Reversible catalytic dehydrogenation of alcohols for energy storage

Article abstract of DOI:10.1073/pnas.1420199112

Reversibility of a dehydrogenation/hydrogenation catalytic reaction has been an elusive target for homogeneous catalysis. In this report, reversible acceptorless dehydrogenation of secondary alcohols and diols on iron pincer complexes and reversible oxidative dehydrogenation of primary alcohols/reduction of aldehydes with separate transfer of protons and electrons on iridium complexes are shown. This reactivity suggests a strategy for the development of reversible fuel cell electrocatalysts for partial oxidation (dehydrogenation) of hydroxyl-containing fuels.

Full text of DOI:10.1073/pnas.1420199112

Reversible catalytic dehydrogenation of alcohols for
energy storage
Peter J. Bonitatibus Jr.a, Sumit Chakrabortyb, Mark D. Dohertya,1, Oltea Siclovana, William D. Jonesb,
and Grigorii L. Soloveichika
aChemical & Environmental Systems Organization, GE Global Research, Niskayuna, NY 12309; and bDepartment of Chemistry, University of Rochester,
Rochester, NY 14627
Edited by Maurice Brookhart, The University of North Carolina at Chapel Hill, Chapel Hill, NC, and approved December 17, 2014 (received for review October
21, 2014)
Reversibility of a dehydrogenation/hydrogenation catalytic reaction
has been an elusive target for homogeneous catalysis. In this report,
reversible acceptorless dehydrogenation of secondary alcohols and
diols on iron pincer complexes and reversible oxidative dehydroge-
nation of primary alcohols/reduction of aldehydes with separate
transfer of protons and electrons on iridium complexes are shown.
This reactivity suggests a strategy for the development of reversible
fuel cell electrocatalysts for partial oxidation (dehydrogenation) of
hydroxyl-containing fuels.
LHn ⇌ L + n=2ꢀ H2;
[1]
[2]
LHn ⇌ L + nꢀ H+ + nꢀ e−;
and
n=2ꢀ O2 + nꢀ H+ + nꢀ e− ⇌ n=2ꢀ H2O:
[3]
A thermodynamic analysis of a variety of potential LOHCs
showed that cyclic hydrocarbons exhibit high hydrogen contents
but that nitrogen heterocycles exhibit lower reaction enthalpies
(9, 10). Nitrogen heterocycles are also more practical, because
they display energy densities that are comparable with those of
liquid hydrogen, and the theoretical open cell potentials of these
materials are calculated to be close to or exceed the potential of
the hydrogen–oxygen fuel cell (11). In addition, the overpotential
of their electrooxidation is also smaller compared with cyclic
hydrocarbons (12). However, basic nitrogen heterocycles are not
compatible with commonly used acidic proton exchange mem-
branes because of the formation of a nonconductive salt. One
alternative class of LOHCs that does not suffer from this problem
is hydroxyl-containing compounds (e.g., alcohols and diols) that
feature reasonable hydrogen content and low oxidation potentials.
Both mono- and polysubstituted alcohols have been proposed as
hydrogen storage materials (13) and used as fuel for direct alcohol
fuel cells, usually in the form of an aqueous alkaline solution (14).
Homogeneous catalysts for the acceptorless dehydrogenation
of primary and secondary alcohols for the most part contain pre-
cious metals, such as Ru (19), Rh (20), and Ir (21). By comparison,
catalysis energy storage reversibility hydrogenation dehydrogenation
|
|
|
|
ydrogenation and dehydrogenation reactions are funda-
mental in synthetic organic chemistry and used in a variety
H
of large- and small-scale processes for manufacturing chemicals,
pharmaceuticals, foods, and fuels. An increased need for energy
storage technologies, in large part because of the recent de-
ployment of intermittent renewable energy sources, has gener-
ated a renewed interest in hydrogen as a form of chemical energy
storage. Hydrogen, which may be used in fuel cells or internal
combustion engines, is best-suited for longer-term energy storage
and can be stored in the form of a compressed gas or a cryogenic
liquid or chemically bonded in hydrides (1). The most attractive
hydrogen storage media are liquid organic hydrogen carriers
(LOHCs), because they have relatively high hydrogen content
and can be transported and distributed using the existing liquid
fuel infrastructure (2–5).
Two strategies for coupling the chemical energy stored
in these LOHCs with an energy storage device include
thermal dehydrogenation to provide H2(g) for a polymer elec-
trolyte membrane (PEM) fuel cell (expression 1) (6) and elec-
trochemical dehydrogenation to yield protons and electrons in
a direct alcohol fuel cell (expressions 2 and 3) (2, 7). In the
former (acceptorless strategy), hydrogen release from LOHCs
frequently requires high reaction temperatures and expensive
platinum group metal (PGM) catalysts. Regeneration of the
hydrogen-depleted compounds can be achieved with both PGM
and less expensive non-PGM catalysts (e.g., nickel-based);
however, elevated hydrogen pressures are needed (8, 9). The
latter was the basis for an Energy Frontier Research Center
around Electrocatalysis, Transport Phenomena, and Materials
for Innovative Energy Storage funded by the Department of
Energy (Acknowledgments), which assumed the use of a single
electrocatalyst for dehydrogenation and hydrogenation of
LOHCs that would also simplify the conventional hydrogen
storage process (10). The envisioned partial electrochemical
dehydrogenation of LOHCs typically involves expensive PGM
catalysts (11–18), with weak bases to serve as proton scavengers in
lieu of a proton exchange membrane. Reversing the applied po-
tential in the presence of these same components provides a
mechanism for regenerating the LOHCs without the need for ele-
vated hydrogen pressure:
Significance
Catalytic hydrogenation and dehydrogenation reactions are ex-
tremely important in organic chemistry and recently for energy
storage in the form of chemical bonds. Although catalysts are
known which catalyze both reactions, the rates and conditions re-
quired for the two are frequently very different due to the differ-
ences associated with the bonds to be activated (C–H/O–H/N–H and
C = O/C = N/H–H). The use of a bifunctional catalyst would
substantially simplify the design of processes related to energy
storage. In this work, organometallic complexes of iron and
iridium are shown to act as catalysts for reversible de-
hydrogenation of alcohols to carbonyl compounds. This finding
opens a pathway to the development of catalysts for direct
reversible electrochemical dehydrogenation of organic fuels in
energy generation and storage reactions.
