) are in excellent agreement, indicating that the etheric ring
opening occurs via a proton transfer from the α-CH3 group
upon activation of the etheric oxygen by the electrophilic metal
ion. Based on the concept of this activation of ether oxygen
atoms, we speculate that the metal ions with higher positive
charge density (M4+ ions, small M3+ ions) would be ideal
candidates for more efficient C−O bond-breaking catalysts,
which require future experimental and theoretical investiga-
tions. Calculations using various ethers suggest that the C−O
bonds of linear ethers cleaved more rapidly than those of cyclic
ethers, consistent with the increase in entropy due to the
increase in the number of particles/degrees of freedom from
the scission of the C−O bond.
ASSOCIATED CONTENT
* Supporting Information
■
S
Details of computations, enthalpy profiles for the liquid phase
Ln(OTf)3-catalyzed hydrogenolysis of 1A ether (Ln = La, Ce,
Sm, Gd), enthalpy profile for gas phase C−O cleavage of 1A.
Details of KIE and MP2 energetics for the uncatalyzed
reactions of ethers 1A−8A. This material is available free of
charge via the Internet at http://pubs.acs.org.
Figure 5. Computed enthalpic barriers (top), free energies (middle)
for the conversion of ether to alkenol catalyzed by Yb(OTf)3, and
enthalpies of binding of the ether substrates to the catalyst (bottom) in
solution at the B3LYP level of theory. Various ethers (1A−8A) are
shown in Figure 4.
AUTHOR INFORMATION
Corresponding Author
*E-mail: assary@anl.gov; curtiss@anl.gov; t-marks@
northwestern.edu.
■
ethers, alkenols, and alkanols can also be found in Table S3 of
the Supporting Information. From Figure 5, it is evident that
the computed activation enthalpies for all the substrates are
below 40 kcal/mol. Since the uncatalyzed reactions have an
activation barrier of ∼60 kcal/mol (Figure 4), it is clear that
Yb(OTf)3 substantially lowers the barriers for all the ether
substrates.
Two ether substrates that show quantitatively lower
activation enthalpies are 1A and 5A. For these two species,
which have phenyl groups adjacent to the etheric oxygen, note
that the B3LYP level of theory underestimates the reaction
barriers compared to the MP2 level of theory (Figure S7). As
noted above, experimental kinetic measurements show that
linear ethers such as 8A undergo more rapid ring opening than
cyclic ethers. The calculations presented here are consistent
with these observations, indicating that 8A incurs a relatively
low activation barrier compared to cyclic ethers 1A−7A. This is
consistent with expected entropic effects due to the increased
number of particles/degrees of freedom accompanying the
scission of the C−O bonds in linear ethers.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the U.S. Department of Energy
under contract DE-AC0206CH11357. This material is based
upon work supported as part of the Institute of Atom Efficient
Chemical Transformation (IACT), an Energy Frontier
Research Center funded by the U.S. Department of Energy,
Office of Sciences, and Office of Basic Energy Sciences. We
gratefully acknowledge the computing resources provided on
“Fusion”, a 320-node computing cluster operated by the
Laboratory Computing Resource Center at Argonne National
Laboratory. Use of the Center for Nanoscale Materials was
supported by the U.S. Department of Energy, Office of Science,
Office of Basic Energy Sciences, under Contract No. DE-AC02-
06CH11357. This research also used resources of the National
Energy Research Scientific Computing Center (NERSC),
which is supported by the Office of Science of the U.S.
Department of Energy under Contract No. DE-AC02-
05CH11231. We also acknowledge grants of computer time
from EMSL, a national scientific user facility located at the
Pacific Northwest National Laboratory. Z.L. was supported
under NSF grant CHE-1213235 on basic f-element chemistry,
which also provided necessary equipment.
5. CONCLUSIONS
Selective and efficient cleavage of etheric C−O bonds is an
important challenge in future processes for the conversion of
biomass to platform chemicals. The selective tandem catalytic
Ln(OTf)3/Pd nanoparticle-mediated etheric C−O bond hydro-
genolysis process in ionic liquids represents a new approach. In
this paper, detailed computational studies were performed to
screen the activities of various lanthanide triflates toward C−O
bond cleavage. The density functional studies (B3LYP) suggest
that the activity of catalysts increases as the ionic radius of the
REFERENCES
(1) Regalbuto, J. R. Science 2009, 325, 822.
■
1913
dx.doi.org/10.1021/cs400483q | ACS Catal. 2013, 3, 1908−1914

ACS Catalysis
Research Article
(2) Corma, A.; Iborra, S.; Velty, A. Chem. Rev. 2007, 107, 2411.
(3) Huber, G. W.; Iborra, S.; Corma, A. Chem. Rev. 2006, 106, 4044.
(4) Chheda, J. N.; Huber, G. W.; Dumesic, J. A. Angew. Chem., Int.
Ed. 2007, 46, 7164.
(5) Bond, J. Q.; Alonso, D. M.; Wang, D.; West, R. M.; Dumesic, J. A.
Science 2010, 327, 1110.
W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 09, Rev. C.01;
Gaussian, Inc.: Wallingford CT, 2009.
(33) Dzudza, A.; Marks, T. J. J. Org. Chem. 2008, 73, 4004.
