Regeneration of Aqueous Periodate Solutions by Ozone Treatment: A Sustainable Approach for Dialdehyde Cellulose Production

Article abstract of DOI:10.1002/cssc.201501639

A method for easy and fast regeneration of aqueous periodate solutions from dialdehyde cellulose (DAC) production by ozone treatment is presented, along with a direct and reliable simultaneous quantification of iodate and periodate by reversed-phase HPLC. The influence of iodate and ozone concentration, solution pH, and reaction time on the regeneration efficiency was studied, as well as the reaction kinetics. Regeneration of spent periodate solutions by ozone was successfully performed in alkaline medium, which favors the formation of free ^.OH radicals, as supported by the addition of radical scavengers and quantum mechanical calculations. At pH 13 and an ozone concentration of approximately 150 mg L^-1, periodate was completely regenerated from a 100 mm solution of iodate within 1 h at room temperature. A cyclic process of cellulose oxidation and subsequent regeneration of spent periodate with 90 % efficiency has been developed. So far, commercial applications of DAC have been hampered by difficulties in reusing the costly periodate. This work overcomes this hurdle and presents a highly efficient, clean, and low-cost protocol for the preparation of DAC with integrated periodate recycling, with the possibility of scaling the process up.

Full text of DOI:10.1002/cssc.201501639

DOI: 10.1002/cssc.201501639
Full Papers
Regeneration of Aqueous Periodate Solutions by Ozone
Treatment: A Sustainable Approach for Dialdehyde
Cellulose Production
[
a]
[a]
[a]
[b]
Slavica Koprivica, Martin Siller, Takashi Hosoya, Walter Roggenstein,
[a]
[a]
Thomas Rosenau, and Antje Potthast*
À1
A method for easy and fast regeneration of aqueous periodate
solutions from dialdehyde cellulose (DAC) production by
ozone treatment is presented, along with a direct and reliable
simultaneous quantification of iodate and periodate by re-
versed-phase HPLC. The influence of iodate and ozone concen-
tration, solution pH, and reaction time on the regeneration effi-
ciency was studied, as well as the reaction kinetics. Regenera-
tion of spent periodate solutions by ozone was successfully
performed in alkaline medium, which favors the formation of
free COH radicals, as supported by the addition of radical scav-
ozone concentration of approximately 150 mgL , periodate
was completely regenerated from a 100 mm solution of iodate
within 1 h at room temperature. A cyclic process of cellulose
oxidation and subsequent regeneration of spent periodate
with 90% efficiency has been developed. So far, commercial
applications of DAC have been hampered by difficulties in re-
using the costly periodate. This work overcomes this hurdle
and presents a highly efficient, clean, and low-cost protocol for
the preparation of DAC with integrated periodate recycling,
with the possibility of scaling the process up.
engers and quantum mechanical calculations. At pH 13 and an
Introduction
Salts of periodic acid are commonly used to oxidize vicinal cis-
diols of carbohydrates in an aqueous medium through simulta-
neous cleavage of a carbon–carbon bond and generation of
two reactive aldehyde functions, which might also be present
in a masked form, for example, as a hydrate or hemiacetal.
This oxidation reaction is particularly suitable for polysaccha-
rides and has been used to oxidize starch, cellulose, and xylan,
but also other glucans, fructans, and their derivatives, provid-
ing they contain vicinal diols. Periodate oxidation of cellulose
involves selective attack on the anhydroglucose unit that splits
the C-2ÀC-3 bond, causing the formation of two aldehyde
Sodium periodate, which is typically used as the oxidizing
agent for DAC preparations, is commercially generated by elec-
trochemical oxidation of iodate. Although electrochemical re-
[
6,7]
generation of spent iodate has been described,
the high
process costs and sensitivity of the electrodes render this pro-
cess less attractive in combination with DAC production. Nev-
ertheless, iodine compounds emerging from periodate oxida-
tion reactions (such as sodium iodate or iodine) are environ-
mentally hazardous and cannot be removed by standard in-
dustrial wastewater treatments. Therefore, any industrial pro-
duction of DAC will have to rely on effective routes for
periodate recycling.
[
1]
groups and, at the same time, periodate is reduced to iodate.
The product obtained, commonly named dialdehyde cellulose
DAC), has gained much attention recently because it can be
Several alternative methods for the regeneration of spent
periodate, mainly from dialdehyde starch preparation, have
been proposed in the literature. They mainly use secondary ox-
(
used to immobilize proteins and dyes or can serve as a starting
material towards the production of biobased specialty fibers
idants, such as hypochlorite and ozone, both under alkaline
[
2–5]
[8–11]
and materials.
conditions,
or peroxosulfate under rather neutral condi-
[
12]
tions. Major drawbacks of these pathways arise from high
energy costs due to either elevated temperatures or long treat-
ment times required to reach high conversion. The input of
chlorinated organic compounds and corrosive halide addition-
ally disqualifies hypochlorite usage for periodate recycling. In
summary, the lack of suitable methods for efficient, large-scale,
and low-cost periodate recycling was the main factor that has
hampered commercial DAC production so far.
[
a] S. Koprivica, Dr. M. Siller, Dr. T. Hosoya, Prof.Dr. T. Rosenau,
Prof.Dr. A. Potthast
University of Natural Resources and Life Sciences (BOKU)
Department of Chemistry
Division of Chemistry of Renewable Resources
UFT Campus Tulln, Konrad-Lorenz-Strasse 24
3
430 Tulln (Austria)
E-mail: antje.potthast@boku.ac.at
[
b] W. Roggenstein
Recently, another primary oxidation agent that may offer
a practicable recycling solution has been examined. In this ap-
proach, an acetic acid derivative of iodobenzene was substitut-
ed for periodate, and a regeneration route with peracetic acid
Kelheim Fibres GmbH, Regensburger Straße 109
9
3309 Kelheim (Germany)
Supporting Information and the ORCID identification number(s) for the
author(s) of this article can be found under http://dx.doi.org/10.1002/
cssc.201501639.
