Determining the Catalyst Properties That Lead to High Activity and Selectivity for Catalytic Hydrodeoxygenation with Ruthenium Pincer Complexes

Article abstract of DOI:10.1021/acs.organomet.9b00816

Ten ruthenium pincer complexes were evaluated as catalysts for the hydrodeoxygenation (HDO) reaction on a lignin monomer surrogate, vanillyl alcohol. Four of these complexes are reported herein with the synthesis and full characterization data for all and single-crystal X-ray diffraction data for three complexes bearing OH/O^-, NMe2, and Me substituents on the pincer. A systematic study of these CNC pincer complexes revealed that the π-donor substituent on the pyridine ring plays a key role in enhancing the yield of the desired deoxygenated product. While OMe, OH, and NMe2 are all effective as π-donor substituents on the central pyridine ring in the pincer, the highest conversion to products and the best selectivity was observed with OH substituents and added sodium carbonate as a base. Base serves to deprotonate the OH group and form 1^O- as observed spectroscopically. Furthermore, efforts to use other catalysts have revealed that free or labile sites are needed on the ruthenium center and an electronically rich and nonbulky CNC pincer is optimal. At low catalyst loadings (0.01 mol %), the OH-substituted catalyst 1^OH in the presence of base serves as a homogeneous catalyst and is able to achieve quantitative and selective conversion of vanillyl alcohol to desired the HDO product, creosol, with up to 10000 turnovers. With this knowledge in hand, we can design the next generation of homogeneous catalysts with increased reactivity toward all of the oxygenated sites on lignin-derived monomers.

Full text of DOI:10.1021/acs.organomet.9b00816

pubs.acs.org/Organometallics
Article
Determining the Catalyst Properties That Lead to High Activity and
Selectivity for Catalytic Hydrodeoxygenation with Ruthenium Pincer
Complexes
Wenzhi Yao, Sanjit Das, Nicholas A. DeLucia, Fengrui Qu, Chance M. Boudreaux, Aaron K. Vannucci,*
and Elizabeth T. Papish*
Cite This: https://dx.doi.org/10.1021/acs.organomet.9b00816
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Ten ruthenium pincer complexes were evaluated as
catalysts for the hydrodeoxygenation (HDO) reaction on a lignin
monomer surrogate, vanillyl alcohol. Four of these complexes are
reported herein with the synthesis and full characterization data for
all and single-crystal X-ray diffraction data for three complexes
bearing OH/O−, NMe2, and Me substituents on the pincer. A
systematic study of these CNC pincer complexes revealed that the
π-donor substituent on the pyridine ring plays a key role in
enhancing the yield of the desired deoxygenated product. While
OMe, OH, and NMe2 are all effective as π-donor substituents on
the central pyridine ring in the pincer, the highest conversion to
products and the best selectivity was observed with OH
‑
substituents and added sodium carbonate as a base. Base serves to deprotonate the OH group and form 1O as observed
spectroscopically. Furthermore, efforts to use other catalysts have revealed that free or labile sites are needed on the ruthenium
center and an electronically rich and nonbulky CNC pincer is optimal. At low catalyst loadings (0.01 mol %), the OH-substituted
catalyst 1OH in the presence of base serves as a homogeneous catalyst and is able to achieve quantitative and selective conversion of
vanillyl alcohol to desired the HDO product, creosol, with up to 10000 turnovers. With this knowledge in hand, we can design the
next generation of homogeneous catalysts with increased reactivity toward all of the oxygenated sites on lignin-derived monomers.
isolation of important industrial chemical feedstocks.12
Selectively deoxygenating lignin-derived compounds without
hydrogenation of the aromatic units is of specific interest
because aromatics and alkenes are higher value chemicals in
INTRODUCTION
■
Total world energy consumption was 575 quadrillion Btu in
2015, and it is predicted to increase to 736 quadrillion Btu in
2040.1 In addition, renewable fuels are predicted to be the
comparison to alkanes, hydrogen use efficiency would be
world’s fastest growing energy source, with consumption
maximized, and carbon loss would be minimized.13 Traditional
increasing by an average 2.3% per year between 2015 and
nanoparticle-based heterogeneous catalysts can achieve upward
2040. Renewables contributed 19% to the total energy
consumption in 2012.2 The largest fraction of renewables is
traditional biomass, which contributed 9% to the total energy
consumption. To convert biomass to a more useful energy
form, there are three main thermal processes available: namely,
pyrolysis, gasification, and combustion.3 The main product
from pyrolysis of biomass is bio-oil. Bio-oil has problems of
thermal instability, affinity for water, corrosivity, high viscosity,
and low heating values, due to its high oxygen content.4 To
expand the utility of the bio-oil, selective deoxygenation can be
applied to reduce oxygen from the compounds. Furthermore,
recent advances in biomass processing and lignin depolyme-
rization have led to a greatly increased need for catalysts
capable of selective deoxygenation of aromatic alcohols.5−10
Deoxygenation of the aromatic alcohols would increase the
energy density of the resulting liquid fuel11 and/or lead to the
of 80−90% selectivity for the hydrodeoxygenation of model
compounds.14−20 Of these model compounds, vanillyl alcohol
has been previously studied as a commonly derived chemical
from lignin depolymerization. Heterogeneous Pd nanoparticle
catalysts have exhibited good product selectivity for the
formation of creosol depending on the reaction additives as
shown in Scheme 1.21−23 Molecular catalysts have also been
examined for catalytic HDO of benzylic alcohols due to the
fact that molecular catalysts lack extended metallic surfaces and
Received: December 2, 2019
https://dx.doi.org/10.1021/acs.organomet.9b00816
© XXXX American Chemical Society
Organometallics XXXX, XXX, XXX−XXX
A

Organometallics
pubs.acs.org/Organometallics
Article
ligands plays an important role in the catalytic activity of these
catalysts, and this work thus lays the groundwork for the
rational design of future molecular HDO catalysts.
