Biomass into chemicals: One pot-base free oxidative esterification of 5-hydroxymethyl-2-furfural into 2,5-dimethylfuroate with gold on nanoparticulated ceria

Article abstract of DOI:10.1016/j.jcat.2009.04.019

2,5-Dimethylfuroate (DMF) is a valuable biomass derivative that can replace oil dependent PET polymers. 5-Hydroxymethyl-2-furfural (HMF) has been selectively converted into DMF (99 mol% yield) under mild conditions (65-130 °C, 10 bar O₂) in the absence of any base, by using gold nanoparticles on nanoparticulated ceria. The catalyst can be reused several times without any loss of activity or selectivity. The absence of metal leaching has been checked by the three-phase test. A full reaction scheme has been established and it has been found that the rate-limiting step of the reaction is the alcohol oxidation into aldehyde. After this, the reaction proceeds via aldehyde conversion into hemiacetal and further oxidation into the corresponding ester. Additionally, the effect of temperature, substrate-to-catalyst ratio, alcohol and water has been studied in an attempt to explain the catalytic behaviour of the Au-CeO₂.

Full text of DOI:10.1016/j.jcat.2009.04.019

Journal of Catalysis 265 (2009) 109–116
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Biomass into chemicals: One pot-base free oxidative esterification
of 5-hydroxymethyl-2-furfural into 2,5-dimethylfuroate with gold
on nanoparticulated ceria
*
O. Casanova, S. Iborra, A. Corma
Instituto de Tecnología Química (UPV-CSIC), Avda dels Tarongers s/n, Universitat Politècnica de València, 46022 Valencia, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
2,5-Dimethylfuroate (DMF) is a valuable biomass derivative that can replace oil dependent PET polymers.
5-Hydroxymethyl-2-furfural (HMF) has been selectively converted into DMF (99 mol% yield) under mild
conditions (65–130 °C, 10 bar O2) in the absence of any base, by using gold nanoparticles on nanopartic-
ulated ceria. The catalyst can be reused several times without any loss of activity or selectivity. The
absence of metal leaching has been checked by the three-phase test. A full reaction scheme has been
established and it has been found that the rate-limiting step of the reaction is the alcohol oxidation into
aldehyde. After this, the reaction proceeds via aldehyde conversion into hemiacetal and further oxidation
into the corresponding ester. Additionally, the effect of temperature, substrate-to-catalyst ratio, alcohol
and water has been studied in an attempt to explain the catalytic behaviour of the Au–CeO2.
Ó 2009 Elsevier Inc. All rights reserved.
Received 28 January 2009
Revised 30 March 2009
Accepted 21 April 2009
Available online 28 May 2009
Keywords:
Catalysis
5-Hydroxymethyl-2-furfural
2,5-Dimethylfuroate
Gold
Ceria
1. Introduction
However, such difficulty can be overcome by substituting FDCA
by its corresponding ester, i.e. dimethylfuroate (DMF), which is
The production of fine chemicals, polymer precursors and pet-
rol-derived commodities from biomass can contribute to diminish
our current dependence on non-renewable energy sources, and
chemical routes for catalytic biomass transformations have been
recently reviewed [1–3]. It has been presented that one of the fore-
most natural sources of potentially relevant chemicals are sugars,
and good results have been achieved in converting fructose into li-
quid alkanes and the production of a key intermediate such as 5-
hydroxymethyl-2-furfural (HMF) by homogeneous Brönsted acid
dehydration [4]. This compound is regarded as a key element for
the development of several derivatives with multiple applications,
e.g. pharmaceuticals, antifungals and polymer precursors [1].
Among these applications, the synthesis of polymer precursors
can be of interest. One of the monomers, whose volume of produc-
tion is constantly increasing is terephthalic acid (commonly used
in the manufacture of PET [5]), which according to Gandini and
co-workers [6] can be replaced by 2,5-furandicarboxylic acid
(FDCA). The latter replacement has been recently emphasized
due to improvements observed in terms of heat resistance and
mechanical properties [7]. Several attempts to synthesize FDCA
have been reported in the literature [8–11], but it is difficult to
handle due to its low solubility in most industrial solvents [5].
readily soluble in most common solvents. Therefore, a process that
is able to convert HMF into DMF can have an impact on the poly-
mer industry.
The idea of synthesizing inexpensive DMF should embody a
process wherein the oxidation of alcohol into aldehyde, the hemi-
acetal formation and further hemiacetal oxidation should occur in
a one-pot system to avoid the costly isolation and purification of
the intermediates. It is well known that alcohol oxidation is usually
a limiting step, whenever successive aldehyde oxidation is desired
[12]. Currently, the former oxidation step can be accomplished by
Pd [13], Au and Au–Pd supported catalysts [14–16], particularly
using gold nanoparticles supported onto nanoparticulated ceria
(Au–CeO2). This type of catalyst is of general use, able to work at
low temperature, short reaction time, low amounts of catalyst,
with environmentally friendly solvents and with O2 as oxidant
[17–21]. The key aspect of the catalytic performance of Au–CeO2
lies on the variation of the chemical properties when particle size
decreases. External atoms have the ability to bind with adsorbates
due to the presence of unsaturations in their valence. Clearly, when
the particle size is reduced, then the external-to-internal atom ra-
tio increases. Therefore, particles in the nanoscale (nanoparticles
thereafter) present a high population of unsaturations, responsible
for specific chemical properties. In the case of CeO2, these unsatu-
rations can be understood as defect sites, mainly oxygen vacancies,
which make the stoichiometric formula CeO2 like CeO2Àx being x
* Corresponding author. Fax: +34 963877809.