ChemSusChem 2016, 9, 825 – 833
825
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
and hydrogen peroxide was suggested, without providing a de-
iodate is the highest oxidation state of iodine species and no
“overoxidation” would be possible.
[
13]
tailed protocol.
An analytical shortcoming of the above-described recycling
processes is that they all rely on indirect quantification of per
iodate, usually performed by a titration method. An analytical
method that provides direct monitoring of both iodate and
periodate simultaneously would be of great importance for
process optimization. To date, a variety of methods for the de-
termination of the iodate/periodate couple have been report-
As mentioned above, an attempt to use ozone for the re-
generation of spent periodate has already been reported in
[
11]
a patent application from Kesselmans, which did not result
in a patent being granted. In this patent application, periodate
was regenerated within 5 h from 140 mm solutions of iodate
at mildly alkaline pH between 8 and 10.5. However, the reac-
tion yield of periodate was not measured, but only calculated
based on the amount of ozone spent, and recycled periodate
showed poor oxidation efficiency, which implied only partial
regeneration.
[
14–19]
ed in the literature.
To perform simple and reliable analy-
sis, we used reversed-phase HPLC with direct UV detection;
this is a widely available technique in standard laboratories.
A successful, large-scale implementation of periodate oxida-
tion of cellulose requires an efficient and low-cost recycling
process with direct monitoring capability. Herein, we focus on
the optimization of the ozone-based regeneration method and
its subsequent application in the recycling of spent periodate
from DAC production. An experimental design approach was
applied to elucidate the parameters governing the regenera-
tion process. Simultaneous quantification of periodate and
iodate allowed monitoring of the reaction kinetics of both cel-
lulose oxidation and periodate regeneration. Mechanistic as-
pects of iodate oxidation with ozone are discussed and corre-
lated with results from DFT calculations. The efficiency of per
iodate regeneration is analyzed over three recycling steps, and
general considerations for method utilization are presented.
Effect of pH on periodate recycling
In the optimization steps, fresh solutions of sodium iodate
were used. All experiments were performed at room tempera-
ture with a volume of 85 mL. At first, the reaction performance
was tested over the pH range of 2–14 by using a 25 mm buf-
fered solution of iodate and moderate ozone concentration. To
ensure equal pH of all samples analyzed by HPLC, the reaction
solution was acidified with concentrated H PO to pH 2 after
3
4
ozonation, prior to appropriate dilution. Reaction yields of per
iodate were given relative to the initially measured iodate con-
centration. Figure 1 shows the pH dependence of the reaction
yields of periodate after 30 min of reaction time at an ozone
À1
concentration of 94 mgL .
Results and Discussion
Utilization of ozone for periodate regeneration
Ozone is a powerful oxidizing agent. It reacts with most organ-
ic compounds containing double or triple bonds in a formal
[
3+2]-cycloaddition reaction with follow-up processes, and it
can convert inorganic ions into oxyanions of the highest oxida-
[
20]
tion state. Ozone has been widely used in the treatment of
effluents from pulp and paper mills, textile dying, the pharma-
[21–23]
ceutical industry, and for water disinfection.
Herein, we
employed ozone as a secondary oxidant to convert iodate
formed during cellulose oxidation back into periodate
(
Scheme 1). The high oxidation power of ozone was no prob-
lem in this case. On the contrary, it was desired because per
Figure 1. Effect of pH on periodate reaction yield (c
o
=25 mm NaIO
3
,
À1
c(O
3
)=94 mgL , t=30 min, 100 mm phosphate buffer).
Neutral and acidic conditions did not cause any formation of
periodate. From pH 8 upwards, conversion was observed. No-
tably, the reaction yield steeply increased above pH 10 and
showed full conversion at pH 13 and above, which was in stark
[
11]
contrast to previous findings.
An experimental design approach was applied to study the
influence of alkaline pH, iodate concentration, ozone concen-
tration, and treatment time on periodate formation in more
detail. Regarding the influence of pH, solutions of pH 8 and 14
were used as the lower and upper levels. Figure 2a illustrates
the reaction yields of eight sets of different reaction conditions
Scheme 1. General scheme for recycling of periodate by ozone treatment in
DAC production.
ChemSusChem 2016, 9, 825 – 833
www.chemsuschem.org
826
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
centrated H PO dissolved the precipitate, but, in turn, intro-
3
4
duced an extra quantity of salts that interfered with iodate
sampling during HPLC analysis. This precipitate formation
made accurate online monitoring of the reaction kinetics im-
possible. To prevent possible side reactions and perform relia-
ble iodate/periodate analysis, we tested whether the use of
less alkali could avoid precipitate formation without a decline
in reaction yields. At the same time, applying less alkali would
be more economical for the whole recycling process.
Reaction kinetics
Having identified pH as the most important parameter that
governed the formation of periodate, we continued investigat-
ing the iodate to periodate conversion in alkaline media with
a high ozone input. We examined, in particular, the correlation
between iodate concentration and available alkali, the role of
oxygen radicals, and the effect of higher temperature in the
conversion process. For this purpose, aliquots were taken at
distinct time intervals, immediately diluted for chromatograph-
ic analysis to avoid the formation of sodium periodate precipi-
tate, and directly measured by HPLC. Furthermore, changes of
pH over time were monitored. The reaction was terminated at
4À
Figure 2. Experimental design approach: IO yields at a) pH 8 and b) pH 14.
at pH 8. The low periodate yield in slightly alkaline media
could hardly be improved by any combination of the parame-
ters examined. Compared with the experimental conditions at
the same pH previously shown in Figure 1, a minor increase in
yield from 5.1 to 11.5% was only observed when a lower
iodate concentration of 10 mm was treated with a higher
1
h or sooner, if considerable amounts of precipitate were
formed in the reaction flask, which indicated complete oxida-
tion. Blank reactions were performed by excluding one of the
components (O , NaOH, or NaIO ).
