Ni-catalyzed cleavage of aryl ethers in the aqueous phase

Article abstract of DOI:10.1021/ja309915e

A novel Ni/SiO₂-catalyzed route for selective cleavage of ether bonds of (lignin-derived) aromatic ethers and hydrogenation of the oxygen-containing intermediates at 120 C in presence of 6 bar H₂ in the aqueous phase is reported. The C-O bonds of α-O-4 and β-O-4 linkages are cleaved by hydrogenolysis on Ni, while the C-O bond of the 4-O-5 linkage is cleaved via parallel hydrogenolysis and hydrolysis. The difference is attributed to the fact that the C_aliphatic-OH fragments generated from hydrolysis of α-O-4 and β-O-4 linkages can undergo further hydrogenolysis, while phenol (produced by hydrolysis of the 4-O-5 linkage) is hydrogenated to produce cyclohexanol under conditions investigated. The apparent activation energies, E_a(α-O-4) < E_a(β-O-4) < E_a(4-O-5), vary proportionally with the bond dissociation energies. In the conversion of β-O-4 and 4-O-5 ether bonds, C-O bond cleavage is the rate-determining step, with the reactants competing with hydrogen for active sites, leading to a maximum reaction rate as a function of the H₂ pressure. For the very fast C-O bond cleavage of the α-O-4 linkage, increasing the H₂ pressure increases the rate-determining product desorption under the conditions tested.

Full text of DOI:10.1021/ja309915e

Article
pubs.acs.org/JACS
Ni-Catalyzed Cleavage of Aryl Ethers in the Aqueous Phase
Jiayue He, Chen Zhao,* and Johannes A. Lercher*
Department of Chemistry and Catalysis Research Center, Technische Universitat Munchen, 85747 Garching, Germany
̈
̈
S
* Supporting Information
ABSTRACT: A novel Ni/SiO2-catalyzed route for selective
cleavage of ether bonds of (lignin-derived) aromatic ethers and
hydrogenation of the oxygen-containing intermediates at 120
°C in presence of 6 bar H2 in the aqueous phase is reported.
The C−O bonds of α-O-4 and β-O-4 linkages are cleaved by
hydrogenolysis on Ni, while the C−O bond of the 4-O-5
linkage is cleaved via parallel hydrogenolysis and hydrolysis.
The difference is attributed to the fact that the Caliphatic−OH
fragments generated from hydrolysis of α-O-4 and β-O-4
linkages can undergo further hydrogenolysis, while phenol (produced by hydrolysis of the 4-O-5 linkage) is hydrogenated to
produce cyclohexanol under conditions investigated. The apparent activation energies, Ea(α-O-4) < Ea(β-O-4) < Ea(4-O-5), vary
proportionally with the bond dissociation energies. In the conversion of β-O-4 and 4-O-5 ether bonds, C−O bond cleavage is the
rate-determining step, with the reactants competing with hydrogen for active sites, leading to a maximum reaction rate as a
function of the H2 pressure. For the very fast C−O bond cleavage of the α-O-4 linkage, increasing the H2 pressure increases the
rate-determining product desorption under the conditions tested.
appreciable rates of conversion and high stability for hydro-
deoxygenation of ethers under moderate conditions (250 °C
and 50 bar H2).8 Similarly, it was shown for Pd-based catalysts
that having acid sites in the proximity of the metal sites led to
high selectivity for hydrodeoxygenation of substituted phenols.9
Combining this insight with the finding that Ni molecular
catalysts can hydrogenolyze aromatic ether bonds,6 we chose to
explore systematically the properties of supported Ni in the
absence of adjacent acid sites on the solid support under much
milder conditions.
The representative fragment structure of hardwood lignin
depicted in Figure 1 illustrates that the β-O-4, α-O-4, and 4-O-
5 linkages are three of the most frequent types of C−O bonds,
contributing to 45−62% (the most abundant), 3−12%, and 4−
9% of the ether bonds in hardwood lignin, respectively, and
would thus be expected to determine the primary products
when large clusters of lignin are depolymerized.10 2-Phenylethyl
phenyl ether, benzyl phenyl ether, and diphenyl ether were
selected as model compounds for the β-O-4, α-O-4, and 4-O-5
linkages, respectively, in our exploration of the reaction
mechanism of C−O bond cleavage in the aqueous phase, as
our previous experiments have shown that more complex
substitutions hardly influence the principal chemistry.11 Here
we show that supported Ni catalysts can selectively and
quantitatively cleave C−O bonds in aromatic ethers such as
benzyl phenyl ether (α-O-4 linkage), 2-phenylethyl phenyl
ether (β-O-4 linkage), and diphenyl ether (4-O-5 linkage) to
form smaller aromatic molecules, cycloalkanes, and cyclo-
INTRODUCTION
■
Lignin is a three-dimensional, highly branched, polyphenolic
substance containing an array of hydroxy- and methoxy-
substituted phenylpropane units.1 As lignin makes up 20−30 wt
% of lignocellulose and especially contains a very high fraction
of its energy content (i.e., ca. 40% in the total lignocellulosic
biomass2), its conversion to energy carriers has recently
received much attention. However, because of the high strength
and stability of its aryl ether bonds (see Figure 1), selective
depolymerization of lignin into small molecules is challenging.3
Recently, a series of interesting molecular catalysts, including
Ru,4 V,5 and Ni6 complexes, has shown high selectivity in
cleaving the C−O bonds of dimeric lignin model compounds to
produce aromatic molecules in organic solvents (CD3CN,
toluene, and m-xylene) under rather mild conditions (80−135
°C, 1 bar H2). In principle, molecular catalysts are logical
choices for depolymerization fo lignin, as their mobility and
flexibility appear to allow them to reach individual ether bonds
without high steric limitation. However, these reported
molecular catalysts are sensitive to large concentrations of
water, which is costly to separate from raw biomass.
