Selective oxidation of D-glucose on gold catalyst

Article abstract of DOI:10.1006/jcat.2001.3497

The selective oxidation of D-glucose to D-gluconic acid was performed at both controlled (7-9.5) and free pH values in an aqueous solution in the presence of a gold on carbon catalyst using dioxygen as the oxidant under mild conditions (323-373 K, pO₂ = 100-300 kPa). No isomerization of glucose to fructose was observed during the reaction, and total selectivity to D-gluconate was obtained. A comparative study of gold and commercial palladium and platinum-derived catalysts highlighted the unique property of supported gold, i.e., it is active at low pH (2.5), whereas at a buffered higher pH (9.5), gold and bismuth-doped platinum-palladium on carbon catalysts show comparable selectivity, though gold has a higher activity. Moreover, upon recycling, gold is more stable toward deactivation, although this also depends on the operative pH.

Full text of DOI:10.1006/jcat.2001.3497

Journal of Catalysis 206, 242–247 (2002)
doi:10.1006/jcat.2001.3497, available online at http://www.idealibrary.com on
Selective Oxidation of D-Glucose on Gold Catalyst
Serena Biella, Laura Prati, and Michele Rossi
Dipartimento di Chimica Inorganica Metallorganica e Analitica e Centro C.N.R., via Venezian 21, 20133 Milano, Italy
Received July 17, 2001; revised November 20, 2001; accepted December 8, 2001
(
15). Slightly alkaline conditions are apparently necessary
The selective oxidation of D-glucose to D-gluconic acid was to increase the reaction rate and to avoid drastic deacti-
performed at both controlled (7–9.5) and free pH values in an vation of the catalyst; conversely, such conditions are also
aqueous solution in the presence of a gold on carbon catalyst us-
ing dioxygen as the oxidant under mild conditions (323–373 K,
responsible for side reactions that reduce gluconate pro-
ductivity. Another important economic drawback of using
pO
2
= 100–300 kPa). No isomerization of glucose to fructose was
bases is the production of gluconate instead of the free acid.
Thus, from an industrial point of view, the fermentation pro-
cess is still preferred to the catalytic process, despite prob-
lems related to microbe separation, control of by-products
and the disposal of waste water.
observed during the reaction, and total selectivity to D-gluconate
was obtained. A comparative study of gold and commercial pal-
ladium and platinum-derived catalysts highlighted the unique
property of supported gold, i.e., it is active at low pH (2.5), whereas
at a buffered higher pH (9.5), gold and bismuth-doped platinum-
palladium on carbon catalysts show comparable selectivity, though
Therefore, only the development of catalysts with ex-
gold has a higher activity. Moreover, upon recycling, gold is more cellent activity/selectivity (i.e., productivity) and durability
stable toward deactivation, although this also depends on the will facilitate the oxidation of glucose with a catalyst, using
operative pH.
ꢀc 2002 Elsevier Science (USA)
dioxygen as the oxidant. In our previous studies on the ox-
Key Words: selective oxidation; gold catalyst; liquid-phase oxi- idation of 1,2-diols, we emphasized the unusual selectivity
dation; gluconic acid.
of gold catalysts in oxidizing the primary alcoholic func-
tion to the corresponding carboxylate, in contrast to the
lower selectivity shown by palladium or platinum on car-
bon catalysts (16–18). Moreover, we concluded that gold is
less sensitive to overoxidation and/or self-poisoning than
platinum or palladium.
As in the oxidation of alcohols the formation of an inter-
mediate aldehyde was postulated (17). This prompted us to
investigate gold as the catalyst for the selective oxidation
of the aldehydic group. Therefore, given the reported valu-
able application of catalytic oxidation to the transforma-
tion of carbohydrates, we decided to investigate D-glucose
as first substrate. The presence of both alcoholic and alde-
hydic (hemiacetal) groups in such a substrate also enabled
us to study the competitive oxidation of these two different
functional groups in the presence of the gold catalyst and,
in comparison, in the presence of platinum- and palladium-
based catalysts. Moreover, the oxidation of a free aldehy-
dic group in competition with a primary alcoholic group
1
. INTRODUCTION
Gluconic acid and its salts are important industrial prod-
ucts, they are used as water-soluble cleansing agents or as
additives in food and beverages (1). For commercial pur-
poses these products are exclusively prepared by the oxi-
dation of glucose or glucose-containing raw materials.