3
3
À1
ozone concentration of 151 mgL for twice as long. Strongly
alkaline solutions, however, resulted in high conversion rates in
the range of 94.6–102.8% after 60 min of ozone treatment for
all iodate and ozone concentrations (Figure 2b). These findings
were, once more, in disagreement with data previously report-
Effect of iodate concentration
The use of diluted alkali indeed delayed precipitate formation
to a certain extent, which allowed for the recording of iodate
and periodate concentration over time. Figure 3 shows the
[11]
ed in the patent application, in which pH 8.5 was claimed to
give optimal regeneration conditions, whereas the reaction
was allegedly not feasible above pH 10.5 due to ozone decom-
position.
yield of NaIO for various initial concentrations of iodate in al-
4
kaline solutions when a maximum ozone concentration of
À1
1
51 mgL was applied. At pH 13, periodate was completely
regenerated from a 10 mm solution of iodate already after
0 min. Up to 100 mm iodate could be converted within 1 h.
Multivariate analysis showed correlation coefficients of 0.91,
À0.12, 0.12, and 0.13 for the four abovementioned variables—
pH, iodate concentration, ozone concentration, and treatment
time—revealing the highest impact of pH on ozonation per-
formance. High ozone concentration at pH 14 caused 91.5%
regeneration of periodate from a 100 mm solution of iodate,
and 100% regeneration from a 10 mm solution of iodate, after
only 5 min. A decrease in ozone concentration slowed down
the conversion process to some extent, necessitating longer
reaction times to attain comparable yields. Interestingly, pro-
longed ozone treatment of a 10 mm solution of iodate resulted
in a slight decrease of periodate yield. This suggests that
severe alkaline conditions cause side reactions that initiate de-
composition of periodate. Because this reaction must be a re-
duction, it is reasonable to assume the involvement of inter-
mediate hydrogen peroxide or superoxide anion radicals as re-
ductants.
1
Interestingly, if the concentration of iodate was higher than
Another downside of working at pH 14 was the lowered sol-
ubility of NaIO , which caused periodate precipitation during
4
Figure 3. Effect of initial iodate concentration in alkaline media on periodate
À1
or after the reaction. Neutralization of the solution with con-
reaction yield (c(O
3
)=151 mgL , T=258C, unbuffered system).
ChemSusChem 2016, 9, 825 – 833
www.chemsuschem.org
827
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
the initial hydroxide concentration (i.e., 150 mm), complete
periodate regeneration was not possible even at prolonged re-
action times (data not shown).
follow two general pathways: ozone reacts by direct oxidation
as molecular ozone or by indirect reaction through the forma-
tion of secondary oxidants. The mechanism of ozone decom-
position includes a multitude of successive reactions, generat-
Similar to pH 13, complete periodate regeneration was ob-
served at pH 12 for an iodate concentration of 10 mm within
[
25,26]
ing a vast number of radicals;
all of them potent oxidants
1
0 min. However, when an iodate concentration of 50 mm was
with different oxidative power. However, only the oxidation
potential of hydroxyl radicals would be sufficiently high to
effect oxidation of iodate to periodate. The rate of decomposi-
tion is greatly enhanced at increased pH due to the catalytic
used, thus exceeding the amount of hydroxide ions in solution,
a yield of only 26% was reached in the first 10 min of the reac-
tion, after which time the concentration of periodate leveled
off. These observations indicated a strong correlation between
the starting iodate concentration and solution alkalinity.
[
27,28]
effect of hydroxide ions [Eqs. (4)–(8)]:
À
2
À þ O2
O þ OH ! HO
ð4Þ
ð5Þ
ð6Þ
ð7Þ
ð8Þ
The pH of the (unbuffered) reaction solution during ozona-
tion was followed. At an initial pH of 13 and an iodate concen-
tration of 100 mm, the pH of the solution mixture decreased
over time as the periodate concentration increased, whereas
no change of pH was observed for 10 and 50 mm iodate. In
less alkaline media, at pH 12, this initial value stayed constant
only in the case of a 10 mm solution of iodate. Using a 50 mm
solution of iodate, we observed a pH drop from 12 to 8 within
the first 5 min of the reaction, thereby stopping the regenera-
tion process. No change in pH was observed when ozone was
purged through alkaline solutions in the absence of iodate.
These results showed that the consumption of hydroxide ions
stemmed from the oxidation of iodate and might have had
a strong impact on the performance of the ozone reaction.
The change in pH during the reaction was assumed to be
caused by the presence of different periodate species in the
solution. It is known that, in aqueous media, periodate is pre-
dominantly hydrated and found in different ionic forms or
dimers of orthoperiodic acid H IO (pK =1.64, pK =8.36, pK =
3
À
À
O þ HO2 ! COH þ O C þ O
3
2
2
À
À
O þ O C ! O C þ O
3
2
3
2
À
À
O C þ H O ! COH þ OH þ O
3
2
2
O þ COH ! HO C þ O
3
2
2
Thus, in reactions that involve ozone above a certain pH, ox-
idation is an indirect process, according to radical-type reac-
[
29,30]
tion mechanisms.
The observed efficient periodate regen-
eration through ozone under alkaline conditions suggests that
reactive oxygen species, such as hydroxyl radicals, which are
even more powerful oxidants than molecular ozone itself, are
the active species involved in iodate oxidation.
