Glucose oxidation to formic acid and methyl formate in perfect selectivity

Article abstract of DOI:10.1039/d0gc01169j

We report the highly remarkable discovery that glucose oxidation catalysed by polyoxometalates (POMs) in methanolic solution enables formation of formic acid and methyl formate in close to 100percent combined selectivity, thus with only negligible sugar oxidation to CO2. In detail, we report oxidation of a methanolic glucose solution using H8[PV5Mo7O40] (HPA-5) as catalyst at 90 °C and 20 bar O2 pressure. Experiments with 13C-labelled glucose confirm unambiguously that glucose is the only source of the observed formic acid and methyl formate formation under the applied oxidation conditions. Our results demonstrate a very astonishing solvent effect for the POM-catalysed glucose oxidation. In comparison to earlier work, a step-change in product yield and selectivity is achieved by applying an alcoholic reaction medium. The extremely high combined yields of formic acid and methyl formate greatly facilitate product isolation as low-boiling methyl formate (bp = 32 °C) can simply be isolated from the reaction mixture by distillation.

Full text of DOI:10.1039/d0gc01169j

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: S. Maerten, C.
Kumpidet, D. Voß, A. Bukowski, P. Wasserscheid and J. Albert, Green Chem., 2020, DOI:
1
0.1039/D0GC01169J.
Volume 20
Number
May 2018
Pages 1919-2160
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
9
7
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

Page 1 of 20
Green Chemistry
View Article Online
DOI: 10.1039/D0GC01169J
Glucose oxidation to formic acid and methyl formate in perfect
selectivity
a
a
a
a
Stephanie Maerten , Chiraphat Kumpidet , Dorothea Voß , Anna Bukowski , Peter
a
a
Wasserscheid and Jakob Albert*
aLehrstuhl für Chemische Reaktionstechnik, Friedrich-Alexander-Universität Erlangen-
Nürnberg, Egerlandstraße 3, 91058 Erlangen, Germany.
*
Email: jakob.albert@fau.de
Abstract
We report the highly remarkable discovery that glucose oxidation catalysed by
polyoxometalates (POMs) in methanolic solution enables formation of formic acid and methyl
formate in close to 100% combined selectivity, thus with only negligible sugar oxidation to
CO . In detail, we report oxidation of a methanolic glucose solution using H [PV Mo O ]
2
8
5
7
40
13
(
HPA-5) as catalyst at 90°C and 20 bar O pressure. Experiments with C-labelled glucose
2
confirm unambiguously that glucose is the only source of the observed formic acid and
methyl formate formation under the applied oxidation conditions. Our results demonstrate a
very astonishing solvent effect for the POM-catalysed glucose oxidation. In comparison to
earlier work, a step-change in product yield and selectivity is achieved by applying an
alcoholic reaction medium. The extremely high combined yields of formic acid and methyl
formate greatly facilitate product isolation as low-boiling methyl formate (bp= 32°C) can
simply be isolated from the reaction mixture by distillation.
Keywords:
Biomass oxidation, polyoxometalates, methanol, formic acid, methyl formate
1

Green Chemistry
Page 2 of 20
View Article Online
DOI: 10.1039/D0GC01169J
Introduction
So far, the industrial production of formic acid uses fossil resources. The most common
pathway for industrial formic acid production is carbonylation of methanol followed by
hydrolysis of the resulting methyl formate. This process currently covers 90% of the global
formic acid production with a total capacity of 770 kton/a (2014) [1]. Alternative pathways for
formic acid production include the electrochemical reduction of CO [2-4] and the catalytic
2
hydrogenation of CO [5]. Both processes suffer from scale-up problems and operation
2
challenges as well as thermodynamic restrictions in the case of the thermocatalytic
conversion [6, 7].
Formic acid is a broadly used industrial chemical. Prominent application examples include its
use as feed additive, for leather tanning, as drilling fluid, as acidifier, as cleaning agent or as
anti-icing substance for roads and airport runways [8, 9]. Formic acid is also discussed as
hydrogen and as CO storage molecule. Its catalytic decomposition has been shown to
deliver hydrogen and carbon dioxide in excellent selectivity [10, 11], and even the use as fuel
for direct formic acid fuel cells (DFAFCs) has been proposed [12-14]. Thermal decomposition
of formic acid in the presence of an acid delivers CO in high selectivity. Thus, depending on
the decomposition conditions, formic acid can serve as precursor for syngas production in
various H /CO ratios, depending on the applied process scheme [15].
2
In general, biomass is a promising feedstock for the production of high-value chemicals with
high oxygen content [16]. Glucose, in particular, has been examined in detail as starting
material for oxidation reactions leading to formic acid, acetic acid, glycolic acid or gluconic
acid [17]. Several recent studies demonstrate the oxidative conversion of carbohydrate
biomass into formic acid [18-21]. Typically, a two step procedure is applied: First, acid
hydrolysis of the biomass leads to the formation of monosaccharides, such as e.g. glucose,
fructose, or xylose. Second, C-C bond cleavage by catalytic oxidation produces formic acid
[
22]. Vanadium-containing catalysts, for instance NaVO or VOSO or polyoxometalates
3 4
(POM), have been shown to exhibit activity for the second reaction step, but formic acid
yields are low to date [23-28]. To improve the performance, strong mineral acids such as
H SO or HCl were applied to reduce the pH-value of the reaction media to support
2
4
dissociation and re-oxidation of the catalyst [29-31].
More specific, the conversion of glucose to formic acid via oxidation in aqueous media has
been successfully promoted by various vanadium-based Keggin-type POM catalysts
(Scheme 1) [32-36]. The applied catalysts activate oxygen for the conversion and are re-
oxidized by molecular oxygen after each reaction cycle. The POMs provide strong Brønsted
acidity, high proton mobility, fast multi-electron transfer, high solubility in various solvents,
and excellent resistance against hydrolytic or oxidative degradations [37, 38]. A remarkable
2