E-mail address: acorma@itq.upv.es (A. Corma).
0021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.04.019

110
O. Casanova et al. / Journal of Catalysis 265 (2009) 109–116
related to oxygen vacancies and the presence of CeIII. Similarly, if
metal nanoparticles (such as Au) are supported onto nanoparticu-
lated ceria, then the immediate consequence is again an increase in
the population of defects, which will enhance a given property
(adsorption capacity, redox potential, etc.) [22].
An important role played by nanoparticulated ceria in alcohol
oxidation was noticed when different catalysts having the same
composition and gold particle size, but different supports (TiO2,
Fe2O3, C, etc.) showed different catalytic performances. This was
related to the fact that nanoparticulated ceria was able to adsorb
oxygen, due to its oxygen vacancies, hence boosting reoxidation
steps within the mechanism. On the other hand, the Lewis acid
character of non-fully saturated cerium atoms becomes beneficial
for alcohol oxidation, and furthermore its interaction with gold
nanoparticles stabilizes positive gold atoms, responsible for the
initiation steps of the reaction [17].
Recently, Christensen and co-workers [23] have reported very
interesting results on HMF oxidation–esterification into DMF
(99 mol% yield) using Au–TiO2 catalyst at 130 °C in MeOH with
the aid of MeONa. Therein, a clear emphasis was laid on the role
played by MeONa, so that in its absence the reaction may not be
completed after longer reaction times, while the addition of a base
to HMF oxidation would make the process less green and more
expensive. Moreover, it has been reported that storage conditions
(light and temperature) may negatively alter the Au–TiO2 proper-
ties and this has been illustrated in oxidative reactions [24]. All
these factors led us to consider the possibility of searching for a
material with better stability and catalytic activity.
In the present work we will show that base-free environmentally
friendly one-pot synthesis of DMF (and other esters thereof) has
been achieved with Au–CeO2 and the catalytic activity will be com-
pared with that of Au–TiO2, which is regarded as the most efficient
one in the literature [23]. The molecular intermediates of the reac-
tion have been determined, and a global network has been estab-
lished. Kinetic parameters of the process were studied together
with the reusability and stability of Au–CeO2 by means of successive
reuses, chemical analysis and the three-phase reaction test.
was 3.5 nm from TEM. This Au–CeO2 catalyst is commercially
available at http://www.upv.es/itq.
Synthesis of 1.5 wt% Au–C catalyst: A colloidal solution of gold
nanoparticles stabilized by polyvinylalcohol was deposited on acti-
vated carbon (KB-B-100, provided by Aldrich), following the proce-
dure reported by Porta et al. [26], under vigorous stirring an
aqueous solution of HAuCl4 (2 L, 100 l
g mLÀ1). To this, a fresh solu-
tion of NaBH4 (38 mL, 0.1 M) was added. The Au nanoparticles gen-
erated were immobilized simply by adding the active carbon (2 g)
into the metal dispersion. After 1 h the slurry was filtered and the
total gold adsorption was checked by atomic absorption spectros-
copy of the filtrate. The average metal size of the nanoparticles was
3.5 nm.
1 wt% Au–TiO2 and 4.5 wt% Au–Fe2O3 catalysts: They were di-
rectly purchased from TEK and WGC, respectively, and were used
as received. In both cases, the average metal size of the nanoparti-
cles was 3.5 nm.
2.2. Synthesis of compounds 5-hydroxymethyl furoic acid methyl ester
and 5-hydroxymethyl furfural dimethyl acetal
5-Hydroxymethyl furoic acid methyl ester was prepared as fol-
lows: 400 mg HMF was dissolved in 40 mL MeOH, 50 mg MeONa
(35 wt% solution in MeOH) and 140 mg Au–TiO2 were added under
vigorous stirring and refluxed at (65 °C). Then, air was bubbled into
the reaction mixture at 0.5 mL/s for ca. 12 h. The reaction was com-
plete after this time as checked by GC/MS. After reaction the cata-
lyst was filtered off and MeOH contained in the filtrate was
evaporated and the remaining reddish concentrate (>95% yield)
was used as prepared.
5-Hydroxymethyl furfural dimethyl acetal was prepared as fol-
lows: 500 mg HMF was dissolved in 25 mL MeOH and 100 mg cal-
cined and dehydrated Al–beta zeolite (Si/Al = 12.5 mol/mol, CP806)
were added under vigorous stirring and refluxed at (65 °C) for
15 min. After reaction the catalyst was filtered off and MeOH con-
tained in the filtrate was evaporated and the remaining yellowish
concentrate (>96% yield according to GC/MS analysis) was used as
prepared without any further purification.
