Polysaccharide depolymerization from TEMPO-catalysis: Effect of TEMPO concentration

Carbohydrate Polymers 170 (2017) 140–147

Contents lists available at ScienceDirect

Carbohydrate Polymers
journal homepage: www.elsevier.com/locate/carbpol

Polysaccharide depolymerization from TEMPO-catalysis: Effect of
TEMPO concentration
Vivian C. Spier a,b , Maria Rita Sierakowski a , Wayne F. Reed b , Rilton A. de Freitas a,∗
a
b

BioPol, Chemistry Department, Federal University of Paraná, Curitiba, Paraná, 81531-980, Brazil
Tulane Center for Polymer Reaction Monitoring and Characterization (PolyRMC), Tulane University, New Orleans, LA, 70118, USA

a r t i c l e

i n f o

Article history:
Received 6 February 2017
Received in revised form 11 April 2017
Accepted 23 April 2017
Available online 26 April 2017
Keyword:
Xyloglucan
N-oxil-2,2,6,6-tetramethylpiperidine
(TEMPO)
Automatic continuous online monitoring of
polymerization reactions (ACOMP)

Automatic continuous mixing (ACM)
Size exclusion chromatography (SEC)
Depolymerization

a b s t r a c t
Polysaccharide TEMPO-oxidation was monitored using automatic continuous online monitoring of polymerization reactions (ACOMP). The products of oxidation, obtained at different pHs (9, 7 and 5) and
different concentrations of catalyst TEMPO, were evaluated by Automatic Continuous Mixing (ACM)
and Size Exclusion Chromatography (SEC). The degree of oxidation was higher at pH 9 and polysaccharide degradation was observed under different pH conditions, but was much higher without catalyst
TEMPO. The rate constant (k) was dependent on reaction pH and TEMPO concentration. The amount of
−COOH per g of polysaccharide, at pH 9, in the presence and absence of TEMPO was different, 0.215 and
0.395 mmol g−1 , respectively. This suggested a secondary and non-selective polysaccharide oxidation
occurring at a lower rate in the absence of catalyst. TEMPO protects the polysaccharide from degradation caused by secondary oxidant species, acting as a catalyst and “sacrificial molecule” at higher
concentrations.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction
The compound 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO)
has been used as a catalyst in polysaccharide oxidation reactions, acting selectively on primary alcohols. The main advantage
of TEMPO-catalyzed oxidation, compared to other catalysts, is its
selectivity, low reactivity in the presence of air, light, humidity, and
storage without special conditions (Kato, Matsuo, & Isogai, 2003;
Sierakowski, Milas, Desbrières, & Rinaudo, 2000; Sierakowski, de
Freitas, Fujimoto, & Petri, 2002; Bragd, Besemer, & Van Bekkum,
2000; de Nooy, Besemer, & Van Bekkum, 1995a; de Nooy, Besemer,
& Van Bekkum, 1995b; de Nooy, Besemer, Van Bekkum, Van Dijk,
& Smit, 1996).
Polysaccharide TEMPO-oxidation is usually performed under
alkaline conditions. However, there are few reports in the literature about the use of acid reaction conditions, affecting directly
the selectivity and the degree of oxidation (Watanabe, Tamura,
Saito, Habu, & Isobai, 2014). To selectively oxidize the primary

alcohols, reactive oxygen species NaClO/NaBrO are used as a secondary oxidizing agent, promptly oxidizing TEMPO to nitrosonium

∗ Corresponding author.
E-mail addresses: , (R.A. de Freitas).
/>0144-8617/© 2017 Elsevier Ltd. All rights reserved.

ions, the effective oxidant in the catalysis. Nitrosonium ion acts
as a catalyst, oxidizing the polysaccharide primary alcohols to
aldehydes and is reduced to hydroxylamine. The hydroxylamine
can be re-oxidized to nitrosonium ion by the secondary oxidants,
and a second round of oxidation converts the aldehyde groups
to carboxylic acids (Isogai, Saito, & Fukuzumi, 2011; Bragd et al.,
2000; de Nooy et al., 1995a, 1995b, 1996; Sierakowski et al., 2000;
Sakakibara, Sierakowski, Lucyzyn, & de Freitas, 2016).
The most common alkaline conditions, pH >9.0, used by several authors to induce alkoxide formation on polysaccharides,
also demonstrated depolymerization using TEMPO as a catalyst,
even at very low temperature (0–4 ◦ C). Such molar mass reduction was reported for different kinds of polysaccharides (Cunha,
Maciel, Sierakowski, Paula, & Feitosa, 2007; de Freitas, Martin,
Paula, Feitosa, & Sierakowski, 2004; Sakakibara et al., 2016 and de
Souza, Lucyszyn, Ferraz, & Sierakowski, 2011).
Some authors identify methods to avoid depolymerization during TEMPO reactions. For example, Shibata & Isogai (2003) observed
that hydroxyl radicals formed from NaBrO and TEMPO at pH
10–11 can cause depolymerization during oxidation, and that some
scavengers of reactive species could only partially suppress the
depolymerization. Shinoda, Saito, Okita, and Isogai (2012) related
the concentration of NaClO concentration with carboxylate content and the degree of polymerization (DP) of cellulose nanofibrils.
The authors assumed that C6-aldehyde formed as an interme-