Scheme 1. Hydrodeoxygenation of Organic Substrates with
Heterogeneous Catalysts and Ruthenium Pincer Catalysts
RESULTS AND DISCUSSION
■
Catalyst Structures. Catalysts of type 1R (Scheme 2)
contain a CNC pincer featuring an imidazole-based NHC ring
bonded to a pyridine derivative. The advantage of this class of
catalysts is ease of synthetic preparation and ability to vary the
R group on the pyridine ring. The para position on the
pyridine ring (R) can be OMe or H, as described in our
published work,29 or it can be OH, NMe2, or Me as recently
synthesized and characterized herein (further details below and
in Scheme 3). For catalysts of type 2R, the imidazole-based
NHC ring is replaced with benzimidazole, which weakens the
NHC donor strength.27−29,31 Complex 2H is previously
unreported and was characterized by H NMR, 13C NMR,
1
19F NMR, MS, and IR methods (see the Supporting
Information). Catalyst 3 builds upon 2OMe by replacing methyl
wingtips with phenyl wingtips on the NHC rings. Catalysts 4A
and 4B replace two acetonitrile ligands with a bidentate chelate
on ruthenium and were synthesized by adding 2,2′-bipyridine
(bipy) to 3 using a multistep route described previously.30
Four of these catalysts (2OMe, 3, 4A,B) were synthesized and
characterized in a recent paper on photocatalytic self-sensitized
CO2 reduction to form CO.29 On comparison of some of these
catalysts to each other, the donor strength of the pincer (and
extent of metal to ligand back-bonding from Ru to NCCH3 by
IR spectroscopy) decreases in the order 1OMe > 2OMe > 3 with
the same substituents on Ru and on the pyridine of the pincer.
Furthermore, the presence of π-donor R groups (e.g., OH,
OMe) results in a more electron rich pincer.
thus can avoid unwanted ring hydrogenation products.24
A
molecular palladium catalyst in homogeneous methanol
solution has exhibited complete selectivity for HDO over
ring hydrogenation for benzylic substrates,25 and a molecular
catalyst attached to the surface of oxide particles has exhibited
high selectivity and activity toward the formation of creosol
from vanillyl alcohol and vanillin.26
Here we examine the ability of a series of molecular
ruthenium catalysts (Scheme 2)27−30 to perform selective
HDO on vanillyl alcohol. The coordination environment
around the Ru center and the electron-donating ability of the
catalysts were systematically varied to gain an understanding of
the catalyst reactivity and how it depends upon ligand design.
The results show that the electron donor strength of the
Synthesis of 1OH, 1NMe , and 1Me. The pincer ligand
2
precursors were synthesized in two steps by treating 4-R-2,6-
dihalopyridine (R = Me, NMe2) with deprotonated imidazole
followed by methylation with methyl triflate. The bis-
(imidazolium) salt is then treated with base in acetonitrile to
generate the free carbene in situ, and metalation is achieved
with [Ru(p-cym)Cl2]2, leading to complex 1R (Scheme 3).
This procedure has been used previously to form 1OMe and 1H,
but we recently found that using triethylamine as the base (vs
Cs2CO3) in step c led to a cleaner reaction and an improved
yield.
Scheme 2. Catalyst Structures Tested for
Hydrodeoxygenation of Organic Substrates
The synthesis of 1OH proved more challenging because
NHC ligands are known to be sensitive to protic groups. Thus,
the OH functionality needed to be masked with a protecting
group until after metalation was achieved. As shown in the
Supporting Information, using a benzyl protecting group with
the starting material 4-OCH2Ph-2,6-difluoropyridine led to
1OCH Ph (following the method in Scheme 3), which then could
2
be cleanly deprotected by hydrogenation with Pd on C in
acetonitrile to produce 1OH. Further experimental and full
characterization details for all new compounds are shown in
the Supporting Information.
Crystal Structures. The crystal structures for complexes
OCH2Ph
1Me, 1NMe , and 1
are shown in Figure 1. All of these
2
complexes adopt a distorted-octahedral geometry with the
CNC pincer occupying a meridional plane. Complex 1OH was
recrystallized from acetonitrile and diethyl ether, leading to a
‑
dimer with one molecular unit deprotonated (1OH + 1O ), as
shown in Figure 1. The OH-bearing pincer ligand is expected
to be acidic, and thus this result is not surprising. For
B
https://dx.doi.org/10.1021/acs.organomet.9b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
pubs.acs.org/Organometallics
Article
a
Scheme 3. General Synthetic Procedure for 1R
a
Reagents and solvents: (a) 1H-imidazole, base, DMF; (b) methyl trifluoromethanesulfonate, DMF; (c) [Ru(p-cym)Cl2]2, base, acetonitrile.
Figure 1. Molecular diagrams of complexes 1Me1, 1NMe 1, 1OCH Ph1, and (1OH + 1O−) based on crystallographic data with most hydrogen atoms
2
2
−
(except one in (1OH + 1O )) and counter-anions removed for clarity. Thermal ellipsoids are drawn at the 40% probability level.
comparison, the same OH-bearing pincer ligand bound to
Ni(II) exhibits a pKa value of 5.4(4) in DMSO and was readily
isolated in the O− form.30 The O1···O2 distance of 2.540(8) Å
reflects a strong hydrogen-bonding interaction.32 The Cpy−O
distances reflect the charge on each pyridinol ring, with C24−
O2 = 1.35(1) Å for 1OH indicating some π donation into the
show some loss of aromaticity with long C−C distances of
∼1.42 Å and short C−C distances of ∼1.36 Å for 1NMe . This
2
effect is strongest with NMe2 as a strong π donor, and the
effect is less pronounced for 1OCH Ph. The other bond lengths
2
and angles for these compounds are tabulated in the
Supporting Information and are unremarkable.
Spectroscopy Showing Acid−Base Reactions on 1OH
.
ring (with a C−O bond order between single and double) and
−
C7−O1 = 1.29(1) Å for 1O indicating more CO character
Since the catalysis results described below will use acid or base
to control the protonation state of the pincer ligand, it is
important to first describe the fundamental acid−base
chemistry in the absence of substrate. Upon deprotonation
for the anionic pincer ring.
The structures of 1NMe and 1OCH Ph show the π donor effect
of the pyridine substituent on the bond lengths in the pincer.