Author contributions: W.D.J. and G.L.S. designed research; P.J.B., S.C., M.D.D., and O.S.
performed research; M.D.D. contributed new reagents/analytic tools; P.J.B., S.C., W.D.J.,
and G.L.S. analyzed data; and P.J.B., S.C., M.D.D., W.D.J., and G.L.S. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1To whom correspondence should be addressed. Email: doherty@ge.com.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1420199112/-/DCSupplemental.
www.pnas.org/cgi/doi/10.1073/pnas.1420199112
PNAS Early Edition
|
1 of 6

the same reaction with nonprecious, earth-abundant metal catalysts
is, so far, a relatively unexplored area in the literature. The first
cobalt catalyst bearing a noninnocent bis(dicyclohexylphosphino)
amine (PNPCy) ligand for acceptorless dehydrogenation of alco-
hols was reported by Hanson and coworkers (22, 23). Recently,
Beller and coworkers (24) have shown hydrogen production from
methanol in the presence of KOH with octahedral iron complexes
(PNPiPr)Fe(H)(CO)X [X = BH4 (1) and X = Br (2)]. Remarkably,
very low catalyst loadings (parts per million level) were used, and
catalysis was performed at 91 °C, which suggests high thermal
stability of these iron complexes and related intermediate species.
However, Guan and coworkers (25) have accomplished ester
hydrogenation using catalyst 1. In addition to these studies, it
was recently shown that the same iron complexes can also effi-
ciently catalyze the reversible dehydrogenation–hydrogenation
of N-heterocycles (26). Yamaguchi et al. (27) have also reported
tetrahydroquinoline to quinoline in boiling xylene. Dehydrogenation
is harder and routinely produced lower yields than hydrogenation,
but with the 5-trifluoromethylpyridonate ligand, a quantitative yield
was achieved in both directions (27). A computational analysis of the
proposed catalytic cycle showed that two major pathways are possi-
ble: through a bifunctional species with coordinated pyridonate li-
gand or through a monomeric Cp*Ir(H)Cl complex (28).
As part of our ongoing effort to develop another strategy,
specifically electrocatalysts for reversible partial oxidation
(dehydrogenation) of alcohol-based fuels, we concentrated on
the possibility of converting a known dehydrogenation catalyst
to an electrocatalyst through separation of protons and elec-
trons. We regard separately extracting protons and electrons in
a catalytic dehydrogenation reaction together with the micro-
scopic reverse (29) (i.e., separately injecting protons and elec-
trons to effect substrate hydrogenation) as fundamental in the
Cp*Ir complexes with substituted pyridonate ligands that development of a reversible alcohol dehydrogenation–hydroge-
catalyze the reversible acceptorless dehydrogenation of 1,2,3,4- nation electrocatalyst. It was clear at the outset that the possibility
Table 1. Iron-catalyzed acceptorless dehydrogenation of alcohols
Entry
1
Substrate
Product
Time (h)
24
Isolated yield (%)
89
2
3
24
12
92
78
4
12
87
5
6
24
8
64
96
7
24
59
8
9
24
24
67
61
[1] = 0.005 M, [substrate] = 0.5 M.
2 of 6
|
www.pnas.org/cgi/doi/10.1073/pnas.1420199112
Bonitatibus et al.

Table 2. Iron-catalyzed hydrogenation of 4’-methoxyacetophenone
Entry
Catalyst
Solvent
Additive
Time (h)
NMR conversion (%)
1
2*
3
1
2
3
Toluene
THF
Toluene
None
KOtBu
None
8
8
8
100
100
100
*10 mol% KOtBu was added. [Catalyst] = 0.005 M, [KOtBu] = 0.05 M, [4′-methoxyacetophenone] = 0.5 M.
of oxidation of intermediate species containing metals in low-
amount of time. Therefore, we focused on complex 1 to investigate
oxidation states during dehydrogenation and the competing the scope of substrates for the iron-catalyzed acceptorless de-
hydrogenation of alcohols.
proton reduction to H2 in the reverse reaction substantially limited
our selection of possible candidates. We recently reported elec-
trocatalytic properties of an iridium amino-olefin complex [Ir
(trop2DACH)][OTf], which is capable of catalyzing alcohol
dehydrogenation with chemical oxidants as well as electro-
catalytic dehydrogenation of primary alcohols with excellent far-
adaic efficiency (30). Two catalytic systems capable of oxidizing
alcohols with a chemical oxidant (ferrocenium cation) in the
presence of a base as a proton acceptor have been very recently
described in literature (30, 31).
