Biomacromolecules 2004, 5, 1671-1677 1671

Articles
Effect of Sulfate Groups from Sulfuric Acid Hydrolysis on the
Thermal Degradation Behavior of Bacterial Cellulose
Maren Roman† and William T. Winter*,‡
Department of Chemistry, Pulp and Paper Research Centre, McGill University,
Montréal, Quebec H3A 2A7, Canada, and Cellulose Research Institute and Department of Chemistry,
SUNY College of Environmental Science and Forestry, Syracuse, New York 13210
Received December 10, 2003; Revised Manuscript Received April 16, 2004

When used as fillers in polymer composites, the thermostability of cellulose crystals is important. Sulfate
groups, introduced during hydrolysis with sulfuric acid, are suspected to diminish the thermostability. To
elucidate the relationship between the hydrolysis conditions, the number of sulfate groups introduced, and
the thermal degradation behavior of cellulose crystals, bacterial cellulose was hydrolyzed with sulfuric acid
Downloaded by VIRGINIA TECH on October 27, 2009 | http://pubs.acs.org

under different hydrolysis conditions. The number of sulfate groups in the crystals was determined by
Publication Date (Web): June 29, 2004 | doi: 10.1021/bm034519+

potentiometric titration. The thermal degradation behavior was investigated by thermogravimetric analysis.
The sulfate group content increased with acid concentration, acid-to-cellulose ratio, and hydrolysis time.
Even at low levels, the sulfate groups caused a significant decrease in degradation temperatures and an
increase in char fraction confirming that the sulfate groups act as flame retardants. Profile analysis of the
derivative thermogravimetric curves indicated thermal separation of the degradation reactions by the sulfate
groups into low- and high-temperature processes. The Broido method was used to determine activation
energies for the degradation processes. The activation energies were lower at larger amounts of sulfate
groups suggesting a catalytic effect on the degradation reactions. For high thermostability in the crystals,
low acid concentrations, small acid-to-cellulose ratios, and short hydrolysis times should be used.

Introduction decrease the yield of levoglucosan,14 and increase the yield

of carbon.15 Consequently, it might be expected that the
Sulfuric acid hydrolysis of native cellulose fibers causes sulfate groups in the crystals prepared by sulfuric acid
breakdown of the fibers into rodlike fragments. These highly hydrolysis diminish the thermostability of the crystals, in
crystalline cellulose needles form stable aqueous suspensions which case hydrochloric acid might be a better choice for
due to sulfate groups, which have been introduced during the hydrolysis.
the hydrolysis through esterification of surface hydroxyl
The present study was conducted to elucidate the relation-
groups.1 The colloidal suspensions of cellulose crystals first
ship between the hydrolysis conditions, the amount of sulfate
gained attention as model systems for rigid rods2 and have
groups introduced, and the thermal degradation behavior of
then been extensively studied for their ability to form ordered,
cellulose crystals prepared by sulfuric acid hydrolysis. The
birefringent phases when the crystal content in the suspen-
aim was to optimize the hydrolysis conditions for the
sions exceeds a critical value.3-8 In recent years, the
application of the crystals as reinforcements in polymer
mechanical properties of cellulose crystals have attracted
nanocomposites. Bacterial cellulose was hydrolyzed under
attention for application as reinforcing fillers in polymer
different conditions and then characterized by potentiometric
nanocomposites.9-11 With typical processing temperatures
titration and thermogravimetric analysis. Even small amounts
for thermoplastics often exceeding 200 °C, the thermosta-
of sulfate groups caused a significant decrease in degradation
bility of the crystals is important for this application.
temperatures. Larger amounts resulted in a stepwise degrada-
Inorganic salts and acids are known to act as flame
tion behavior.
retardants in the pyrolysis and combustion of cellulose; that
is, they increase the char yields at the expense of flammable
tars by promoting the dehydration reactions.12,13 In ac- Experimental Section
cordance, impregnation of cellulose with sulfuric acid has
been shown to lower the onset of thermal degradation,14-16 Materials. Bacterial cellulose (PrimaCel) was provided
by the Nutrasweet Kelco Company, Chicago, IL (Now
* To whom correspondence should be addressed. E-mail: wtwinter@ CPKelco, Wilmington, DE). Sodium hydroxide for the
syr.edu. Telephone: 315-470-6876. Fax: 315-470-6856.
† McGill University. cellulose purification was purchased from MCB, acetic acid
‡ SUNY College of Environmental Science and Forestry. from EM Science, and sulfuric acid from Fisher Scientific.
10.1021/bm034519+ CCC: $27.50 © 2004 American Chemical Society
Published on Web 06/29/2004
 1672 Biomacromolecules, Vol. 5, No. 5, 2004 Roman and Winter

