Regioselective oxidation of hydroxyl groups of sugar and its derivatives using silver catalysts mediated by TEMPO and peroxodisulfate in water

Article abstract of DOI:10.1006/jcat.2000.2927

Primary hydroxyl groups were oxidized regioselectively to carboxylic acid using organic nitrosonium salts generated on supported silver catalysts, which promote disproportionation of 2,2,6,6-tetramethylpiperidinyl-1-oxy (TEMPO) in aqueous solution. The oxidation reactions were performed at pH 9.5 in a batch reactor at room temperture using heterogeneous silver catalysts and peroxides as primary co-oxidants; e.g., 99 mol% selectivity to methyl-α-D-glucopyrasiduronic acid was obtained at 90% conversion of the pyranoside using a silver carbonate-celite catalyst . The efficiency of the system was increased by adding carbonates to the silver catalyst. This result is explained by the increase of the electron charge deficiency on silver in the presence of carbonate, which accelerates the nucleophilic attack of hydroxyls and/or TEMPO. In the case of the Ag-Al₂O₃ catalyst, this result was proved by temperature-programmed desorption measurements using ammonia. With primary/secondary polyols, the selectivity for the primary hydroxyl groups is high. In addition, primary hydroxyl groups, as in the case of pyranosides, were oxidized more selectively than those of the furanosides. The observed regioselectivity is due to the sterical hindrance caused by the four methyl groups in TEMPO.

Full text of DOI:10.1006/jcat.2000.2927

Journal of Catalysis 194, 343–351 (2000)
doi:10.1006/jcat.2000.2927, available online at http://www.idealibrary.com on
Regioselective Oxidation of Hydroxyl Groups of Sugar and Its Derivatives
Using Silver Catalysts Mediated by TEMPO and Peroxodisulfate in Water
,1
H. Kochkar, L. Lassalle, M. Morawietz,† and W. F. Ho¨ lderich
Department of Chemical Technology and Heterogeneous Catalysis, University of Technology, RWTH-Aachen, Worringerweg 1,
52074 Aachen, Germany; and †Degussa-Hu¨ls, Rodenbacher Chaussee 4, 64304 Hanau, Germany
Received February 7, 2000; revised May 10, 2000; accepted May 17, 2000
That means that such oxidized compounds have an impor-
tant economic impact.
Primary hydroxyl groups were oxidized regioselectively to car-
boxylic acid using organic nitrosonium salts generated on sup-
ported silver catalysts, which promote disproportionation of 2,2,6,6-
tetramethylpiperidinyl-1-oxy (TEMPO) in aqueous solution. The
oxidation reactions were performed at pH 9.5 in a batch reactor at
room temperture using heterogeneous silver catalysts and perox-
ides as primary co-oxidants; e.g., 99 mol% selectivity to methyl- -
D-glucopyrasiduronic acid was obtained at 90% conversion of the
pyranoside using a silver carbonate–celite catalyst . The efficiency
of the system was increased by adding carbonates to the silver cata-
lyst. This result is explained by the increase of the electron charge
deficiency on silver in the presence of carbonate, which accelerates
the nucleophilic attack of hydroxyls and/or TEMPO. In the case
of the Ag–Al2O3 catalyst, this result was proved by temperature-
programmed desorption measurements using ammonia. With pri-
mary/secondary polyols, the selectivity for the primary hydroxyl
groups is high. In addition, primary hydroxyl groups, as in the case
of pyranosides, were oxidized more selectively than those of the
furanosides. The observed regioselectivity is due to the sterical hin-
c
The oxidation reaction should advantageously be con-
ducted in a catalytic manner. Among the known cata-
lytic methods, the oxidation of alcohols with 2,2,6,6-tetra-
methylpiperidine-1-oxyl (TEMPO) (1) and a co-oxidant,
such as sodium hypochlorite or sodium bromide, became
a very attractive tool for organic synthesis (2). The au-
thors used the stable organic nitroxyl radical 2,2,6,6-tetra-
methylpiperidine-1-oxyl (TEMPO) (1) as a mediator. The
actual oxidant is the nitrosonium ion (2) (ammonium ox-
ide), which can be obtained from (1) by several methods.
During the oxidation, species (2) is reduced to hydrox-
ylamine (3) (Fig. 1).
The nitroxyl radical (1) can be used in stoichiomet-
ric amounts by treatment with the following methods: p-
toluenesulfonic acid or 1(S)-(+)-camphor-10-sulfonic acid
(3) and methyl iodide (4). However, the high chemical re-
activity generally hinders its isolation and purification. For-
tunately, radical (1) can also be used in a catalytic amount
as a mediator. For example, Fe(III) ions oxidize TEMPO
to the nitrosonium ion (2), which shows good activity in
the oxidation of benzyl alcohol to form benzaldehyde in
a two-phase system (aqueous phase/acetonitrile or methy-
lene chloride) (5). In this case, nitrosonium ion (2), which is
a kind of ammonium salt, accelerates the reaction by acting
as an anion transport. The disadvantage of this method is
that the Fe(III) ion was used in excess (3 equiv of Fe(III)
ions, 2 mol% of TEMPO, and 1 mol of substrate), which is
not feasible for industrial purposes in view of environmen-
tal protection and costs.
drance caused by the four methyl groups in TEMPO.
2000 Academic
Press
Key Words: regioselective; oxidation; TEMPO; carbohydrates;
silver; peroxides; carbonate.
1. INTRODUCTION
The oxidation of alcohols is very common among organic
chemical reactions and is widely used in many processes
in industry. Especially, this holds true for the oxidation of
primary alcohols to aldehydes or carboxylic acids.
Metal-catalyzed oxidation of alcohols to carboxylic com-
pounds, in conjunction with co-oxidants, is an important
step in the synthesis of fine chemicals (1). The oxidation of
sugars and their derivatives as well as of starch and cellu-
lose is particularly important. The oxidized carbohydrates
can be used as thickening and gelling agents, in paints and
resins, and as detergent co-builders and superabsorbers.
The nitroxyl radical (1) can also be oxidized by cupric
salts to the nitrosonium ion (2), which oxidizes alcohols
such as benzyl alcohol and n-hexanol to the corresponding
aldehydes, such as benzaldehyde and n-hexanal, the hy-
droxylamine (3), and a protonic acid. The hydroxylamine
was in turn easily oxidized by cupric salts to give a ni-
troxyl radical (1) again. But it was necessary to trap
the acid because it inhibits the catalytic oxidation sys-
tem. Therefore, cupric hydroxide or pyridine were used
1 To whom correspondence should be addressed. Fax: 49 241 88 88 29.
E-mail: hoelderich@rwth-aachen.de.
343
0021-9517/00 $35.00
c
Copyright
2000 by Academic Press
All rights of reproduction in any form reserved.

344
KOCHKAR ET AL.
More recently, the oxidation of primary hydroxyl groups
of polysaccharide by means of TEMPO in water has been
described. These investigations are related to a process for
producing completely carboxylated carbohydrates. In this
process, carbohydrates having a carbonyl content of at least
75% were oxidized using the hypochlorite/bromide sys-
tem (14). The amount of hypochlorite solution used was 2–
2.4 mol/mol of monosaccharide unit. This clearly shows that
the process produces a lot of halogenated salts, which cause
some environmental pollution and corrosion problems.
The catalytic oxidation processes of primary hydroxyl
groups cited above have the following disadvantages:
FIGURE 1
as acid-trapping agents, again in stoichiometric amounts
(6, 7).
Cella et al. (8) performed the oxidation in the presence
of m-chloroperbenzoic acid. Thereby, the nitrosonium ion
(2) was generated in methylene chloride as a solvent. Using
this system, a large range of alcohols could be oxidized to
the corresponding aldehydes or ketones. However, alcohols
having oxygen in the -position cannot be oxidized by the
nitrosonium ion and those in the -position react slowly.
The drawback of this method is the use of an excess of
m-chloroperbenzoic acid (1.5 mol/mol of alcohol). This fact
results in many purification steps for the products.
Primary alcohols are quantitatively oxidized to alde-
hydes under two-phase conditions (methylene chloride–
aqueous sodium hypochlorite) in the presence of catalytic
amounts of 4-methoxy-2,2,6,6-tetramethylpiperidine-1-
oxyl (4-MeO–TEMPO). Co-catalysis by sodium bromide
and buffering at pH 8.6 with NaHCO3 are also required
(9, 10).
In two-phase systems,
(i) the oxidation of alcohol leads to the aldehyde only,
(ii) the nitroxyl radical and the co-oxidant were used in
stoichiometric amounts, and
(iii) many purification steps for the products are re-
quired.
In the aqueous phase,
(i) a large excess of the co-oxidant, e.g., hypochlorite,
was employed and
(ii) problems associated with this method are, sooner
or later, loss of the oxidizing power of the nitroonium
salts due to the concurrently generated hydrogen perox-
ide and molecular chlorine (15) and an accumulation of
halogenated salts.
The present paper is to show another method for the
catalytic oxidation of alcohols by means of nitroxyl radicals
(1) using a heterogeneous system.
We report here a new method for the regioselective ox-
idation of primary hydroxyl groups over a silver catalyst/
peroxodisulfate system mediated by TEMPO. The oxida-
tion was performed first with methyl- -D-glucopyranoside
since it has been used as a model molecule.
Many authors described a method for the oxidation of
primary hydroxyl groups in an aqueous phase. The reac-
tion is mediated by TEMPO. Thereby, hypobromite is used
as the regenerating oxidant, which in turn is regenerated by
another hypohalite as the oxidant (see Scheme 1), prefer-
ably in the form of a salt, such as lithium hypochlorite,
sodium hypochlorite, potassium hypochlorite, or calcium
hypochlorite (11–13).
SCHEME 1. Proposed mechanism for the oxidation of primary hydroxyl groups11.

OXIDATION OF HYDROXYL GROUPS OF SUGAR/DERIVATIVES
345
2. EXPERIMENTAL
potassium hydrogen carbonate (0.15 g) in distilled water
(10 ml) was slowly added. After complete addition, stir-
ring was continued for 10 min. The yellow-green precipitate
formed was collected by filtration and dried under reduced
pressure on a rotary evaporator for several hours. The cata-
lyst contains 0.17 mmol of silver/g of celite (0.085 mmol of
Ag2CO3/g of reagent).
2.1. Materials
Methyl- -D-glucopyranoside ( -MDG), AgO, Ag2O,
AgNO3, pyridine, hexamethyldisilazane, trifluoroacetic
acid, perborate, percarbonate, oxone, and peroxodisulfate
were bought from Aldrich. NaY was kindly provided by
Zeolyst International and AlPO4 and Al2O3 by BASF-AG
(D10-10). Celite was purchased from Merck and was used
after purification with hydrochloric acid (10 wt% ).
Experiments were performed in a thermostated glass
batch reactor of 250 ml volume, equipped with a con-
denser, a thermometer, and a magnetic stirrer. The pH was
kept constant using a pH meter (Titroline Alpha, Schott)
coupled to a pH control unit and an automatic burette
(Metrohm 655, 20-ml piston) containing 1 M KOH, which
consumption was recorded during the reactions.
2.3. Physicochemical Characterisation
The composition of the catalysts was determined by in-
ductively coupled plasma (ICP) after dissolution of the
solids. The specific BET surface and average pore diameter
were obtained from the adsorption isotherms of nitrogen
at 77 K, using a Micromeritics automatic volumeter (ASAP
2000), after evacuation of the samples at 473 K for 16 h. An
X-ray diffraction study was performed on a Siemens 5000
a
instrument equipped with a Cu lamp (
spectra were recorded from 2 = 5 to 2 = 100 .
= 1.54 A). The
cu
The amount of acidic sites was determined by measure-
ments of temperature programmed desorption (TPD) of
ammonia. On samples previously calcined at 673 K under
helium flow (overnight), the ammonia was adsorbed for
15 min at 373 K. The temperature was increased using a
ramp of 10 C/min to 1073 K. The desorbed ammonia was
titrated by HCl (0.025 M) using a special titrometer coupled
to a pH control unit and an automatic burette containing
the acid. During the desorption, the temperature, pH, and
HCl consumption were recorded.
2.2. Preparation of the Catalysts
2.2.1. Metal-supported on Al2O3, AlPO4, and NaY. Ag–
Na–Y, Ag–AlPO4, and Ag–Al2O3 were prepared by incip-
ient wetness impregnation. Silver nitrate was added to a
suspension containing 10 g of the support. The mixture
was stirred for 15 h and then filtrated, washed, and dried
overnight at 373 K. Then the catalyst was calcined under air
at 773 K for 6 h. All the surface areas, porosity, and silver
content are reported in Table 1.
2.2.2. Preparation of the silver carbonate on celite. The
supporting material celite was purified by washing with
methanol containing 10 wt% concentrated hydrochloric
acid at 25 C and afterward treated with distilled water until
the mother liquor was neutral. Then the material was dried
at 393 K.
To a mechanically stirred solution of silver nitrate
(1.85 mmol, 0.31 g) in distilled water (50 ml), purified celite
(10 g) was added. To the stirred suspension, a solution of
2.4. Oxidation Procedure
A defined amount of substrate (5 mmol) was dissolved in
100 ml of water. Then 30 mg (0.19 mmol) of TEMPO and
10 mmol of co-oxidant were added. The system was equili-
brated at 298K and pH 9.5. The catalyst wasinstantaneously
added to the mixture (time zero). The pH of the solution
during the reaction was kept constant at 9.5 by automatic
titration with 1 M KOH solution. At the end of the reaction,
the mixture was neutralized by adding hydrochloric acid to
a pH of 8. Then the catalyst was recovered by filtration, and
water and TEMPO were removed under vacuum at 313 K.
The product mixture was dried overnight under vacuum at
room temperature. The oxidation degree of the substrate
was determined by gas chromatography after silylation.
For silylation, e.g., 10 mg of the oxidized material was
introduced into a gas chromatography tube and 600 l of
pyridine was added. The solution was stirred until the mix-
ture was totally dissolved. Then 600 l of hexamethyldisi-
lazane and 150 l of trifluoroacetic acid were added. The
mixture was placed in an oven at 363 K for 30 min. After
addition of n-decane as the internal standard, quantitative
analysis was carried out. The products were analysed by a
GC Siemens RGC 202 equipped with a capillary column
ultra chromapack 25 m in length and a flame ionization
detector (FID).
TABLE 1
Characterisations of the Catalysts
b
Catalyst/
support
Ag
(wt% )a
Surface area
Pore diameter h i
1
(m2 g
)
(A)
˚
Al2O3
253
A
B
C
D
1.1
2.0
2.8
5.0
65
65
64
64
247
249
241
NaY
573
E
5.8
<12
AlPO4
41
F
4.5
20 < < 40
a Determined by ICP-AES.
b Determined by using the BJH method.

346
KOCHKAR ET AL.
TABLE 2
Oxidation of -MDG over Silver Catalysts Mediated by TEMPO,
at pH 9.5, 298 K, Molar Ratio -MDG/TEMPO = 26 [( ), Molar
Ratio of the Substrate to Silver; Solvent, Water]
SCHEME 2
(NH4)2S2O8/
-MDG
Conversion
(% )
Selectivity
(mol% )
Catalyst
was used in the presence of a catalytic amount of nitroxyl
radicals, with a molar ratio of -MDG to silver 108, a
very high selectivity to methyl- -D-glucopyrasiduronic acid
(97 mol% ) was obtained at a conversion of 70% . In this
procedure the silver has the function of a catalyst.
No
2
0
0
2
9
49
71
70
86
95
96
97
Ag2O (0.5)
AgO (0.5)
Ag–Al2O3 (108)
In the absence of TEMPO radicals but with Ag–Al2O3
and PDS the conversion was lower than 2% . This allows
us to conclude that the absence of TEMPO radicals in this
type of reaction makes it ineffective and nonselective, while
the presence of a catalytic amount of radical increases the
3. RESULTS AND DISCUSSIONS
3.1. Preliminary Experiments
First, a kind of model reaction, the oxidation of methyl- - regioselectivity of the oxidation to a very high degree. The
D-glucopyranoside ( -MDG), was performed (Table 2). It chosen catalysts are good for the regioselective oxidation
can be pointed out that in the presence of silver oxide cata- of carbohydrates; they give high selectivities to carboxylic
lysts the conversion depends on the oxidation state of silver: acid salt without cleavage of the pyranoside ring.
high conversion of pyranoside ( -MDG) was obtained in
For the oxidation of -MDG, Mombarg et al. (16) re-
the case of silver in a high oxidation state. These experi- ported a conversion of about 60% after 20 h using [Ti]–
ments show clearly the ability of the silver oxide to activate MCM-41 and H2O2, but the selectivity to the hydroxyl
and regenerate the TEMPO. But we noted that silver oxide group in the C6 position is low due to the decomposition of
was reduced to metallic silver as shown by XRD (Fig. 2), H2O2 and the oxidative cleavage of the vicinal diol groups.
whatever the oxidation state of the silver oxide. These re- Earlier experiments(17) byour group showed that V–Y and
sults support the idea that the reaction takes place in a Co–Y zeolites gave 15% and 60% conversion, respectively,
two-step procedure (see Scheme 2).
with 99% selectivity to the acid.
Using the peroxodisulfate (PDS) in combination with
Heyns and Paulsen (18) reviewed the platinum-catalyzed
TEMPO but in the absence of silver oxides only a low con- selective oxidation of monosaccharides with oxygen. They
version of 9% was obtained. However, when Ag–Al2O3 concluded that the order of reactivity of the functional
FIG. 2. X-ray spectra of (a) Ag(I)oxide before the catalytic test and (b) after the catalytic test indicating Ag0.

OXIDATION OF HYDROXYL GROUPS OF SUGAR/DERIVATIVES
347
TABLE 4
groups toward oxygen is as follows: hemiacetal > primary
hydroxyl > axial hydroxyl > equatorial hydroxyl. Hence, to
oxidize selectively the primary hydroxyl groups, a protec-
tion of the hemiacetal function is generally required. For
the oxidation of L-sorbose, Baiker (19) proposed the use
of acetone as a complexing agent. The selectivity to the ke-
toacid can also be improved by applying strongly adsorbing
amines, so-called supported bases, instead of increasing the
bulk pH of the medium. The bases used are cycloaliphatic
and aromatic amines, phosphines, and aminophosphines.
Nevertheless, these base modifiers have a bad effect on the
activity of the catalyst (20).
Effect of the PDS/TEMPO Molar Ratio on the Catalytic
Properties of Ag–Al2O3 Catalyst
Conversion
(% )
Selectivity
(mol% )
PDS/TEMPO
25
50
75
125
250
27
70
42
30
19
98
97
96
87
70
Note. -MDG /TEMPO = 26; pH 9.5; RT; solvent, water.
On carbon-supported platinum, Schuurman et al. (21)
showed that the activation energy for the formation of car-
boxylic acid from primary alcohol and oxygen at pH 10 is
about 110 kJ mol 1. de Nooy et al. (12) demonstrated, us-
ing TEMPO as a mediator, that the reaction requires only
58 kJ mol 1. So this low enthalpy of activation makes the
use of TEMPO very attractive. However, comparison with
the present results obviously shows that a silver catalyst in
combination with PDS as the co-oxidant yields better re-
sults. In addition, we do not need to protect the hemiacetal
function of the monosaccharide.
in turn decomposes in the presence of silver to give mainly
O2 (22). To verify this hypothesis, perborate and percarbon-
ate were replaced by H2O2 (35 wt% ). As shown in Table 3,
no activity was obtained and we noted fast decomposition
of hydrogen peroxide (few minutes).
However, when oxone (2KHSO5 KHSO4 K2SO4) was
used, a very high selectivity of 99% is also obtained at 6%
conversion of the -MDG. In this case, oxone decomposes
to give mainly O2 and SO24 with small amounts of PDS
(22). The activity observed in this case is assumed to be
related to the in situ formation of the PDS. Nevertheless, the
conversion remained very low here due to the deposition
of sulfate anions on the surface of the catalyst.
3.2. Co-oxidant Effect
We studied the effect of the co-oxidant on the oxida-
tion of -MDG in the presence of TEMPO and silver sup-
ported on alumina (2 wt% ). The results are reported in
Table 3. Ammonium PDS achieves a very high selectivity
to the acid, 99% , at 78% conversion of the -MDG. In con-
trast, when the reaction was carried out in the presence of
sodium percarbonate or sodium perborate as the primary
oxidant, very low conversions were achieved, maintaining
99% selectivity. The reason might be that the perborate and
percarbonate aqueous solutions decompose to H2O2, which
For the oxidation of -MDG, de Nooy et al. (23) re-
ported selectivity to methyl- -D-glucopyranosiduronic acid
higher than 95% using the hypochlorite/bromide system
and TEMPO at pH 10, but no indication of the conversion
is given. Mombarg et al. (16) reported 20% selectivity to
methyl- -D-glucopyranosiduronic acid at 60% conversion
using [Ti]–MCM-41 and H2O2 as the oxidant. Using the
de Nooy system, the relative rate of the reaction is about
0.02 mmol min 1. In contrast, our system demonstrated an
activity higher than 0.1 mmol min 1. A comparison with
the present results shows that the Ag–Al2O3 catalyst with
TEMPO and PDS as the co-oxidant obviously yields better
results.
TABLE 3
Effect of the Co-oxidant
Selectivitiesc (mol% )
Conversion
3.3. Effect of Peroxodisulfate/TEMPO Ratio
Catalyst
Oxidant
PDS
Oxone
Perborate
H2O2 (35 wt% )
(% )
(i)
(ii) (iii) (iv)
The PDS/TEMPO molar ratio was varied between 25 and
250 (Table 4). No conversion was observed in the absence of
PDS. In contrast, the conversion reaches a maximum at an
approximate PDS/TEMPO ratio close to 50 and decreases
at higher ratios. This result can be explained by complex
formation between the cationic oxidizing agent and the sul-
fate anion (Scheme 3). The hydrogen bond is expected to
Ag–Al2O3
78
6
<1
0
99
99
99
—
—
—
—
—
—
—
—
—
—
—
—
Ti–MCM-41a H2O2 (35 wt% )
TEMPO
60
n.i.
21
>95
51.6 17.6 9.6
ClO–/Br– b
—
—
—
Note. Reaction conditions: (methyl- -D-glucopyranoside) 5 mmol, mo- lower the positive charge on the nitrogen, which would be
lar ratio oxidant/ -MDG = 2; (TEMPO) 0.1 mmol, 0.50 g of catalyst,
298 K, water as the solvent, pH 9.5.
a See Ref. (16).
the driving force in this reaction path. Miyazawa et al. (4)
demonstrated that the nucleophilicity of the counteranion
of this cationic oxidizing agent has an important effect on
the catalytic properties; e.g., secondary alcohols are oxi-
dized faster than primary alcohols in the presence of Br .
b See Ref. (23).
c (i) Methyl- -D-glucopyrasiduronic acid; (ii) formic acid; (iii) glycolic
acid; (iv) tartronic acid.

348
KOCHKAR ET AL.
SCHEME 3. Supposed complex between TEMPO and sulfate anion.
The opposite results were observed in the presence of Cl .
On the other hand, the adsorption of polar substrates or
products can modify this charge (12). The selectivity to the
acid decreases at high PDS ratios; this might be explained
by the sulfation effect, which increases the acidity of the
catalyst, resulting in the ring opening of the pyranoside.
FIG. 4. Conversion of -MDG and selectivity to methyl- -D-
glucopyuronic acid in five successive runs on Ag/Al2O3 catalyst: -MDG/
TEMPO = 26 (mol); oxidant/ -MDG = 2 (mol); 298 K; solvent, water;
pH 9.5.
3.4. Effect of the Percentage of Silver
D-glucopyranoside, PDS, and TEMPO solutions. Figure 4
shows the conversion of methyl- -D-glucopyranoside and
the selectivity to methyl- -D-glucopyrasiduronic acid as a
function of the number of runs. The conversion decreases
dramatically after the third run, from about 50% to 17% .
These results can be explained by the formation of sulfate,
which remains adsorbed on the metal surface, blocking the
catalytically active centers. The effect of sulfate was veri-
fied as follows: the parent catalyst was treated by sulfate an-
-MDG oxidation reactions were conducted versus the
silver content supported on alumina, adjusting a molar ratio
-MDG/PDS/TEMPO = 1/2/0.1. Figure 3 illustrates con-
version and selectivity to methyl- -D-glucopyranosiduronic
acid as a function of the silver percentage. -MDG oxida-
tion is very low in the absence of silver around 9% , whereas
the conversion increases with the percentage of silver. At
2 wt% Ag, the conversion and selectivity to methyl- -D-
glucopyranoside are 78% and 98 mol% respectively. This
means the silver works as a catalyst. We noted also that in-
creasing the silver content does not affect the surface area.
The BET surfaces remain around 250 m2/g. The silver con-
centration was measured in the reaction medium and in the
catalyst after the reaction. These measurements show that
the amount of silver measured on the catalyst was the same
as that introduced initially in the solution. Therefore, silver
is quantitatively fixed on the catalyst.
2
ions by wetness impregnation (SO4 :Ag = 20 : 1) and then
checked in the same conditions. The conversion obtained
in this case is lower than that of the parent catalyst 40% .
By means of FT-IR spectroscopy, we noted the presence
of a band at 600 cm 1, which is attributed to the vibration
of sulfate anion (24). In conclusion, the presence of sulfate
anions deactivates the silver catalyst.
3.6. Effect of the Support
Four supports with different textures were checked:
Al2O3, NaY, celite, and AlPO4 (Fig. 5). The activities
obtained vary in the order Ag–carbonate–celite AgY >
Ag–AlPO4 > Ag–Al2O3, with selectivities to methyl- -D-
3.5. Stability of the Catalysts
The Ag–Al2O3 catalyst with 5 wt% silver was used in five
successive oxidation runs with a fresh charge of methyl- -
FIG. 3. Effect of the percentage of silver on the conversion of -MDG.
FIG. 5. Effect of the carrier.

OXIDATION OF HYDROXYL GROUPS OF SUGAR/DERIVATIVES
349
glucopyrasiduronic acid of 99 mol% , except that the Ag–
AlPO4 catalyst demonstrated a 86 mol% selectivity.
The Ag–carbonate–celite catalyst, which is commonly
known as Fe´tizon reagent, is very active. It was reported
(25) that this catalyst is a very versatile and useful oxida-
tion agent and is one of the most important among the so-
called supported reagents. Similar regiospecific oxidations
of related compounds have been reported (26–28). Olefinic
sugars are readily oxidized at the allylic hydroxyl groups in
favour of free nonallylic groups. However, the oxidation
fails completely with the corresponding acetylenic carbo-
hydrates, which very likely undergo a cleavage leading to
acetylene and aldehydes (29, 30).
SCHEME 4. Supposed complex between silver and carbonate anion.
creases by a factor of 10 in the case of Ag–Al2O3–(CO23 ).
Thisresult can be interpreted bythe fact that in the presence
of carbonate ions, the charge density on silver decreases
(Scheme 4) and then silver is more ionic. This result is sup-
ported by the following observations:
The acidity of the used supports can be classified in the
following order: AlPO4 Al2O3 >Na–Y> celite. The low
conversion of the -MDG (20% ) observed in the case of
Ag–AlPO4 is probably due to the low stability of the sup-
port under basic conditions (pH 9.5), whereas the final con-
versions obtained using Ag–Y or Ag–Al2O3 are nearly the
same.
We suggest that the high activity observed in the case of
silver carbonate on a celite catalyst, which presents a very
small surface area (6 m2 g 1), is related to the carbonate
ions. Carbonate ions can modify the electron density on
silver to generate the activity.
(i) TPD of ammonia (see Fig. 7) shows clearly that the
absorbed ammonia increases after treatment of the catalyst
with carbonate. This result is consistent with the idea that
the silver is more ionic.
(ii) When water is replaced by a less polar solvent (cy-
clohexane or benzene), the conversion decreases 6% and
3% , respectively, due to less solvation of silver. This result
confirms the ionicity of silver, which can be stabilized here
by carbonate anions.
3.8. Substrate Effect
We were especially interested in the influence of the ring
size because it was found that primary hydroxyl groups in
pyranosides were oxidized more selectively than those in
furanosides (11). The model substrates, 1,2-propanediol,
-MDG, saccharose, and potato starch, were oxidized
(see Table 5). Obviously, primary alcohols are oxidized
more rapidly than secondary ones. With primary/secondary
polyol (1,2-propanediol), the selectivity for the primary hy-
droxyl groups is high and reaches 75 mol% at 90% con-
version. During oxidation, the lactic acid reacts with the
1,2-propanediol to form the corresponding ester.
3.7. Role of Carbonate
Methyl- -D-glucopyranoside oxidation reactions were
conducted on Ag–Al2O3 and Ag–Al2O3–(CO23 ) catalysts.
Because the selectivity to methyl- -D-glucopyrauronic acid
is extremely high, the conversion can be conveniently mea-
sured from the amounts of potassium hydroxide added to
the reaction medium to maintain a constant pH. In Fig. 6,
the conversion as a function of reaction time is demon-
strated over the parent Ag–Al2O3 catalyst and for Ag–
Al2O3–(CO23 ) with CO23 /Ag = 0.5 mol. The initial rate in-
FIG. 7. Temperature-programmed desorption of ammonia from sup-
FIG. 6. Methyl- -D-glucopyranoside conversion as a function of re-
ported silver on alumina.
action time.

350
KOCHKAR ET AL.
TABLE 5
Substrate Effect
Selectivities (mol% )
Conversion
(% )
Substrate
acid
others
1,2-Propanediol
-MDG
Saccharose
Starch
90
78
20
<1d
20d,e
75
99
100b
100c
99c
25a
1
—
SCHEME 5. Supposed complex between TEMPO and substrate.
4. MECHANISM
Starch
1
Note. 0.50 g of Ag–Al2O3 (2 wt% ); pH 9.5; 298 K; molar ratio of
substrate/TEMPO = 26; solvent, water.
2,2,6,6-Tetramethylpiperidinyl-1-oxy (TEMPO) can be
activated under basic conditions using silver catalysts to
nitrosonium salt. The latter is a strong oxidant and shows
selectivity toward primary hydroxyl groups over secondary
ones.
We demonstrated a modified method for the regioselec-
tive oxidation of polysaccharides. The reaction mediated
by TEMPO and a silver catalyst was used as the regener-
ating oxidant; the reduced silver (or its equivalent) was in
turn regenerated by peroxodisulfate (Scheme 6). For this
reason, the TEMPO skeleton can be used as a recyclable
oxidation catalyst in the presence of a secondary oxidant.
a Products of self-esterification of lactic acid and esterification with 1,2-
propanediol.
b Tricarboxylic acid, identified by GC-MS.
c Determined by 13C NMR (176 ppm, carboxylic acid), FT-IR spec-
troscopy (1742 cm 1, carbonyl function).
d Determined by titration.
e PDS is replaced by NO2 as oxidant.
For pyranoside rings, the selectivity depends on the ster-
ical demand of the primary hydroxyl groups. In general,
the observed regioselectivity for different substrates seems
to depend on the accessibility of the alcohols (Scheme 5).
Primary hydroxyl groups of -MDG were oxidized more
rapidly than those of saccharose (100 mol% to the tricar-
boxylic acid at 20% conversion). In the oxidation of various
pyranosides it appeared that the size of the ring played an
important role. The oxidation of starch was not successful.
This might be explained by diffusion limitations.
5. CONCLUSIONS
As compared with the conventional methods, the pro-
cess of the present work has the following merits: Ex-
tremely high activity and selectivity are obtained in the
oxidation of primary alcohols or primary hydroxyl groups,
e.g., in the case ofcarbohydrates. More particularly, primary
hydroxyl groups of methyl- -D-glucopyranoside are oxi-
dized regioselectively in the presence of secondary groups.
Side reactions are diminished. The catalytic oxidation pro-
cess presents the same advantages as the method used by
For comparison reasons, oxidation of starch was per-
formed on a fluidized bed usingNO2 asan oxidant. From 13
C
NMR, UV–visible, and FT-IR spectroscopy, we concluded
that oxidation takes place regioselectively in primary hy-
droxyl groups (31).
SCHEME 6. Supposed mechanism using silver catalyst, TEMPO, and peroxodisulfate as a primary oxidant.

OXIDATION OF HYDROXYL GROUPS OF SUGAR/DERIVATIVES
351
Van Bekkum et al., namely, high selectivities achieved and 11. de Nooy, A. E. J., Besemer, A. C., and van Bekkum, H., Recl. Trav.
Chim. Pays-Bas. 113, 165–166 (1994).
12. de Nooy, A. E. J., Besemer, A. C., and van Bekkum, H., Tetrahedron
51(29), 8023–8032 (1995).
13. Chang, P. S., and Robyt, J. F., J. Carbohydr. Chem. 15(7), 819–830
mild reaction conditions (oxidation at atmospheric pres-
sure, room temperature (RT), and water as the solvent).
However, the catalyst is deactivated by sulfate anions. Fur-
ther investigations of the use of silver and TEMPO in car-
(1996).
bohydrate oxidation in combination with other oxidants are 14. Besemer, A. C., and de Nooy, A. E. J., Patent, WO 95/07303, 1995.
15. Endo, T., Miyazawa, T., Shiihashi, S., and Okawara, M., J. Am. Chem.
in progress.
Soc. 106, 3877–3878 (1984).
16. Mombarg, E. J. M., Osnabrug, S. J. M., van Rantwijk, F., and van
ACKNOWLEDGMENTS
Bekkum, H., Heterog. Catal. Fine Chem. IV. 385 (1997).
17. Kochkar, H., and Ho¨lderich, W., Patent DE 99118600.8 (priority
This work was carried out in the frame of the European Project
09/21/99), Degussa AG.
BRITE/CARBOPOL-CT96-1208. The authors are very grateful to the
18. Heyns, K., and Paulsen, H., Adv. Carbohydr. Chem. 17, 169 (1962).
European Community for financial support. Furthermore, they thank
19. Baiker, A., Presented at 5th International Symposium on Heteroge-
Dr. R. Vanheertum (Degussa–Hu¨ls AG) for his steady interest in this
neous Catalysis and Fine Chemicals, Lyon, France, 8/30–9/3, 1999.
work and for fruitful discussions.
20. Mallat, T., Bro¨nnimann, C., and Baiker, A., J. Mol. Catal. A 117, 425
(1997).
REFERENCES
21. Schuurman, Y., Kuster, B. F. M., van der Wiele, K., and Marin, G. B.,
Appl. Catal. 89, 31–46 (1992).
1. Sheldon, R. A., and Kochi, J. K., “Metal-Catalyzed Oxidations of Or-
ganic Compounds.” Academic Press, New York, 1981.
2. Semmelhack, M. F., Schmid, C. R., Cortes, D. A., and Chou, C. S.,
J. Am. Chem. Soc. 105, 4492–4494 (1983).
22. Cotton, F. A., and Wilkinson, G., “Advanced Inorganic Chemistry,”
4th ed. Wiley, New York, 1980.
23. de Nooy, A. E. J., Besemer, A. C., and van Bekkum, H., Carbohydr.
Res. 269, 89–98 (1995).
3. Ma, Z. H., and Bobbitt, J. M., J. Org. Chem. 56(21), 6111–6114 (1991).
4. Miyazawa, T., Endo, T., Shiihashi, S., and Okawara, M., J. Org. Chem.
50(8), 1332–1334 (1985).
5. Miyazawa, T., and Endo, T., J. Mol. Catal. 31, 217–220 (1985).
6. Semmelhack, M. F., Schmid, C. R., Cortes, D. A., and Chou, C. S.,
J. Am. Chem. Soc. 106, 3374–3376 (1984).
7. Miyazawa, T., and Endo, T., J. Mol. Catal. 32, 357–360 (1985).
8. Cella, J. A., Kelly, J. A., and Kenehan, E. F., J. Org. Chem. 40, 62 (1975).
9. Anelli, P. L., Biffi, C., Montanari, F., and Quici, S., J. Org. Chem. 52,
2259–2562 (1987).
10. Inokuchi, T., Matsumoto, S., Nishigama, T., and Torii, S., J. Org. Chem.
55, 462–466 (1990).
24. Bellamy, L. J., in “The Infra-red Spectra of Complex Molecules,”
pp. 345–346. Wiley, New York, 1962.
25. Fe´tizon, M., and Golfier, M., C. R. Acad. Sci. Paris C267, 900 (1968).
26. Pelyvas, I., Ztricskai, F. S., and Bognar, R., Carbohydr. Res. 76, 257
(1979).
27. Talapatra, B., Ray, T., and Talapatra, S. K., J. Indian. Chem. Soc. 55,
1204 (1978).
28. Leboul, J., Cleophax, J., Gero, S. D., Rolland, A., and Forchioni, A.,
Tetrahedron 33, 965 (1977).
29. Horton, D., and Tsai, J.-H., Carbohydr. Res. 75, 151 (1979).
30. Mijs, W. J., and De Jonge, C. R. H. I., “Organic Syntheses by Oxidation
with Metal Compounds,” pp. 542–544. Wiley, New York, 1986.
31. Kochkar, H., Morawietz, M., and Ho¨lderich, W., to be submitted.