Efficient transfer hydrogenation of carbonate salts from glycerol using water-soluble iridium N-heterocyclic carbene catalysts

Article abstract of DOI:10.1039/d0gc01958e

The transfer hydrogenation of CO₂and carbonates from biomass-derived alcohols, such as glycerol, to afford formic and lactic acid is a highly attractive path to valorizing two waste streams, and is significantly more thermodynamically favorable than direct carbonate hydrogenation. Expanding on our seminal report of the first homogeneous catalyst for this process, here we show that thermally-robust and water-soluble Ir(i) and Ir(iii) N-heterocyclic carbene (NHC) complexes with sulfonate-functionalized wingtips are highly prolific and robust catalysts for carbonate transfer hydrogenation from glycerol, requiring no additives in aqueous media. The most prolific catalyst of the nine examined, [Ir(NHC-Ph-SO₃^?)₂CO₂]Na (cat7), effectively facilitates the reaction at low catalyst loading (10 ppm) at 150 °C using microwave or conventional heating. The cation of the carbonate salt significantly impacts catalytic activity, with highest activity observed with Cs₂CO₃(27?850 and 13?350 TONs for lactate and formate respectively in 6 hours, compared to 15?400 and 8120 with K₂CO₃). Catalytic amounts of Cs⁺were found to significantly enhance activity with K₂CO₃. Catalyst7is even more prolific with conventional heating under a positive N₂pressure, reaching TOFs of >3000 h^?1and >2100 h^?1respectively for lactate and formate with K₂CO₃. The high activity of this catalyst compared to non-sulfonated and cyclooctadiene analogs is attributed to a combination of catalyst solubility in aqueous media and presence of π-acceptor carbonyl ligands. A catalytic mechanism is proposed for7involving O-H oxidative addition of glycerol, β-hydride elimination, bicarbonate dehydroxylation, insertion and reductive elimination.

Full text of DOI:10.1039/d0gc01958e

View Article Online
View Journal
Green
Chemistry
Cutting-edge research for a greener sustainable future
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: D. Ainembabazi,
K. Wang, M. T. Finn, J. A. Ridenour and A. Voutchkova, Green Chem., 2020, DOI: 10.1039/D0GC01958E.
Volume 20
This is an Accepted Manuscript, which has been through the
Number
May 2018
Pages 1919-2160
9
7
Royal Society of Chemistry peer review process and has been
accepted for publication.
Green
Chemistry
Cutting-edge research for a greener sustainable future
rsc.li/greenchem
Accepted Manuscripts are published online shortly after acceptance,
before technical editing, formatting and proof reading. Using this free
service, authors can make their results available to the community, in
citable form, before we publish the edited article. We will replace this
Accepted Manuscript with the edited and formatted Advance Article as
soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes to the
text and/or graphics, which may alter content. The journal’s standard
Terms & Conditions and the Ethical guidelines still apply. In no event
shall the Royal Society of Chemistry be held responsible for any errors
or omissions in this Accepted Manuscript or any consequences arising
from the use of any information it contains.
ISSN 1463-9262
PAPER
Paul T. Anastas et al.
The Green ChemisTREE: 20 years after taking root with the 12
principles
rsc.li/greenchem

Please do not adjust margins
Page 1 of 11
Green Chemistry
View Article Online
DOI: 10.1039/D0GC01958E
ARTICLE
Efficient Transfer Hydrogenation of Carbonate Salts from Glycerol
using Water-Soluble Iridium N-Heterocyclic Carbene Catalysts
Diana Ainembabazia, Kai Wanga, Matthew Finna, James Ridenoura, Adelina Voutchkova-Kostala*
Received 00th January 20xx,
Accepted 00th January 20xx
The transfer hydrogenation of CO2 and carbonates from biomass-derived alcohols, such as glycerol, to afford formic and
lactic acid is a highly attractive path to valorizing two waste streams, and is significantly more thermodynamically favorable
than direct carbonate hydrogenation. Expanding on our seminal report of the first homogeneous catalyst for this process,
here we show that thermally-robust and water-soluble Ir(I) and Ir(III) N-heterocyclic carbene (NHC) complexes with
sulfonate-functionalized wingtips are highly prolific and robust catalysts for carbonate transfer hydrogenation from glycerol,
DOI: 10.1039/x0xx00000x
-
requiring no additives in aqueous media. The most prolific catalyst of the nine examined, [Ir(NHC-Ph-SO3 )2CO2]Na (cat 7),
effectively facilitates the reaction at low catalyst loading (10 ppm) at 150 °C using microwave or conventional heating. The
cation of the carbonate salt significantly impacts catalytic activity, with highest activity observed with Cs2CO3 (27850 and
13350 TONs for lactate and formate respectively in 6 hours, compared to 15400 and 8120 with K2CO3). Catalytic amounts of
Cs+ were found to significantly enhance activity with K2CO3. Catalyst 7 is even more prolific with conventional heating under
a positive N2 pressure, reaching TOFs of >3000 h-1 and >2100 h-1 respectively for lactate and formate with K2CO3. The high
activity of this catalyst compared to non-sulfonated and cyclooctadiene analogs is attributed to a combination of catalyst
solubility in aqueous media and presence of -acceptor carbonyl ligands. A catalytic mechanism is proposed for 7 involving
O-H oxidative addition of glycerol, -hydride elimination, bicarbonate dehydroxylation, insertion and reductive elimination.
ESI Scheme S1 summarizes the efficiency of recent catalytic examples
for direct hydrogenation of both substrates.
Introduction
The development of methods to convert CO2 and carbonates to
valuable chemicals, such as formic acid, formaldehyde and methanol,
has attracted increased interest due to the abundance of this C1
feedstock. Formic acid is a particularly valuable target, given the
recent increase in its demand for applications ranging from chemical
feedstocks,[1] fuels,[2] hydrogen storage media,[3] as well as
commodity chemicals for food and agriculture.[1] The primary
synthetic route to formic acid form CO2 is direct hydrogenation,
which since its first report in the 70’s[4] continues to be an active area
of research. Among the most active homogeneous catalysts reported
for CO2 hydrogenation are Rh,[5, 6] Ru,[7, 8] Ir,[9] Fe,[10] Ni[11] and Co.[12]
However, direct hydrogenation of CO2 is a thermodynamically
unfavourable process (G = 13.4 kcal/mol),24 requiring base, amine
additives and/or aqueous conditions to drive the reaction by
stabilizing the product as the formate salt.[8, 13, 14] Given that in basic
Given that commercial H2 gas is most commonly produced
by steam reforming of natural gas, the use of renewable sources
of hydrogen using transfer hydrogenation (TH) strategies is
highly desirable form a sustainability perspective, as long as
high selectivity and activity can be achieved. TH strategies also
have process advantages relative to direct hydrogenation, such
as the use of milder reaction conditions, elimination of use of
flammable hydrogen, and ability to produce secondary
products of commercial value. The generation of such by-
products has the potential to improve the economics of the
overall process, provided the by-product is more valuable than
the hydrogen donor used, and is formed selectively.
Recently, we highlighted another key advantage of transfer
hydrogenation (TH) of CO2 vs direct hydrogenation: overcoming
the unfavourable thermodynamics of the latter.[22] The
thermodynamics of the TH process strongly depend on choice
of H-donor under basic conditions: isopropanol, for example,
lowers G by 9.7 kcal/mol vs direct hydrogenation of CO2, while
glycerol lowers it by 22.6 kcal/mol. The significant driving force
of the TH from glycerol is largely due to the conversion of the
initial dehydrogenation product of glycerol, dihydroxyacetone
(DHA), to lactate via dehydration and intramolecular Cannizzaro
reaction.[25, 26, 27] Consistent with this, the calculated Gaq for
TH of bicarbonate from glycerol is also significantly more
favourable than the direct hydrogenation of bicarbonate under
basic conditions (-83.9 vs 4.44 kcal/mol).
