Catalysis and Reactivation of Ordered Mesoporous Carbon-Supported Gold Nanoparticles for the Base-Free Oxidation of Glucose to Gluconic Acid

Article abstract of DOI:10.1021/cs502093b

An ordered mesoporous carbon (CMK-3)-supported gold catalyst was prepared and used in the aerobic oxidation of glucose to gluconic acid under base-free conditions with molecular oxygen. XRD and TEM results revealed that gold nanoparticles were uniformly dispersed on the surface and in the channels of CMK-3. Catalytic tests showed that conversion remarkably increased with decreased selectivity when oxygen pressure and reaction temperature were increased. Glucose conversion to gluconic acid reached over 92% with 85% selectivity under the conditions of 383 K reaction temperature, 0.3 MPa oxygen pressure, and 2 h reaction time. Hydrogen peroxide was generated during reaction, and the relationship between hydrogen peroxide and the byproduct fructose was discussed. Low glucose/Au molar ratio minimized fructose formation. A 92% gluconic acid yield was obtained after reaction for 15 min when the molar ratio of glucose/Au was set to 100. The spent catalyst treated with an aqueous solution of NaOH at 363 K could produce glucose conversion up to 87%, which was close to the result of as-prepared catalyst and excluded the effect of alkaline residues. (Chemical Equation Presented).

Full text of DOI:10.1021/cs502093b

Research Article
pubs.acs.org/acscatalysis
Catalysis and Reactivation of Ordered Mesoporous Carbon-
Supported Gold Nanoparticles for the Base-Free Oxidation of
Glucose to Gluconic Acid
Puyu Qi, Shasha Chen, Jin Chen, Jianwei Zheng, Xinlei Zheng, and Youzhu Yuan*
State Key Laboratory of Physical Chemistry of Solid Surfaces and National Engineering Laboratory for Green Chemical Production of
Alcohols-Ethers-Esters, Collaborative Innovation Center of Chemistry for Energy Materials, College of Chemistry and Chemical
Engineering, Xiamen University, Xiamen 361005, China
*
S Supporting Information
ABSTRACT: An ordered mesoporous carbon (CMK-3)-
supported gold catalyst was prepared and used in the aerobic
oxidation of glucose to gluconic acid under base-free
conditions with molecular oxygen. XRD and TEM results
revealed that gold nanoparticles were uniformly dispersed on
the surface and in the channels of CMK-3. Catalytic tests
showed that conversion remarkably increased with decreased
selectivity when oxygen pressure and reaction temperature
were increased. Glucose conversion to gluconic acid reached
over 92% with 85% selectivity under the conditions of 383 K
reaction temperature, 0.3 MPa oxygen pressure, and 2 h
reaction time. Hydrogen peroxide was generated during reaction, and the relationship between hydrogen peroxide and the
byproduct fructose was discussed. Low glucose/Au molar ratio minimized fructose formation. A 92% gluconic acid yield was
obtained after reaction for 15 min when the molar ratio of glucose/Au was set to 100. The spent catalyst treated with an aqueous
solution of NaOH at 363 K could produce glucose conversion up to 87%, which was close to the result of as-prepared catalyst
and excluded the effect of alkaline residues.
KEYWORDS: gold, mesoporous carbon, glucose oxidation, base-free condition, gluconic acid
1
. INTRODUCTION
with traditional Pt, Pd, and Pt/Pd bimetallic catalysts. The
results revealed that the Au catalyst was obviously superior to
others because it exhibited low sensitivity to oxygen poisoning.
Another advantage of the Au catalyst was that it showed activity
in a wide range of pH values.
The transformation of biorenewable feedstock to highly
attractive chemical intermediates is becoming an effective
method of minimizing pollution and improving the efficiency of
energy utilization. Gluconic acid, mostly produced from the
selective oxidation of glucose, is widely used in many fields,
including construction industries, food additives, pharmaceut-
1
,2
The oxidation of glucose to gluconic acid was mostly
undertaken under alkaline conditions because the presence of a
15−17
3
−5
base could significantly improve the catalytic activity.
Rossi
ical raw materials, and paper industries.
Gluconic acid is
14
et al. found that the solution conditions could affect the
stability of the catalyst because the decline degree of conversion
was found to be related to the pH value in the recycling
experiments. A relatively high pH condition (pH = 9.5) was
beneficial for the stability without any leaching of Au. By
contrast, almost 16% of the Au was lost when the pH was fixed
at 7.0. The loss of Au and the aggregation of Au nanoparticles
basically manufactured through fermentation using Aspergillus
niger and Gluconobacter suboxydan, but economic and environ-
mental problems may be a serious weakness in the case of large-
3
,6,7
scale production.
By contrast, heterogeneous catalysis for
such transformation is receiving considerable attention because
8
−10
it is eco-friendly.
Developing heterogeneous catalysts as
replacement for enzymes is becoming increasingly important.
In early reports, platinum (Pt) and palladium (Pd) were
commonly used in this reaction with lead or bismuth as the
second metal. Although beneficial effects were achieved with
high selectivity and a conversion above 99%, low durability was
(
Au NPs) could lead to a drastic drop in conversion after four
runs.
Aside from carbon-supported catalysts, metal oxide supports
18−21
have also been intensively researched.
A 0.45% Au/TiO2
1
0,11
catalyst was tested for stability under 313 K, pH 9.0, and 250
observed from the leaching of the second metal.
Gold (Au)
is also widely used in the aerobic oxidation of carbohydrates
1
2,13
with high catalytic activity and selectivity.
Rossi and co-
Received: December 26, 2014
Revised: March 7, 2015
1
4
workers studied the performance of Au/C for glucose
oxidation under various reaction conditions and compared it
©
XXXX American Chemical Society
2659
DOI: 10.1021/cs502093b
ACS Catal. 2015, 5, 2659−2670