In this report, we describe catalytic systems that address both
strategies outlined above as part of a unified effort to develop
catalysts for reversible dehydrogenation of organic fuels in energy
generation and storage reactions. These catalysts show reversible
acceptorless (expression 1) and partial oxidative (expression 2)
dehydrogenation of alcohols using non-PGM (iron-based) and
PGM (iridium-based) catalysts, respectively.
As depicted in Table 1, secondary benzylic alcohols (Table 1,
entries 1–4) were successfully dehydrogenated by catalyst 1, and
the corresponding acetophenone derivatives were isolated in
good yields (64–92%). A variety of functionalities, such as -OMe,
-NO2, and -Cl, was unaffected under the catalytic conditions.
Electronic influence on the dehydrogenation activity seemed to
be significant, because a substrate containing an electron-with-
drawing -NO2 group (Table 1, entry 3) reacted much faster than
the one with an electron-donating -OMe group (Table 1, entry 2)
based on respective NMR conversions after 12 h. For the -OMe
derivative, 68% conversion was achieved after 12 h, whereas
100% conversion was observed for the -NO2 derivative after the
same time. A more detailed kinetic analysis of the mechanism is
required to rationalize these electronic effects, and these studies
are currently underway. In addition to the aromatic substrates,
cyclohexanol (Table 1, entry 5) was also successfully dehydro-
genated to give cyclohexanone. In contrast to 1, activation of 2
with a base is required (SI Appendix, Table S1).
Results and Discussion
Because diols can potentially release 4 eq of H to afford lac-
tones, we investigated the reactivity of complex 1 toward several
diol substrates. When 1,2-benzenedimethanol (Table 1, entry 6)
was subjected to our catalytic conditions, it readily produced the
corresponding lactone phthalide. Because this reaction suggested
the formation of a transient aldehyde intermediate from a pri-
mary alcohol, we tried to verify its formation by performing de-
hydrogenation of benzyl alcohol itself. However, benzyl benzoate
was isolated as the sole product as a result of an intermolecular
condensation process; 1,5-pentanediol (Table 1, entry 7) was also
dehydrogenated successfully to afford δ-valerolactone. Several
research groups have previously reported homogeneous catalysts
for the base-free oxidation of diols to lactones in the presence of
a hydrogen acceptor, such as acetone (32–35). However, reports
related to the acceptorless conversion of the same are extremely
rare in the literature (36–39). To the best of our knowledge, only
recently have a few iron-based homogeneous catalysts been re-
ported for this transformation (40). Fe-catalyzed acceptorless
alcohol dehydrogenation with a base as cocatalyst was also re-
cently reported (41).
Reversible Acceptorless Dehydrogenation of Alcohols with Iron
Complexes. We initially studied acceptorless dehydrogenation of
1-phenylethanol with catalysts iPr[PN(H)P]Fe(H)(CO)(HBH3)
(1) and iPr[PN(H)P]Fe(H)(CO)(Br) (2). On investigating a variety
of conditions, it was determined that the catalytic reaction was
best carried out in toluene at 120 °C. Under these conditions,
when 1 mol% 1 was used as the catalyst, 1-phenylethanol was
quantitatively converted to acetophenone within 24 h as judged
by GC-MS and NMR spectroscopy. In a closed system, only 28%
of acetophenone formed after 24 h. Complex 2 was also found
to be an active catalyst in the presence of a catalytic amount of
KOtBu; however, a lower conversion was achieved after the same
Reaction of diols with both primary and secondary hydroxyl
groups with catalyst 1 results in similar yield and chemoselectivity
(exclusive dehydrogenation of the secondary alcohol moiety) for
compounds with different acidity of the methine C–H bond (Table
1, entry 8 compared with entry 9). This preference can be explained
by the difference in thermodynamics for dehydrogenation of pri-
mary and secondary hydroxyl groups. For example, the Gibbs free
energy of dehydrogenation for 1-butanol is >30 kJ/mol more
positive than that for 2-butanol [+33.5 vs. +2.5 kJ/mol at 120 °C
based on values from the National Institute of Standards and
Technology database (42)].
To show the possibility of using these iron catalysts for the
reverse reaction of alcohol dehydrogenation, we examined their
activity toward the hydrogenation of a corresponding ketone
(Table 2). Although Guan and coworkers (25) have shown that
Scheme 1. Proposed pathway for the alcohol dehydrogenation on
iron PNP complexes.
Bonitatibus et al.
PNAS Early Edition
|
3 of 6

electron transfer processes (instead of simultaneous transfer of
protons and electrons to the same molecule, such as in BQ) to
show the first necessary step in a broader strategy to transition
Grützmacher’s system to an electrocatalytic (i.e., electrode-
driven) dehydrogenation and use such complexes as electro-
catalysts (30). To verify the possibility of operating these catalysts
in both directions, we have also studied the reverse reaction,
separately injecting protons and electrons for aldehyde hydro-
genation using 5 and 6 combined with stoichiometric reductants
(e.g., cobaltocene) and weak conjugate acid (phenol).
Considering the high efficiency reported in the original system
using 5 with stoichiometric BQ (example, TOF greater than
150,000 h−1 for benzyl alcohol dehydrogenation in chloroben-
zene at 90 °C) (44, 46), we used phenolate anion as a substitute
base combined with various chemical oxidants for the catalytic
dehydrogenation of 4-methoxybenzyl alcohol to p-anisaldehyde
(Table 3). Oxidations were carried out anaerobically at room
temperature with 0.03 mol% 5 or 0.1 mol% 6 and required 2 eq
each phenolate and oxidant relative to 4-methoxybenzyl alcohol.
Oxidants were surveyed with each catalyst over a relatively nar-
row range of chemical potentials in o-dichlorobenzene (o-DCB).