−
aqueous media CO2 is in equilibrium with HCO3 (pKa1 = 6.35), the
species undergoing hydrogenation could be either bicarbonate or
carbonate.[8] Explicit hydrogenation of bicarbonate has been shown
15, 16]
18, 19]
with Rh,[6,
Ru,[17,
Ir,[20] and Fe[21] catalysts. The
thermodynamic driving force of bicarbonate hydrogenation in
aqueous media is greater than that of CO2 (Gaq = 4.4 kcal/mol for
HCO3− vs 13.4 kcal/mol for CO2,[22] consistent with trends reported by
Papai et al[23]). Despite that, reported catalytic activities for
carbonates are generally lower than those for CO2 (maximum TOF
reported is 551 hr-1 for bicarbonate,[18] vs > 23,000 hr-1 for CO2).[24]
a. 800 22nd St NW, Suite 4000, Washington D.C. 20052
† Footnotes relating to the title and/or authors should appear here.
Electronic Supplementary Information (ESI) available: catalyst syntheses,
crystallographic data for 7, and additional data on catalytic activity. See
DOI: 10.1039/x0xx00000x
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 2 of 11
ARTICLE
Journal Name
Scheme 1. Catalysts reported for the transfer hydrogenation of
carbonates from glycerol.
Results and discussion
Catalyst synthesis.
View Article Online
DOI: 10.1039/D0GC01958E
Eight iridium(I) and iridium(III) N-heterocyclic carbene (NHC)
complexes, six of which bear sulfonate-functionalized wingtips (2 –
7, Scheme 2), were synthesized and characterized with the goals of
(i), assessing their activity for the TH of carbonate salts from glycerol
to afford lactate and formate salts; and (ii), identifying structural
features of the pre-catalyst, including presence of sulfonate
moieties, which are associated with higher catalytic activity and
selectivity. The sulfonated iridium compounds were compared to the
ruthenium sulfonated NHC complex previously reported by us (1)24,
as well as two non-sulfonated iridium NHC complexes (8 and 9).
While exact analogues of Ir(I) and Ir(III) pre-catalysts could not be
accessed synthetically in some cases (e.g. Ir(I) analogue of 2 proved
unstable in our hands), the series depicted in Scheme 2 provides
ample grounds for comparison of the effects of ligands and metal
oxidation state.
Iridium(III) compounds 2 and 3 were included based on their high
activity for transfer hydrogenation of aldehydes, imines, and ketones
from glycerol,[25, 34] which was first reported by Peris et al. Iridium(I)
compounds 4 – 7, on the other hand, were included in light of our
previous reports of their high activity for acceptorless
In addition to the above process advantages of using TH
from glycerol, the use of glycerol as H-donor provides a route to
valorization[28] of this abundant by-product of the biodiesel
industry[29] to lactic acid. Demand and price for the latter
continue to increase, driven by new applications in cosmetics,
polymer synthesis,[30] fine chemicals and food preservation.[31]
We recently reported the first homogeneous catalyst for the TH
of CO2 and carbonates from glycerol, consisting of water-
soluble Ru N-heterocyclic carbene (NHC) complex.[22] The
catalyst afforded 49 turnovers/h with 26 bar CO2 and 180 C,
but showed higher efficiency for carbonate (179 h-1 with 2 M
K2CO3 at 150 C, Scheme 1). Since then, Choudhury et al
reported an Ir-NHC (abnormal NHC) complex, achieving a TOF
of 90 h-1 at 150 C with K2CO3 and ambient CO2 pressure.[32]
Notably, the latter catalyst did not afford lactic acid as a by-
product, but rather DHA. Lin et al. also recently reported the TH
of carbonates from glycerol at 240 C with Pd/C.[33] affording a
TOF of 45 h-1 (Scheme 1). While the TH of carbonates has
potential applications on a large scale, it requires the
development of more efficient, robust catalysts and the
minimization of the use of precious metals.
Towards this goal, here we report our progress on the design of
highly prolific and robust homogeneous catalysts for the transfer
hydrogenation of carbonate salts from glycerol, affording lactate and
formate salts under basic conditions with efficiencies that compare
favourably to the precedents (Scheme 1). We also report mechanistic
insights on the process that will continue to inform the design of
prolific catalysts precursors for transfer hydrogenation reactions of
from renewable H-donors. We note that the catalysts reported here
are designed to serve as precursors to site-isolated heterogeneous
catalysts (SIHCs) for this reaction, which will minimize the use of the
precious metal in continuous processes.
[14, 35]
dehydrogenation of glycerol.
Comparisons between
compounds 4 and 5 elucidate the effect of the N-sulfonate spacer on
catalytic activity and stability, while comparisons between 4 and 6,
and 5 and 7 respectively speak to the effect of stronger -acceptor
CO ligands versus the weaker and more labile cyclooctadiene (cod)
ligands. Finally, compound 8 and 9 provide insight into the effect of
the sulfonate functionalization on catalyst activity.
Scheme 2. NHC catalysts examined for the transfer hydrogenation of
carbonate salts from glycerol.
2K
SO3
Na
SO3
SO3
SO3
N
I
Cl
N
N
N
Ru
N
N
N
O
O
N
SO3
Ir
I
N
Ir Cl
N
N
N
N
SO3
1 (Heltzel et al 2018)24
2
3
SO3
Na
-O3S
Na
O3S
N
Na
CO
N
N
Ir CO
Ir
NMe
MeN
Ir
SO3
NMe
MeN
3
N
N
-
SO3
Na
Me
MeN
SO3
N
N
4
5
6
SO3
-
-
PF6
PF6
Me
N
CO
CO
CO
N
N
Ir CO
Ir CO
Ir CO
N
SO3
NMe
MeN
NMe
MeN
Me
MeN
N
NMe
N
7
8
9
The molecular structure of 7 was determined by single crystal X-
ray diffraction (Figure 1) and compared to our previously reported
structure of the cod analogue 5. The crystal structure shows a
disordered square planar coordination with two phenyl wingtips in a
2 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 3 of 11
Green Chemistry
Please do not adjust margins
Journal Name
ARTICLE
staggered orientation. A closely coordinating sodium ion and several on the screening of the catalysts for TH of carbonatesVfireowmArtgicllyecOenrloinle,
co-crystalized water molecules can also be resolved. The torsion with no additional base. Products were identified and quantitated
DOI: 10.1039/D0GC01958E
angles between the CO-Ir plane and the carbene carbons (C(16) and with HPLC and 1H NMR using an internal standard, as described in the
C(6) respectively) are 65.865° (N4 C16 Ir1 C1) and 65.875° (N2 C6 Ir1 Experimental section and ESI.
C2) respectively. In comparison, the symmetric analogue with N-
methyl wingtips (9), reported by Crabtree et al,[35] has a larger
torsion angle than 7 (80.78°). The Ir-C (NHC) bond distances of 7
measure 2.093(8) and 2.088(8) Å, marginally longer than those of the
analogous cod complex by approximately 0.045 Å,[14] but still
Carbonate Salts.
Initial screen for TH of potassium carbonate (K2CO3) from glycerol
was performed with 2.7 M K2CO3, 1:1 water/glycerol and 0.001 mol%
(10 ppm) catalyst using microwave heating to 150 C for 6 hours. The
use of microwave heating allows for more efficient screening of the
catalysts and precise reaction temperature control. Carbonate was
selected over bicarbonate in order to maximize initial pH (12.1). At
comparable to those reported for analogous Ir-NHC compounds,
36]
such as 9.[35,
The Ir-CO bond lengths, C(1)-Ir(1) and C(2)-Ir(1),
measure 1.880(10), and 1.876(10) Å respectively, in the typical range
of Ir-CO bond lengths and comparable to those in non-sulfonated
compound 9.[35] The phenyl wingtips stagger in opposite directions
with torsion angles of 41.428° (C8 C7 N1 C5) and 45.526° (C18 C17
N3 C16), compared to those of the cod analogue at 60.40° and
45.02°.[14] The differences in torsion angle between the two ligands
can be attributed to the coordinating sodium ion and closely
associated water molecules that act as a bridge between two
molecular units (see ESI Figure S2 and S3 for more information).
this pH the theoretical ratios of HCO3 :CO32- at pH 12.1 is 2:98, which
-
is notable because bicarbonate is likely the only species that can
undergo transfer hydrogenation.15
Control reactions without catalyst afford no appreciable
conversion of glycerol or carbonate salts, while those without
glycerol afford no carbonate conversion. In the absence of carbonate
salts (or CO2), the catalytic reactions afford only lactate salts via
acceptorless dehydrogenation (AD) of glycerol, as reported
previously.[14] The dehydrogenation reaction does not proceed in the
absence of base. We had suggested that the base is also important
for suppressing the reverse reaction, formate dehydrogenation,
which we observed at pH < 4.[22] Given that many of these catalysts
are active for glycerol AD, it was not surprising to observed AD as a
competing process, resulting in the formation of excess lactate
compared to formate. The TH and AD processes have distinct
activation energies and thus temperature-dependences; as a result,
the ratio of lactate:formate varies with temperature and with choice
of catalyst. For example, at 150 C with catalyst 1 the concentrations
of lactate and formate are comparable, but above 150 C lactate
exceeds formate (ESI Figures S4a and Figure S4b).