V.C. Spier et al. / Carbohydrate Polymers 170 (2017) 140–147

diate structure of TEMPO-oxidation could be associated with a
␤-elimination process, and that the elimination of such intermediates, due to post-oxidation or reduction of aldehydes formed,
reduced such effect.
The alkaline oxidation medium suggests at least one main
hypothesis to explain the polysaccharide degradation during
TEMPO catalyzed oxidation, related to ␤-elimination. Such reaction occurs mainly under alkaline conditions and in the presence of
an aldehyde, acid or ester group at polysaccharide C6, contributes
to increase the acidity of the hydrogen at C5, usually deprotonated
during the alkaline reaction. The reaction intermediate product is a
double linkage between the carbohydrate C4-C5. The end products
are related to an elimination of the group linked to carbon ␤ (C4),
forming a reducing carbohydrate (aldehyde functional group) and
an unsaturated carbohydrate (Anet, 1964).
Another mechanism of depolymerization can be associated
to non-specific oxidation of polysaccharides by the secondary
oxidants. The reactive oxygen species (NaClO/NaBrO), can promote an increase in the reducing carbohydrates, diminishing the
polysaccharide molar mass. Boruch (1985) observed hydrolytic
degradation of starch molecules, with reduction of viscosity and
increasing reducing sugars, during starch oxidation with NaClO.
Hiraoki, Ono, Saito, and Isogai (2015) also observed that the greater
the amount of NaClO during TEMPO reaction conditions, the lower
the molar mass of TEMPO-oxidized celluloses.
Based on the two hypotheses presented above, a secondary
function is discussed, related to catalyst TEMPO, as a “sacrificial molecule”, protecting polysaccharides from depolymerization
processes induced by the reactive oxygen species (NaClO/NaBrO)
present in the reaction medium for catalysis. This was confirmed
using automatic continuous online monitoring of polymerization
reactions (ACOMP), size exclusion chromatography (SEC), different TEMPO concentrations and pHs (9, 7 and 5), to evaluate
␤-elimination and unselective oxidation and depolymerization,

respectively.
2. Material and methods
2.1. Plant material and polysaccharide extraction
Seeds of Tamarindus indica L. were provided by Conceic¸ão de
Almeida, Bahia state, Brazil. The ground seeds were submitted to an
extraction of pigments and lipids, and the isolation of the xyloglucan (XG) was performed as described. Ground seeds were defatted
and depigmented under ethyl ether reflux, using a Soxhlet apparatus, and dried in a fume hood at room temperature. The XG
was isolated after aqueous extraction using a blender, followed
by nylon cloth filtration. The filtered solution was centrifuged at
10,000g, 40 ◦ C for 30 min. After centrifugation, the supernatant was
filtered sequentially by cellulose acetate membranes (3.0, 0.8, 0.45
e 0.22 ␮m) and followed by precipitation using ethanol. After drying under vacuum at 40 ◦ C, the powder obtained was termed XG.
The extraction, chemical and oligosaccharide characterization of
this polysaccharide was published elsewhere (Spier et al., 2015).
2.2. Selective oxidation of xyloglucan
The TEMPO selective oxidation reaction was performed according to studies reported by de Nooy et al. (1995a, 1995b, 1996).
Briefly, a XG solution (0.001 g cm−3 ) was dissolved overnight in
ultrapure water (MilliQ system, USA) at 25 ◦ C, and filtered through
0.45 ␮m cellulose acetate filters (Millipore, Merck KGaA, Germany).
The polymer solution was cooled and the reaction was performed at temperature of 3 ± 1 ◦ C under continuous flow of N2 .
In a reactor under stirring, 8.61 mg mL−1 of a 10% (v/v) sodium

141
®

hypochlorite solution (NaClO – SIGMA ) and 0.077 mg mL−1 NaBr
®
®
(SIGMA ) were added together with TEMPO (SIGMA ), at concen−1
trations of 0, 0.015, 0.030, 0.060 and 0.150 mol L . Previously to

adding the oxidant mixture to the reaction, the secondary oxidants
(NaClO/NaBrO) and TEMPO mixture were mixed, and the solution filtered through cellulose acetate of 0.45 ␮m. The oxidation
was performed at different pH values; 9, 7, and 5, adjusted with
1 mol L−1 HCl solution. During the reaction, the pH values were
maintained using 0.05 mol L−1 NaOH solution.
After the oxidation process, the reaction was stopped using
an alcoholic solution of NaBH4 (0.0015 g cm−3 ). Then, the pH was
adjusted to 7. The end products of oxidation were purified over 48 h
dialysis against ultrapure water, precipitated in ethanol and dried at
room temperature. The oxidation products were analyzed through
automatic continuous mixing (ACM) and size exclusion chromatography (SEC) to evaluate the polyelectrolyte effect and the oxidation
effect on molar mass (Bayly, Brousseau, & Reed, 2002; Sorci & Reed,
2002).

2.3. Reaction monitoring and characterization of the end product
2.3.1. Real time reaction monitoring
During the XG selective oxidation (item 2.2) the reaction was
continuously monitored using an ACOMP system, with a Shimadzu LC-10AD pump and Shimadzu quaternary mixing module
FCV10AL with 1.0 mL min−1 flow rate. The pump continuously
extracted sample from the reactor and flowed it through the following detector train: a Brookhaven BI-MwA multi-angle light
scattering detector (MALS), a Shimadzu RI detector (RID 10A),
and a custom built single capillary viscometer was a custom built
reviewed previously (Reed, 2003). Additionally, this work presents
first time polarimetric detection (AUTOPOL VI Automatic Polarimeter, Rudolph Research Analytical), and DLS detection (NanoDLS
Particle Size Analyzer Brookhaven Instruments) in the ACOMP platform. It is noted that using DLS in a flow cell requires stop-flow
capability, since there is a velocity dependent term in the autocorrelation function not related to diffusion, so that motion of the
scattering liquid must stop during the DLS measurement. The BI
Nanosizer is equipped with stop flow. Measures of pH and conductivity were also monitored (Jenway 3540 pH & Conductivity
meter).