Although the currently used oxidation method is based
on biochemical transformation, recent developments have
indicated that the catalytic route may be a valid alternative
for producing gluconate on an industrial scale. In fact, early
studies at the beginning of the 1970s (2–5) led to many pa-
pers and patents, the most important of which reported on
the beneficial effect of doping platinum metal-group cata-
lysts with heavy metals such as bismuth or lead (6–9). It
has been reported that by using bismuth on palladium or
platinum/palladium on carbon catalysts, some of the factors
that limit the use of the catalytic route to produce gluconate
appear to have been overcome; high selectivity and excel-
lent conversion (>99%) were obtained. However, aspects
such as the nature of the catalyst and the role of bismuth
are still under discussion despite the many papers on this
subject (10–14). Moreover, catalyst durability is also related
to bismuth leaching, as was observed under the commonly
used oxidative conditions (temperature range 303–333 K,
(
n-propanol and n-butanal) was also investigated to gain
better insight into the reaction mechanism.
2. EXPERIMENTAL
2.1. Reagents and Apparatus
D-Glucose, n-propanol, and n-butanal (Fluka) were used
pH 8–10, glucose 20–30 wt%, catalyst/glucose ratio 0.5–2%) without further purification. NaOH (Merck) was 99.9%
242
0
ꢀ
021-9517/02 $35.00
c 2002 Elsevier Science (USA)
All rights reserved.

SELECTIVE OXIDATION OF GLUCOSE
243
pure and stored under nitrogen. Gaseous oxygen (SIAD) Glucose and the catalyst (S/M = 1000) were typically mixed
was 99.99% pure. Gold powder (Fluka) was of the in distilled water (4 wt% solution). The temperature of the
highest purity grade. Commercial 5% Pd/C (MPB5) was stirredmixturewaskeptat323Kand, assoonasthereaction
supplied by S u¨ d Chemie-M.T., 5% Pt/C (ESCAT10) by started, a solution of NaOH 0.3M was added automatically
Engelhard and 1%Pt–4%Pd–5%Bi/C (CEF 196XRA/W) to maintain the pH of the solution at a fixed value (7, 8, or
by Degussa. BiONO3 was used in a hydrochloric acid solu- 9.5 ± 0.1). The samples were analyzed at various times by
−1
13
tion (16.2 g liter of bismuth). Polyvinyl alcohol (M 10,000)
and NaBH4 were purchased from Merck. The possible re-
C-NMR and/or HPLC.
When the pH was not controlled, the reactions were car-
action products (gluconic acid, δ-glucono-lactone, 2-keto ried out in a temperature-controlled glass reactor (30 ml)
gluconic acid, fructose, glucose, saccharic acid, oxalic acid, with an electronic magnetic stirrer connected to a large
gulonic acid, glucuronic acid, tartaric acid, tartronic acid, reservoir (5,000 ml) containing oxygen at 300 kPa. A flow-
propanal, propionic and butyric acids), all from Fluka, were time diagram of the oxygen uptake was plotted by a mass
used as reference compounds. Activated carbon (X40S flow controller connected to a PC through an A/D board.
2
from CAMEL) (5–100 µ) with a specific area of 1100 m /g
Glucose and the catalyst (reactant/metal = 1000) were
was used after washing with distilled water until a neutral mixed in distilled water (4 wt% solution). The reactor was
pH was reached.
pressurized at 300 kPa of O2 and the temperature was kept
at 363 K. After a period of equilibration (15 min), the mix-
ture was stirred and the products were analyzed by time
2
.2. Preparation of the Catalysts
(
every 10 min). In recycling tests, following the first run,
2
.2.1. Gold on carbon. Gold on carbon was prepared the catalyst was filtered off and, after washing with distilled
byimmobilizing goldsolasreported previously (19, 20). An water or another specified treatment, was reused in the next
aqueous HAuCl4 solution of the desired concentration was run with a freshly prepared solution of glucose.
prepared (100 µg/ml). To this was added the 2 wt% solu-
tion of the protective agent polyvinylalcohol (PVA), while 2.4. Analysis of Products
stirringing vigorously. Then, dropwise, a freshly prepared
The products were identified and quantified by compar-
solution of NaBH4 (0.1M) was added to reduce the gold.
ing them with actual samples. Both the quantitative and
The sol was immobilized on carbon by adding the support
1
3
qualitative analyses were performed by HPLC; C-NMR
was also used for qualitative analysis. The quantitative anal-
ysis of the acidic products was deduced from the amount of
NaOH that was added for neutralization and performed by
titration (inverse).
to the metal dispersion while stirringing vigorously until the
solution was clear (about 1 h). The amount of support was
calculated as having a final gold loading of 1 wt%. After
1
h, the slurry was filtered and the total absorption of gold
was checked by ICP analysis of the filtrate.