All Ir catalysts except for 1,5-cyclooctadiene (cod) complexes 4
and 5 were more active for TH than our previously reported Ru (1)
catalyst, which afforded 2754 and 506 turnovers for lactic and formic
acid respectively in 6 hours (Figure 2). The low activity of 4 and 5 is
likely associated with the labile cod ligand, whose dissociation may
provide a route to decomposition products, such as iridium dimers
previously observed by Crabtree et al.[27] Interestingly, cod
complexes 4 and 5, which differ in N-substituents, also differ in TH
activity: while the N-propyl-sulfonate analogue (4) is inactive for
transfer hydrogenation, the N-phenylsulfonate analogue (5) shows
some (albeit low) TH activity (508 turnovers of formic acid, Figure 2).
Figure 1. ORTEP diagram of one of two crystallographically unique
catalyst molecules in 7, containing the iridium metal centre Ir1.
Lattice water molecules and sodium atoms are excluded for clarity.
See ESI Figure S1 and S2 for additional ORTEP illustrations. Select
bond lengths (Å) and angles (deg): C(1)-Ir(1) 1.880(10), C(2)-Ir(1)
1.876(10), C(6)-Ir(1) 2.093(8), C(16)-Ir(1) 2.088(8), C(6)-Ir(1)-C(16)
86.9(3)°, C(1)-Ir(1)-C(2) 92.5(4)°.
The activity of the iridium (III) sulfonated catalysts 2 and 3 was
comparable: ~ 5200 - 5300 turnovers for lactic acid and 890 - 1350
for formic acid, with 2 having higher activity than 3. We had
previously shown that adding an excess of NaOAc doubles the
activity of 3 for AD of glycerol, suggesting that slow initial
substitution of the chloride may be responsible for its inferior activity
relative to 2.[14] The relative activity of the Ir(I) catalysts (4 – 9) was
dependent on the ancillary ligand and the N-wingtips: as noted
earlier, the cod complexes were significantly less active than the
bis(carbonyl) analogues. This is consistent with both our prior report
on glycerol AD,[14] as well as a report by Crabtree et al demonstrating
that [(IMe)2Ir(CO)2]BF4 (IMe is 1,3-dimethylimidazole-2-ylidene) had
higher activity than corresponding cod complex for this process.[35]
While both sulfonated carbonyl derivatives 6 and 7 were more active
than their cod precursors (4 and 5, Figure 2), the relative
Glycerol Transfer Hydrogenation to Carbonate Salts
We next compared the activity of the catalysts in Scheme 2 for the
transfer hydrogenation for carbonate salts from glycerol. In a
previous report we discussed the optimization of reaction
conditions,[22] and suggested that catalytic activity for the gaseous
reactions with CO2 is often limited by mass transfer related to CO2 or
base solubility. This is further complicated by the fact that the
catalyst can also facilitate acceptorless dehydrogenation, as well as
pH-dependent formate dehydrogenation at low pH.[22] As a result,
reaction rates for transfer hydrogenation and product selectivity
have a strong dependence on temperature, pressure and base
solubility. Given that CO2, carbonate and bicarbonate are in
equilibrium in the presence of aqueous base, our main focus here is
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 3
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 4 of 11
ARTICLE
Journal Name
improvements were not equal. Compound 7, with N-phenylsulfonate suggesting this pathway is less likely than those via pyrVuievwalAdretichleydOenlioner
wingtips, showed ~ 7-fold higher activity for production of lactic acid
DOI: 10.1039/D0GC01958E
-hydroxyacetone (Scheme 4).
than cod analogue 5, and ~16-fold for formic acid, far exceeding the
The fact that no 1,3-PDO was observed suggests that glycerol
dehydration is selective. This observed selectivity is likely associated
improvement in activity of 6 over 4. We attribute this to the higher
thermal stability of the N-phenylsulfonate in 7 to degradation via
Hoffman elimination compared to the N-propylsulfonte in 6.[37, 38]
More thermally robust catalysts should have a significant advantage
in catalyst lifetime under microwave conditions, where superheating
can result in formation of hotspots.
with
the
more
thermodynamically
favourable
dehydrogenation/hydrogenation to the 1,2-PDO vs the 1,3-PDO, as
suggested by calculations of the Gibbs free energies of reaction
(Gaq 5.4 kcal/mol, ESI Table S3).
Consequently, the overall highest activity for transfer
We sought further insight into selectivity trends by comparing the
hydrogenation was observed with 7 (15400 and 8120 turnovers for catalysts in terms of the excess of acceptorless hydrogenation (AD)
lactic and formic acid respectively, compared to 2780 and 810 activity relative to transfer hydrogenation (TH) activity. As such, we
turnovers respectively for 6. The TOF of formic acid afforded by 7 can define two relative excess terms, as follows:
(1353 h-1) is ~ 16-fold higher than the activity of our previously
reported Ru catalyst (1) under these conditions (84 h-1) and the Ir-
excessAD/TH = [LA] – ([FA] + [1,2-PDO]) x 100
(1)
NHC catalyst by Choudhury et al (90 h-1).[32]
[FA] + [1,2-PDO]
and
Scheme 3.
excessAD/FA = [LA] – [FA] x 100
(2)
cat.
OH
O
O
K2CO3
H2O
150oC, 6h
[FA]
HO
OH
OH
H
OH
OH
OH
OH
LA
where [LA], [FA] and [1,2-PDO] refer to the concentrations of lactic
acid, formic acid and 1,2-PDO at a given time in the reaction.
Equation 1 quantitates the relative excess of AD over TH resulting in
both observed TH products (FA and 1,2-PDO), whereas equation 2
refers specifically to transfer hydrogenation to yield formic acid. The
two values are identical for all catalysts except 7, given that only 7
affords 1,2-PDO as a second TH product. The values for excessAD/TH
are lowest for 7 (26%), and highest for the Ir(III) complex 3 (503%),
while excessAD/FA for 7 is 89% (ESI Table S4). The latter analysis
suggests that catalyst 7 is not only the most efficient for AD of
glycerol, but also for transferring the H-equivalents to an acceptor.
However, the trends for TH and AD are not parallel for the catalyst
series, as summarised below:
FA
1,2-PDO
TOF (per hour)
1500
0
500
343
1000
2000
2500
9
8
7
6
5
4
3
2
1
153
107
630
1,353
2564
135
463
85
346
212
148
891
865
226
84
1,2-PDO
FA
LA
459
4000
TH activity (formic acid): 7 > 2 > 3 > 9 > 6 > 8 > 5 > 1 > 4
AD activity: 7 > 3 > 2 > 8 > 6 > 1 > 5 > 9 > 4
0
8000
TON
12000
16000
Figure 2. Activity of NHC catalysts 1 – 9 (Scheme 2) for transfer
hydrogenation of K2CO3 from glycerol, based on observed turnover
numbers (TON) and turnover frequencies (TOF) of for FA, LA and 1,2-
PDO (labels in grey refer to TOF per hour). Conditions: 1: 1 water:
glycerol (6.84 M), K2CO3 (2.7M), <0.001 mol% catalyst, 150 °C, 6 h.
Quantified by HPLC and NMR.
Scheme 4. Proposed route to lactic acid, formic acid and 1,2-PDO
from glycerol transfer hydrogenation of carbonates.
Catalyst 7 was also the only one that produced significant amounts
of 1,2-propanediol (1,2-PDO) as a side product, with 4080 turnovers,
resulting in a selectivity ratio of ~1:2:4 for 1,2-PDO:formic:lactic acid.