More robust solid catalysts would therefore be better choices
for the aqueous-phase conversion of lignin. The strategies
explored to date using heterogeneous catalysts such as Ni/
K2CO3/ZrO2, however, require high temperatures (240−400
°C) and pressures (250−315 bar) to cleave the aromatic ethers
of lignin dissolved in the aqueous phase.7 The achievable yields
of the cleaved products were limited by recombination of the
products via free radical reactions, even when a large excess of
H2 was present. Thus, the higher reactivity of the potential
products necessitates that the reactions be carried out under
very mild conditions. Ni supported on zeolite HZSM-5 showed
Received: October 8, 2012
Published: November 28, 2012
© 2012 American Chemical Society
20768
dx.doi.org/10.1021/ja309915e | J. Am. Chem. Soc. 2012, 134, 20768−20775

Journal of the American Chemical Society
Article
Figure 1. Fragment structure of hardwood lignin.
Figure 2. (A) TEM images of Ni/SiO2 catalysts prepared by the DP method at different calcination temperatures. (B) TOF as a function of Ni
particle size for C−O bond cleavage of 2-phenylethyl phenyl ether (β-O-4) over Ni/SiO2 in water at 120 °C under 6 bar H2.
a
Table 1. Conversion of 2-Phenylethyl Phenyl Ether (β-O-4) Using Ni/SiO2 Catalysts in Water
selectivity (%)
b
c
Tcalcin. (°C)
Ni loading (wt %)
dNi (nm)
TOF (h−1
)
conv. (%)
ethylbenzene
cyclohexanol
cyclohexanone
phenol
400
500
600
700
800
42
41
42
46
42
4.5
4.7
5.4
5.9
7.9
7
12
17
26
20
5.6
12
51
55
50
56
52
18
25
29
31
4.1
2.3
2.1
2.0
3.1
27
18
19
11
26
18
21
8.1
19
a
b
Conditions: β-O-4 (1.98 g), Ni/SiO2 (0.30 g), H2O (80 mL), 120 °C, 6 bar H2, 90 min, stirring at 700 rpm. Ni particle sizes as determined by
c
TEM. The TOFs were normalized by the fractions of accessible Ni atoms on the surface, which were measured by H2 chemisorption to be 0.050,
0.057, 0.051, 0.040, and 0.021 for the Ni/SiO2 samples calcined at increasing temperatures. The reaction conditions were chosen to allow
determination of the rate below 20% conversion.
20769
dx.doi.org/10.1021/ja309915e | J. Am. Chem. Soc. 2012, 134, 20768−20775

Journal of the American Chemical Society
Article
Figure 3. (a) Product distributions for the conversion of 2-phenylethyl phenyl ether (β-O-4) over Ni/SiO2 as functions of time. Reaction conditions:
2-phenylethyl phenyl ether (β-O-4) (1.98 g, 0.010 mol), 57 wt % Ni/SiO2 (0.30 g, 2.91 × 10−3 mol of Ni), H2O (80 mL), 120 °C, 6 bar H2, stirring
at 700 rpm. (b) Reaction pathway for the cleavage of 2-phenylethyl phenyl ether (the model compound for the β-O-4 linkage in lignin) over Ni/
SiO2 in the aqueous phase.
hexanol under very mild conditions (120 °C and 6 bar H2) in
ethylcyclohexane being lower than 2%. The selectivity for
hydrogenated single-aromatic-ring products was below 2%
independent of the conversion. The C−O−C bond was
selectively cleaved at the position of the aliphatic carbon,
producing phenol and ethylbenzene (Figure 3b). The
selectivity for benzene and 2-phenylethanol was less than 1%
even after 600 min. This indicates a preferred interaction of the
aliphatic C−O bond in 2-phenylethyl phenyl ether (β-O-4) on
the Ni surface. In summary, hydrogenolysis is the dominant
reaction pathway for cleavage of 2-phenylethyl phenyl ether (β-
O-4) with Ni/SiO2 in the aqueous phase, affording ethyl-
benzene and phenol as the primary products (Figure 3b).
This result also highlights the importance of the hydroxyl
group of phenol for adsorption on Ni/SiO2 as well as the
different solubilities of phenol (8.3 g/100 mL of H2O at 20 °C)
and ethylbenzene (0.018 g/100 mL of H2O at 20 °C) in hot
water. Separate experiments showed that the rate of phenol
hydrogenation over Ni/SiO2 (130 mol·molNi(surf)−1·h−1) was 2
orders of magnitude higher than that of ethylbenzene (0.8
mol·molNi(surf)−1·h−1) in water (Table 2). In nonpolar
water.
RESULTS AND DISCUSSION
■
Catalyst Synthesis Strategy and Catalytic Measure-
ments. We synthesized five supported Ni/SiO2 catalysts using
the deposition precipitation (DP) method with urea as the
hydrolysis agent. The synthesized Ni/SiO2 catalysts had Ni
loadings of ca. 40 wt %. To alter the Ni particle size, the
calcination temperature was varied from 400 to 800 °C, which
led to gradual growth of Ni particles from 4.5 to 7.9 nm [as
determined by transmission electron microscopy (TEM);
Figure 2A] due to Ni sintering at higher treatment temper-
atures. The hydrogenolysis of 2-phenylethyl phenyl ether (β-O-
4) was conducted over these five Ni/SiO2 catalysts at 120 °C
with 6 bar H2 for 90 min in the aqueous phase. The results
consistently showed that almost equal concentrations of C6
(phenol, cyclohexanone, and cyclohexanol) and C8 (ethyl-
benzene) compounds were formed under the selected
conditions (Table 1). The C−O bond cleavage rate normalized
to accessible Ni atoms [i.e., the turnover frequency (TOF)]
first increased from 7 to 12, 17, and 26 h−1 as the Ni particle
size increased from 4.5 to 4.7, 5.4, and 5.9 nm, respectively, and
then dropped to 20 h−1 with a further increase in particle size to
8 nm (Figure 2B). To obtain the best compromise between
optimum catalytic stability and activity, we selected a catalyst
with larger particles for the in-depth study. The Ni/SiO2
catalyst (57 wt % Ni loading) had a Brunauer−Emmett−Teller
(BET) surface area of 140 m2·g−1 with a pore volume of 0.1826
cm3·g−1 and a particle size of ca. 3 μm as determined by
scanning electron microscopy (SEM). The supported Ni
particles had an average particle diameter of 8.0 1.8 nm as
determined by TEM, and the atomic fraction of Ni available for
catalysis was 5% as measured by H2 chemisorption (see the
Supporting Information).