To shed more light onto this issue, we performed ozonation
of 10 mm iodate at pH 13 in the presence of tBuOH as a radical
scavenger. tBuOH reacts rapidly with COH radicals and only
[
31,32]
very slowly with ozone [Eqs. (9) and (10)]:
5
6
1
2
3
[
24]
1
4.98). Under alkaline conditions, in particular, from pH 9 on-
2À
wards, the dominant periodate species is H IO . Because the
3
6
tBuOH þ COH ! tBuOC þ H O
2
ozone oxidation of iodate stoichiometrically leads to the for-
ð9Þ
8
À1 À1
À
k ¼ 5 Â 10 m
s
mation of H IO6 , the subsequent transition towards equilibri-
4
um is followed by a release of protons. Therefore, we can
assume that the decrease in pH is caused by periodate genera-
tion, as shown in Equations (1)–(3):
tBuOH þ O ! products
3
ð10Þ
À1 À1
k ¼ 0:03 m
s
I
O
À
þ
O
!
I
O
À
3
3
4
þ O2
ð1Þ
ð2Þ
A possible hydrogen-transfer reaction involving ozone and
tBuOH, according to Equation (10), is strongly unfavored com-
pared with the hydrogen-transfer reaction shown in Equa-
tion (9), which is largely exothermic because of much higher
À
À
IO4 þ 2 H O ! H IO
2
4
6
K ¼ 29
[
33,34]
H IO6 ! H IO6 þ Hþ
À
2À
C
C
stability of tBuO than that of HO .
Direct ionic oxidation of
4
3
ð3Þ
[35]
tertiary alcohol tBuOH by ozone can also be ruled out.
pK ¼ 8:36
2
The gas was dispersed through a 2 mm inner diameter (i.d.)
tube instead of a frit because small amounts of tBuOH already
induced foam formation that made the reaction difficult to
handle. Without a gas disperser and without adding a radical
scavenger, periodate regeneration was complete after 40 min
(Figure 4). After the addition of 100 mm tBuOH, the reaction
significantly slowed down to consume only 29% of iodate
within 60 min. A further increase of tBuOH completely imped-
ed iodate oxidation, which indicated that excess scavenger
promptly trapped all COH radicals. These findings strongly sup-
port the assumption that COH radicals are the active oxidant
Due to the decrease in pH during the ozonation process,
the alkalinity of the reaction solution should be carefully ad-
justed according to the concentration of iodate to be convert-
ed. Based on our study, we propose that the appropriate hy-
droxide concentration for efficient periodate regeneration
should be at best equal to the initial concentration of iodate.
Mechanistic studies
It is known that ozone is highly reactive in aqueous media,
and therefore, rather unstable. The ozonation process can
species, or are at least strongly involved in the periodate re-
generation process.
ChemSusChem 2016, 9, 825 – 833
www.chemsuschem.org
828
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
À
À
3
2
.6, 0.78, 2.2, 1.5, and 2.1 V for the COH radical, O C radical, O C
2
radical, HO C radical, and ozone, respectively, in which the po-
2
tential of the COH radical is clearly the highest, and hence, this
[
36–39]
radical is the strongest oxidant.
Temperature effect
Finally, regeneration efficiency was tested at increased temper-
atures, which have been shown to accelerate the rate of ozone
decomposition and, in turn, to be conducive to faster cellulose
[
40]
oxidation. A temperature of 558C was chosen to mimic the
conditions for DAC preparation. Only 60% of periodate was
formed at 558C in 10 min, before the concentration of per
iodate leveled off (see the Supporting Information). Compared
with complete conversion at RT within the same time, elevated
temperatures proved to be rather disadvantageous for per
iodate regeneration. This observation is explained by very poor
solubility of ozone in water at 558C and possible side reac-
Figure 4. Effect of tBuOH on the ozonation of iodate (c
o
=10 mm NaIO
3
,
À1
c(O
3
)=151 mgL , pH 13, unbuffered system).
To lend further support to this theory, computations were
carried out on the interactions between iodate and the above-
mentioned chemical species that arose from ozone decompo-
sition. Figure 5 shows potential energy curves regarding the
OÀI distances in the interactions between iodate and COH radi-
[
41,42]
tions.
cal, and iodate and ozone, calculated at the DFT(UM06-2X)
level. It became clear that, among all chemical species present
in the system, only the COH radical had an energy minimum in
Recycling of periodate from DAC preparation
To investigate whether ozone treatment could be utilized for
the recycling of periodate in DAC production, the efficiency of
periodate regeneration and its impact on cellulose oxidation
was examined over three cycles (Scheme 1).
the interaction with iodate, with a binding energy of around
À1
7
kcalmol (for details, see the Supporting Information). It is
thus likely that the COH radical reacts with iodate with no acti-
vation barrier for the initial step, whereas the reactions of the
other species exhibit significantly higher barriers. Without in-
vestigation of further elementary steps towards periodate for-
mation, these computational results corroborate our assump-
tion of an COH radical based reaction mechanism in the ozone
oxidation of solutions of iodate; this is in agreement with our
experimental results. The computational results are also rea-
sonable when considering the difference in oxidation poten-
tials of the five oxidants: their standard redox potentials are
Fresh periodate was used as an initial oxidant for the cellu-
lose pulp in the first cycle. Periodate consumption and iodate
formation during oxidation of cellulose at 558C to a degree of
oxidation (DO) of 2% is shown in Figure 6. The pH of the solu-
tion decreased over time from 4.3 to 3.3. The absence of per
iodate after 2 h was taken as indication of the reaction being
completed. A linear dependence was observed from the semi-
logarithmic plot of concentration over time, showing that the
kinetics of cellulose oxidation to a low DO followed a pseudo-
first-order rate law with respect to periodate concentration
(
see the Supporting Information). We observed the same kinet-
ic model in a recent study in which the oxidation of cellulose
to a DO of 50% was monitored at RT over a longer reaction
[
43]
time.
