(Hexamethylbenzene)Ru catalysts for the Aldehyde-Water Shift reaction

Article abstract of DOI:10.1039/d0gc03809a

The Aldehyde-Water Shift (AWS) reaction uses H₂O as a benign oxidant to convert aldehydes to carboxylic acids, producing H₂, a valuable reagent and fuel, as its sole byproduct. (Hexamethylbenzene)Ru^IIcomplexes are demonstrated to have higher activity and selectivity (up to 95%) for AWS over disproportionation than previously reported catalysts.

Full text of DOI:10.1039/d0gc03809a

Green Chemistry
View Article Online
View Journal | View Issue
COMMUNICATION
(Hexamethylbenzene)Ru catalysts for the
Aldehyde-Water Shift reaction†
Cite this: Green Chem., 2021, 23,
1609
Alexander S. Phearman, a,b Jewelianna M. Moore, a Dayanni D. Bhagwandin,b,c
Received 10th November 2020,
Accepted 14th January 2021
Jonathan M. Goldberg, b D. Michael Heinekey b and Karen I. Goldberg
*
a,b
DOI: 10.1039/d0gc03809a
rsc.li/greenchem
The Aldehyde-Water Shift (AWS) reaction uses H2O as a benign dehydrogenation of aldehydes and alcohols to generate esters
oxidant to convert aldehydes to carboxylic acids, producing and lactones also reported that carboxylic acids formed when
H2,
a valuable reagent and fuel, as its sole byproduct. water and a hydrogen acceptor were present with RuH2(PPh3)4
(Hexamethylbenzene)RuII complexes are demonstrated to have as a catalyst (up to 91% yield of butyric acid from butyralde-
higher activity and selectivity (up to 95%) for AWS over dispropor- hyde at 180 °C).9 In the absence of a hydrogen acceptor, esters
tionation than previously reported catalysts.
were formed preferentially and H2 was not detected. In 2004,
Stanley reported a dinuclear rhodium catalyst for tandem
hydroformylation and AWS.6 While activity was high (TOF ≈
1900 h−1) and heptanoic acid was the only observed product,
the reaction required a CO headspace purge to prevent catalyst
decomposition and remove H2.10
Introduction
Carboxylic acids are widely used in the production of consu-
mer products, including polymers, esters, and amides.
Carboxylic acids are commonly synthesized using stoichio-
metric harsh oxidants such as permanganate, chromate, or
chlorite, although sometimes milder oxidants such as H2O2 or
O2 are used.1–5 These latter two oxidants are also more atom
efficient, reducing waste production. Another attractive syn-
thetic route uses H2O as the oxidant, and has been termed the
“Aldehyde-Water Shift” (AWS, Scheme 1a) due to its similarity
to the water gas shift reaction.6 In addition to using a cheap,
readily available, and benign oxidant, formation of a carboxylic
acid by the AWS is accompanied by release of H2, a valuable
coproduct.
In the early 1980′s, while investigating aldehyde dispropor-
tionation into carboxylic acid and alcohol under neutral or
strongly basic conditions (Cannizzaro reaction), Maitlis noted
that a Ru-catalyzed reaction afforded significantly more carbox-
ylate than alcohol, while Ir- and Rh-complexes gave the
expected 1 : 1 ratio.7,8 A 1987 Murahashi study examining the
More recently, Boncella demonstrated the use of a first-row
transition metal complex, (iPrPNHP)Mn(CO)2(OH) (iPrPNHP =
(HN-{CH2CH2(PiPr2)}2), to promote the AWS reaction, however
strong product inhibition prevented catalytic turnover.11
A
variety of supported Cu, Pt, and Au catalysts have also been
investigated for heterogeneous AWS reactivity in flow reactors,
with up to 75% acid selectivity reported.12 The AWS reaction
has also been invoked as the second step of acceptorless
alcohol dehydrogenation of primary alcohols to carboxylic
acids or carboxylates.13–20
Our initial studies of catalysts for the AWS reaction focused
on a series of Ir, Rh, and Ru half-sandwich complexes.21 While
Ir and Rh complexes were significantly more active as catalysts
than the Ru compounds, their selectivity for carboxylic acid
product was low and the competing aldehyde disproportiona-
tion reaction (Scheme 1b) dominated. While aldehyde dispro-
aDepartment of Chemistry, University of Pennsylvania, 231 South 34th Street,
Philadelphia, PA 19104, USA. E-mail: kig@sas.upenn.edu
bDepartment of Chemistry, University of Washington, Box 351700, Seattle,
WA 98195-1700, USA
cDepartment of Chemistry, Hunter College CUNY, 695 Park Avenue, New York,
NY 10065, USA
†Electronic supplementary information (ESI) available. CCDC 2034882 (9). For
ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/
d0gc03809a
Scheme 1 (a) The Aldehyde-Water Shift reaction and (b) aldehyde dis-
proportionation under neutral conditions.