The catalyst was filtered and then washed with hot
2.4.1. HPLC analysis. Analyses were performed on a
water until the filtrate was chloride free (AgCl test); it was Varian 9010 instrument equipped with a Varian 9050 UV
used wet. The metal contents of the fresh and used cata- (210 nm) and a Waters 2410 R.I. detectors. An Alltech OA-
lysts were determined by ICP analysis (Jobin Yvon JY24). 1000 column (300×6.5 mm) was used with aqueous H2SO4
XRPD (Rigaku D III-MAX horizontal-scan powder diff- 0.01M (pH 2.1) (0.4 ml/min) as the eluent. Glucose and
ractometer with CuKα radiation) and XPS (M-Probe In- gluconic acid peaks overlap but glucose does not adsorb at
strument [SSI] equipped with a monochromatic AlKα the used λ for UV detection (210 nm); thus, the amount
source [1486.6 eV]) analyses were performed to determine of glucose was determined by subtracting the gluconic acid
the size of the metal crystallites and the atomic percentage contribution quantified at 210 nm. Alternatively, the glu-
of gold on the carbon surface (2–5 nm).
cose quantitative analysis was carried out using a column
(
Supelcosil LC-NH2 [250 × 4.6 nm]) specially designed for
2
.2.2. Bimetallic catalysts. The 5%Pd–5%Bi on carbon
sucrose with CH3CN/H2O 45/55 as the eluent. This latter
analysis also confirmed the absence of isomerization of glu-
cose to fructose.
catalyst was prepared according to the impregnation
method, as previously reported (21) on commercial 5% Pd
on carbon.
2
.4.2. NMR analysis. 13C-NMR spectra were recorded
in water on a Bruker 300 MHz without adjust the pH. The
assignment of peaks in the oxidation of glucose was made
by comparing with standard spectra.
2
.3. Oxidation Procedure
All the experiments at a controlled pH were carried out
at atmospheric pressure by bubbling dioxygen through the
2.4.3. Titration procedure. If necessary, a sample of the
3
−1
slurry (20 Ncm min ) and controlling the pH (7, 8, and reaction mixture was neutralized (until pH 10–11) with
.5). The experiments were performed by using a 718 STAT NaOH 1M and then back-titrated with HCl 0.1N. The
Titrino (Metrohom) equipped with a 0.3M NaOH reserve. titration curve (pH versus ml HCl added) highlighted two
9

2
44
BIELLA, PRATI, AND ROSSI
equivalent points, one corresponding to the excess of
NaOH, the second to an organic acid at pH about 3.
a
100
8
6
4
2
0
0
0
0
0
1%Pt-4%Pd-5%Bi/C
5%Pd-5%Bi/C
3
. RESULTS AND DISCUSSION
5
%Pt/C
%Au/C
1
3
.1. Oxidation at Controlled pH
We recently reported on the use of a gold on carbon
catalyst in the selective oxidation of 1,2-diols to α-hydroxy-
acids (16–18). Although we proved a difference in the ac-
tivity of the gold catalysts, depending on the preparation
procedure (20), an intrinsic higher selectivity of gold to-
ward primary alcoholic function with respect to palladium
or platinum catalysts was noted. Strong alkaline conditions
were required for gold to speed up the reaction, whereas in
the other cases this also enhanced selectivity. However, we
also noted that unlike Pd and Pt, the presence of an alkali
to oxidize the alcoholic group is necessary when using gold-
supported catalysts. Aldehyde is the supposed intermediate
of the reaction; thus, we tested gold on carbon as the cata-
lyst in the oxidation of glucose, keeping in mind that, when
glucose is used as the reagent, a strong alkaline solution
cannot be used, because the reagent starts to decompose
at a pH above 11. Therefore, our experiments were con-
ducted at a pH of 7–9.5 in order to avoid alkali-promoted
reactions such as retro-Claisen, Cannizzaro, condensation,
or isomerization.
