Bromide-free TEMPO-mediated oxidation of primary alcohol groups in starch and methyl α-D-glucopyranoside

Article abstract of DOI:10.1016/S0008-6215(00)00109-9

TEMPO (2,2,6,6-tetramethylpiperidine-1-oxyl)-mediated oxidation of potato starch and methyl α-D-glucopyranoside (MGP) was performed in the absence of sodium bromide (NaBr) as co-catalyst, solely using sodium hypochlorite (NaOCl) as the primary oxidant. The low reaction rate associated with a bromide-free process was increased by performing the oxidation at increased temperatures. The reaction proceeded stoichiometrically and with high selectivity and with only minor depolymerisation, provided that temperature and pH were kept ≤20°C and < 9.0, respectively. At 20°C and pH 8.5, the reaction rate was comparable to that of a corresponding oxidation catalysed by NaBr at 2°C. Consequently, this is a simple approach to raise the TEMPO/NaOCl reaction rate under bromide-free conditions while still maintaining good product properties. At higher oxidation temperatures (≥25°C) and under more alkaline conditions (pH ≥ 9.0) degradation of the starch skeleton occurred. Simultaneously, side-reactions of the nitrosonium ion lowered the yield of the oxidation. Despite the absence of the NaBr catalyst, the reaction rate-controlling step was found to be the oxidation of the primary hydroxyl groups with the nitrosonium ion. The reaction was first- order in MGP and in TEMPO. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0008-6215(00)00109-9

www.elsevier.nl/locate/carres
Carbohydrate Research 328 (2000) 355–363
Bromide-free TEMPO-mediated oxidation of primary
alcohol groups in starch and methyl
a-
D
-glucopyranoside
Petter L. Bragd a,*, Arie C. Besemer a, Herman van Bekkum b
a SCA Research, Utrechtseweg 48, NL-3700 AJ Zeist, The Netherlands
b Laboratory of Organic Chemistry and Catalysis, Delft Uni6ersity of Technology, Julianalaan 136,
NL-2628 BL Delft, The Netherlands
Received 13 January 2000; received in revised form 27 March 2000; accepted 4 April 2000
Abstract
TEMPO (2,2,6,6-tetramethylpiperidine-1-oxyl)-mediated oxidation of potato starch and methyl a-
D-glucopyra-
noside (MGP) was performed in the absence of sodium bromide (NaBr) as co-catalyst, solely using sodium
hypochlorite (NaOCl) as the primary oxidant. The low reaction rate associated with a bromide-free process was
increased by performing the oxidation at increased temperatures. The reaction proceeded stoichiometrically and with
high selectivity and with only minor depolymerisation, provided that temperature and pH were kept 520 °C and
B9.0, respectively. At 20 °C and pH 8.5, the reaction rate was comparable to that of a corresponding oxidation
catalysed by NaBr at 2 °C. Consequently, this is a simple approach to raise the TEMPO/NaOCl reaction rate under
bromide-free conditions while still maintaining good product properties. At higher oxidation temperatures (]25 °C)
and under more alkaline conditions (pH]9.0) degradation of the starch skeleton occurred. Simultaneously,
side-reactions of the nitrosonium ion lowered the yield of the oxidation. Despite the absence of the NaBr catalyst,
the reaction rate-controlling step was found to be the oxidation of the primary hydroxyl groups with the nitrosonium
ion. The reaction was first-order in MGP and in TEMPO. © 2000 Elsevier Science Ltd. All rights reserved.
Keywords: TEMPO; Oxidation; Starch; Methyl a-D-glucopyranoside; Kinetics
1. Introduction
Stable nitroxyl radicals like TEMPO
(2,2,6,6-tetramethylpiperidine-1-oxyl) (I) and
its analogues have gained interest during the
last 15 years as versatile and useful reagents
for conversions of alcohols to carbonyl com-
pounds [1]. The actual oxidising species is the
nitrosonium ion (II) [2], the oxidised form of
TEMPO.