Table 1. Amounts of Introduced Sulfate Groups, nSO3H, and Surface Charge Densities, σ, at Different Reaction Conditions for the Sulfuric
Acid Hydrolysis of Bacterial Cellulose
hydrolysis conditionsa
sample (%(w/v)) (mmol‚g-1) (°C) (h) nSO3H ( 0.8 (mmol‚kg-1) σ ( 0.003 (e‚nm-2)b
A unhydrolyzed 0 0.000
B 12 188 104 2 2.1 0.007
C 65 66 40 1 6.5 0.021
D 65 199 40 1 8.5 0.027
E 65 663 40 1 9.8 0.031
F 65 66 40 3 10.3 0.033
G 61 618 40 3 50.8 0.162
H 65 66 60 2 73.0 0.233
a Sulfuric acid concentration, acid-to-cellulose ratio, hydrolysis temperature, hydrolysis time. b We calculated a specific surface area of 189 m2 g-1

using an assumed average rectangular-solid particle with dimensions 8 × 40 × 1000 nm.

Sodium hydroxide for the titration was purchased from

Mallinckrodt and the hydrochloric acid standard solution
(0.2015 N) from J. T. Baker. All chemicals were reagent
grade or higher in purity and used as received. The water
for the titrations was of ASTM type I.
Sulfuric Acid Hydrolysis. Bacterial cellulose was purified
Downloaded by VIRGINIA TECH on October 27, 2009 | http://pubs.acs.org