(34) Stein, A. R. Can. J. Chem. 1994, 72, 1789.
(35) Proctor, A. J.; Beautement, K.; Clough, J. M.; Knight, D. W.; Li,
Y. Tetrahedron Lett. 2006, 47, 5151.
(36) Bonho
̂
te, P.; Dias, A.-P.; Papageorgiou, N.; Kalyanasundaram,
(6) Huber, G. W.; Chheda, J. N.; Barrett, C. J.; Dumesic, J. A. Science
2005, 308, 1446.
K.; Gratzel, M. Inorg. Chem. 1996, 35, 1168.
̈
(37) Cotton, A. F.; Wilkinson, G.; Murillo, C. A.; Bochmann, M.
(7) Kunkes, E. L.; Simonetti, D. A.; West, R. M.; Serrano-Ruiz, J. C.;
Gartner, C. A.; Dumesic, J. A. Science 2008, 322, 417.
(8) van Putten, R.-J.; van der Waal, J. C.; de Jong, E.; Rasrendra, C.
B.; Heeres, H. J.; de Vries, J. G. Chem. Rev. 2013, 11, 1499.
(9) Lange, J.-P.; Price, R.; Ayoub, P. M.; Louis, J.; Petrus, L.; Clarke,
L.; Gosselink, H. Angew. Chem., Int. Ed. 2010, 49, 4479.
(10) Bozell, J. J.; Petersen, G. R. Green Chem. 2010, 12, 539.
(11) Lange, J.-P.; van der Heide, E.; van Buijtenen, J.; Price, R.
ChemSusChem 2012, 5, 150.
Advanced Inorganic Chemistry, 6th ed.; Wiley: New York, 1999.
(12) Bermejo-Deval, R.; Assary, R. S.; Nikolla, E.; Moliner, M.;
Roman-Leshkov, Y.; Hwang, S.-J.; Palsdottir, A.; Silverman, D.; Lobo,
R. F.; Curtiss, L. A.; Davis, M. E. Proc. Natl. Acad. Sci. U.S.A. 2012, 109,
9727.
(13) Zhang, Z.; Dong, K.; Zhao, Z. ChemSusChem 2011, 4, 112.
(14) Gonzalez Maldonado, G. M.; Assary, R. S.; Dumesic, J.; Curtiss,
L. A. Energy Environ. Sci. 2012, 5, 6981.
(15) Gonzalez Maldonado, G. M.; Assary, R. S.; Dumesic, J. A.;
Curtiss, L. A. Energy Environ. Sci. 2012, 5, 8990.
(16) Zhao, H.; Holladay, J. E.; Brown, H.; Zhang, Z. C. Science 2007,
316, 1597.
(17) Geboers, J. A.; Van de Vyver, S.; Ooms, R.; Op de Beeck, B.;
Jacobs, P. A.; Sels, B. F. Catal. Sci. Technol. 2011, 1, 714.
(18) Van de Vyver, S.; Geboers, J.; Jacobs, P. A.; Sels, B. F.
ChemCatChem 2011, 3, 82.
(19) Van de Vyver, S.; Thomas, J.; Geboers, J.; Keyzer, S.; Smet, M.;
Dehaen, W.; Jacobs, P. A.; Sels, B. F. Energy Environ. Sci. 2011, 4, 3601.
(20) Zakzeski, J.; Bruijnincx, P. C. A.; Jongerius, A. L.; Weckhuysen,
B. M. Chem. Rev. 2010, 110, 3552.
(21) Climent, M. J.; Corma, A.; Iborra, S. Green Chem. 2011, 13, 520.
(22) Atesin, A. C.; Ray, N. A.; Stair, P. C.; Marks, T. J. J. Am. Chem.
Soc. 2012, 134, 14682.
(23) Dzudza, A.; Marks, T. J. Org. Lett. 2009, 11, 1523.
(24) Dzudza, A.; Marks, T. J. Chem.Eur. J. 2010, 16, 3403.
(25) Lu, J.; Stair, P. C. Langmuir 2010, 26, 16486.
(26) Elam, J. W.; Zinovev, A.; Han, C. Y.; Wang, H. H.; Welp, U.;
Hryn, J. N.; Pellin, M. J. Thin Solid Films 2006, 515, 1664.
(27) Becke, A. D. J. Chem. Phys. 1993, 98, 1372.
(28) Feller, D. J. Comput. Chem. 1996, 17, 1571.
(29) Schuchardt, K. L.; Didier, B. T.; Elsethagen, T.; Sun, L.;
Gurumoorthi, V.; Chase, J.; Li, J.; Windus, T. L. J. Chem. Inf. Model.
2007, 47, 1045.
(30) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B
2009, 113, 6378.
(31) Assary, R. S.; Curtiss, L. A. Energy Fuels 2012, 26, 1344.
(32) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A.; Vreven, T.; Kudin,
K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone,
V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa,
J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene,
M.; Li, X.; Knox, J. E.; Hratchian, H. P.; 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.; Ayala, P. Y.;
Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.;
Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.;
Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman,
J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.;
Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Martin, R. L.; Fox, D. J.; Keith, T.; Laham, A.; Peng, C. Y.;
Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen,
1914
dx.doi.org/10.1021/cs400483q | ACS Catal. 2013, 3, 1908−1914