ACS Catalysis
Research Article
mmol/L glucose concentration. The value of the turnover
frequency generally stayed constant for 17 runs when the
reaction time was prolonged for the last batches. Au NPs
showed almost no agglomeration after the recycling experi-
pressures. The appearance of fructose was obviously odd
because it was commonly produced under alkaline conditions
in this reaction. By eliminating the fructose production and
minimizing it at the bottom, a gluconic acid yield of ∼92%
could be achieved. Furthermore, the catalyst stability and
different regeneration methods were studied intensively. The
results showed that the spent catalyst treated with an aqueous
NaOH solution at 363 K could give a glucose conversion of
87%, which was quite close to that of the as-prepared catalyst
and excluded the effect of alkaline residues.
1
8
ments. A 0.3% Au/Al O catalyst was used in a continuous-
2
3
flow system of glucose oxidation, showing outstanding
19
performance and stability in 110 days of operation. Wang
2
0
et al. studied the stability of Au supported on metal oxides for
the base-free oxidation of glucose. Reactions were carried out at
338 K and 2.3 bar for 6 h with a glucose/Au molar ratio of 140.
Approximately 30% and 60% decreases in conversion were
observed for Au/n-CeO (for nanosized CeO ) and Au/μ-
2. EXPERIMENTAL SECTION
2
2
CeO (for microsized CeO ), respectively. According to the
report, the relatively more stable nature of the Au/n-CeO2
catalyst was due to the low surface density of Au NPs and more
anchoring sites that could stabilize the Au NPs against particle
growth.
2
2
2.1. Catalyst Preparation. The preparation of SBA-15
28
followed the procedure described by Zhao et al. The synthesis
of CMK-3 was carried out using sucrose as the carbon source
and SBA-15 as the hard template according to known
29
procedures. Briefly, 1.0 g of SBA-15, 1.25 g of sucrose, 5 g
of deionized water, and 0.14 g of concentrated sulfuric acid
were mixed. The mixture was successively treated at 373 and
Another factor that caused deactivation was the absorbed
carboxylic acids, which could be removed by calcination.
However, the calcination step would lead to a severe loss of
activity because the diameter of the Au NPs was obviously
4
33 K for 6 h. When the white mixture turned dark brown,
another 0.8 g of sucrose, 5.0 g of deionized water, and 0.07 g of
concentrated sulfuric acid were added. Afterward, thermal
treatment was performed again. The black solid substance was
calcined at 1223 K for 5 h with 2 K min− heating rate under
argon. The silicon skeleton was removed with 30% hydrofluoric
acid. Finally, the resultant mesoporous carbon material was
filtered, washed several times with ethanol, and dried overnight
at 383 K.
21
larger than that of the as-prepared catalyst. Miedziak et al.
prepared 0.5% Au−Pd/MgO and carried out the reaction under
a mild condition of 333 K and atmospheric pressure for 24 h.
The highlight was the use of air as oxidant instead of pure
oxygen, and the catalyst achieved a conversion of 62% under a
base-free condition. However, the catalyst presented a drastic
decrease in activity, giving a conversion of less than 20% in the
second use. The conversion could be improved by only ∼5%
through a calcination treatment. The author ascribed the
catalyst deactivation to product inhibition and support
dissolution.
1
A suitable amount of the stabilizing agent sodium gluconate
was added into 125 mL of diluted HAuCl aqueous solution
4
and stirred at room temperature for 4 h. Freshly prepared
NaBH aqueous solution was rapidly injected into the mixture.
4
The stabilities of the above-mentioned catalysts in the base-
free oxidation of glucose are not satisfactory, and a useful
regeneration method has not been established. In fact, except
for glucose oxidation, gold catalysts also suffered from
deactivation caused by carboxylic compounds or intermediates
When the solution turned bright orange, CMK-3 was added.
The solution was further stirred vigorously at room temper-
ature for 2 h, filtered, and then washed with distilled water
several times until no trace of chloride was found by the
AgNO test. The sample was finally dried overnight at 383 K
3
22
in the oxidation of carbohydrates such as octanol and benzyl
for the activity test without calcination. Given that the
stabilizing agent itself was a sodium salt, it was assumed to
have only a slight influence, even if traces remained on the
catalyst surface.
2
3
alcohol, etc. Adding sacrificial base could effectively solve this
problem; however, the addition of base would generally
produce the corresponding salt instead of pure acid and
cause difficulties to the post-treatment of the reaction liquid
and therefore being environmentally unfriendly. The problem
of corrosion and waste base disposal was also presented in the
The other carbon supports including carbon nanotubes
(CNTs), activated carbon (AC), graphite, and graphene were
calcined at 1223 K under argon to remove the surface oxidation
groups and keep the carbon supports similar to the CMK-3
support. Prior to thermal treatment, the as-received CNTs and
AC were first purified with concentrated nitric acid (68 wt %)
at 353 K for 16 h under refluxing to remove the residues left in
the synthesis step, such as nickel and iron cations. Afterward,
the treated CNTs and AC were washed with distilled water
several times until the filtrate became neutral and dried
overnight at 383 K. The graphite and graphene support were
directly calcined at 1223 K without the nitric acid treatment
because their purities were acceptable. The metal-loading step
was completely the same as above. As for the preparation of
Au/SBA-15 catalyst, SBA-15 was first functionalized with 3-
aminopropyltriethoxysilane (APTES) to modify the surface
24
oxidation of aliphatic alcohols; hence, developing a gold
catalyst that has excellent performance and favorable stability
for the base-free oxidation of carbohydrates is highly desired.
Given the high surface area and uniform mesopore size of the
ordered mesoporous carbon, this material has been intensively
25−27
2
explored as a catalyst support.
Ma et al. discovered the
superior performance of Au supported on ordered mesoporous
carbon in the aerobic oxidation of glucose. The superior activity
was ascribed to the mesopore system that could not be blocked
by metal, and the remaining space was enough for the diffusion
of molecules in the reaction. Thus, the contact between Au and
glucose and the mass transfer of gluconic acid were facilitated.
14
Previously, Rossi et al. reported that high temperature and
pressure could improve the performance of Au/C catalyst, but
they did not discuss the effects in detail. To clarify this effect,
CMK-3, a kind of ordered mesoporous carbon, was chosen in
the present study as the catalyst support to load Au NPs
without any calcination step for the base-free oxidation of
glucose to gluconic acid under different temperatures and
group. The solid formed, which was named SBA-15-NH , was
2
then stirred with HAuCl aqueous solution for 30 min. A
4
freshly prepared NaBH solution was rapidly injected into the
4
mixture. After stirring for another 1 h, the solution was filtered,
washed, and dried overnight. For comparison, ZrO support
2
was prepared using Zr(NO ) as ingredient according to the
3
4
2
660
DOI: 10.1021/cs502093b
ACS Catal. 2015, 5, 2659−2670

ACS Catalysis
Research Article
Figure 1. (a) Small-angle XRD patterns of CMK-3 and Au/CMK-3 catalyst; (b) TEM images of CMK-3; (c) TEM image of Au/CMK-3 catalyst;
and (d) Au particle size distribution of Au/CMK-3.
3
0
procedure described in the literature. A prepared ZrO was
NCS Plasma1000. All the catalysts with carbon supports were
calcined in a muffle furnace at 873 K (graphene and graphite at
1223 K, for they could not be burned away at 873 K) for 4 h
with 2 K min− heating rate to remove the support. The
remaining Au was dissolved with aqua regia at room
temperature and further diluted with 5% HCl in a 25 mL
volumetric flask before measurement. The Au/SBA-15 and Au/
2
−1
mixed with HAuCl aqueous solution, and then 0.1 mol L
ammonium hydroxide was added dropwise to adjust the pH to
.0. After stirring for 2 h, the mixture was filtered and washed
with a filter membrane and then dried overnight at 383 K. The
pale yellow solid was finally reduced at 473 K for 2 h with 2 K
min heating rate under N /H flow.
4
1
9
−1
2
2
2
.2. Catalyst Characterization. X-ray diffraction (XRD)
ZrO catalysts were added into excess aqua regia and boiled for
2
patterns were obtained using a Rigaku Ultima IV equipped with
Cu K radiation (λ = 0.15418 nm) scanned from 10° to 90°.
The scanning rate was 10° min , and the voltage and current
20 min. The solution was syringed out and filtered after cooling
down and similarly diluted with 5% HCl in a 25 mL volumetric
flask.
α
−1
were 35 kV and 15 mA, respectively.
The ultraviolet−visible (UV−vis) spectra were recorded on a
Shimadzu UV2550 UV−vis spectrometer. The reaction mixture
was syringe filtered with a microfiltration membrane (0.22 μm),
and then 15 mL of the clarified solution was added into a
volumetric flask and filled to 25 mL with excess titanyl sulfate.
The solution turned yellow and was kept in a refrigerator prior
to the measurement. A series of standard solutions (10, 20, 30,
N adsorption−desorption isotherms were measured by the
2
static volumetric method at 77 K using a Micromeritics TriStar
II 3020 porosimetry analyzer. Prior to the measurements, all
samples were degassed at 573 K for 4 h. The specific surface
area (SBET) was calculated according to the Brunauer−
Emmett−Teller (BET) method, adopting the data within the
−1
relative pressure range P/P of 0.05−0.2. The total pore volume
40, and 50 μg mL ) were measured to quantify the exact
amount of hydrogen peroxide.
0
was determined from the adsorbed amount of N at a relative
2
pressure of ∼0.995. The average pore diameter and pore size
distributions were evaluated using the desorption branch of
isotherm in line with the Barrett−Joyner−Halenda method.
Transmission electron microscopy (TEM) images were
derived on a Philips Analytical FEI Tecnai 30 electron
microscope operated at a 300 kV acceleration voltage. Before
observation, the samples were ultrasonically dispersed for 30
min in ethanol at room temperature and then dropped on a
carbon-coated copper grid. Ethanol was dried off under infrared
light irradiation.
Cyclic voltammetry (CV) experiments were carried out with
a LabNet UI5020 electrochemical workstation with a three-
electrode cell according to the method described in Cui’s
31
work. A glassy carbon electrode was used as the working
electrode. It was polished with alumina polishing powder
(Gaoss Union, 0.05 μm) and ultrasonically treated with
ultrapure water. This procedure was repeated three times
before dropping the catalyst. A platinum electrode and
saturated calomel electrode were used as the counter and
reference electrodes, respectively. Exactly 10 mg of catalyst was
added into the mixed liquor that contained 0.5 mL of ultrapure
water, 0.45 mL of ethanol, and 0.05 mL of naphthol. The
The exact Au loadings were calculated by inductively coupled
plasma atomic emission spectrometry (ICP-AES) using an
2
661
DOI: 10.1021/cs502093b
ACS Catal. 2015, 5, 2659−2670