The formation of 1,2-PDO is attributed to transfer hydrogenation of
a dehydration products of glycerol, -hydroxyacetone (Scheme 4), or
the double hydrogenation of pyruvaldehyde. Neither of these could
be identified as an intermediate by HPLC or NMR, likely because
under reaction conditions both would undergo transfer
hydrogenation from isopropanol. In fact, pyruvaldehyde afford a
95:5 product mixture of lactate to 1,2-PDO using catalyst 7,
isopropanol and base. While this result does not provide direct
evidence for the fact that 1,2-PDO is formed via pyruvaldehyde, it is
consistent with the latter hypothesis. We also considered the
possibility that the 1,2-PDO is produced from lactic acid through
transfer hydrogenolysis. However, a comparable experiment with
lactic acid and isopropanol as hydrogen donor afforded no 1,2-PDO,
Sulfonate effect
We further examined the effect of the sulfonate group by
comparing the dehydrogenation activity of 7 to that of the non-
sulfonated analogue, 8, as well as 6 to 9 (Scheme 2). Compound 8
affords 3780 and 644 turnovers of lactic and formic acid respectively,
which is approximately a 4-fold decrease in lactic and 12-fold
decrease in formic acid compared to 7. Catalyst 8 also has a much
4 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 5 of 11
Green Chemistry
Please do not adjust margins
Journal Name
ARTICLE
higher excessAD/FA (487%, see eq 1 and Table S4). Interestingly, the water: glycerol (6.84M), K2CO3 (specified concentratioVnie),w<A0rt.i0cl0e1Online
activity of 6 vs 9 was much more comparable (2780 vs 2060 turnovers mol% 7, 150 °C, 6 h.
DOI: 10.1039/D0GC01958E
of lactic acid, and 810 vs 918 of formic acid for 6 and 9 respectively).
Carbonate salt effect
Sulfonated and non-sulfonated Iridium(III) NHC catalysts have shown
similar trends for the acceptorless dehydrogenation of glycerol. We
had previously shown that the non-sulfonated analogue of 3,
possessing n-propyl wingtips, was ~50% less active in acceptorless
dehydrogenation of glycerol relative to 3.26 We suggest that the
effect of the sulfonate group is driven by solubility differences in
aqueous media: thus, in the case of 7 vs 8, the increase in aqueous
solubility that results from addition of the sulfonate groups is
expected to be more significant than that of in 6 vs 9 due to the more
lipophilic nature of the phenyl vs propyl groups, as indicated by the
octanol-water partition coefficients for the propyl vs phenyl
analogues (-5.57 and -3.81 respectively, calculated by ChemAxon
Marvin Suite). The lower aqueous solubility is likely to favour
formation of inactive CO-bridging dimers, as observed by Crabtree et
al for an NHC catalyst.[27]
The initial results with potassium carbonate led us to consider
the potential effect of changing the carbonate cation. To that end,
carbonate salts of the alkali earth metal series (Li, Na, K and Cs) were
tested using catalyst 7 under the microwave conditions described
above (150 C, 6 hours). However, we observed significant
differences in solubility of these salts in the reaction medium. These
differences were justified by measuring the saturation
concentrations of the carbonates in the reaction medium (Li2CO3:
0.33 M, Na2CO3: 1.3 M, K2CO3: 3.6 M and Cs2CO3: 5.3 M). Thus, not
surprisingly, the TONs obtained associated with the
dehydrogenation product (lactic acid) and the sum of hydrogenation
products (formic acid and 1,2-PDO) were found to increase as we
change the salt cation from Li to Cs (Figure 4 and Table S6): for
example, relative to the TONs afforded by Li2CO3 (2530 and 146 for
lactic and total transfer hydrogenation products respectively),
Cs2CO3 affords 27 850 and 27 840 turnovers respectively. Glycerol
conversion correspondingly increases – for example, 2% for Li2CO3 vs
26% for Cs2CO3 (ESI Table S6). Given that only the Cs2CO3 and K2CO3
reactions were fully homogeneous, we can only compare their
kinetics. Given that the concentration of salt used was below
saturation for both K2CO3 and Cs2CO3, differences in reaction rates
between these two are not accounted for by solubility of the
carbonate salt. Interestingly, the relative acceleration of TH is higher
than that of dehydrogenation (e.g. moving down the group from K to
Cs increases lactic acid TONs by 180%, and those associated with the
dehydrogenation products (formic acid and 1,2-PDO) by 228%. As a
result, the relative excess of AD to TH (excessAD/TH) drops from 1636%
for Li to ~0% for Cs, as we change cations down the alkali metal series
(Figure 3). The latter implicates the cation in participating in a rate-
determining step in the catalytic cycle associated with
hydrogenation, consistent with prior reports of cation effects on CO2
and carbonate hydrogenation.[16, 19, 24]
Carbonate concentration.
We briefly explored the effect of carbonate concentration to the
catalytic activity of 7. While maintaining a glycerol concentration of
6.84 M in a 1:1 glycerol:water mixture, K2CO3 concentration was
varied from 0 to 2.7 M, at which point saturation was reached. As
carbonate concentration is increased, rates of formation of lactic and
formic acid increase correspondingly (Figure 3). In addition, we
observe a decrease in the relative excess of TH vs AD (excessAD/TH),
from 117% at 0.5 M K2CO3 to 26% at 2.7 M. The increase in overall
reaction rate is consistent with the involvement of hydroxide in the
last step of the Cannizzaro reaction leading to lactic acid formation
(Scheme 4). The lactic acid formation also provides additional
thermodynamic driving force for the overall process.24 The increase
in TH rate is also consistent with the expected effect of increasing
substrate (i.e. carbonate) concentration. The rate of formation of the
secondary TH product, 1,2-PDO, also increases at higher carbonate
concentration but by significantly less than the that of formate
production. The latter is not surprising, since higher carbonate
concentration results in higher abundance of carbonate as a
competing hydrogen acceptor. The small increase in 1,2-PDO could
also reflect the need for base to facilitate glycerol dehydrogenation,
the first step in 1,2-PDO formation (Scheme 4).
In addition, we observe that the rates of formation of the two
transfer hydrogenation products (formic acid and 1,2-PDO) are not
accelerated equally: the increase in 1,2-PDO TONs is approximately
double that of FA TONs from K to Cs (Li2CO3 affords no 1,2-PDO). This
results in a decrease in the selectivity for formic acid vs 1,2-PDO from
to 48% for Cs. The greater acceleration associated with 1,2-PDO
relative to FA implicates the cation in both the dehydration of
glycerol to -hydroxyacetone and the TH of the latter to 1,2-PDO
(Scheme 4).
LA
FA
1,2-PDO
Rel. excess(AD/TH)
16,000
12,000
8,000
4,000
0
140
120
100
80
60
40
20
0
We also considered the fact that the reaction rate would be
affected by the difference in aqueous solubility of the salts (ESI Table
S5) with increase in the effective ionic radius of the cation. As
previously noted, Li2CO3 and Na2CO3 are not fully dissolved at the
beginning of the reaction. Furthermore, the stability of the product
lactate and formate salts are likely to be more favourable with more
polarizable cations, such as Cs+. While the latter arguments could
account for the increase in formic and lactic acid, they do not readily
account for the increase in rate of 1,2-PDO formation. As observed
previously, increasing K2CO3 concentration had only a limited effect
on the rate of 1,2-PDO formation (Figure 3). In contrast, changing the
cation down the group results in an increase of 1,2-PDO double that
of the corresponding increase in TONs of FA (Figure 4).
0
0.5
1
1.5
2
2.5
3
[K2CO3], M
Figure 3. TON of FA, LA and 1,2-PDO observed in reactions at
different concentration of K2CO3 using catalyst 7. Conditions: 1: 1
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 5
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 6 of 11
ARTICLE
Journal Name
To probe whether Cs+ can be used in catalytic quantities to obtain and formate production (~50% reduction in both), and no 1,2-PDO
View Article Online
similar effects, we performed an experiment with a 9:1 ratio of production. The suppression of 1,2-PDO DfOoIr:m10a.t1i0o3n9/Du0nGdCer019n5e8aEt
K2CO3:Cs2CO3, which affords activity comparable to that obtained condition suggests a strategy that can be used to increase selectivity
with Cs2CO3 alone (22 200 and 12 660 turnovers of lactic and formic for formic acid, especially in the reaction with Cs2CO3.
respectively, compared to 27 850 and 13 350 – entries 6 and 3, ESI
Mechanistically, the effect of neat conditions on suppressing 1,2-
PDO production could be due to the fact that in the absence of water,
carbonate cannot hydrolyse to form bicarbonate and hydroxide:
Table S6). The addition of 10% Cs2CO3 increases lactic and formic acid
production ~ 1.5-fold compared to an experiment with equivalent
concentration of K2CO3 (Figure 4, right panel), and only 4% relative
excess of AD over TH. Glycerol conversion follows suit -– for example,
26% for Cs2CO3 vs 21% for 9:1 K2CO3:Cs2CO3.These results further
suggest that Cs+ can be used in catalytic quantities to accelerate lactic
acid and formic acid formation, but effect on 1,2-PDO formation is
more concentration-dependent. We also examined whether the
presence of harder Lewis acids, such as Li+, suppresses activity via
inhibitory effects. A reaction with 9:1 ratio of Li2CO3:Cs2CO3 resulted
in only a ~ 10% reduction in lactate and formate compared to the
control with K2CO3 alone (Figure 4, right panel), suggesting that the
main driver of the low TONs obtained with Li2CO3 is low solubility.