2. Experimentals and methods
2.3. Three-phase test
2.1. Synthesis of catalysts
The reactant for the three-phase test was prepared using the
following procedure: first, the imine reactant was obtained by con-
tacting 4 g aminofunctionalized silica (Aldrich, 1 mmol NH2/g so-
lid) with 1 g HMF dissolved in 50 mL CH2Cl2, at its reflux
temperature for 4 h; second, the imine was reduced with NaBH4
(1.5 equiv.) in ethanol at 5 °C for 2 h. The solids were collected
by vacuum filtration at room temperature, exhaustively washed
with the same solvent and dried overnight at 60 °C. The product
for the three-phase test was prepared in a similar manner to the
reactant, but using diformylfurane instead of HMF and the reduc-
tion step was not carried out.
The three-phase test was carried out by mixing 1 g of the three-
phase test reactant, as explained before, under the conditions indi-
cated in the catalytic experiments. For the homogeneous test with
the three-phase test reactant, AuCl3 was used as a catalyst, under
the same conditions as indicated in the catalytic experiments.
Synthesis of 2.1 wt% Au–CeO2 catalyst: The preparation of nanop-
articulated ceria was carried out following the reported procedure
[25]. In short, an aqueous solution of Ce(NO3)4 (375 mL, 0.8 M) was
treated, under stirring and at ambient temperature, with an aque-
ous solution of ammonia (1.12 L, 0.8 M). The colloidal dispersion of
CeO2 nanoparticles was heated in a PET vessel at 100 °C for 24 h.
The resulting yellow precipitate was filtered and dried under vac-
uum overnight. The cerium oxide synthesized has, owing to the
small size of the nanoparticles, a very high surface area (180 m2/g).
Gold was deposited on the nanoparticulated ceria by using the
following procedure (same as when no-nanometric ceria was
used): a solution of HAuCl4Á3H2O (350 mg) in deionized water
(160 mL) was brought to pH 10 by the addition of a solution of
NaOH 0.2 M. Once the pH value was stable the solution was added
to a gel containing colloidal CeO2 (4.01 g) in H2O (50 mL). After
adjusting the pH of the slurry at a value of 10 by the addition of
a 0.2 M solution of NaOH 0.2 M, the slurry was continuously stirred
vigorously for 18 h at RT. The Au–CeO2 solid was then filtered and
exhaustively washed with several litres of distilled water until no
traces of chlorides were detected by the AgNO3 test. The catalyst
was dried under vacuum at room temperature for 1 h. The total
Au content of the final catalyst was 2.1 wt% as determined by
chemical analysis and the average metal size of the nanoparticles
2.4. Catalytic experiments
0.2 g HMF, 16 g (20 mL) methanol and the appropriate amount
of freshly synthesized catalyst (HMF/Au mol ratio of 300, unless
otherwise stated) were loaded into a teflon-lined autoclave (vol-
ume 30 mL). Afterwards, the reactor was sealed, pressurized at
10 bar with N2 and heated up to the required temperature. Then,
oxygen was bubbled into the reaction mixture at a constant flow-

O. Casanova et al. / Journal of Catalysis 265 (2009) 109–116
111
100
90
80
70
60
50
40
30
20
10
0
rate of 0.33 mL/s keeping constant pressure by means of a back-
pressure-regulator. The mixture was stirred at ca. 1000 rpm. The
progress of the reaction was followed by taking samples at regular
periods and was analyzed by GC. Nitrobenzene was used as the
external standard.
2.5. Analytical systems
1 Conversion (mol%)
3 Yield (mol%)
10 Yield (mol%)
4 Yield (mol%)
9 Yield (mol%)
IR spectra were performed using a Nicolet 750 spectrophotom-
eter. The spectra were obtained in the region between 3800 and
400 cmÀ1 with a resolution of 4 cmÀ1 by accumulating 40 scans,
prior in situ dehydration of the sample at 100 °C.
The content of organic (wt% C, H, N, S) was measured by ele-
mental analysis using EA-1108 CHNS Fisons analyzer and sulpha-
nilamide as standard. Metal loading in the case of Au–C was
determined by atomic absorption spectrophotometry, previous
sample disgregation with a mixture of HF/HNO3. After dilution of
the resulting solution with Milli Q water, analyses were performed
4
3
5
1
2
0
Reaction time (hr)
with
a Varian-10 plus Atomic Absorption Spectrometer (+/
À0.01 ppm error). Chemical analysis of Au–CeO2 was carried out
using a Philips MiniPal 25 fm Analytic X-ray apparatus and the cor-
responding calibration plot.
Fig. 1. Kinetics plot for the oxidation–esterification of 1 into 10 at 130 °C, 10 bar
bubbled O2 at 0.33 mL/s using Au–CeO2 (1/Au mol ratio 300) as a catalyst.