Upon acting as oxidant species, II is con-
verted into the hydroxylamine (III), the re-
duced form of TEMPO. In a TEMPO-
mediated process regeneration of the nitroso-
nium ion takes place in situ by oxidants, such
as m-chloroperbenzoic acid, copper salts,
sodium bromite and sodium hypochlorite/
sodium bromide [3–6]. Recently, additional
* Corresponding author. Tel.: +31-30-6944855; fax: +31-
30-6991199.
E-mail address: petter.bragd@hygiene.sca.se (P.L. Bragd).
0008-6215/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(00)00109-9

356
P.L. Bragd et al. / Carbohydrate Research 328 (2000) 355–363
oxidants such as N-chlorosuccinimide and
[bis(acetoxy)iodo]benzene were described [7,8].
The TEMPO oxidation process has been
applied to the oxidation of various polysac-
charides, including potato starch, inulin,
guar gum, pullulan, dextrins, cellulose, car-
boxymethyl cellulose and chitosan, into the
C-6 aldehyde and carboxy derivatives [9–13].
In all these studies, sodium hypochlorite/
sodium bromide was used as the regenerating
oxidising system of II, and hypobromite,
which is the active oxidising species in this
process, is formed in situ by oxidation of
sodium bromide by sodium hypochlorite.
TEMPO (I) is regenerated in situ via the
hydroxylamine (III), which reacts with II to
form TEMPO. It is generally accepted that
catalytic cycle proceeds as depicted in Scheme
1(A). It cannot be ruled out that TEMPO is
regenerated through a direct oxidation of the
hydroxylamine by action of the primary oxi-
dant [1]. Oxidation of the primary alcohol
proceeds via a reactive C-6 aldehyde interme-
diate to the C-6 carboxy product. Both steps
are probably catalysed by TEMPO [11]. In
general, 10% sodium bromide (w/w) with re-
spect to the substrate was added in previous
works. From an industrial point of view, a
process without sodium bromide would be
more attractive, since its presence in the waste
stream is highly undesired, due to environ-
mental and corrosion concerns. Very recently,
a procedure solely using hypochlorite as pri-
mary oxidant was described [14]. The low
reaction rate associated with sodium bromide-
free oxidation in conventional systems, was
increased by the use of ultrasound. Indepen-
dently, we have made a study of TEMPO
oxidation in the absence of sodium bromide
which we present here. In a bromide-free pro-
cess, the nitrosonium ion is generated by di-
rect oxidation by hypochlorite (Scheme 1(B)).
In the previous studies, the TEMPO/sodium
hypochlorite/sodium bromide oxidation was
usually carried out at 0–5 °C. Our approach
to accelerate the reaction was to perform the
oxidation at increased, but still moderate tem-
peratures (20–30 °C). In these experiments
potato starch was used as the substrate. The
implications on parameters like degree of oxi-
dation, molecular weight and reaction rate
were studied. For comparison, a few oxida-
tions catalysed by sodium bromide were also
performed at increased temperature. Further-
more, to elucidate reaction orders of the reac-
tants and the rate-determining step in the
absence of sodium bromide, a kinetic study
was conducted, using methyl a- -glucopyra-
D
noside as the substrate.
2. Experimental
General methods.—The molecular mass dis-
tribution was determined with a Pharmacia
8
HPLC AKTA EXPLORER 10, equipped with
a UV–Vis detector (UV-900). Uronic polysac-
charides were monitored at 215 nm. The SEC-
column system was a TOSOHAAS TSKgel G
4000 PW, MW 1000–700,000 connected in
series with a TSKgel G 6000 PW, MW
500,000–50,000,000. As eluent, 0.05 M NaCl
was used. Flow rate was set at 0.5 mL/min.
Under the given conditions V0 is ca. 12 mL.
NMR spectra were recorded on a Varian
VXR-400 S spectrometer using D2O as sol-
vent. As internal standard, t-BuOH was used
(CH3 at 31.2 ppm). Quantitative 13C NMR
spectra were recorded with a 0.33° flip angle,
an acquisition delay of 10 s, 50,000 datapoints
Scheme 1. (A) Simplified pathway of the TEMPO oxidation
cycle with NaOCl/NaBr as primary oxidant (the conversion
of the aldehyde intermediate to carboxylate proceeds via the
same reaction sequence). (B) Oxidation of TEMPO by
hypochlorite to the nitrosonium ion in the absence of NaBr
(the substrate oxidation is the same as in (A).
