Selective aerobic oxidation of hmf to 2,5-diformylfuran on covalent triazine frameworks-supported ru catalysts

Article abstract of DOI:10.1002/cssc.201403078

The selective aerobic oxidation of 5-hydroxymethylfurfural (HMF) to 2,5-diformylfuran has been performed under mild conditions at 80 8C and 20 bar of synthetic air in methyl t-butyl ether. Ru clusters supported on covalent triazine frameworks (CTFs) allowed excellent selectivity and superior catalytic activity compared to other support materials such as activated carbon, γ-Al₂O₃, hydrotalcite, or MgO. CTFs with varying pore size, specific surface area, and N content could be prepared from different monomers. The structural properties of the CTF materials influence the catalytic activity of Ru/CTF significantly in the aerobic oxidation of HMF, which emphasizes the superior activity of mesoporous CTFs. Recycling of the catalysts is challenging, but promising methods to maintain high catalytic activity were developed that facilitate only minor deactivation in five consecutive recycling experiments.

Full text of DOI:10.1002/cssc.201403078

DOI: 10.1002/cssc.201403078
Full Papers
Selective Aerobic Oxidation of HMF to 2,5-Diformylfuran
on Covalent Triazine Frameworks-Supported Ru Catalysts
[a]
[b]
[a]
Jens Artz, Sabrina Mallmann, and Regina Palkovits*
The selective aerobic oxidation of 5-hydroxymethylfurfural
HMF) to 2,5-diformylfuran has been performed under mild
from different monomers. The structural properties of the CTF
materials influence the catalytic activity of Ru/CTF significantly
in the aerobic oxidation of HMF, which emphasizes the superi-
or activity of mesoporous CTFs. Recycling of the catalysts is
challenging, but promising methods to maintain high catalytic
activity were developed that facilitate only minor deactivation
in five consecutive recycling experiments.
(
conditions at 808C and 20 bar of synthetic air in methyl t-butyl
ether. Ru clusters supported on covalent triazine frameworks
(CTFs) allowed excellent selectivity and superior catalytic activi-
ty compared to other support materials such as activated
carbon, g-Al O , hydrotalcite, or MgO. CTFs with varying pore
2
3
size, specific surface area, and N content could be prepared
Introduction
As fossil resources have been depleted drastically and energy
consumption has increased rapidly during the last century,
chemical processes independent of fossil resources have re-
ceived growing attention in recent years. As a renewable
carbon-containing resource, biomass has been the focus of nu-
merous studies. The goal is to implement an environmentally
benign continuous production of platform chemicals that does
not depend on the diminishing fossil resources to enable sus-
mer industry. As an example, FDCA has been used successfully
as a substitute for terephthalic acid in the production of poly-
[9]
ethylene terephthalate (PET).
DFF can be used as a precursor for pharmaceuticals, antifun-
[10–13]
gal agents, furanic biopolymers, and furan-urea resins.
However, the selective oxidation of HMF to DFF is challenging
because of the high reactivity of the aldehyde function. There-
fore, DFF is transformed easily to 5-formyl-2-furancarboxylic
acid (FFCA) and FDCA (Scheme 1). Additionally, the oxidation
of HMF to the monocarboxylic acid 5-hydroxymethylfuran-2-
carboxylic acid (HMFCA) can occur.
[
1,2]
tainable industrial processes. More recently, several attempts
have been made to transform these biomass-derived platform
chemicals selectively to highly valuable monomers, pharma-
[
2]
ceuticals, and fine chemicals.
-Hydroxymethylfurfural (HMF) is a prominent biomass-de-
The methodology used most commonly to produce DFF se-
[14]
5
lectively is the use of classical oxidants, homogeneous Co
rived platform chemical that can be obtained by the
dehydration of hexoses such as glucose or fruc-
[
3,4]
tose.
HMF can be hydrogenated to the corre-
sponding diols, for example, 2,5-dihydroxymethylte-
[
5,6]
[6]
trahydrofuran,
used as a solvent or monomer, or
even 1,6-hexanediol, a monomer for the production
[
7]
of caprolactone.
Furthermore, several attempts
have been made to oxidize HMF selectively to 2,5-di-
formylfuran (DFF) and 2,5-furandicarboxylic acid
[
8]
(
FDCA). Both are potential precursors for the poly- Scheme 1. Selective aerobic oxidation of HMF to DFF and its byproducts.
[
15]
[
a] J. Artz, Prof. Dr. R. Palkovits
and Mn catalysts, and V-based heterogeneous catalysts, such
as V O /TiO or vanadyl pyridine complexes supported on poly-
Chair of Heterogeneous Catalysis & Chemical Technology
Institut fꢀr Technische und Makromolekulare Chemie
RWTH Aachen University
2
5
2
[16]
(
vinylpyrrolidinone) (PVP). However, most of these catalysts
suffer from low activities, high catalyst-to-substrate ratios and
difficult recyclability because of their homogenous nature or
the leaching of active species. Therefore, these systems are not
suitable for sustainable DFF production.