Figure 5. Potential energy curves in the interactions between iodate and
a) COH radical and b) ozone, as calculated at the DFT(UM06-2X) level of
theory. The 6-311+G(d), 6-311G(d), and 6-311G (d, p) basis sets were em-
ployed for O, I, and H atoms, respectively.
Figure 6. Reaction kinetics of periodate oxidation of cellulose (c
NaIO , T=558C) and observed decrease in pH over time.
o
=10 mm
4
ChemSusChem 2016, 9, 825 – 833
www.chemsuschem.org
829
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
The reaction solution of about 10 mm iodate after cellulose
oxidation was treated with ozone at pH 12 for 20 min. Recy-
cled periodate solutions were reused for the subsequent oxida-
tion of cellulose pulp over three cycles, as illustrated in
Figure 7. Thereby, the starting amount of cellulose was adjust-
ed after each recycling step to keep the DO of the final prod-
uct constant and, at the same time, preserving a pulp consis-
tency of around 7.3%. Adding wet cellulose to a recycled solu-
tion led to a slight decrease of periodate concentration (and
pulp consistency) in every oxidation step.
formed with equal stoichiometric amounts of fresh periodate.
A comparison of control reactions and those run with recycled
oxidant showed no influence of the periodate origin on
iodate–periodate interconversion (Table 1). Likewise, fluores-
cence labeling revealed comparable carbonyl group contents
of DACs in all cases. An average carbonyl content of
À1
100 mmolg showed good oxidation performance, thus con-
firming successful utilization of ozone for periodate recovery in
the process of DAC production.
General considerations
The overall efficiency of the process for three consecutive recy-
cling steps was 90, 84, and 90%, respectively. The observed
losses exclusively result from the workup procedure after oxi-
dation. In industrial applications, the loss of iodate in the wet,
solid product could be minimized, if the solutions from subse-
quent washing steps of DAC fibers were collected and residual
iodate was concentrated, for example, by vacuum evaporation.
This step was not considered necessary for the proof of con-
cept in our approach. The decrease of periodate concentration
and pulp consistency, emerging from the addition of the wet
cellulose pulp at the start of oxidation, could be compensated
for by the addition of fresh periodate to a recycled solution.
Sodium hydroxide and sulfurous/sulfuric acid, which may be
used for pH adjustment in the recycling process, are already
readily used in pulp and fiber production. The drop in pH ob-
served during ozonation proved to be beneficial for the recy-
cling process because it decreased the amount of acid re-
quired for the next oxidation step. Due to the limiting gas–
liquid mass transfer, an excess of ozone has to be applied in
the range that is dependent on the particular setup in use.
Ozonation time cannot be reduced through the use of in-
creased amounts of alkali due to the formation of periodate
precipitate, which can be dissolved only below pH 2; this, in
turn, causes additional consumption of acid and dilution of the
solution.
Figure 7. Periodate regeneration by ozone treatment over three recycling
steps.
In all experiments, the final concentration of iodate after oxi-
dation corresponded to the initial (diluted) concentration of
periodate within the error of measurement. However, about
1
0–15% of iodate remained trapped in the wet DAC fibers
after they were pressed out at the end of the oxidation. This
À
amount of iodate, corresponding to “fresh/recycled IO ”
4
À
minus “IO ” given in Table 1, was accounted for as oxidation
3
loss.
The efficiency of three consecutive regeneration steps, taken
À
À
as the ratio of “recycled IO ” and “IO ” amounts, was found
4
3
to be 107, 101, and 109%. These somewhat higher values are
due to small amounts of unconsumed periodate that adds to
the regenerated periodate. Clearly, ozone treatment resulted in
full periodate recovery throughout all recycling steps.
All recycling methods so far utilize a higher pH for the per
iodate regeneration than that for the oxidation reaction. This is
necessary due to the pH dependency of the redox potential of
the periodate/iodate couple: with 1.6 V at a low pH versus
0.7 V at high pH, the oxidation of iodate is facilitated under al-
kaline conditions. Therefore, accumulation of salt cannot be
avoided in any recycling scenario. Other inorganic secondary
oxidants would end up as salts in the regeneration process
Carbonyl group labeling of all DACs prepared was carried
out to see if recycled periodate might have any hitherto un-
known effects on the produced DAC. Parallel to cellulose oxi-
dation with recycled periodate, control reactions were per-
4À
Table 1. Effect of ozone treatment on cellulose oxidation performance. Controls 1 and 2 represent reactions with fresh solutions of IO run in parallel to
[
a]
those with recycled periodate.
Start
cellulose
c(C=O)
Cycle I
Cycle II
Cycle III
[
b]
[c]
[b]
[c]
[b]
[c]
ox.
n(IO
I
À
DAC
c(C=O)
[mmolg
reg.
I
)
ox. II
DAC
c(C=O)
[mmolg
reg. II
ox. III
DAC
c(C=O)
[mmolg
reg. III
4À
3
4À
3À
4À
3À
4À
Sample
n(IO
[mmol]
)
)
n(IO
n(IO
)
n(IO
)
n(IO
)
n(IO
)
À1
À1
À1
À1
[
mmolg
]
[mmol]
]
[mmol]
[mmol]
]
[mmol]
[mmol]
]
[mmol]
[
a]
Recycling
Control 1
Control 2
26
1.23
1.04
98
1.11
1.11
0.92
0.93
103
109
0.93
0.93
0.93
0.77
92
0.84
[a]
[a]
0.82
103
0.83
4À
[
a] Fresh IO . [b] Oxidation. [c] Regeneration .