2.3.2. Reaction end product characterization
The XG end products of oxidation after purification (item 2.2),
were characterized by Automatic continuous mixing (ACM) and
SEC. ACM experiments were performed to observe the polyelectrolyte properties of the oxidized products. XG was evaluated at
0.001 g cm−3 , and two solutions were prepared for the same sample, one in purified water and the other in NaNO3 0.1 mol L−1 . Both
samples were filtered through cellulose acetate filters of 0.45 ␮m.
Using a Shimadzu LC-10AD pump and Shimadzu quaternary mixing
module FCV10AL with 1.0 mL min−1 flow rate, a gradient was created from 0 to 0.1 mol L−1 of NaNO3 . The samples passed through
a light scattering detector at ␭ = 660 nm (Brookhaven Instrument
MwA), and the static light scattering signal was followed during
the salt ramp with constant XG concentration.
SEC analyses were also carried out using a Shimadzu LC-10AD
pump, Brookhaven Instruments Corp. BI-MwA MALS detector, a
Shimadzu RI, a custom-built viscometer, and Shodex OHpak SB806 HQ columm, using 0.1 mol L−1 NaNO3 with 0.02% (w/v) NaN3
as an eluent, and 0.8 mL min−1 flow rate. Solutions (0.001 g cm−3 )
were prepared in the mobile-phase, during 16 h and passed through
a 0.22 ␮m cellulose acetate filter (Millipore).


142

V.C. Spier et al. / Carbohydrate Polymers 170 (2017) 140–147

XGoxipH7 and XGoxipH5 are shown in the Supplementary material
(Fig. S1).
To obtain the Zimm data a diluted polysaccharide concentration
(c = 0.001 g cm−3 ) and q2 <S2 >z 1, was used, according to Eq. (1).
1
Kc
=

R␪
Mw

1+

q2 < S 2 > z
3

+ 2A2 c

(1)

where RÂ is the Rayleigh scattering ratio, at a scattering vector
amplitude defined as q = (4 n/ )sin(Â/2), where  is the scattering
angle, c is the polymer concentration, <S2 > z the z-averaged square
of the radius of gyration and A2 is the second virial coefficient (obs:
A2 effect was considered small enough and ignored), and Mw the
weight average molar mass. K is an optical constant (Eq. (2)) given
for vertically polarized incident light by
2n 2
0

4
Fig. 1. mmol −COOH.g−1 of polymer as a function of time, at pH values of 9, 7, and
5 in the presence of catalyst TEMPO (0.015 mol L−1 ), and also without TEMPO at pH
9 (XG oxipH9,NT ).

2.4.

1H

and 13 C-1 H NMR analysis of xyloglucan

The native and oxidized XG were analyzed by monodimensional NMR spectrum (hydrogen – 1 H) and bidimensional (HSQC
– heteronuclear single quantum coherence), in a BRUKER, DRX400 model, AVANCE series. A 5 mm inverse probe was utilized,
with deuterated water (D2 O) as solvent and TMS-p (2,2,3,3-tetradeuterium-3-trimethysilyl sodium propionate salt) as reference for
the calibration spectra (␦ = 0 ppm). All analyses were performed at
60 ◦ C.

K=

NA

2

(2)

4

n0 is the refraction index of the solvent, dn⁄dc is the differential
index of refraction, NA is Avogadro’s number and the laser wavelength.
The reduced viscosity (Áred ) was obtained by
VSample − VSolvent

Áred =

VSolvent − VZero Flow

/cSample

(3)

where VSample , VSolvent and Vzeroflow are the voltage signals from the
viscometer (differential of pressure) for the sample, solvent and
zero flow, respectively.
Assuming, at diluted concentrations that intrinsic viscosity ([Á])
is ∼
= Áred , it was possible to determine the viscometric radius <S>␩
from Flory-Fox (1953), by

3. Results and discussion

3⁄2

[Á] = 6
TEMPO-catalyzed polysaccharide oxidation (mmol −COOH)
was monitored from the amount of NaOH solution titrated during the reaction. As observed in Fig. 1, comparing the reactions
at pH values of 9, 7, and 5, the amount in mmols −COOH per
g of polymer was reduced, respectively from basic to acid conditions. The rate, using first order kinetics, was 1.62 × 10−4 s−1
at pH 9, and at pHs 7 and 5 the rates are almost the same,
∼0.52 × 10−4 s−1 . These three first reactions were made using the
secondary oxidants (NaClO/NaBrO) and TEMPO at concentration
of 0.015 mol L−1 . However, when reactions were performed without TEMPO (XGoxipH9,NT ), the oxidation process occurred at a rate
of 0.58 × 10−4 s−1 , with a total degree of polysaccharide oxidation
higher than the reactions catalyzed by TEMPO, suggesting some
non-selective oxidation in absence of TEMPO (Table 1).
The first observation is that apparently the secondary oxidants
present in the medium, NaClO/NaBrO, can be responsible for a nonselective oxidation of the polysaccharide. In fact, Table 1 shows that
at pH = 9 over 1.8x more oxidation occurs without TEMPO than with
TEMPO as catalyst.