Worringerweg 1
5
2074 Aachen (Germany)
E-mail: palkovits@itmc.rwth-aachen.de
[
b] S. Mallmann
DWI Leibniz-Institut fꢀr Interaktive Materialien
Forckenbeckstr. 50
Recently, the first Ru-based solid catalysts have been report-
ed for the oxidation of HMF to DFF. In 2011, Ebitani et al.
showed that Ru on hydrotalcite can be used in the one-pot
synthesis of DFF from glucose and fructose to yield 25 and
5
2074 Aachen (Germany)
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201403078.
ChemSusChem 0000, 00, 0 – 0
1
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
[
17]
4
9% of DFF, respectively. More recently the same group ach-
ieved DFF directly in a one-pot reaction from raffinose with
7% yield at 1208C under a flow of O . Antonyraj et al. ach-
Results and Discussion
2
2
Catalyst preparation
ieved the full conversion of HMF with 97% selectivity to DFF
at 1308C and 2.8 bar O after 4 h reaction time using Ru/g-
We used 1,3-dicyanobenzene (1,3-DCB), 2,6-pyridinedicarboni-
trile (2,6-DCP), 1,4-dicyanobenzene (1,4-DCB), and 4,4’-biphe-
nyldicarbonitrile (4,4’-DCBP) as monomers to obtain CTF mate-
rials that contained numerous N moieties with different prop-
erties such as specific surface area, pore size and structure, and
N content (Scheme 3). All these materials were synthesized
2
[
18]
Al O . Still, the conversion decreased over five consecutive
2
3
recycling steps.
Nie et al. tested Ru on several supports at 1108C with only
a slight decrease in selectivity (96.2%) and activity over five
[
19]
cycles for Ru/C, which is the best catalyst to date. However,
the catalyst had to be reactivated by hydrothermal treatment
for 4 h after each reaction cycle. They also showed that Ru/C
was much more active than Pt, Pd, Rh, and Au-based catalysts.
Moreover, a pressure of 20 bar of pure O was necessary to
2
achieve these results.
Zhang et al. used Co-Ce-Ru to reach an 82.6% DFF yield
with 96.5% conversion after 12 h at 1208C and only an atmos-
[
20]
pheric pressure of oxygen. Furthermore, they enhanced the
catalyst recycling by using magnetic separation with
III
a Fe O @SiO -NH -Ru catalyst, which attained 86.4% yield at
3
4
2
2
full conversion after 4 h at 1208C. Additionally, they used air at
atmospheric pressure, which led to the same results but re-
quired prolonged reaction times of 16 h.
Scheme 3. Monomers applied as linkers in CTF synthesis.
Herein, we present a versatile method for the preparation of
nanoparticulate Ru catalysts stabilized on covalent triazine
frameworks (CTF), a class of highly stable polymers formed by
using molten ZnCl as the solvent in a ZnCl /monomer molar
2
2
[21]
the trimerization of aromatic dinitriles in molten ZnCl2.
ratio of 5:1. Heating the monomer/salt mixture to 4008C and
subsequently to 6008C for at least 10 h each leads to the for-
mation of a fully amorphous black solid with an extremely
high surface area and bimodal micro- and mesoporosity. How-
ever, if the temperature was increased to 6008C, partial car-
bonization is observed, which leads to decreased N contents
compared to the applied monomer. This effect is especially
visible for prolonged reaction times at 6008C as the N content
further decreases. Still, the number of N moieties is sufficient
to enable the immobilization of homogeneous Ru precursors
as will be discussed later. The N contents, specific surface
areas, pore volumes, and the amounts of stabilized Ru in the
These materials are temperature stable up to 4008C and in-
soluble in most common solvents. As a result of their desirable
physical properties, CTFs meet the modern demands placed
upon a solid catalyst suitable for use in sustainable chemis-
[
22]
try. We used various dinitrile monomers to access porous
CTF materials that contain numerous N moieties, which allow
the coordination of different molecular catalysts before reduc-
tion (Scheme 2). This approach allows both a molecular disper-
[
23]
sion of metal species on the solid support and a narrow par-
ticle size distribution of the metal nanoparticles formed upon
reduction. Following this methodology, Ru/CTF materials were
accessed that show high activity and selectivity in the aerobic
oxidation of HMF to DFF even at low temperatures using syn-
thetic air as the oxidant.
prepared catalysts are summarized in Table 1. N physisorption
2
isotherms for all materials studied in this work are illustrated in
Figure 1.
N physisorption isotherms of all CTF polymers based on
2
1,3-DCB (CTF-a) correspond to type IV isotherms typical of
3 2
Scheme 2. a) Idealized synthesis of a CTF based on 1,3-DCB as a monomer (CTF-a). b) Proposed coordination approach of RuCl ·xH O to form immobilized
III
2
Ru @CTF-a. c) Proposed stabilized metal nanoparticles after reduction in the presence of pure H .
&
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
2
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Table 1. Elemental analysis, specific surface area, total and micropore
volume of monomers and CTF materials as well as Ru-content of the
metal loaded Ru/CTF catalysts.
[
a]
[b]
[c]
[d]
[e]
Monomer/
material
N
S
BET
2
V
P(total)
3
V
P(micro)
3
Ru
À1
À1
À1
[%]
[m g
]
[cm g
]
[cm g
]
[%]
[
g]
1
,3-DCB
21.9
9.5
11.1
10.1
9.3
32.5
17.2
21.9
10.4
13.7
3.7
–
2439
2342
2255
2045
–
1179
–
2071
–
–
–
–
CTF-a
1.96
1.65
1.56
1.45
–
0.64
–
1.36
–
2.63
–
0.2
–
0.47
0.45
0.42
0.34
–
0.64
–
0.43
–
0.30
–
0.2
–
4.32
3.71
3.38
3.71
–
3.34
–
3.91
–
3.99
5.00
6.08
6.58
[
[
[
f]
f]
f]
CTF-a (20/20)
CTF-a (30/30)
CTF-a (40/40)
[
[
[
g]
g]
g]
2
,6-DCP
CTF-b
,4-DCB
CTF-c
,4’-DCBP
CTF-d
1
4
1683
900
8
[
h]
Ru/C
HT
MgO
–
–
–
60
[
a] Determined by elemental analysis. [b] Surface area identified using the
BET method. [c] Total pore volume determined at p/p =0.98. [d] Micro-
pore volume calculated by N -DFT model. [e] Determined by ICP-OES
analysis after immobilization and reduction of the RuCl ·xH O precursor
under H atmosphere, 3 h, 3508C. [f] CTF-a X/Y was synthesized for
0
2
3
2
2
X hours at 4008C and further Y hours at 6008C. [g] Theoretical value for N
content; [h] Data provided by Sigma–Aldrich.
mesoporous materials (Figure 1a). Prolonged synthesis times
lead to a decrease in N uptake and smaller hysteresis, which
2
Figure 1. a) N
2
physisorption of CTF-a materials after different synthesis
can be correlated with a decrease in specific surface areas
times denoted as CTF-a (X/Y) for synthesis for X h at 4008C and Y h at
6008C; (offset for CTF-a and CTF-a (20/20): +100 cm g
a (30/30): +50 cm g ) and b) N physisorption of different CTF materials
based on various linker molecules after synthesis for 10 h at 4008C and 10 h
(
SBET) as well as total and micropore volumes. The structural pa-
3
À1
; offset for CTF-
3
À1
rameters of the CTF materials can be controlled by varying the
monomer. The CTF based on 2,6-DCP (CTF-b, Figure 1b) exhib-
its a type I isotherm characteristic of microporous materials.
This effect is most probably caused by a stable coordination of
the pyridinic structure element of the monomer to the Lewis
2
at 6008C.
hensive characterization of the interaction of metal species
with these N-rich support materials. However, SEM with
energy-dispersive X-ray spectroscopy (EDX) mapping demon-
strates that Ru is dispersed finely throughout the support
upon reduction to the nanoparticulate state (Figure 3). Further-
more, TEM images indicate that neither metal nanoparticles
nor agglomerates are formed during coordination. Even after
acidic ZnCl during synthesis, which leads to a more dense co-
2
[
21b]
ordination geometry during polymerization.
The SBET and
micropore volume of CTF-b are rather small compared to that
of CTF-a. Nevertheless, because of the pyridine linkers, it con-
tains a significantly higher amount of N compared to the other
CTF materials. The N physisorption isotherm of CTF-c (based
2
0
on para-substituted 1,4-DCB) corresponds to a type IV iso-
reduction to Ru species, no nanoparticle formation could be
therm, which emphasizes the mesoporous structure of the ma-
observed by TEM, which indicates a high metal dispersion
(Figure 2). In line with this, powder XRD patterns of the cata-
terial. Interestingly, the amount of N adsorbed and the hyste-
2
0
resis are much less pronounced according to the significantly
lower SBET and pore volume values of CTF-c than that of CTF-a.
The N content is comparable to that of CTF-a, as expected
from the theoretical data for the monomers. CTF-d based on
lysts show no reflections of Ru nanoparticles formed after re-
4
,4’-DCBP presents a highly mesoporous material with a moder-
2
À1
ate surface area of 1683 m g . However, because of the low N
content of its monomer, CTF-d exhibits a low number of coor-
dination sites.
All of the prepared CTF materials show a comparable uptake
of Ru during metal coordination followed by reduction under
a H atmosphere. The coordination mode adopted by Ru in
2
III
these Ru @CTF systems before reduction has not yet been
III
0
identified completely. Future studies will focus on a compre-
Figure 2. TEM images of a) Ru @CTF-a and b) Ru @CTF-a.
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
3
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Figure 3. SEM/EDX mapping images of Ru/CTF-a. a) SEM of Ru/CTF-a. b) C
mapping. c) N mapping. d) Ru mapping. Chloride was present throughout
III
the whole sample, which indicates that Ru species were not reduced fully
0
to Ru species.
Figure 4. Dependence of the catalytic conversion on various parameters
such as a) temperature, b) initial pressure of air, c) stirring speed, and d) HMF
concentration in MTBE. Only one parameter was changed at a time and the
others were maintained as follows. Reaction conditions: 1 h, 808C, 20 bar of
air, 15 mL MTBE, 500 rpm stirring speed, Ru/C, HMF/metal molar ratio 40:1.
duction of the coordinated complexes at 3508C under a H at-
2
mosphere. Clearly, Ru nanoparticles formed upon reduction
are small and, therefore, X-ray amorphous, which confirms the
efficient pre-coordination and stabilization on the N functional-
ities of the CTF supports. Comparable effects of N-containing
support materials have been reported previously, for example,
for Pd nanoparticles supported on comparable triazine frame-
works as well as for Ru nanoparticles on N-functionalized
boiling point, MTBE is easily separated from the product at low
temperature, and the conversion and yield were highest
among the used solvents. The use of DMSO or water as the
solvent led to lower yields of DFF caused by overoxidation to
FFCA. Therefore, subsequent experiments to determine the in-
fluence of various reaction parameters were performed in
MTBE (Figure 4).
[
24,25]
carbon nanofibers.
Close to full conversion could be attained after 1 h at 1108C
and an initial pressure of 20 bar of synthetic air. Nevertheless,
at 1108C the yield of DFF decreases slightly. An explanation
could be polymerization reactions of the product that already
take place at these elevated temperatures. To study the influ-
ence of the different support materials presented in this work,
reactions were performed at 808C to allow a moderate conver-
sion of 54.5% of HMF with a DFF yield of 43.7% for commer-
cial Ru/C. We varied the initial pressure of synthetic air to con-
firm minor changes of conversion and yield for pressures
above 40 bar (Figure 4). The stirring speed does not have
a great influence on the catalytic activity above 500 rpm. Fur-
thermore, the HMF concentration in solution does not influ-
ence the catalytic conversion and DFF yield drastically as long
as a molar ratio of 40:1 of HMF to the catalyst remains con-
stant. This circumstance grants the use of concentrated solu-
tions, reduces solvent needs, and enables easy and cost-effi-
cient separation of the product, and the efficiency of the trans-
formation of HMF into DFF is not affected.
Catalytic oxidation of HMF to DFF
Initially, the dependence of the catalytic activity of commercial-
ly available Ru/C on different parameters and solvents was ex-
amined. Solvent screening showed that methyl t-butyl ether
(MTBE) is the most appropriate solvent for both high catalytic
activity and product separation (Table 2). As a result of its low
Table 2. Dependence of the catalytic activity on the solvent using Ru/C
[
a]
as a catalyst.
[
b]
[b]
[b]
Solvent
Conversion
%]
DFF yield
[%]
C balance
[%]
[
toluene
MTBE
54.1
54.5
49.4
41.9
55.1
47.9
49.5
31.4
43.7
36.9
10.3
28.9
37.0
38.2
77.3
89.2
87.5
68.4
73.8
89.1
88.7
1
,4-dioxane
[c]
DMSO
[
c]
2
H O
acetone
acetonitrile
With this in mind, catalytic test reactions for several catalyst
supports were performed under optimized conditions for 1 h
at 808C and 20 bar of air using MTBE as the solvent at
[
a] Reaction conditions: 1 h, 808C, 20 bar of air, 15 mL solvent, 500 rpm
stirring speed, HMF/metal molar ratio 40:1. [b] Determined by HPLC anal-
ysis. [c] In the presence of water, oxidation to FFCA takes place; hygro-
scopic DMSO seems to contain traces of water, which explains the lower
selectivity to DFF.
500 rpm stirring speed. No DFF formation occurred without
catalyst. Both Pd/C and Pt/C resulted in poor DFF yields com-
&
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
4
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
CTF-a, a support with a superior specific surface area and
a moderate total pore volume (Table 1). Ru supported on mi-
croporous CTF-b, which has a significantly lower specific sur-
face area, was nearly inactive, whereas Ru/CTF-c shows compa-
rable results to Ru/CTF-a because of the comparable structural
properties of these support materials. For Ru supported on
CTF-d, a mesoporous support with a remarkably high total
pore volume but only moderate specific surface area, a lower
conversion was observed. Consequently, a high specific surface
area of the support material seems to be essential for high cat-
alytic activity. Furthermore, mesoporosity is advantageous for
high catalytic activity compared to purely microporous support
materials.
Table 3. Dependence of the catalytic activity on the metal catalyst, sup-
port material, and oxidation state of Ru.
[
a]
[
b]
[b]
[b]
Entry
Catalyst
Conversion
DFF yield
[%]
C balance
[%]
[%]
1
2
3
4
5
6
7
8
9
blank
Pd/C
Pt/C
Ru/C
Ruox/C
Ru/g-Al
Ru/HT
Ru/MgO
Ru/CTF-a
Ru/C (RT)
CTF-a
8.3
0.1
8.8
1.3
43.7
11.4
38.9
1.0
1.6
63.6
2.1
91.8
89.4
72.8
89.2
80.5
87.7
85.7
85.8
77.3
87.3
79.3
19.4
28.5
54.5
30.9
51.2
15.3
15.8
86.3
14.8
21.7
2
O
3
[
c]
10
11
1.0
A similar trend can be observed for a single CTF material by
varying the synthesis parameters to alter the structural param-
eters. The specific surface area and total pore volume of CTF-
a decrease with prolonged synthesis times (Table 1). Accord-
ingly, conversion and DFF yields decrease (Figure 5b). For sub-
[
a] Reaction conditions: 1 h, 808C, 20 bar of air, 15 mL MTBE, 500 rpm stir-
ring speed, HMF/metal molar ratio 40:1. [b] Determined by HPLC analysis.
c] Adsorption test at RT, 500 rpm, atmospheric pressure.
[
pared to most of the Ru-based catalysts (Table 3, entries 2 and
). Previous studies on Ru-based catalysts for the oxidation of
3
0
HMF to DFF suggest Ru as the active species. To support this
hypothesis, Ru/C was calcined for 4 h at 3008C under an air
flow to form Ru /C. This catalyst enabled only 30.9% conver-
ox
sion and a significantly lower DFF yield of 11.4% compared to
0
4
3.7% for Ru /C under the same conditions (Table 3, entry 4 vs.
), which indicates the superior activity of reduced Ru catalysts.
5
Nevertheless, a comprehensive investigation of the nature of
the catalytically active sites is certainly necessary and will be
targeted in future studies. Commercially available Ru/g-Al O3
2
exhibited comparable conversions and DFF yields to Ru/C. In
contrast, freshly prepared Ru catalysts on hydrotalcite (HT) and
MgO with a rather high metal loading of 6.1 wt% for Ru/HT
and 6.6 wt% for Ru/MgO were nearly inactive and resulted in
poor yields of DFF (Table 3, entries 6–8). A reason for this
might be the low specific surface area of both supports com-
pared to all the other catalysts included in this study.
Figure 5. Dependence of the catalytic conversion (X: conversion; Y: yield) on
the CTF support material. a) CTFs based on different monomers. b) CTF-a syn-
thesized in various reaction times (X/Y). Reaction conditions: 1 h, 808C,
2
4
0 bar of air, 15 mL MTBE, 500 rpm stirring speed, HMF/metal molar ratio
0:1.
The most active catalyst, Ru/CTF-a, was prepared by reduc-
sequent investigations, CTF-a synthesized with 10 h time inter-
vals was utilized, which reduced the overall synthesis time as
well as the energy demand during synthesis.
tion under H atmosphere (Table 3, entry 9). Notably, for both
2
Ru/CTF-a and Ru/C, DFF was the major product and only trace
amounts of FFCA were observed. Nevertheless, in all experi-
ments mass balances are not closed. Previous studies have
shown that HMF is adsorbed strongly on the surface of differ-
To compare the catalytic performance of Ru/CTF-a to com-
mercial Ru/C, time-resolved measurements were conducted
(Figure 6). A certain conversion occurs during the heating of
[
26]
[27]
ent solid supports
and oligomeric byproducts
can be
formed. Experiments with Ru/C at room temperature and only
CTF-a under the standard reaction conditions were conducted
(Table 3, entries 10 and 11). These experiments confirm that
HMF consumption occurs even at room temperature and in
the absence of a catalytically active species. At the same time
no DFF is formed, which supports HMF adsorption as the
origin for the gaps in the mass balances.
The dependence of the activity of several Ru catalysts sup-
ported on CTF materials on the material precursor and the syn-
thesis conditions was investigated (Figure 5). Interestingly,
HMF conversion and DFF yield can be correlated with porosity
and specific surface area of the support material. In line, the
maximum conversion could be achieved for Ru supported on
Figure 6. Time-resolved a) conversion (X) and b) DFF yield (Y) using Ru/CTF-
a (black) and Ru/C (gray). Reaction conditions: 808C, 20 bar of air, 15 mL
MTBE, 500 rpm stirring speed, HMF/metal molar ratio 40:1.
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
5
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
the autoclave to the desired temperature of 808C in approxi-
mately 10 min. Surprisingly, full conversion is already achieved
after 3 h for Ru/CTF-a. In contrast, 6 h is required until full con-
version with Ru/C. At approximately 84% conversion, Ru/CTF-
active species into solution. The conversion did not increase if
the filtered reaction mixture was used under the reaction con-
ditions. Nevertheless, the current reactivation procedure relies
on rather harsh conditions. The agglomeration of metal nano-
particles together with an associated loss of external metal sur-
face area could cause the observed decrease in conversion and
À3
À1 À1
a achieves a productivity of 25.4ꢁ10 molg h , which is ap-
proximately seven times higher than the productivity of 3.3ꢁ
À3
À1 À1
10
molg
h
for Ru/C. DFF yields remain limited to approxi-
DFF yield. Furthermore, N physisorption studies on the Ru/
2
mately 78%, which corresponds to our earlier observations. As
no significant byproduct formation occurred and the reaction
solutions remained colorless, we suggest the adsorption of
HMF and DFF on the catalyst surface. Nevertheless, the forma-
tion of polymeric byproducts that adsorb on the catalysts
cannot be excluded. For further comparison, the effect of the
temperature on the activity and selectivity has been studied
for Ru/CTF-a. After 1 h at 908C, a conversion of 92.3% and
a 73.4% DFF yield were obtained. A further increase of the
temperature to 1008C led to 99.2% conversion and 77.9% DFF
yield. Interestingly, after washing the catalyst with 15 mL of
acetone and considering the extracted compounds in the
CTF-a catalyst as prepared and after five cycles reveal a signifi-
cant loss of pore volume and surface area (Figure S1 and
Table S1). This effect emphasizes a strong substrate and prod-
uct adsorption as mentioned previously together with poten-
tial polymer formation. Consequently, further optimization of
catalyst preparation and regeneration has to aim for reduced
adsorption properties and suitable means to facilitate the com-
plete removal of polymeric deposits. Therefore, future studies
will aim to optimize recycling and reactivation conditions to-
gether with a continuous operation for HMF oxidation.
mass balance, DFF yields could be further increased to 86.8 Conclusions
908C, conversion (X)=90.5%) and 92.0% (1008C, X=99.1%),
(
An efficient catalyst system based on Ru supported on cova-
lent triazine frameworks (CTFs) was developed for the selective
oxidation of 5-hydroxymethylfurfural (HMF) to 2,5-diformylfur-
an (DFF) under aerobic conditions. The bimodal and mesopo-
rous Ru/CTF catalysts showed high conversions and DFF yields
at low temperatures using air as the sole oxidant and methyl t-
butyl ether as an easily separable and reusable solvent. The
catalytic activity depends strongly on the structural parameters
of the CTF materials, such as specific surface area and total
pore volume. These parameters can be controlled by the
choice of the linker as well as the synthesis time for the frame-
work. High conversions of 97.3% and DFF yields of 72.7%
could be obtained after only 3 h at 808C using 20 bar of air. At
which leads to a nearly closed mass balance. This gives further
evidence for the strong adsorption of both product and sub-
strate as suggested before. At the same time, the conversion
decreases slightly, as both DFF and HMF are adsorbed on the
catalyst surface.
Ru/CTF-a and Ru/C have been recycled to investigate their
stability (Figure 7). Both catalysts showed a strong loss in activ-
ity if simply washed with an organic solvent, dried under
84% HMF conversion, the productivity of Ru/CTF-a was nearly
seven times higher than that of Ru/C. The recycling of these
Ru-based catalysts is still challenging. Nevertheless, Ru/CTF-
a exhibits only minor deactivation if a reactivation procedure is
applied under H flow. Therefore, this concept paves the way
2
for an environmentally benign continuous production of bio-
mass-derived chemicals that does not depend on diminishing
fossil resources and enables sustainable industrial processing.
Figure 7. Recycling study of a) Ru/C and b) Ru/CTF-a. Reaction conditions:
1
h, 808C, 20 bar of air, 15 mL MTBE, 500 rpm stirring speed, HMF/metal
molar ratio 40:1. Catalysts were reactivated after each run.
Experimental Section
vacuum, and reused without further treatment. Our findings
suggest that not only polymeric surface species but also the
surface oxidation of the supported metal nanoparticles cause
the observed loss of activity. Therefore, catalysts were reacti-
Catalyst preparation
For the synthesis of CTF-a, 1,3-dicyanobenzene (0.621 g,
4.85 mmol, 1 equiv.) and ZnCl (3.305 g, 24.25 mmol, 5 equiv.) were
2
vated at 3508C for 3 h under H flow. With this strategy, a sig-
mixed and ground together, transferred into a quartz ampoule,
and dried under vacuum for at least 3 h. The ampoule was then
flame-sealed and placed inside a furnace for 10 h of heat treatment
2
nificantly reduced deactivation could be achieved together
with the stable catalytic activity of Ru/CTF-a after two recycling
steps. Overall, Ru/CTF-a exhibits not only superior activity but
also minor deactivation compared to Ru/C. We assign this ob-
servation to the N functionalities of the support that provide
a stabilizing effect to the Ru species and hinder agglomeration
À1
at 4008C and a further 10 h at 6008C (heating rate: 10 Kmin ).
After cooling to RT, the ampoule was broken open (CAUTION: the
ampoules are under pressure, which is released during opening),
and the solid product was ground and washed thoroughly with
water and dilute HCl (0.1m). The solid material was then ground in
a ball mill (Fritsch Pulverisette23, 5 min, 30 Hz) to obtain a black
powder, which was washed successively with water, dilute HCl,
[
24,25]
and leaching of metal species.
Additionally, hot filtration
tests were performed to exclude the leaching of catalytically
&
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
6
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
dilute NaOH, water, and THF and dried under vacuum for at least
2 h. Materials based on 2,6-pyridinedicarbonitrile (CTF-b), 1,4-di-
cyanobenzene (CTF-c), and 4,4’-biphenyldicarbonitrile (CTF-d) were
laboratory Kolbe (Mꢀlheim an der Ruhr, Germany) for CHN analy-
sis.
1
synthesized as described for 1,3-DCB. For Ru impregnation, CTF
Keywords: biomass · heterogeneous catalysis · oxidation ·
ruthenium · supported catalysts
(
600 mg) was added to
a solution of RuCl ·xH O (0.079 g,
3 2
0
.381 mmol) in EtOH (400 mL) that was heated to reflux and stirred
III
for 6 h. After cooling to RT, the Ru @CTF material was collected by
filtration and washed with EtOH to remove uncoordinated Ru pre-
cursor. After drying under vacuum at 608C for at least 12 h, the
[1] a) J. N. Chheda, G. W. Huber, J. A. Dumesic, Angew. Chem. Int. Ed. 2007,
46, 7164–7183; Angew. Chem. 2007, 119, 7298–7318; b) M. J. Climent,
A. Corma, S. Iborra, Green Chem. 2011, 13, 520–540.
III
Ru @CTF material was reduced in a tube furnace under a H at-
2
À1
À1
mosphere (heating rate: 10 Kmin , 3508C, H2 flow 100 mLmin
h) to obtain Ru/CTF (for Ru loading, see Table 1). Ru/C, Ru/g-
Al O , Pd/C, and Pt/C catalysts (5 wt%) were purchased from
,
[
[
2] A. Corma, S. Iborra, A. Velty, Chem. Rev. 2007, 107, 2411–2502.
3] a) J. N. Chheda, Y. Romꢂn-Leshkov, J. A. Dumesic, Green Chem. 2007, 9,
3
2
3
342–350; b) H. Zhao, J. E. Holladay, H. Brown, Z. C. Zhang, Science
Sigma Aldrich and were used as received. Ru/HT (HT composition:
CH Al Mg O ·4H O) and Ru/MgO were prepared by wet impreg-
2
007, 316, 1597–1600; c) I. Delidovich, K. Leonhard, R. Palkovits, Energy
1
6
2
6
19
2
Environ. Sci. 2014, 7, 2803–2830.
nation from RuCl ·xH O (0.1313 g, 0.633 mmol) in EtOH (20 mL)
3
2
[
4] a) A. A. Rosatella, S. P. Simeonov, R. F. M. Frade, C. A. M. Afonso, Green
Chem. 2011, 13, 754–793; b) R.-J. van Putten, J. C. van der Waal, E.
de Jong, C. B. Rasrendra, H. J. Heeres, J. G. de Vries, Chem. Rev. 2013,
113, 1499–1597.
with the particular support (1.0 g). All supported metal precursors
III
were reduced using the method described for Ru @CTF. The CTF
materials were characterized by thermogravimetric analysis (TGA),
elemental analysis, N sorption measurements, TEM, and XRD. The
[5] R. Alamillo, M. Tucker, M. Chia, Y. Pagꢂn-Torres, J. Dumesic, Green Chem.
012, 14, 1413–1419.
2
2
Ru-doped materials were analyzed using inductively coupled
plasma optical emission spectroscopy (ICP-OES), SEM/EDX, TEM,
and XRD.
[
6] a) A. J. Sanborn, P. D. Bloom, US Pat, 2008, 7579490; b) C. Moreau, M. N.
Belgacem, A. Gandini, Top. Catal. 2004, 27, 11–30.
[
7] T. Buntara, S. Noel, P. H. Phua, I. Meliꢂn-Cabrera, J. G. de Vries, H. J.
Heeres, Angew. Chem. Int. Ed. 2011, 50, 7083–7087; Angew. Chem. 2011,
1
23, 7221–7225.
[
8] a) P. Verdeguer, N. Merat, A. Gaset, J. Mol. Catal. 1993, 85, 327–344;
b) C. Carlini, P. Patrono, A. M. R. Galletti, G. Sbrana, V. Zima, Appl. Catal.
A 2005, 289, 197–204; c) A. S. Amarasekara, D. Green, E. McMillan,
Catal. Commun. 2008, 9, 286–288; d) X. Xiang, L. He, Y. Yang, B. Guo,
D. M. Tong, C. W. Hu, Catal. Lett. 2011, 141, 735–741.
9] a) A. Gandini, M. N. Belgacem, Prog. Polym. Sci. 1997, 22, 1203–1379;
b) E. de Jong, A. Dam, L. Sipos, G.-J. M. Gruter in Biobased Monomers,
Polymers and Materials, Vol. 1105 (Eds.: P. B. Smith, R. A. Gross), American
Chemical Society, Washington DC, 2012, pp. 1–13; c) A. J. J. E. Eerhart,
A. P. C. Faaij, M. K. Patel, Energy Environ. Sci. 2012, 5, 6407–6422.
Selective oxidation of HMF to DFF
Typically, a stainless-steel autoclave (75 mL) with a glass inlet was
charged with a solution of HMF (0.1261 g, 1 mmol) in solvent
(
15 mL). The catalyst (HMF/metal molar ratio: 40:1) was added, and
[
the autoclave was equipped with a stirring bar and temperature
sensor. It was sealed, pressurized to 20 bar with synthetic air (hy-
drocarbon free), and heated to 808C with stirring at 500 rpm. After
a certain time, the autoclave was cooled and depressurized. The
catalyst was removed by filtration with a syringe filter (CHROMAFIL
Xtra, PA-20/25, 0.20 mm), and the reaction solution was analyzed
by HPLC (Shimadzu 2020, 300ꢁ8.0 mm organic acid resin column,
T=408C, UV detector at l=254 nm for HMF and DFF, refractive
index detector (RID-10A) detector for HMFCA, FFCA, and FDCA)
[10] K. T. Hopkins, W. D. Wilson, B. C. Bender, D. R. McCurdy, J. E. Hall, R. R.
Tidwell, A. Kumar, M. Bajic, D. W. Boykin, J. Med. Chem. 1998, 41, 3872–
3878.
[
11] M. Del Poeta, W. A. Schell, C. C. Dykstra, S. K. Jones, R. R. Tidwell, A.
Kumar, D. W. Boykin, J. R. Perfect, Antimicrob. Agents Chemother. 1998,
4
2, 2503–2510.
with 154 mL trifluoroacetic acid in 1 L water as an eluent (flow rate:
[
[
12] A. Gandini, N. M. Belgacem, Polym. Int. 1998, 47, 267–276.
13] A. S. Amarasekara, D. Green, L. D. Williams, Eur. Polym. J. 2009, 45, 595–
À1
1
mLmin ). For recycling studies, the catalysts were collected by
filtration by using a Whatman filtration system equipped with Ano-
598.
disc 25 (0.20 mm) membranes, washed thoroughly with solvent,
[14] L. Cottier, G. Descotes, E. Viollet, J. Lewkowski, R. Skowronski, J. Hetero-
cycl. Chem. 1995, 32, 927–930.
dried overnight under vacuum at 608C and reactivated in a tube
À1
[
15] W. Partenheimer, V. V. Grushin, Adv. Synth. Catal. 2001, 343, 102–111.
[16] a) C. Moreau, R. Durand, C. Pourcheron, D. Tichit, Stud. Surf. Sci. Catal.
997, 108, 399–406; b) O. C. Navarro, A. C. Canꢃs, S. I. Chornet, Top.
furnace under a H atmosphere (10 Kmin , 3508C, H2 flow
1
2
À1
00 mLmin , 3 h) to regain activity. Conversions (X) and yields (Y)
1
were calculated as follows: X(HMF)=(n (HMF)Àn (HMF))/
0
1
Catal. 2009, 52, 304–314.
n (HMF)·100%, in which n is the initial molar amount of HMF and
0
0
[
17] a) A. Takagaki, M. Takahashi, S. Nishimura, K. Ebitani, ACS Catal. 2011, 1,
n1 is the molar amount of unreacted HMF, and Y(DFF)=n(DFF)/
n0(HMF)ꢁ100% for the formation of one molecule of DFF per mol-
1562–1565; b) S. Dabral, S. Nishimura, K. Ebitani, ChemSusChem 2014,
7, 260–267.
ecule of HMF.
[18] C. A. Antonyraj, J. Jeong, B. Kim, S. Shin, S. Kim, K.-Y. Lee, J. K. Cho, J.
Ind. Eng. Chem. 2013, 19, 1056–1059.
[
19] a) J. Nie, J. Xie, H. Liu, J. Catal. 2013, 301, 83–91; b) J. Nie, J. Xie, H. Liu,
Chin. J. Catal. 2013, 34, 871–875.
Acknowledgements
[20] a) Y. Wang, B. Liu, K. Huang, Z. Zhang, Ind. Eng. Chem. Res. 2014, 53,
1313–1319; b) S. Wang, Z. Zhang, B. Liu, J. Li, Ind. Eng. Chem. Res. 2014,
5
3, 5820–5827.
We acknowledge financial support from the DFG (PA1689/1-1)
and funding from the Excellence Initiative of the German federal
and state governments. SEM/EDX mapping was performed at the
Center for Chemical Polymer Technology CPT, which is supported
by the EU and the federal state of North Rhine-Westphalia (grant
no. EFRE 30 00 883 02). We thank Mr. B. Spliethoff and Mr. A.
Dreier for the TEM measurements, as well as the microanalytic
[
[
[
21] a) P. Kuhn, M. Antonietti, A. Thomas, Angew. Chem. Int. Ed. 2008, 47,
450–3453; Angew. Chem. 2008, 120, 3499–3502; b) P. Kuhn, A.
3
Thomas, M. Antonietti, Macromolecules 2009, 42, 319–326.
22] a) J. Roeser, K. Kailasam, A. Thomas, ChemSusChem 2012, 5, 1793–1799;
b) P. Katekomol, J. Roeser, M. Bojdys, J. Weber, A. Thomas, Chem. Mater.
2
013, 25, 1542–1548.
23] a) R. Palkovits, M. Antonietti, P. Kuhn, A. Thomas, F. Schꢄth, Angew.
Chem. Int. Ed. 2009, 48, 6909–6912; Angew. Chem. 2009, 121, 7042–
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
7
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
7
045; b) M. Soorholtz, R. J. White, T. Zimmermann, M.-M. Titirici, M. An-
[26] C. Detoni, C. H. Gierlich, M. Rose, R. Palkovits, ACS Sustainable Chem.
Eng. 2014, 2, 2407–2415.
[27] K. Pupovac, R. Palkovits, ChemSusChem 2013, 6, 2103–2110.
tonietti, R. Palkovits, F. Schꢄth, Chem. Commun. 2013, 49, 240–242.
24] C. E. Chan-Thaw, A. Villa, G. M. Veith, K. Kailasam, L. A. Adamczyk, R. R.
Unocic, L. Prati, A. Thomas, Chem. Asian J. 2012, 7, 387–393.
25] S. Armenise, L. Roldꢂn, Y. Marco, A. Monzꢃn, E. Garcꢅa-Bordejꢆ, J. Phys.
Chem. C 2012, 116, 26385–26395.
[
[
Received: October 2, 2014
Published online on && &&, 0000
&
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
8
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

FULL PAPERS
Oxidize and conquer: 5-Hydroxyme-
thylfurfural (HMF) has been oxidized se-
lectively to 2,5-diformylfuran (DFF)
using air under mild conditions with Ru
supported on covalent triazine frame-
works (CTFs) as catalysts. These catalysts
result in higher conversions and yields
compared to commercially available Ru/
C and show superior stability in recy-
cling studies.
J. Artz, S. Mallmann, R. Palkovits*
& – &&
&
Selective Aerobic Oxidation of HMF to
2,5-Diformylfuran on Covalent Triazine
Frameworks-Supported Ru Catalysts
ChemSusChem 0000, 00, 0 – 0
www.chemsuschem.org
9
ꢀ 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