Gas chromatography (GC) analyses were performed with a Var-
ian 3300 chromatograph equipped with a flame ionization detector
and the capillary column was TRB-5 (5% crosslinked phenyl-
methyl silicone) (sizes 30–0.25–0.25) (Teknokroma). Mass spectra
were performed by GC–MS (HP Agilent 5973 with a 6980 mass
selective detector).
reached at ca. 4 mol% after 2.5 h, and then it decreases, showing
its unstable behaviour. Finally, diester 10, which is a secondary
product increases gradually its concentration in the reaction med-
ium achieving 99 mol% after 5 h. Other compounds such as alde-
hyde ester 5 and aldehyde acetal 6 were detected at the trace level.
Taking into account the kinetic behaviour of the different com-
pounds presented in Fig. 1, a reaction network is presented in
Scheme 2. It can be seen that 3 and 4 are derived from the same
hemiacetal intermediate 2 (non-detectable through GC), i.e. they
follow the oxidation and acetalization of 2, respectively. It is
remarkable that between these two reactions, 2 is more prone to
oxidation rather than to acetalization, as the maximal concentra-
tion of 3 is much higher than 4. On the other hand, monoester ace-
tal 9, which ends up being further oxidized into the diester 10, can
be formed from the two pathways opened separately by the dis-
guised primary products indicated before. More specifically, after
3 is formed, it is gradually oxidized into the monoester aldehyde
5. Then, after its hemiacetalization into 7, most of it is converted
into the final desired product 10, while a fraction may be trans-
formed into 9, which is reoxidized into 10, as shown in the right-
hand pathway in Scheme 2. In order to find out if the oxidation
of the reactant into 3 or, on the contrary, it is the transformation
of 3 into the final product the controlling reaction step, we have re-
3. Results and discussion
3.1. Reaction network for HMF oxidation–esterification to DMF
HMF oxidation–esterification was performed in methanol, at
130 °C and 10 bar O2 using Au–CeO2 as a catalyst. Following the
evolution of the reaction with time, different compounds were de-
tected which are presented in Scheme 1, whereas Fig. 1 displays
the kinetic curves. Firstly, it can be seen that the monoester alcohol
3 concentration increases up to a maximum of ca. 60 mol% yield,
decreasing after the first hour of reaction. This behaviour allows
us to consider the monoester alcohol 3 as a disguised primary
and unstable product. Secondly, the acetal alcohol 4 appears as a
disguised primary and unstable compound [27]. The monoester
acetal 9 appears in the kinetics profile, when 3 and 4 concentra-
tions are decreasing, showing their disguised secondary character.
However, 9 is a minoritary product, because its maximal yield is
O
+
O
+
O
O
O
OH
O
OH
O
OH
O
1
4
3
Disguised primary unstable
Disguised primary unstable
O
O
O
O
+
+
O
O
O
O
O
O
9
10
Disguised secondary unstable
Disguised secondary unstable
O
O
+
O
O
O
O
O
O
6
5
trace
trace
Scheme 1. Detected products in HMF oxidative esterification with MeOH.

112
O. Casanova et al. / Journal of Catalysis 265 (2009) 109–116
O
OH
OH
O
1
O
O
O
O
O
O
O
OH
O
OH
O
OH
O
O
4
2
3
O
O
O
O
O
O
O
O
O
O
O
O
5
6
10
O
O
O
O
OH
O
O
O
O
O
7
OH
O
O
8
9
Scheme 2. Proposed 1 oxidation–esterification pathway.
acted the compound 3 under the same reaction conditions. Then,
from the results given in Table 1, one can see that, when the oxida-
tion was started from either 1 or 3, a similar reaction time is re-
quired to reach 99 mol% yield to 10, which means that the
alcohol oxidation into aldehyde is the limiting step.
If now we follow the second reaction route going through the
compound 4, then the acetal aldehyde 6 is formed, which is rapidly
transformed into the hemiacetal 8. Most of this is oxidized into 9
and finally converted into 10. In order to rule out the possibility
that compound 4 reacts to 3, we carried out the reaction starting
from pure compound 4. The kinetics of this reaction route is shown
in Fig. 2, therein the accumulation of 9 can be clearly seen, fol-
lowed by the acetal oxidation. As seen in Scheme 2, compound 9
is an obligatory stage for the two reaction pathways.
The effect of the presence of different substituents in the 5-fur-
ane on the rate of oxidation of the 2-hydroxymethyl group can be
seen from the results given in Table 2. When using 4-substituted-
benzyl alcohol type substrates [19,28], the presence of ring deacti-
vating groups, such as dimethoxymethyl and methylcarboxy (in 4
and 3, respectively), decreases the rate of oxidation. In fact, after
2 h, yields to 10 are lower (74.6 and 72 mol%, respectively) than
when no ring substituent was present as in 11 (Scheme 3)
(91 mol% yield to 15). Certainly, 11 follows an analogous oxida-
tion–esterification pathway to 1, as shown in Scheme 3.