Page 3 of 20
Green Chemistry
View Article Online
DOI: 10.1039/D0GC01169J
feature of this oxidation process is that formic acid and CO are the only products when the
2
oxidation process is completed. At full glucose conversion, formic acid yields of up to 60%
can be achieved, a value that is further improved to 85% by working in a water/1-heptanol
biphasic system with in situ product extraction [39] Vanadium-based POM catalysts are also
active with other starting materials, such as pomace or humins. However, formic acid yields
are lower with these more complex substrates, e.g. 55% in case of pomace oxidation [40, 41]
Scheme 1: POM catalysed oxidation of glucose to formic acid in aqueous media [29, 36].
In recognition of the significant advantages offered by the POM-catalysed, selective biomass
oxidation, some of us have developed the so-called OxFA-process [22, 32-35]. This
continuous process applies H [PV Mo O ] (HPA-5) at low pressure (≤ 20bar) of molecular
8
5
7
40
oxygen or synthetic air as the oxidant in aqueous solution. The process can be applied to a
broad range of biogenic raw material including a lot of complex water-insoluble biomass
nd
rd
compounds of the 2 and 3 generation if para-toluenesulfonic acid (p-TSA) is used as an
additive to assist feedstock hydrolysis [32, 41] Despite its versatility, the classical OxFA
process – as all other POM catalysed conversion of biomass to date - suffers from a
maximum formic acid yield of 60% in a monophasic reaction system. This means that ca.
4
0% of the carbon present in the biomass substrate is converted to CO [17, 22].
2
In this contribution we describe the incidental and very remarkable discovery that a simple
change in the reaction medium of the POM-catalysed glucose oxidation leads to a step-
change in performance. In detail, we report a remarkable influence of methanol as solvent
onto the formic acid selectivity in glucose oxidation at 20 bar initial O and 90 °C reaction
2
temperature.
3

Green Chemistry
Page 4 of 20
View Article Online
DOI: 10.1039/D0GC01169J
Results and Discussion
Solvent selection
Our initial intention was to study the influence of the reaction medium on the POM-catalysed
glucose oxidation reaction. We selected solvents of interest according to the following
requirements: i) The applied solvent should dissolve the HPA-5 catalyst and the substrate
glucose at room temperature; ii) the mixture should be liquid at room temperature; iii) the
solvent must be stable under the oxidative reaction conditions (for all stability tests of the
applied solvents see Supporting Information, Table S1); iv) the solvent should enable straight-
forward isolation of the desired formic acid product. In the light of these requirements the
glucose oxidation was performed in methanol, ethanol, n-propanol and n-butanol, and the
results were compared to the original reaction in water. The results are shown in Figure 1.
Figure 1: Solvent screening. Reaction conditions: 1 mmol glucose, 0.1 mmol HPA-5 catalyst
dissolved in 10 mL solvent, 20 bar initial O , 90 °C, 24 h, 1000 rpm.
2
All systems achieved full glucose conversion within 24 h reaction time. It is important to note
that for all experiments using alcoholic solvents, the organic acids formed during oxidation
were present in the reaction mixture as either free acids or esterification products with the
respective alcohol solvent. With all alcoholic solvents, formic acid and its respective ester were
formed as the dominant product. Very surprisingly, however, the yield in methyl formate
exceeded 99% in the case of methanol being the solvent. In this case neither CO nor CO2
were detectable in our gas phase analysis by gas chromatography. In contrast, the combined
yield of formic acid and the respective esters using C -C alcohols was only between 30-50%,
2
4
significantly lower than the aqueous benchmark system. Moreover, in ethanol, n-propanol and
n-butanol as reaction solvent, significant amounts of CO , CO and acetic acid (and its
2
respective esters) were found as side products. The carbon mass balance (expressed as total
carbon yield TOC, for details see Equation (3) in the Supporting Information) could be closed
4

Page 5 of 20
Green Chemistry
View Article Online
to over 95% for each experiment. Blank experiments without the HPA-5 catalyst didn’t show
DOI: 10.1039/D0GC01169J
any product formation in any of the solvents under investigation.
Labelling experiments
In order to further elucidate the outstanding catalytic performance in methanol, the reaction in
this solvent was investigated in more detail. Our aim was to unambiguously prove that the
formic acid equivalent in the formed methyl formate origins solely from glucose oxidation and
not from oxidation of the solvent methanol followed by esterification. In order to address these
13
questions we performed oxidation experiments with C labelled glucose. Scheme 2 shows the
13
12
expected reaction for the combination of C labelled glucose and C methanol.
Scheme 2: POM-catalysed oxidation of 13C labelled glucose in 12C methanol as solvent –
expected reactivity for the case of glucose serving as sole source of the formic acid equivalent
formed.
The analysis of the GC-MS data of the respective reaction is shown in Figure 2. Four
experiments were performed: Figure 2A shows the GC-MS for the reaction product of the
benchmark experiment with 12C glucose in 12C methanol indicating the mass of 60 g/mol for
12
the exclusively formed C methyl formate. Figure 2 B shows the mass spectrum of the reaction
of glucose in a 9:1 12C/ C methanol mixture indicating only 12C methyl formate formation.
Figure 2C shows the GC-MS for the same product but this time obtained from a 1:1 mixture of
13
1
2
13
12
C glucose and C glucose in C methanol. As a result, a roughly 1:1 mixture of the masses
0 and 61 g/mol was found for the reaction product. Figure 2D, finally, shows the GC-MS for
6
the main product in the experiments in which solely 13C glucose was used as the feedstock.
As expected, the product with the mass of 61 g/mol was found almost exclusively, the
13
remaining traces of 60 g/mol may be explained with the fact that the applied C glucose was
only of 99% isotopic purity. No doubly labelled product of mass 62 g/mol was detected in any
of the samples indicating that the formic acid unit origins from glucose while the methyl ester
group origins uniquely from the solvent as indicated in Scheme 2. Consequently, our labelling
experiments provide convincing evidence that methanol is not the source of formic acid in the
applied reaction system. Instead, the methanolic reaction medium changes in a very
remarkable manner the selectivity of the HPA-5 catalyst in glucose oxidation, avoiding almost
5