Table 2. Hydrogenation of Aromatic Compounds on Ni/
a
SiO2 in the Aqueous Phase at 120 °C
catalyst
conv.
(%)
TOF
entry
reactant
phenol
solvent
water
amount (g)
(h−1
130
)
1
2
0.03
0.03
15
b
b
phenol +
ethylbenzene
water
20
173
0
c
c
0
3
4
5
6
7
phenol
hexadecane
0.03
0.3
0.3
0.3
0.3
10
87
13
ethylbenzene hexadecane
16
benzene
toluene
water
water
3.4
2.1
1.0
2.9
1.8
0.8
ethylbenzene water
a
Reaction conditions: reactant (0.010 mol), H2O (80 mL), 6 bar H2,
b
c
50 min, stirring at 700 rpm. For phenol. For ethylbenzene.
Kinetics of 2-Phenylethyl Phenyl Ether (β-O-4)
Conversion. Figure 3a shows the product distribution for 2-
phenylethyl phenyl ether (β-O-4) conversion as a function of
time at 120 °C in the presence of 6 bar H2. Cyclohexanol and
ethylbenzene were the major products at reaction time t = 0,
each being obtained with 50% selectivity during the conversion.
Phenol was hardly detected because it was completely
hydrogenated to cyclohexanol, while ethylbenzene remained
nearly unconverted over Ni/SiO2, with the selectivity for
hexadecane, the rate of phenol hydrogenation (87 mol·-
molNi(surf)−1·h−1) was lower than that in water (130
mol·molNi(surf)−1·h−1) at 120 °C because of the difference in
the solubilities of the reactant. The ethylbenzene hydrogenation
rate in hexadecane (13 mol·molNi(surf)−1·h−1) was still lower
than that of phenol (87 mol·molNi(surf)−1·h−1), indicating that
the increased capability for adsorption on Ni/SiO2 due to the
phenol hydroxyl group plays a more significant role than the
20770
dx.doi.org/10.1021/ja309915e | J. Am. Chem. Soc. 2012, 134, 20768−20775

Journal of the American Chemical Society
Article
Figure 4. (a) Product distributions for the conversion of benzyl phenyl ether (α-O-4) over Ni/SiO2 as functions of time. Reaction conditions: benzyl
phenyl ether (α-O-4) (1.84 g, 0.010 mol), 57 wt % Ni/SiO2 (0.030 g, 2.91 × 10−4 mol of Ni), H2O (80 mL), 120 °C, 6 bar H2, stirring at 700 rpm.
(b) Reaction pathway for the cleavage of benzyl phenyl ether (the model for α-O-4 bond linkage in lignin) over Ni/SiO2 in the aqueous phase.
Figure 5. (a) Product distributions for the conversion of diphenyl ether (4-O-5) over Ni/SiO2 as functions of time. Reaction conditions: diphenyl
ether (4-O-5) (1.70 g, 0.010 mol), 57 wt % Ni/SiO2 (0.30 g, 2.91 × 10−3 mol of Ni), 120 °C, 6 bar H2, stirring at 700 rpm. (b) Reaction pathway for
the cleavage of diphenyl ether (the model for the 4-O-5 linkage in lignin) over Ni/SiO2 in the aqueous phase.
solvent. When phenol and ethylbenzene were added as co-
reactants in water, ethylbenzene was unconverted, but phenol
achieved a high hydrogenation rate with a TOF of 173
mol·molNi(surf)−1·h−1. These results show that phenol is highly
reactive whereas ethylbenzene hardly reacts at all on Ni/SiO2
under the selected conditions.
Kinetics of Benzyl Phenyl Ether (α-O-4) Conversion.
The reaction of benzyl phenyl ether (α-O-4) on Ni/SiO2 at 120
°C with 6 bar H2 (Figure 4a) showed a very high reaction rate,
even in the presence of only a tenth of the catalyst used for 2-
phenylethyl phenyl ether (β-O-4) conversion. The initial TOF
for the C−O cleavage (1017 mol·molNi(surf)−1·h−1) was much
higher than that for the β-O-4 bond (13 mol·molNi(surf)−1·h−1)
(Table 4). The results agree well with the reports that the α-O-
4 bond is thermally very unstable.12 The primary products were
toluene and phenol, each of which was obtained with nearly
50% selectivity at t = 0; as above, these products resulted from
direct hydrogenolysis of the aliphatic C−O bond of benzyl
phenyl ether.
phenyl ether (α-O-4) conversion using Ni/SiO2 in the aqueous
phase is dominated by the hydrogenolysis of benzyl phenyl
ether to toluene and phenol, which is followed by a small extent
of hydrogenation of phenol to cyclohexanone and cyclohexanol
(Figure 4b).