This journal is © The Royal Society of Chemistry 2021
Green Chem., 2021, 23, 1609–1615 | 1609

View Article Online
Communication
Green Chemistry
Chart 1 Best p-cymene supported ruthenium complexes previously
screened for the AWS.
portion does produce carboxylic acid, the maximum yield is
50%, as it also produces an equivalent of alcohol. This initial
study prompted further investigation into
a series of
(p-cymene)Ru complexes bearing different bidentate ligands
with the goal of increasing the activity of the more selective Ru
catalysts.22 It was determined that the use of the catalyst pre-
cursor [(η6-p-cymene)RuCl2]2 dimer (1) and complexes with
chelating bidentate amine ligands such as [(η6-p-cymene)Ru(o-
PDA)Cl]Cl (2) (o-PDA = ortho-phenylenediamine) (Chart 1)
resulted in improved activity while maintaining good selecti-
vity using acetaldehyde as a substrate in aqueous solution (e.g.
92% conversion and 85% acid selectivity after 20 h submerged
in a 105 °C oil bath). The only other product observed was
ethanol, formed from aldehyde disproportionation.
Chart 2 Hexamethylbenzene-supported
screened for AWS reactivity in this study.
ruthenium
complexes
Table 1 Comparison of precatalyst performance for acetaldehyde
oxidation
Time
(h)
Acid
yield (%)
Alcohol
yield (%)
Acid
selectivity (%)
Entry
Precatalyst
It was postulated that the selectivity-determining step in
the catalytic cycle was the reaction of a M–H intermediate with
either H+ or aldehyde. The subsequent development of an
aqueous hydricity scale by the Miller group23 revealed an inter-
esting correlation between M–H hydricity and the AWS selecti-
vity observed with Ir, Rh and Ru catalysts. Complexes that
exhibited low AWS selectivity (Cp*Ir and Cp*Rh, Cp* = penta-
methylcyclopentadienyl) were characterized as less hydridic
than the corresponding (p-cymene)Ru complexes, which
exhibited higher AWS selectivity. While it is unclear if the
thermodynamic hydricity is a crucial factor for selectivity, this
empirical observation motivated us to test more hydridic Ru
complexes for the AWS reaction. Specifically, we examined
more electron-donating (η6-C6Me6)Ru complexes, as [(η6-
C6Me6)Ru(2,2′-bipyridine)H]+ proved to be the most hydridic
complex of the complexes studied by Miller and coworkers.23
Herein, we report the use of (η6-C6Me6)Ru complexes to
1
2
3
4
5
6
7
8
9
3
3
4
4
5
5
6
7
8
9
3
5
3
5
5
50
5
5
5
5
68(7)
83(2)
55.1(6)
75(8)
0.5(1)
4.2
1.2(3)
0.8(2)
10.8
3.8(8)
4.8(7)
2.7(1)
6(1)
0
3.1
0
0
6.6
16.5(10)
94.3(2)
94.6(9)
95(1)
93(1)
100
57.3
100
100
61.9
76.4(15)
10
53.4(12)
Reaction conditions: 5 mL H2O, 2.5 mmol acetaldehyde, 0.2 mol%
dimeric precatalyst or 0.4 mol% monomeric precatalyst, 95 °C, N2
atmosphere. Quantification was by 1H NMR spectroscopy using phenol
as internal standard and data with standard deviations (contained in
parentheses) were repeated in at least triplicate.
aqueous solution was observed after 20 hours at 105 °C24 in
enable highly selective catalytic conversion of aldehydes to car- the presence of either 3 or 4, with >95% selectivity for acetic
boxylic acids by the Aldehyde-Water Shift reaction with low acid. Production of hydrogen was confirmed by GC-TCD
(Fig. S3†). Since full conversion was achieved under these con-
catalyst loadings of 0.4 mol%. The use of these complexes
resulted in improved activity along with the best AWS selecti- ditions, further experiments were carried out at lower tempera-
vity to date of any reported catalytic system. Results of catalysis ture and with shorter reaction times to assess differences in
activity between catalysts. When the temperature was lowered
dependence on H2 pressure and pH further contribute to the to 95 °C and the reaction time was limited to 5 h, 3 and 4
using a variety of bidentate ligands, as well as selectivity
understanding of this reaction pathway.