1
and H decoupling during acquisition only.
The uronic acid content in the oxidised mate-

P.L. Bragd et al. / Carbohydrate Research 328 (2000) 355–363
357
Oxidation of methyl h-
D
-glucopyranoside
rials was determined with the Blumenkrantz
assay [15], which method allows quantitative
determination without interfering with neutral
sugars. The latter assay is assumed to repre-
sent the carboxylic groups arising from
TEMPO–C-6 oxidation. The total carboxylic
acid content of the oxidised materials (C-6–
TEMPO oxidation and non-selective oxida-
tion) was determined by acid–base titration.
After dissolution of the oxidised starch (0.2 g)
in distilled water, NaOH was added (15 mL
0.05 M) and the soln was back-titrated to pH
7 with 0.1 M HCl to give the carboxyl
content.
Materials and reagents.—Potato starch (wa-
ter content 18%) was obtained from AVEBE
(Groningen, The Netherlands). Sodium
hypochlorite ca. 15% (w/w) was obtained from
Akzo Nobel (Arnhem, The Netherlands).
TEMPO was obtained from Sigma (St. Louis,
MO, USA). All other chemicals were of com-
mercial grade and were used without further
purification.
(MGP).—All kinetic experiments were carried
out using MGP as a model compound. After
dissolution of MGP (1.94 g, 0.01 mol) in 250
mL of distilled water, NaOCl was added in
10% excess with respect to the amount needed
for complete conversion of C-6 hydroxyl to
carboxylic acids (11 mL of 15% NaOCl). The
pH was adjusted to 8.5 and kept at this value
throughout the oxidation by addition of 0.5
M NaOH, controlled by a pH stat. The reac-
tion flask was kept under N2 to avoid CO2
contamination. The reaction was started by
adding TEMPO (0.065 mmol from a 5 mg/mL
soln). The oxidations were monitored by the
consumption of NaOH as a function of time.
This consumption was assumed to represent
the formation of uronic acid. In some experi-
ments aliquots were taken throughout the re-
action and uronic acid formation was
determined using the Blumenkrantz assay.
Oxidation of potato starch.—If not other-
wise stated, all experiments were carried out
according to the following procedure: potato
starch (12.2 g) was gelatinised in demi-water
(450 mL) at ca. 95 °C under mechanical stir-
ring. The soln was then slowly cooled to the
oxidation temperature. In most experiments,
the oxidation was performed at 2090.5 °C in
a thermostatted water bath. TEMPO (4–8 mg
TEMPO/g starch) was dissolved in the starch
soln. NaOCl was added in portions of 2 mL
throughout the oxidation in order to minimise
unwanted side reactions, such as glycolic
cleavage of the C-2–C-3 vicinal diol moiety.
The pH was kept at 8.5 by addition of 0.5 M
NaOH, controlled by a pH stat. After comple-
tion of the reaction, remaining carbonyl inter-
mediates were reduced to hydroxyl groups by
NaBH4 (0.2 g). The oxidised materials were
precipitated in 2 volumes of ethanol and then
subsequently filtrated and rinsed with acetone.
The materials were dried in a vacuum oven at
30 °C for 10–15 h. All materials were re-dis-
solved in water, ion-exchanged (Dowex
50WX8-100 cation exchanger, Sigma) and
subsequently freeze dried in order to remove
salts and to convert carboxylates to the proto-
nated form.
3. Results and discussion
TEMPO oxidation of potato starch —
effect of temperature and pH.—In a compar-
ative study, starch was oxidised by
TEMPO/sodium hypochlorite with and with-
out sodium bromide as the catalyst. The re-
sults are summarised in Table 1. As the starch
samples only were partially oxidised (1 and 1.4
mol sodium hypochlorite/mol primary alco-
hol, respectively), some aldehydes will be
present at the end of the oxidations. Conse-
quently, the uronic acid content is expected
not to exceed ca. 42 and 65 mol%, respec-
tively. The results indicate that the oxidations
have proceeded stoichiometrically in the ab-
sence of sodium bromide at all tested tempera-
tures, except at 25–30 °C. At otherwise
identical conditions, no variation of the uronic
acid content was observed for different con-
centrations of TEMPO (4–8 mg TEMPO/g
starch). The uronic content of the starch mate-
rials oxidised in the presence of the sodium
bromide catalyst also suggested stoichiometric
conversion, but the obtained values were
somewhat higher.
In an attempt to further investigate the
selectivity of the oxidations, the products were

358
P.L. Bragd et al. / Carbohydrate Research 328 (2000) 355–363
Table 1
Test matrix for the TEMPO oxidation of potato starch at different temperatures with and without NaBr a
Experiment
TEMPO
(g/g starch)
NaOCl
NaBr
(% w/w)
Temperature
(°C)
Time
(min)
Uronic acid
(mol%) c
Carboxyltotal
(mol%) d
(mol/mol) b
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
4
8
4
8
4
8
4
8
4
8
4
8
4
4
4
4
4
1
1
1.4
1.4
1
2
2
2
115
110
330
275
120
100
250
180
65
48
46
72
70
49
45
65
63
42
43
61
60
60
62
72
73
78
44
48
62
68
46
42
62
60
45
45
66
65
70
70
74
75
70
2
10
10
10
10
20
20
20
20
25
30
2
1
1.4
1.4
1
1
45
1.4
1.4
1.4
1.4
1.4
1.4
1.4
185
120
140
115
170
108
64
10
10
10
10
20
a See Section 2 for further conditions.
b The amount of sodium hypochlorite added (mol NaOCl/mol primary alcohol), which is equivalent to an oxidation degree of
42 and 65 mol%, respectively.
c Determined by Blumenkrantz assay oxidation (taken as 6-CH2OH oxidation).
d Determined by acid–base titration (total COOH content).
Table 2
13C NMR data for oxidised starch samples (l in ppm)
Temperature
(°C)
NaBr
(% w/w)
pH
C-1
(C-2–C-5) a
C-6
C-6ox
C-6ox/(C-6ox+C-6) area
ratio of integrals (%)
2
20
2
20
20
8.5
98.5
99.8
99.2
99.8
100.5
73.2–74.5
73.5–75.5
73.0–74.6
72.0–74.5
172.0–176.2
61.7
62.0
62.0
177.6
177.6
177.6
75
68
70
67
46
8.5
8.5
8.5
10
10
62.0
177.6
10
59.2–62.0 b
176.0–180.0 b
a No attempts were made to assign the four ring carbon signals.
b Several small signals.
investigated by 13C NMR (Table 2), since
pyranoside saccharides can be oxidised by hy-
pohalites at near neutral pH to give 2,3-dicar-
boxy-systems [16–19]. However, in all the
oxidised samples partially C-6 oxidised starch
was the only product that could be detected.
The integrals of the C-6ox and C-6 peaks also
indicated stoichiometric conversions of the re-
actions. Consequently, the specificity for C-6
oxidation is not seriously deteriorated at
sodium bromide-free conditions when the
temperature is kept at 520 °C.
The differences in uronic acid and total
carboxylic acid content are to be attributed to
the accuracy of the methods. The rate of
sodium bromide-free oxidation at 15–20 °C
was in the same range as for the correspond-
ing sodium bromide catalysed oxidation at
2 °C (Table 1). The concentration of TEMPO
and the pH also have a large influence on the

P.L. Bragd et al. / Carbohydrate Research 328 (2000) 355–363
359
Table 3
Test matrix and results for the TEMPO/NaOCl oxidation of potato starch at different pH
pH
8.5
9
9.5
NaOCl (mol%) a
Temperature (°C)
Time (min)
Uronic acid (mol%) b
1.4
1.4
1.4
1.4
20
20
20
20
185
135
213
180
66
58
44
29
10
a Equivalent to an oxidation degree of ca. 65 mol%.
b Determined by the Blumenkrantz assay.
TEMPO/(H)OCl oxidation of methyl h-
glucopyranoside (MGP) — kinetics and
D
-
reaction rate (see Section 3.2), but an in-
crease of temperature is probably the most
simple approach to increase the reaction rate
in a sodium bromide-free process. Some oxi-
dations in the absence of sodium bromide
were carried out at higher pH (up to 10) and
gave materials with a lower degree of oxida-
tion while the effect was more pronounced at
higher pH (Table 3).
mechanism.—In the absence of TEMPO, no
consumption of NaOH could be detected at
the given conditions (pH 8.5). Therefore, in
the kinetic experiments presented below gly-
colic cleavage of C-2, C-3 diols or any other
side reactions are not taken into consider-
ation. Further, an experiment was conducted
to investigate whether sodium bromide-free
This trend is probably not due to non-se-
lective oxidation or to disproportionation of
hypochlorite, since the optimum rate of both
reactions is ca. pH 7 [20,21]. Instead, the
likely explanation is that the reaction yield of
the TEMPO-mediated oxidation is lower at
more alkaline pH. The pH dependence is dis-
cussed more detailed in the reaction kinetics
section below. Furthermore, there appears to
be a major depolymerisation at higher pH
since the 13C NMR C-1, C-6ox and C-6 sig-
nals in the sample oxidised at pH 10 were
splitted into several smaller peaks. As alkalin-
ity and temperature were expected to affect
the extent of polysaccharide degradation, all
oxidised potato starch materials were
analysed by size exclusion chromatography
(SEC) to determine the molecular mass distri-
bution. It was seen that starch, oxidised in
the absence of sodium bromide at pH 8.5 at
520 °C, degrades only to a limited extent
(Fig. 1). Materials which were oxidised in the
presence of sodium bromide under the same
conditions appear to be slightly more de-
graded. A broadening of the signal occurs at
20 °C (Fig. 2). At higher pH, there is a severe
depolymerisation (Fig. 3). It is well known
that C-6 oxidised glucopyranosides are sus-
ceptible to b-alkoxy elimination reactions un-
der alkaline conditions [22].
Fig. 1. Relative molecular mass distribution (SEC retention
time vs. absorbance at 215 nm) for TEMPO/NaOCl oxidised
starch at different temperatures at pH 8.5. For comparison, a
spectrum of a reference sample oxidised at 2 °C with 10%
NaBr (w/w) is given.
Fig. 2. Relative molecular mass distribution (SEC retention
time vs. absorbance at 215 nm) for TEMPO/NaOCl oxidised
starch catalysed by NaBr at different temperatures at pH 8.5.
The amount of NaBr is 10% (w/w) with respect to the
substrate.

360
P.L. Bragd et al. / Carbohydrate Research 328 (2000) 355–363
It was assumed that the TEMPO-mediated
reaction is responsible for both consecutive
oxidation steps [11]. Furthermore, when
sodium bromide was present, De Nooy et al.
[11] observed that oxidation of the aldehyde
to the carboxylic acid is ca. seven times faster
than the oxidation of the primary alcohol to
the carbonyl intermediate. In order to simplify
the kinetic expression, a steady state approxi-
mation was applied here and, after an initial
brief period, the rate of formation and disap-
pearance of reactive carbonyl intermediate is
assumed to be balanced. It was also assumed
that the reaction rate-determining step is the
oxidation of methyl a- -glucopyranoside with
D
the nitrosonium ion (II). The following kinetic
expression applies:
Fig. 3. Relative molecular mass distribution (SEC retention
time vs. absorbance at 215 nm) for TEMPO/NaOCl oxidised
starch at different pH at 20 °C.
d
−
=k[MGP]t[nitrosonium]t
(1)
dt
TEMPO is cycled in the course of the reac-
tion. Consequently, the concentration of II is
assumed to be constant and the expression
simplifies to
d
dt
−
=kobs[MGP]t
(2)
Integration gives
!