Consistent with this conjecture, we recover ~ 90% of the original
mass of Li2CO3 at the end of the reaction as undissolved carbonate.
CO32- + H2O  HCO3- + HO-
The Bronsted acidic bicarbonate may be necessary to facilitate the
dehydration of glycerol to -hydroxyacetone, the precursor to 1,2-
PDO. Although there is also a dehydration step in the generation of
lactic acid (glycerealdehyde to pyruvaldehyde, Scheme 4), that step
is more thermodynamically favourable due to formation of a
conjugated product, and thus be may be facilitated by weak Lewis
acids. If hydrolysis to bicarbonate is indeed suppressed in neat
glycerol, we must reconsider whether bicarbonate is indeed the true
substrate for TH to formate, or whether carbonate can bind to the
catalyst and be protonated from glycerol.
Thus, we conclude that the stark effect observed in reaction rate
by changing the alkali earth metal ion of the carbonate salts is a
combination of increasing solubility with increasing polarizability of
the cation, as well as specific roles in the facilitating a rate-
determining step in the catalytic cycle associated with TH and the
dehydration of glycerol to -hydroxyacetone, the H-acceptor
precursor to 1,2-PDO.
Figure 5. Comparison of TONs of formic acid (FA), lactic acid (LA) and
1,2-propanediol (1,2-PDO) obtained using aqueous glycerol (1:1) and
neat glycerol conditions using catalyst 7. [Conditions: Neat glycerol
or 1:1 water:glycerol, as indicated; <0.001 mol% 7, carbonate salt
(2.7 M) 150 C, 6 h, microwave.]
TH of carbonates using conventional heating.
Figure 4. Effect of carbonate source on the TONs of formic acid (FA),
lactic acid (LA) and 1,2-propanediol (1,2-PDO) using catalyst 7. Raw
data presented in ESI Table S4. [Conditions: 1: 1 water: glycerol (6.84
M), [M2CO3]: 0.03 M for Li2CO3, 1.3 M for Na2CO3 and 2.7 M for K2CO3
and Cs2CO3, <0.001 mol% 7, 150 C, 6 h, microwave.]
We further explored the activity of catalyst 7 under conventional
heating conditions at the same temperature as the microwave
reactions. In order to keep all components in the liquid phase the
reactions were performed in a Parr reactor under N2 pressure,
keeping all other conditions consistent with the microwave reactions
and extended the reaction time (1:1 glycerol/water, 2.7 M K2CO3 at
150 C). Under conventional heating and 26 bar N2 the TONs of LA
and FA observed after 6 hours are approximately double compared
to those obtained in the microwave reaction (27 760 vs 15 380 for LA
and 16 630 vs 8120 for FA, 8% conversion of glycerol (Figure 6 and
ESI Table S5). The glycerol conversion closely matches the sum of
lactic acid and 1,2-PDO yields, suggesting high selectivity for these
two products. This is accompanied by a higher selectivity for FA over
1,2-PDO under the conventional conditions (~3:1 vs 2:1). In 24 hours,
the reaction reaches 72 250 turnovers of LA and 52 030 of FA (TOF of
2170 h-1). The 1,2-PDO concentration reaches a steady state after the
first 6 hours of reaction, while LA and FA concentrations steadily rise
(Figure 6). As a result, the selectivity for FA vs 1,2-PDO increases over
the course of the reaction, reaching 23:1 after 24 hours (Figure 6a).
The relative excess of AD over the sum of the two TH products, which
Solvent effect.
We also investigated the effect of water on catalyst activity and
selectivity. The choice of solvent for initial screening was 1:1
water:glycerol, selected so as increase solubility of carbonate salts
and reduce viscosity of neat glycerol. Although these effects should
benefit reaction rate, the mechanism in Scheme 4 involves a number
of dehydration steps, which could make the effect of water on the
system more complex. Reactions performed in neat glycerol all
showed small decrease in amounts of lactate produced relative to
the 1:1 water:glycerol reactions with the same salt (12% and 30% for
Na2CO3 and K2CO3 respectively) and a small increase in formate
produced (Figure 5, 33% for Na2CO3). Notably, however, neither neat
reaction produced any 1,2-PDO. In the case of the neat reaction with
Cs2CO3 we observed more significant decreases in both the lactate
6 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 7 of 11
Green Chemistry
Please do not adjust margins
Journal Name
ARTICLE
should be minimized in order to optimize catalyst selectivity, steadily
Thus, we propose the mechanism shown in Figure 7. In order to
View Article Online
decreases over the course of the reaction from 286% to 33%, tracking associate the H-donor (glycerol), the 16-eDleOcIt:r1o0n.10Ir3(9I)/Ds0pGecCi0e1s9c5a8nE
the decrease in pH from 12.2 to 11.4 (ESI Figure S5).
undergo O-H oxidative addition of the primary or secondary hydroxyl
to form an alkoxy hydride complex, which has precedents for Ir(I)
complexes.[44] If the mechanism involves ß-hydride elimination, the
resulting saturated Ir(III) alkoxy hydride complex (previously
observed by NMR)[14] must open up a site.[45] The latter could more
readily dissociate a carbonyl compared to the Ir(I) species, given that
expectedly lower M-CO bond dissociation energy due to decreased
backbonding from the more oxidized iridium. To seek evidence for
CO dissociation we performed LC-MS at various times during the
reaction to identify prevalent iridium fragments. We were able to
identify a fragment consistent with mass of [(NHC)2Ir(CO)(DHA)H2]-
(DHA = dihydroxyacetone), suggesting CO loss upon glycerol
coordination and ß-hydride elimination. Thus, the CO dissociation
would allow the alkoxy glycerol complex to have an open site
necessary for ß-hydride elimination, forming DHA or glyceraldehyde.
The identity of the intermediate has not been determined here as we
have not been able to spectroscopically observe either in sufficient
concentration. This is likely because both DHA and glyceraldehyde
rapidly react under reaction conditions to form lactic acid. However,
we postulate the intermediate is more likely to be DHA, as the
glycerol dehydrogenation at the secondary alcohol is
thermodynamically more favourable than that of primary by (-5.11
kcal/mol vs 0.17 kcal/mol respectively). Furthermore, even if
glyceraldehyde is the kinetic product, rapid isomerization to DHA will
be driven by the favourable thermodynamics.
(a)
(b)
75,000
60,000
45,000
30,000
15,000
0
20,000
15,000
10,000
5,000
0
LA
FA
1,2-PDO
LA
FA
1,2-PDO
0
4
8
12 16 20 24
0
4
8
12 16 20 24
Time, h
Time, h
Figure 6. Time course for production of formate, lactate and 1,2-
PDO from the reaction of glycerol using catalyst 7, 6.85 M aqueous
glycerol, 2.70 M K2CO3, 0.001 mol% 7, 150 °C, N2 pressure (a) 26 bar
and (b) 50 bar.
We hypothesized that the reason for the increase in activity under
the latter conventional conditions vs microwave conditions was
related to the applied N2 pressure. The effect of the positive pressure
is complex, especially at elevated temperatures. Initially we
postulated that the higher pressure may result in decreased
solubility of the carbonate salt; however, this effect is likely to be
small for a pressure difference of 26 to 50 bar relative to the internal
pressure of liquids.[39] Potential evidence for this was found in pH
measurements over the course of both reactions, which show that
the 50-bar reaction has pH values 0.6 – 0.8 units lower than that of
the 26-bar reaction (ESI Figure S5). While the pH drop could also be
due to formation of acid products, the higher reaction rate in the 26-
bar reaction contradicts that. In sum, we conclude that the TH to
afford FA proceeds more efficiently and selectively in a pressure
reactor, where the effect of external pressure can be optimized to
increase availability of carbonate without compromising solubility of
the base.