ChemSusChem 2016, 9, 825 – 833
www.chemsuschem.org
830
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
themselves. Thus, with oxygen as the only byproduct, ozone
keeps salt generation at a minimum. Nevertheless, a small
amount of salt has been shown to improve the oxidation effi-
ciency by disrupting the hydrogen-bond network of cellulose,
and thereby, making some of the hydroxyl groups better ac-
cessible to periodate. For example, Sirvio et al. performed salt-
Spent periodate from DAC preparation was completely re-
generated by ozone treatment over three consecutive recy-
cling steps. The observed consumption of hydroxyl ions during
ozonation could be an advantage in the recycling process be-
cause it decreased the amount of acid necessary to return to
the acidic conditions desired for cellulose oxidation. Minor
losses of iodate in the recycling process result from the
workup procedure after the oxidation step, which can be over-
come by collection of the DAC washing solutions. Recycled
periodate maintained good oxidation performance throughout
subsequent DAC preparations, and showed no differences
from fresh periodate. The presented recycling process for
periodate offers an environmentally friendly, economically fea-
sible, and sustainable pathway for the production of oxidative-
ly functionalized celluloses and other polysaccharides.
assisted oxidation with an amount of NaCl that corresponded
À
to the concentration of salt which would accumulate from IO
4
[44]
recycling after 17 cycles. This indicates that the presence of
salt could be tolerated to a certain extent, after which per
iodate would be precipitated and recovered as a salt, given
[
6]
a sufficient amount of alkali, or purified by other means.
Another particular benefit of using ozone in alkaline
medium is the removal of organic impurities, such as low-mo-
lecular-weight acids and hydroxyacids from cellulose or poly-
saccharide degradation, by oxidation. These impurities are
completely oxidized, that is, converted into CO2 and water,
under the strongly oxidizing conditions. They would thus not
further contribute to the total organic carbon (TOC) content in
the effluents and also could not undergo any side reactions,
such as condensation to chromophoric substrates, with cellu-
lose or DAC.
Experimental Section
Materials
All chemicals used for optimization of the regeneration process
were obtained as p.a. grade from Sigma–Aldrich (Schnelldorf, Ger-
many) and were used as received. Acetonitrile and phosphoric acid
Regardless of the change in pH, recycling of periodate evi-
dently has to be carried out separately from cellulose oxidation
to prevent side reactions of the polymer with the secondary
oxidation agent, for example, cellulose degradation or further
oxidation of aldehydes to carboxyl groups. Recycling of per
iodate with ozone also allows the preparation of DAC of vari-
ous degrees of oxidation by using higher periodate concentra-
(
85%) were of the highest purity available and obtained from
Merck (Vienna, Austria) and Roth (Graz, Austria), respectively.
Fully bleached sulfite dissolving pulp (beech) from Lenzing AG
(
weight-average molar mass of the pulp was 290 kgmol with
a starting carbonyl content of 24.3 mmolkg . Full characterization
Lenzing, Austria) was used as the cellulosic test substrate. The
À1
À1
[45]
of this pulp was previously reported by our group.
[
43]
tions. In our study, we evaluated the regeneration of per
iodate up to 100 mm solutions of iodate. If desired, even satu-
rated solutions of iodate (e.g., 400 mm) can be recycled by
ozone treatment, providing the pH is kept high enough, if only
in the form of precipitated salt.
Preparation of DAC
Air-dried cellulose sheets (10 g) were torn into small pieces of 1–
2
2
cm and disintegrated in water by using a kitchen blender. After
The proposed ozone-based regeneration method should be
suitable for recycling of spent periodate not only from oxida-
tion of cellulose, but also from oxidation of starch and other
glucans. If the oxidation product was soluble, for example,
fully oxidized DAC or low-molecular-weight carbohydrates, the
precipitation of iodate prior to regeneration could be achieved
removing excess water by filtration with a Büchner funnel, the
water content of the cellulosic pulp was approximately 50%wt.
The wet cellulose was immersed in a solution of sodium periodate
(
0.26 g) to a total volume of 120 mL and was oxidized at 558C to
[43]
a DO of 2%, according to a method reported previously. Ali-
quots were taken every 10 min to monitor both periodate con-
sumption and iodate formation, by measuring their absorbance at
l=290 nm after ion separation on a reversed-phase column. When
all periodate was consumed, the reaction solution was filtered off,
and the pulp was thoroughly washed with deionized water to
remove residual iodate. Pulp samples were frozen without drying
and stored for further analysis by gel permeation chromatography
[9]
by the addition of metal ions or organic solvents.
Conclusions
This study demonstrates efficient periodate recycling by ozone
treatment with a great potential for large-scale applications in
DAC production. Alkaline conditions that promote the forma-
tion of oxygen radicals and a high ozone concentration at
room temperature proved to be most beneficial for high con-
version efficiency. Direct monitoring of reaction kinetics
showed that 100% yields were attained when the concentra-
tion of alkali was at least equal to the initial iodate concentra-
tion to be converted, or higher. Insight into the mechanism of
ozone oxidation, by means of hydroxyl radical scavengers and
DFT calculations of potential energies, assisted better under-
standing of the recycling process and confirmed that hydroxyl
radicals were centrally involved in the oxidation system.
(
GPC).