The XG oxidation products at pH 9 with the system
TEMPO/NaClO/NaBrO will be termed XGoxipH9 and for pHs 7 and 5,
XGoxipH7 and XGoxipH5 , respectively. The reaction without TEMPO
as catalyst, but in presence of the secondary oxidants NaClO/NaBrO
at pH 9, will be termed XGoxipH9,NT .
The ACOMP reaction end product values are presented in
Table 2, as a function of the degree of oxidation (mmol -COOH.g−1
of polymer), monitoring the results of oxidation from native to oxidized XG at different pHs: 9, 7 and 5, with and without TEMPO as
catalyst. In Table 2 dn/dc is the differential index of refraction of XG
in solution. In Fig. 2, only the results of XGoxipH9 in the presence of
TEMPO 0.015 mol L−1 and in the absence of TEMPO are shown. For

dn⁄dc

ϕ0

S2

3⁄2
Á

MW

(4)

where ϕ0 = 2.56 × 1023 is the Flory constant.
Because of solubility issues, reactions could only be carried out
at low concentrations of XG ≤0.001 g cm−3 . Hence, this presents
an unusual context for ACOMP, which normally dilutes concentrated media from reactor by factors ranging from 10 to over 1000
times. Here, XG dilution was unnecessary and so it was possible to

simply circulate the reactor contents directly through the detector
train. The recirculation also permits full recovery of the final product, whereas conventional ACOMP normally loses a fraction of the
material in the extraction and dilution stream, which is normally
wasted.
The normal advantage of dilution is that the supporting solvent
under which detection occurs can be modified at will; e.g. ionic
strength (IS) and pH can be changed, solvent mixtures made, etc.
In the case of XG, any dilution of the already very dilute reactor
content degrades detector signals. The disadvantage of the recirculation with no dilution is that light scattering cannot be lowered,
which negatively affects the ability to monitor reaction kinetics due
to the build-up of polyelectrolyte properties.
As observed by ACOMP experiments (Fig. 2 and Table 2) the
sample XGoxipH9 presented a Mw reduction of 8.3% and <S>z of
24%, suggesting that some depolymerization and some increasing
in chain flexibility was observed during oxidation. Sakakibara et al.
(2016) also observed that at pH 9 there is some reduction in the
molar mass of galactomannans and an increase in chain flexibility, due to reduction of persistence length of oxidized products at
moderate ionic strength.
The acid groups formed during XG oxidation were continuously
neutralized by titration with NaOH solution and, as expected, the
polysaccharide even during oxidation processes was maintained


V.C. Spier et al. / Carbohydrate Polymers 170 (2017) 140–147

143

Table 1
First order rate constant (k) of oxidation and oxidation degree in presence and absence of TEMPO at 0.015 mol L−1 .
Sample

k/10−4 (s−1 )

Oxidation degree (mmol −COOH.g−1 )*

R

Oxidation with TEMPO
XGoxipH9
XGoxipH7
XGoxipH5

1.62
0.53
0.52

0.215
0.062
0.046

0.998
0.997
0.995

Oxidation without TEMPO
XGoxipH9,NT

0.58

0.395

0.997

*at 1.1 × 104 s of reaction.
Table 2
End product values from ACOMP; dn/dc, weight average molar mass (Mw ), radius of gyration (<S>z ), intrinsic viscosity ([␩]), Flory-Fox radius of gyration (<S>␩ ), hydrodynamic
radius (Rh ) and optical rotation ([∝]) of Xyloglucan (XG) native and oxidized, after oxidation time of 1.1 × 104 s.
Sample

dn/dc (cm3 g−1 )

Mw (105 g mol−1 )

<S>z (nm)

[␩] (cm3 g−1 )

<S>␩ (nm)

Rh (nm)

25
[∝]D (◦ )

Native sample
XG

0.140

4.8

120

517

39

194

+121

Oxidation with TEMPO
0.154
XGoxipH9
0.153
XGoxipH7
0.153
XGoxipH5

4.4
4.3
4.3

91
95
96

265
300
370

33
35
37

146
118
127

+83
+95
+101

Oxidation without TEMPO
XGoxipH9,NT
0.153

1.6

46

165

10

35

+96

continuously in a low excluded volume state. The change of the

light scattering signal during the oxidation was not large, and small
decreases are associated with small molar mass reduction. Samples monitored by ACOMP at pH values of 7 and 5 yielded almost
the same value of Mw and <S>z (Table 2). However, for the sample
XGoxipH9 , NT the Mw and <S>z reduced 66.7% and 74%, respectively,
confirming that in the absence of catalyst TEMPO, the degradation
of XG was much more significant.
Two mechanisms can be used here to explain the Mw reduction observed by ACOMP experiments. The first one was related to
␤-elimination and the second one due to reactive oxygen species
(NaClO/NaBrO) present in the reaction medium. Some authors previously observed that an increase in the amount of NaClO can be
related to an increase in the degree of oxidation of the products
(Milanovic, Kostic, Milanovic, & Skundric, 2012; Xu, Li, Cheng, Yang,
& Qin, 2014). de Freitas et al. (2004), Milanovic et al. (2012) and
Sakakibara et al. (2016) observed a molar mass reduction during
TEMPO mediated oxidations, for different polysaccharides, almost
in the same experimental conditions.
␤-elimination was minimized due to reduction of pH, however,
our results proved that even at acid pH, the Mw reduction was
also observed. This clearly suggested that ␤-elimination is not the
only mechanism of polysaccharide depolymerization from TEMPO
oxidation reactions using the experimental conditions here (Isogai
et al., 2011). In parallel, non-selective oxidation reactions from the
reactive oxygen species present in the medium can be responsible
for some depolymerization. This hypothesis was confirmed in the
experiments without TEMPO.
This is the first report that finds a function for TEMPO besides
catalysis. Here, it is deduced that TEMPO acts as a “sacrificial
molecule”. In the absence of TEMPO that is promptly oxidized by
NaClO/NaBrO to nitrozonium ion, the secondary oxidants reacted
with the polysaccharide, promoting a non-selective oxidation
and depolymerization. Based on that, TEMPO competes with the

polysaccharide during oxidation, protecting it from depolymerization and non-selective oxidations. Even catalytic amounts of
TEMPO partially protected the extensive depolymerization of the
polysaccharide.
The viscometer and DLS detectors also showed the same tendency of degradation, with [␩] and Rh decreasing during oxidation.
In both cases, the higher values of <S>z are associated to a z-average,