Table 1
Catalytic activity of Au–CeO2 towards 10 formation starting from 1 and 3.a
Exp.
Reactant
Product
Time (h)
Conversion (mol%)
Selectivity (mol%)
1
2
1
3
10
10
5
2
4
>99
70
>99
>99
>99
>99
a
Reaction conditions: 1.59 mmol
1 or 3 in 20 mL MeOH; 50 mg Au–CeO2
(2.1 wt% Au) (1/Au mol ratio of 300) at 130 °C, 10 bar bubbled O2 at 0.33 mL/s
constant flowrate.
Therefore, it can be said that Au–CeO2 is a selective catalyst that
favours the oxidation of the hemiacetal into ester and the oxida-
tion of alcohol into aldehyde, rather than the reoxidation of the
acetal into the ester. Undoubtedly, the former process is thermody-
namically more favoured owing to the high stability of acetals.
100
90
80
3.2. Comparative study of Au–CeO2 performance with Au
nanoparticles onto different supports
70
4 Conversion (mol%)
60
9 Yield (mol%)
50
Table 3 presents the initial activity of different catalysts (Au–
CeO2, Au–TiO2, Au–Fe2O3 and Au–C) for the oxidation and esterifi-
cation of 1–10 in MeOH (entries 1–4). Meanwhile in Fig. 3, we de-
scribe the kinetic behaviour of Au–CeO2 and Au–TiO2 without the
presence of a base. As can be deduced from the results in Table
3, in the absence of base (entry 4), Au–CeO2 is the most active cat-
alyst of the group (>99 mol% yield to 10, after 5 h). The values of
TOF1 and TOF2 calculated from the conversion of 1 and yield to
10, respectively, reveal (Table 3) that although Au–C (entry 3) is
the most active catalyst (TOF1 480 hÀ1), both TOF2 and the yield to-
wards 10 after 24 h are rather low (0.5 hÀ1 and 7.6 mol%, respec-
tively). Au–TiO2 catalyst (entry 2) is less active than Au–C and it
10 Yield (mol%)
6 Yield (mol%)
40
30
20
10
0
20
12
14
18
16
10
0
2
4
6
8
Reaction time (hr)
Fig. 2. Kinetics plot for the oxidation-esterification of 4 into 10 at 130 °C, 10 bar
bubbled O2 at 0.33 mL/s using Au–CeO2 (4/Au mol ratio 300) as a catalyst.

O. Casanova et al. / Journal of Catalysis 265 (2009) 109–116
113
Table 2
Table 3
Influence of the furane substituent group on hydroxymethyl substituent group
Oxidation–esterification of 1 using Au nanoparticles onto different supports at 130 °C,
10 bar O2 in methanol.
oxidation into methyl ester.a
a
b
Exp. Reactant Product Time
(h)
Conversion
(mol%)
Selectivity to product
(mol%)
Exp. Catalyst
TOF1
TOF2
T
(h)
Conv.
(mol%)
Yield (mol%)c
10
34.7 6.2
(hÀ1
)
(hÀ1
)
3
4
1
2
3
11
4
3
15
9
10
2
2
2
>99
91
72
91
82
>99
1
2
Au–Fe2O3
Au–TiO2
120
305
4
57
24
3
24
24
2
94
97
>99
96
98
>99
58
61
17.5
2
0
17.4
1
0
49
30
3.7
96.3
a
3
4
Au–C
Au–CeO2
480
320
0.5
137
20.2 7.6
Reaction conditions: 1.59 mmol reactant in 20 mL MeOH; 50 mg Au–CeO2
37
0
3.8
0
54
>99
0
(2.1 wt% Au) (reactant/Au mol ratio of 300) at 130 °C, 10 bar bubbled O2 at 0.33 mL/
s constant flowrate.
5
d
5
6
7
Au–CeO2
CeO2
Au–CeO2
–
–
155
–
–
10
24
24
72
51.8
53.1
15
0
e
>99
0
85
displays lower TOF2 (57 hÀ1) than Au–CeO2 (137 hÀ1). Further-
more, the kinetic curves given in Fig. 3 clearly shows that Au–
CeO2 allows to obtain 100% yield to 10 in 5 h reaction time. Mean-
while with the Au–TiO2 reported in [23], only 40% yield to 10 is
achieved in 5 h, and after 24 h the yield is still 96%. Au–Fe2O3 cat-
alyst (entry 1) showed very poor activity. The sole presence of
nanoparticulated CeO2 under oxidizing conditions (entry 6) re-
sulted in the formation of large amounts of acetal 4, together with
4 mol% yield diformylfurane (DFF). Similarly, when N2 was bubbled
with Au–CeO2 catalyst (entry 5), the acetal 4 was the predominant
product. Despite the absence of oxygen, 3.8 mol% of 3 was detected
in the latter experiment. The fact that without oxygen some of the
substrate could be oxidized is due to the fact that nanoparticulated
cerium oxide acts as oxygen supplier [29–31], something consis-
tent with the idea of considering this nanometric support as an
oxygen pump, releasing and adsorbing oxygen through a redox
process where Ce+4/Ce+3 is involved. As seen in Table 3, when no-
nanometric CeO2 (entry 7) was used to support gold nanoparticles,
it took 3 days to reach 85 mol% 10. Results on the comparative
behaviour of Au–CeO2 point at the formation of peroxo and super-
oxo species in monoelectronic defects of nanocrystalline ceria.