All were greater than −0.13 V vs. the ferrocenium/ferrocene
couple (Fc+/0), which is the peak oxidation potential (observed to
be pseudoreversible at scan rates ≥ 1 V/s) that we reported for
monodeprotonated 5, namely the neutral Ir-amido-amine com-
plex [Ir(trop2DACH-1H)]0 (30).
From Table 3, we found 5 to be more active and efficient for
the dehydrogenation of 4-methoxybenzyl alcohol compared with
6, exhibiting higher rates with lower catalyst loadings while main-
taining comparable moderate-to-high conversions to aldehyde. We
observed a 94% yield of p-anisaldehyde after 2 h when the reaction
was carried out with 5, phenolate, and AgOTf (Table 3, entry 4),
whereas the same reaction with 6 gave only 63% yield after 4 h
(Table 3, entry 9). Oxidant solubility played an important role in
catalysis—acetylated ferrocenium oxidants ([Fe(η5-C5H4C(O)Me)
Cp]+ and [Fe(η5-C5H4C(O)Me)2]+ in Table 3, entries 3, 5, 8, and 10)
showed very poor solubility and performance in o-DCB compared
with [FeCp2]+ and [Fe(η5-C5H4Me)2]+ oxidants. Of significant note
was that, by separating proton and electron transfer events in
4-methoxybenzyl alcohol dehydrogenation, no penalties were ob-
served in terms of reaction rate or percentage conversion to aldehyde
(for Table 3, entry 4 specifically) relative to the concerted process
(i.e., Grützmacher’s 5-BQ system exhibited a TOF of 625 h−1 at
25 °C in chlorobenzene for oxidation of 4-methoxybenzyl alcohol).
Electrochemical studies of 5 and related complexes ([Ir
(trop2DACH-1H)]0 and [Ir(trop2DACH-2H)]−) (46) allowed for
refinement of our understanding around catalyst activation and
determination of a thermochemical cycle (details in SI Appendix).
For example, calculation of H atom removal from 5 (homolytic
Scheme 2. Reversible oxidative dehydrogenation of 4-methoxybenzyl alcohol
catalyzed by complexes 5 and 6.
complex 1 reduces ester functionalities with 10 atm H2, hydroge-
nation of a simple ketone has not been reported so far. In our in-
vestigation, we found that 1 mol% complex 1 catalyzes quantitative
hydrogenation of 4-methoxyacetophenone with 5.4 atm H2 pressure
in toluene at room temperature (Table 2, entry 1). Both complexes
2 and iPr[PNP]Fe(H)(CO) (3) also served as equally effective cat-
alysts in this hydrogenation reaction (Table 2, entries 2 and 3).
A plausible mechanism for the iron-catalyzed alcohol de-
hydrogenation is outlined in Scheme 1. Based on our current
understanding of the dehydrogenation of N-heterocycles (26), we
propose that the amido-iron complex 3 is the active catalyst in
the alcohol dehydrogenation. Consistent with this hypothesis,
catalytic hydrogenation of 4-methoxyacetophenone (Table 2,
entry 3) and dehydrogenation of 1-phenylethanol (in toluene at
120 °C) were quantitative when 1 mol% 3 was directly used as
the catalyst. The other iron species that is possibly involved in the
catalytic cycle is a trans-dihydride complex (4). The existence of
this species, which tends to lose H2 even at room temperature
and thereby, regenerates the active dehydrogenation intermediate
3, was previously supported by NMR spectroscopy and trapping
experiments (26). A computational analysis using density func-
tional theory (DFT) suggests that the dehydrogenation of alcohol,
a key elementary step, occurs in a concerted fashion through
metal–ligand cooperation (43). A stepwise pathway involving an
iron-alkoxide intermediate (shown in Scheme 1) is also possible,
although less likely, because it would require higher activation
energy (DFT calculations are shown in ref. 43). Catalytic results
mentioned in this report indicate that iron complexes 1–3 repre-
sent a rare class of homogeneous catalysts for reversible de-
hydrogenation/hydrogenation of alcohols and ketones.
Reversible Oxidative Dehydrogenation of Primary Alcohols with Iridium
Complexes. To drive a dehydrogenation reaction toward separate
generation of protons and electrons requires a stoichiometric
oxidant combined with a weak base as a proton sink, and for the
reverse reaction (catalytic hydrogenation), a stoichiometric amount
of reductant is necessary combined with a (weak) conjugate acid
as a proton source. Herein, we studied two iridium(I) complexes,
[Ir(trop2DACH)][OTf] (5), a highly efficient catalyst used in
the dehydrogenation of a broad range of primary alcohols to
aldehydes (44), and the related [Ir(trop2DAD)][OTf] (6) (45), which
features an unsaturated and sterically less encumbered ligand
framework, as reversible alcohol dehydrogenation–hydrogenation
catalysts (Scheme 2).