by dispersing 10 g (dry weight basis) for 6-7 days in 1 L

Publication Date (Web): June 29, 2004 | doi: 10.1021/bm034519+

0.5% aqueous NaOH solution and repeatedly washed with

water and filtered until the filtrate pH reached neutrality.
Then the cellulose was redispersed for 1-4 days in 1 L 0.5%
acetic acid, and the washing/filtration process was repeated.17
The cellulose was freeze-dried from aqueous suspension prior
to hydrolysis. The purified cellulose was placed in sulfuric
acid of known concentration and acid-to-cellulose ratio at a
constant temperature for a period of time. The applied
hydrolysis conditions are listed in Table 1. After completion
of the hydrolysis, the reaction was quenched by addition of
ice-cold water except for the 12% (w/v) sulfuric acid for
which the reaction mixture was cooled externally with an
ice bath. The crystals were collected in a filter and washed
with water by the washing/filtration process described above.
Sample F was washed by successive centrifugation until
peptization occurred prior to filtration. Centrifugation was
not used for the other samples because it was more tedious
than filtration. Finally, the crystals were freeze-dried from
aqueous suspension.
Potentiometric Titration. A fine aqueous suspension
(0.1%) of the sample was prepared by ultrasound treatment
in a plastic jar, to avoid release of ions from glass-jar walls,
and under ice-bath cooling, to avoid desulfation upon heating.
The suspension (50 g) was weighed into a 100-mL three-
neck flask and 1 mL of 0.05 M aqueous NaCl solution was
added. An ORION combination electrode (ROSS) and a
thermometer were inserted through the side-necks. Before
each measurement, the electrode was calibrated with both
pH 4 and 7 buffers. The suspension was held at 25 °C and
stirred under argon gas flow, to minimize carbon dioxide
interference, until the reading of the pH meter (Corning,
model 440) was stable. Then aqueous NaOH solution (0.0011
N, determined by titration with a hydrochloric acid standard
solution) was added dropwise from a microburet and the pH
recorded after each drop.
An example of the resulting titration curves is shown in
Figure 1a. The equivalence points in the titrations were
determined using Gran’s Plot procedure.18 Figure 1b shows
the Gran plot for the titration curve in Figure 1a.
Figure 1. Determination of the amount of sulfate groups in sulfuric
acid hydrolyzed bacterial cellulose by potentiometric titration (sample
H): (a) titration curve, (b) Gran plot; V0: initial volume of suspension;
V: volume of NaOH added; k and k′: arbitrary constants.
The equivalence-point volume is given by the abscissa
intercepts of the linear regression lines. Ideally, the lines for
the two branches of the titration curve intercept the volume
axis at the same value so that either branch can be used for
the determination of the equivalence-point volume. In our
titrations, the abscissa intercepts, and thus the equivalence-
point volumes for the left and right branch differed by up to
0.82 mL. We attributed this difference to carboxylic acid
groups in the samples, which start to dissociate once the
stronger sulfate groups have been neutralized. The actual
equivalence-point volume was taken as the average of the
two volumes determined from both branches of the titration
curve. Titration of the unhydrolyzed bacterial cellulose gave
an equivalence-point volume of 0.27 mL translating into a
carboxyl group content of 6.0 mmol kg-1. The sulfate content
in the hydrolyzed samples was calculated from the additional
base required for neutralization.
Thermogravimetric Analysis. Thermogravimetric (TG)
curves were recorded with a TA Instruments model 2950
Hi-Res TGA. The sample (20 mg) was heated to 500 °C in
 Hydrolysis of Bacterial Cellulose
Figure 2. Deconvolution of the DTG curve between 150 and 350 °C
of sulfuric acid hydrolyzed bacterial cellulose (sample H) using split
Pearson VII line shapes.
an air current of 100 cm3 min-1. The heating rate was 10 °C
min-1 up to 110 °C and 2 °C min-1 from 110 to 500 °C.
Derivative TG curves (DTG) express the weight-loss rate