Therefore, the catalytic behaviour of gold promoted by the
appropriate support containing Lewis acid sites resulted to be an
optimal combination to perform the reaction in high yields.
a
Calculated as moles of converted 1 divided by reaction time (15 min) and moles
of Au.
Calculated as moles of formed 10 divided by reaction time (15 min) and moles
of Au.
Other by-products which complete molar balance are the following: 5-
methoxymethyl-2,2-dimethoxylfurane, 5-methoxymethyl-2-methylfuroate and 6.
b
c
d
Oxygen was replaced by nitrogen.
No-nanometric CeO2 was used as support.
e
100
90
80
70
60
50
Au-CeO2
Au-TiO2
40
30
20
10
0
3.3. Stability and reusability of Au–CeO2 catalysts
5
15
0
10
20
25
30
Reaction time (hr)
Catalyst recyclability was checked by successive reuses of the
catalyst under the same reaction conditions. To do that, the cata-
lyst was recovered by vacuum filtration and successive washing
with methanol and drying overnight. Calcination of adsorbed or-
ganic compounds was avoided since it could cause gold agglomer-
ation into larger and inactive particles. The results of catalyst
recycling are shown in Fig. 4. As it can be seen, after the first use,
the time required to reach similar levels of yield towards 10
(>96 mol%) than the fresh catalyst was nearly twice and the initial
rate of 1 conversion was almost halved. This is probably due to the
presence of 11 wt% organic material after the first use that in-
creases up to 15 wt% after the fourth use. Unfortunately, the ad-
sorbed organic was resistant to Sohxlet extraction with
Fig. 3. Kinetics plot for the yield of 10 at 130 °C, 10 bar bubbled O2 at 0.33 mL/s
using Au–CeO2 and Au–TiO2 (1/Au mol ratio 300) as catalysts.
methanol. Nevertheless, a full oxidation of the organic on the cat-
alyst was performed at 250 °C in air for 12 h. The regenerated cat-
alyst contained <5% organic and the activity was almost totally
restored, as seen in Fig. 4 (cycle 5).
Catalyst stability from the point of view of gold leaching has
been tested by two different procedures. Firstly, the gold content
in the used catalysts was determined by X-ray fluorescence, and
taking into account the contribution in weight of adsorbed organic
O
O
O
O
O
O
12
OH
OH
O
O
13
11
15
O
O
O
14
Scheme 3. Proposed 11 oxidation–esterification pathway.

114
O. Casanova et al. / Journal of Catalysis 265 (2009) 109–116
4
10
1.3
10
1.1
10
1.2
10
1.2
time (hr)
metal leaching and further redeposition on the surface, known as
‘‘boomerang effect”.
2.1
rate (mmol/hr)
100
90
80
70
60
50
40
30
20
10
0
3.4. Optimization of reaction conditions: HMF/Au ratio and
temperature
The evolution of the yield of the desired diester product 10 with
reaction time, working at 185, 300 and 760 reactant to gold molar
ratio is shown in Fig. 6. As can be observed there, the time required
to reach >99 mol% yield to 10, was ca. 2.5, 5 and 30 h, respectively,
being the initial reaction rate, directly proportional to the concen-
tration of catalyst.
The variation of the conversion of 1 with time is plotted in Fig. 7
for three different temperatures (80, 100 and 130 °C) at a 1/Au mol
ratio of 300. Clearly, the higher the temperature, the higher initial
rate and, consequently, 130 °C was regarded here as a suitable
temperature to perform the reaction. An Arrhenius plot of the log-
arithm of the initial rate gives an apparent activation energy of
1
2
3
4
5
Number of Cycles
Fig. 4. Yield of 10 (mol%) with successive reuses of Au–CeO2. The numbers above
the bar chart stand for the reaction time required for achieving said yield and the
initial rates of 1 conversion for each cycle.
100
90
80
70
60
50
material, leaching of gold was not detected after catalyst reuses. In
agreement with this, no traces of gold were detected in the reac-
tion filtrate. In second and definitive leaching test, the three-phase
reaction test was applied [32,33]. According to the procedure ex-
plained in the experimental part, the IR results on the three-phase
test are shown in Fig. 5. Compound 1 reacted with aminofunction-
alized silica to yield the corresponding imine. Due to the instability
of imine-supported-1 (line c), further reduction into secondary
amine (line d) was necessary in order to avoid the substrate from
passing to the solution. The latter curve corresponds to the actual
three-phase test reactant. After the test was accomplished, the IR
spectrum of used solids (line e) is shown. If we compare the initial
and final solid, no trace of bands corresponding to the ester or alde-
hyde (line f) can be observed. However, when the reaction was per-
formed using homogeneous gold (AuCl3), the bands corresponding
to oxidation products appeared (line g). Therefore, it can be stated
that under the reaction conditions, Au–CeO2 did not suffer from
1/Au 185
40
30
20
10
0
1/Au 300
1/Au 760
15
Reaction time (hr)
30
35
0
5
10
20
25
Fig. 6. Effect of 1/Au mol ratio on 10 yield at 130 °C and 10 bar bubbled O2 in MeOH
using Au–CeO2 catalyst.