Dehydrogenation of benzylic and allylic alcohols proceeds at
room temperature with very low loadings of 5 (0.01 mol%) and
requires catalytic amounts of strong base (e.g., KOtBu; 0.03 mol%)
and stoichiometric 1,4-benzoquinone (BQ) as a hydrogen (H2)
scavenger (44, 46). We now report our findings with 5 and 6
for alcohol dehydrogenation in terms of separately extracting
protons and electrons through chemical catalysis results that
used stoichiometric oxidants (e.g., ferrocenium) combined with
a weak base (phenolate). We targeted separation of proton and
Table 3. Dehydrogenation of 4-methoxybenzyl alcohol with
chemical oxidants in the presence of sodium phenolate as base in
o-DCB using 5 (0.03 mol%) and 6 (0.1 mol%) as catalysts
E1/2
Time GC yield
Entry Catalyst
Oxidant
(V)
(h)
(%)
1
2
3
5
5
5
[Fe(η5-C5H4Me)2]BF4
[FeCp2]PF6
−0.11
0.00
0.25
2
2
2
75
70
11
[Fe(η5-C5H4C(O)Me)Cp]
BF4
4
5
6
7
8
5
5
6
6
6
AgOTf
0.39*
0.46
−0.11
0.00
2
2
4
4
4
94
4
75
76
17
[Fe(η5-C5H4C(O)Me)2]BF4
[Fe(η5-C5H4Me)2]BF4
[FeCp2]PF6
[Fe(η5-C5H4C(O)Me)Cp]
BF4
0.25
9
10
6
6
AgOTf
0.39*
0.46
4
4
63
16
[Fe(η5-C5H4C(O)Me)2]BF4
*Reported here as the observed peak potential and not formal E1/2
.
4 of 6
|
www.pnas.org/cgi/doi/10.1073/pnas.1420199112
Bonitatibus et al.

hydrogenation. Likewise, [Cr(η6-C6H6)2] was not expected or
observed to activate both 5 and 6. Our initial hydrogenation
attempts that used 2.5 eq each phenol and [CoCp2] provided low
yields of p-anisaldehyde (48% and 27% with 6 and 5, respec-
tively) and produced large amounts of precipitate that indicated
acid-starved conditions of this reaction. To avoid this acid-
starved condition, all hydrogenation reactions in Table 4 were
run as a 2:1 ratio of phenol to reductant. We identified a de-
finitive sweet spot at the [CoCp2] redox potential and found 6 to
be more effective and active compared with 5. An 89% yield of 4-
methoxybenzyl alcohol was observed after 24 h at room tem-
perature with 6, phenol, and [CoCp2] (Table 4, entry 8), whereas
the same reaction with 5 afforded a 60% yield after 72 h (Table
4, entry 3). Such as was observed in the dehydrogenation re-
action, solubility of the reductants used played an important role
in the effectiveness of catalytic hydrogenation. [CoCp*] has
a reduction potential sufficient to reduce 5 and 6; howeve2r, very
poor solubility in o-DCB rendered it largely ineffective with 5
(despite stirring for 3 d) and incapable of reducing 6. (Note that
Scheme 3. Products of [Ir(trop2DACH)] oxidation.
bond dissociation) to form [Ir(trop2DACH-1H)]+ (Scheme 3,
Center) was determined to require 79.2 kcal/mol, with comparable
energy necessary to repeat the deprotonation/oxidation step (77.6
kcal/mol). (Note that [Ir(trop2DACH-2H)]+ may also be consid-
ered as an Ir(III)-bis-amide complex.) By derivation of additional
thermochemical data, the pKa of a monodeprotonated mono-
oxidized species, [Ir(trop2DACH-1H)]+, was calculated to be 9.2
1
(in DMSO), which is actually lower than the pKa of 10.5 mea-
sured for 5 (46). Together, these data are consistent with activa-
tion of 5 beyond simple monodeprotonation and oxidation under
our catalysis conditions. Likely, a second deprotonation by phenolate
base occurs to engender [Ir(trop2DACH-2H)]0—an intermediate
also proposed by Grützmacher in the original KOtBu-initiated
BQ-mediated system. Additional oxidation of [Ir(trop2DACH-2H)]0
to [Ir(trop2DACH-2H)]+ is reasonable given that the potentials of
chemical oxidants in Table 3 are much higher than the E1/2 measured
[CoCp*] remained in a reduced state at the end of the reaction
2
and did not consume/reduce phenol to any significant extent.)
Conversely, [Co(η5-C5H4Et)2] is a liquid, and hydrogenations
were observed to be homogeneous with near-complete con-
sumption of reductant. With high o-DCB solubility, excess
phenol consumed [Co(η5-C5H4Et)2], and low conversions resulted
(despite negligible side products).
for interconversion between these two species (−0.20 V vs. Fc+/0
)
(SI Appendix).
Summary
Although aldehyde yields for 5 (Table 3, entries 1 and 2) and 6
(Table 3, entries 6 and 7) are comparable, albeit under slightly
different conditions, our understanding around activation of 6 as
a dehydrogenation catalyst is relatively limited. Accommodation
of a fifth ligand (i.e., phenolate) by 6 is expected given that the
thiophenolate [Ir(trop2DAD)(SPh)]0 is known and has been
crystallographically characterized (47). Cyclic voltammetry (CV)
of 1 mM 6 in o-DCB/0.08 M n-Bu4NPF6 electrolyte shows two
In conclusion, we have developed a rare class of homogeneous
iron catalysts (1–3) for efficient reversible acceptorless de-
hydrogenation–hydrogenation of alcohols and diols under mild
conditions. Moreover, we showed that it is possible to decouple
proton and electron transfer processes in both directions through
use of metallocene-based chemical oxidants/reductants, presum-
ably as surrogates for catalytic electrode-driven (de)hydrogenation
reactions. This work further illustrates the importance of non-
innocent ligands in enabling reversible (de)hydrogenation cataly-
sis. In all cases, catalytically active metal complexes bear such
ligands, with nitrogen atoms (capable of forming/breaking N–H
bonds) acting as a proton sink/source. The obtained results are
of importance for the development of reversible dehydrogenation
electrocatalysts and ultimately, the application of such catalysts
in regenerative liquid fuel cells using chemically bound hydrogen
for energy storage.