as a function of temperature. Profile analysis of DTG curves
was done with a peak fitting program using asymmetric split
Pearson VII line shapes.19 A deconvolution example is shown
Downloaded by VIRGINIA TECH on October 27, 2009 | http://pubs.acs.org
in Figure 2.
Publication Date (Web): June 29, 2004 | doi: 10.1021/bm034519+
The activation energies, Ea, of the degradation processes
were determined with the Broido method using the equation20
Ea
( 1y) ) - R T1 + C
ln ln
where R was the gas constant, C was a temperature
independent term, and y was the fraction of undecomposed
material remaining at time t, (W - Wf)/(Wi - Wf). Wi and
Wf are the initial and final sample weight, and W is the
sample weight at time t. The activation energy is obtained
from the slope of a plot of ln(ln 1/y) versus 1/T.
Transmission Electron Microscopy (TEM). One drop
(8 µL) of a 0.001% aqueous suspension of crystals was
allowed to dry on a Formvar and carbon coated grid (200
mesh, treated with poly-L-lysine solution). The crystals were
stained with uranyl acetate. Micrographs were taken at 80
kV with a JEOL JEM-2000 EX electron microscope.
X-ray Diffraction. X-ray diffraction was carried out with
a Rigaku DMAX-1000 diffractometer and fiber attachment
using Ni-filtered Cu KR radiation, generated with a rotating
anode generator operating at 50 kV and 120 mA. The
diffraction angle was calibrated with sodium fluoride. The
crystals were loosely inserted into the sample holder. Before
each measurement, the random orientation of the crystals in
the sample holder was verified by rotation of the sample
about the transmitted beam direction and also from rotation
in θ while recording intensities at fixed 2θ values corre-
sponding to the principal reflections in a native cellulose
pattern (2θ ) 14.4, 16.6, 22.5°). To determine the sample
crystallinity, profile analysis was carried out with a peak
fitting program using Gaussian line shapes.19 The deconvo-
lution included the interference maxima at 2θ ) 14.4, 16.6,
20.3, and 22.5° and a broad peak at the interference minimum
at 2θ ) 18.5° accounting for the amorphous scattering. The
crystallinity was taken as the ratio of the sum of areas under
the crystalline diffraction peaks to the total area under the
curve between 2θ ) 10 and 27°.
Biomacromolecules, Vol. 5, No. 5, 2004 1673
FTIR Spectroscopy. FTIR spectra (500 scans, data
spacing 1.928 cm-1) were recorded with a Nicolet Impact
400 FTIR spectrometer. KBr pellets contained 2% of sample
and were dried under vacuum in a desiccator for 2 days.
Results and Discussion
Sulfate Content and Crystal Aggregation. The sulfate
contents of the samples, after hydrolysis, were determined
by potentiometric titration, and the values are listed in Table
1.
Sample B, hydrolyzed under the conditions used in ref 1,
contained a very small number of sulfate groups. The amount
was much smaller than, for example, in sample D, which
was hydrolyzed at roughly the same acid-to-cellulose ratio
but with a higher acid concentration. At low acid concentra-
tions, water is present in large amounts. Because water is a
reaction product in esterifications, an excess of it pushes the
equilibrium of the reversible reaction to favor the reactants.
Thus, higher acid concentrations lead to larger amounts of
sulfate groups in the crystals. The amounts of sulfate groups
were also larger for larger acid-to-cellulose ratios (samples
C-E and samples F and G) and longer hydrolysis times
(samples C and F and samples E and G). The effect of the
hydrolysis time was more pronounced at a higher acid-to-
cellulose ratio. Hydrolysis at 60 °C (sample H) gave higher
sulfation than hydrolysis at 40 °C (sample F).
The sulfate content of sample H was 73 mmol kg-1. Araki
and Kuga hydrolyzed bacterial cellulose with 65% sulfuric
acid (by weight) at 70 °C for 30 min and measured a sulfate
content of 5 mmol kg-1.21 The acid-to-cellulose ratio was
not specified but was 103 mmol g-1 in the author’s reference
to previously published experimental details. Considering that
the acid concentration and probably the acid-to-cellulose ratio