10
(a)
100
90
8
(b)
(c)
80
(d)
6
130 ºC
70
(e)
100 ºC
60
4
80 ºC
50
40
30
20
10
0
(f)
2
(g)
0
1750
1700
1650
1600
1550
1500
1450
Wavenumber (cm-1)
0
0.2
0.4
0.6
0.8
1
1.2
Fig. 5. IR spectra for (a) 1 reagent; (b) SiO2–NH2 initial support; (c) silica-supported
imine reagent 1; (d) silica-supported amine reagent before the three-phase-test; (e)
silica-supported amine reagent after the three-phase test; (f) silica-supported imine
product (formyl-furane); (g) silica-supported amine reagent after the homogeneous
reaction.
Reaction time (hr)
Fig. 7. Effect of temperature on 1 conversion at 1/Au mol ratio 300 and 10 bar
bubbled O2 in MeOH using Au–CeO2 catalyst.

O. Casanova et al. / Journal of Catalysis 265 (2009) 109–116
115
100
90
80
70
60
50
40
30
20
10
0
34 kJ/mol which is in the same order as for the oxidation of other
alcohols such as p-methylbenzyl alcohol using Au–CeO2 [18].
3.5. Effect of the alcohol
Several short chain alcohols (methanol, ethanol and n-butanol)
were tested for the oxidation–esterification of 1, and the results are
given in Fig. 8. It can be clearly seen that methanol was the most
suitable alcohol (which here plays the role of reactant and solvent
simultaneously) to obtain the corresponding ester. It has to be
noted that, as reported in the literature [23], between 0.3 and
0.5 mol MeOH was oxidized per mol of reactant, being the typical
oxidation products such as methyl formate, formaldehyde di-
methyl acetal and CO2. It was noticed that when ethanol or butanol
was loaded, then initial rate decreased and time required for full
conversion was enhanced considerably. Also large amounts of ace-
tals from ethanol and n-butanol were formed. Au–CeO2 catalyst, by
being a good alcohol oxidizer, was capable of performing these
side-reactions due to the easiness to form metal–alcoholate ad-
ducts when increasing the alkyl chain.
Water free
10 % WATER
20 % WATER
0
5
10
15
20
25
30
Reaction time (hr)
Fig. 9. Effect of water on 10 yield at 1/Au mol ratio 300, 130 °C and 10 bar bubbled
O2 in MeOH using Au–CeO2 catalyst.
3.6. Effect of water
4. Conclusions
Although Au–CeO2 has proved to withstand water to perform
catalytic oxidations of alcohols [17], we have studied the effect
of water concentration on the catalytic performance, and the re-
sults are given in Fig. 9. A negative effect of water can be observed
when the initial rate was smaller and the concentration of water
was larger. When 20% water was loaded, the reaction was not com-
pleted (68 mol% yield to 10) even after 30 h. Nevertheless, no trace
of carboxylic acid was found. Possibly if acid was formed it would
rapidly react to give the methyl ester due to the presence of a Lewis
acid. It seems that the presence of water may avoid metal–alcohol-
ate adducts from forming on the active site (adsorption step) with-
out which the reaction cannot occur and disfavours the hemiacetal
formation.
Au–CeO2 has proved to be an excellent catalyst for the aerobic
oxidation of HMF into DMF and able to perform this reaction with-
out base. Au–CeO2 can be easily recovered and reused with small
loss of activity but maintaining high selectivity towards DMF. Gold
leaching was ruled out by means of chemical analysis of the mate-
rial and the three-phase test. The reaction kinetics shows that the
oxidative pathway encounters its limiting step for the oxidation of
the alcohol to aldehyde. Once the aldehyde is formed, the corre-
sponding hemiacetal is obtained, which is rapidly oxidized into
the ester. From our results, we can conclude that the alcohol oxida-
tion step is facilitated, if the 5-substituent group is a furane-ring
activating function, in accordance with the previous results in
the literature. Our study on the effect of temperature and the sub-
strate to catalyst ratio demonstrates that the higher the tempera-
ture and the amount of catalyst gives rise to a higher initial
reaction rate, but nevertheless 100% yield of DMF is always ob-
tained. Unfortunately, any alcohol may not be suitable to carry
out this oxidation, because they may also be oxidized and enter
into a similar reaction pathway as the target reactant. Finally,
water has a negative effect on the catalytic behaviour towards
the formation of DMF, as it may impede adsorption of reactant
onto the active site and subsequent oxidation.
100
90
80
70
60
50
40
Acknowledgments
The authors wish to acknowledge the Spanish Ministry of Edu-
cation and Science for the financial support in the project MAT
2006-14274-C02-01 and the PROMETEO project of the Generalitat
Valenciana. O.C. is grateful to the Spanish Council for Scientific Re-
search (CSIC) for a Ph.D. scholarship.
MeOH
30
n-ButOH
20
EtOH
References
10
[1] A. Corma, S. Iborra, A. Velty, Chem. Rev. 207 (2007) 2411.
[2] P. Mäki-Arvela, B. Holmbom, T. Salmi, D.Y. Murzin, Catal. Rev. 49 (2007) 197.
[3] G. Budroni, A. Corma, J. Catal. 257 (2008) 403.
[4] Y. Román-Leshkov, C.J. Barret, Z.Y. Liu, J.A. Dumesic, Nature 447 (2007) 982.
[5] K. Weissermel, H.J. Arpe, Industrial Organic Chemistry, VCH–Wiley, Weinheim,
1997, p. 255.
[6] C. Moreau, M.N. Belgacem, A. Gandini, Top. Catal. 27 (2004) 11.
[7] A. Fushiya, Japan Patent 075068 (2008).
[8] M.L. Ribeiro, U. Schuchardt, Catal. Commun. 4 (2003) 83.
0
0
5
10
15
20
25
Reaction time (hr)
Fig. 8. Effect of the alcohol on the yield towards the corresponding ester from 1 at
1/Au mol ratio 300, 130 °C and 10 bar bubbled O2 using Au–CeO2 catalyst. Acetals
from acetaldehyde and butyraldehyde were also formed in high concentrations
when EtOH and n-ButOH were used.

116
O. Casanova et al. / Journal of Catalysis 265 (2009) 109–116
[9] W. Partenheimer, V.V. Grushin, Adv. Synth. Catal. 343 (2001) 102.
[22] A. Corma, H. García, Chem. Soc. Rev. 37 (2008) 2096.
[23] E. Taaring, I.S. Nielsen, K. Egeblad, R. Madsen, C.H. Christensen, Chem. Sus.
Chem. 1 (2008) 75.
[10] P. Verdeguer, N. Merat, A. Gaset, J. Mol. Catal. 85 (1993) 327.
[11] M. Kröger, U. Prüsse, K.D. Vorlop, Top. Catal. 13 (2000) 237.
[12] T. Mallat, A. Baiker, Chem. Rev. 104 (2004) 3037.
[13] K. Mori, T. Hara, T. Mizugaki, K. Evitan, K. Kaneda, J. Am. Chem. Soc. 126 (2004)
10657.
[14] W. Hou, N.A. Dehm, R.W.J. Scott, J. Catal. 253 (2008) 22.
[15] D.I. Enache, J.K. Edwards, P. Landon, B. Solsona-Espriu, A.F. Carley, A.A. Herzing,
M. Watanabe, C.J. Kiely, D.W. Knight, G.J. Hutchings, Science 311 (2006) 4776.
[16] N. Dimitratos, A. Villa, D. Wang, F. Porta, D. Su, L. Prati, J. Catal. 244 (2006) 113.
[17] A. Abad, P. Concepción, A. Corma, H. García, Angew. Chem., Int. Ed. 44 (2005)
4066.
[24] F. Moreau, G.C. Bond, Appl. Catal. A: Gen. 302 (2006) 110.
[25] L. Zhang, S.A. Kozmin, J. Am. Chem. Soc. 126 (2004) 11806.
[26] F. Porta, L. Prati, M. Rossi, S. Coluccia, G. Martra, Catal. Today 61 (2000) 65.
[27] L.L. Santos, V.R. Ruiz, M.J. Sabater, A. Corma, Tetrahedron 64 (2008) 7902.
[28] P. Fristrup, L.B. Johansen, C.H. Christensen, Catal. Lett. 120 (2008) 184.
[29] Q. Fu, H. Saltsburg, M. Flytzani-Stephanopoulos, Science 301 (2003) 935.
[30] J. Guzman, S. Carretin, A. Corma, J. Am. Chem. Soc. 127 (2005) 3286.
[31] J. Guzman, S. Carretin, J.C. Fierro-González, L. Hao, B.C. Gates, A. Corma, Angew.
Chem., Int. Ed. 44 (2005) 4778.
[18] A. Abad, C. Almela, A. Corma, H. García, Tetrahedron 62 (2006) 6666.
[19] A. Abad, A. Corma, H. García, Chem. Eur. J. 14 (2008) 212.
[20] A. Abad, C. Almela, A. Corma, H. García, Chem. Commun. (2006) 3178.
[21] A. Abad, A. Corma, H. García, Pure Appl. Chem. 79 (2007) 1847.
[32] B.H. Lipshutz, S. Tasler, W. Chrisman, B. Spliethoff, B. Tesche, J. Org. Chem. 68
(2003) 1177.
[33] I.W. Davies, L. Matty, D.L. Hughes, P.J. Reider, J. Am. Chem. Soc. 123 (2001)
10139.