Green Chemistry
Page 6 of 20
View Article Online
all side product formation and leading to an extremely high methyl formate selectivity. Note
DOI: 10.1039/D0GC01169J
that TOC almost reached 100% in all experiments.
A)
B)
C)
D)
Figure 2: Mass spectra of the main product in glucose oxidation experiments as analysed by
1
2
12
GC-MS: A) Mass spectrum of the main product obtained from C-glucose oxidation in
C
1
2
methanol, B) Mass spectrum of the main product obtained from C-glucose oxidation in a 9:1
12
13
C/ C methanol mixture, C) Mass spectrum of the main product obtained from oxidising a
C/ C-glucose mixture (1:1) in C methanol, D) Mass spectrum of the main product obtained
from C-glucose oxidation in C methanol. Reaction conditions: 1 mmol C-glucose/ C-
12
13
12
13
12
13
12
glucose, 0.1 mmol HPA-5 catalyst dissolved in 10 mL methanol, 20 bar initial O , 90 °C, 24 h,
2
1000 rpm.
Mechanistic investigations – the pathway from glucose to methyl formate
To learn more about the oxidation pathway leading to methyl formate formation in almost
perfect selectivity in the methanolic HPA-5 reaction solution, we performed a time-resolved
glucose conversion experiment (Figure 3). Already after one hour reaction time, full conversion
of glucose was achieved. Erythrose, glyoxal and glycolaldehyde were detected as reaction
intermediates while around 50% of the initial glucose was already converted to formic
6

Page 7 of 20
Green Chemistry
View Article Online
acid/methyl formate. With increasing reaction time, erythrose and glyoxal decreased faster
DOI: 10.1039/D0GC01169J
than glycolaldehyde which is a hint that the latter is a secondary intermediate on the way to
formic acid. After 24 hours, all intermediates were fully oxidised to formic acid that forms methyl
formate under the reaction conditions in almost quantitative manner. The mass balance was
always closed to 100% in the liquid phase and no gaseous side products were detected
throughout the entire course of the experiment.
Figure 3: Glucose oxidation catalysed bv HPA-5 in methanol as solvent – product formation
over reaction time. Reaction conditions: 1 mmol glucose, 0.1 mmol HPA-5 catalyst dissolved
in 10 mL methanol, 20 bar initial O , 90 °C, 1-24 h, 1000 rpm.
2
Based on these time-resolved experiments, we propose the following reaction mechanism for
the HPA-5-catalysed glucose oxidation in methanolic solution to formic acid and further
esterification to methyl formate (Scheme 3). In the first step, glucose undergoes an oxidative
C-C bond cleavage under water elimination to form one molecule of erythrose and one
molecule of glyoxal. Hereafter, erythrose is cleaved by a similar oxidative reaction step to form
one molecule glyoxal and one molecule glycolaldehyde. All C -intermediates are subsequently
2
oxidised to formic acid. The latter is rapidly transformed into methyl formate via esterification
with methanol. The esterification is fast and almost complete due to the enormous excess of
the solvent methanol. Obviously, no undesired total oxidation of any carbon intermediate to
CO or CO takes place in methanol. The amount of water that is overall produced (six
2
equivalents per glucose substrate) correspond well to the amount of water that we have
measured after full conversion in our Karl-Fischer titration experiments.
7

Green Chemistry
Page 8 of 20
View Article Online
DOI: 10.1039/D0GC01169J
Scheme 3: Proposed reaction mechanism for the HPA-5-catalysed glucose oxidation to methyl
formate in methanol.
Methanol stability under oxidative conditions
To critically evaluate the stability of the solvent methanol under the reaction conditions further,
we performed methanol oxidation experiments with the same HPA-5 catalyst under the same
conditions but without adding glucose. For comparison, we also performed an experiment
under nitrogen atmosphere to clarify whether oxygen is transferred from the POM catalyst to
the methanol. The resulting 13C-NMR spectra (Figure 4) and headspace GC-MS
measurements (see Supporting Information, Figures S1 and S2) show that under oxidative
conditions small amounts of dimethoxymethane are formed from methanol at 90 °C and
110 °C, respectively. At 110 °C under oxygen atmosphere small amounts of dimethylether are
formed in addition. Both products are not detectable under nitrogen atmosphere.
Figure 4: Methanol stability under oxidative conditions in presence of HPA-5. Reaction
conditions: 0.1 mmol catalyst, 10 g MeOH, 20 bar initial O or N pressure, 24
2
2
To clarify the role of the HPA-5 catalyst in methanol oxidation, we performed control
experiments without catalyst both under oxygen and nitrogen atmosphere at 90°C.
Interestingly, we could not find neither dimethoxymethane nor dimethylether without the POM
8