was higher by about one-third and the temperature by 10
°C, the value of Araki and Kuga seemed to be low even
after allowing for the 75% decrease in the hydrolysis time.
TEM pictures of samples B and G are shown in Figure 3.
In sample B, the crystalline fragments were aggregated to a
great extent. Sample G still showed some aggregation, but
a larger number of isolated fragments could be found. Our
results agree with the observation of Mukherjee and Woods
that cellulose hydrolyzed under more drastic conditions
aggregates less.22 Compared to cellulose crystals from
ramie,22 cotton,22 filter paper,6,8 or bleached kraft wood
pulp,4,7 the fragments from bacterial cellulose appeared to
be less uniform, which was also apparent in pictures of ref
2, 21, and 23. The length of the fragments ranged from about
200 nm to several micrometers. The reason for the nonuni-
form fragmentation could be related to the predominance of
the cellulose IR allomorph in bacterial cellulose or to
morphological differences in the microfibrils.24
Thermal Degradation Behavior and Morphology. The
TG and DTG curves of unhydrolyzed bacterial cellulose are
shown in Figure 4.
Three maxima in the weight-loss rate were apparent. The
broad DTG peak centered at 50 °C, corresponding to a weight
loss of 3.8%, was due to the evaporation of adsorbed water.
 1674 Biomacromolecules, Vol. 5, No. 5, 2004
Downloaded by VIRGINIA TECH on October 27, 2009 | http://pubs.acs.org
Publication Date (Web): June 29, 2004 | doi: 10.1021/bm034519+
Figure 3. TEM pictures of sulfuric acid hydrolyzed bacterial cellu-
loses: sample B (top), sample G (bottom). Arrows indicate isolated
fragments, bar: 200 nm.
Figure 4. TG (s) and DTG (- -) curves of unhydrolyzed bacterial
cellulose.
The DTG peak between 250 and 325 °C was caused by
concurrent cellulose degradation processes such as depo-
lymerization, dehydration, and decomposition of glycosyl
units followed by the formation of a charred residue. The
DTG peak above 425 °C was attributed to the oxidation and
Roman and Winter
Figure 5. (a) TG and (b) DTG curves of unhydrolyzed (s) and sulfuric
acid hydrolyzed bacterial celluloses, samples: B (- - -), F (- -), G
(s -), H (- - -).
breakdown of the charred residue to lower molecular weight
gaseous products.
The TG and DTG curves between 150 and 400 °C of
various hydrolyzed samples are shown in Figure 5. The
degradation behavior of the hydrolyzed samples showed
significant differences from that of unhydrolyzed bacterial
cellulose. With increasing sulfate content degradation started
at lower temperature and occurred over a broader temperature
range. Similar changes in the degradation behavior of
cellulose were observed with an increase in amorphous
content.25 Such an increase could occur during the hydrolysis
because sulfuric acid, above a certain concentration (g 65
wt %),26 is a cellulose solvent and at lower concentration
still causes swelling.
To rule out that an increase in amorphous content was
the reason for the changes in degradation behavior, the
morphology of the crystals was investigated by X-ray
diffraction. The diffraction patterns of selected samples are
shown in Figure 6.
The diffraction patterns of the hydrolyzed samples still
showed the typical reflections of cellulose I indicating that
the crystal integrity had been maintained. The crystallinity
in the unhydrolyzed sample was 85%. In most hydrolyzed
samples, the same amount of or a slight increase in
crystallinity was observed. Sample F had a lower crystallinity
(72%) than the unhydrolyzed sample. A possible explanation
is that the sample may have started to dissolve during
hydrolysis. However, neither sample G or H showed a similar
decrease in crystallinity and they were hydrolyzed under
more drastic conditions. Hence, we attributed the increase
in amorphous content in sample F to swelling during the
time-consuming washing process by centrifugation, which
was only applied to sample F. The fact that the amorphous
content decreased in most samples during the hydrolysis
 Hydrolysis of Bacterial Cellulose Biomacromolecules, Vol. 5, No. 5, 2004 1675

Figure 6. X-ray diffraction patterns of unhydrolyzed (sample A) and
sulfuric acid hydrolyzed bacterial celluloses.
Downloaded by VIRGINIA TECH on October 27, 2009 | http://pubs.acs.org
Between 150 and 350 °C, the DTG curves were investi-
gated by profile analysis. The degradation temperatures
(temperatures at maximum weight-loss rate), Tmax, and
maximum weight-loss rates, WLRmax, of the degradation
processes, obtained from profile analysis, are listed in Table
3.
With progressively higher sulfation, a more complex
degradation pattern was observed. Starting with sample C,
the degradation between 150 and 350 °C is best described
in terms of two processes. We propose that the lower
temperature process (termed “process 1” in Table 3) corre-
sponds to degradation of the more accessible, and therefore
more highly sulfated, amorphous regions, whereas the higher
temperature process (“process 2” in Table 3) relates to the
breakdown of the unsulfated crystal interior. At the highest
sulfation levels (samples G and H), a two-process degradation
fails to totally explain the experimental data. In the case of
sample H, we tested this possibility by adding an intermediate
temperature process and obtained an excellent fit to the

observed DTG data. For convenience, the third process in

Table 2. Water Content and Char at 350 °C of Sulfuric Acid sample H is listed in Table 3 as a second entry under process
Publication Date (Web): June 29, 2004 | doi: 10.1021/bm034519+