achieved 88 and 81% conversion of acetaldehyde to acetic acid
with selectivities of 95 and 93%, respectively (Table 1, entries
2 and 4). In contrast, the p-cymene complex 2 required 20 h at
105 °C to reach similar conversions with only 85% selectivity
for the acid product.22
Results and discussion
Informed by our previous work, initial studies focused on [(η6-
Despite Miller’s report of [(η6-C6Me6)Ru(2,2′-bipyridine)H]+
C6Me6)RuCl2]2 (3) and [(η6-C6Me6)Ru(o-PDA)Cl]Cl (4) (Chart 2; being the most hydridic of the complexes they analyzed, hydri-
Table 1, entries 1–4). Complete conversion of acetaldehyde in city doesn’t appear to be the only important parameter for the
1610 | Green Chem., 2021, 23, 1609–1615
This journal is © The Royal Society of Chemistry 2021

View Article Online
Green Chemistry
Communication
AWS reaction, as various monomeric (η6-C6Me6)Ru complexes
bearing substituted bipyridine ligands (5–8) showed very little
activity (Table 1, entries 5–9). For example, even after 50 h reac-
tion times, <8% conversion was observed with precatalyst 5
(Table 1, entry 6). Use of the more electron donating 4,4′-
dimethoxy-2,2′-bipyridine ligand (precatalyst 6) only resulted
in 1.2% acid yield after 5 hours at 95 °C (Table 1, entry 7).
Exchanging the inner- and outer-sphere chlorides of 6 for
weakly coordinating triflate anions (7) had no effect (Table 1,
entry 8). The hydroxyl groups of the 6,6′-dihydroxy-2,2′-bipyri-
dine have previously been suggested to participate in bifunc-
tional catalysis, and Ru and Ir half sandwich catalysts bearing
this ligand have been shown to catalyze transfer hydrogen-
ation,25 CO2 reduction,26 alcohol dehydrogenation,27–30 and
water splitting.31 In the AWS reaction of acetaldehyde, complex
8, which bears this hydroxy substituted bipyridine, mainly
facilitated disproportionation, resulting in only 62% selectivity
for acetic acid with 17% conversion (Table 1, entry 9). A new
Table 2 Substrate scope studied for aldehyde oxidation
Acid yield
(%)
Alcohol yield
(%)
Acid selectivity
(%)
Entry Substrate
1a
83(2)
69(2)
4.8(7)
10(1)
94.6(9)
87(1)
2
3
4
27.8(13)
26.3(5)
8.7(5)
6.5(2)
76.0(4)
80.3(3)
5b
13.3(2)
7
1.9(1)
87.6(4)
100
6b,c
—
Reaction conditions: 3.5 mL H2O, 1.5 mL 1,4-dioxane, 2.5 mmol sub-
strate, 0.005 mmol 3, 95 °C, 20 h, N2 atmosphere. Reaction products
were quantified by GC-FID. All experiments were repeated in at least
triplicate with standard deviation provided in parentheses. a Reaction
run in 100% H2O for 5 h and products were quantified by 1H NMR
spectroscopy. b Reaction run in 50 : 50 H2O : 1,4-dioxane. c Quantitative
13C{1H} NMR analysis using inverse-gated pulse sequence to determine
yield and selectivity.
(η6-C6Me6)Ru complex supported by
a pyridine alkoxide
ligand, (η6-C6Me6)Ru(2-(2′-pyridyl)-2-propanoate)Cl (9), was
synthesized to test the effect of a monoanionic bidentate
ligand on the AWS reaction (Scheme 2). Characterization of 9
by X-ray diffraction (Fig. S8†) shows that the complex has the
expected three-legged piano stool structure with the pyridine
alkoxide ligand coordinated in a bidentate fashion. As a cata-
lyst for AWS with acetaldehyde, 9 afforded acetic acid in 53.4%
yield (5 h at 95 °C), but with an acid selectivity of only 76% benzaldehyde was very similar to that of pivaldehyde at 80%
(Table 1, entry 10).
and 33%, respectively.33 While there is less steric bulk close to
the carbonyl group for benzaldehyde, the phenyl ring conju-
gation renders the α-carbonyl carbon less electrophilic. While
less impressive than the results for the AWS with acetaldehyde,
these results are marked improvements over the reactions
carried out with the p-cymene-supported 2 as the precatalyst
for these substrates. Even under more forcing conditions
(105 °C), the p-cymene-supported 2 resulted in only 5% conver-
sion of pivaldehyde and <1% conversion of benzaldehyde in
the same time.22
Renewable substrates furfural and hydroxymethylfurfural
(HMF), which are derived from the dehydration of sugars, have
been extensively studied as bio-based platform chemicals.34,35
When furfural was used as the substrate with catalyst 3 (50 : 50
H2O : 1,4-dioxane, 20 h, 95 °C), high selectivity of 88% for oxi-
dation was observed, albeit in low yield (13% furanoic acid)
(Table 2, entry 5). Using the same conditions with HMF,36
clean conversion to hydroxymethylfuroic acid (7% yield) was
observed (Table 2, entry 6). No competing oxidation of the
alcohol moiety was observed under these reaction conditions,
demonstrating selectivity for aldehyde groups over alcohols
under mild conditions.