one-electron (quasi)reversible waves (E11/2 = −0.63 V and E12/2
=
−1.35 V vs. Fc+/0), which can be attributed to consecutive re-
duction of the diazabutadiene ligand and is consistent with prior
observations in THF (45). Addition of 4 eq phenolate base to 6
results in the appearance of two additional one-electron (quasi)
reversible waves at more positive potentials (E13/2 = −0.33 V and
E14/2 = −0.06 V vs. Fc+/Fc) (SI Appendix). It is presumed that
activation of 6 for dehydrogenation catalysis involves formation
of [Ir(trop2DAD)(OPh)]0 followed by oxidation of the coor-
dinated phenolate complex. (Note that 4 mM phenolate does
not show electrochemical activity up to +0.6 V vs. Fc+/0, which is
well above any potential used in these experiments.)
Methods
All organometallic compounds were prepared and handled under a nitrogen
atmosphere using standard Schlenk and glove box techniques. E1/2 values are
reported relative to Cp2Fe+/Cp2Fe and were measured in o-DCB/0.08 M
n-Bu4NPF6 as electrolyte at 0.1 V/s with Pt working and counter electrodes
and Ag wire/10 mM AgNO3 solution as reference electrode in CH3CN/0.1 M
n-Bu4NPF6 electrolyte. Experimental details are in SI Appendix.
We also studied the reverse reaction (catalytic hydrogenation
of p-anisaldehyde to 4-methoxybenzyl alcohol) using 5 or 6
combined with phenol as a proton source and stoichiometric
reductants (Table 4). It was necessary to run hydrogenations with
excess phenol (acid), because we observed formation of large
amounts of a precipitate, which was identified by 1H NMR, CV,
HPLC-TOF/MS, and thermogravimetric analysis (TGA) coupled
FTIR as an adduct of cobaltocenium phenolate with phenol
([CoCp2][OPh]·HOPh). Therefore, hydrogenation reactions were
carried out with 1 mol% 5 or 6, 5 eq phenol, and 2.5 eq reductant
relative to p-anisaldehyde to achieve high yields. Reductants were
surveyed with each catalyst over a relatively wide range of redox
potentials in o-DCB (Table 4), most being strong enough to
activate (reduce) 5 and 6 by one and two electrons, respectively.
Reduction of 5 to metalloradical [Ir(trop2DACH)] was reported
previously in DMSO (46), and in o-DCB/0.08 M n-Bu4NPF6
electrolyte, the peak potential observed by CV at −1.36 V vs. Fc+
/0 indicates that its formation should be possible by [CoCp2], [Co
(η5-C5H4Et)2], and [CoCp*]. Based on CV studies of 6 under
the same electrochemical2conditions, reductants [FeCp*] and
[Cr(η6-C6H6)2] listed in Table 4 are not strong enough 2to re-
duce both C = N bonds in 6.
Table 4. Hydrogenation of p-anisaldehyde with chemical
reductants in the presence of phenol as a proton source in o-DCB
using 5 and 6 (1 mol%) as catalysts
Entry
Catalyst
Reductant
E1/2 (V)*
Time (h)
Yield (%)
1
2
3
4
5
6
7
8
9
5
5
5
5
5
6
6
6
6
6
[FeCp*2]
[Cr(η6-C6H6)2]
[CoCp2]
−0.54
−1.19
−1.33
−1.44
−1.95
−0.54
−1.19
−1.33
−1.44
−1.95
72
72
72
72
72
24
24
24
24
24
0
10
60
25
36
0
[Co(η5-C5H4Et)2]
[CoCp*2]
[FeCp*2]
[Cr(η6-C6H6)2]
[CoCp2]
3
89
49
0
[Co(η5-C5H4Et)2]
[CoCp*2]
10
Data from Table 4 support our expectations that [FeCp*2] was
not strong enough to activate either 5 or 6 for catalytic aldehyde
*Relative to Cp2Fe+/Cp2Fe.
Bonitatibus et al.
PNAS Early Edition
|
5 of 6

General Procedure for the Iron-Catalyzed Dehydrogenation of Alcohols. In a glove
box, a 50-mL flame-dried Schlenk flask equipped with a condenser was charged
with an iron catalyst (25 μmol), an alcohol substrate (2.5 or 25 mmol), and 5 mL
toluene. The solution was stirred at 120 °C for a specific time under a constant N2
flow. After the reaction, the solution was allowed to cool to room temperature,
filtered through a short silica gel column, and eluted with THF. The resulting
filtrate was evaporated under vacuum to afford the pure product. 1H and 13C{1
H} NMR spectra of the products were recorded in CDCl3 and matched with the
chemical shifts reported in the literature. The results are summarized in Table 1.
General Procedure for the Iridium-Catalyzed Oxidative Dehydrogenation of
Alcohols. Dehydrogenation reactions were carried out with 0.03 mol% (5) or
0.1 mol% (6) catalyst; however, both used 1.5 mmol alcohol, 3 mmol oxidant,
and 3 mmol sodium phenolate as the base at room temperature in o-DCB (re-
action volume ∼ 5 mL). 1H NMR yields were determined against hexame-
thylbenzene as an internal standard based on peak integration of either
methoxy or aromatic protons.
General Procedure for the Iridium-Catalyzed Reductive Hydrogenation of Alcohols.