Either product can be displaced by bicarbonate via ligand
substitution. Unfortunately, we have not been able to observe either
(DHA nor glyceraldehyde) under basic conditions, likely because both
quickly react to form lactate via dehydration and intramolecular
Cannizzaro reaction.[35]
Coordinated bicarbonate undergoes hydroxide elimination; such
a step has been previously proposed by DFT calculations for the Ru-
catalysed hydrogenation of bicarbonate.[46] The resulting H-Ir-CO2
47]
complex undergoes insertion[46,
to generate iridium formate,
Mechanism of Ir-catalysed TH
possibly trapped by re-coordination of CO (Figure 7). It is also
possible though that CO re-coordination may not be feasible at the
very low catalyst concentrations (and thus CO concentrations) used
here. Thus, an alternative to the cycle is possible, where CO does not
re-coordinate. In the latter scenario after initial CO loss, subsequent
cycles would not involve CO dissociation. Finally, reductive
elimination releases the formic acid, which is rapidly deprotonated
under basic conditions, and reforms the Ir(I) species.
The identification of the Ir(I) compound 7 as the most active
catalyst for both TH and AD led us to consider possible mechanisms
for both reactions and how they may differ from established
mechanisms for the more extensively studied Ir(III) catalysts. For
Ir(III) and Ir(I) catalysts with open coordination sites, TH typically
proceeds via ligand substitution by an alkoxide, followed by ß-
hydride elimination to form a hydride and insertion of the H-acceptor
into the hydride. Release of H2 has been proposed to proceed via
hydride abstraction using H+.[40] Such a mechanism cannot be applied
to 7, as the 16-electron square planar d8 complex lacks open
coordination sites. Furthermore, dissociation of CO is likely to be
prohibitive due to the high calculated and experimentally-
determined BDEs for similar Iridium species.[41, 42] Dissociation of the
NHCs is also expected to be prohibitive, with typical BDEs > 60
kcal/mol.[43] However, in search of experimental evidence that for
potential ligand dissociation, we performed a reaction with excess
imidazolium salt, which under the strongly basic conditions would be
expected to facilitate NHC formation. This would lead to decreased
reaction rate if M-NHC dissociation were involved in catalyst
activation. However, no significant difference was observed in the
rate from the control experiment, suggesting that the catalytic cycle
does not involve NHC dissociation.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 7
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 8 of 11
ARTICLE
Journal Name
NMR spectra were recorded on an Agilent NMR VsiepweAcrttricolemOentliener
operating at 400 MHz. GC-FID analyses wereDpOerI:fo10rm.10e3d9o/Dn0aGnCA0g19il5e8nEt
6890N GC System with Agilent 7693 autosampler. GC-MS analyses
were performed on a Shimadzu QP2010S GC-MS, while LC-MS
analyses were performed on a Shimadzu LC-MS-105 single quad with
a 2020 detector. Compound identity and purity were established
1
based on H and 13C NMR data, HPLC, GC-FID and GC-MS. Air-free
manipulations were performed using Schlenk technique or in an
MBraun Pure Lab HE Inert glove box system.
Catalyst synthesis
Compounds 1, 2, 3 and 5 were prepared by silver transmetallation
from the corresponding imidazolium salts, while 4 was synthesized
by direct metalation using [Ir(cod)OMe]2.[25] Compounds 1-7 and 9
were synthesized as previously described protocols by us and
others,[49] while 8, a new compound, was prepared from the
[Ir(cod)OEt]2 dimer using a method analogous to that reported by
Hermann and Jimenez[14, 50] For bis-NHC Ir(I) complexes 4 – 8, we
found that the addition of the imidazolium salt to the iridium dimer
affords cleaner formation of the target compounds than the more
common Ag transmetalation method. All catalyst syntheses are
described in the ESI.
Figure 7. Proposed catalytic cycle for transfer hydrogenation of
bicarbonate with glycerol by catalyst 7. Note that it has not been
determined definitely whether the intermediate dehydrogenation
product of glycerol is DHA or glyceraldehyde.
General procedure for transfer hydrogenation of carbonate salts
from glycerol
Microwave reactions: In a typical procedure, a 10-mL microwave vial
was charged with a stir bar, 2 mL glycerol, 2 mL H2O, K2CO3 (2.7 M)
and catalyst (200 µL of 0.63 mM aqueous solution). The reaction
mixture was heated to 150 °C for 6 hours in an Anton Paar
Monowave microwave with a stir rate of 600 rpm. The products were
quantified using HPLC and NMR.
The latter analysis suggests that catalyst 7 is not only the most
efficient for AD of glycerol, but also for transferring the H-
equivalents to an acceptor.
The lower activity of [(NHC-Ph-SO3)2Ir(cod)]- (5) vs [(NHC-Ph-
SO3)2Ir(CO)2]- (7) for both acceptorless dehydrogenation, but
especially for transfer hydrogenation (see Figure 2) could be
attributed to several factors: (i) more favourable reductive
elimination of formic acid from Ir center in 7 with stronger -
acceptor (CO) ligands; (ii) more favourable insertion of CO2 into Ir-H
with more electron-withdrawing CO ligands, or (iii) the fact that 5
undergoes an entirely different mechanism than 7 via displacement
of the cycooctadiene ligand. Although the presence of carbonyl
ligands in 7 also opens the possibility for cluster formation[37, 41] this
is not as likely at the low catalyst loadings (10 ppm) used compared
to prior reports. [35, 48] It is also possible that cluster formation is more
favoured for catalysts that have low solubility in the reaction
medium.
Batch reactions: All reactions were carried out in a high temperature-
pressure autoclave (Parr®, 4564 series) fitted with a glass insert,
standard mechanical agitator, and liquid sampling tube. The glass
insert was loaded with catalyst, 25 mL of aqueous K2CO3 (of desired
concentration) and 25 mL of glycerol. The glass insert was then
placed into the autoclave. The autoclave was sealed, and the stirrer
turned on and set to 50% power. The autoclave was purged 3 times
with N2 (Roberts Oxygen, industrial grade) and pressurized to 10 bar.
When the reaction reached the desired temperature, the pressure
was adjusted to the desired operating conditions, typically 26 Bar.
Product Characterization
Reaction aliquots were analysed by HPLC and NMR. HPLC was
performed using a Shimadzu Prominence-I (LC-2030C 3D) instrument
equipped with a PDA and RI detectors using a mobile phase of 0.005
M H2SO4 with a flow rate of 0.44 mL/min at 35 °C. Samples for
quantification of lactic acid (LA), formic acid (FA), 1,2- propanediol
(1,2-PDO) and glycerol were prepared by adding a 1-mL aliquot of
sample to 0.22 mL of 5 M H2SO4 and filtering with a syringe filter. The
PDA detector scanned the range of 190 - 800 nm, affording traces at
190, 218, 254, and 284 nm for analysis. The 190-nm wavelength trace
included glycerol and all the desired products, while the 218-nm
trace excluded glycerol and 1,2-PDO. Glycerol was quantitated using
the Refractive Index (RI) detector. Typical HPLC trace and PDA
chromatograms are shown in Figure S6. The retention times for LA,
glycerol, FA and 1,2 PDO are 28.7, 29.5, 30.5, and 36 min respectively.
Experimental
General considerations and instrumentation
The syntheses of the catalysts were carried out under nitrogen using
standard Schlenk technique. Commercial chemicals were used
without further purification. [RuCl2(p-cymene)]2 and IrCl3 were
purchased from Acros-Organics. Solvents were dried using a solvent
purification system (SPS MBraun) or 4 Å molecular sieves. Glycerol
(>99%, Alfa Aesar) was dried over activated 4 Å molecular sieves.
KOH (purity 85%) was obtained from VWR BDH, while NaOH (purity
97%) was obtained from VWR Life Science. Potassium formate (99%)
was obtained from Alfa Aesar. Isopropanol (HPLC grade) was
obtained from Fischer Scientific.