Kinetics of Diphenyl Ether (4-O-5) Conversion. The
conversion of diaryl ethers has been reported to be challenging,
requiring severe conditions such as hydrolysis in supercritical
water to cleave the C−O bond.8 The C−O bond was also
cleaved with Raney Ni in boiling methanol, but the achieved
rate was extremely low (0.01 mol·molNi−1·h−1).13 Diphenyl
ether (4-O-5) was reacted under the same conditions as
reported above for β-O-4 (Figure 5a). The major products were
cyclohexanol (25% selectivity), benzene (10% selectivity), and
cyclohexyl phenyl ether (65% selectivity) at t = 0, suggesting a
different reaction mechanism than for α-O-4 and β-O-4, in
which hydrogenolysis and hydrogenation were sequential. As
the reaction proceeded, the yields of cyclohexanol and benzene
increased steadily, and the yield of cyclohexyl phenyl ether first
increased to a maximum of 13% and then decreased to zero.
Cyclohexanol was formed from phenol at a high rate of 130
mol·molNi(surf)−1·h−1, which was much faster than the rate of
benzene hydrogenation (2.9 mol·molNi(surf)−1·h−1) (Table 2).
The increasing concentrations of phenol and benzene did not
result from the hydrogenolysis of cyclohexyl phenyl ether, as a
separate reaction with cyclohexyl phenyl ether over Ni/SiO2
showed that the C−O bond cleavage rate was relatively low
(0.8 mol·molNi(surf)−1·h−1), while dicyclohexyl ether did not
The phenol hydrogenation rate (130 mol·molNi(surf)−1·h−1)
was much higher than that of toluene (1.8 mol·molNi(surf)−1·h−1)
(Table 2). The very low catalyst concentration, however, was
sufficient to catalyze the ether hydrogenolysis effectively but
not the hydrogenation of the phenol aromatic ring. When the
catalyst concentration was increased to that used for the β-O-4
conversion (0.30 g), 46% selectivity for toluene and 54%
selectivity for cyclohexanol at 93% conversion was observed
within 50 min. In summary, the reaction pathway for benzyl
20771
dx.doi.org/10.1021/ja309915e | J. Am. Chem. Soc. 2012, 134, 20768−20775

Journal of the American Chemical Society
Article
react under the conditions used (Table 3). The rate of diphenyl
ether cleavage (26 mol·molNi(surf)−1·h−1) was much higher, and
we conclude that the hydrogenation/hydrogenolysis route is
only a minor pathway (Figure 5b).
Relation of the Thermodynamic Properties of the
Reactant to the Kinetic Parameters. After exploring the
differences in the kinetic behavior of the three model
compounds on Ni/SiO2, we next explored how the strength
of the ether bond, characterized by the bond dissociation
energy (BDE),14 correlates with the kinetic parameters (Table
4). The aryl ether bond (BDE = 314 kJ·mol−1) of 4-O-5 is
much stronger than the aliphatic ether bonds of α-O-4 (218
kJ·mol−1) and β-O-4 (289 kJ·mol−1). Thus, hydrogenolysis
favors selective cleavage of the weaker aliphatic ether bonds.
The aliphatic ether C−O bond of α-O-4 is weaker than the aryl
ether C−O bond of 4-O-5 by 96 kJ·mol−1 and is more easily
cleaved under hydrothermal conditions above 250 °C even in a
blank test (see Figure S6 in the Supporting Information). The
low BDE of α-O-4 (218 kJ·mol−1) leads to a very high initial
TOF (1017 mol·molNi(surf)−1·h−1) and a lower apparent
activation energy (Ea = 72 kJ·mol−1), as determined from an
Arrhenius plot (Figure 6a). By comparison, the higher BDE of
the aliphatic ether bond of β-O-4 (289 kJ·mol−1) leads to a low
TOF (13 mol·molNi(surf)−1·h−1) and a much higher Ea (86
kJ·mol−1) (Figure 6a).
Relative to α-O-4 and β-O-4, 4-O-5 showed very different
reactivity on Ni/SiO2. In principle, the C−O bond of the 4-O-5
linkage should be the most difficult ether bond to cleave, as it
has the highest BDE (314 kJ·mol−1). Previous work would also
support this conclusion.8,11 The fact that 4-O-5 has the highest
Ea (97 kJ·mol−1) agrees with the fact that it has the highest
BDE. However, the reaction rate for 4-O-5 (26
mol·molNi(surf)−1·h−1) was significantly higher than that for β-
O-4 (13 mol·molNi(surf)−1·h−1) with the Ni/SiO2 catalysts. This
may be attributed to the different C−O bond cleavage
mechanisms for 4-O-5 and β-O-4. The C−O bonds in benzyl
phenyl ether (α-O-4) and 2-phenylethyl phenyl ether (β-O-4)
were directly cleaved by Ni catalyzed hydrogenolysis. However,
the C−O bond of 4-O-5 was cleaved by parallel hydrogenolysis
and hydrolysis, in which hydrolysis appears to increase the
overall conversion rate. It should be noted that the apparent
activation energies for cleavage of α-O-4, β-O-4, and 4-O-5
increased almost linearly with their C−O BDEs (Figure 6b),
indicating that the aryl C−O bond cleavage barrier is heavily
influenced by the C−O BDE.
Table 3. Rates of C−O Bond Cleavage in Diphenyl Ether (4-
O-5), Cyclohexyl Phenyl Ether, and Dicyclohexyl Ether at
a
120 °C
a
Reaction conditions: reactant (0.010 mol), 57 wt % Ni/SiO2 (0.30 g,
2.91 × 10−3 mol of Ni), H2O (80 mL), 6 bar H2, 110 min, stirring at
700 rpm. CPE = cyclohexyl phenyl ether.
b
The selectivity for cyclohexanol (70%) was 2.3 times higher
than the selectivity for benzene (30%) at complete conversion.