Substrate scope
Previously-studied half sandwich complexes for the AWS reac-
tion exhibited high acid selectivity for only linear aldehydes.
Using benzyl and branched alkyl aldehydes as substrates
resulted in greatly reduced conversion and acid selectivity. A
brief substrate scope using complex 3 was investigated. A less
polar solvent system, 70 : 30 H2O : 1,4-dioxane, was used to
improve substrate solubility and eliminate potential biphasic
reaction mixtures. A longer reaction time of 20 h at 95 °C was
also employed to achieve appreciable conversions.32
Overall, the increased steric bulk of isobutyraldehyde and
pivaldehyde resulted in decreased catalyst activity and selecti-
vity relative to acetaldehyde. Isobutyraldehyde was converted
with 87% selectivity to isobutyric acid (79% conversion), while
pivaldehyde was converted with 76% selectivity (37% conver-
sion, Table 2, entries 2–3). The selectivity and conversion for
trans-Cinnamaldehyde was also used as a substrate to test
the tolerance of internal alkene functionalities to AWS con-
ditions, including H2 formation (Scheme 3). Although conver-
sion was low after 20 hours, trans-cinnamic acid was the major
product (5.9% yield) when 3 was used as the catalyst at 95 °C
in 70 : 30 H2O : 1,4-dioxane solvent. The saturated carboxylic
acid, 3-phenylpropanoic acid, was also observed in 3.6% yield,
Scheme 2 Synthesis of (η6-C6Me6)Ru(2-(2’-pyridyl)-2-propanoate)Cl
(9).
This journal is © The Royal Society of Chemistry 2021
Green Chem., 2021, 23, 1609–1615 | 1611

View Article Online
Communication
Green Chemistry
Scheme 3 Oxidation of cinnamaldehyde using 3. Standard deviation in parentheses.
indicating that the internal alkene can act as an intra- yield initially rises rapidly before the slope (Δbenzoic acid
molecular hydrogen acceptor. Only trace 3-phenylpropanol was yield/Δtime) decreases (red triangles). As benzoic acid is pro-
observed and the direct product of aldehyde disproportiona- duced through both the AWS reaction and aldehyde dispropor-
tion, cinnamyl alcohol, was not detected by GC. Formation of tionation, we subtracted the amount of benzoic acid produced
the saturated carboxylic acid is perhaps not surprising as through disproportionation (i.e. subtracted the amount of
transfer hydrogenations using (η6-arene)Ru complexes have benzyl alcohol produced as each are generated in a 1 : 1 ratio
been well-studied.37,38 In addition, when Murahashi studied during disproportionation) to more clearly visualize the
the reaction of water with crotonaldehyde, another change in AWS reactivity over time. As demonstrated through
α,β-unsaturated substrate, the saturated carboxylic acid was the subtraction of benzoic acid produced via disproportiona-
produced in 75% selectivity after 24 hours at 180 °C with tion (green open triangles, Fig. 1), AWS has essentially stopped
RuH2(PPh3)4.9 The results with catalyst 3 can be compared to after ∼20 h. Furthermore, the linear increase in the alcohol
those obtained with the p-cymene catalyst, [(η6-p-cymene)Ru(o- yield also supports that alcohol production is solely a product
PDA)Cl][Cl] (2) where no cinnamaldehyde conversion was of transfer hydrogenation (disproportionation) and not hydro-
observed after 20 h at 105 °C.22
genation with H2 produced from the AWS reaction under these
conditions. If hydrogenation played a significant role, the
slope of the alcohol yield would be expected to change over
time as H2 production from the AWS reactions slows to near
zero at 20 h.
We considered two possibilities for the decrease in AWS
reaction over time. First, because the AWS reaction produces
carboxylic acid product, the pH of the reaction solution
decreases with higher conversion. Several protonation and
deprotonation events occur throughout the proposed catalytic
cycle, including the proposed selectivity-determining step (see
later discussion). Second, the released hydrogen builds up in
the reaction vessels over time. Thus, the increased H2 pressure
could inhibit the AWS pathway.
Time-dependent selectivity
Monitoring the conversion of benzaldehyde to benzoic acid
and benzyl alcohol by 3 at 95 °C in 70 : 30 H2O : 1,4-dioxane
over time revealed a clear time-dependent selectivity profile
(Fig. 1, Table S2†). The background disproportionation reac-
tion remains constant over 40 h as evidenced by the steadily
increasing and linear benzyl alcohol yield with respect to time
(blue circles, linear fit R2 = 0.998). In contrast, the benzoic acid
To examine the effect of pH on the reaction, we attempted
to study the catalytic reactions at constant pH by using sodium
phosphate buffered solutions (0.25 M, pH 6.0 or 7.0).