0
50
100
150
200
250
300
time (min)
b 100
8
0
1
%Pt-4%Pd-
5%Bi/C
60
5%Pd-5%Bi/C
5%Pt/C
4
2
0
0
0
1%Au/C
0
50
100
150
200
250
time(min)
c
100
8
0
1
%Pt-4%Pd-5%Bi/C
%Pd-5%Bi/C
60
5
5%Pt/C
%Au/C
4
2
0
0
0
1
In addition, most patents and papers dealing with this
subject report experiments conducted in a weakly alkaline
mediumatacontrolledpHof9to10, becauseitwasrecently
shown that the catalytic activity of Pd or Pt on carbon, with
or without Bi-doping strongly depends on the pH (22–24).
In fact, carrying out the oxidation at 323 K, without control-
ling the pH, led to the rapid poisoning of the catalysts, which
was ascribed to strong adsorption of the reaction product.
At a pH of 9.5 and by bubbling oxygen through the slurry
0
50
100
150
200
250
300
350
400
time(min)
FIG. 1. Glucose oxidation under pH-controlled conditions. Reaction
conditions: glucose 4 wt%; glucose/M = 1000; O2 flow = 20 ml/min; T =
23 K: (a) the pH was maintained at 9.5 by adding dropwise NaOH 0.3M;
b) the pH was maintained at 8 by adding dropwise NaOH 0.3M; (c) the
pH was maintained at 7 by adding dropwise NaOH 0.3M.
3
(
3
−1
at 20 Ncm min and a temperature of 323 K, we confirmed
the beneficial effect of adding bismuth to monometallic Pd the reaction rate decreases to around zero after about 2 h
or Pt catalysts (6–9) (Fig. 1a). Surprisingly we also obtained (corresponding to 55–70% glucose conversion). Thus, the
a reversed activity in the series Au/C, Pd/C, and Pt/C with overall TOF value is considerably lower at pH 8 than at
respect to that observed in alcoholic group oxidation in pH 9.5 and, most importantly, 100% conversion of glucose
ethane-1,2-diol (16–18), gold being superior to Pt. Pd was cannot be reached.
found to be inactive. Moreover, gold was much more ac-
tive than commercial Pt–Pd–Bi or Pd–Bi catalysts that were reaction was complete in little more than an hour, with al-
specifically designed for glucose oxidation. most the same initial rate observed at pH 9.5 or 8. However,
In contrast, in the case of gold on carbon at pH 8, the
The difference in the various catalysts increases when the in this case too, a slight decrease in the reaction rate was
pH of the reaction decreases. At pH 8, a drastic decrease observed above 95% glucose conversion (Fig. 1b), resulting
in the TOF of the Pt–Pd–Bi and Pd–Bi on carbon catalysts in a lower TOF value at pH 8 than that obtained at pH 9.5.
occurred; this decrease corresponded to the catalyst activity
When the pH was decreased to 7, we found that deactiva-
as dependent on time or conversion: at about 50 to 70% tion of Pt–Pd–Bi and Pd–Bi on the carbon catalyst proceded
conversion of glucose, the activity of these catalysts was at considerably lower conversion than at pH 8 (Fig. 1c); at
near zero (Fig. 1b). This shows that the catalyst was clearly pH 7 this deactivation appeared to be more rapid for Pt–
poisoned, probably due to the reaction products. TOFs do Pd–Bi (<20% conversion) than for Pd–Bi on carbon (40%
not differ very much from the value obtained at pH 9.5 at conversion), both being less resistant than monometallic
the beginning of the reaction, but, as the reaction proceeds, Pt/C. Previous studies correlated the behavior of Pd- and

SELECTIVE OXIDATION OF GLUCOSE
245
lyst revealed a leaching of metal which was important at
pH 7 (18%), whereas no leaching of metal was observed at
pH 9.5. The diameter of the metal particle also increased
(
Table 1): when metal leaching increases, the particle size
decreases. This behavior was explained by the dissolution
and reprecipitation of the gold species. Since XPS analyses
revealed only the presence of metallic gold, we assume that
Au(I, III) species, which may be present, reduced, probably
as a result of carbon or other reducing species.
a 100
8
6
4
2
0
0
0
0
0
FIG. 2. TOF (h 1) as a function of pH.
−
1°run
2
°run
°run
3
Pt-based catalysts with the poisoning effect of the formed
gluconate, which becomes more evident as the pH de-
creases (13). Thus as the amount of free acid increases
with decreasing pH, the poisoning of the Bi-doped cata-
lysts was considered to be a temporary inhibition, related
to the concentration of gluconic acid on the catalyst surface
and, therefore, to the concentration of the acid in solution.