Considering that the selectivity for cyclohexanol (formed from
phenol hydrogenation) was much higher than that for benzene,
we therefore conclude that hydrolysis of diphenyl ether (4-O-5)
occurred in parallel to hydrogenolysis, producing two molecules
of phenol. Thus, cyclohexanol was obtained from diphenyl
ether via two routes, one relying on sequential hydrolysis/
hydrogenation and the other replying on sequential hydro-
genolysis/hydrogenation (Figure 5b). It is estimated that the
ratio of the hydrogenolysis rate to the hydrolysis rate is ca. 3:2.
In summary, the overall reaction pathway for diphenyl ether
conversion (Figure 5b) follows two major routes: (i)
hydrogenolysis of diphenyl ether to benzene and phenol,
after which phenol is rapidly hydrogenated to cyclohexanol, and
(ii) hydrolysis of diphenyl ether to two molecules of phenol,
which in turn are hydrogenated to cyclohexanol. Another route
of minor importance is the hydrogenation of diphenyl ether to
cyclohexyl phenyl ether, which in turn is hydrolyzed or
hydrogenolyzed to cyclohexanol, benzene, and phenol. Phenol
is in turn hydrogenated to cyclohexanol, and a small fraction of
benzene is hydrogenated to cyclohexane.
To explore the specific contributions of the Ni/SiO2
components in water, the C−O cleavage rates for the three
model compounds were compared with those in pure water,
with SiO2, and with Ni/SiO2 in the presence of 6 bar H2 at 120
°C (Table 4). None of the ethers reacted in pure water or in
the presence of SiO2 alone. With the Ni/SiO2 catalyst, the C−
Table 4. Bond Dissociation Energies (BDEs), Initial Turnover Frequencies (TOFs) of C−O Cleavage, and Apparent Activation
Energies (Ea) for C−O Bond Cleavage in Lignin Model Compounds
a
Data from ref 14.
20772
dx.doi.org/10.1021/ja309915e | J. Am. Chem. Soc. 2012, 134, 20768−20775

Journal of the American Chemical Society
Article
Figure 6. (a) Arrhenius plots [ln(TOF) vs 1/T] for the three model compounds. Reaction conditions: H2O (80 mL); 6 bar H2; 50 min; and either
(i) α-O-4 (1.84 g), 57 wt % Ni/SiO2 (0.03 g, 2.91 × 10−4 mol of Ni); (ii) β-O-4 (1.98 g), 57 wt % Ni/SiO2 (0.30 g, 2.91 × 10−4 mol of Ni); or (iii)
4-O-5 (1.70 g), 57 wt % Ni/SiO2 (0.30 g, 2.91 × 10−4 mol of Ni). (b) Relationship between apparent Ea and BDE for the three model compounds.
O bond cleavage rates for α-O-4 (1017 mol·molNi(surf)−1·h−1),
β-O-4 (13 mol·molNi(surf)−1·h−1), and 4-O-5 (26 mol·-
molNi(surf)−1·h−1) increased dramatically, verifying that the Ni
center is the active site for C−O bond cleavage.
Impact of the H2 Pressure. The product distributions for
the β-O-4, α-O-4, and 4-O-5 model compounds were recorded
at conversions lower than 15% over Ni/SiO2 at 120 °C with
varying H2 pressure (Figures S7−S9 in the Supporting
Information). In the absence of H2, none of the model
compounds reacted. Thus, the presence H2 is also required for
the hydrolysis route of diphenyl ether (4-O-5). With increasing
respectively (Figure 7). This causes product desorption to be
rate-determining at lower PH , and a higher H2 pressure
2
facilitates product removal via hydrogen addition and hydro-
gen-aided desorption. When PH was increased to 100 bar, the
2
TOF for C−O bond cleavage of α-O-4 decreased because of
the decrease in reactant coverage derived from the competition
between reactant and H2 for adsorption on the Ni sites.
Variation of the H2 pressure also led to differences in the
hydrogenation pathways of the model compounds. With
increasing PH , the selectivity for hydrogenated single-
2
aromatic-ring products increased from 0 to 45 and 20% for
β-O-4 and 4-O-5, respectively (Figures S7 and S9 in the
Supporting Information). However, a hydrogenated single-
aromatic-ring product was not observed for α-O-4 even at 100
bar H2 (Figure S8 in the Supporting Information). This implies
the following: (i) Hydrogenolysis and hydrogenation are
competing reactions, and higher H2 pressure favors the
hydrogenation of a single aromatic ring of 4-O-5 and β-O-4.
(ii) α-O-4 shows a different behavior because its hydrogenolysis
barrier (72 kJ·mol−1) is much lower than the Ea for
hydrogenation of the phenolic ring over Ni/SiO2 (ca. 100
kJ·mol−1),17 and therefore, hydrogenolysis dominates over
hydrogenation at low temperatures.
PH , the rates of C−O bond cleavage in β-O-4 and 4-O-5 first
increased to maxima at 6 and 10 bar H2, respectively, and then
gradually decreased (Figure 7). The C−O bond cleavage on Ni
2
Comparison of the Mechanisms of C−O Bond
Cleavage for the Three Model Compounds. On the
basis of the above results and discussion, the α-O-4 and β-O-4
Caliphatic−O−Caromatic linkages are cleaved on the Caliphatic−O
side by the Ni-catalyzed hydrogenolysis. However, for 4-O-5,
the Caromatic−O−Caromatic linkage is cleaved by hydrogenolysis
and hydrolysis in parallel. It has been shown that hydrolysis of
the 4-O-5 linkage does not occur in the absence of H2 (Figure
7) or Ni/SiO2 (Table 4). Hydrolysis is hypothesized to occur
along the same reaction path as hydrogenolysis, with cleavage
by Ni and subsequent addition of H· and OH· (from water
dissociation) instead of addition of two H· (Figure 8B). For the
α-O-4 and β-O-4 compounds, the weaker Caliphatic−O bonds
rather than the stronger Caromatic−O bonds are cleaved (Figure
8A). As the BDE of the Caliphatic−O bond in the α-O-4 linkage
(218 kJ·mol−1) is much lower than that in the β-O-4 linkage
(289 kJ·mol−1), the rate of hydrogenolysis of the α-O-4 linkage
over Ni/SiO2 is 2 orders of magnitude higher than that of β-O-
4 linkage (Table 4).