Unfortunately, the mass balance of products and starting
material was consistently low (70–90%) for these reactions
after workup, making any conclusions based on these experi-
ments suspect.39 A different strategy to investigate the role that
the increased concentration of benzoic acid and lower pH
might have on the AWS reaction was then employed.
Experiments were carried out with benzoic acid added to the
starting solution. Thus, 5 or 25 mol% benzoic acid (0.125 or
0.625 mmol) relative to substrate was added to the standard
2.5 mmol benzaldehyde loading and the reaction progress was
measured after 20 h at 95 °C in 70 : 30 H2O : 1,4-dioxane. The
amount of additional acid (greater than the original 0.125 or
0.625 mmol) was recorded (Table 3). In general, increasing
Fig. 1 Time-dependent selectivity profile for benzaldehyde oxidation.
Reaction conditions: 3.5 mL H2O, 1.5 mL 1,4-dioxane, 2.5 mmol benz-
aldehyde, 0.005 mmol 3 (0.2 mol%), 95 °C, N2 atmosphere. Quantified
by GC-FID. All entries repeated in at least triplicate. Yields in tabular
format can be found in Table S2.†
1612 | Green Chem., 2021, 23, 1609–1615
This journal is © The Royal Society of Chemistry 2021

View Article Online
Green Chemistry
Communication
Table 3 Effect of added benzoic acid on oxidation of benzaldehyde
that purging the reaction headspace of their system with CO
increased AWS selectivity.6 In addition, within error, acid
selectivity in the vented experiments does not change at the 20
and 70 h timepoints, while it decreases from 84.0(3) to 77.1(6)
in a closed system, further demonstrating the benefit of releas-
ing H2 pressure on acid selectivity.
Benzyl
alcohol yield
(%)
Acid
selectivity
(%)
Added benzoic
acid (mol%)
Benzoic acid yield
(additional, %)
Proposed catalytic cycle
The proposed catalytic cycle in Scheme 4 is based on mechan-
istic proposals for similar acceptorless alcohol dehydrogena-
tion reactions,14,18,29,41 as well as computational studies
carried out by the Cundari group.21,42 The cycle is depicted
with a generic monodentate half-sandwich complex, as the
speciation of 3 under catalytic conditions is unknown and
different species bearing the (η6-C6Me6)Ru manifold may share
a common reactive intermediate.
0
5
25
26.3(5)
17.2(14)
14.9(4)
6.5(2)
4.3(1)
4.4(3)
80.3(3)
80.0(9)
77.2(14)
Reaction conditions: 3.5 mL H2O, 1.5 mL 1,4-dioxane, 2.5 mmol benz-
aldehyde, 0.005 mmol 3 (0.2 mol%), added benzoic acid as shown in
table, 95 °C, 20 h, N2 atmosphere. Reaction products were quantified
by GC-FID. Experiments were repeated in triplicate with standard devi-
ation in parentheses.
Dissociation of a ligand from the precatalyst (A) generates
an unsaturated Ru complex (B) (Scheme 4). Activation of an
aldehyde by coordination to Ru (C) facilitates nucleophilic
attack by H2O and tautomerization to give a gem-diol (D). With
some substrates, significant amounts of gem-diol are present
in aqueous solutions, which could associate directly with B to
form D. We note that substrates giving high conversion in this
study form significant amounts of gem-diol in H2O at neutral
amounts of added benzoic acid led to lower consumption of
benzaldehyde after 20 h. Although both AWS and disproportio-
nation are inhibited by the addition of benzoic acid, as can
been seen in the experiments with 5 mol% and 25 mol%
added benzoic acid, the AWS experiences a slightly greater
inhibition than disproportionation.
Likewise, the effect of H2 pressure on the Aldehyde-Water
Shift was investigated. To determine the effect of H2 pressure,
reactions were performed in a side-by side fashion, one in the
normal closed system and the other with a vent needle to an
outlet bubbler. The outlet tubing to the bubbler was purged
thoroughly with N2 immediately before use to prevent intro-
duction of O2 into the reaction.40 To study H2 dependence, a
moderately lower reaction temperature (85 °C) was used to
minimize solvent loss. This experimental design was carried
out in triplicate at both 20 and 70 h reaction times. The selecti-
vity for benzoic acid in the open system was significantly
higher after both 20 and 70 h, confirming that H2 pressure
inhibits the AWS reaction (Table 4). This result is consistent
with the work of Stanley and co-workers, who demonstrated
Table 4 Effect of hydrogen pressure release
Time
(h)
Acid
yield (%)
Alcohol
yield (%)
Total
yield (%)
Acid
selectivity (%)
Vented
Vented
Closed
Closed
20
70
20
70
15.2(12)
27.9(7)
18.1(8)
25.6(8)
1.4(1)
3.4(4)
3.5(2)
7.6(5)
16.6(12)
31.3(8)
21.6(8)
33.2(9)
91.3(8)
89.3(12)
84.0(3)
77.1(6)
Reaction conditions: 3.5 mL H2O, 1.5 mL 1,4-dioxane, 2.5 mmol benz-
aldehyde, 0.005 mmol 3 (0.2 mol%), 85 °C. Reaction products were
quantified by GC-FID. Experiments were repeated in triplicate with
standard deviation in parentheses.