ACS Catalysis
Research Article
2
−1
mixture was ultrasonically treated for 30 min and then coated
onto the glassy carbon electrode. All experiments were
conducted at room temperature under stirring. Current−time
data were adopted when the current was stabilized.
value decreased to 956 m g after loading about 1% Au
(Table 1 and Table S1). This result confirmed that a part of the
Table 1. Structural Properties of the Tested Catalysts
Thermogravimetric analysis (TGA) was carried out on a
TG_209F1 thermogravimetric analyzer. After drying overnigh1t
at 373 K, the catalysts were treated at 1073 K with 10 K min−
heating rate in a nitrogen flow.
S
,
Vtotal
,
Vmeso,
−1
Dmeso
,
Au
D ,
BET
m g
Au
nm
2
−1
−1
sample
mL g
mL g
nm
content, %
Au/CMK-3
Au/SBA-15
Au/CNTs
Au/graphene
Au/graphite
Au/AC
956
491
128
1.17
0.62
0.38
0.12
0.04
0.44
0.01
0.84
0.51
0.34
0.11
0.03
0.16
4.0
4.6
0.94
0.91
0.96
0.83
0.86
0.89
0.95
2.98
3.56
2.92
3.22
3.73
3.50
2.52
XPS measurements were conducted on a Phi Quantum 2000
Scanning ESCA Microprobe instrument using an Al Kα
radiation source (hν = 1486.6 eV). Prior to the measurements,
each sample was pressed into a thin disk. The XPS spectra of all
the samples were recorded at room temperature. The binding
energy was calibrated using C 1s peak at 284.6 eV as reference
with an uncertainty of ±0.2 eV.
12.5
9.2
34.2
11.6
778
29.0
10.3
4.0
Au/ZrO2
0.005
42.6
Au was confined in the pore channels. The XRD patterns in
Figure S2 show that Au NPs were uniformly dispersed on other
supports, such as SBA-15, ZrO , CNTs, graphite, and graphene,
because no obvious peak belonging to Au species could be
seen.
Furthermore, a strong glucose response was observed for the
Au/CMK-3 catalyst by CV. Figure 2 shows the CVs data of 0.1
The catalysts were also analyzed by CO2 temperature-
programmed desorption (CO -TPD) on a Micromeritics
2
2
Autochem II 2920 instrument with a thermal conductive
detector (TCD). All the tested catalysts were dried overnight at
3
83 K prior to the measurement. Typically, 40 mg of the
catalyst sample was put into a quartz U tube, and CO was
2
passed through the catalyst bed for 1 h at room temperature.
Weakly adsorbed CO was removed by pure Ar sweeping until
2
the baseline of TCD signal stabilized. Under an atmosphere of
−1
pure Ar with a rate of 50 mL min , CO -TPD was conducted
2
from room temperature to 973 K at a heating rate of 10 K
−
1
min .
2
.3. Catalytic Testing. The catalytic performances were
tested in a 48 mL pressure bottle. In a typical run, 20 mL of 0.1
mol L− aqueous glucose solution and 0.0396 g of catalyst were
used. The system was first vacuumized with a vacuum pump
and then fed with oxygen. This process was repeated several
times before the reaction. In addition to the normal reaction
time, the reaction was allowed to proceed for an additional 10
min to ensure that the system reached the exact reaction
temperature. The bottle was rapidly immersed in a water bath
to cool, and the oxygen was turned off at the same time to stop
the reaction. Afterward, the reaction liquid was immediately
syringed out, filtered, and analyzed using high-performance
liquid chromatography (HPLC) (Shimadzu LC20A) with
glycol as internal standard. The HPLC was equipped with a
Shodex Sugar Column (SH 1011) and a refractive index
detector.
1
−1
Figure 2. Cyclic voltammetric responses to 0.1 mol L NaOH with
1
glucose (dotted line) for different catalysts.
−1
−1
8 mmol L glucose (solid line) and to 0.1 mol L NaOH without
mol L−1 NaOH with 18 mmol L−1 glucose for different carbon
catalysts. A strong cathodic peak at about −0.6 V could be
observed for the Au/CNTs catalyst, regardless of the existence
of glucose. To check the origin of this peak, we tested the cyclic
voltammogram of pure CMK-3 and CNTs in NaOH without
glucose, and the results are presented in Figure S3a. Obviously,
both of them exhibited almost the same CV shape as the
corresponding catalysts, indicating the cathodic peak at about
In all cases, the catalysts were recovered after reaction by
centrifuging for 5 min (3000 rpm). The black solid was washed
with distilled water several times until the supernatant became
neutral. The washed catalysts were dried at 383 K for at least 4
h prior to reuse.
−
0.6 V originated from the CNTs. The peak at −0.6 V was
considered because the electrochemical reduction of oxygen
32
existed in the NaOH solution. Because all the carbon
supports were calcined at 1223 K under argon, their surface
chemical properties were supposed to show little difference.
Previously, it was reported that the CNTs exhibited more
excellent catalytic activity than graphite in the electrochemical
reduction of oxygen in KOH solution because of the cracks
3
. RESULTS AND DISCUSSION
.1. Catalyst Characterization. The physical properties
3
and morphology of the mesoporous carbon CMK-3 were
studied by a series of characterizations. An ordered 2D-
hexagonal structure was represented in the small angle XRD
patterns in Figure 1, which was well maintained even after Au
loading. Typical P6mm symmetry structure could also be
observed in the TEM images. In addition, Au NPs that were
uniformly dispersed on the surface and in the mesopore
channels of the CMK-3 support could clearly be identified. The
N adsorption−desorption isotherms and pore size distribution
patterns are shown in Figure S1. The obvious hysteresis loops
also verified the typical mesostructure of the material. The
support showed a BET surface area of ∼1110 m g , and the
32
existing on the surface of the CNTs. The roughness (for
example cavities and cracks) of the electrode surface was also
33
reported to affect the interactions with analytes in solution.
For the four different carbon supports, CNTs were meant to be
advantageous in the electrochemical reduction of oxygen
because of the fishbone-shaped structure in Figure S3b which
was supposed to bring in more defects on the surface. In
addition, we could see that except for the Au/CNTs catalyst,
the Au/graphite catalyst also exhibited a small peak at about
2
2
−1
2
662
DOI: 10.1021/cs502093b
ACS Catal. 2015, 5, 2659−2670