Figure 7. TOFs for the three model compounds over Ni/SiO2 as
functions of H2 pressure. Detailed conditions are given in the
Supporting Information.
is the rate-determining step,15 so these results suggest that H2
competes with the organic reactant for adsorption sites. Up to
medium PH , the increasing surface coverage leads to an
2
increased hydrogenolysis rate, while at high PH , the rate is
2
retarded because of the lower ether coverage.16 Thus, Figure 7
represents a typical change in the hydrogenolysis rate for β-O-4
and 4-O-5.
In contrast, the rate of C−O bond cleavage for α-O-4
increased with PH up to 80 bar. This is attributed to the fact
Once the Caliphatic−O bonds in the α-O-4 and β-O-4 linkages
are cleaved into Caliphatic· and ·O−Caromatic fragments, the
abundant dissociated H· atoms on the Ni surface are in turn
added to form methyl- or ethylbenzene, respectively, and
2
that Ni/SiO2 leads to a much higher rate of C−O bond
cleavage in α-O-4, with the maximum cleavage rate being 150
and 70 times faster than those for α-O-4 and 4-O-5,
20773
dx.doi.org/10.1021/ja309915e | J. Am. Chem. Soc. 2012, 134, 20768−20775

Journal of the American Chemical Society
Article
and β-O-4 linkages to give PhCH3 and PhCH2CH3,
respectively, while hydrogenation converts phenol derived
from the 4-O-5 linkage to cyclohexanol under the present
conditions.
The rates for C−O bond cleavage decreased in the order (α-
O-4) > (4-O-5) > (β-O-4), and the apparent energies of
activation increased in the order (α-O-4) < (β-O-4) < (4-O-5),
tracking with the variation in the bond dissociation energies,
(α-O-4) < (β-O-4) < (4-O-5). The C−O bond of 4-O-5 is
cleaved by parallel hydrogenolysis and hydrolysis, and the
additional hydrolysis route accelerates the conversion. In the
hydrogenolysis of β-O-4 and 4-O-5, an optimal H2 pressure
maximized the reaction rates for C−O bond cleavage, because
the organic reactant competes with H2 for the active sites. In
contrast, the rate of C−O bond cleavage increased continuously
up to 80 bar for α-O-4 conversion, probably because increasing
PH facilitates product desorption from the Ni surface (the rate-
2
determining step for α-O-4 at low PH ). As PH is further
2
2
increased, the TOF of C−O bond cleavage of α-O-4 decreases
because of the lower reactant coverage resulting from the
competition between reactant and H2 adsorption on the Ni
sites.
The results indicate that Ni-based solid catalysts are
interesting and stable for ether cleavage and exhibit significantly
higher rates than related molecular catalysts. Tailoring the Ni
catalyst via variation of the particle size and eventually the
support is expected to lead to a new group of catalysts for
cleaving carbon−heteroatom bonds.
Figure 8. Proposed mechanisms for cleavage of C−O bonds in the α-
O-4, β-O-4, and 4-O-5 model compounds.
phenol (Figure 8A). As the BDEs of PhCH2CH2−H and
PhCH2CH2−OH are 367 and 333 kJ·mol−1,14a respectively, it
appears that PhCH2CH2· prefers to combine with H· but not
OH· (from water dissociation) because of the much higher (by
34 kJ·mol−1) bond formation energy for this process, which
makes hydrogenolysis (H· addition) dominant. In addition, the
BDE of Ph−H (477 kJ·mol−1) is also much higher than that of
Ph−OH (437 kJ·mol−1),14a so hydrogenolysis rather than
hydrolysis contributes a substantial fraction to the overall 4-O-5
conversion (Figure 5). In essence, hydrolysis combined with
OH· and H· addition could occur on all three model
compounds in water, but the low concentrations of PhCH2OH
and PhCH2CH2OH produced can be further hydrogenolyzed
to PhCH3 and PhCH2CH3, respectively (Table S1 in the
Supporting Information), while hydrogenation to produce
cyclohexanol is the major reaction of phenol derived from the
4-O-5 linkage under the present conditions (Table 2).
Therefore, hydrogenolysis is the major reaction pathway for
the overall conversions of the α-O-4 and β-O-4 linkages, and
parallel hydrogenolysis/hydrolysis is the major route for
conversion of the 4-O-5 linkage with Ni/SiO2 catalysts in water.
MATERIALS AND METHODS
■
Materials. The chemicals were purchased from commercial
suppliers and used as provided: benzyl phenyl ether (TCI, >98%
GC assay), 2-phenylethyl phenyl ether (Frinton Laboratories, >99%
GC assay), diphenyl ether (Sigma-Aldrich, >99% GC assay),
cyclohexyl phenyl ether (Sigma-Aldrich, >95% GC assay), benzene
(Fluka, >99.5% GC assay), toluene (Sigma-Aldrich, >99.5% GC
assay), ethylbenzene (Fluka, >99% GC assay), phenol (Sigma-Aldrich,
>99% GC assay), ethyl acetate (Sigma-Aldrich, >99.5% GC assay),
nickel(II) nitrate hexahydrate (Sigma-Aldrich, ≥98.5%), urea (Sigma-
Aldrich, BioReagent), HNO3 (Sigma-Aldrich, >65%), 5 wt % Pd/C
(Sigma-Aldrich), SiO2 (Aerosil 200, Evonik-Degussa), H2 (Air
Liquide, >99.999%), N2 (Air Liquide, > 99.999%).