Page 9 of 20
Green Chemistry
View Article Online
DOI: 10.1039/D0GC01169J
catalyst under O atmosphere but methoxymethanol (Figure 5). As expected, no products
2
could be detected under nitrogen atmosphere without catalyst.
Figure 5: 13C-NMR spectra of the product mixture obtained from methanol oxidation at 90 °C
without catalyst. Further reaction conditions: 10 g MeOH, 20 bar, 24
Based on a published mechanism for gas-phase methanol oxidation [42], we propose that
formaldehyde must be formed as an intermediate on the way to methoxymethanol (Scheme
4), although we couldn’t detect this compound in our experiments due to its fast consecutive
reaction with methanol. To verify this hypothesis, we performed control experiments using all
suggested intermediates and products as substrates under the same oxidative reaction
13
conditions of 90°C and 20 bar oxygen pressure. The corresponding C-NMR spectra can be
found in the Supporting Information (Figures S6-S8).
As expected, the use of dimethoxymethane and dimethylether as substrates did not show any
conversion under the applied reaction conditions in absence of the catalyst confirming that
these compounds are stable under the reaction conditions in methanolic solution. Using
formaldehyde resulted in the formation of methoxymethanol as the only product, as expected.
Compared to the experiment in pure methanol this compound was formed in much larger
quantities with added formaldehyde confirming that formaldehyde is indeed an intermediate
on the way to methoxymethanol (without HPA-5 catalyst) and dimethoxymethane (with HPA-5
catalyst). Based on these results, we propose the following reaction scheme for methanol
oxidation in the absence of glucose in our HPA-5 reaction system (Scheme 4). Note that these
experiments were only performed on a qualitative manner. Therefore we cannot comment on
the mass balance here.
9

Green Chemistry
Page 10 of 20
View Article Online
DOI: 10.1039/D0GC01169J
Scheme 4: Reaction pathway of oxidative methanol conversion in the HPA-5 catalysed
reaction system in the absence of glucose.
Product composition depending on water content
As six moles of water are formed per conversion of one mol of sugar to six moles of methyl
formate (see Scheme 3), we were interested to understand in more detail the role of small
amounts of water in our methanolic reaction system. In the following set of experiments, we
therefore varied the amount of water from zero (pure methanol) to 100% (pure water). The
results are depicted in Figure 6.
As expected, glucose conversion is not a strong function of the water content, the reaction
converts all glucose in pure methanol, pure water and most of the solvent mixtures (see Figure
6
A). The amount of CO formed shows, however, a very interesting dependency on the actual
2
water content in the methanolic reaction mixture. Up to 90% water content, the degree of
undesired CO formation stays as low as 3.3% and only increases to the typical values of 43%
2
with pure water as a solvent. This shows impressively that already a small amount of methanol
in the reaction system has a quite significant influence on the suppression of the unwanted
CO production during glucose oxidation.
2
1
0

Page 11 of 20
Green Chemistry
View Article Online
1
00
50 DOI: 10.1039/D0GC01169J
A)
80
60
40
20
0
40
30
20
10
0
Glucose
CO2
0
20
40
60
80
100
Water content / %
1
00
B)
methyl formate
formic acid
80
60
40
20
0
0
20
40
60
80
100
Water content / %
Figure 6: Influence of water in the HPA-5-catalysed oxidation of glucose in methanolic solution:
A) Glucose conversion and CO yield as function of the water content. B) Methyl formate and
2
formic acid yields as function of the water content. Reaction conditions: 1 mmol glucose,
0
.1 mmol HPA-5 catalyst dissolved in 10 mL solvent, 20 bar initial O , 90 °C, 24 h, 1000 rpm.
2
The yield of methyl formate depends also on the water content of the reaction mixture and
decreases linearly from almost 100% yield in pure methanol to almost 0% yield in pure water,
as expected (Figure 6B). The yield of formic acid depends also strongly on the water content.
Without water the formic acid yield is almost zero as it immediately reacts with methanol to
methyl formate. The higher the water content, the higher is the formic acid yield. The maximum
formic acid selectivity is found at 90% water with 78 % (formic acid yield of 72%). At even
higher water content, the selectivity decreases slightly due to the increasing CO formation.
2
Moreover, again no intermediates were left after 24 h reaction time whereby CO2 and CO
1
1

Green Chemistry
Page 12 of 20
View Article Online
could be detected as byproducts in the gas phase (for details see Supporting information,
DOI: 10.1039/D0GC01169J
Table S2).The carbon mass balance was always closed to >99% in this study.
Reaction rate comparison of HPA-5-catalysed glucose oxidation in water vs. methanol
Finally, we were interested to compare the kinetics of our new HPA-5-catalysed glucose
oxidation in methanol with the same reaction in the state-of-the-art aqueous system. Therefore,
we performed for both reaction systems time-resolved experiments with sampling of both the
gas and the liquid phase over the course of the reaction. In order to further approach reaction
conditions of technical interest, we performed these comparison experiments at reduced
oxygen pressure using 5 bar synthetic air (21 vol% oxygen), i.e. a twenty-fold reduced oxygen
partial pressure compared to our standard reaction conditions, at 90°C reaction temperature.
Figure 7 shows that the glucose concentration drops much faster in methanol compared to the
reaction in water. The fact that under these relatively oxygen-lean conditions a full glucose
oxidation is possible in methanol while only 40 % of the glucose is oxidised in water
demonstrates that re-oxidation of the HPA-5 catalyst is possible at lower oxygen partial
pressure in methanol compared to water. The fact that re-oxidation of the HPA-5 catalyst at
the here applied low O partial pressure of only 1.02 bar is indeed possible in methanol as
2
solvent is visually confirmed by the different colour of the reaction solution. While the reaction
mixture in water is dark blue at the end of the reaction in water (indicating vanadium in its +IV
oxidation state), it stayed in its original orange colour (indicating vanadium in its +V oxidation
state) throughout the experiment in methanolic solution. This interpretation was further
confirmed by our 51V-NMR spectroscopy measurements (see Figures S9-S10 in the
Supporting Information). Thus, the HPA-5-catalysed glucose oxidation is not only significantly
faster in methanol than in water (for detailed experimental data see Supporting Information
Tables S3, S4 and Figure S11, respectively). It is also feasible at significantly lower oxygen
partial pressure which is another big advantage. The reason can be found in the higher oxygen
solubility in methanol compared to water (2.15 mM in methanol compared to 0.26 mM in water
at 20°C and 1 atm) [44, 45].
The use of our new methanolic reaction system enables operation with air (instead of pure
oxygen) and/or at significantly lower total pressure. Both aspects reduce equipment and
operation cost of a potential glucose oxidation process and add significantly to the economic
attractiveness of our new methanolic oxidation process.
1
2