Recycling of periodate
After cellulose oxidation, the reaction solution containing about
1
0 mm iodate was allowed to cool and the pH was adjusted to 12
with 4.5m NaOH. An amount of about 100 mL was treated with
ozone for 20 min at room temperature. Finally, 12m HCl was used
to acidify the regenerated solution of periodate to pH 4.5, which
was required for the next oxidation cycle (Scheme 1). Recycled
periodate was reused to repeatedly oxidize cellulose to DAC, as de-
scribed above. Control reactions with fresh solutions of periodate
were run in parallel. The starting amount of cellulose was adjusted
ChemSusChem 2016, 9, 825 – 833
www.chemsuschem.org
831
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[
17]
to keep the DO of the final product constant. Dilution of recycled
periodate caused by the water content of the cellulosic pulp was
accounted for in each oxidation step.
lined by Sajonz et al. with a few modifications. Optimum separa-
tion was achieved with a mobile phase consisting of 0.15% aque-
ous H PO (ꢀpH 2) and acetonitrile (90/10, v/v), and isocratic elu-
3
4
À1
tion at a flow rate of 1 mLmin . The column temperature was
kept at 258C. The injection volume was 5 mL. All samples for HPLC
Ozone treatment of iodate
analysis were prepared with 0.15% H PO as the diluent and fil-
3 4
tered through a 0.45 mm filter prior to injection. The system was
controlled by ChemStation software Rev B.04.03 SP1.
An oxygen-based ozone generator (Anseros COM-AD-04, Tübingen,
Germany) was used. Ozone concentration in the gas phase was in
the range from 0.1–151.3 mgL and was controlled by the genera-
À1
The iodate/periodate quantification method was analytically vali-
dated (see the Supporting Information). The limit of detection and
limit of quantification for iodate were 0.02 and 0.06 mm, respec-
tively, and 0.03 and 0.09 mm, respectively, for periodate. The aver-
age relative standard deviations were 1.6 and 1.9% for iodate and
periodate, respectively.
tor power level, according to product specifications from the man-
[46]
ufacturer, which we verified by an iodometric titration method.
The ozone treatment was performed in a 100 mL glass bottle with
a three-way dispenser (ozone gas inlet and outlet, sampling inlet).
The ozone generator was equilibrated under oxygen flow for 1 h
before the ozone gas stream was switched on to indicate the start
of the reaction.
Acknowledgements
The ozone-containing gas was purged at a constant flow of
À1
5
0 Lh through a polytetrafluoroethylene (PTFE) frit (20”9 mm)
with a pore size of 10 mm to ensure a high number of small bub-
bles, that is, a large reactive surface area. Furthermore, the reaction
solution was vigorously mixed with a magnetic stirrer to enhance
mass transfer into the liquid phase.
This work was supported by Kelheim Fibres GmbH, Germany and
the Christian Doppler Research Society, Austria. We thank Prof.
Paul Kosma, BOKU University, for providing the ozone generator.
Dr. Thomas Zweckmair is gratefully acknowledged for his assis-
tance in analytical method development. The computational re-
sults presented have been achieved by using the Vienna Scientific
Cluster (VSC).
Design of experiments
Four variables were explored to probe the conversion efficiency:
pH, iodate concentration, ozone concentration, and reaction time.
Thus, a two-level full factorial design of given variables was ap-
Keywords: biomass · iodine · ozone · radicals · reaction
mechanisms
2
plied, that is, 4 experiments were carried out in total to determine
the effect of the parameters selected and their respective interac-
tions. The lower and higher levels of the reaction variables are
given in Table 2. The response variable evaluated was the reaction
yield of periodate. The software JMP 12.0.1 was used for multivari-
ate analysis.
[
[
1] A. S. Perlin in Advances in Carbohydrate Chemistry and Biochemistry,
Vol. 60 (Ed.: D. Horton), Academic Press, San Diego, 2006, pp. 183–250.
2] T. Nikolic, M. Kostic, J. Praskalo, B. Pejic, Z. Petronijevic, P. Skundric, Car-
bohydr. Polym. 2010, 82, 976–981.
[
[
3] D. M. Lewis, K. N. Tapley, Textile Chem. Colorist 1999, 31, 20–26.
4] B. Sun, Q. Hou, Z. Liu, Y. Ni, Cellulose 2015, 22, 1135–1146.
Table 2. Reaction variables used in the experimental design.
[5] J. A. Sirviç, H. Liimatainen, J. Niinimäki, O. Hormi, RSC Adv. 2013, 3,
6590–16596.
[6] V. F. Pfeifer, V. E. Sohns, H. F. Conway, E. B. Lancaster, S. Dabic, E. L. Grif-
1
3À
3
Level
pH
c(IO
)
c(O
[mgL
)
t
À1
À1
[
mmolL
]
]
[min]
fin, Ind. Eng. Chem. 1960, 52, 201–206.
[
7] L. J. J. Janssen, M. H. A. Blijlevens, Electrochim. Acta 2003, 48, 3959–
lower
upper
8
14
10
100
23
151
5
60
3964.
[
[
8] T. A. McGuire, C. L. Mehltretter, Starch/Staerke 1971, 23, 42–45.
9] A. C. Besemer, J. M. Jetten, EP1189834 B1, 2004.
[10] H. Liimatainen, J. Sirviç, H. Pajari, O. Hormi, J. Niinimäki, J. Wood Chem.
GPC analysis
Technol. 2013, 33, 258–266.
[
[
[
11] R. P. W. Kesselmans, (Avebe), WO1998027118 A1, 1998.
12] A. C. Besemer, (SCA Hygiene Prod. Zeist ), EP1341717 B1, 2004.
13] K. A. J. Huvaere, J. B. F. G. Joos, K. J. A. Van Aken, P. E. M. J. Willems, (Eco-
synth, Orineo), PCT WO2013030264, 2013.