Ni/SiO2 Catalyst Preparation Using the DP Method. An
aqueous solution (250 mL) containing Ni(NO3)2·6H2O (0.14 M, 10.2
g) was divided in two parts. To one part (50 mL) was added urea
(0.42 M, 6.3 g), and the other part (200 mL) together with SiO2 (1.9
g) and HNO3 (65%, 0.02 M, 0.32 mL) was put into a flask
thermostatted at 80 °C. The urea solution was then slowly added to
the flask, and the suspension was rapidly heated to 90 °C, after which
the suspension was magnetically stirred for 10 h. The suspension was
then cooled to 25 °C, and the solids were filtered and washed three
times with distilled water (5/1 water/slurry). Finally, the sample was
dried at 90 °C for 24 h; calcined in flowing air (100 mL·min−1) at 400,
500, 600, 700, or 800 °C; and reduced in flowing H2 (100 mL·min−1)
at 460 °C.
Catalytic Test. The detailed reaction conditions are described in
the figure captions and table footnotes. In a typical experiment, the
catalytic reactions were carried out in a slurry autoclave reactor loaded
with Ni/SiO2 using water as the solvent at 120 °C in the presence of 6
bar H2. The lignin model compound (0.010 mol), 57 wt % Ni/SiO2
(0.30 g, 2.91 × 10−3 mol of Ni for β-O-4 and 4-O-5 or 0.030 g, 2.91 ×
10−4 mol of Ni for α-O-4), and H2O (80 mL) were charged into a Parr
reactor (Series 4848, 300 mL). After the reactor was flushed with H2
three times, the autoclave was charged with 6 bar H2, and the reaction
was conducted at 120 °C with a stirring speed of 700 rpm. Because it is
a two-phase reaction, the kinetic data were collected at time durations.
CONCLUSION
■
The C−O bond linkages of α-O-4, β-O-4, and 4-O-5 model
compounds representing typical bonds in lignin were selectively
cleaved by a silica-supported Ni catalyst to produce aromatic
and aliphatic molecules as well as cyclohexanol in aqueous
media under mild conditions (120 °C, 6 bar H2). The C−O
bonds of α-O-4 and β-O-4 are cleaved by direct hydrogenolysis,
which has been shown to be structure-sensitive on Ni-based
catalysts, while the C−O bond of 4-O-5 appear to be cleaved by
parallel hydrogenolysis and hydrolysis. The difference is
attributed to the hydrogenolysis of the low concentrations of
PhCH2OH and PhCH2CH2OH intermediates from the α-O-4
20774
dx.doi.org/10.1021/ja309915e | J. Am. Chem. Soc. 2012, 134, 20768−20775

Journal of the American Chemical Society
Article
After the reaction, the reactor was quenched to ambient temperature
using ice, and the organic products were extracted using ethyl acetate
and analyzed by gas chromatography (GC) and GC−mass spectros-
copy (GC−MS) analysis on a Shimadzu 2010 gas chromatograph with
flame ionization detector and a Shimadzu QP 2010S GC−MS
instrument, both equipped with HP-5 capillary columns (30 m × 250
μm). 2-Isopropylphenol was used as an internal standard to calibrate
the liquid product concentrations and carbon balances. The carbon
balances for all of the reported experiments were better than 95 3%.
The calculations of conversion and selectivity were on a carbon mole
basis. Conversion is defined as the amount of change in raw materials
during the reaction divided by the total amount of starting material,
multiplied by 100%. The selectivity is defined as the number of C
atoms in the product of interest divided by the total number of C
atoms in the products, multiplied by 100%.
school NanoCat). C.Z. is grateful for the support from
European Graduate School for Sustainable Energy.
REFERENCES
■
(1) Stocker, M. Angew. Chem., Int. Ed. 2008, 47, 9200−9211.
̈
(2) (a) Regalbuto, J. R. Science 2009, 325, 822−824. (b) Perlack, R.
D.; Wright, L. L.; Turhollow, A. F.; Graham, R. L.; Stokes, B. J.;
Erbach, D. C. Biomass as Feedstock for a Bioenergy and Bioproducts
Industry: The Technical Feasibility of a Billion-Ton Annual Supply; U.S.
Department of Energy: Washington, DC, 2005. (c) Huber, G. W.;
Corma, A. Angew. Chem., Int. Ed. 2007, 46, 7184−7201. (d) Huber, G.
W.; Iborra, S.; Corma, A. Chem. Rev. 2006, 106, 4044−4098.
(3) (a) Ragauskas, A. J.; Williams, C. K.; Davison, B. H.; Britovsek,
G.; Cairney, J.; Eckert, C. A.; Frederick, W. J., Jr.; Hallett, J. P.; Leak,
D. J.; Liotta, C. L.; Mielenz, J. R.; Murphy, R.; Templer, R.;
Tschaplinski, T. Science 2006, 311, 484−489. (b) Chang, M. C. Y.
Curr. Opin. Chem. Biol. 2007, 11, 677−684. (c) Sanderson, K. Nature
2011, 474, S12−S14. (d) Matson, T. D.; Barta, K.; Iretskii, A. V.; Ford,
P. C. J. Am. Chem. Soc. 2011, 133, 14090−14097.
Catalyst Characterization. Atomic Absorption Spectroscopy. A
UNICAM 939 atomic absorption spectrometer was used to determine
the concentration of nickel in Ni/SiO2. Before measurement, a mixture
of 0.5 mL of 48 wt % hydrofluoric acid and 0.1 mL of
nitrohydrochloric acid was used to dissolve 20−40 mg of the sample
at 383 K (the boiling point of the mixture).