Scheme 4 Proposed catalytic cycle. Selectivity determining step high-
lighted in blue (AWS) and red (aldehyde disproportionation).
This journal is © The Royal Society of Chemistry 2021
Green Chem., 2021, 23, 1609–1615 | 1613

View Article Online
Communication
Green Chemistry
pH (Khydration = 1.25 for acetaldehyde),43 while substrates so informs future precatalyst design and optimization.
giving low conversion exist primarily in the aldehyde form Furthermore, selectivity was demonstrated to be dependent on
(Khydration = 0.01 for benzaldehyde and 0.25 for pivaldehyde).44 headspace H2 pressure and catalyst activity is reduced with
Deprotonation of the gem-diol (D) then forms a Ru-alkoxide increasing acid concentration. This behavior is thus consistent
(E). This step would be inhibited by increasing acid concen- with the “shift” nature of the AWS reaction. Overall our results
tration, consistent with the observed decrease in conversion demonstrate how an improved understanding of the AWS reac-
with added acid (Table 3). Subsequent β-hydride elimination tion can facilitate improvement in the reactivity and selectivity
or dissociative β-hydride abstraction45–48 results in a Ru–H (G) of catalysts for this relatively underdeveloped environmentally-
and the carboxylic acid product. The low activity of complexes friendly aldehyde oxidation method.
6–8, which have strongly-bound bidentate ligands, suggests
that an open site (F) is required, consistent with a β-hydride
elimination pathway. Ligands such as (o-PDA, 4) or Cl− are Conflicts of interest
likely labile enough to allow formation of F. Although not all
of the complexes studied here have been crystallographically
There are no conflicts to declare.
characterized, the p-cymene derivatives of 4–7 have been
(Fig. S13†).25,49–51 It is evident that the Ru–N bonds are signifi-
cantly elongated with primary amine donors compared to
Acknowledgements
bipyridine nitrogen donors, indicating that o-PDA is not as
The authors thank Dr. Louise M. Guard for helpful discussions
tightly bound to Ru and would be more likely than bipyridine
and the group of Professor Robert Crabtree (Yale University)
to act as a hemilabile ligand.
for the gift of 2-(2′-pyridyl)-2-propanol. Accurate mass
measurement analysis was carried out by Dr. Charles W. Ross
III (University of Pennsylvania) and X-ray crystallography was
carried out by Dr. Patrick J. Carroll (University of
Selectivity for the AWS vs. aldehyde disproportionation
would then arise from the reactivity of the Ru–H (G)
(Scheme 4, top left). Protonation of G would release H2 and
regenerate a 16e− unsaturated Ru complex, B, to complete the
Pennsylvania). This work was supported by the National
cycle. Alternatively, hydride transfer to an aldehyde and sub-
Science Foundation under the CCI Center for Enabling New
sequent protonation in acidic aqueous conditions would yield
Technologies through Catalysis (CENTC) (CHE-1205189), the
alcohol product, thereby formally hydrogenating an equivalent
Department of Energy (DE-SC0018057), and the Penn
of aldehyde and completing the disproportionation pathway.
Undergraduate Research Mentoring Program (J.M.M.). The
The demonstrated increase in acid selectivity when H2
NSF Major Research Instrumentation Program (NSF
pressure is allowed to escape is consistent with this proposal,
CHE-1827457, 400 MHz NMR spectrometer), the NIH sup-
as only the AWS pathway releases H2 as a product. By Le
plemental awards 3R01GM118510-03S1 and 3R01GM087605-
Chatelier’s principle, release of H2 would be expected to
06S1 (600 MHz NMR spectrometer), and Vagelos Institute for
Energy Science and Technology are recognized for their
change the selectivity in favor of the AWS, especially at later
reaction times. Further improvements in selectivity could
support of NMR instrumentation used in this study.
likely be achieved by purging the reaction headspace to remove
H2 rather than simply releasing H2 pressure. The slight
decrease in acid selectivity when additional benzoic acid is
References
added (Table 3) may be due to differences in [H+] dependency
in the selectivity-determining step or change in speciation
of 3.
1 M. B. Smith and J. March, March’s Advanced Organic
Chemistry: Reactions, Mechanisms and Structure, John Wiley
and Sons, Inc., Hoboken, NJ, 6th edn, 2007.
2 H. Yu, S. Ru, G. Dai, Y. Zhai, H. Lin, S. Han and Y. Wei,
Angew. Chem., Int. Ed., 2017, 56, 3867–3871.