ACS Catalysis
0.5 V. Considering the morphology of graphite, this might be
Research Article
−
Table 3. Catalytic Behaviors of Au/CMK-3 Catalyst for Base-
Free Oxidation of Glucose under Different Temperatures
ascribed to its layer structure, which would cause defects on the
edge of each layer. By contrast, the border of CMK-3 and AC
were relatively more saponaceous, thus being disadvantageous
in the electrochemical reduction of oxygen. The glucose
oxidation peak currents for the Au/CMK-3 and Au/CNTs
catalysts were much larger than those of the Au/graphite and
Au/AC catalysts. The anodic sweep began at −0.6 V for the
Au/CMK-3 catalyst, which was more negative than that of the
others. Therefore, the Au/CMK-3 and Au/CNTs catalysts have
potential for the selective oxidation of glucose.
a
and Oxygen Pressures
selectivity, %
temp, pressure
conv, %
gluconic acid
fructose
others
333 K, 0.1 MPa
333 K, 0.3 MPa
363 K, 0.1 MPa
363 K, 0.3 MPa
383 K, 0.1 MPa
383 K, 0.3 MPa
22.5
29.3
59.2
73.1
66.3
92.4
100
100
93.7
92.1
90.1
87.5
0
0
0
0
4.5
5.1
6.4
7.4
1.8
2.8
3.5
5.1
3
.2. Catalytic Results. Table 2 summarizes the catalytic
a
performances of several catalysts, including Au supported on
Reaction conditions: glucose = 0.36 g, H O = 20 mL, catalyst =
2
0
.0396 g, glucose/Au (molar ratio) = 1000/1, reaction time = 2 h.
Table 2. Base-Free Oxidation of Glucose with Au NPs
a
Supported on Different Carriers
relatively low, which was consistent with the results in
1
4,34
references;
however, as the temperature increased, certain
selectivity, %
byproducts appeared. Among them, fructose was the most
conv, % gluconic acid fructose others TOF s−1
b
catalyst
14,34
dominant. According to previous reports,
fructose was
Au/CMK-3
Au/SBA-15
Au/CNTs
Au/graphene
Au/graphite
Au/AC
92.4
67.0
62.0
55.6
54.5
20.8
12.7
87.5
92.4
82.7
74.0
84.1
91.4
91.9
7.4
3.8
5.1
3.8
4.92
4.76
2.40
1.75
1.68
0.61
0.35
produced through the isomerization of glucose, which occurred
only under alkaline conditions. Given that no sacrificial base
was added into the reaction system, the appearance of fructose
was worth discussing. The same tendency was also observed
with regard to pressure. The improvement was more obvious
when the reaction pressure was increased at a high temperature.
When the temperature was relatively low at 333 K and the
pressure was increased from atmospheric pressure to 0.3 MPa,
the conversion ascended from 22% to 29%. However, the
enhancements were 14% for 363 K and 26% for 383 K,
indicating that high conversion is achieved much more easily
with the combination of high temperature and pressure
compared with increasing only one of them. However, with
gradually increasing pressure, the selectivity toward gluconic
acid showed a measurable reduction. The reason why fructose
was generated and its relationship with the reaction conditions
would be discussed hereinafter.
15.5
11.0
10.6
6.1
1.8
15.0
5.3
2.5
Au/ZrO2
5.6
2.5
a
Reaction conditions: glucose = 0.36 g, H O = 20 mL, catalyst =
2
0
.0396 g, glucose/Au (molar ratio) = 1000/1, T = 383 K, P(O ) = 0.3
2
b
MPa, reaction time = 2 h. Calculated by the conversion at 5 min.
CMK-3, SBA-15, ZrO , CNTs, graphite, and graphene for the
2
base-free oxidation of glucose. Because the exact Au amount
was less than the theoretical one, the TOF values of these
catalysts were calculated. On the basis of the data in Table 2,
the catalytic activity of Au/CMK-3 was much better than those
of the other chosen materials. A correlation could be observed
that the conversion of glucose was basically positively
correlated to the volume of mesopore (Table 1). This
phenomenon might have resulted because the mesostructure
could facilitate the contact between glucose molecules and Au
NPs. In addition, the mass transfer process was also favored.
The mesopore size of Au/CMK-3 was ∼4.0 nm, which was
slightly larger than the average size of Au NPs. According to the
3.3. Catalyst Deactivation and Regeneration. Recycling
experiments were conducted to investigate the stability and
deactivation of the catalyst. Figure S4 shows the conversion and
selectivity data of four subsequent runs under the reaction
conditions of 383 K and 0.3 MPa. Compared with the
2
0,21
performance of catalysts in previous works
under base-free
conditions, the Au/CMK-3 catalyst was relatively more
favorable; however, its activity eventually decreased to a certain
degree and had a conversion of 70% at the fourth run.
2
theoretical model, mesopores with a suitable size that are
partially occupied by Au NPs but still have space for the
diffusion of glucose and gluconic acid molecules would show a
good activity. Among all the tested catalysts, only Au/CMK-3
and Au/AC had the most suitable pore size of 4.0 nm. Given
that the mesopore volume of AC was too low, CMK-3 was
obviously the most beneficial support for the reaction. The Au/
SBA-15 catalyst, which had a pore size of 4.6 nm, also showed a
satisfactory conversion of ∼70%. Particularly, the TOF value
with the Au/SBA-15 catalyst was 4.76 s , which meant that the
reaction proceeded quite fast in the initial stage; however, the
final conversion over it was 20% lower than the Au/CMK-3
catalyst, indicating the glucose oxidation was retarded for the
rest of the reaction time. The reason will be discussed in
Section 3.3.
Because of the large specific surface area and the poor
hydrophilicity of CMK-3 support, the mass loss of the catalyst
was actually very serious through the washing steps, which was
considered as one of the main reasons why the conversion of
glucose dropped from 92% in the first run to 70% in the fourth
run. In general, ∼30% mass loss of the catalyst was observed
after the consecutive four runs, so the actual weight of catalyst
in the fourth run could be regarded as about 0.03 g {0.0396 ×
(1−0.3) ≈ 0.03}. We carried out a comparison experiment with
0.03 g of fresh catalyst to evaluate the impact of weight loss and
achieved a conversion of 87%, which was a little lower than
92% but much higher than 70%. Hence, we could say that the
mass loss of the catalyst was actually a factor that caused the
decrease of conversion, although not the crucial one.
−
1
A series of experiments were carried out over a range of
different temperatures and pressures using Au/CMK-3 as
catalyst. On the basis of the data in Table 3, increasing the
temperature could effectively improve glucose conversion. No
byproduct was produced when the reaction temperature was
To explore the other factors that cause deactivation, TEM
and ICP-AES measurements were carried out to test the size
and content of Au. The results of the particle size distribution
in Figure 3 (more than 100 particles were accounted) revealed
2
663
DOI: 10.1021/cs502093b
ACS Catal. 2015, 5, 2659−2670

ACS Catalysis
Research Article
Figure 3. (a) TEM image of Au/CMK-3 catalyst after reuse for four runs; (b) Au particle size distribution of Au/CMK-3 catalyst after reuse for four
runs.
that Au NPs grew in a very small degree after the reaction. The
difference at this level could just be ignored, considering the
system and human errors. The ICP-AES results indicated that
corresponds to the elimination of hydroxyl functionalities, and
the last thermal decline is caused by the degradation of
disordered or amorphous carbon. The weight loss percentages
at different stages are respectively arranged in Table 4
according to the above information.
30% of Au was lost after four runs. However, given that the
adsorption of carboxylic compounds was commonly observed
in the selective oxidation of alcohols to acids, gluconic acid
might occupy a certain weight in the tested sample. Hence, this
result was not that accurate.
Table 4. Weight Loss of Fresh, Used, and Recovered Au/
a
CMK-3 Catalysts at Different Temperature Ranges
To eliminate the negative influence of the adsorbed
carboxylic acids, two possible methods were assumed, and
their regenerating effects were compared. The used catalysts
were divided into two groups. One group was treated with
NaOH aqueous solution. Exactly 0.12 g of used catalyst was
weight loss at different temperature ranges, wt %
treatment condition
fresh
RT−423 K 423−623 K 623−773 K >773 K
0.7
0.5
1.6
1.1
1.6
0.6
0.9
23.1
5.1
0.6
4.7
4.0
4.2
3.8
0.8
1.9
4.0
4.6
4.1
4.0
4.5
used
−1
added into 50 mL of 0.1 mol L NaOH, and the mixture was
stirred for 30 min at room temperature, 323, and 363 K,
separately. These catalysts were centrifuged later and washed
with distilled water until the supernatant became neutral.
Finally, they were dried overnight at 383 K. The other group
NaOH
treatment
RT
323 K
363 K
5.2
4.4
calcination
treatment
773 K in
Ar
0.5
−1
773 K in
N2
1.6
1.5
1.3
9.7
was calcined at 773 K for 5 h with 4 K min heating rate under
argon and nitrogen, separately.
a
Assignments for the species at different temperature ranges: RT−423
The TGA curves in Figure 4 show the thermal weight loss of
K, residual water; 423−623 K, carboxylic groups; 623−773 K, hydroxyl
3
5,36
fresh, used, and all the recovered catalysts. It has reported
groups; >773 K, amorphous carbon.
that the weight loss from room temperature to 423 K
corresponds to the evaporation of residual water, the fell
stage at 423 to 623 K belongs to the decarboxylation of
carboxylic groups, the third stage between 623 and 773 K
A comparison between the fresh and used catalysts clearly
shows that the amount of both surface carboxylic and hydroxyl
groups increased. Particularly, the carboxylic groups gained
their amount quite largely from only 0.9% to 23.1%. This result
was direct and obvious proof that the produced carboxylic
compounds were strongly adsorbed on the surface of the
catalyst. It might remind us of the previous problem that, with a
−1
higher initial TOF value of 4.72 s , the Au/SBA-15 catalyst yet
exhibited a relatively lower conversion than Au/CMK-3 at final.
Because the adsorption of carboxylic acid was severe in this
reaction, considering the APTES modifying procedure through
catalyst preparation, the abundant surface amino groups might
help adsorb more acidic compounds and accelerate catalyst
deactivation. Actually, we have also prepared the Au/SiO2
catalyst with the same modifying step using APTES and tested
the catalytic activity under identical reaction conditions.
Because of its large mesopore volume (1.03 mL g ), the
conversion was supposed to be considerable. However, the
same phenomenon as with the Au/SBA-15 catalyst could be
observed, although the conversion was 37.9% at a time of 5
min, and it increased only to 74.9% at 2 h. Furthermore, the
Figure 4. TGA curves of (a) fresh Au/CMK-3 catalyst; (b) the Au/
CMK-3 catalyst after reuse for four times; (c) sample b treated with
NaOH aqueous solution at 298 K; (d) sample b treated with NaOH
aqueous solution at 323 K; (e) sample b treated with NaOH aqueous
solution at 363 K; (f) sample b calcined at 773 K in argon; and (g)
sample b calcined at 773 K in nitrogen atmosphere.
−1
2
664
DOI: 10.1021/cs502093b
ACS Catal. 2015, 5, 2659−2670

ACS Catalysis
Research Article
quantity of disordered carbon doubled from 1.9% to 4.0%,
which was possibly caused by the physical damage of vigorous
stirring during the subsequent runs, and it was not affected by
either NaOH or calcination treatment. Concerning the removal
of carboxylic groups, the effects were all favorable. For the
NaOH treatment group, the residue of the carboxylic group was
almost the same in number for the catalysts treated at room
temperature and 323 K. However, the value for the 363 K
sample was lower, possibly because the carboxylic compounds
were strongly adsorbed, and a high temperature was necessary
to break the force acting between the compounds and catalyst
surface. Although a gap still existed between the recovered
catalyst and the fresh catalyst, the used catalyst showed
significant improvement. The hydroxyl group formed through
the reaction was almost not influenced by the NaOH treatment,
as expected. By contrast, the calcination group showed good
performance in removing both carboxylic and hydroxyl groups,
especially for the catalyst calcined under argon, which had
almost the same weight loss as the fresh catalyst.
From the TEM images and size distributions of Au NPs in
Figure S5, the sizes of Au NPs in the NaOH treatment group
remained almost constant as the used catalyst at ∼3.0 nm.
However, the calcination group showed larger sizes of over 3.5
nm. The value might not seem to increase very noticably, but
some extraordinary large particles that appear in these two
catalysts indicated that agglomeration truly happened. The 20
nm scale TEM images in Figure S5g,h clearly show that these
particles could achieve a size of over 15 nm, which could not be
found in other catalysts. Furthermore, the ICP-AES results
showed that the recovered catalysts had a much higher Au
content than the used one, thereby confirming our previous
assumption that the relatively low amount of Au for the used
catalyst was caused mainly by the occupation of the adsorbed
carboxylic compounds.
The conditions of the surface elements analyzed with XPS
are discussed below. The XPS spectra of Au 4f are shown in
Figure 5a and fitted into two peaks. A study has reported that
Figure 5. XPS spectra of (a) Au 4f, (b) O 1s, and (c) C 1s of the
catalysts of fresh, reused four times, and regenerated with NaOH
aqueous solution at 363 K.
37
the binding energies of Au 4f and Au 4f7/2 were located at
5
/2
0
8
7.7 and 83.9 eV for Au and at 89.6 and 86.3 eV for Au O ,
2 3
respectively. For the fresh catalyst, the peaks were located at
4.5 and 88.2 eV, revealing the valence of Au species between 0
area that corresponds to the carbonyl oxygen in quinones
dropped drastically from 35% to 20%.
8
39
and +3. After the experiments, the peaks shifted slightly to the
high binding energy right at 84.7 and 88.4 eV, which might
possibly be explained by the abundant oxygen atoms with
stronger electron negativity than the carbon existing in the
adsorbed carboxylic acids. However, the peaks returned to 84.5
and 88.2 eV after regeneration (the catalyst treated by NaOH at
According to Wan et al., different surface groups affect the
reaction to a great extent. For example, the carbonyl of
quinones and phenol groups could promote the adsorption of
reactants and intermediates, thereby becoming beneficial for
the reaction. However, the carboxyl groups had an opposite
effect and were unfavorable for the product formation. After the
NaOH treatment at 363 K, the peak area percentage of COOH
dropped to 14%, which means that about half of the adsorbed
carboxylic compounds were removed. Meanwhile, the peak of
the carbonyl oxygen of quinones returned to ∼27%.
363 K was chosen here).
The O 1s spectra in Figure 5b could be divided into four
3
8
peaks as reported. The carbonyl oxygen of quinones was
located at 531−531.9 eV; the carbonyl oxygen atoms in esters
and anhydrides and the oxygen atoms in phenolic groups at
In Figure 5c, the C 1s spectra could also be resolved into four
40
2
532.3−532.8 eV; the noncarbonyl oxygen atoms in esters and
peaks. The sp -hybridized graphitic carbon was located at
284.5 eV; the carbon atoms in the C−O single bond of phenols
and ethers, at 286.1 eV; the carbon atoms doubly bonded to the
oxygen in ketones and quinones, at 287.5 eV; and the carbon in
the CO of carboxyls, carboxylic anhydrides, and esters, at
288.7 eV. The graphitic carbon formed the majority of ∼65% in
the fresh catalyst, whereas the latter two peaks comprised 23%
of the area. As expected, the proportion of the carbonyl in
ketones, quinones, and carboxylic compounds increased in
quantity to ∼32% after the recycling experiments, and as a
anhydrides at 533.1−533.8 eV; and the COOH at a binding
energy of 534.3−535.4 eV. The peak area of the used catalyst
was relatively much bigger than that of the fresh one, indicating
that a large amount of oxygen atoms exists in the catalyst.
However, it was significantly reduced after the NaOH
treatment but was still slightly larger than that of the as-
prepared catalyst. The peak that corresponds to COOH at
5
34.1 eV was quite small and took only ∼5% in the fresh
catalyst. After the reactions, it rose to 21%, obviously because of
the strongly adsorbed carboxylic acids. Meanwhile, the peak
2
result, the area percentage of the sp -hybridized graphitic
2
665
DOI: 10.1021/cs502093b
ACS Catal. 2015, 5, 2659−2670