Page 13 of 20
Green Chemistry
View Article Online
0
.12
.1
DOI: 10.1039/D0GC01169J
0
0
0
0
0
.08
.06
.04
.02
0
0
5
10
15
20
time/h
Figure 7: Glucose concentration over time in water (blue) and methanol (orange). Reaction
conditions: 1 mmol glucose, 0.1 mmol HPA-5 catalyst dissolved in 10 mL solvent, 5 bar
synthetic air (21 vol% O ), 90 °C, 1-24 h, 1000 rpm.
2
The initial reaction rates are very fast. Consequently, we have further investigated the influence
of possible mass transfer limitations. In detail, we have performed variations of oxygen partial
pressure and stirring speed (for details see Supporting Information, Figures S13 and S14).
Both studies confirmed that sufficient oxygen is available for catalyst re-oxidation in methanol
under the applied reaction conditions whereby it is limited in water.
It remains to be noted that we could already demonstrate the stability of our HPA-5 catalyst
system in pure methanolic solution as evidenced by three consecutive catalytic runs without
any catalyst treatment or product removal in-between and without any loss in performance
(see Supporting information, Figure S12).
Experimental section
Materials
D(+)-glucose (99.4%) was purchased from Merck KGaA and used without further purification.
13
The C-glucose (99%) was obtained from Eurisotop and used as received. A list of all
solvents previously tested in the POM catalysed glucose oxidation and the corresponding
sources can be found in the Supporting Information, Table S1. The solvents used in this
work, i.e. demineralized water, methanol (99.0%), ethanol (99.5%), n-propanol (98%) and n-
butanol (99.9%) were all purchased from Merck KGaA. For our mechanistic investigations
formic acid (99.5%) from VWR, methyl formate (97%) from Alfa Aesar, dimethoxymethane
(98%) from Alfa Aesar, and formaldehyde (50 wt% aqueous solution) from Merck KGaA were
used.
1
3

Green Chemistry
Page 14 of 20
View Article Online
DOI: 10.1039/D0GC01169J
All gases applied for the described experiments (oxygen 5.0, helium 5.0, nitrogen 5.0, and
synthetic air) were purchased from Linde AG. Argon 5.0 and hydrogen 5.0 from Linde AG
were additionally used as carrier gases for the GC analysis.
For the synthesis of the HPA-5 (H PV Mo O ) catalyst, V O (99.5%) from Alfa Aesar, MoO
3
8
5
7
40
2
5
(
99%) from Alfa Aesar, H PO (25 wt% aqueous solution) from AppliChem and H O (30 wt%
3
4
2
2
aqueous solution) from Alfa Aesar were used.
Catalyst preparation and characterisation
Our synthesis of Keggin heteropolyoxometalates was based on a method published by
Odyakov and Zhizhina [43] and was optimised for the synthesis of HPA-5 by Albert et al
.
[33]. This procedure consists of a two-step synthesis, with the first step producing the stable
vanadium precursor H [PV O ]. HPA-5 is obtained in the second step by synthesising the
9
14 42
molybdenum precursor (HPA-0) followed by subsequent addition of the vanadium precursor.
For the synthesis of the vanadium precursor, V O is first suspended in about 8°C cold water
2
5
(<10°C) with constant stirring. Slow dropwise addition of 30 wt% hydrogen peroxide solution
at lower than 10°C results in the dark red coloration of the vanadium (V) peroxo species
formed. Reaction induced warming to 30-35 °C (due to removing of the cooling bath) causes
the release of O accompanied by a colour change to brown orange by forming the unstable
2
H [V O ]. When the evolution of the gas stops, the stable dark brown H [PV O ] vanadium
6
10 28
9
14 42
precursor is obtained by slow addition of 25 wt% phosphoric acid.
To prepare the molybdenum precursor (HPA-0), MoO is first suspended in water at room
3
temperature with constant stirring. The subsequent addition of 25 wt% H PO and heating to
3
4
the boiling point of about 140 °C leads to the formation of the yellow H [PMo O ] precursor
3
12 40
(HPA-0). Thereafter, the reaction mixture is further boiled while adding portions of the
H [PV O ] precursor from the first reaction step. Depending on the stoichiometric ratio of
9
14 42
the starting materials, the finished polyoxometalate with the corresponding degree of
substitution n is formed. The reaction mixture is then concentrated by evaporation and
cooled to room temperature. This process is followed by a dead-end filtration of the
concentrated mixture to remove unreacted starting materials (V O or MoO ) or other
2
5
3
impurities. The remaining filtrate is then dried by rotary evaporation at 55 °C in the water bath
under a vacuum (the end pressure is ca. 50 mbar) to obtain a crystalline solid. The obtained
crystals are finally dried overnight at 30 °C and 50 mbar vacuum. As a result, HPA-5 is
obtained as a powdery, crystalline, reddish-brown solid in high yields of over 90%.
The characterisation of the so-prepared HPA-5 catalyst was carried out by ICP-OES, FT-IR,
31
51
TGA, P- and V-NMR spectroscopy. Elemental analysis by ICP-OES (Perkin Elmer
Plasma 400) resulted in a ratio of P/V/Mo of 1/5/7. Infrared spectroscopy of the solid catalyst
1
4