Table 3
Determination of molar mass (Mw ), viscometric radius of gyration <S > Á , dispersion
(Ð = Mw /Mn ) and recovery of XG and XGoxi from SEC experiments.
SAMPLE

Mw (105 g mol−1 )

XG

4.8

With TEMPO
XGoxipH9
XGoxipH7
XGoxipH5

3.1
2.9
3.0

Without TEMPO
1.3
XGoxipH9,NT

S

Ð

Recovery (%)

39

1.4

95

12
20
18

1.8
2.0
2.1

97
93
96

9

1.7

92

Á

(nm)

suggesting that large aggregates formed during the secondary
oxidation/depolymerization weight the ACOMP light scattering
measurements towards higher values. Viscometry is much less sensitive to the presence of aggregates. This technique was also used
to provide S Á , which is a more reliable size parameter than light
scattering <S2 >z when aggregates are present (Table 2, Fig. 2). The
same contamination by aggregation was observed in experiments
of de Freitas, Drenski, Alb, and Reed (2010) analyzing chitosan carboxymethylation, by Spier et al. (2015) analyzing XG enzymatic
depolymerization, due to self-association of the fragments of XG
and by Mkedder et al. (2013) studding the cellulase depolymerization of xyloglucan.
The values of [∝]25
D measured from ACOMP were less positive
for all samples (Table 2, Fig. 2), and apparently, were much more
affected at higher pH or in the absence of TEMPO. The optical
rotation was modified, compared to native polymer, due to formation of glucuronic acid (GlcA) and galacturonic acid (GalA). Isbell &
Frush (1943) observed an optical rotation reduction of 31%, at pH
9.0, from ␤-d-galactose to ␤-d-galacturonic acid, confirming that
oxidation was occurring during real time measurements. Depolymerization and mutarotation of the reducing carbohydrate have a
non-negligible effect on optical rotation.
To confirm the aggregation the end products of oxidation at different pHs and in presence or absence of TEMPO were precipitated,
purified and characterized by SEC (Table 3 and Fig. 3A and B). As
clearly observed by SEC all the pH values led to Mw reduction, in the
presence and absence of TEMPO, respectively. This confirms that,


V.C. Spier et al. / Carbohydrate Polymers 170 (2017) 140–147

5

4.5 10

5

4 10

5

3.5 10

5

3 10

5

2.5 10

5

2 10

5

A1

5 10

5

4.5 10

5

140

B1

140
120

-1

120

MOLAR MASS (g.mol )

5 10

100
80

40
20

1.5 105
1 10

5

0

0.05

0.1

0.15

0.2

3.5 10

5

3 10

5

2.5 10

5

2 10

5

100
80

z

z

60

4 105

60
40
20

1.5 105

0
0.25

1 10

5

0
0

-1

0.1

0.2

0.3
-1

0.4

0.5

mmol COOH.g polymer

mmol COOH.g polymer
A2

560

560

B2

-3

35
30

320

25

15
160

10
80
5
0
0

0.05

0.1

0.15

0.2

0
0.25

35
30

400

25

320

η

η

20

240

480

20
240
15
160
10
80

5

0
0

-1

A3

0.2

0.3

0.4

300

120

B3

140
120

250

100

100
200

h

h

60
100

80
150
60

[α]D (°)

D

[α] (°)

80
150

R (nm)

200

R (nm)

0
0.5

mmol COOH.g polymer

140

250

0.1

-1

mmol COOH.g polymer

300

<S> (nm)

400

Reduced Viscosity (g.cm )

40

<S> (nm)

Reduced Viscosity (g.cm-3)

40
480

<S> (nm)

<S> (nm)

MOLAR MASS (g.mol-1)

144

100
40

50

20

0
0

0.05

0.1

0.15

0.2

0
0.25

-1

mmol COOH.g polymer

40
50

20

0
0

0.1

0.2

0.3
-1

0.4

0
0.5

mmol COOH.g polymer

Fig. 2. ACOMP of XGoxipH9 in presence of 0.015 mol L−1 TEMPO (A) and XGoxipH9 , NT (B). 1–Data from Zimm equation (Eq. (1)), 2–data from viscometer (Eqs. (3) and (4)) and
3- data from DLS and optical rotation, all as a function of mmol COOH.g−1 polymer.

during oxidation and online monitoring by ACOMP the presence of
aggregates are affecting the end values of Mw and S z .
The end-products of oxidation were much better characterized
using SEC analysis than from ACOMP experiments, mainly due to
the presence of aggregates in solution. The polyelectrolyte behavior
of XG oxidized samples was determined by ACM experiments, as
presented in Fig. 4, confirming the oxidation of the polysaccharide
due to increasing of the dependence of Kc/R␪ as a function of NaNO3
concentration.
In the Supplementary material (Fig. S2.1) presented the 1 H
NMR characterization of the anomeric hydrogen for native and oxidized XGs obtained at different values of pH (9, 7 and 5) with TEMPO
and also the sample at pH 9 without TEMPO.