This was in part confirmed by removing the solution from a
deactivated catalyst and adding a fresh, acid-free solution of
glucose: the activity was partially restored (80%), probably
owing to the dilution of the adsorbed gluconic acid.
Gold on carbon, unlike the other tested catalysts, enabled
a quantitative transformation of glucose even at pH 7. How-
ever, the TOF value was considerably lower than that at
pH 9.5, partly because of the decrease in the reaction rate
above 85 to 90% conversion.
0
50
100
150
200
250
time(min)
1
00
b
80
60
40
20
0
1
°run
2°run
3°run
4°run
Figure 2 shows the trend of the TOF values at different
pH by using the three different catalysts (Au, Pt–Pd–Bi,
Pd–Bi). Given that we report TOF values calculated at 30%
conversion, but knowing that gold is unique in achieving to-
tal conversion at any pH, it can be seen that gold on carbon
shows the lowest decrease in activity with decreasing pH.
Our study on glycol oxidation using Au/C had revealed
the resistance of this catalyst to the poisoning by hydroxy
carboxylate (16–18); thus, we first ascribed the decrease in
the reaction rate above 85% conversion to a kinetic effect
of the lower concentration of the reactant. Adding a new
sample of glucose at the end of the reaction resulted in
a kinetic profile identical to the first run. However, more
detailed experiments of recycling (Figs. 3a–3c), carried out
by removing the reacted solution and washing the cata-
lyst with water before reusing it in the following run with
a freshly prepared solution of glucose, showed a different
behavior. In four runs, we observed a decrease in activity,
the degree of which depended on the pH of the reaction
0
50
100
150
200
250
time(min)
c 100
80
60
40
20
0
1°run
2
°run
°run
3
4°run
0
50
100
150
200
250
time(min)
FIG. 3. Recycling tests under pH-controlled conditions using
%Au/C. Reaction conditions as in Fig. 1. The catalyst was washed
with distilled water every run and then used in a fresh glucose solution:
a) the pH was maintained at pH 7 by adding dropwise NaOH 0.3M;
b) the pH was maintained at 8 by adding dropwise NaOH 0.3M; (c) the
1
(
(
(
pH 7 > pH 8 > pH 9.5). Analyses on fresh and used cata- pH was maintained at 9.5 by adding dropwise NaOH 0.3M.

2
46
BIELLA, PRATI, AND ROSSI
TABLE 1
1
00
Data of Fresh and Used Au/C Catalyst
9
0
catalyst washed with
water
Used at
Used at
pH 8a
Used at
pH 7a
pH 9.5a
80
Fresh
catalyst washed with
H2SO4 2N
7
6
5
0
0
0
catalyst washed with
NaBH4
%
Au [ICP]
0.89
4.42
0.89
5.85
0.81
5.71
0.73
5.26
d(nm) [XRPD]
a
Used in four runs (see Figs. 3a–3c) and washed with water every run.
0
1
2
3
4
5
6
7
3
.2. Oxidation at Uncontrolled pH
run
As previously reported, platinum-based catalysts rapidly
FIG. 5. Recycling tests at uncontrolled pH using 1%Au/C with dif-
deactivate in the liquid-phase oxidation of glucose under ferent treatment. Reaction conditions: glucose 4 wt%; glucose/M = 1000;
uncontrolled pH (13, 23). Based on the observation that pO2 = 300 kPa; T = 373 K; treact = 6 h.
gold enables 100% conversion, even at neutral pH, we also
tested the behavior of gold on carbon under uncontrolled
pH.
Based on the assumption that gold on carbon does not
3.3. Selectivity of the Reaction
Using fresh catalysts based on gold or even on platinum/
bismuth-modified metals, we obtained very high (>99%)
selectivity to gluconate. Differences between gold and
platinum-based catalysts were detected only by recycling,
with gold always showing total selectivity.
suffer from oxygen poisoning, we used a closed vessel at
an oxygen pressure of 300 kPa to increase oxygen dissolu-
tion, thus speeding up the reaction. As shown in Fig. 4,
gold on carbon shows a lower reaction rate than under
pH-controlled conditions, but it also enabled a quantita-
tive transformation without controlling the pH. This led us
to study the recycling of the catalyst under the same condi-
tions (Fig. 5). Washing the catalyst with water in each cycle,
after the third run we observed a decrease in activity which
increased markedly upon recycling. ICP analysis failed to
detect leaching of gold in a single reactant solution; leaching
of gold was obvious, however, after considering the metal
content following a few runs: 10% of the metal is lost in
only two runs and about the 70% in six runs.