Hydrolyzed Bacterial Celluloses, from TG Curves

2. Under the more drastic hydrolysis conditions, applied in
sample water contenta (%) char at 350 °C (%)
the cases of samples G and H, sulfation might not only occur
A 3.8 11.6 in the amorphous regions but might also affect cellulose
B 3.1 16.5 chains at the splayed ends of the crystals. Accordingly, the
C 2.8 15.3
additional degradation process could be related to enhanced
D 2.7 15.6
E 2.8 14.1 access to the crystal interior due to degradation of sulfated
F 2.8 17.5 regions at the ends of crystals. The degradation temperatures
G 3.1 21.1 and maximum weight-loss rates of the different degradation
H 4.0 21.9 processes decreased with increasing sulfate content. The
a Weight loss at 110 °C.
differences in sulfate content in the samples C, D, and E
were too small to define a noticeable trend.
Table 3. Degradation Temperatures, Tmax, and Maximum A stepwise degradation behavior was also observed by
Weight-Loss Rates, WLRmax, of the Thermal Degradation
Processes of Sulfuric Acid Hydrolyzed Bacterial Celluloses, from Kim et al. in TGA curves of sulfuric acid impregnated
Profile Analysis of DTG Curves cellulose.15 They concluded that cellulose pyrolysis in the
process 1 process 2 presence of sulfate is divided into a low-temperature process
Tmax WLRmax Tmax WLRmax
between 110 and 200 °C and a high-temperature process
sample (°C) (% min-1) (°C) (% min-1) between 300 and 600 °C. Sample impregnation was done
by immersing the cellulose for a few minutes in sulfuric acid
A 297 7.45
B 288 5.10 with concentrations ranging from one to 20%. Under these
C 255 1.29 280 3.13 conditions, breakdown of the cellulose fibers is not expected
D 262 1.43 284 3.23 and the sulfate will be located primarily in the amorphous
E 260 1.27 283 3.33 regions on the fiber surface. Thus, the interpretation of our
F 249 1.03 280 2.78 three-step degradation behavior is still consistent with the
G 239 1.07 269 1.66 two-step behavior observed by Kim et al.
H 226 0.79 256 1.01
283 0.52
To elucidate the chemical events associated with process
1, sample H was heated to 228 °C and held isothermally for
suggested that the changes in thermal degradation behavior 10 min. The residue was then investigated by FTIR spec-
were due to the sulfate groups. troscopy. The FTIR spectra of the sample before and after
The water content and char at 350 °C of the different heating are shown in Figure 7.
samples, obtained from the TG curves, are listed in Table 2. The final weight loss after 10 min at 228 °C was 37%.
The water content in all samples was around 3%. In sample The spectrum, after heating, showed bands at 1732 and 1631
H, degradation started below 150 °C. Therefore, the weight cm-1 indicating the presence of carbonyl groups and unsatur-
loss at 110 °C might be due, in part, to degradation processes, ated carbon-carbon bonds in the sample. During thermo-
and the actual weight loss from water evaporation might be oxidative degradation of cellulose, there are a number of
less than the value shown in Table 2. The amount of charred reactions that lead to CdO and CdC bond formation. At
residue at 350 °C was larger in samples with a larger number low temperatures, CdC bonds arise from elimination of
of sulfate groups, confirming that the sulfate groups act as water involving the ring hydroxyl groups. Carbonyl groups
flame retardants. are formed by rearrangement of intermediary enols.27 A
 1676 Biomacromolecules, Vol. 5, No. 5, 2004 Roman and Winter

Table 4. Onset Temperatures and Activation Energies,

Determined with the Broido Method, of the Thermal Degradation
Processes in Sulfuric Acid Hydrolyzed Bacterial Celluloses
onset temp activation
process tempa range energy R2
no. sample (°C) (°C) (kJ mol-1) valueb
1 C 212 205-264 123.0 1.0000
D 220 212-269 130.3 0.9999
E 215 206-268 122.1 1.0000
F 203 200-247 111.0 0.9997
G 183 191-241 74.5 0.9998
H 180 198-235 71.2 0.9998
Figure 7. FTIR spectra of sulfuric acid hydrolyzed bacterial cellulose
2 A 272 284-300 283.4 0.9998
(sample H): (a) before and (b) after heating in air to 228 °C for 10
B 258 262-292 297.5 0.9999
min, normalized to the same intensity of the C(3)-O(3)H stretching
vibration at 1059 cm-1. C 254 264-283 138.9 0.9998
D 260 269-288 148.2 0.9998
E 257 268-287 146.6 0.9998
decrease in the size of the OH band between 3000 and 3800
F 252 265-284 127.1 0.9998
cm-1 upon heating indicated that partial dehydration had G 243 254-277 74.1 0.9999
taken place. H 212 240-260 60.8 0.9995
Since the elimination of sulfuric acid in sulfated anhy- 258 270-287 34.5 0.9995
Downloaded by VIRGINIA TECH on October 27, 2009 | http://pubs.acs.org