8 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 9 of 11
Green Chemistry
Please do not adjust margins
Journal Name
ARTICLE
NMR was used to confirm HPLC yields and identity of products. For [1]
NMR analysis, a 0.100-mL aliquot of reaction solution was mixed with
equal volume of standard solution of 3-(trimethylsilyl)propionic- [2]
2,2,3,3-d4 acid sodium salt (TSP) and 0.5 mL D2O. The only products
identified by NMR were glycerol, LA, FA, 1,2-PDO and pyruvaldehyde
(in minute amounts). A comparison was made with reactions in [3]
which the TSP standard was added at the beginning of reaction,
rather than to each aliquot, and no major differences were observed. [4]
M. G. Mura, L. D. Luca, G. Giacomelli, A. Porcheddu, Adv.
View Article Online
Synth. Catal. 2012, 354, 3180-3186.
DOI: 10.1039/D0GC01958E
J. Albert, A. Jess, C. Kern, F. Pöhlmann, K. Glowienka, P.
Wasserscheid, ACS Sustain. Chem. Eng. 2016, 4, 5078-
5086.
K. Sordakis, C. Tang, L. K. Vogt, H. Junge, P. J. Dyson, M.
Beller, G. Laurenczy, Chem. Rev. 2018, 118, 372-433.
P. Haynes, L. H. Slaugh, J. F. Kohnle, Tett. Lett. 1970, 11,
365-368.
[5]
P. G. Jessop, T. Ikariya, R. Noyori, Chem. Rev. 1995, 95, 259-
272; C. M. Jens, M. Scott, B. Liebergesell, C. G. Westhues,
P. Schäfer, G. Franciò, K. Leonhard, W. Leitner, A. Bardow,
Adv. Synth. Catal. 2019, 361, 307-316.
Conclusion
[6]
[7]
I. Jószai, F. Joó, J. Mol. Cat. A 2004, 224, 87-91.
Here we report the design of highly prolific and robust homogeneous
catalysts for the transfer hydrogenation of carbonate salts from
P. Munshi, A. D. Main, J. C. Linehan, C.-C. C. Tai, P. G.
Jessop, J. Am. Chem. Soc. 2002, 124, 7963-7971; C. A. Huff,
M. S. Sanford, ACS Catal. 2013, 3, 2412-2416; G. A.
Filonenko, M. P. Conley, C. Copéret, M. Lutz, E. J. M.
Hensen, E. A. Pidko, ACS Catal. 2013, 3, 2522-2526; G. H.
Gunasekar, K.-D. Jung, S. Yoon, Inorg. Chem. 2019, 58,
3717-3723; W. C. SHEN Zheng, LIU Shiyang，ZHANG Wei,
HUANG Xin, ZHANG Yalei, ACTA PETROL. SIN. 2018, 34,
897-903.
glycerol, affording lactate and formate salts under basic conditions
with efficiencies that exceed those of the few literature precedents.
The most active catalyst identified is a water-soluble Ir(I) N-
heterocyclic carbene complex (catalyst 7) whose activity was
examined at low catalyst loading (10 ppm) at 150 °C under
microwave and conventional heating in 1:1 water:glycerol. The
microwave reaction with 7 affords lactic acid, formic and some
smaller quantity of a dehydration-transfer hydrogenation product, 1-
[8]
K. Rohmann, J. Kothe, M. W. Haenel, U. Englert, M.
Hölscher, W. Leitner, Angew. Chem. Int. Edit. 2016, 55,
8966-8969.
2-propanediol (1,2-PDO). The cation of the carbonate salt impacts
catalytic activity and selectivity significantly, with highest activity
observed with Cs2CO3 (27850 and 13350 TONs for lactate and
[9]
J. F. Hull, Y. Himeda, W. H. Wang, B. Hashiguchi, R. Periana,
D. J. Szalda, J. T. Muckerman, E. Fujita, Nat. Chem. 2012, 4,
383-388; N. Onishi, S. Xu, Y. Manaka, Y. Suna, W.-H. Wang,
J. T. Muckerman, E. Fujita, Y. Himeda, Inorg. Chem. 2015,
54, 5114-5123; R. Tanaka, M. Yamashita, K. Nozaki, J. Am.
Chem. Soc. 2009, 131, 14168-+; S. Takaoka, A. Eizawa, S.
Kusumoto, K. Nakajima, Y. Nishibayashi, K. Nozaki,
Organometallics 2018, 37, 3001-3009.
A. Boddien, D. Mellmann, F. Gärtner, R. Jackstell, H. Junge,
P. J. Dyson, G. Laurenczy, R. Ludwig, M. Beller, Science
2011, 333, 1733-1736; Y. Zhang, A. D. MacIntosh, J. L.
Wong, E. A. Bielinski, P. G. Williard, B. Q. Mercado, N.
Hazari, W. H. Bernskoetter, Chem. Sci. 2015, 6, 4291-4299;
C. Federsel, A. Boddien, R. Jackstell, R. Jennerjahn, P. J.
Dyson, R. Scopelliti, G. Laurenczy, M. Beller, Angew. Chem.
Int. Ed. 2010, 49, 9777-9780.
S. Muthuramalingam, M. Sankaralingam, M. Velusamy, R.
Mayilmurugan, Inorg. Chem. 2019, 58, 12975-12985.
Y. M. Badiei, W. H. Wang, J. F. Hull, D. J. Szalda, J. T.
Muckerman, Y. Himeda, E. Fujita, Inorg. Chem. 2013, 52,
12576-12586; S. A. Burgess, K. Grubel, A. M. Appel, E. S.
Wiedner, J. C. Linehan, Inorg. Chem. 2017, 56, 8580-8589.
S. Ogo, R. Kabe, H. Hayashi, R. Harada, S. Fukuzumi, Dalton
T. 2006, 4657-4663.
formate respectively in 6 hours, compared to 15400 and 8120 with
K2CO3, but lower selectivity for formic acid over 1,2-PDO. However,
catalytic amounts of Cs+ significantly enhance activity with K2CO3 to
levels comparable to those obtained with Cs2CO3. The reaction
affords high selectivity for FA more efficiently in a pressure reactor
with conventional heating, where the external pressure can be
manipulated to optimize carbonate concentration and solubility.
High TONs for LA and FA (72,245 for LA and 52,032 for FA) were
obtained with 26 bar N2 at 150 C in 24 hours. Under neat
conditions we observed an increase in the selectivity for FA over 1,2-
PDO, especially in the reaction with Cs2CO3. A mechanism is
proposed for TH of carbonate from glycerol using 7, which involves
O-H oxidative addition of glycerol, -hydride elimination,
bicarbonate dehydroxylation, CO2 insertion and reductive
elimination of formic acid. In sum, we report a highly efficient
catalytic reaction that uses carbonate salts, in place of carbon
dioxide, to generate formic acid via transfer hydrogenation
processes from renewable H-donors such as glycerol.
[10]
[11]
[12]
Conflicts of interest
No conflicts of interest are reported.
[13]
[14]
[15]
M. Finn, J. A. Ridenour, J. Heltzel, C. Cahill, A. Voutchkova-
Kostal, Organometallics 2018, 37, 1400-1409.
Acknowledgements
F. Joó, G. Laurenczy, P. Karády, J. Elek, L. Nádasdi, R. Roulet,
Appl. Orgmet. Chem. 2000, 14, 857-859; D. Jantke, L.
Pardatscher, M. Drees, M. Cokoja, W. A. Herrmann, F. E.
Kuhn, ChemSuschem 2016, 9, 2849-2854; W. Gan, D. J. M.
Snelders, P. J. Dyson, G. Laurenczy, ChemCatChem 2013, 5,
1126-1132.
The authors sincerely thank NSF for funding through NSF CAREER
award 1554963. We thank the Prof. Cahill (GWU Chemistry) and his
group for access to X-ray crystallography and assistance in data
analysis.
[16]
[17]
G. Zhao, F. Joó, Cat. Commun. 2011, 14, 74-76.
F. Joó, L. Nádasdi, J. Elek, G. Laurenczy, Chem. Commun.
1999, 971-972; H. Horváth, G. Laurenczy, Á. Kathó, J.
Orgmet. Chem. 2004, 689, 1036-1045; S. S. Bosquain, A.