Page 15 of 20
Green Chemistry
View Article Online
DOI: 10.1039/D0GC01169J
was recorded using a Shimadzu Prestige-21 FT-IR spectrophotometer in combination with a
-1
diamond crystal ATR unit from L.O.T., Golden Gate in the range of 4000 to 400 cm . The
spectra showed the expected stretching vibrations for the Keggin-oxoanion (see Supporting
Information, Figure S3). Thermogravimetric analyses (TGA) were conducted using a
SETSYS-1750 CS Evolution from SETARAM Instruments resulting in the two-step loss of
water that is typical for Keggin-heteropolyacids (see Supporting Information, Figure S4). 11
wt% loss of crystallisation water and 5 wt% loss of constitutional water resulted in 14 moles
of hydrated water per mole of HPA-5. Using this value, a molar mass of MHPA-5=1857 g/mol
can be calculated. The NMR spectra were recorded using a Jeol ECX-400 MHz
31
spectrometer (9.4 Tesla) at 293 K. The P NMR spectra were recorded at 161.98 MHz with
12 scans in the range of −7 to +5 ppm, resulting in a resolution of 0.24 Hz. The field
frequency stabilisation was locked to deuterium by placing a coaxial inner tube with 1 wt%
5
51
H PO in D O into a 5 mm tube containing the sample. The V NMR spectra were measured
3
4
2
with 2024 scans in a range of −580 to −460 ppm with an excitation frequency of 105.25 MHz
and a resolution of 0.77 Hz. The field frequency stabilisation was locked to deuterium by
placing a coaxial inner tube with D O into a 10 mm tube containing the sample. The
2
corresponding spectra are shown in Figure S5 in the Supporting Information. The
experimental data confirmed that HPA-5 was synthesized successfully.
Experimental setup and procedure for the glucose oxidation experiments
The catalytic oxidation experiments were performed in a dedicated screening device, a
parallel autoclave reactor system consisting of ten parallel 20 mL Hastelloy C-276
autoclaves. All ten autoclaves were connected individually to an oxygen/nitrogen supply line.
For safety reasons, a rupture disk was installed with a maximum burst pressure of 100 bar.
All pipes, fittings, and valves were made of stainless steel 1.4571. A PTFE sealing was
located between headcover and vessel of the reactor for protection against pressure leaks.
The reactors were placed into a heating hole on the top of a magnetic stirrer system. The
system works with magnetic bars placed in each of the ten 20 mL autoclaves and can be
heated up to 200°C. The kinetic experiments were carried out in a 600 mL Hastelloy C276
autoclave equipped with a gas entrainment impeller.
It should be generally noted that all oxidation experiments involving oxidisable solvents, such
as methanol, ethanol, n-propanol or n-butanol come with some intrinsic risk of thermal
runaway if the oxidation reaction is not carefully controlled. In the here described case all
autoclaves used were constructed to withstand a complete oxidation of all organic material in
the reactor or a complete consumption of the oxygen provided to the reaction.
For a typical glucose oxidation experiment, 0.1 mmol of HPA-5 catalyst and 1 mmol of
glucose were filled into the 20 ml autoclave, subsequently solvent (10 mL) was added. A
1
5

Green Chemistry
Page 16 of 20
View Article Online
DOI: 10.1039/D0GC01169J
magnetic bar was added into the reactor for mixing. Before closing the reactor, a PTFE
gasket ring was inserted between head and vessel. The oxygen supply tube was coupled
with the head part of the reactor. The off-gas of the rupture disk was also connected to the
venting line. Then the pressure transmitter were connected to the system for pressure
monitoring. The reactor was purged with oxygen (10 bar) twice and then pressurised with
oxygen. The stirrer was set to 300 rpm and the heating was switched on. When the oxidation
temperature of 90 °C was reached, the reaction pressure was adjusted and the stirrer was
set to 1000 rpm. The reactor was kept at the oxidation conditions for 24 h. During this time
the reactor pressure and the temperature were monitored and recorded.
For the kinetic experiments in the 600 ml Hastelloy C autoclave, water, glucose and catalyst
were charged into the vessel. The system was purged three times with oxygen to remove the
remaining air, the stirrer speed was set to 300 rpm and heating was switched on. As soon as
the desired temperature of 90°C was reached, the oxygen pressure was adjusted to the
desired value. At the beginning of the experiment, a liquid and a gas sample were taken in
order to determine the initial substrate and product concentrations. Then the stirrer speed
was set to 1000 rpm to start the reaction. Liquid and gas samples were taken at regular
intervals to follow the reaction. To interrupt the oxidation reaction during sampling, the gas
entrainment was stopped by reducing the stirrer speed to 300 rpm. Each kinetic experiment
was carried out for 24 h reaction time.
For the catalyst stability, three consecutive runs with only adding fresh substrate (1 mmol of
glucose) after each run were performed in a similar manner as described before.
Product analysis
After cooling down, all gaseous products (CO, CO ) were analysed with a gas chromatograph
2
Varian GC 450-TCD-FID equipped with a Shin Carbon ST column (2 m x 0.75 mm ID). The
yields of CO and CO were determined by calculating n(CO )/n(C atoms feedstock) and
2
2
n(CO)/n(C atoms feedstock), respectively. The liquid phase was analysed by 13C-NMR
spectroscopy (Jeol ECX-400 MHz Spectrometer, 9,4 Tesla) at 20°C as well as high-
performance liquid chromatography (HPLC, Jasco instrument with a Shodex SUGAR SH1011
300 mm x 8 mm column). The calculation of glucose conversion and product selectivity was
carried out on a carbon basis and is shown in detail in the Supporting Information, Equations
13
S1 and S2. The C-labelling experiments were performed using a GC-MS Agilent 7890B GC/
977A MSD-system with HP-5ms Ultra Inert column (30 m x 250 µm x 0.25 µm) and a single
5
quadrupole MS. For double checking of the 13C-labelling measurements, the samples were
analysed in addition by headspace GC-MS (Shimadzu QC 2010/ QP2010 SE GCMS-system
equipped with CTC combi PAL headspace and separated with CP-Sil PONA CB column, 50
m x 210 µm x 0.5 µm) with liquid injection method. Moreover, water content of the alcoholic
1
6