ACS Catalysis
Research Article
carbon dropped to 54%. As surmised, the NaOH treatment
would be quite useful to remove the adsorbed carboxylic acids,
as clearly seen in the spectra. The newly brought in CO
species through the reactions was clearly eliminated and took
the share back to 25%, which was close to that of the fresh
catalyst. At the same time, the area percentage of graphitic
carbon returned to 60%. The proportion of the carbon singly
bonded to oxygen was almost constant in the range of 13−15%
throughout the entire process.
concerning why fructose could be produced under acidic
conditions. To illuminate the reaction path to the byproduct
fructose, the amount of hydrogen peroxide produced in the
reaction system was quantitatively measured. Given that
hydrogen peroxide can react with titanyl sulfate and form a
yellow complex, which has a strong adsorption at approximately
410 nm in UV−vis (Figure 7), the peak intensity can be
All recovered catalysts were tested with the same weight and
reaction conditions as the fresh catalyst. Their regenerating
performances listed in Figure 6 show that the NaOH treatment
Figure 6. Conversion of glucose and selectivity of gluconic acid over
catalysts as-prepared, reused four times, and regenerated using
different methods. Reaction conditions: glucose = 0.36 g, H O = 20
2
mL, catalyst = 0.0396 g, glucose/Au (molar ratio) = 1000/1, T = 383
K, P(O ) = 0.3 MPa, reaction time = 2 h.
2
made significant improvements in the activity. The samples
treated at 363 K showed the best conversion of over 87%,
which was close to that of the fresh catalyst, and the result was
consistent with the above characterizations. Given that the
increase in activity might be affected by the promotion of
surface basicity, the fresh catalyst was treated in the same way
with NaOH at 363 K. After treatment, the catalyst showed a
slightly reduced conversion, which was within the margin of
error. Hence, the result revealed that the excellent activity of
the recovered catalyst was not caused by the added basic sites.
This conclusion was also confirmed by the CO -TPD
2
experiment (Figure S6). The as-prepared Au/CMK-3 catalyst
showed an almost negligible TCD signal at ∼713 K whenever a
prior CO purge was given. The weak peak at ∼713 K is
2
essentially due to the decomposition of oxygen-containing
species such as carboxyl and carbonyl groups rather than the
Figure 7. UV−vis spectra of the mixture of titanyl sulfate with (a)
standard solutions and (b) reaction liquid at different reaction times;
contribution of CO -adsorption species on the catalyst surfaces.
2
The spent catalyst after the NaOH treatment also represented a
similar peak with negligible TCD intensity, indicating that no
enhancement in the surface basicity occurred in this case. In
contrast to the superior effects of the NaOH treatment, the
calcination did not show any improvement on activity either
under argon or nitrogen atmosphere, which might be ascribed
to the agglomeration of Au NPs.
(c) the photograph of the mixture of titanyl sulfate with reaction liquid
at different reaction times.
converted into concentration using the standard curves and
fitted equation. According to the data in Table 5, as the reaction
time lapsed with the gradually increasing conversion, the
selectivity toward gluconic acid tapered off regularly. At the
same time, the concentration of hydrogen peroxide also
dropped drastically, in keeping with the variation tendency of
the gluconic acid selectivity. This phenomenon could also be
observed intuitively in Figure 7c, which presents the photo-
3
.4. Mechanism Aspect. The above research demon-
strated that the Au/CMK-3 catalyst had an advantage in the
base-free oxidation of glucose and was very convenient to
regenerate with NaOH treatment. Hence, the details of the
reaction mechanism were studied to address questions
2
666
DOI: 10.1021/cs502093b
ACS Catal. 2015, 5, 2659−2670

ACS Catalysis
Research Article
Table 5. Base-Free Oxidation of Glucose and Concentration
of Hydrogen Peroxide Found over Au/CMK-3 Catalyst As a
Function of Reaction Time
carried out at 383 K and 0.3 MPa. When no catalyst was added,
no gluconic acid was produced. Consequently, no hydrogen
peroxide was formed, and no fructose would be generated.
Indeed, fructose formation was not observed in the actual
situation. However, when extra alkaline was added into the
system in the absence of catalyst, fructose was generated as the
major product without any oxidizing products. This phenom-
enon indicated that fructose was produced with existing
hydroxyl ions, and if no sacrificial alkaline was added into the
system, fructose formation depended on the generated hydroxyl
ions from hydrogen peroxide. The proposed model for reaction
mechanism is presented in Figure 8.
a
selectivity, %
reaction
time, min
gluconic
acid
concn (H O )found,
2
2
−
1
conv, %
fructose others
mmol L
1
3
6
9
5
0
0
0
53.0
63.8
79.6
88.4
92.4
94.6
94.0
91.2
89.5
87.5
4.1
4.8
5.5
6.7
7.4
1.3
1.2
3.3
3.8
5.1
1.80
1.10
0.48
0.24
0.18
1
20
a
Reaction conditions: glucose = 0.36 g, H O = 20 mL, catalyst =
2
From the variation tendency of hydrogen peroxide
concentration and gluconic acid selectivity in Table 5, a
corollary that in the inception phase, the main reaction had a
comparatively fast rate and a great advantage over the
isomerization reaction might possibly be obtained. Thus, the
gluconic acid selectivity and the hydrogen peroxide concen-
tration were relatively high at the reaction time of 15 min.
However, as the reaction proceeded, given the decrease in
reactant concentration and the deactivation of the catalyst, the
main reaction was retarded, which could be observed from the
pH course in Figure S7, and the accumulated hydrogen
peroxide offered a favorable environment for glucose isomer-
ization. As a result, the disparity between oxidation and
isomerization became small. Accordingly, both the selectivity
toward gluconic acid and the amount of hydrogen peroxide
became less simultaneously. From a certain perspective,
isomerization relied on oxidation, and the amount of fructose
was positively correlated with the reaction conversion. This
conclusion could be set up only when the catalyst was fixed.
Because various catalysts would have different performances for
the glucose oxidation reaction, we could predict that for two
catalysts that achieved the same final conversion, the one that
consumed less time would have better selectivity. For example,
if the performances of the Au/CMK-3 and Au/CNTs catalysts
were compared, the conversions were clearly almost the same
0
.0396 g, glucose/Au (molar ratio) = 1000/1, T = 383 K, P(O ) = 0.3
2
MPa.
graph of the mixture of the reaction liquid and titanyl sulfate.
When the conversion reached ∼93% in 2 h, which meant the
completion of the reaction, the amount of hydrogen peroxide
left was 0.18 mmol L , and the mixture also turned almost
colorless.
−1
16
According to a previous report, when glucose molecules
react with oxygen and water, aside from gluconic acid,
hydrogen peroxide is also produced. Thus, the more rapidly
the main reaction occurs, the more hydrogen peroxide is
produced. The author proved that the generated hydrogen
peroxide would act as an oxidant to oxidize glucose and
decompose to oxygen. However, apart from these two
consumption reactions, hydrogen peroxide could also produce
hydroxyl ions on the surface of Au by a series of chain
reactions. The isomerization reaction between glucose and
fructose occurs only under the effect of hydroxyl ions. Given
that no alkaline was added into the system in advance, and the
reaction liquid became acidic because of the generated gluconic
acid, the needed hydroxyl ions for the production of fructose
could be afforded by hydrogen peroxide. This deduction was
also confirmed by the results of the comparison experiments
4
1
Figure 8. Proposed reaction mechanism for the base-free oxidation of glucose with molecular oxygen.
2
667
DOI: 10.1021/cs502093b
ACS Catal. 2015, 5, 2659−2670