All hydrogenation reactions used 1 mol% catalyst (5 or 6), 1.25 mmol aldehyde,
3.13 mmol reductant, and 6.26 mmol phenol as acid at room temperature in o-
DCB (reaction volume ∼ 10 mL). 1H NMR yields were determined against hex-
amethylbenzene as an internal standard based on peak integration of either
methoxy or aromatic protons.
General Procedure for the Iron-Catalyzed Hydrogenation of 4′-Methoxy-
acetophenone. In a glove box, a 25-mL stainless steel Parr pressure reactor was
loaded with an iron complex (25 μmol), KOtBu (if required), 4′-methox-
yacetophenone (375 mg, 2.5 mmol), and 5 mL toluene (or THF). The reactor was
sealed, flushed with H2 three times, and finally, placed under 80 psig H2 pressure.
The solution was then stirred at room temperature for 8 h. After the reaction,
the solution was filtered through a short silica gel column and eluted with THF.
The resulting filtrate was evaporated to dryness to afford the pure hydroge-
nation product. 1H and 13C{1H} NMR spectra of the product were recorded in
CDCl3 and matched with the reported spectra in the literature. For all of the
reactions, quantitative conversions were achieved (Table 2).
ACKNOWLEDGMENTS. This material is based on work supported as part
of the Center for Electrocatalysis, Transport Phenomena, and Materials
for Innovative Energy Storage (an Energy Frontier Research Center
funded by Department of Energy, Office of Science, Office of Basic
Energy Sciences Award DE-SC0001055), and funding was provided by
a grant from Empire State Development.
1. Weidenthaler C, Felderhoff M (2011) Solid-state hydrogen storage for mobile appli-
cations: Quo vadis? Energy Environ Sci 4(7):2495–2502.
2. Crabtree RH (2008) Hydrogen storage in liquid organic heterocycles. Energy Environ
Sci 1(1):134–138.
3. Ichikawa M (2008) Organic liquid carriers for hydrogen storage. Solid-State Hydrogen
Storage: Materials and Chemistry, ed Walker G (CRC, Boca Raton, FL), pp 500–532.
4. Teichmann D, Arlt W, Wasserscheid P, Freymann R (2011) A future energy supply based
on liquid organic hydrogen carriers (LOHC). Energy Environ Sci 4(8):2767–2773.
5. Yadav M, Xu Q (2012) Liquid-phase chemical hydrogen storage materials. Energy
Environ Sci 5(12):9698–9725.
26. Chakraborty S, Brennessel WW, Jones WD (2014) A molecular iron catalyst for the
acceptorless dehydrogenation and hydrogenation of N-heterocycles. J Am Chem Soc
136(24):8564–8567.
27. Yamaguchi R, Ikeda C, Takahashi Y, Fujita K (2009) Homogeneous catalytic system for
reversible dehydrogenation-hydrogenation reactions of nitrogen heterocycles with
reversible interconversion of catalytic species. J Am Chem Soc 131(24):8410–8412.
28. Li H, Jiang J, Lu G, Huang F, Wang Z-X (2011) On the “reverse gear” mechanism of the
reversible dehydrogenation/hydrogenation of a nitrogen heterocycle catalyzed by
a Cp*Ir Complex: A computational study. Organometallics 30(11):3131–3141.
29. Tolman RC (1938) The Principles of Statistical Mechanics (Dover, New York).
30. Bonitatibus PJ, Jr, Rainka MP, Peters AJ, Simone DL, Doherty MD (2013) Highly se-
lective electrocatalytic dehydrogenation at low applied potential catalyzed by an Ir
organometallic complex. Chem Commun (Camb) 49(90):10581–10583.
31. Weiss CJ, Das P, Miller DL, Helm ML, Appel AM (2014) Catalytic oxidation of alcohol
via nickel phosphine complexes with pendant amines. ACS Catal 4(9):2951–2958.
32. Lin Y, Zhu X, Zhou Y (1992) A convenient lactonization of diols to γ- and δ-lactones
catalyzed by transition metal polyhydrides. J Organomet Chem 429(2):269–274.
33. Endo Y, Bäckvall J-E (2011) Aerobic lactonization of diols by biomimetic oxidation.
Chemistry 17(45):12596–12601.
6. Pez GP, Scott AR, Cooper AC, Cheng H (2008) US Patent 7429372.
7. Soloveichik GL, Zhao J-C (2008) US Patent Appl 20080248345.
8. Biniwale RB, Rayalu S, Devotta S, Ichikawa M (2008) Chemical hydrides: A solution to
high capacity hydrogen storage and supply. Int J Hydrogen Energy 33(1):360–365.
9. Bond GC (2006) Metal-Catalysed Reactions of Hydrocarbons (Springer, Berlin).
10. Soloveichik GL, Lemmon JP, Zhao J-C (2008) US Patent Appl 2008/0248339.
11. Rodríguez M, et al. (2001) Synthesis, structure, and redox and catalytic properties of
a new family of ruthenium complexes containing the tridentate bpea ligand. Inorg
Chem 40(17):4150–4156.
34. Buntara T, et al. (2011) Caprolactam from renewable resources: Catalytic conversion
of 5-hydroxymethylfurfural into caprolactone. Angew Chem Int Ed Engl 50(31):
7083–7087.
35. Murahashi S, Naota T, Ito K, Maeda Y, Taki H (1987) Ruthenium-catalyzed oxidative
transformation of alcohols and aldehydes to esters and lactones. J Org Chem 52(19):
4319–4327.
36. Zhang J, Balaraman E, Leitus G, Milstein D (2011) Electron-rich PNP- and PNN-type
ruthenium(II) hydrido borohydride pincer complexes. Synthesis, structure, and cata-
lytic dehydrogenation of alcohols and hydrogenation of esters. Organometallics
30(21):5716–5724.