Page 17 of 20
Green Chemistry
View Article Online
solutions was quantified by Karl-Fischer Titration (KFT) using a Coulometer 756 from Metrohm.
DOI: 10.1039/D0GC01169J
By electrochemical means, iodine is generated directly in the iodide-containing electrolyte. The
generator electrode without diaphragm was used for producing iodine at the anode. Karl
Fischer reagent was supplied by Merck KGaA, with a type of “CombiCoulomat fritless” for
coulometric water. With the limitation of the method and the Karl Fischer reagent used, it is
suitable for a range of water content between 0.1 and 20 wt.%
1
7

Green Chemistry
Page 18 of 20
View Article Online
DOI: 10.1039/D0GC01169J
Conclusions
We have discovered a very remarkable solvent effect in the polyoxymetalate-catalysed
oxidation of glucose. Replacing the well-investigated aqueous reaction system by a
methanolic solution, we demonstrate HPA-5-catalysed glucose oxidation under mild
conditions (90 °C) in almost perfect yield to methyl formate (> 99% yield). Undesired side
products that have been typically found in the traditional aqueous oxidation system, such as
CO or CO , can be completely avoided in this way. As the mass balance in the here-
2
presented glucose oxidation in methanol is perfectly closed, we claim a new option for
13
glucose valorisation in hitherto unmatched efficiency. Experiments with labelled C glucose
have unambiguously proved that glucose is converted completely to formic acid which
subsequently forms methyl formate by esterification in the excess of methanolic reaction
medium. Stability tests of the solvent methanol (under the identical oxidation conditions but
without added glucose) showed that methanol is converted to small amounts of
dimethylether and dimethoxymethane but not to methyl formate.
Based on the detailed analysis of all reaction intermediates formed during glucose oxidation,
we are able to propose a conclusive reaction pathway from glucose to methyl formate via
formic acid. As water forms in the reaction system during esterification from the latter to
methyl formate, we have also studied methanol/water mixtures as reaction medium for HPA-
5-catalysed glucose oxidation. Remarkably, a 90:10 water/methanol reaction system still
shows a very pronounced enhancement of combined formic acid+methyl formate selectivity
and a very effective suppression of gaseous side products. Higher water contents shift the
esterification equilibrium towards the free acid, as expected. A highly important finding is that
glucose oxidation in methanol benefits from a faster re-oxidation of the HPA-5 catalyst. This
enables effective re-oxidation at lower oxygen partial pressures and allows to work with air
as oxidant at pressures as low as 5 bar. In conclusion, our findings constitute a new,
technically highly attractive approach to convert glucose in perfect selectivity to methyl
formate. As the latter is a valuable intermediate this opens new attractive routes for biomass
valorisation, in particular if the here presented oxidation catalysis can be extended in the
future to more technical carbohydrate feedstocks, such as industrial sugar solutions,
hemicellulose fractions of the pulping process or cellulose hydrolysates.
Acknowledgements
The authors acknowledge financial support from the Cluster of Excellence “Engineering of
Advanced Materials (EAM)”. CK acknowledges in addition support of the Office of the Civil
Service Commission (OCSC) scholarship of the Republic of Thailand. Moreover, we thank
Birgit Brunner from Bayreuth University for performing GC-MS measurements and Peter
Schulz from CRT, University Erlangen-Nuremberg for Headspace-GC measurements.
1
8