"
[MGP]0−[uronic acid]t
ln
=kobst
(3)
(4)
[MGP]0
Fig. 4. Course of the TEMPO/NaOCl oxidation at 20 °C and
pH 8.5. ", MGP; ꢀ, aldehyde intermediate; , 1-o-methyl-
a-D-glucuronic acid.
where
[MGP]t=[MGP]0−[uronic acid]t
TEMPO oxidation proceeds as a consecutive
two-step reaction via the C-6 carbonyl inter-
mediate to the C-6 carboxy product. Aliquots
were taken throughout the reaction and
quenched in 95% ethanol. Uronic acid ([acidt])
was determined with the Blumenkrantz assay.
Consumed hypochlorite was determined with
iodometric titration and subsequently corre-
lated to concentrations of formed uronic acid
and aldehyde with [ald]t=[NaOCl]−2[acid]t.
Non-reacted alcoholic substrate was calcu-
lated with [MGP]t =[MGP]0−[ald]t−[acid]t.
It followed from this experiment that the car-
bonyl intermediate reaches its maximum con-
centration at ca. 30–50% reaction, in
agreement with the theory for first-order con-
secutive reactions (Fig. 4).
Both the influence of pH and of the respec-
tive concentrations of TEMPO, sodium
hypochlorite, and methyl a-
D
-glucopyranoside
-glucopyranoside
on the rate of methyl a-
D
oxidation was studied. The progress of the
oxidation for an experiment performed under
standard conditions (20 °C, pH 8.5) is shown
in Fig. 5. The consumption of NaOH gave a
good estimation of uronic acid formation
throughout the reaction. Thus, this technique
has been used in the experiments when other-
wise not stated.
Apart from the initial phase of the reaction,
possibly due to building up the steady-state
concentration of nitrosonium ion, the oxida-
tions appear to be first-order in substrate (Fig.
6). Accordingly, kobs in Eq. (2) is independent

P.L. Bragd et al. / Carbohydrate Research 328 (2000) 355–363
361
Scheme 2.
Fig. 5. Uronic acid formation vs. time in the TEMPO/NaOCl
oxidation of MGP as monitored by the Blumenkrantz assay
(ꢀ) and by the consumption of NaOH (—).
of the initial concentration of methyl a-
D
-glu-
copyranoside within the tested concentration
interval (10–64 mM). The oxidation rate ap-
pears also to be independent of the amount of
sodium hypochlorite (the molar ratio of
sodium hypochlorite/methyl a- -glucopyra-
D
noside varied between 1.4 and 3.9) (Fig. 7).
There is a linear dependence between the over-
all rate constant (kobs) and the amount of
TEMPO (the molar ratio of TEMPO/methyl
a- -glucopyranoside varied between 0.003 and
D
0.012), in accordance with Eq. (1) (Fig. 8).
Under the conditions described, the rate-de-
termining step of the TEMPO/sodium
hypochlorite oxidation is the oxidation of the
alcohol to the carbonyl intermediate by the
nitrosonium ion. Thus steps (5) and (7) are
distinctly faster than step (6) (Scheme 2). The
same behaviour was observed by De Nooy et
al. in the presence of sodium bromide [11]. At
forehand, we presumed that oxidation of
TEMPO to the nitrosonium ion using sodium
hypochlorite would be much slower in the
absence of NaOBr, becoming the rate-deter-
mining step and responsible for the lower rate
in a bromide-free experiment. However, also
in the absence of sodium bromide the results
clearly show that step (6) is rate determining.
The catalytic effect of sodium bromide in a
TEMPO/(H)OCl mediated oxidation remains
to be explained.
Fig. 6. ln([MGP0]−[acidt]) vs. time for three initial concen-
trations of MGP: ꢀ, 0.01 M; , 0.04 M; ", 0.064 M.
Fig. 7. Rate constants (104kobs (s−1)) vs. NaOCl/MGP (molar
ratio).
Effects of pH on the reaction rate.—The pH
has a large impact on the rate of oxidation
both in the presence and absence of sodium
bromide. An increase of pH starting from
mildly alkaline conditions is accompanied by
a distinct increase of the reaction rate. How-
ever, under more severe alkaline conditions,
the rate is lowered again. The pH optimum is
somewhat lower in the absence of sodium
Fig. 8. Rate constants (104kobs) vs. 103[TEMPO]/[MGP].