Carbonyl group profiling of DACs produced with fresh and regen-
erated solutions of periodate, as well as of the nonoxidized refer-
ence sample, was achieved by selective labeling of the aldehyde
groups with the marker carbazole-9-carboxyloxyamine followed by
size exclusion chromatography multiangle laser light scattering
[14] A. A. Ensafi, M. A. Chamjangali, Spectrochim. Acta Part A 2003, 59,
2897–2903.
[
15] Y. Ni, Y. Wang, Microchem. J. 2007, 86, 216–226.
(
SEC-MALLS) fluorescence analysis (CCOA method), as described
[47]
[16] S. Honda, K. Suzuki, K. Kakehi, Anal. Biochem. 1989, 177, 62–66.
[
[
previously.
17] P. Sajonz, J. Bookalam, R. A. Miller, Chromatographia 2006, 64, 635–640.
18] H. Lee, H. Y. Yoo, J. Choi, I. H. Nam, S. Lee, S. Lee, J. H. Kim, C. Lee, J.
Lee, Environ. Sci. Technol. 2014, 48, 8086–8093.
19] M. Gupta, A. Jain, K. K. Verma, Talanta 2007, 71, 1039–1046.
20] J. HoignØ, H. Bader, Water Res. 1983, 17, 173–183.
HPLC analysis
[
[
[
[
An Agilent 1100 Series HPLC system (Agilent Technologies, Wald-
bronn, Germany) was used. Samples were separated on a J’sphere
ODS-M80 reversed-phase column (250”4.6 mm, 4 mm particle size)
with a precolumn, both from YMC (Dinslaken, Germany). Identifica-
tion and quantification of iodate and periodate were performed by
direct UV detection at l=220 nm, according to the method out-
21] L. Bijan, M. Mohseni, Water Res. 2005, 39, 3763–3772.
22] M. B. Roncero, A. L. Torres, J. F. Colom, T. Vidal, Bioresour. Technol. 2003,
8
7, 305–314.
[
23] P. R. Gogate, A. B. Pandit, Adv. Environ. Res. 2004, 8, 501–551.
[24] A. F. Holleman, E. Wiberg, N. Wiberg, Lehrbuch der Anorganischen
Chemie, Walter de Gruyter, Berlin, 1985, pp. 432–434.
ChemSusChem 2016, 9, 825 – 833
www.chemsuschem.org
832
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[
25] B. Kasprzyk-Hordern, M. Ziółek, J. Nawrocki, Appl. Catal. B 2003, 46,
39–669.
[39] A. S. Dutton, J. M. Fukuto, K. N. Houk, Inorg. Chem. 2005, 44, 4024–
4028.
6
[
[
[
[
[
26] U. von Gunten, Water Res. 2003, 37, 1443–1467.
27] J. HoignØ, H. Bader, Water Res. 1976, 10, 377–386.
28] J. Staehelin, J. Hoigne, Environ. Sci. Technol. 1982, 16, 676–681.
29] L. Forni, D. Bahnemann, E. J. Hart, J. Phys. Chem. 1982, 86, 255–259.
30] K. Sehested, J. Holcman, E. Bjergbakke, E. J. Hart, J. Phys. Chem. 1984,
[40] B. G. Ershov, P. A. Morozov, Russ. J. Phys. Chem. A 2009, 83, 1295–1299.
[41] J. A. Roth, D. E. Sullivan, Ind. Eng. Chem. Fundam. 1981, 20, 137–140.
[42] H. L. Clever, R. Battino, H. Miyamoto, Y. Yampolski, C. L. Young, J. Phys.
Chem. Ref. Data 2014, 43, 033102-193–033102-206.
[43] M. Siller, H. Amer, M. Bacher, W. Roggenstein, T. Rosenau, A. Potthast,
Cellulose 2015, 22, 2245–2261.
[44] J. Sirvio, U. Hyvakko, H. Liimatainen, J. Niinimaki, O. Hormi, Carbohydr.
Polym. 2011, 83, 1293–1297.
[45] U. Henniges, M. Kostic, A. Borgards, T. Rosenau, A. Potthast, Biomacro-
molecules 2011, 12, 871–879.
8
8, 269–273.
31] Q. Z. Dai, J. Y. Wang, L. L. Chen, J. M. Chen, Ind. Eng. Chem. Res. 2014,
3, 11593–11600.
[
[
5
32] I. Stranic, G. A. Pang, R. K. Hanson, D. M. Golden, C. T. Bowman, J. Phys.
Chem. A 2013, 117, 4777–4784.
[33] Y. Li, J. Wang, H. Yao, L. Dang, Z. Li, J. Mol. Catal. A 2011, 334, 116–122.
[34] L. Gu, X. Zhang, L. Lei, Ind. Eng. Chem. Res. 2008, 47, 6809–6815.
[35] N. Y. Shafikov, A. A. Gusmanov, Yu. S. Zimin, V. D. Komissarov, Kinet.
Catal. 2002, 43, 799–801.
[46] K. Rakness, G. Gordon, B. Langlais, W. Masschelein, N. Matsumoto, Y. Ri-
chard, C. M. Robson, I. Somiya, Ozone Sci. Eng. 1996, 18, 209–229.
[47] A. Potthast, J. Rçhrling, T. Rosenau, A. Borgards, H. Sixta, P. Kosma, Bio-
macromolecules 2003, 4, 743–749.
[
36] K. Honda, Y. Yamaguchi, Y. Yamanaka, M. Yoshimatsu, Y. Fukuda, A. Fu-
jishima, Electrochim. Acta 2005, 51, 588–597.
[
[
37] P. S. Rao, E. Hayon, J. Phys. Chem. 1975, 79, 397–402.
38] W. H. Koppenol, FEBS Lett. 1982, 140, 169–172.
Received: December 13, 2015
Published online on March 15, 2016
ChemSusChem 2016, 9, 825 – 833
www.chemsuschem.org
833
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim