Selective Catalytic Dehydrogenative Oxidation of Bio-Polyols to Lactic Acid

Article abstract of DOI:10.1002/anie.202004174

The global demand for lactic acid (LA) is increasing due to its successful application as monomer for the manufacture of bioplastics. Although N-heterocyclic carbene (NHC) iridium complexes are promising molecular catalysts for LA synthesis, their instabilities have hindered their utilization especially in commercial applications. Here, we report that a porous self-supported NHC-iridium coordination polymer can efficiently prevent the clusterization of corresponding NHC-Ir molecules and can function as a solid molecular recyclable catalyst for dehydrogenation of bio-polyols to form LA with excellent activity (97 percent) and selectivity (>99 percent). A turnover number of up to 5700 could be achieved in a single batch, due to the synergistic participation of the Ba²⁺ and hydroxide ions, as well as the blockage of unwanted pathways by adding methanol. Our findings demonstrate a potential route for the industrial production of LA from cheap and abundant bio-polyols, including sorbitol.

Full text of DOI:10.1002/anie.202004174

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Selective Catalytic Dehydrogenative Oxidation of Bio-polyols to
Lactic Acid
Authors: Jiajie Wu, Lingyun Shen, Sai Duan, Zhe-Ning Chen, Qingshu
Zheng, Yaoqi Liu, Zheming Sun, James H. Clark, Xin Xu, and
Tao Tu
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.202004174
Link to VoR: https://doi.org/10.1002/anie.202004174

10.1002/anie.202004174
Angewandte Chemie International Edition
RESEARCH ARTICLE
Selective Catalytic Dehydrogenative Oxidation of Bio-polyols to
Lactic Acid
Jiajie Wu,[a] Lingyun Shen,[a] Sai Duan,[b] Zhe-Ning Chen,[b] Qingshu Zheng,[a] Yaoqi Liu,[a] Zheming
Sun,[a] James H. Clark,[c] Xin Xu,*[b] and Tao Tu*[a,d,e]
Dedicated to Prof. Dr. Karl Heinz Dötz.
[a]
[b]
J. Wu, L. Shen, Dr. Q. Zheng, Y. Liu, Dr. Z. Sun, and Prof. Dr. T. Tu
Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, Department of Chemistry
Fudan University
2005 Songhu Road, Shanghai 200438 (China)
E-mail: taotu@fudan.edu.cn
S. Duan, Dr. Z.-N. Chen, Prof. Dr. X. Xu
Collaborative Innovation Center of Chemistry for Energy Materials, MOE Laboratory for Computational Physical Science
Fudan University
2005 Songhu Road, Shanghai 200438 (China)
E-mail: xxchem@fudan.edu.cn
[c]
[d]
[e]
Prof. J. H. Clark
Green Chemistry Centre of Excellence
University of York
York YO105DD, UK
Prof. Dr. T. Tu
State Key Laboratory of Organometallic Chemistry
Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences
354 Fenglin Road, Shanghai 200032 (China)
Prof. Dr. T. Tu
College of Chemistry and Molecular Engineering
Zhengzhou University
100 Kexue Avenue, Zhengzhou 450001 (China)
Supporting information for this article is given via a link at the end of the document.
Abstract: The global demand for lactic acid (LA) is increasing due to
promising renewable carbon sources for this purpose, as all of
them are readily accessed by simple hydrolysis of agricultural
residues, lignin, cellulose, and hemicellulose.[2] An emerging
technique for their utilization is the direct conversion of bio-polyols
such as sorbitol, using multifunctional catalysts.[2] Apart from its
applications in food, cosmetics, and paper industries as
additives,[3] sorbitol is also a platform molecule useful for the
synthesis of various valuable fine chemicals, including ethylene
glycol (EGO),[4] 1,2-propanediol,[4b] and lactic acid (LA).[5] Given
the steady increase in the price and annual consumption of LA
over the last ca. 20 years, largely as a result of the increasing
global demands for biodegradable polylactide,[6] there is a
particular merit in developing efficient and selective approaches
for LA production from sorbitol and other bio-polyols.
Although a great deal of effort has been devoted to this
important transformation, only two strategies, i.e., hydrogenolysis
and dehydrogenative oxidation, have been realized so far.[4a,5] But
in both cases, the yield and the selectivity of LA were quite low
and well below the industrial requirement even at elevated
pressures and temperatures. As shown in Figure 1a, the best
result for this challenging transformation is a yield of 40% with a
selectivity of 82% to LA. This was based on bis-N-heterocyclic
carbene (NHC) iridium complex 1,[7] which is easily converted to
various inactive species (like 2 in Figure 1a, a dimer with Ir-Ir
bond) due to the possible clusterization (dimerization,[5]
trimerization, tetramerization, and even multimerization)[8] of
NHC-Ir complexes, could not be avoided by simply increasing the
bulkiness of the NHC ligands.[5] Despite the low performance, this
is still a very attractive protocol because clean energy H2 as well
its prominence as a bio-based platform molecule and its successful
application as
a monomer for bio-plastics manufacture. The
development of new, selective, and sustainable approaches for LA
synthesis using highly active catalysts is becoming critically important.
Although N-heterocyclic carbene (NHC) iridium complexes have been
proven as promising molecular catalysts for this purpose, their
instabilities have seriously hindered their potential utilization
especially in commercial applications. Here, we report for the first time
that a porous self-supported NHC-iridium coordination polymer can
efficiently prevent the clusterization of corresponding NHC-Ir
molecules and can function as a solid molecular recyclable catalyst
for dehydrogenation of a number of bio-polyols to make LA with
excellent activity (97%) and selectivity (>99%), and up to 5700
turnover number could be achieved in a single batch, due to the
synergistic participation of the Ba2+ and hydroxide ions, as well as the
blockage of unwanted pathways by adding methanol. Our findings
demonstrate a potential route for the industrial production of LA from
cheap and abundant bio-polyols, including sorbitol.
Introduction
With the goal of alleviating our serious dependence on fossil fuels,
the catalytic conversion of inexpensive bio-renewables into fuels
and value-added chemicals has received considerable attention
during the last decade. [1] Bio-polyols such as sorbitol, glycerol,
xylitol, and mannitol have been regarded as among the most
1
This article is protected by copyright. All rights reserved.

10.1002/anie.202004174
Angewandte Chemie International Edition
RESEARCH ARTICLE
Figure 1. Oxidative dehydrogenation of sorbitol to LA. a) Previous work: Catalyzed by homogenous catalyst NHC-Ir 1 (yield was calculated based on 1 equiv. LA
formed per sorbitol according to Ref 8a). b) This work: Catalyzed by self-supported catalysts 5-6 prepared by coordination assembly (yield was calculated based
on 2 equiv. LA formed per sorbitol).
as LA was produced.[9] Presumably, the prevention of multi-
nuclear iridium clustering may provide a viable way to resolve the
problems and move this bio-resource conversion towards
commercial viability.
production from sorbitol. Furthermore, the reaction mechanisms
have been disclosed at the state-of-the-art first-principles level. It
is also noteworthy that one molecule of sorbitol generates two
molecules of LA in this protocol rather than that one sorbitol
produced only one LA in the previous study.[5] The present
protocol is thus obviously more atom economic. These significant
advances have greatly increased the industrial potential for this
important transformation.
Recently, the “self-supporting” approach has emerged as a
new and efficient strategy for catalyst immobilization using
coordination assembly of the corresponding ligands with metal
ions.[10] The resulting self-supported catalysts can function as
solid molecular catalysts and can even exhibit a higher catalytic
activity than the corresponding homogenous catalysts as found in
several cases.[11] In our experience, the “self-supporting” strategy
looks promising for our purpose here, as the ditopic ligands can
isolate the catalytically active metal ions and block the possible
clusterization of the iridium catalysts.[5,8] Therefore, in the present
work, we have synthesized a series of three-dimensional (3D)
porous NHC-Ir coordination polymeric assemblies (5a-d and 6a-
d) by using the “self-supporting” strategy (Figure 1b). These
coordination assemblies efficiently hamper the possible
clusterization of iridium species, and function as solid molecular
catalysts with excellent catalytic activity and selectivity for LA
Results and Discussion
Syntheses and characterizations of the three-
dimensional NHC-Ir coordination assemblies. With the “self-
supporting” strategy in mind, tetraimidazolium iodides 3 and 4 with
two different rigid cores adamantane (Ad) or carbon (C) were
synthesized via Suzuki coupling and N-methylation reactions (see
Supporting Information (SI) for more details). The diphenyl
fragment was introduced to ensure that after deprotonation with a
This article is protected by copyright. All rights reserved.

10.1002/anie.202004174
Angewandte Chemie International Edition
RESEARCH ARTICLE
base, the resultant tetradentate terminal NHCs were
geometrically well-separated from each other and could only
further bind to different Ir centers to generate an extended 3D
supramolecular structure via the coordination assembly (Figure
1b). Subsequently, tetraimidazolium salts with different ions were
prepared by treating the iodides 3a and 4a with NH4PF6 or
Et3OBF4 to produce salts 3b-c and 4b-c, respectively. After
deprotonation by lithium bis(trimethylsilyl)amide (LiHMDS) and
further coordination with the selected iridium precursors
([Ir(acac)(CO)2], acac = acetylacetonato or [Ir(COD)Cl]2, COD =
1,5-cyclooctadiene), a number of 3D coordination assemblies 5a-
matrix of solid 5a (Figure 1d). All these data clearly indicated that
the coordination environment of the Ir center in the solid 5a is
quite similar to that in the molecular complex 8. Thus assemblies
such as solid 5a can be expected to function as molecular
catalysts.
To measure the porosity of the NHC-Ir coordination
assemblies, N2 adsorption/desorption studies were carried out
(Figure S34). The 3D coordination assembly 5a has a surface
area of 388 m2g-1 that is almost 50 times greater than that of 1D
coordination assembly 7a (8 m2g-1), highlighting the porous
structure of the solid assembly 5a.
d and 6a-d containing different counter-ions (I-, BF4- or PF6 ) and
-
ancillary ligands (CO or COD) were readily obtained in excellent
yields (95-99%). For comparison, one dimensional (1D) linear
NHC-Ir coordination assemblies 7a-b were also synthesized from
the corresponding bis-imidazolium salts based on our previous
reports.[11]
All obtained coordination assemblies 5a-d, 6a-d and 7a-b are
essentially insoluble in water and all organic solvents tested. Thus,
their compositions can only be fully characterized by using solid-
state analytical techniques. The morphologies of the freshly
prepared 3D coordination assemblies 5a-d and 6a-d were
investigated by scanning electron microscopy (SEM). Similar to
the 1D NHC-Ir assemblies 7a-b,[11] the 3D NHC-Ir assemblies 5a-
d and 6a-d are composed of some irregular nano-particles (Figure
2a, and Figures S2-S10). The transmission electron microscopy
(TEM) confirmed this observation (Figure 2b and Figures S11-13).
The average size for NHC-Ir assembly 5a is ca. 450 nm as
measured by dynamic light scattering (DLS, Figure S35), which is
slightly larger than the diameter observed by SEM (around ca.
300 nm, Figure 2a). This difference might be caused by a solvent
effect in the hydrated state.[12]
Furthermore, taking solid 5a as an example, the amorphous
nature of NHC-Ir coordination assemblies was shown by the
powder X-ray diffraction (PXRD) patterns (Figure S18). For
comparison, a homogeneous NHC-Ir complex 8 was synthesized
from 1-methyl-3-phenylimidazolium iodide and [Ir(COD)Cl]2, and
was fully characterized by 1H and 13C NMR spectra (Figures S61-
62), and high resolution mass spectrometry (see SI for details).
Pleasingly, most peaks in the solid 13C NMR spectra of assembly
5a are clearly assignable to the structural carbon atoms (Figure
S63). The peak at 174.4 ppm is attributed to the CNHC of solid 5a,
and the corresponding signal in the spectra of the molecular
complex 8 is at 175.8 ppm. X-ray photoelectron spectroscopy
(XPS) signals of the agglomerated iridium nano-particles were not
observed in solid 5a (Figure S20). Instead, only Ir(I) signals at
62.2 (4f5/2) eV and 65.2 (4f7/2) eV that correspond to the two
characteristic bands of molecular iridium species emerge for both
solid 5a and complex 8 (Figure S29). Energy dispersive X-ray
spectroscopy (EDX) quantification of 5a revealed a uniform
distribution of all expected elements, including C, Ir and, N, in the
assembly matrix (Figure 2c). Furthermore, magnified high-angle
annular dark-field scanning transmission electron microscopy
(HAADF-STEM) images showed well dispersed white dots,
indicating that individual Ir atoms are atomically dispersed in the
Figure 2. Morphology studies of catalyst 5a. a) SEM image of 5a. b) TEM image
of 5a. c) EDS mapping of 5a. d) Magnified HAADF-STEM image of 5a, white
dots represent individual Ir atoms, indicating that the iridium centers are
atomically dispersed in the assembly matrix.
The catalytic activity of the NHC-Ir coordination assemblies.
The catalytic activities of the characterized NHC-Ir coordination
assemblies were then tested for the oxidative dehydrogenation of
bio-polyols initially using D-sorbitol as the substrate. In the
presence of 0.2 mol % NHC-Ir coordination assemblies, 3 mmol
D-sorbitol and 1.5 equiv. of Ba(OH)2 were heated in a sealed tube
at 160 °C for 6 hours.[2a] Compared to the molecular NHC-Ir
complex 1, linear 1D NHC-Ir coordination assembly 7b showed a
slight improvement in the yield and selectivity to LA (48% vs 35%
for 7b and 1, respectively, Figure 3a). These observations are
important because: (i) a supported catalyst normally results in a
worse yield and selectivity than the corresponding homogenous
privileged catalysts;[13] (ii) an extended reaction time or increased
This article is protected by copyright. All rights reserved.

10.1002/anie.202004174
Angewandte Chemie International Edition
RESEARCH ARTICLE
Figure 3. Optimization of reaction conditions and reusability of the catalyst for oxidative dehydrogenation of sorbitol to LA. a) Catalyst screening with 1.5 equiv.
Ba(OH)2, 3 mmol D-sorbitol, 0.2 mol % catalyst, in 0.3 mL H2O at 160 °C for 6 hours. b) Influence of alcohol additives and temperatures. Reactions were carried
out with 3 mmol D-sorbitol, 0.2 mol% 5a, 2.5 equiv. Ba(OH)2·8H2O, and 0-10 equiv. of selected alcohols as additional additives at 150 °C (left 9 purple bars); 5
equiv. MeOH at 140 °C (blue bar) and at 130 °C (right purple bar) for 12 hours. c) Reusability of catalyst 5a. Reactions were carried out with 3 mmol D-sorbitol, 0.2
mol% 5a and 2.5 equiv. Ba(OH)2·8H2O at 140 °C for 12 hours (1-12 runs, the blue bars), or with extended reaction time to 24 hours (after the 12th run, the turquois
bar). All the conversions, yields and selectivities were determined by an HPLC analysis and calculated based on 2 equiv. LA formed per sorbitol.
temperature will not improve the product yields due to
deactivation of the molecular catalysts through clusterization.[5,8]
Furthermore, when other iridium and ruthenium complexes 8-12,
privileged molecular catalysts in various dehydrogenative
transformations,[8a,14] were applied to this reaction, inferior yields
and selectivities were observed (Figure 3a). Therefore, the “self-
supporting” strategy can indeed be considered as an effective
approach to increase the yield and selectivity for the
dehydrogenative transformation from sorbitol to LA.
As porous catalysts can accelerate mass and heat transfer[15]
and thus facilitate reactions, we then tested the porous 3D NHC-
Ir coordination assemblies 5a-d and 6a-d (See Figure 3a).
Compared to the 1D assembly 7, the 3D assemblies 5 and 6
exhibited slightly higher yields and selectivities (57%-70% vs
48%-50%, Figure 3a), suggesting that a high porosity of the
catalysts was beneficial to the synthesis of LA. Higher selectivities
Catalyst composition has a significant effect on the reaction, and
the bidentate COD ligand was found to be better than the
monodentate CO ligand in both cases (See 5a vs 5b and 6a vs
6b in Figure 3a). In contrast with non-coordinating BF4− and PF6
−
anions, the iodide anion may well be located close to the iridium
center, leading to enhanced activity and selectivity (See 5a vs 5c-
d and 6a vs 6c-d in Figure 3a). With all these outcomes, solid
assemblies 5a has been identified as the most efficient catalyst
for the oxidative dehydrogenation of sorbitol with a yield of 70%
for LA within 6 hours.
The other reaction parameters, including base, additives, and
reaction temperature, were then investigated (See Table S1 and
Figure S1). We found that bases played a critical role in the
transformation. Weak bases, such as K3PO4 and K2CO3, all
resulted in only traces of LA (See entries 2 and 3 in Table S1). A
strong base, e.g., t-BuOK, was also unable to convert sorbitol. We
found that the presence of the hydroxide ion OH− is essential to
the dehydrogenation process (See entry 4 in Table S1). Therefore,
other alkali hydroxides (i.e., NaOH and CsOH) and some alkaline-
earth metal hydroxides (e.g. Ca(OH)2 and Ba(OH)2) were tested
were obtained by using the assemblies 5a-d with
a
tetrakisphenyladamantane core (R = Ad, Figure 3a). This may
indicate that the substrate can contact the catalytic sites better for
a catalyst that has a higher surface area and a larger pore size.
This article is protected by copyright. All rights reserved.

10.1002/anie.202004174
Angewandte Chemie International Edition
RESEARCH ARTICLE
under otherwise identical reaction conditions. The use of Ca(OH)2
led to equal selectivity between LA and EGO. However, in
contrast to other metal hydroxides, up to 90% conversion of
sorbitol was achieved using Ba(OH)2, with an LA selectivity up to
53%. On the other hand, little reaction occurred when BaCO3 or
BaCl2 was used (See entries 14 and 15 in Table S1). The quantity
of Ba(OH)2 and the reaction temperature were also found to be
important parameters in the reaction. Up to 98% conversion of
sorbitol with up to 70% yield of LA was observed using 2.5 equiv.
Ba(OH)2 at 150 °C (See entries 11 to 13 in Table S1).
thermodynamics (A comparison with other possible catalytic
species is shown in the SI, Schemes S6 - S8). In addition, it is
worth noting that, according to the DFT calculations, the NHC
ligands of the catalysts are firmly bonded to the Ir(I) center. The
dissociation free energy of the NHC ligand from the Ir(I) center is
23.6 kcal/mol. On the other hand, the conversion from η2 to η1 for
the COD ligand is more feasible (19.5 kcal/mol), which further
confirms the structural stability of the solid catalyst 5a.
The second step of the transformation involves the retro-aldol
reaction of glucose and fructose, forming smaller carbohydrate
molecules including glycolaldehyde plus erythrose (via the [2+4]
route for the former, Figure 4b) or dihydroxyacetone plus
glyceraldehyde (via the [3+3] route for the latter, Figure 4c). The
reaction mechanisms for the conversion of these two C3
molecules to LA are well established. [16] It is also known that the
C4 molecule (i.e., erythrose) can be further converted into lactate
and methanol[5] (Figure 4b,c). DFT calculations give a barrier of
46.8 kcal/mol for the [2+4] route via TS4 and 45.0 kcal/mol for the
[3+3] route via TS5, showing that the direct C-C cleavage via
these two routes are not really feasible. With the participation of
Ba2+/hydroxide, the corresponding barriers are dramatically
reduced to 22.6 and 20.2 kcal/mol via TS4Ba and TS5Ba,
respectively, demonstrating undoubtedly the key role played by
Mechanistic studies. We carried out a time-dependent study
on the oxidative dehydrogenation of sorbitol using 0.2 mol % solid
catalyst 5a in order to better understand the possible reaction
pathways. The liquid products in the reaction mixtures, ranging
from the unreacted sorbitol and the products LA, glyceraldehyde
(GLY), EGO, methanol and ethanol, etc., were analyzed directly
by High Performance Liquid Chromatography (HPLC) with a
refractive index (RI) detector. The concentration profiles were
plotted as a function of time in Figure S72. As expected, under
alkaline conditions and with hydrogen liberation, no sugar product
was observed. The reaction rate was found to strongly depend on
sorbitol concentration: after 1 hour over 60% of sorbitol was
consumed, a 90% conversion was reached after 5 hours.
Interestingly, the LA selectivity stayed ca. 70% during the reaction, the Ba2+ ion and the base (Figs. 4b and c). Considering that the
while EGO increased almost linearly during the first 30 minutes
and then decreased slowly, indicating that EGO may be one of
the reaction intermediates. By extending the reaction time, short
aliphatic alcohols (MeOH and EtOH) were generated, possibly
formed by further cleavage of LA[5] and hydrogenation of EGO
along with the liberation of hydrogen.
participation of Ba2+/hydroxide can facilitate the generation of
lactate species both kinetically and thermodynamically, a high
activity and a high selectivity towards lactate in our experiments
has thus been confirmed.
These proposed reaction mechanisms involve several
reversible reactions, e.g., the retro-aldol reaction, and the
isomerization between glucose and fructose. Controlling reaction
reversibility provides an additional tool to optimize selectivity. It is
interesting to note that two equivalents of LA are generated from
one equivalent of sorbitol via the [3+3] route, while, only one
equivalent of LA is obtained via the [2+4] route. In other words, it
is advantageous to enhance the [3+3] route and depress the [2+4]
route for a greater selectivity to LA.
The first step in the sorbitol transformation is an iridium-
catalyzed dehydrogenation step leading to glucose or fructose
generation. Here complex 8 was used as a molecular model of
the catalytic center of the NHC-Ir coordination assemblies (Figure
4a). Our density functional theory (DFT) calculations[16]
demonstrated that the NHC-Ir-catalyzed dehydrogenative
reaction here is actually a redox-neutral process (i.e., from an Ir(I)
species to an Ir(I) species without any valence change). As shown
in Figure 4a, partial dissociation of the COD ligand from η2 to η1
is the initial step to generate the active catalytic Ir(I) species A.
Subsequently, with the participation of the base, a) glucose and
b) fructose are generated with the activation free energies of 25.2
and 26.5 kcal/mol (TS1 and TS2), respectively. The primary (α)
and the secondary (β) hydroxyl groups in sorbitol provide the
proton to the base, and concertedly H-bond to the Cα or the Cβ
site of sorbitol,[17] forming an iridium hydride intermediate B. When
the COD ligand bis-chelates to the Ir(I) center (from η1 to η2), a
penta-coordinated hydride intermediate C is then formed. After
hydrogen liberation, the initial NHC-Ir catalyst 8 is regenerated.
As shown in Figure 4a, although the generation of glucose seems
slightly more favorable than that of fructose (TS1 vs. TS2), their
activation free energies for the catalyst regeneration are very
similar (TS3: 35.5 vs. 35.9 kcal/mol), diminishing the energy
difference for these two possible reaction pathways. This
similarity in kinetics is also supported by the calculated
With this mechanistic understanding in mind, we have
selected several primary alcohols as additives in the model
reaction under the standard reaction conditions (Figure 3b).
Remarkably, the yields of LA were increased dramatically when
methanol was added (Figure 3b). Full conversion and an
impressive yield of 95% for LA were obtained when 5 equivalents
of methanol were used. However, other alcohols have little or no
influence on the selectivity (Figure 3b). These observations can
be attributed to the depression the [2+4] route by adding methanol,
as methanol is the product using this route according to our
proposed mechanisms. More methanol was seen to decrease the
activity and the selectivity, presumably because this destroyed
the solute-free reaction conditions. When ethylene glycol was
added, the total conversion was significantly decreased but with
a complete selectivity to LA, which suggested that the addition of
ethylene glycol blocks the [2+4] route and pushes the reaction
equilibrium over to the [3+3] route. Figure 3b also summarizes the
effects of reaction temperature. Decreasing the temperature to
This article is protected by copyright. All rights reserved.

10.1002/anie.202004174
Angewandte Chemie International Edition
RESEARCH ARTICLE
Figure 4. a) The calculated catalytic dehydrogenation from sorbitol to glucose or fructose with NHC-Ir 8 (Values are relative Gibbs free energies (∆G, in kcal/mol)
with respect to sorbitol and catalyst 8. In addition, ∆G for the pathway to produce fructose or glucose is labeled in dark blue or dark red, respectively. Clusters of
[OH-(H2O)2] and (H2O)3 are used to simulate the base and water, respectively, since these clusters represent a good compromise between accuracy and efficiency
as shown[15b,18] for the description of the proton exchange reaction. Calculated mechanisms for the generation of lactate and other by-products via b) [2+4] and c)
[3+3] routes without (dark red) and with (dark green) the participation of Ba2+/hydroxide.
140 °C led to a slightly higher yield (97%), while lowering the
temperature led to a reduction in selectivity (88%).
MeOH. Full conversion and almost quantitative yields of LA
(>99%) were obtained in the first 11 runs under the optimized
reaction conditions, demonstrating its high catalytic stability and
efficiency. A slight decline of the conversion and yield was
observed in the 12th run (82%) with the selectivity decreased to
89% (Figure 3c; see SI for details). When the reaction time was
extended to 24 hours after these multiple runs, the conversion and
yield increased slightly, while the selectivity went down slightly to
98%.
Reusability of the solid catalyst. Being insoluble in water
and most organic solvents, solid catalyst 5a is well suited to
recycling. Practical testing proved that this is particularly helping
to build the case for an industrially viable catalytic system. After
simple centrifugation, filtration, and washing with water and
methanol, the solid molecular catalysts were readily recovered for
the next run, when simply recharged with sorbitol, base, and
This article is protected by copyright. All rights reserved.

10.1002/anie.202004174
Angewandte Chemie International Edition
RESEARCH ARTICLE
In order to understand the recycling process better, the
content of iridium in the filtrate after the multiple runs was
analyzed by inductively coupled plasma emission spectroscopy
(ICP-AES, See Table S3). Approximately, 0.1 μg/mL iridium
leaching was observed after six runs, presumably due to a trace
amount of inactive iridium species or precursors entrapped in the
matrix cavities of the assemblies. The amount of leaching
afterwards was negligible. The Ir content of the recovered solid
5a after 3 runs was measured as 19.20%, which was practically
the same as in the freshly prepared solid catalyst 5a (19.82%)
determined by ICP-AES and was also consistent with the
theoretical value (20.02%) of [IrC45H44IN4]n. As investigated by
SEM and TEM, the morphology of the recovered solid catalyst 5a
after 16 runs was found to be very similar to that of the newly
prepared polymer (Figure S2 vs Figure S3, and Figure S11 vs
S12), further confirming the stability of the solid molecular
catalysts.
sugar alcohols, except two disaccharides, were selectively
converted to LA in good to excellent yields with full conversions
with or without MeOH addition. Remarkably, the addition of
methanol did improve the selectivity of the oxidative
dehydrogenation of sugar alcohols in all cases. Mannitol or
dulcitol, as the isomer of sorbitol, gave similar yields with a full
substrate conversion. These similar outcomes indicated that the
different orientation of the hydroxyl groups hardly affected the
dehydrogenation process. For all three kinds of bio-polyols with
five carbons, a full conversion with a high selectivity to LA was
achieved with the addition of methanol. However, the four carbon
bio-polyol erythritol gave a poorer result possibly due to a different
pathway to form LA. As the cleavage of glucosidic bonds is
unlikely to happen under the alkaline condition,[19] the yield of LA
is unable to exceed 50% when using maltitol and lactitol as
substrates.
Encouraged by the remarkable performance of the solid
molecular catalysts, a scale-up reaction using only 500 ppm 5a
was carried out. With a reaction temperature of 160 °C, sorbitol
was fully converted with a complete selectivity to LA in 48 hours.
Reducing the catalyst loading to 150 ppm led to a large drop to
only 43% conversion of sorbitol after 100 hours under the same
reaction conditions, although this corresponded to an increased
turn-over-number (TON) of 5700.
Conclusion
We have made a significant advance in the production of LA
from simple, low cost bio-polyols. Efficient and selective catalytic
oxidative dehydrogenation was achieved using a 3D self-
supported NHC-iridium catalyst described for the first time. The
solid catalysts synthesized by coordination assembly showed a
significantly enhanced stability with high activity and perfect
selectivity for dehydrogenation of sorbitol to produce the
commercially important product LA with the help of Ba2+/
hydroxide ions and methanol. Computational studies revealed
that in the dehydrogenation of sorbitol under alkaline conditions,
the ligand dissociation of the NHC-iridium catalysts was avoided
via an outer-sphere catalytic pathway, which was beneficial to
their structural stability. The high activity and selectivity towards
LA were facilitated by the participation of Ba2+/hydroxide both
kinetically and thermodynamically. By blocking the [2+4] route to
favor the [3+3] route, the right amount of methanol could further
increase the selectivity to LA. At the optimal reaction conditions,
the yield of LA has reached a remarkable 97%, which sheds light
on the potential industrial production of LA from cheap and
abundant biopolyols.
O
[Cat.], Ba(OH)2
H
Cn 2n+2On
OH
HO
n = 4, 5, 6, 12
OH
OH OH
OH OH
OH OH
OH
OH
OH
HO
HO
OH OH
OH OH
OH OH
Sorbitol
Mannitol
Dulcitol
70% / 97%
66% / 95%
60% / 99%
OH
OH
OH
HO
OH
HO
OH
HO
OH
OH OH
OH OH
OH OH
Xylitol
69% / 92%
Adonitol
81% / 88%
Arabitol
73% / 99%
HO
HO
OH
OH
OH
OH
OH
OH
OH
O
O
O
O
OH
HO
OH
HO
OH
HO
OH
OH
OH
OH
OH
OH
OH
Erythritol
48% / 67%
Maltitol
29% / 51%
Lactitol
26% / 40%
Acknowledgements
Figure 5. Substrate scope without (left values) or with (right values) the addition
of 5 equiv. of methanol. Reactions conditions: sugar alcohol (3 mmol), solid
catalyst 5a (0.2 mol%), Ba(OH)2·8H2O (2.5 equiv.), 140 °C, 12 hours.
Conversion and selectivity were determined by an HPLC analysis and
calculated on a molar carbon basis.
Financial supports from the National Key R&D Program of China
(Nos. 2016YFA0202902 and 2018YFA0208600), the National
Natural Science Foundation of China (Nos. 21861132002,
21871059, 21572036 and 21688102), and Department of
Chemistry, Fudan University are gratefully acknowledged.
Substrate scope. With a full substrate conversion and an
excellent selectivity to LA in the oxidative dehydrogenation of
sorbitol using only 0.2 mol % NHC-Ir solid molecular catalyst 5a,
the substrate scope of different bio-polyols containing 4, 5, 6 and
12 carbon atoms was then explored at 140 °C for 12 hours with
or without the addition of 5 equiv. methanol (Figure 5). All selected
Keywords: Lactic acid • bio-polyols • N-heterocyclic carbene
iridium complex • oxidative dehydrogenation • self-supporting
[1]
a) G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044-4098.
b) J. N. Chheda, G. W. Huber, J. A. Dumesic, Angew. Chem. Int. Ed.
This article is protected by copyright. All rights reserved.

10.1002/anie.202004174
Angewandte Chemie International Edition
RESEARCH ARTICLE
2007, 46, 7164-7183; c) M. Besson, P. Gallezot, C. Pinel, Chem. Rev.
2014, 114, 1827-1870; d) B. Zada, M. Chen, C. Chen, L. Yan, Q. Xu, W.
Li, Y. Fu, Sci. China Chem. 2017, 60, 853-869; e) S. Shylesh, A. A.
Gokhale, C. R. Ho, A. T. Bell, Acc. Chem. Res. 2017, 50, 2589-2597; f)
T. A. Bender, J. A. Dabrowski, M. R. Gagné, Nat. Rev. Chem. 2018, 2,
35-46; g) P. Sudarsanam, R. Zhong, S. Van den Bosch, S. M. Coman,
V. I. Parvulescu, B. F. Sels, Chem. Soc. Rev. 2018, 47, 8349-8402; h) Z.
Zhang, Q. Yang, H. Chen, K. Chen, X. Lu, P. Ouyang, J. Fu, J. G. Chen,
Green Chem. 2018, 20, 197-205; i) H. C. Ong, W. H. Chen, A. Farooq,
Y. Y. Gan, K. T. Lee, V. Ashokkumar, Renew. Sust. Energ. Rev. 2019,
113, 109266; j) Y. Jing, Y. Guo, Q. Xia, X. Liu, Y. Wang, Chem 2019, 5,
2520-2546.
[2]
a) L. N. Ding, A. Q. Wang, M. Y. Zheng, T. Zhang, ChemSusChem 2010,
3, 818-821; b) X. Jin, B. Yin, Q. Xia, T. Fang, J. Shen, L. Kuang, C. Yang,
ChemSusChem 2019, 12, 71-92.
[3]
[4]
[5]
[6]
S. Esposito, B. Silvestri, V. Russo, B. Bonelli, M. Manzoli, F. A. Deorsola,
M. Di Serio, ACS Catal. 2019, 9, 3426-3436.
a) Y. Jia, H. Liu, Chin. J. Catal. 2015, 36, 1552-1559; b) X. Guo, J. Guan,
B. Li, X. Wang, X. Mu, H. Liu, Sci. Rep. 2015, 5, 1-9.
M. G. Manas, J. Campos, L. S. Sharninghausen, E. Lin, R. H. Crabtree,
Green Chem. 2015, 17, 594-600.
M. Dusselier, P. Van Wouwe, A. Dewaele, E. Makshina, B. F. Sels, Energ.
Environ. Sci. 2013, 6, 1415-1442.
[7]
[8]
M. Iglesias, L. A. Oro, Chem. Soc. Rev. 2018, 47, 2772-2808.
a) L. S. Sharninghausen, J. Campos, M. G. Manas, R. H. Crabtree, Nat.
Commun. 2014, 5, 5084; b) J. Campos, L. S. Sharninghausen, R. H.
Crabtree, D. Balcells, Angew. Chem. Int. Ed. 2014, 53, 12808-12811;
Angew. Chem. 2014, 126, 13022-13025; c) L. S. Sharninghausen, R. H.
Crabtree, Isr. J. Chem. 2017, 57, 937-944.
[9]
P. Nikolaidis, A. Poullikkas, Renew. Sust. Energ. Rev. 2017, 67, 597-611.
[10] a) L. X. Dai, Angew. Chem. Int. Ed. 2005, 43, 5726-5729; Angew. Chem.
2005, 116, 5846-5850; b) Z. Wang, G. Chen, K. Ding, Chem. Rev. 2009,
109, 322-359.
[11] a) Z. Sun, Y. Liu, J. Chen, C. Huang, T. Tu, ACS Catal. 2015, 5, 6573-
6578; b) J. Chen, J. Wu, T. Tu, ACS Sustain. Chem. Eng. 2017, 5, 11744-
11751; c) Z. Sun, J. Chen, T. Tu, Green Chem. 2017, 19, 789-794; d) Y.
Zhang, J. Wang, H. Zhu, T. Tu, Chem. Asian J. 2018, 13, 3018-3021.
[12] W. She, K. Luo, C. Zhang, G. Wang, Y. Geng, L. Li, B. He, Z. Gu,
Biomaterials 2013, 34, 1613-1623.
[13] a) Q.-H. Fan, K. Ding, Enantioselective catalysis with structurally tunable
immobilized catalysts. In Asymmetric Catalysis from
a Chinese
Perspective (pp. 207-245). Springer, Berlin, Heidelberg 2011; b) R.
Zhong, A. C. Lindhorst, F. J. Groche, F. E. Kühn, Chem. Rev. 2017, 117,
1970-2058.
[14] a) O. Blum, D. Milstein, J. Am. Chem. Soc. 2002, 124, 11456-11467; b)
A. Azua, J. A. Mata, E. Peris, Organometallics 2011, 30, 5532-5536; c)
W. Kuriyama, T. Matsumoto, O. Ogata, Y. Ino, K. Aoki, S. Tanaka, T.
Saito, Org. Process Res. Dev. 2012, 16, 166-171; d) J. J. A. Celaje, Z.
Lu, E. A. Kedzie, N. J. Terrile, J. N. Lo, T. J. Williams, Nat. Commun.
2016, 7, 11308; e) S. Movassaghi, S. Singh, A. Mansur, K. K. Tong, M.
Hanif, H. U. Holtkamp, C. G. Hartinger, Organometallics, 2018, 37, 1575-
1584.
[15] a) B. Fang, J. H. Kim, M. S. Kim, J. S. Yu, Acc. Chem. Res. 2013, 46,
1397-1406.; b) Y. Shen, Q. Zheng, H. Zhu, T. Tu, Adv. Mater. 2020, 32,
1905950.
[16] I. Y. Zhang, X. Xu, Y. Jung, W. A. III. Goddard, Proc. Natl. Acad. Sci. U.
S. A. 2011, 108, 19896-19900.
[17] Y. Wang, W. Deng, B. Wang, Q. Zhang, X. Wan, Z. Tang, H. Wan, Nat.
Commun. 2013, 4, 2141.
[18] J. Joubert, F. Delbecq, Organometallics 2006, 25, 854-861.
[19] a) W. Deng, Q. Zhang, Y. Wang, Catal. Today 2014, 234, 31-41; b) Y.
Dai, C. Shao, Y. Piao, H. Hu, K. Lu, T. Zhang, S. Man, Carbohydr. Polym.
2017, 178, 34-40.
This article is protected by copyright. All rights reserved.

10.1002/anie.202004174
Angewandte Chemie International Edition
RESEARCH ARTICLE
Entry for the Table of Contents
Bio-polyols conversion: A series of 3D porous self-supported NHC-iridium coordination polymers were prepared and could function
as recyclable solid molecular catalysts for dehydrogenation of bio-polyols to lactic acid with excellent activity and selectivity due to the
synergistic participation of the Ba2+ and hydroxide ions, as well as the blockage of unwanted pathways by adding methanol.
This article is protected by copyright. All rights reserved.