362
P.L. Bragd et al. / Carbohydrate Research 328 (2000) 355–363
bromide (ca. 9.3 compared to ca. 10.5 for a
sodium bromide-catalysed oxidation) (Fig. 9).
This is advantageous in connection with the
occurrence of b-alkoxy elimination reactions.
De Nooy et al. [10] observed that the reaction
was retarded at higher alkaline pH, but no
actual pH optimum was found. The retarda-
tion at higher pH was ascribed to the low
concentration of hypochlorous acid due to
virtually complete dissociation to hypochlorite
at pH 11 and it was suspected that step (5)
would be retarded and would become rate
limiting under these conditions.
However, in addition to the retardation of
step (5), we also believe that side-reactions of
the nitrosonium ion account for the rate de-
crease. Endo et al. [23] reported that nitroso-
nium ion reacts with hydroxyl ions under
alkaline conditions (pH\9) and reverts to the
TEMPO radical, while forming hydrogen per-
oxide (Scheme 3).
significant above ca. pH 9.5, and in the pres-
ence of sodium bromide above pH 10. Devia-
tions from the first-order kinetics occur also at
pH57, probably due to the slower regenera-
tion of TEMPO (step (7)) at neutral or lower
pH. Another explanation is that side reac-
tions, such as the glycolic cleavage of C-2–C-3
diols might contribute to the overall hydrox-
ide consumption at this pH, and thus cause
deviations.
Semmelhack et al. [24] proposed that reac-
tion between an alcoholic substrate and nitro-
sonium ion takes place through a nucleophilic
addition following
a
cyclic elimination
(Scheme 4). De Nooy et al. [11] postulated
that this pathway is likely under alkaline con-
ditions. The hydroxyl ion could facilitate the
rate-determining step by adopting a proton
from the CH2OH group, thus explaining the
pH dependence of the oxidation under weakly
alkaline conditions.
The latter reaction could explain the consid-
erable deviations from the first-order kinetics
occurring at alkaline conditions. In the ab-
sence of sodium bromide this effect becomes
4. Conclusions
In the present study, starch and methyl
a- -glucopyranoside have been selectively oxi-
D
dised by the bromide-free TEMPO/(H)OCl
system at somewhat higher temperatures (5–
30 °C), in order to increase the reaction rate.
In the oxidation of starch, the temperature
had to be increased to 20 °C to obtain similar
reaction rate as a bromide-catalysed system at
2 °C. An increased reaction rate was also ob-
tained at higher concentrations of TEMPO.
The oxidation of potato starch was found to
proceed stoichiometrically to C-6 carboxy
starch, except at more alkaline conditions
(pH]9.0), presumably due to side reactions
of the nitrosonium ion. The specificity for C-6
oxidation was not lost at bromide-free condi-
tions when the temperature was kept at 5
20 °C and pHB9.0. Analogously, at tem-
peratures 520 °C and pHB9.0, the depoly-
merisation of the polysaccharide backbone
was only minor, but at higher tem-
peratures and more alkaline conditions severe
depolymerisation occurred. Fortunately, pH
8.5 is near the optimal pH to activate the
system. Thus, good product properties
and reaction rates can be obtained after a
Fig. 9. Relative oxidation rate vs. pH at 20 °C for TEMPO/
(H)OCl (") and TEMPO/(H)OCl/NaBr (ꢀ). The oxidations
catalysed by NaBr (10% (w/w)) were carried out at 2 °C, in
order to avoid competitive hypohalite oxidations.
Scheme 3. Reaction of nitrosonium ion with hydroxyl groups
at high pH.
Scheme 4. The reactive intermediate formed in the oxidation
of the CH2OH group of MGP to the C-6 aldehyde.

P.L. Bragd et al. / Carbohydrate Research 328 (2000) 355–363
363
[4] M.F. Semmelhack, C.R. Schmid, D.A. Corte´s, C.S.
Chou, J. Am. Chem. Soc., 106 (1984) 3374–3376.
[5] T. Inokuchi, S. Matsumoto, T. Nishiyama, S. Torii, J.
Org. Chem., 55 (1990) 462–466.
[6] P.L Anelli, C. Biffi, F. Montanari, S. Quici, J. Org.
Chem., 52 (1987) 2559–2562.
careful choice of experimental conditions.
Some kinetic experiments on TEMPO oxida-
tion of methyl a-
D
-glucopyranoside revealed
that the reaction is first-order in methyl a-
D-
glucopyranoside and first-order in TEMPO. It
was concluded that the rate-determining step
is the oxidation of the substrate with the
nitrosonium ion. The obtained first-order ki-
netics apply only to weakly alkaline condi-
tions (pH 8.5–9.0). At lower, near neutral pH,
large deviations occur as the regeneration step
of TEMPO radical becomes rate limiting. Un-
der more alkaline conditions, the nitrosonium
ion reverts to the TEMPO radical causing
large deviations.
[7] J. Einhorn, C. Einhorn, F. Ratajczak, J.-L. Pierre, J.
Org. Chem., 61 (1996) 7452–7454.
[8] A. de Mico, R. Margarita, L. Parlanti, A. Vescovi, G.
Piancatelli, J. Org. Chem., 62 (1997) 6974–6977.
[9] A.E.J. De Nooy, A.C. Besemer, H. van Bekkum, Recl.
Tra6. Chim. Pays-Bas, 113 (1994) 165–166.
[10] A.E.J. De Nooy, A.C. Besemer, H. van Bekkum, Carbo-
hydr. Res., 269 (1995) 89–98.
[11] A.E.J. De Nooy, A.C. Besemer, H. van Bekkum, Tetra-
hedron, 51 (1995) 8023–8032.
[12] P.S. Chang, J.F. Robyt, J. Carbohydr. Chem., 15 (1996)
819–830.
[13] A. Isogai, Y. Kato, Cellulose, 5 (1998) 153–164.
[14] G. Descotes, Y. Queneau, in W. Praznik, A. Huber
(Eds.), Carbohydrates as Organic Raw Materials IV,
Verlag, Vienna, 1998, p. 39.
[15] N. Blumenkrantz, G. Asboe-Hansen, Anal. Biochem., 54
(1973) 484–489.
Acknowledgements
[16] M. Floor, A.P.G. Kieboom, H. van Bekkum, Starch, 41
(1989) 348–354.
[17] A.C. Besemer, H. van Bekkum, Starch, 46 (1994) 95–
101.
[18] A.C. Besemer, H. van Bekkum, Starch, 46 (1994) 101–
106.
[19] A.C. Besemer, H. van Bekkum, Recl. Tra6. Chim. Pays-
Bas, 113 (1994) 398–402.
The authors thank A. Sinnema (Delft Uni-
versity of Technology) for measuring and in-
terpreting the 13C NMR spectra and Dr J.W.
Thornton (SCA Research) for fruitful
discussions.
[20] R.L. Whistler, R. Schweiger, J. Am. Chem. Soc., 79
(1957) 6460–6464.
[21] J. Schmorak, D. Mejzler, M. Lewin, Starch, 14 (1962)
278–290.
References
[22] R.L. Whistler, J.N. BeMiller, Ad6. Carbohydr. Chem., 13
(1958) 289–329.
[1] A.E.J. De Nooy, A.C Besemer, H. van Bekkum, Synthe-
sis, 10 (1996) 1153–1174.
[23] T. Endo, T. Miyazawa, S. Shiihashi, M. Okawara, J.
Am. Chem. Soc., 106 (1984) 3877–3878.
[24] M.F. Semmelhack, C.R. Schmid, D.A. Corte´s, Tetra-
hedron Lett., 27 (1986) 1119–1122.
[2] J.M. Bobbit, C.L. Flores, Heterocycles, 27 (1988) 509–
533.
[3] J.A. Cella, J.A. Kelley, E.F. Kenehan, J. Org. Chem., 40
(1975) 1860–1862.
.