In order to reduce metal leaching we tentatively changed
the treatment of the catalyst in the recycling step (Fig. 5).
Nevertheless by operating both acidic or reductive washing
we still obtained a strong deactivation of the catalyst, faster
than when water was employed; thus we concluded that the
loss of metal does not involve the treatment of the catalyst.
We separated two contributions leading high selectivity:
first, avoiding isomerization of glucose to fructose (obtain-
able by operating at a pH as low as possible) and, second,
enhancing the oxidation rate of the aldehydic group with
respect to the alcoholic group. In the case of gold, as stated
above, the presence of a base is compulsory in alcohol oxi-
dation. Thustheabsenceofoxidationinthealcoholicgroup,
observed under acidic (uncontrolled pH) conditions, was
not surprising. On the contrary, we expected that at con-
trolled pH, especially when pH increases (9.5), the selectiv-
ity decreases since the oxidation of the alcoholic group also
begins to become possible. It is difficult to detect real selec-
tivity in glucose oxidation due to analytical limits. Thus,
in a parallel experiment, we tested the competitive oxi-
dation of an aldehyde (n-butanal) and a primary alcohol
(
n-propanol). We observed that aldehyde oxidation pro-
ceeds quicker than alcohol oxidation (Table 2) and, in the
case of the Au/C catalyst, we confirmed the total inactiv-
ity of the Au/C catalyst toward alcohol oxidation. More-
over, the oxidation of the aldehyde is reported to proceed
throughanintermediateacetalform(10). Thus, weassumed
that starting from the hemiacetal form, as in glucose, the
aldehyde oxidation rate would be promoted, while that of
the alcohol would remain the same. Considering that (i)
at 100% conversion, i.e., in the absence of glucose, HPLC
analyses excluded the presence of byproduct derived from
alcohol oxidation (<1%), and the analysis by titration fit-
ted with the presence of 100% acidic product, (ii) in the
test on competitive oxidation of aldehydic and alcoholic
groups only the aldehydic is oxidized, we can propose for
glucose oxidation in the presence of Au/C a total selectivity
to gluconic acid.
1
00
8
6
4
2
0
0
0
0
0
0
1
2
3
4
5
6
7
time (h)
FIG. 4. Glucose oxidation at uncontrolled pH conditions using
%Au/C catalyst. Reaction conditions: glucose 4 wt%; glucose/M = 1000;
1
pO2 = 300 kPa; T = 373 K.

SELECTIVE OXIDATION OF GLUCOSE
247
TABLE 2
Oxidation of n-Butanal/n-Propanol Mixture
Butyrate
or palladium-based catalysts, which, however, remain in-
ferior to the results obtained with the monometallic gold
catalyst.
Propanoate
(mol%)
Run
Catalyst
pH
(mol%)
ACKNOWLEDGMENT
a
1
1%Au/C
9.5
9.5
9.5
9.5
8
8
8
8
74
76
69
54
63
70
57
36
63
70
70
58
0
0
0
5
0
0
0
2
0
a
We gratefully acknowledge the financial support provided by LONZA
S.A.
2
1%Pt–4%Pd–5%Bi/C
5%Pd–5%Bi/C
5%Pt/C
a
3
a
4
a
5
1%Au/C
REFERENCES
a
6
1%Pt–4%Pd–5%Bi/C
5%Pd–5%Bi/C
5%Pt/C
a
7
1. Kirk-Othmer, “Encyclopedia of Chemical Technology,” 4th ed. Wiley,
New York, 1993; Datta, R., in “Encyclopedia of Chemical Technol-
ogy,” Vol. 13, 4th ed. Wiley, New York, 1995.
a
8
b
9
1%Au/C
No control
No control
No control
No control
b
1
1
1
0
1
2
1%Pt–4%Pd–5%Bi/C
5%Pd–5%Bi/C
5%Pt/C
0
13
37
2. Johnson, Matthey and Co., Ltd., GB 1208101 (1970).
3. Kawaken Fine Chemicals Co. Ltd., JP 8007230 (1980).
4. Mitsui Toatsu Chemicals Inc., JP 7652121 (1976).
b
b
5
6
7
. Kawaken Fine Chemicals Co. Ltd./Kao Corp., JP 5872538 (1983).