3 M. Liu and C.-J. Li, Angew. Chem., Int. Ed., 2016, 55, 10806–
10810.
4 Y. Zhang, Y. Cheng, H. Cai, S. He, Q. Shan, H. Zhao,
Y. Chen and B. Wang, Green Chem., 2017, 19, 5708–5713.
5 J. Kubitschke, H. Lange and H. Strutz, Carboxylic Acids,
Aliphatic. Ullmann’s Encyclopedia of Industrial Chemistry,
2014, pp. 1–18.
6 G. G. Stanley, D. A. Aubry, N. Bridges, B. Barker and
B. Courtney, Prepr. Pap. – Am. Chem. Soc., Div. Fuel Chem.,
2004, 49, 712–713.
Conclusions
The use of Ru half-sandwich complexes with more electron
donating C6Me6 arene rings resulted in significant improve-
ments in selectivity and activity compared to p-cymene ana-
logues for the Aldehyde-Water Shift reaction. The use of H2O
as both the oxygen atom source and solvent is unique, and
concomitant H2 release results in an atom efficient process
under mild reaction conditions (95 °C) with low catalyst load-
ings (0.4 mol% Ru). While the use of electron rich arene
ligands boosts selectivity for AWS, strongly-bound bidentate
ancillary ligands drastically decrease reactivity. This obser-
vation supports that ligand dissociation (allowing for
β-hydride elimination) is a key step in the catalytic cycle, and
7 J. Cook, J. E. Hamlin, A. Nutton and P. M. Maitlis, J. Chem.
Soc., Chem. Commun., 1980, 144–145, DOI: 10.1039/
C39800000144.
1614 | Green Chem., 2021, 23, 1609–1615
This journal is © The Royal Society of Chemistry 2021

View Article Online
Green Chemistry
Communication
8 J. Cook, J. E. Hamlin, A. Nutton and P. M. Maitlis, J. Chem. 31 J. DePasquale, I. Nieto, L. E. Reuther, C. J. Herbst-
Soc., Dalton Trans., 1981, 2342–2352, DOI: 10.1039/
dt9810002342.
9 S. Murahashi, T. Naota, K. Ito, Y. Maeda and H. Taki,
J. Org. Chem., 1987, 52, 4319–4327.
10 A. R. Barnum, Ph. D. Dissertation, Louisiana State
University, 2012.
11 A. M. Tondreau, R. Michalczyk and J. M. Boncella,
Organometallics, 2017, 36, 4179–4183.
12 W.-C. Wen, S. C. Eady and L. T. Thompson, Catal. Today,
2020, 355, 199–204.
13 N. Xiang, P. Xu, N. Ran and T. Ye, RSC Adv., 2017, 7, 38586–
38593.
14 E. Balaraman, E. Khaskin, G. Leitus and D. Milstein, Nat.
Chem., 2013, 5, 122–125.
15 R. E. Rodriguez-Lugo, M. Trincado, M. Vogt, F. Tewes,
G. Santiso-Quinones and H. Grutzmacher, Nat. Chem.,
2013, 5, 342–347.
Gervasoni, J. J. Paul, V. Mochalin, M. Zeller, C. M. Thomas,
A. W. Addison and E. T. Papish, Inorg. Chem., 2013, 52,
9175–9183.
32 The reported boiling point of a 70 : 30 water : 1,4-dioxane
azeotrope is 88.2 °C at atmospheric pressure (E. R. Smith
and M. Wojciechowski, Journal of Research of the National
Bureau of Standards, 1937, 18, 461–465). In a small volume
(20 mL) sealed system the headspace pressure exceeds
atmospheric pressure due to vapor pressure and the dihy-
drogen produced by the AWS, allowing the temperature of
the solution to exceed 88.2 °C.
33 Papish and Brewster recently reported that a [(Cp*)Ir(6,6′-
dihydroxy-2,2′-bipyridine)]2+ complex is capable of decarbo-
nylating benzoic acid (ref. 20). No evidence of decarbonyla-
tion is observed when complex 3 is used with benzaldehyde
as substrate for the AWS reaction.
34 B. Saha and M. M. Abu-Omar, Green Chem., 2014, 16, 24–
38.
16 J.-H. Choi, L. E. Heim, M. Ahrens and M. H. G. Prechtl,
Dalton Trans., 2014, 43, 17248–17254.
35 A. A. Rosatella, S. P. Simeonov, R. F. M. Frade and
C. A. M. Afonso, Green Chem., 2011, 13, 754–793.
36 Control reactions without Ru, showed that an aqueous
solution of HMF, when heated to 95 °C for 20 h, resulted in
a dark brown solution, while a 50 : 50 water : 1,4-dioxane
mixture resulted in no color change.