(4) Nichols, J. M.; Bisshop, L. M.; Bergman, R. G.; Ellman, J. A. J.
Am. Chem. Soc. 2010, 132, 12554−12555.
NMR Spectroscopy. The 1H, 13C, and correlation spectroscopy
(COSY) NMR measurements were performed on a JEOL 400 MHz
spectrometer. CDCl3 was used as the solvent.
(5) Son, S.; Toste, F. D. Angew. Chem., Int. Ed. 2010, 49, 3791−3794.
(6) Sergeev, A. G.; Hartwig, J. F. Science 2011, 332, 439−443.
(7) (a) Roberts, V. M.; Knapp, R. T.; Li, X.; Lercher, J. A.
ChemCatChem 2010, 2, 1407−1410. (b) Roberts, V. M.; Stein, V.;
Reiner, T.; Li, X.; Lemonidou, A.; Li, X.; Lercher, J. A. ChemEur. J.
2011, 17, 5939−5948. (c) Roberts, V. M.; Fendt, S.; Reiner, T.; Li, X.;
Lemonidou, A. A.; Li, X.; Lercher, J. A. Appl. Catal., B 2010, 95, 71−
77.
BET Analysis. The BET surface areas and pore diameters were
determined from N2 adsorption measurements carried out at 77 K on
a PMI automated BET sorptometer. The samples were outgassed at
523 K for 20 h before measurement.
H2 Chemisorption. The catalysts were activated at 733 K for 3 h in
H2 and 1 h in vacuum and then cooled to 313 K. H2 adsorption
(chemisorption and physisorption) was then measured over the
pressure range from 1 to 40 kPa. Next, the physisorbed H2 was
removed by outgassing the sample at the same temperature for 1 h,
and another isotherm (physisorption) was measured. The concen-
tration of chemisorbed hydrogen on the metal was determined by
extrapolating the differential isotherm to zero H2 pressure, and this
value was used to calculate the Ni dispersion.
Transmission Electron Microscopy. TEM images were measured
on a JEOL JEM-2010 transmission electron microscope operating at
120 kV. Before measurement, the catalyst was ground, suspended in
ethanol, and dispersed by ultrasonic treatment. The obtained
dispersion was transferred to a copper-grid-supported carbon film.
TEM images were recorded with a magnification of 200 000. The
average cluster size was calculated by counting 300 metal particles.
Scanning Electron Microscopy. SEM images were recorded on a
JEOL REM 5900 LV microscope operating at 25 kV. The powder
samples were used without any pretreatment.
(8) Zhao, C.; Lercher, J. A. Angew. Chem., Int. Ed. 2012, 51, 5935−
5941.
(9) Zhao, C.; Lercher, J. A. ChemCatChem 2012, 4, 64−68.
(10) (a) Zakzeski, J.; Bruijinincx, P. C. A.; Jongerius, A. L.;
Weckhuysen, B. M. Chem. Rev. 2010, 110, 3552−3599. (b) Furimsky,
E. Appl. Catal., A 2000, 199, 147−190.
(11) (a) Zhao, C.; Kou, Y.; Lemonidou, A. A.; Li, X.; Lercher, J. A.
Chem. Commun. 2010, 46, 412−414. (b) Zhao, C.; Kou, Y.;
Lemonidou, A. A.; Li, X.; Lercher, J. A. Angew. Chem., Int. Ed. 2009,
48, 3987−3990. (c) Zhao, C.; Camaioni, D. M.; Lercher, J. A. J. Catal.
2012, 288, 92−103. (d) Zhao, C.; He, J.; Lemonidou, A. A.; Li, X.;
Lercher, J. A. J. Catal. 2011, 280, 8−16. (e) Zhao, C.; Kazacov, S.; He,
J.; Lercher, J. A. J. Catal. 2012, 296, 12−23. (f) Zhao, C.; Song, W.;
Lercher, J. A. ACS Catal. 2012, 2, 2714−2723.
(12) (a) Dorrestijn, E.; Laarhoven, L.; Arends, I.; Mulder, P. J. Anal.
Appl. Pyrolysis 2000, 54, 153−192. (b) Rylander, P. N. Hydrogenation
Methods; Academic Press: London, 1985; pp 157−163.
(13) Chandler, G.; Sasse, W. Aust. J. Chem. 1963, 16, 20−30.
(14) (a) Luo, Y. R. In Comprehensive Handbook of Chemical Bond
Energies; CRC Press: Boca Raton, FL, 2007. (b) Jarvis, M. W.; Daily, J.
W.; Carstensen, H.; Dean, A. M.; Sharma, S.; Dayton, D. C.;
Robichaud, D. J.; Nimlos, M. R. J. Phys. Chem. A 2011, 115, 428−438.
(15) Shang, S. B.; Kenney, C. N. J. Catal. 1992, 134, 134−150.
(16) Bond, G. C.; Cunningham, R. H. J. Catal. 1997, 166, 172−185.
(17) Hichri, H.; Accary, A.; Andrieu, J. Chem. Eng. Process. 1991, 30,
133−140.
ASSOCIATED CONTENT
■
S
* Supporting Information
SEM and TEM images of the catalyst; synthesis and H, 13C,
1
and COSY NMR characterization of dicyclohexyl ether;
product yields for the conversion of benzyl phenyl ether in
the absence of H2; and product distributions with varying H2
pressure. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
chenzhao@mytum.de; johannes.lercher@ch.tum.de.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
J.H. gratefully acknowledges support from the graduate school
(Faculty Graduate Center of Chemistry) of Technische
■
Universitat Munchen and Elitenetzwerk Bayern (graduate
̈
̈
20775
dx.doi.org/10.1021/ja309915e | J. Am. Chem. Soc. 2012, 134, 20768−20775