. Asahi Chemical Industry Co. Ltd. DE-OS2936652 (1980).
. Kawaken Fine Chemicals Co. Ltd./Kao Corp., U.S. Patent 4,843,173
a
Reaction conditions: n-butanal/n-propanol (mol/mol) = 1; O2 flow =
2
0 ml/min; T = 323 K; treact = 1 h. The pH was maintained at the indicated
constant value by dropping NaOH 0.3M.
(
1989).
b
Reaction conditions: n-butanal/n-propanol (mol/mol) = 1; pO2 =
8
9
. Degussa Aktiengesellschaft, U.S. Patent 5,132,452 (1992).
. For a review see also: Besson, M., and Gallezot, P., Catal. Today 57,
3
00 kPa; T = 363 K; treact = 1 h.
127 (2000).
1
1
0. Mallat, T., and Baiker, A., Catal. Today 19, 247 (1994).
1. Wenkin, M., Renard, C., Ruiz, P., Delmon, B., and Devillers, M., Stud.
Surf. Sci. Catal. 110, 517 (1997).
4
. CONCLUSION
The catalytic oxidation of D-glucose to D-gluconic acid
remains an attractive target in industrial chemistry.
1
1
1
1
2. Besson, M., Lahmer, F., Gallezot, P., Fuertes, P., and Fleche, G.,
J. Catal. 152, 116 (1995).
3. Abbadi, A., and van Bekkum, H., Appl. Catal. A: General 124, 409
The present study showed that gold on carbon catalyst is
a valid alternative to most of the investigated multimetallic
catalysts based on palladium and/or platinum metals. More-
over, it was shown that gold has a unique property, i.e., it
operates without the external control of pH, thus ensuring
total conversion at all pH values. We already reported on
the resistance of gold to oxygen poisoning (16–18) during
alcohol oxidation, but gold is probably also less sensitive to
chemical poisoning than was proven for platinum catalysts
(
1995).
4. Bronnimann, C., Bodnar, Z., Hug, P., Mallat, T., and Baiker, A.,
J. Catal. 150, 199 (1994).
5. Wenkin, M., Touillaux, R., Ruiz, P., Belmon, B., and Devillers, M.,
Appl. Catal. A: General 148, 181 (1996) and references cited therein.
6. Prati, L., and Rossi, M., Stud. Surf. Sci. Catal. 110, 509 (1997).
7. Prati, L., and Rossi, M., J. Catal. 176, 552 (1998).
1
1
1
8. Bianchi, C., Porta, F., Prati, L., and Rossi, M., Topics in Catal. 13, 231
(
2000).
19. Prati, L., and Martra, G., Gold Bull. 32, 96 (1999).
(
13). Nevertheless, metal leaching was proved to be more 20. Coluccia, S., Martra, G., Porta, F., Prati, L., and Rossi, M., Catal. Today
6
1, 165 (2000).
1. Tsubota, S., Cunningham, D. A. H., Bando, Y., and Haruta, M., in
Preparation of Catalysts VI” (Studies on Surface Science and Catal-
consistent as pH decreased and corresponded to a gradual
decrease in activity.
The high selectivity of the gold catalyst to D-gluconate
2
“
ysis) (G. Poncelet, J. Martens, B. Delmon, P. A. Jacobs, and P. Grange,
Eds.), Vol. 91, p. 227. Elsevier, Amsterdam, 1995.
is explained by considering (i) the large gap in the oxida-
tion rate of the aldehydic and the alcoholic groups, (ii) the 22. Despeyroux, B. M., Deller, K., and Peldszus, E., in “New Develop-
ments in Selective Oxidation” (G. Centi and F. Trifir o` , Eds.), p. 159.
Elsevier, Amsterdam, 1990.
inertness toward oxidation of secondary alcoholic group,
and (iii) the operative pH at which isomerization of glu-
cose is avoided. Comparing gold with platinum or palla-
dium on carbon catalysts, we conclude that bismuth must
be involved to obtain good results in the cases of platinum-
2
3. Abbadi, A., and van Bekkum, H., J. Mol. Catal. A: Chemical 97, 111
1995).
(
2
4. Abbadi, A., Makkee, M., Visscher, W., van Veen, J. A. R., and van
Bekkum, H., J. Carbohydr. Chem. 12, 573 (1993).