17 K.-i. Fujita, R. Tamura, Y. Tanaka, M. Yoshida, M. Onoda
and R. Yamaguchi, ACS Catal., 2017, 7, 7226–7230.
18 A. Sarbajna, I. Dutta, P. Daw, S. Dinda, S. M. W. Rahaman,
A. Sarkar and J. K. Bera, ACS Catal., 2017, 7, 2786–2790.
19 R. Yamaguchi, D. Kobayashi, M. Shimizu and K.-i. Fujita,
J. Organomet. Chem., 2017, 843, 14–19.
37 T. Ikariya and A. J. Blacker, Acc. Chem. Res., 2007, 40, 1300–
1308.
20 W. Yao, A. R. DeRegnaucourt, E. D. Shrewsbury,
K. H. Loadholt, W. Silprakob, F. Qu, T. P. Brewster and 38 R. Noyori, Angew. Chem., Int. Ed., 2002, 41, 2008–2022.
E. T. Papish, Organometallics, 2020, 39, 3656–3662.
21 T. P. Brewster, W. C. Ou, J. C. Tran, K. I. Goldberg,
S. K. Hanson, T. R. Cundari and D. M. Heinekey, ACS
Catal., 2014, 4, 3034–3038.
22 T. P. Brewster, J. M. Goldberg, J. C. Tran, D. M. Heinekey
and K. I. Goldberg, ACS Catal., 2016, 6, 6302–6305.
39 In a mass balance control experiment with just benz-
aldehyde, buffer, and solvent, only 90% of the starting
material was detected after workup.
40 In preliminary studies with acetaldehyde, it was noted that
acid selectivity did not change greatly with reaction temp-
erature (Fig. S11 and 12†).
23 C. L. Pitman, K. R. Brereton and A. J. Miller, J. Am. Chem. 41 T. Zweifel, J. V. Naubron and H. Grutzmacher, Angew.
Soc., 2016, 138, 2252–2260. Chem., Int. Ed., 2009, 48, 559–563.
24 Temperatures listed in this communication are that of a 42 W. C. Ou and T. R. Cundari, ACS Catal., 2014, 5, 225–232.
thermocouple-controlled hotplate and aluminum heating 43 J. L. Kurz, J. Am. Chem. Soc., 1967, 89, 3524–3528.
block. The heating block is half the height of the reaction 44 S. H. Hilal, L. L. Bornander and L. A. Carreira, QSAR Comb.
vials.
Sci., 2005, 24, 631–638.
25 I. Nieto, M. S. Livings, J. B. Sacci, L. E. Reuther, 45 O. Blum and D. Milstein, J. Organomet. Chem., 2000,
M. Zeller and E. T. Papish, Organometallics, 2011, 30, 6339–
6342.
26 S. Siek, D. B. Burks, D. L. Gerlach, G. Liang, J. M. Tesh,
593–594, 479–484.
46 N. A. Smythe, K. A. Grice, B. S. Williams and K. I. Goldberg,
Organometallics, 2009, 28, 277–288.
C. R. Thompson, F. Qu, J. E. Shankwitz, R. M. Vasquez, 47 O. Blum and D. Milstein, J. Am. Chem. Soc., 1995, 117,
N. Chambers, G. J. Szulczewski, D. B. Grotjahn, 4582–4594.
C. E. Webster and E. T. Papish, Organometallics, 2017, 36, 48 C. M. Fafard and O. V. Ozerov, Inorg. Chim. Acta, 2007, 360,
1091–1106. 286–292.
27 G. Zeng, S. Sakaki, K.-i. Fujita, H. Sano and R. Yamaguchi, 49 T. Bugarcic, A. Habtemariam, R. J. Deeth, F. P. Fabbiani,
ACS Catal., 2014, 4, 1010–1020.
28 R. Zhu, B. Wang, M. Cui, J. Deng, X. Li, Y. Ma and Y. Fu,
Green Chem., 2016, 18, 2029–2036.
29 R. Kawahara, K. Fujita and R. Yamaguchi, Angew. Chem.,
Int. Ed., 2012, 51, 12790–12794.
30 R. Kawahara, K. Fujita and R. Yamaguchi, J. Am. Chem.
Soc., 2012, 134, 3643–3646.
S. Parsons and P. J. Sadler, Inorg. Chem., 2009, 48, 9444–
9453.
50 X. Wu, J. Liu, D. Di Tommaso, J. A. Iggo, C. R. Catlow,
J. Bacsa and J. Xiao, Chem. – Eur. J., 2008, 14, 7699–7715.
51 C. Aliende, M. Pérez-Manrique, F. A. Jalón, B. R. Manzano,
A. M. Rodríguez and G. Espino, Organometallics, 2012, 31,
6106–6123.
This journal is © The Royal Society of Chemistry 2021
Green Chem., 2021, 23, 1609–1615 | 1615