The distances are C6−N8 = 1.358(5) Å in 1NMe2 and C6−O1
2
2
1
of 1OH with Na2CO3 in CD3OD, the H NMR spectrum
= 1.353(4) Å in 1OCH Ph which both reflect substantial CX
2
−
exhibits an upfield shift due to the formation of 1O (Scheme
double-bond character. Similarly, other methoxy-substituted
pincers show C−O bond distances (1.34−1.36 Å) between
single (∼1.43 Å) and double (∼1.23 Å) bond lengths.28,29
Furthermore, the C−C distances within the aromatic ring
4). The pyridine C−H protons exhibit the greatest shift from δ
−
6.97 ppm for 1OH to 6.58 ppm for 1O . Upfield shifts were also
seen for the imidazole-derived NHC ring protons. A similar
C
https://dx.doi.org/10.1021/acs.organomet.9b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
pubs.acs.org/Organometallics
Article
Scheme 4. Activation of Protic HDO Catalysts by Base
Table 1. Screening Ruthenium Pincer Catalysts for
Hydrodeoxygenation of Vanillyl Alcohol
a
b
c
c
entry
cat.
conversn (%)
yield of A (%)
yield of B (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
1OH
91(1)
39.1(5)
30.7(4)
24.2(4)
20.9(2)
15.0(4)
16.8(5)
16.3(4)
21.4(7)
2(2)
2.1(2)
95.8(7)
88(5)
89(4)
76(3)
52.2(8)
69.2(4)
69.4(3)
78.6(4)
1.9(1)
83.2(5)
83.7(4)
78.2(7)
11(2)
97.9(2)
0.2(1)
3(2)
1.6(3)
1.8(4)
1OMe
99.9(1)
93.6(2)
99.5(4)
18.5(4)
1NMe
1Me
1H
2OMe
2H
2
d
100.0
99.9(1)
99.6(1)
14(1)
3
4A
d
4B
100.0
pattern was seen in the 1H NMR for a Ni(II) complex bearing
the same OH-containing pincer, with an upfield shift from δ
6.96 to 6.21 ppm upon deprotonation in DMSO-d6.30 The
magnitude of the change in chemical shift is likely sensitive to
solvent effects. The IR data also support deprotonation of 1OH
with base, as there are several changes in the CN stretching
frequencies involving the pyridine ring. A peak at 1524 cm−1 is
e
1OH
98.0(4)
92(4)
91(4)
79(3)
e
1OMe
e
1NMe
1Me
2
e
f
d
1NMe
1Me
100.0
2(1)
98(1)
2
f
99.9(1)
19.0(8)
45.9(4)
14(1)
2.5(1)
1.8(4)
1.9(7)
3(1)
97.4(1)
15.5(7)
43.6(6)
84(1)
none
none
none
none
g
O−
18
19
20
tentatively ascribed to a CO stretch for 1 , which is similar
h
to the observed CO stretch at 1568 cm−1 for a Ni(II)
f
d
complex of the same pincer ligand.30
100.0
0.16(9)
99.95(9)
a
All experiments were done in triplicate and analyzed by GC. The
Hydrodeoxygenation of Substrates. Vanillyl alcohol
(VA) was used as a surrogate for lignin-derived monomers, and
catalytic conversion of VA was carried out as shown in Scheme
5. For all catalysts and conditions studied, the catalytic reaction
estimated standard deviation in the last digit is reported in
parentheses. Conditions: 0.0642 M vanillyl alcohol in methanol, 1
mol % of catalyst, 290 psi of H2, 100 °C for 1 h. See the Supporting
b
Information for further details. Conversion is calculated on the basis
c
of starting material consumption. Yield is calculated from the GC
d
Scheme 5. Catalytic Conversion of Vanillyl Alcohol
data. Quantitative conversion was observed in all three experiments.
a
e
f
g
Showing the Major Products Obtained
50 mol % of Na2CO3 was added. 1 mol % of HOTf was added. 10
h
mol % of Na2CO3 was added. 110 mol % of Na2CO3 was added.
accelerating the methylation of VA (B formation), it appears
that the presence of a strong π-donor group on the pyridine
ring is needed for the efficient hydrodeoxygenation and
formation of product A. When catalysts of types 2 and 3 are
considered next (entries 6−8), these catalysts are competent at
the methylation of VA (78−84% yield of product B), but they
fail to produce large yields of the HDO product A (16−21%
yield). Furthermore, by a comparison of 1OMe, 2OMe, and 3, all
of which bear methoxy groups on the pyridine rings, it is
observed that the substitution of the imidazole-based NHC
(1OMe) for the benzimidazole-based NHC (2OMe and 3) was
detrimental, perhaps due to the weaker donor properties for
the benzimidazole-derived NHC rings. Finally, catalysts 4A
and 4B were tested (entries 9 and 10). Catalyst 4A is not
competent at the methylation or the HDO reaction (cf. entry
17 in Table 1), and it seems reasonable to suggest that a lack of
labile ligands on the Ru center is detrimental to the catalytic
reaction. Catalyst 4B appears to serve as a Lewis acid for the
effective methylation of VA, but it fails at the HDO reaction.
Comparing 4B to 3 (2% vs 21% yield or A) would suggest that
the addition of a bipy ligand is detrimental. This result suggests
that multiple free sites are needed for the HDO reaction.
Overall, entries 1−10 in Table 1 illustrate that the presence of
two labile ligands combined with electronic factors of the
pincer ligand appear to enhance the HDO reaction with 1OH as
the best catalyst in Scheme 2 under neutral conditions.
a
“cat.” refers to the catalyst structures shown in Scheme 2.
yielded two major products labeled as A and B in Scheme 5
and ring hydrogenation products were not observed. Product A
(creosol) reduces the oxygen content of VA through the
desired hydrodeoxygenation (HDO) reaction. Further reduc-
tion in the oxygen content of A was not observed in this work;
specifically, we did not observe the HDO reaction on phenolic
oxygen atoms. Product B (methyl vanillyl ether) is the
alkylation product (Williamson ether synthesis) in methanol
solvent and is considered an undesired product, as it does not
reduce the oxygen content of VA. In fact, control experiments
in which the transition-metal catalyst was omitted (Table 1,
entries 18−20) illustrate that product B can be obtained
readily as the major product by treating VA with methanol
under hydrogen (290 psi) in the presence of base (44−84%
yield) or acid (>99% yield). However, no significant yield of
product A is obtained without a transition-metal catalyst
(Table 1, entry 17).
A survey of the catalytic ability of the catalysts in Scheme 2
was performed, and the results and reaction conditions are
shown in Table 1. In entries 1−10, we first investigated the
activity of all the catalysts in the absence of base. The catalysts
of type 1R are organized in order of decreasing yield of A
(entries 1−5) with R = OH > OMe > NMe2 > Me > H. While
most of these catalysts (except 1H) are competent at
One catalyst (1NMe ) contains a basic group that can
2
potentially be modified by the addition of external acids.
However, triflic acid is detrimental to the HDO reaction in
general, whether or not a basic group is present on the catalyst
D
https://dx.doi.org/10.1021/acs.organomet.9b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
pubs.acs.org/Organometallics
Article
structure (as in 1NMe and 1Me in entries 15 and 16, Table 1)
because it promotes the formation of product B. These results
are similar to entry 20 in Table 1, which shows the effect of
triflic acid alone.
Table 2. Hydrodeoxygenation of Vanillyl Alcohol with 1OH
Evaluating the Identity and Quantity of Base
:
2
a
conversn
b
c
c
entry
base (mol %)
(%)
yield of A (%) yield of B (%)
In relation to the above studies with HOTf, attempts to
1
2
3
4
5
6
7
8
9
NaOtBu (10)
NaOH (10)
NaHCO3 (10)
K2CO3 (10)
Na2CO3 (1.1)
Na2CO3 (10)
Na2CO3 (25)
Na2CO3 (50)
Na2CO3 (110)
41.6(4)
50.8(8)
29.2(8)
51.1(5)
20(2)
51(2)
73(1)
98.0(4)
99.73(6)
38.1(4)
47(1)
27(1)
46.8(6)
16(2)
48(2)
69.5(4)
95.8(7)
98.8(3)
1.65(6)
2.1(1)
0.6(2)
0.8(6)
0.5(2)
1.1(2)
2(1)
protonate 1NMe with HOTf did not lead to any substantial
2
1
changes in the H NMR spectrum in DMSO. Some slight
changes were observed in the IR spectrum upon adding HOTf,
but it is possible that these changes are due to hydrogen
bonding with HOTf or H3O+ formed from adventitious water
or incomplete protonation. While the pKa value of the
conjugate acid of 4-dimethylaminopyridine (DMAP) is 9.6 in
water33 and DMAP would be protonated readily by HOTf, it is
0.2(1)
0.4(3)
possible that 1NMe is less basic than DMAP due to
2
a
All experiments were done in triplicate and were analyzed by GC.
delocalization of the lone pair on NMe2 into the pyridine
Conditions: 0.0642 M vanillyl alcohol in methanol, 1 mol % of 1OH
,
ring. In fact, the crystal structure above for 1NMe suggests that
2
290 psi of H2, 100 °C for 1 h. See the Supporting Information for
b
there is substantial double-bond character for CNMe2.
External base serves to promote the HDO reaction with
several catalysts. Entry 11 (Table 1) shows that the presence of
base (50 mol % of Na2CO3) with 1OH facilitated the HDO
reaction and led to a 96% yield for product A in just 1 h. Thus,
nearly complete conversion to A is obtained by deprotonating
1OH, which enhances the π-donor properties of the pincer
further details. Conversion is calculated on the basis of starting
c
material consumption. Yield is calculated from the GC data.
conversion. Thus, the ideal conditions (50 or 110 mol % of
Na2CO3 and 1 mol % of 1OH) led to selective and nearly
complete formation of A. Using these optimal conditions (50
mol % of Na2CO3), we also explored the hydrodeoxygenation
of less activated substrates, but we see that 1OH (1 mol %) is
not effective at converting benzyl alcohol to toluene (see the
Supporting Information for details).
ligand (Scheme 4). The addition of base to 1OMe, 1NMe , and
2
1Me was also explored. The base should not affect the
Me
structures of 1OMe, 1NMe , or 1 ; thus, any changes in the
2
observed reactivity would not be attributed to changes in the
catalyst. As illustrated by entry 12 of Table 1, the use of base
with 1OMe generates an 88% yield of A, showing that base
accelerates the HDO reaction even in the absence of a protic
ligand (cf. entry 2 with a 31% yield of A). A base also enhances
Xu et al. have investigated the role of pH, promoters (e.g.,
formic acid), and remote directing groups in the HDO reaction
of a variety of alcohols and observed optimum yields at pH 1.6
in water. They propose an SN1 mechanism with OH2 loss from
substrate followed by hydride transfer from the catalyst to the
substrate.34 Formic acid served as a source of protons to
modulate pH, and the resulting formate was decarboxylated to
produce an iridium hydride catalyst. In our case, increased
HDO reactivity is observed with a weak base present, which
suggests that a different mechanism is operative in our study.
There appears to be an optimum balance for base strength in
our case. Sodium carbonate is strong enough to facilitate
catalyst deprotonation but does not generate any species
capable of direct binding to the active Ru centers of our
catalysts. Conversely, bases such as NaOtBu and NaOH both
generate species that may bind to the Ru centers and inhibit
catalysis; these species include tert-butoxide, hydroxide, and
methoxide from solvent deprotonation.
As described in the Supporting Information, exploring
product B as a substrate and under optimal catalytic conditions
(with 1OH or 1OMe as the catalyst) led to slower formation of A
in comparison to the conversion of VA directly to A. In view of
our data, we propose that product B formation does not
facilitate the formation of A. This suggests two possibilities. (1)
Perhaps B must be converted to VA by any adventitious water
present before the HDO reaction can occur. (2) Alternatively,
B goes directly to A, but by a mechanism that is different from
that employed when we start with VA and it must be
inherently slower. The methylated substrate B certainly cannot
bind to ruthenium as readily as VA.
catalysis with 1NMe , and an 89% yield (entry 13) of the HDO
2
product A is obtained (vs 24% without base, entry 3).
Similarly, adding base to 1Me leads to an increased yield of A
(76% in entry 14) but a somewhat decreased percent
conversion. Comparing these results shows that the selectivity
to the desired product is increased with base present for all
four catalysts: 1R where R = OH, OMe, NMe2, Me. However,
base alone does not lead to the desired product A (entries 18
and 19, Table 1). Hydrogen activation most likely occurs via
the well-established reaction Ru + H2 → Ru−H + H+. The
generation of H+ is detrimental to the reaction selectivity, likely
by promoting the formation of the undesired product B. Thus,
the base can play two roles: it can prevent acid buildup and
undesired pathways and, when the catalyst is designed
properly, the base can further activate the catalyst to favor
formation of A vs B. Accordingly, 1OH with base is our most
selective catalyst, with an A:B ratio of 479:1. The other
Me
catalysts (1OMe, 1NMe , 1 ) do not come close to this
2
selectivity, with at best 57:1.
With the knowledge in hand that base is advantageous to the
reaction and that catalyst 1OH can be further activated by base,
the experimental conditions were systematically varied to
further enhance catalytic activity. Increasing the reaction
temperature did not have a significant effect on reaction
selectivity or activity (see the Supporting Information for
details). In addition, catalyst decomposition was observed at
temperatures >150 °C. Next, the identity and loading of the
base was explored (Table 2). Strong bases such as NaOH and
NaOtBu were detrimental to the reaction (entries 1 and 2).
Weak bases such as Na2CO3 gave optimal conversion and
selectivity at high base loadings (entries 4−9). The use of a
very weak base (NaHCO3, entry 3) did not lead to good
Once the optimum base loadings were established, a lower
catalyst loading of 1OH was investigated to probe whether the
catalyst can operate efficiently under very dilute conditions.
Without an increase in the reaction time beyond 1 h, the
lowest catalyst loading that results in quantitative conversion to
product A is 0.05 mol % (Na2CO3 2.5 mol %, T = 150 °C,
TON = 2000). When the temperature is lowered to 100 °C,
E
https://dx.doi.org/10.1021/acs.organomet.9b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
pubs.acs.org/Organometallics
Article
quantitative conversion can be obtained with 0.01 mol %
catalyst loading in 3 days (Na2CO3 0.5 mol %, TON = 10000).
Delightfully, this increases the turnover number 5-fold, and it
may be further increased at either lower catalyst loadings or
longer reaction times. Catalyst 1OH performs better than
heterogeneous catalysts in the literature, which only achieve
90% yield of product A21 or which achieve similar results
(>99% yield of A and selectivity) but only at much higher (e.g.,
5 wt % for Zn/Pd/C) catalyst loadings.25,26
Scheme 6. Mechanistic Proposal for Hydrodeoxygenation of
Vanillyl Alcohol with Ruthenium Pincer Complexes
a
For every catalytic system, there is a need to interrogate
whether the active catalyst is homogeneous or heterogeneous
in nature. To probe this issue for 1OH, we performed the
mercury test, since mercury is known to coat the surface of
nanoparticles and typically results in lower activity for such
heterogeneous systems.35−37 When we repeated entry 8 of
Table 2 with a few drops of mercury added to the reaction
vessel, we obtained 95% yield of A by GC (done in triplicate).
Within experimental error, this result is the same as that for
entry 8, and thus mercury does not alter the catalytic activity of
1OH with 50 mol % of Na2CO3 at 100 °C. This suggests that
the catalyst is homogeneous and molecular under these
conditions, though we caution that the nature of the true
catalyst in solution is often sensitive to the specific conditions
employed.38,39 Similarly, since catalyst 1OMe with 50 mol % of
Na2CO3 present (entry 12, Table 1) showed a large run to run
variation in results, we performed the mercury test on this
system as well and obtained similar results (91(4)%
conversion, 88(4)% yield of A, 2(1)% yield of B). This
suggests that the variation seen is not due to nanoparticle
formation for 1OMe with base.
a
The CNC pincer is represented as [Ru] here.
Thus, with the data obtained herein we can propose a
−
mechanism for the HDO reaction as catalyzed by 1O (Scheme
1
6). H NMR data on 1OH mixed with Na2CO3 in CD3OD
−
acidic and coordination of the phenolic O− to the metal center
may occur, this does not lead to products due to the inherent
challenges in performing a substitution reaction on an sp2
carbon (no HDO reaction on phenolic OH groups was
observed herein). Furthermore, the presence of OH and OMe
groups on the aromatic ring in VA serve to activate the
benzylic position and may explain the reactivity at this site (vs
a lack of reactivity for benzyl alcohol).34
supports the formation of 1O , and slowly over 20 h at room
temperature the exchange of acetonitrile ligands for solvent
occurs (see the Supporting Information). This exchange
should be much faster at 100 °C under our typical HDO
conditions. This would be followed by the binding of H2 to a
free site by most likely displacing the solvent trans to the
pyridine nitrogen. Here a π-donor ligand should help to
increase the acidity of the transient dihydrogen complex via a
more electron rich metal which donates into the σ* orbital on
bound H2.40,41 This serves to weaken or break the H−H bond
(perhaps forming a dihydride intermediate), and then this
complex can be deprotonated by base to form a metal hydride.
In fact, both experimental and computational studies support
that π-donor groups para to nitrogen on a pyridine ring can
accelerate (de)hydrogenation reactions.42−47 (As an aside, the
rate of formation and the stability of this metal hydride should
be much less under acidic conditions, thus explaining why
triflic acid inhibits the formation of A in our studies.)
Next, we propose that VA displaces the solvent and binds to
the metal. The hydride can then attack the benzylic position to
produce product A and a metal-bound hydroxide ligand. We
cannot rule out an outer-sphere mechanism, in which hydride
attacks VA in solution, but the need for multiple free sites
suggests an inner-sphere mechanism. Release of water or
hydroxide can allow the catalyst to begin another cycle. These
steps may occur in a different order, or perhaps chloride loss
occurs at some point during catalysis (binding H2 is often
faster at cationic and electron-deficient metal centers).48
Additionally, while the phenolic protons of VA are more
CONCLUSIONS
■
In this study, the electronic properties of ruthenium pincer
complexes along with the ability to provide free sites for
substrate binding were related to the ability for these
complexes to function as HDO catalysts. At least one labile
site was necessary for any catalytic activity (e.g., for formation
of the methylation product, B) to be observed. Two to three
labile ligand sites, however, proved necessary but not sufficient
for good yields of the HDO product A. The best yields and
selectivity for A were achieved with the most electron rich
pincer ligand (1 rather than 2 or 3) with π-donor substituents
(1NMe , 1OMe, 1OH) in the presence of a weak base (Na2CO3).
2
At low catalyst loadings (0.01 mol %), 1OH in the presence of
base serves as a homogeneous catalyst that is able to achieve
quantitative and selective conversion of vanillyl alcohol to the
desired HDO product, A. With this knowledge in hand, we can
design the next generation of homogeneous catalysts with
increased reactivity toward all of the oxygenated sites on lignin-
derived monomers.
F
https://dx.doi.org/10.1021/acs.organomet.9b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
pubs.acs.org/Organometallics
Article
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.organomet.9b00816.
REFERENCES
■
■
sı
(1) US_Energy_Information_Administration International Energy
Outlook 2017. https://www.eia.gov/outlooks/ieo/pdf/0484(2017).
pdf (accessed 9/11/19).
(2) Renewable_Energy_Policy_Network_for_21st_Century Re-
newables 2014 Global Status Report. https://www.ren21.net/wp-
content/uploads/2019/05/GSR2014_Full-Report_English.pdf (ac-
cessed 9/11/2019).
(3) Bridgwater, A. V. Renewable fuels and chemicals by thermal
processing of biomass. Chem. Eng. J. 2003, 91 (2), 87−102.
(4) Raymundo, L. M.; Mullen, C. A.; Strahan, G. D.; Boateng, A. A.;
Trierweiler, J. O. Deoxygenation of Biomass Pyrolysis Vapors via in
Situ and ex Situ Thermal and Biochar Promoted Upgrading. Energy
Fuels 2019, 33 (3), 2197−2207.
(5) Rahimi, A.; Ulbrich, A.; Coon, J. J.; Stahl, S. S. Formic-acid-
induced depolymerization of oxidized lignin to aromatics. Nature
2014, 515, 249.
Experimental details on the synthesis and character-
ization of all compounds, optimization data, hydro-
deoxygenation of various substrates, and control experi-
ments (PDF)
Accession Codes
CCDC 1981237−1981240 contain the supplementary crys-
tallographic data for this paper. These data can be obtained
free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by
emailing data_request@ccdc.cam.ac.uk, or by contacting The
Cambridge Crystallographic Data Centre, 12 Union Road,
Cambridge CB2 1EZ, UK; fax: +44 1223 336033.
́
(6) Huang, X.; Atay, C.; Koranyi, T. I.; Boot, M. D.; Hensen, E. J. M.
Role of Cu−Mg−Al Mixed Oxide Catalysts in Lignin Depolymeriza-
tion in Supercritical Ethanol. ACS Catal. 2015, 5 (12), 7359−7370.
AUTHOR INFORMATION
Corresponding Authors
■
̌
̌
́
̈
(7) Gasser, C. A.; Cvancarova, M.; Ammann, E. M.; Schaffer, A.;
Shahgaldian, P.; Corvini, P. F.-X. Sequential lignin depolymerization
by combination of biocatalytic and formic acid/formate treatment
steps. Appl. Microbiol. Biotechnol. 2017, 101 (6), 2575−2588.
(8) Partenheimer, W. The Aerobic Oxidative Cleavage of Lignin to
Produce Hydroxyaromatic Benzaldehydes and Carboxylic Acids via
Metal/Bromide Catalysts in Acetic Acid/Water Mixtures. Adv. Synth.
Catal. 2009, 351 (3), 456−466.
Aaron K. Vannucci − Department of Chemistry and
Biochemistry, University of South Carolina, Columbia, South
Carolina 29208, United States; orcid.org/0000-0003-
0401-7208; Email: vannucci@mailbox.sc.edu
Elizabeth T. Papish − Department of Chemistry and
Biochemistry, University of Alabama, Tuscaloosa, Alabama
35487, United States; orcid.org/0000-0002-7937-8019;
Email: etpapish@ua.edu
(9) Renders, T.; Van den Bosch, S.; Koelewijn, S. F.; Schutyser, W.;
Sels, B. F. Lignin-first biomass fractionation: the advent of active
stabilisation strategies. Energy Environ. Sci. 2017, 10 (7), 1551−1557.
(10) Das, A.; Rahimi, A.; Ulbrich, A.; Alherech, M.; Motagamwala,
A. H.; Bhalla, A.; da Costa Sousa, L.; Balan, V.; Dumesic, J. A.; Hegg,
E. L.; Dale, B. E.; Ralph, J.; Coon, J. J.; Stahl, S. S. Lignin Conversion
to Low-Molecular-Weight Aromatics via an Aerobic Oxidation-
Hydrolysis Sequence: Comparison of Different Lignin Sources. ACS
Sustainable Chem. Eng. 2018, 6 (3), 3367−3374.
Authors
Wenzhi Yao − Department of Chemistry and Biochemistry,
University of Alabama, Tuscaloosa, Alabama 35487, United
States; orcid.org/0000-0002-4874-0169
Sanjit Das − Department of Chemistry and Biochemistry,
University of Alabama, Tuscaloosa, Alabama 35487, United
States
Nicholas A. DeLucia − Department of Chemistry and
Biochemistry, University of South Carolina, Columbia, South
Carolina 29208, United States
Fengrui Qu − Department of Chemistry and Biochemistry,
University of Alabama, Tuscaloosa, Alabama 35487, United
States; orcid.org/0000-0002-9975-2573
(11) Gollakota, A. R. K.; Reddy, M.; Subramanyam, M. D.; Kishore,
N. A review on the upgradation techniques of pyrolysis oil. Renewable
Sustainable Energy Rev. 2016, 58 (C), 1543−1568.
(12) Luo, H.; Abu-Omar, M. M. Lignin extraction and catalytic
upgrading from genetically modified poplar. Green Chem. 2018, 20
(3), 745−753.
(13) Nolte, M. W.; Shanks, B. H. A Perspective on Catalytic
Strategies for Deoxygenation in Biomass Pyrolysis. Energy Technology
2017, 5 (1), 7−18.
Chance M. Boudreaux − Department of Chemistry and
Biochemistry, University of Alabama, Tuscaloosa, Alabama
35487, United States; orcid.org/0000-0003-1322-9878
(14) Zakzeski, J.; Bruijnincx, P. C. A.; Jongerius, A. L.; Weckhuysen,
B. M. The Catalytic Valorization of Lignin for the Production of
Renewable Chemicals. Chem. Rev. 2010, 110 (6), 3552−3599.
(15) Shao, Y.; Xia, Q.; Dong, L.; Liu, X.; Han, X.; Parker, S. F.;
Cheng, Y.; Daemen, L. L.; Ramirez-Cuesta, A. J.; Yang, S.; Wang, Y.
Selective production of arenes via direct lignin upgrading over a
niobium-based catalyst. Nat. Commun. 2017, 8 (1), 16104.
(16) Guo, T.; Xia, Q.; Shao, Y.; Liu, X.; Wang, Y. Direct
deoxygenation of lignin model compounds into aromatic hydro-
carbons through hydrogen transfer reaction. Appl. Catal., A 2017, 547,
30−36.
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.organomet.9b00816
Notes
The authors declare the following competing financial
interest(s): E.T.P. and A.K.V. are filing a patent application
related to this work.
(17) Baddour, F. G.; Witte, V. A.; Nash, C. P.; Griffin, M. B.; Ruddy,
D. A.; Schaidle, J. A. Late-Transition-Metal-Modified β-Mo2C
Catalysts for Enhanced Hydrogenation during Guaiacol Deoxygena-
tion. ACS Sustainable Chem. Eng. 2017, 5 (12), 11433−11439.
(18) Hsu, P.-J.; Jiang, J.-W.; Lin, Y.-C. Does a Strong Oxophilic
Promoter Enhance Direct Deoxygenation? A Study of NiFe, NiMo,
and NiW Catalysts in p-Cresol Conversion. ACS Sustainable Chem.
Eng. 2018, 6 (1), 660−667.
(19) Ju, C.; Li, M.; Fang, Y.; Tan, T. Efficient hydro-deoxygenation
of lignin derived phenolic compounds over bifunctional catalysts with
optimized acid/metal interactions. Green Chem. 2018, 20 (19), 4492−
4499.
ACKNOWLEDGMENTS
■
We thank the NSF (CHE-1800214) for funding this research.
We thank NSF CHE MRI 1828078 and UA for purchase of the
SC XRD instrument. Preliminary data was also obtained with
NSF OIA-1539035 support. S.D. thanks the University of
Alabama’s Graduate Council Fellowship (GCF). C. M. B.
thanks AL EPSCoR for a graduate fellowship. The support of
the National Science Foundation through the EPSCoR R-II
Track-2 grant number OIA-1539105 is gratefully acknowl-
edged by A.K.V.
G
https://dx.doi.org/10.1021/acs.organomet.9b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
pubs.acs.org/Organometallics
Article
(20) Song, W.; Zhou, S.; Hu, S.; Lai, W.; Lian, Y.; Wang, J.; Yang,
W.; Wang, M.; Wang, P.; Jiang, X. Surface Engineering of CoMoS
Nanosulfide for Hydrodeoxygenation of Lignin-Derived Phenols to
Arenes. ACS Catal. 2019, 9 (1), 259−268.
(21) Hao, P.; Schwartz, D. K.; Medlin, J. W. Effect of Surface
Hydrophobicity of Pd/Al2O3 on Vanillin Hydrodeoxygenation in a
Water/Oil System. ACS Catal. 2018, 8 (12), 11165−11173.
(22) Lien, C.-H.; Medlin, J. W. Promotion of Activity and Selectivity
by Alkanethiol Monolayers for Pd-Catalyzed Benzyl Alcohol Hydro-
deoxygenation. J. Phys. Chem. C 2014, 118 (41), 23783−23789.
(23) Parsell, T. H.; Owen, B. C.; Klein, I.; Jarrell, T. M.; Marcum, C.
Autocatalytic Surface-Growth Mechanism of Metal Film Formation. J.
Am. Chem. Soc. 2003, 125 (34), 10301−10310.
(37) Eberhard, M. R. Insights into the Heck Reaction with PCP
Pincer Palladium(II) Complexes. Org. Lett. 2004, 6 (13), 2125−2128.
(38) Stracke, J. J.; Finke, R. G. Water Oxidation Catalysis Beginning
with 2.5 μM [Co 4(H 2O) 2(PW 9O 34) 2] 10−: Investigation of the
True Electrochemically Driven Catalyst at ≥ 600 mV Overpotential at
a Glassy Carbon Electrode. ACS Catal. 2013, 3, 1209−1219.
(39) Bayram, E.; Linehan, J. C.; Fulton, J. L.; Roberts, J. A. S.;
̈
Szymczak, N. K.; Smurthwaite, T. D.; Ozkar, S.; Balasubramanian, M.;
Finke, R. G. Is It Homogeneous or Heterogeneous Catalysis Derived
from [RhCp*Cl2]2? In Operando XAFS, Kinetic, and Crucial Kinetic
Poisoning Evidence for Subnanometer Rh4 Cluster-Based Benzene
Hydrogenation Catalysis. J. Am. Chem. Soc. 2011, 133 (46), 18889−
18902.
̈
L.; Haupert, L. J.; Amundson, L. M.; Kenttamaa, H. I.; Ribeiro, F.;
Miller, J. T.; Abu-Omar, M. M. Cleavage and hydrodeoxygenation
(HDO) of C−O bonds relevant to lignin conversion using Pd/Zn
synergistic catalysis. Chem. Sci. 2013, 4 (2), 806−813.
́
́
(40) Kristjansdottir, S. S.; Norton, J. R. Acidity of Hydrido Transition
Metal Complexes in Solution; VCH: 1992; Chapter 9.
(24) Liu, H.; Jiang, T.; Han, B.; Liang, S.; Zhou, Y. Selective Phenol
Hydrogenation to Cyclohexanone Over a Dual Supported Pd−Lewis
Acid Catalyst. Science 2009, 326, 1250−1252.
(41) Papish, E. T.; Magee, M. P.; Norton, J. R., Protonation of
transition metal hydrides to give dihydrogen complexes. Mechanistic
implications and catalytic applications. In Recent Advances in Hydride
Chemistry; Elsevier: 2001; pp 39−74.
(42) Cho, D.; Ko, K. C.; Lee, J. Y. Catalytic Mechanism for the
Ruthenium-Complex-Catalyzed Synthesis of Amides from Alcohols
and Amines: A DFT Study. Organometallics 2013, 32 (16), 4571−
4576.
(43) Hull, J. F.; Himeda, Y.; Wang, W.-H.; Hashiguchi, B.; Periana,
R.; Szalda, D. J.; Muckerman, J. T.; Fujita, E. Reversible hydrogen
storage using CO2 and a proton-switchable iridium catalyst in aqueous
media under mild temperatures and pressures. Nat. Chem. 2012, 4
(5), 383−388.
(44) Wang, W.-H.; Hull, J. F.; Muckerman, J. T.; Fujita, E.; Himeda,
Y. Second-Coordination-Sphere and Electronic Effects Enhance
Iridium(III)-Catalyzed Homogeneous Hydrogenation of Carbon
Dioxide in Water Near Ambient Temperature And Pressure. Energy
Environ. Sci. 2012, 5 (7), 7923−7926.
(25) DeLucia, N. A.; Das, N.; Overa, S.; Paul, A.; Vannucci, A. K.
Low temperature selective hydrodeoxygenation of model lignin
monomers from a homogeneous palladium catalyst. Catal. Today
2018, 302, 146−150.
(26) DeLucia, N. A.; Jystad, A.; Laan, K. V.; Tengco, J. M. M.;
Caricato, M.; Vannucci, A. K. Silica Supported Molecular Palladium
Catalyst for Selective Hydrodeoxygenation of Aromatic Compounds
under Mild Conditions. ACS Catal. 2019, 9, 9060−9071.
(27) Rodrigues, R. R.; Boudreaux, C. M.; Papish, E. T.; Delcamp, J.
H. Photocatalytic Reduction of CO2 to CO and Formate: Do
Reaction Conditions or Ruthenium Catalysts Control Product
Selectivity? ACS Appl. Energy Mater. 2019, 2, 37−46.
(28) Das, S.; Rodrigues, R. R.; Lamb, R. W.; Qu, F.; Reinheimer, E.;
Boudreaux, C. M.; Webster, C. E.; Delcamp, J. H.; Papish, E. T.
Highly Active Ruthenium CNC Pincer Photocatalysts for Visible-
Light-Driven Carbon Dioxide Reduction. Inorg. Chem. 2019, 58 (12),
8012−8020.
(45) Himeda, Y.; Onozawa-Komatsuzaki, N.; Sugihara, H.; Kasuga,
K. Simultaneous Tuning of Activity and Water Solubility of Complex
Catalysts by Acid−Base Equilibrium of Ligands for Conversion of
Carbon Dioxide. Organometallics 2007, 26 (3), 702−712.
(46) Himeda, Y. Conversion of CO2 into Formate by Homoge-
neously Catalyzed Hydrogenation in Water: Tuning Catalytic Activity
and Water Solubility through the Acid−Base Equilibrium of the
Ligand. Eur. J. Inorg. Chem. 2007, 2007 (25), 3927−3941.
(47) Sandhya, K. S.; Remya, G. S.; Suresh, C. H. Pincer Ligand
Modifications To Tune the Activation Barrier for H2 Elimination in
Water Splitting Milstein Catalyst. Inorg. Chem. 2015, 54 (23), 11150−
11156.
(29) Boudreaux, C. M.; Liyanage, N. P.; Shirley, H.; Siek, S.;
Gerlach, D. L.; Qu, F.; Delcamp, J. H.; Papish, E. T. Ruthenium(II)
complexes of pyridinol and N-heterocyclic carbene derived pincers as
robust catalysts for selective carbon dioxide reduction. Chem.
Commun. 2017, 53, 11217−11220.
(30) Burks, D. B.; Davis, S.; Lamb, R. W.; Liu, X.; Rodrigues, R. R.;
Liyanage, N. P.; Sun, Y.; Webster, C. E.; Delcamp, J. H.; Papish, E. T.
Nickel(II) pincer complexes demonstrate that the remote substituent
controls catalytic carbon dioxide reduction. Chem. Commun. 2018, 54,
3819−3822.
̈
̈
(31) Gradert, C.; Krahmer, J.; Sonnichsen, F. D.; Nather, C.;
Tuczek, F. Molybdenum(0)−carbonyl complexes supported by mixed
benzimidazol-2-ylidene/phosphine ligands: Influence of benzannula-
tion on the donor properties of the NHC groups. J. Organomet. Chem.
2014, 770, 61−68.
(48) Crabtree, R. Iridium compounds in catalysis. Acc. Chem. Res.
1979, 12 (9), 331−337.
(32) Jeffrey, G. A. Hydrogen-Bonding: An update. Crystallogr. Rev.
2003, 9 (2), 135−176.
̈
̈ ̈
(33) Kaljurand, I.; Kutt, A.; Soovali, L.; Rodima, T.; Maemets, V.;
Leito, I.; Koppel, I. A. Extension of the Self-Consistent Spectrophoto-
metric Basicity Scale in Acetonitrile to a Full Span of 28 p KaUnits:
Unification of Different Basicity Scales. J. Org. Chem. 2005, 70 (3),
1019−1028.
(34) Yang, S.; Tang, W.; Yang, Z.; Xu, J. Iridium-Catalyzed Highly
Efficient and Site-Selective Deoxygenation of Alcohols. ACS Catal.
2018, 8 (10), 9320−9326.
(35) Anton, D. R.; Crabtree, R. H. Dibenzo[a,e]cyclooctatetraene in
a proposed test for heterogeneity in catalysts formed from soluble
platinum-group metal complexes. Organometallics 1983, 2 (7), 855−
859.
(36) Widegren, J. A.; Bennett, M. A.; Finke, R. G. Is It
Homogeneous or Heterogeneous Catalysis? Identification of Bulk
Ruthenium Metal as the True Catalyst in Benzene Hydrogenations
Starting with the Monometallic Precursor, Ru(II)(η6-C6Me6)(OAc)2,
Plus Kinetic Characterization of the Heterogeneous Nucleation, Then
H
https://dx.doi.org/10.1021/acs.organomet.9b00816
Organometallics XXXX, XXX, XXX−XXX