Notes and references
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 9
Please do not adjust margins

Please do not adjust margins
Green Chemistry
Page 10 of 11
ARTICLE
Journal Name
Dorcier, P. J. Dyson, M. Erlandsson, L. Gonsalvi, G. [38]
Laurenczy, M. Peruzzini, Appl. Orgmet. Chem. 2007, 21,
947-951; G. Laurenczy, F. Joó, L. Nádasdi, Inorg. Chem.
2000, 39, 5083-5088; F. Joõ, ChemCatChem 2014, 6, 3306-
3308; K. Sordakis, A. Guerriero, H. Bricout, M. Peruzzini, P. [39]
J. Dyson, E. Monflier, F. Hapiot, L. Gonsalvi, G. Laurenczy, [40]
Inorg. Chim. Acta 2015, 431, 132-138.
A. W. Salman, R. A. Haque, S. Budagumpi, H. Zetty Zulikha,
View Article Online
Polyhedron 2013, 49, 200-206; S. Budagumpi, R. A. Haque,
DOI: 10.1039/D0GC01958E
A. W. Salman, M. Z. Ghdhayeb, Inorg. Chim. Acta 2012,
392, 61-72.
Y. Marcus, Chem. Rev. 2013, 113, 6536-6551.
S. Gülcemalꢀ, D. Gülcemalꢀ, G. F. S. Whitehead, J. Xiao,
Chem. Eur. J. 2016, 22, 10513-10522.
[18]
A. Boddien, F. Gärtner, C. Federsel, P. Sponholz, D. [41]
Mellmann, R. Jackstell, H. Junge, M. Beller, Angew. Chem.
S. J. Zhang, S. D. Foyle, A. Okrut, A. Solovyov, A. Katz, B. C.
Gates, D. A. Dixon, J. Phys. Chem. A 2017, 121, 5029-5044.
G. P. Rosini, F. Liu, K. Krogh-Jespersen, A. S. Goldman, C. Li,
S. P. Nolan, J. Amer. Chem. Soc. 1998, 120, 9256-9266.
A. T. Termaten, M. Schakel, A. W. Ehlers, M. Lutz, A. L. Spek,
K. Lammertsma, Chem. Eur. J. 2003, 9, 3577-3582.
F. T. Ladipo, M. Kooti, J. S. Merola, Inorg. Chem. 1993, 32,
1681-1688; O. Blum, D. Milstein, J. Amer. Chem. Soc. 2002,
124, 11456-11467; D. Morales-Morales, R. Redón, Z. Wang,
D. W. Lee, C. Yung, K. Magnuson, C. M. Jensen, Can. J.
Chem. 2001, 79, 823-829.
V. Guiral, F. Delbecq, P. Sautet, Organometallics 2000, 19,
1589-1598; S. M. Ng, C. Zhao, Z. Lin, J. Organomet. Chem.
2002, 662, 120-129; H. Itagaki, N. Koga, K. Morokuma, Y.
Saito, Organometallics 1993, 12, 1648-1654; M.
Yamakawa, H. Ito, R. Noyori, J. Amer. Chem. Soc. 2000, 122,
1466-1478.
Int. Edit. 2011, 50, 6411-6414.
J. L. Drake, C. M. Manna, J. A. Byers, Organometallics 2013,
32, 6891-6894.
[42]
[43]
[19]
[20]
M. Erlandsson, V. R. Landaeta, L. Gonsalvi, M. Peruzzini, A.
D. Phillips, P. J. Dyson, G. Laurenczy, Eur. J. Inorg. Chem. [44]
2008, 2008, 620-627.
[21]
C. Federsel, A. Boddien, R. Jackstell, R. Jennerjahn, P. J.
Dyson, R. Scopelliti, G. Laurenczy, M. Beller, Angew. Chem.
Int. Edit. 2010, 49, 9777-9780; F. Bertini, I. Mellone, A.
Ienco, M. Peruzzini, L. Gonsalvi, ACS Catal. 2015, 5, 1254- [45]
1265; S. Coufourier, S. Gaillard, G. Clet, C. Serre, M. Daturi,
J.-L. Renaud, Chem. Commun. 2019, 55, 4977-4980.
J. M. Heltzel, M. Finn, D. Ainembabazi, K. Wang, A. M.
Voutchkova-Kostal, Chem. Commun. (Camb) 2018, 54,
6184-6187.
[22]
[23]
[24]
G. Kovacs, G. Schubert, F. Joo, I. Papai, Catal. Today 2006, [46]
115, 53-60.
Y. Zhang, A. D. MacIntosh, J. L. Wong, E. A. Bielinski, P. G. [47]
Williard, B. Q. Mercado, N. Hazari, W. H. Bernskoetter,
G. Kovács, G. Schubert, F. Joó, I. Pápai, Catal. Today 2006,
115, 53-60.
Y.-y. Ohnishi, T. Matsunaga, Y. Nakao, H. Sato, S. Sakaki, J.
Amer. Chem. Soc. 2005, 127, 4021-4032.
Chem. Sci. 2015, 6, 4291-4299.
[48]
L. S. Sharninghausen, R. H. Crabtree, Isr. J. Chem. 2017,
10.1002/ijch.201700048.
D. Jantke, M. Cokoja, A. Pöthig, W. A. Herrmann, F. E. Kühn,
Organometallics 2013, 32, 741-744; A. Azua, S. Sanz, E.
Peris, Chem. Eur. J. 2011, 17, 3963-3967; D. Jantke, L.
Pardatscher, M. Drees, M. Cokoja, W. A. Herrmann, F. E.
Kühn, ChemSusChem 2016, 9, 2849-2854.
[25]
[26]
A. Azua, M. Finn, H. Yi, A. Beatriz Dantas, A. Voutchkova-
Kostal, ACS Sustain. Chem. Eng. 2017, 5, 3963-3972.
L. S. Sharninghausen, J. Campos, M. G. Manas, R. H.
Crabtree, Nat. Commun. 2014, 5; Z. Y. Lu, I. Demianets, R.
Hamze, N. J. Terrile, T. J. Williams, ACS Catal. 2016, 6, 2014-
2017; Y. Li, M. Nielsen, B. Li, P. H. Dixneuf, H. Junge, M.
Beller, Green Chem. 2015, 17, 193-198.
L. S. Sharninghausen, B. Q. Mercado, R. H. Crabtree, N.
Hazari, Chem. Commun. 2015, 51, 16201-16204.
M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi, C. Della Pina,
Angew. Chem. Int. Edit. 2007, 46, 4434-4440.
A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Green
Chem. 2008, 10, 13-30.
M. Dusselier, P. Van Wouwe, A. Dewaele, E. Makshina, B.
F. Sels, Energ. Environ. Sci. 2013, 6, 1415-1442.
J. N. Starr, G. Westhoff, Ullmann's Encycl. Ind. Chem. 2014,
1-8.
A. Kumar, S. Semwal, J. Choudhury, ACS Catal. 2019, 9,
2164-2168.
J. Su, L. Yang, X. Yang, M. Lu, B. Luo, H. Lin, ACS Sustain.
Chem. Eng. 2015, 3, 195-203.
A. Azua, J. A. Mata, E. Peris, F. Lamaty, J. Martinez, E.
Colacino, Organometallics 2012, 31, 3911-3919.
L. S. Sharninghausen, J. Campos, M. G. Manas, R. H.
Crabtree, Nat. Commun. 2014, 5, 5084.
U. Hintermair, U. Englert, W. Leitner, Organometallics
2011, 30, 3726-3731; A. R. Chianese, X. Li, M. C. Janzen, J.
W. Faller, R. H. Crabtree, Organometallics 2003, 22, 1663-
1667; G. D. Frey, C. F. Rentzsch, D. von Preysing, T. Scherg,
M. Muhlhofer, E. Herdtweck, W. A. Herrmann, J.
Organomet. Chem. 2006, 691, 5725-5738.
[49]
[50]
C. F. Rentzsch, E. Tosh, W. A. Herrmann, F. E. Kuhn, Green
Chem. 2009, 11, 1610-1617; M. V. Jiménez, J. Fernández-
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
Tornos, J. J. Pérez-Torrente, F. J. Modrego, P. Garcı
Orduña, L. A. Oro, Organometallics 2015, 34, 926-940.
́a-
[37]
M. G. Manas, J. Campos, L. S. Sharninghausen, E. Lin, R. H.
Crabtree, Green Chem. 2015, 17, 594-600.
10 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Page 11 of 11
Green Chemistry
View Article Online
DOI: 10.1039/D0GC01958E