ACS Catalysis
Research Article
for Au/CMK-3 at 30 min and for Au/CNTs at 2 h, whereas the
gluconic acid selectivities showed a disparity of over 10%.
To prove the earlier point of hydrogen peroxide generation
and consumption, a series of experiments were carried out.
Table 7. Base-Free Oxidation of Glucose over Au/CMK-3
Catalyst under Different Glucose/Au Molar Ratios
a
glucose/Au, molar
ratio
reaction
time
selectivity (gluconic
acid), %
conv, %
−1
−1
Solutions of 0.05 mol L H SO and 0.1 mol L NaOH were
2
4
1
1
000
000
1 h
2 h
79.6
92.4
91.2
87.5
88.1
91.9
used to adjust the initial pH of the reaction liquid, and the
connection of conversion, selectivity, and the amount of
hydrogen peroxide were studied. The data in Table 6 indicate a
5
1
00
00
1 h
99.0
15 min
>99.9
a
Reaction conditions: H O = 20 mL, catalyst = 0.0396 g, T = 383 K,
2
Table 6. Oxidation of Glucose over Au/CMK-3 Catalyst at
Different Initial pH Conditions
a
P(O ) = 0.3 MPa.
2
selectivity
(gluconic
acid), %
concn
+
−
H (or OH )/glucose,
molar ratio
(H O )
,
found
−1
2
2
condition
conv, %
mmol L
selectivity were obtained in only 15 min when the ratio was set
to 100.
H SO4
2
1
16.1
19.2
41.8
53.0
70.7
90.7
92.6
98.2
98.0
97.1
94.6
89.3
66.1
63.6
0.95
1.19
1.41
1.80
1.26
0.63
0.14
0
0
0
.5
.1
In addition, we have also tried hydrogen peroxide as oxidant
in this pressure bottle system, and the results are as expected.
When the reaction temperature was set as 333 K and
equimolecular hydrogen peroxide was given, the conversion
was ∼29.0%, with almost full selectivity. As was mentioned
previously, when the temperature was lower than 333 K, no
byproduct was produced, and high selectivity toward gluconic
acid could be achieved. However, when the temperature was
increased to 383 K, with an increasing conversion of 85.4%, the
selectivity toward gluconic acid decreased to 84.7%. This
change was in accordance with the situation when oxygen was
used as oxidant, indicating that isomerization happened as well
on this occasion. We also conducted an experiment at 333 K
with bubbling oxygen. No fructose was generated when no
NaOH was added. However, if equimolecular NaOH was
added before the reaction, fructose could be observed in the
reaction liquid, despite the low temperature and atmospheric
pressure, because there was a large amount of hydroxyl ions
existing in the solution.
It is speculated that the generation of fructose is controlled
by both temperature and hydrogen peroxide. Only in the case
when both the temperature and concentration of hydrogen
peroxide are high enough could the produced hydroxyl ions
reach the amount to make isomerization happen. In addition,
because of the acidic condition of the reaction system, the
produced hydroxyl ions may not exist for a long time.
Therefore, hydroxyl ions may disappear before isomerization
happens when its concentration is not high enough.
NaOH
0.1
0
1
.5
a
Reaction conditions: glucose = 0.36 g, H O = 20 mL, catalyst =
2
0
.0396 g, glucose/Au (molar ratio) = 1000/1, T = 383 K, P(O ) = 0.3
2
MPa, reaction time = 15 min.
volcano-type change in the amount of hydrogen peroxide with
the increase in initial pH. As the amount of H SO gradually
2
4
increased, as expected, glucose conversion dropped drastically
because the main reaction was intensively suppressed. As
mentioned previously, hydrogen peroxide was produced in the
oxidation reaction; consequently, the concentration of hydro-
gen peroxide was relatively low, and isomerization suffered, as
well, leading to the tremendous gluconic acid selectivity. When
alkaline was added, oxidation was promoted, as Rossi et al.
reported, and glucose conversion increased to a large extent.
Similarly, the selectivity decreased as a result. At an initial pH of
14
1
3.0, the reaction was fast, such that almost full conversion was
achieved at the reaction time of 15 min, with ∼60% selectivity;
however, despite the facilitated main reaction, the detected
hydrogen peroxide was extremely low because the generated
hydrogen peroxide was largely decomposed to oxygen under
16
the effect of added hydroxyl ions, as Rossi et al. proved.
With regard to the influence of temperature and pressure,
both of them could affect the selectivity to gluconic acid when
the main reaction was facilitated. A comparison of the data of
the reactions at 363 K and 0.1 MPa for 2 h and at 383 K and
0
.3 MPa for 30 min show that although the conversion of the
latter was higher, the selectivity was also better. Similar to the
law that catalysts with better activity also have better selectivity,
a reaction condition beneficial for conversion would also have
benefits for selectivity. From this aspect, obtaining both high
conversion and selectivity is desirable, and the feasible way is to
finish the reaction as soon as possible. This process can be
optimized by increasing the temperature or pressure, short-
ening the reaction time, and reducing the glucose/catalyst
molar ratio simultaneously. To verify this deduction, a series of
experiments with different glucose/Au molar ratios were
conducted, and the results are listed in Table 7, which shows
an obvious consistency with the above inference. When the
glucose/Au ratio was decreased to 500, the reaction consumed
4
2,43
∼
1 h to achieve 99% conversion with 88% selectivity. More
remarkably, almost full conversion and approximately 92%
2
668
DOI: 10.1021/cs502093b
ACS Catal. 2015, 5, 2659−2670