At 5.43 ppm the chemical shift of the ␣-d-Xyl substituted for ␤d-Gal or ␤-d-GalA was observed, and at 5.23 ppm the ␣-d-Xyl not
substituted by ␤-d-Gal or ␤-d-GalA was observed. The ␦ for ␤-d-Glc
was observed at 4.85 ppm and at 4.84 ppm for ␤-d-Gal. For higher
degree of oxidation another ␦ was observed at 4.77–4.78 ppm (XG
oxipH9.0 and XG oxipH9.0,NT ). For XG oxipH5.0 and XG oxipH7.0 it can be
observed only as a shoulder. Based on the chemical shift above, the
ratio Glc: Xyl: Gal was determined for native XG as 2.6: 2.1: 1.0,

respectively. No reducing sugar was observed, since the polysaccharide was purified by dialysis previously to characterization, as
described in the item 2.2
The ␤-d-Gal and ␤-d-Glc chemical shift at 4.85 ppm and
4.84 ppm were used to estimate the amount of Gal units still


V.C. Spier et al. / Carbohydrate Polymers 170 (2017) 140–147

145

0 mol.L-1
0.015 mol.L-1
0.030 mol.L-1
0.060 mol.L-1
0.150 mol.L-1
mmol COOH/ g polymer

mmol COOH/g polymer

0.4
0.35
0.3
0.25

0.1

0.08

0.06

0.04

0.02

0
0

500

1000

1500

2000

2500

3000

Time (s)

0.2
0.15
0.1
0.05

◦

Fig. 3. Light scattering (LS) @ 90 elution profile from SEC of XG native and TEMPO
oxidized products XGoxipH9 , XGoxipH7 , XGoxipH5 and XGoxipH9,NT .

0
0

2000

4000

6000

8000

1 10

4

Time (s)
Fig. 5. mmol COOH/g of polymer as a function of time of XG oxidation at pH 9 and
TEMPO concentrations of 0, 0.015, 0.030, 0.060 and 0.150 mol L−1 , compared with
XG oxidation at pH 9 without TEMPO (XGoxipH9,NT ). The first 3000 s is inserted, for
TEMPO concentrations of 0.030, 0.060 and 0.150 mol L−1 .

Fig. 4. Kc/R␪ @ 90◦ of XG and oxidized products XGoxipH9 , XGoxipH7 , XGoxipH5 with
tempo, and XGOXIpH9,NT , versus NaNO3 concentration, by ACM.

remaining in the XG. The deconvoluted spectra, Supplementary
material (Fig. S2.2), were used to determine the total area related
to ␤-d-Gal + ␤-d-Glc of XG oxi samples. Such values were normalized to ␤-d-Gal and ␤-d-Glc in native XG. From this approach,
the amount of ␤-d-Gal in XG from 16.7% in native XG reduced to
15,6%, 15.5%, 13.4% and 12.8%, respectively, to XGoxipH5.0 , XGoxipH7.0 ,

XGoxipH9.0 and XGoxipH9.0,NT .
Comparing the amount of -COOH per gram of polysaccharide
determined by titration, and the amount of Gal reduction, a very
interesting correlation can be observed, with approximately the
same amount of oxidation observed per gram of polysaccharide
(Table 1), except to XGoxipH9.0,NT . This suggested that ␤-d-Gal units
were oxidized partially, to ␤-d-GalA in TEMPO samples, however,
other free units of Glc and non-selective oxidation sites can not be
discarded, during TEMPO reaction.

Using the HSQC spectra (13 C-1 H), supplementary material (Fig.
S2.3), were possible to identify at anomeric 1 H and 13 C region,
the ␤-d-Gal chemical shifts at 4.84/104.63 ppm and ␤-d-Glc at
4.83/103.34 ppm. At 5.42 ppm the chemical shift of the ␣-d-Xyl
substituted for ␤-d-Gal and at 5.22 ppm the ␣-d-Xyl not substituted. All these chemical shifts are compatible with the hydrogen
mono-dimensional spectra. For XG oxidized samples, it was possible to observe a new correlation at, approximately, 4.78/103.6 ppm,
attributed by Lucyszyn et al. (2009) to the chemical shift of ␤-dGalA. The chemical shift at 4.22/68.95 ppm was the C6 of ␤-d-Glc
and at 4.06/61.41 ppm the C6 of ␤-d-Glc. The 3.85/61.75 ppm and
4.01/61.73 ppm are related to H5 and C5 of xylose. The other chemical shifts are attributed to C2, C3, C4 and C5 of XG, as described by
Arruda et al. (2015).
To confirm the second effect of TEMPO, protecting the polysaccharide from degradation, different concentrations of the catalyst
were used, keeping constant the amount of the secondary oxidants NaClO/NaBrO. The concentrations used were, 0.030, 0.060
and 0.150 mol L−1 , corresponding to 2, 4 and 10 times of the initial
amount of TEMPO (0.015 mol L−1 ). All these TEMPO concentrations
were compared to experiments in the absence of TEMPO (0 mol L−1 )
Table 4 (Fig. 5)
SEC analysis of the XG oxidized products at pH 9 using different
TEMPO concentrations, are presented in Table 5 and Fig. 6.
As clearly observed on Fig. 6 an increase in the amount of
the catalyst TEMPO, reduced the degree of depolymerization.

Only increasing 10 times the initial TEMPO concentration (from
0.0150 mol L−1 to 0.150 mol L−1 ) the Mw reduction was of only
of 8%. Further increase in TEMPO concentration did not provided
any significant modification of Mw values, and can be related to
␤-elimination process.