37. Zhao J, Hartwig JF (2005) Acceptorless, neat, ruthenium-catalyzed dehydrogenative
cyclization of diols to lactones. Organometallics 24(10):2441–2446.
38. Muñiz K (2005) Bifunctional metal-ligand catalysis: Hydrogenations and new reactions
within the metal-(di)amine scaffold. Angew Chem Int Ed Engl 44(41):6622–6627.
39. Tseng K-NT, Kampf JW, Szymczak NK (2013) Base-free, acceptorless, and chemo-
selective alcohol dehydrogenation catalyzed by an amide-derived NNN-ruthenium(II)
hydride complex. Organometallics 32(7):2046–2049.
12. Hino T, Wada T, Fujihara T, Tanaka K (2004) Redox behavior of new Ru-dioxolene and
-ammine complexes and catalytic activity toward electrochemical oxidation of alcohol
under mild conditions. Chem Lett 33(12):1596–1597.
13. Cheung K-C, Wong W-L, Ma D-L, Lai T-S, Wong K-Y (2007) Transition metal complexes
as electrocatalysts—development and applications in electro-oxidation reactions.
Coord Chem Rev 251(17-20):2367–2385.
14. Serra D, Correia MC, McElwee-White L (2011) Iron and ruthenium heterobimetallic
carbonyl complexes as electrocatalysts for alcohol oxidation: Electrochemical and
mechanistic studies. Organometallics 30(21):5568–5577.
15. Ozawa H, Hino T, Ohtsu H, Wada T, Tanaka K (2011) A new type of electrochemical
oxidation of alcohols mediated with
a ruthenium–dioxolene–amine complex in
neutral water. Inorg Chim Acta 366(1):298–302.
16. Yamazaki S, et al. (2012) Electrocatalytic oxidation of alcohols by a carbon-supported
Rh porphyrin. Chem Commun (Camb) 48(36):4353–4355.
17. Vannucci AK, et al. (2012) Water oxidation intermediates applied to catalysis: Benzyl
alcohol oxidation. J Am Chem Soc 134(9):3972–3975.
18. Hull JF, et al. (2012) Reversible hydrogen storage using CO2 and a proton-switchable
iridium catalyst in aqueous media under mild temperatures and pressures. Nat Chem
4(5):383–388.
40. Kamitani M, Ito M, Itazaki M, Nakazawa H (2014) Effective dehydrogenation of
2-pyridylmethanol derivatives catalyzed by an iron complex. Chem Commun (Camb)
50(59):7941–7944.
19. Muthaiah S, Hong SH (2012) Acceptorless and base-free dehydrogenation of alcohols
and amines using ruthenium-hydride complexes. Adv Synth Catal 354(16):3045–3053.
20. Trincado M, Banerjee D, Grutzmacher H (2014) Molecular catalysts for hydrogen
production from alcohols. Energy Environ Sci 7(8):2464–2503.
41. Song H, Kang B, Hong SH (2014) Fe-catalyzed acceptorless dehydrogenation of sec-
ondary benzylic alcohols. ACS Catal 4(9):2889–2895.
42. NIST (2005) NIST Chemistry WebBook. Available at webbook.nist.gov/chemistry. Ac-
cessed January 2, 2015.
21. Zeng G, Sakaki S, Fujita K-i, Sano H, Yamaguchi R (2014) Efficient catalyst for ac-
ceptorless alcohol dehydrogenation: Interplay of theoretical and experimental stud-
ies. ACS Catal 4(3):1010–1020.
22. Zhang G, Hanson SK (2013) Cobalt-catalyzed acceptorless alcohol dehydrogenation:
Synthesis of imines from alcohols and amines. Org Lett 15(3):650–653.
23. Zhang G, Vasudevan KV, Scott BL, Hanson SK (2013) Understanding the mechanisms
of cobalt-catalyzed hydrogenation and dehydrogenation reactions. J Am Chem Soc
135(23):8668–8681.
24. Alberico E, et al. (2013) Selective hydrogen production from methanol with a defined
iron pincer catalyst under mild conditions. Angew Chem Int Ed Engl 52(52):14162–14166.
25. Chakraborty S, et al. (2014) Iron-based catalysts for the hydrogenation of esters to
alcohols. J Am Chem Soc 136(22):7869–7872.
43. Chakraborty S, et al. (2014) Well-defined iron catalysts for the acceptorless reversible
dehydrogenation-hydrogenation of alcohols and ketones. ACS Catal 4(11):3994–4003.
44. Königsmann M, et al. (2007) Metalloenzyme-inspired catalysis: Selective oxidation of
primary alcohols with an iridium-aminyl-radical complex. Angew Chem Int Ed Engl
46(19):3567–3570.
45. Breher F, et al. (2003) Tropad: A new ligand for the synthesis of water-stable para-
magnetic [16+1]-electron rhodium and iridium complexes. Chemistry 9(16):3859–3866.
46. Donati N, Königsmann M, Stein D, Udino L, Grützmacher H (2007) Iridium aminyl
radical complexes as catalysts for the catalytic dehydrogenation of primary hydroxyl
functions in natural products. C R Chim 10(8):721–730.
47. Häbe
K (2006) Chemie von “Non-Innocent” Liganden, Rhodium und Iridium in
Stickstoff-Olefinkomplexen. Doctor of Science (ETH Zürich, Zurich).
6 of 6
|
www.pnas.org/cgi/doi/10.1073/pnas.1420199112
Bonitatibus et al.