ACS Catalysis
. CONCLUSIONS
Research Article
(
3) Hustede, H.; Haberstroh, H.-J.; Schinzig, E. In Ullmann’s
4
Encyclopedia of Industrial Chemistry, 6th ed.; Wiley-VCH: Weinheim,
000; Vol. A 12, pp 449−456.
4) Thielecke, N.; Aytemir, M.; Pru
15−120.
5) Baatz, C.; Pru
This study showed that the Au/CMK-3 catalyst had an
advantage in the aerobic oxidation of glucose under base-free
conditions. Increasing the temperature and oxygen pressure
could promote the reaction rate and conversion, and the
improving effect was more obvious when the oxygen pressure
was increased at a relatively high temperature. At 383 K and 0.3
MPa, the conversion could exceed 92% in 2 h. Furthermore, the
amount of the byproduct fructose was affected by the
generation of hydrogen peroxide, which was produced
simultaneously in the main reaction. Thus, for a catalyst
under a fixed reaction condition, the selectivity reduced
gradually along with time, but for the different catalysts that
reached the same conversion, the one that cost less time would
have a better selectivity toward gluconic acid.
2
(
1
(
̈
sse, U. Catal. Today 2007, 121,
̈
ße, U. J. Catal. 2007, 249, 34−40.
(6) Lichtenthaler, F. W. In Biorefineries  Industrial Processes and
Products, 1st ed.; Kamm, B.; Gruber, P. R.; Kamm, M., Eds.; Wiley-
VCH: Weinheim, 2008; Vol. 2, pp 2−5.
(
(
(
7) Kobayashi, H.; Fukuoka, A. Green Chem. 2013, 15, 1740−1763.
8) Dirkx, J. M. H.; van der Baan, H. S. J. Catal. 1981, 67, 1−13.
9) Besson, M.; Lahmer, F.; Gallezot, P.; Fuertes, P.; Fleche, G. J.
Catal. 1995, 152, 116−121.
(
Chem. 2002, 180, 141−159.
(
(
10) Wenkin, M.; Ruiz, P.; Delmon, B.; Devillers, M. J. Mol. Catal. A,
11) Gallezot, P. Catal. Today 1997, 37, 405−418.
12) Pina, C. D.; Falletta, E.; Rossi, M. Chem. Soc. Rev. 2012, 41,
The stability of the Au/CMK-3 catalyst was studied. The
conversion dropped gradually from 92% to ∼70% through the
four subsequent runs. Two different kinds of regeneration
method, including NaOH treatment and calcination, were
compared and discussed. On the basis of the results, the used
catalyst treated with 50 mL of 0.1 mol/L NaOH for 30 min at
3
50−369.
(13) Kusema, B. T.; Murzin, D. Y. Catal. Sci. Technol. 2013, 3, 297−
307.
(14) Biella, S.; Prati, L.; Rossi, M. J. Catal. 2002, 206, 242−247.
(
15) Hermans, S.; Devillers, M. Appl. Catal., A 2002, 235, 253−264.
16) Comotti, M.; Pina, C. D.; Falletta, E.; Rossi, M. Adv. Synth.
(
Catal. 2006, 348, 313−316.
17) Zhang, H. J.; Okuni, J.; Toshima, N. J. Colloid Interface Sci. 2011,
54, 131−138.
18) Mirescu, A.; Berndt, H.; Martin, A.; Pru
007, 317, 204−209.
19) Thielecke, N.; Vorlop, K.-D.; Pru
266−269.
(20) Wang, Y. R.; Van de Vyver, S.; Sharma, K. K.; Roman
Y. Green Chem. 2014, 16, 719−726.
3
63 K showed the best performance with 87% conversion. The
(
3
(
2
characterizations indicated that after the recycling experiments,
abundant carboxylic acids were adsorbed on the surface of the
catalyst, which was the predominant factor that caused
deactivation. Both NaOH treatment and calcination could
effectively remove the adsorbates, and atmospheric calcination
was more useful in this aspect. However, given the appearance
of large particles during the calcination procedure, the
consequence was no match for the NaOH treatment method.
Therefore, one can conclude that the NaOH treatment is a very
simple but useful way to regenerate used catalysts that suffered
from adsorbate deactivation in glucose oxidation.
̈
ße, U. Appl. Catal., A
(
̈
ße, U. Catal. Today 2007, 122,
́
-Leshkov,
(21) Miedziak, P. J.; Alshammari, H.; Kondrat, S. A.; Clarke, T. J.;
Davies, T. E.; Morad, M.; Morgan, D. J.; Willock, D. J.; Knight, D. W.;
Taylor, S. H.; Hutchings, G. J. Green Chem. 2014, 16, 3132−3141.
(
22) Dimitratos, N.; Villa, A.; Wang, D.; Porta, F.; Su, D.; Prati, D. J.
Catal. 2006, 244, 113−121.
(
Kreutzer, M. T.; Kooyman, P. J.; Moulijn, J. A.; Kapteijn, F. Catalysts.
2014, 4, 89−115.
(24) Wang, T.; Shou, H.; Kou, Y.; Liu, H. Green Chem. 2009, 11,
23) Skupien, E.; Berger, R. J.; Santos, V. P.; Gascon, J.; Makke, M.;
ASSOCIATED CONTENT
Supporting Information
The following file is available free of charge on the ACS
■
*
S
5
(
2
62−568.
25) Ding, Y.; Li, X. H.; Pan, H. Y.; Wu, P. Catal. Lett. 2014, 144,
68−277.
(26) Sun, F.; Liu, J.; Chen, H. C.; Zhang, Z. X.; Qiao, W. M.; Long,
Publications website at DOI: 10.1021/cs502093b.
N2 adsorption−desorption isotherms, XRD patterns,
TEM images and CO -TPD profiles of some catalysts
2
as well as the recycling experiment data (PDF)
D. H.; Ling, L. C. ACS Catal. 2013, 3, 862−870.
(
(
H.; Chmelka, B. F.; Stucky, G. D. Science 1998, 279, 548−552.
(
27) Saha, D.; Deng, S. G. Langmuir 2009, 25, 12550−12560.
28) Zhao, D. Y.; Feng, J. L.; Huo, Q. S.; Melosh, N.; Fredrickson, G.
AUTHOR INFORMATION
Corresponding Author
Phone: +86-592-2181659. Fax: +86 592 2183047. E-mail:
■
*
29) Jun, S.; Joo, S. H.; Ryoo, R.; Kruk, M.; Jaroniec, M.; Liu, Z.;
Ohsuna, T.; Terasaki, O. J. Am. Chem. Soc. 2000, 122, 10712−10712.
(30) Bi, Q.-Y.; Du, X.-L.; Liu, Y.-M.; Cao, Y.; He, H.-Y.; Fan, K.-N. J.
Am. Chem. Soc. 2012, 134, 8926−8933.
yzyuan@xmu.edu.cn.
Notes
(
31) Cui, H.-F.; Ye, J.-S.; Liu, X.; Zhang, W.-D.; Sheu, F.-S.
The authors declare no competing financial interest.
Nanotechnology 2006, 17, 2334−2339.
(
331−335.
32) Zhe, Y. Q.; Ma, C. A.; Zhu, Y. H. Acta Phys. Chim. Sin. 2004, 20,
ACKNOWLEDGMENTS
We acknowledge the financial supports from the MOST of
China (2011CBA00508), the NSFC (21403178 and
■
(33) Zhang, X.; Cui, Y.; Lv, Z.; Li, M.; Ma, S.; Cui, Z.; Kong, Q. Int. J.
Electrochem. Sci. 2011, 6, 6063−6073.
̈
(
(
34) Onal, Y.; Schimpf, S.; Claus, P. J. Catal. 2004, 223, 122−133.
2
1473145), the Research Fund for the Doctoral Program of
35) Datsyuk, V.; Kalyva, M.; Papagelis, K.; Parthenios, J.; Tasis, D.;
Higher Education (20110121130002), and the Program for
Changjiang Scholars and Innovative Research Team in
University (IRT_14R31).
Siokou, A.; Kallitsis, I.; Galiotis, C. Carbon 2008, 46, 833−840.
36) Aviles, F.; Cauich-Rodríguez, J. V.; Moo-Tah, L.; May-Pat, A.;
Vargas-Coronado, R. Carbon 2009, 47, 2970−2975.
37) Zhang, J.; Jin, Y.; Li, C. Y.; Shen, Y. N.; Han, L.; Hu, Z. X.; Di,
(
́
(
REFERENCES
■
X. W.; Liu, Z. L. Appl. Catal., B 2009, 91, 11−20.
(38) Kvande, I.; Zhu, J.; Zhao, T.-J.; Hammer, N.; Rønning, M.;
Raaen, S.; Walmsley, J. C.; Chen, D. J. Phys. Chem. C 2010, 114,
1752−1762.
(
(
1) Gallezot, P. Catal. Today 2007, 121, 76−91.
2) Ma, C. Y.; Xue, W. J.; Li, J. J.; Xing, W.; Hao, Z. P. Green Chem.
2
013, 15, 1035−1041.
2
669
DOI: 10.1021/cs502093b
ACS Catal. 2015, 5, 2659−2670

ACS Catalysis
Research Article
(
39) Wan, X. Y.; Zhou, C. M.; Chen, J. S.; Deng, W. P.; Zhang, Q.
H.; Yang, Y. H.; Wang, Y. ACS Catal. 2014, 4, 2175−2185.
40) Kundu, S.; Wang, Y. M.; Xia, W.; Muhler, M. J. Phys. Chem. C
008, 112, 16869−16878.
41) Weiss, J. Trans. Faraday Soc. 1935, 31, 1547−1557.
42) Saliger, R.; Decker, N.; Pruße, U. Appl. Catal., B 2011, 102,
84−589.
43) Pru
53−459.
44) Ishida, T.; Kuroda, K.; Kinoshita, N.; Minagawa, W.; Haruta, M.
J. Colloid Interface Sci. 2008, 323, 105−111.
(
2
(
(
̈
5
(
̈
ße, U.; Jarzombek, P.; Vorlop, K.-D. Top Catal. 2012, 55,
4
(
2
670
DOI: 10.1021/cs502093b
ACS Catal. 2015, 5, 2659−2670