Table 4
Rate constant of oxidation (k) and degree of oxidation measured at pH 9, using different concentration of TEMPO.
Sample

TEMPO (mol L−1 )

k/10−4 (s−1 )

Degree of oxidation (mmol g−1 )*

R

XGoxipH9,NT
XGoxipH9
XGoxipH9
XGoxipH9
XGoxipH9

0
0.015
0.030
0.060
0.150

0.58
1.62
18.8
34.2
45.7

0.395
0.215
0.092
0.097
0.095

0.997
0.998
0.976
0.998
0.998

*at 1.1 × 104 s of reaction.


146

V.C. Spier et al. / Carbohydrate Polymers 170 (2017) 140–147

Table 5
Determination of molar mass (Mw ), viscometric radius of gyration <S > Á , dispersion
(Ð = Mw /Mn ) and recovery from SEC experiments of XG and XG oxidized at pH 9 at
different concentrations of TEMPO.
AMOSTRA

TEMPO (mol L−1 )

Mw (105 g mol−1 )

XG
XGoxipH9,NT
XGoxipH9
XGoxipH9
XGoxipH9
XGoxipH9

0
0
0.015
0.030
0.060
0.150

4.8
1.3
3.1
3.6
4.1
4.4

S
39
9
12

18
26
30

Á

(nm)

Ð

Rec (%)

1.4
1.8
1.4
1.6
1.5
1.5

95
92
97
96
94
97

Acknowledgments
Support for this work was provided by Brazilian funding agencies CNPq (Conselho Nacional de Pesquisa), process
no. 477275/2012-5 and 306245/2014-0 Rede Nanobiotec/CapesBrazil, project 34 and Nanoglicobiotec-Ministry of Science and
Technology/CNPq no. 564741/2010-8 and no.555169/2005-7.

Vivian C. Spier was a beneficiary of a doctoral fellowship from
CAPES and by collaboration of the Tulane Center for Polymer Reaction Monitoring and Characterization (PolyRMC).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at />064.
References

Fig. 6. SLS @ 90◦ elution profile from SEC of XG native and TEMPO oxidized products
XGoxipH9 using TEMPO concentration of 0, 0.015, 0.030, 0.060 and 0.150 mol L−1 .

These results demonstrate that one interesting approach
to reduce the depolymerization of polysaccharides, during
TEMPO/NaClO/NaBrO oxidation is to increase the amount of primary catalyst, not affecting the ␤-elimination process elimination
process, but reducing depolymerization induced by secondary
antioxidants as NaClO and NaBrO.

4. Conclusion
In this manuscript, for the first time, TEMPO reactions, used to
selectively oxidize polysaccharide, were monitored through real
time analysis, measuring molar mass, radius of gyration, viscosity, hydrodynamic radius and optical rotation. The last two are
used here for the first time during ACOMP reactions. The optical
rotation confirms the formation of uronic acids in XG, and can
be considered an important tool to monitor real time modification of polysaccharides. Independently of the pH (9, 7 or 5), using
0.015 mol L−1 of TEMPO, some depolymerization was observed,
proving that ␤-elimination is not the only mechanism allowing
molar mass reduction. In experiments without TEMPO as a catalyst, a much more significant molar mass reduction was observed,
and at this point attributed to the secondary oxidants presented in
the medium (NaClO/NaBrO). TEMPO could protect XG from nonselective oxidations that culminate with molar mass reduction,
since depolymerization reduced with crescent amounts of TEMPO.
During XG oxidation, TEMPO acts not only as a catalyst, but also

as a “sacrificial molecule”, reacting with the secondary oxidants.
The nitrosonium ion formed, the effective catalyst, was responsible for the selective primary alcohol oxidation and its oxidation
protected the polysaccharide from non-selective degradative oxidations. Based on this manuscript, the amount of TEMPO should
be increased, to reduce the depolymerization on reactions with
polysaccharides.

Anet, E. (1964). 3-deoxyglycosuloses (3-deoxyglycosones) and the degradation of
carbohydrates. Advances in Carbohydrate Chemistry, 19, 181–218.
Arruda, I. R. S., Albuquerque, P. B. S., Santos, G. R. C., Silva, A. G., Mourão, P. A. S.,
Correia, M. T. S., et al. (2015). Structure and rheological properties of a
xyloglucan extracted from Hymenaea courbaril var. courbaril seeds.
International Journal of Biological Macromolecules, 73, 31–38.
Bayly, E. E., Brousseau, J.-L., & Reed, W. F. (2002). Continuous monitoring of the
effect of changing solvent conditions on polyelectrolyte conformations and
interactions. International Journal of Polymer Analysis and Characterization, 7,
1–18.
Boruch, M. (1985). Transformations of potato starch during oxidation with
hypochlorite. Starch, 37, 91–98.
Bragd, P. L., Besemer, A. C., & Van Bekkum, H. (2000). Bromide-free
TEMPO-mediated oxidation of primary alcohol groups in starch and methyl
␣-D-glucopyranoside. Carbohydrate Research, 328, 355–363.
Cunha, P. L. R., Maciel, J. S., Sierakowski, M. R., Paula, R. C. M. D., & Feitosa, J. P. A.
(2007). Oxidation of cashew tree gum exudate polysaccharide with TEMPO
reagent. Journal of the Brazilian Chemical Society, 18, 85–92.
de Freitas, R. A., Martin, S., Paula, R. C., Feitosa, J. P. A., & Sierakowski, M. R. (2004).
Effect of the oxidation level on the thermogravimetric kinetics of an oxidized
galactoxyloglucan from Hymenaea courbaril (Jatobá) seeds. Thermochimica
Acta, 409, 41–47.
de Freitas, R. A., Drenski, M. F., Alb, A. M., & Reed, W. F. (2010). Characterization of
stability, aggregation, and equilibrium properties of modified natural