Page 19 of 20
Green Chemistry
View Article Online
DOI: 10.1039/D0GC01169J
References
1
.
J. Hietala, A.V., P. Johnsson, I. Pollari, W. Reutemann, H. Kieczka, Ullmann’s Encycl. Ind. Chem.
Weinheim, Germany, Vol. 1. 2016.
2
3
4
5
6
7
.
.
.
.
.
.
Ikeda, S., T. Takagi, K. Ito, Bulletin of the Chemical Society of Japan, 1987, 60, 2517-2522.
X. Lu, D. Y. C. Leung, H. Wang, M. K. H. Leung, J. Xuan, ChemElectroChem, 2014, 1, 836-849.
R. Kortlever, I. Peters, S. Koper, M. T. M. Koper, ACS Catal., 2015, 5, 3916-3923.
J. C. Tsai, K. M. Nicholas, J. Am. Chem. Soc., 1992, 114, 5117-5124.
A. S. Agarwal, Y. Zhai, D. Hill, N. Sridhar, ChemSusChem, 2011. 4, 1301-1310.
J. T. Feaster, C Shi, E. R. Cave, T. Hatsukade, D. N. Abram, K. P. Kuhl, C. Hahn , J. K. Nørskov, T.
F. Jaramillo, ACS Catal., 2017. 7, 4822-4827.
8
.
M. E. M. Berger, D. Assenbaum , N. Taccardi , E. Spiecker and P. Wasserscheid, Green Chem.,
2
011, 13, 1411-1415.
9
1
.
0.
Y. Himeda, Eur. J. Inorg. Chem., 2007, 25, 3927-3941
A. Boddien, B. Loges, F. Gartner, C. Torborg, K. Fumino, H. Junge, R. Ludwig, M. Beller, J. Am.
Chem. Soc., 2010, 132, 8924-8934
1
1
1
1.
2.
3.
A. Boddien, B. Loges, H. Junge, M. Beller, ChemSusChem, 2008, 1, 751–758.
G. A. El-Nagar, M. A. Hassan, I. Lauermann, C. Roth, Scientific reports, 2017, 7, 1-11.
M. Weber, J.-T. Wang, S. Wasmus, R. F. Savinell, Journal of the Electrochemical Society, 1996,
1
43, L158.
1
1
4.
5.
N. V. Rees, R.G. Compton, Journal of Solid State Electrochemistry, 2011, 15, 2095-2100.
J. Albert, A. Jess, A. Kern, F. Pohlmann, K. Glowienka, P. Wasserscheid, ACS Sustain. Chem. Eng.,
2
016, 4, 5078-5086].
1
1
1
1
2
6.
7.
8.
9.
0.
G. W. Huber, S. Iborra, A. Corma, Chem. Rev., 2006. 106, 4044-4098.
Z. Zhang, G.W. Huber, Chem. Soc. Rev., 2018, 47, 1351-1390.
F. Jin, J. Yun, G. Li, A. Kishita, K. Tohji, Enomoto, Green Chem., 2008, 10, 612-615.
R. Wölfel, N. Taccardi, A. Bösmann, P. Wasserscheid, Green Chem., 2011, 13, 2759–2763.
H. Mehdi, V. Fabos, R. Tuba, A. Bodor, L. T. Mika, I. T. Horvath, Topics in Catalysis, 2008, 48,
4
9–54.
2
2
2
1.
2.
3.
J. Xu, Y. Zhao, H. Xu, H. Zhang, B. Yu, L. Hao, Z. Liu, Appl. Catal. B, 2014, 154–155, 267-273.
P. Preuster, J. Albert, Energy Technol., 2018, 6, 501-509.
W. Wang, M. Niu, Y. Hou, W. Wu, Z. Liu, Q. Liu, S. Ren, K. N. Marsh, Green Chem., 2014, 16,
2
614–2618.
Z. Tang, W. Deng, Y. Wang, E. Zhu, X. Wan, Q. Zhang, Y. Wang, ChemSusChem, 2014, 7, 1557-
567.
2
4.
1
2
2
2
2
2
3
3
3
3
5.
6.
7.
8.
9.
0.
1.
2.
3.
M. Wang, J. Lu, X. Zhang, L. Li, H. Li, N. Luo, F. Wang, ACS Catal., 2016, 6, 6086-6090.
J. Albert, J. Mehler, J. Tucher, K. Kastner, C. Streb, ChemistrySelect, 2016, 1, 2889-2894.
J. Albert, Faraday Discuss., 2017, 202, 99-109.
D. Voß, H. Pickel, J. Albert, ACS Sustain. Chem. Eng., 2019, 7, 9754-9762.
J. Li, D.J. Ding, L. Deng, Q. X. Guo, F. Fu, ChemSusChem, 2012, 5, 1313-1318.
M. Niu, Y. Hou, S. Ren, W. Wang, Q. Zheng, W. Wu, Green Chem., 2015, 17, 335-342
T. Lu, M. Niu, Y. Hou, W. Wu, S. Ren, F. Yang, Green Chem., 2016, 18, 4725-4732
J. Albert, R. Wölfel, A. Bösmann, P. Wasserscheid, Energy Environ. Sci., 2012, 5, 7956-7962.
J. Albert, D. Lüders, A. Bösmann, D. M. Guldi, P. Wasserscheid, Green Chem., 2014, 16, 226-
2
37.
3
3
3
3
3
3
4.
5.
6.
7.
8.
9.
J. Reichert, J. Albert, ACS Sustain. Chem. Eng., 2017, 5, 7383-7392.
S. Ponce, M. Trabold, A. Drochner, J. Albert, B. J.M. Etzold, Chem. Eng. J., 2019, 369, 443-450.
J. Albert, M. Mendt, M. Mozer, D. Voß, Appl. Catal. A, 2019, 570, 262-270.
I. V. Kozhevnikov, Chem. Rev., 1998, 98, 171-198.
W. Deng, Q. Zhang, Y. Wang, J. Chem. Soc. Dalton, 2012, 41, 9817-9831.
J. Reichert, B. Brunner, A. Jess, P. Wasserscheid, J. Albert, Energy Environ. Sci., 2015, 8, 2985-
2
990.
4
0.
S. G. Maerten, D. Voß, M. A. Liauw, J. Albert, ChemistrySelect, 2017, 2, 7296-7302.
1
9

Green Chemistry
Page 20 of 20
View Article Online
DOI: 10.1039/D0GC01169J
4
4
4
4
4
1.
2.
3.
4.
5.
J. Albert, P. Wasserscheid, Green Chem., 2015, 17, 5164-5171.
Y.Zhao, Z. Qin, G. Wang, M. Dong, L. Huang, Z. Wu, W. Fan, J. Wang, Fuel, 2013, 104, 22-27.
V.F. Odyakov, E.G. Zhizhina, React. Kinet. Catal. Lett. 2008, 95, 21-28.
R. Battino, T. R. Rettich, T. Tominaga, Phys. Chem. Ref. Data 1983, 12, 163–178.
M. Quaranta, M. Murkovic, I. Klimant, Analyst 2013, 138, 6243–6245.
2
0