products; The case of carboxymethylated chitosans. Materials Science and
Engineering: C, 30, 34–41.
de Nooy, A. E. J., Besemer, A. C., & Van Bekkum, H. (1995a). Highly selective nitroxyl
radical-mediated oxidation of primary alcohol groups in water-soluble
glucans. Carbohydrate Research, 269, 89–98.
de Nooy, A. E. J., Besemer, A. C., & Van Bekkum, H. (1995b). Selective oxidation of
primary alcohols mediated by nitroxyl radical in aqueous solution. Kinetics
and mechanism. Tetrahedron, 51, 8023–8032.
de Nooy, A. E. J., Besemer, A. C., Van Bekkum, H., Van Dijk, J. A. P. P., & Smit, J. A. M.
(1996). TEMPO-mediated oxidation of pullulan and influence of ionic strength
and linear charge density on the dimensions of the obtained polyelectrolyte
chains. Macromolecules, 29, 6541–6547.
de Souza, C. F., Lucyszyn, N., Ferraz, F. A., & Sierakowski, M. R. (2011). Oxidation
and N-alkylation at the C-6 position of galactomannan extracted from
Caesalpinia ferrea var. ferrea seeds. Macromolecular Symposia, 299-300, 66–73.
Flory, P. J., & Fox, T. G. (1951). Treatment of intrinsic viscosities. Journal of American
Chemical Society, 73, 1904–1908.
Hiraoki, R., Ono, Y., Saito, T., & Isogai, A. (2015). Molecular mass and
molecular-mass distribution of TEMPO-oxidized celluloses and
TEMPO-oxidized cellulose nanofibrils. Biomacromolecules, 16, 675–681.
Isbell, H. S., & Frush, H. L. (1943). Ring structures and mutarotations of the
modifications of D-galacturonic acid. Journal of Research of the National Bureau
of Standards, 31, 33–44.
Isogai, A., Saito, T., & Fukuzumi, H. (2011). TEMPO-oxidized cellulose nanofibers.
Nanoscale, 3, 71–85.
Kato, Y., Matsuo, R., & Isogai, A. (2003). Oxidation process of water-soluble starch
in TEMPO-mediated system. Carbohydrate Polymers, 51, 69–75.
Lucyszyn, N., Lubambo, A. F., Matos, K. F., Marvilla, I., Souza, C. F., & Sierakowski, M.
R. (2009). Specific modification of xyloglucan from Hymenaea courbaril seeds.
Materials Science and Engineering: C, 29, 552–558.

Milanovic, J., Kostic, M., Milanovic, P., & Skundric, P. (2012). Influence of
TEMPO-mediated oxidation on properties of hemp fibers. Industrial &
Engineering Chemistry Research, 51, 9750–9759.
Mkedder, I., Travelet, C., Durand-Terrasson, A., Halila, S., Dubreuil, F., & Borsali, R.
(2013). Preparation and enzymatic hydrolysis of nanoparticles made from
singe xyloglucan polysaccharide chain. Carbohydrate Polymers, 94, 934–939.


V.C. Spier et al. / Carbohydrate Polymers 170 (2017) 140–147
Reed, W. F. (2003). Handbook of size exclusion chromatography and related
techniques. In C.-S. Wu (Ed.), (2nd ed., pp. 589–622). Marcel Dekker.
Sakakibara, C. N., Sierakowski, M. R., Lucyzyn, N., & de Freitas, R. A. (2016). Tempo
mediated oxidation on galactomannan: Gal/man ratio and chain flexibility
dependence. Carbohydrate Polymers, 153, 371–378.
Shibata, I., & Isogai, A. (2003). Depolymerization of cellouronic acid during
TEMPO-mediated oxidation. Cellulose, 10, 151–158.
Shinoda, R., Saito, T., Okita, Y., & Isogai, A. (2012). Relationship between length and
degree of polymerization of TEMPO-oxidized cellulose nanofibrils.
Biomacromolecules, 13, 842–849.
Sierakowski, M. R., Milas, M., Desbrières, J., & Rinaudo, M. (2000). Specific
modifications of galactomannans. Carbohydrate Polymers, 42, 51–57.
Sierakowski, M. R., de Freitas, R. A., Fujimoto, J., & Petri, D. F. S. (2002). Adsorption
behavior of oxidized galactomannans onto amino-terminated surfaces and
their interaction with bovine serum albumin. Carbohydrate Polymers, 49,
167–175.

147

Sorci, G. A., & Reed, W. F. (2002). Electrostatically enhanced second and third virial
coefficients, viscosity, and interparticle correlations for linear polyelectrolytes.

Macromolecules, 35, 5218–5227.
Spier, V. C., Sierakowski, M. R., Ibrahim, A. T., Baum, J. C., Silveira, J. L. M., & de
Freitas, R. A. (2015). Time-dependent viscometry strudy of endoglucanase
action on xyloglucan: A real-time approach. International Journal of Biological
Macromolecules, 81, 461–466.
Watanabe, E., Tamura, N., Saito, T., Habu, N., & Isobai, A. (2014). Preparation of
completely C6-carboxylated curdlan by catalytic oxidation with
4-acetamido-TEMPO. Carbohydrate Polymers, 100, 74–79.
Xu, Q., Li, W., Cheng, Z., Yang, G., & Qin, M. (2014). TEMPO/NaBr/NaClO-mediated
surface oxidation of nanocrystalline cellulose and its microparticulate
retention system with cationic polyacrylamide. BioResources, 9, 994–1006.