droglucose units requires less energy than elimination of a From profile analysis of DTG curves. b Regression coefficient from
water,14 desulfation probably also occurred. The released
Publication Date (Web): June 29, 2004 | doi: 10.1021/bm034519+

linear regression analysis of a plot of ln(ln 1/y) vs 1/T, see Experimental

sulfuric acid molecules stay in the heated zone and do not
decompose or volatize until higher temperatures (g380 °C).14
If behaving like phosphoric acid, sulfuric acid facilitates
removal of the ring hydroxyl groups by either direct catalysis
of the elimination of water or esterification of the hydroxyl
groups and subsequent elimination of sulfuric acid.13 Hence,
process 1 probably includes both desulfation and acid-
catalyzed dehydration. The large weight loss of 37% after
10 min at 228 °C suggests that partial decomposition or
depolymerization and volatilization of the products had also
taken place.
Removal of the ring hydroxyl groups reduces depoly-
merization through intramolecular transglycosylation27 and,
thus, the amount of volatile anhydro sugars formed, consti-
tuting the tar fraction. A suppression of intramolecular
transglycosylation could be the reason for the larger char
fraction observed in samples with larger amounts of sulfate
groups. However, the char fraction also increases through
decomposition of anhydro sugars after their formation before
volatilization and through thermal auto-cross-linking of
cellulose chains. Both processes are also acid catalyzed.28,29
In an oxidative atmosphere, degradation processes involving
radicals, such as the thermal auto-oxidation of cellulose via
hydroperoxide groups, also have to be considered. The effect
of sulfuric acid on radical reactions is not obvious.
The activation energies of the degradation processes
between 150 and 350 °C were determined with the Broido
method. The calculated values are listed in Table 4 together
with the temperature ranges and R2 values for the linear
regression. The onset temperatures of the degradation peaks,
obtained from profile analysis of the DTG curves, are also
included.
The activation energies for both processes were signifi-
cantly lower for larger amounts of sulfate groups, suggesting
that degradation reactions had been catalyzed by sulfuric acid.
Catalysis could either be direct through the acid molecules
or indirect by promoting dehydration reactions and increasing
the amount of water released. Water is known to catalyze
Section.
cellulose degradation reactions.30 In addition, removal of the
ring hydroxyl groups leads to a loss of the stabilizing intra-
and intermolecular hydrogen bonds, which could also lower
the activation energy for degradation.
Conclusions
The objective of this study was to optimize the hydrolysis
conditions for the application of the crystals as reinforce-
ments in polymer composites. In view of the considerable
decrease in degradation temperatures caused even by small
amounts of sulfate groups, esterification during hydrolysis
should be kept at a minimum. Thus low acid concentrations,
low acid-to-cellulose ratios, and short reaction times are
preferred. An alternative is to use hydrochloric acid, which
does not introduce any acidic groups. The aggregation of
cellulose crystals from hydrochloric acid hydrolysis,7 due
to the lack of surface charges, was a disadvantage of
hydrochloric acid with respect to sulfuric acid. However,
because the suppression of aggregation by introduction of
charged sulfate groups compromises the thermostability of
the crystals, aggregation of the crystals cannot be avoided
even with sulfuric acid.
Acknowledgment. The authors gratefully acknowledge
generous support by USDA-CSREES awards (W.T.W.) 96-
35501-3454 and 98-35504-6358 and fellowship support for
M.G. by the German Academic Exchange Service (DAAD).
Assistance by Dr. R. B. Hanna and Dr. S. E. Anagnost with
the TEM is also acknowledged.
References and Notes
(1) Rånby, B. G. Acta Chem. Scand. 1949, 3, 649.
(2) Marchessault, R. H.; Morehead, F. F.; Joan Koch, M. J. Colloid Sci.
1961, 16, 327.
(3) Marchessault, R. H.; Morehead, F. F.; Walter, N. M. Nature 1959,
184, 632.
(4) Revol, J.-F.; Bradford, H.; Giasson, J.; Marchessault, R. H.; Gray,
D. G. Int. J. Biol. Macromol. 1992, 14, 170.
 Hydrolysis of Bacterial Cellulose
(5) Revol, J.-F.; Godbout, L.; Dong, X.-M.; Gray, D. G.; Chanzy, H.;
Maret, G. Liq. Cryst. 1994, 16, 127.
(6) Dong, X. M.; Revol, J.-F.; Gray, D. G. Cellulose 1998, 5, 19.
(7) Araki, J.; Wada, M.; Kuga, S.; Okano, T. Colloids Surf. 1998, 142,
75.
(8) Araki, J.; Wada, M.; Kuga, S.; Okano, T. Langmuir 2000, 16, 2413.
(9) Favier, V.; Chanzy, H.; Cavaillé, J. Y. Macromolecules 1995, 28,
6365.
(10) Helbert, W.; Cavaillé, J. Y.; Dufresne, A. Polym. Compos. 1996,
17, 604.
(11) Chazeau, L.; Cavaillé, J. Y.; Canova, G.; Dendievel, R.; Boutherin,
B. J. Appl. Polym. Sci. 1999, 71, 1797.
(12) Tang, W. K.; Neill, W. K. J. Polym. Sci., Part C 1964, 6, 65.
(13) Barker, R. H.; Drews, M. J. In Cellulose chemistry and its
applications; Nevell, T. P., Zeronian, S. H., Eds.; Ellis Harwood:
Chichester, U.K., 1985; p 423.
(14) Julien, S.; Chornet, E.; Overend, R. P. J. Anal. Appl. Pyrolysis 1993,
27, 25.
(15) Kim, D.-Y.; Nishiyama, Y.; Wada, M.; Kuga, S. Cellulose 2001, 8,
29.
(16) Parks, E. J. Tappi J. 1971, 54, 537.
(17) Verlhac, C.; Dedier, J.; Chanzy, H. J. Polym. Sci., Part A: Polym.
Chem. 1990, 28, 1171.
Downloaded by VIRGINIA TECH on October 27, 2009 | http://pubs.acs.org
Publication Date (Web): June 29, 2004 | doi: 10.1021/bm034519+
Biomacromolecules, Vol. 5, No. 5, 2004 1677
(18) Gran, G. Analyst 1952, 77, 661.
(19) Krumm, S. WinFit, Beta Release 1.2.1; Geology Department,
University of Erlangen: Erlangen, Germany, 1997.
(20) Broido, A. J. Polym. Sci., Part A-2 1969, 7, 1761.
(21) Araki, J.; Kuga, S. Langmuir 2001, 17, 4493.
(22) Mukherjee, S. M.; Woods, H. J. Biochim. Biophys. Acta 1953, 10,
499.
(23) Rånby, B. G. Ark. Kemi 1952, 4, 249.
(24) Yamamoto, H.; Horii, F. Cellulose 1994, 1, 57.
(25) Cabradilla, K. E.; Zeronian, S. H. In Thermal Uses and Properties
of Carbohydrates and Lignins; Shafizadeh, F., Sarkanen, K. V.,
Tillman, D. A., Eds.; Academic Press: New York, 1976; pp 73-96.
(26) Jayme, G.; Lang, F. In Methods in Carbohydrate Chemistry; Whistler,
R. L., Ed.; Academic Press: New York, 1963; Vol. 3, pp 75-83.
(27) Shafizadeh, F. In Cellulose chemistry and its applications; Nevell,
T. P., Zeronian, S. H., Eds.; Ellis Harwood: Chichester, U.K., 1985;
pp 266-289.
(28) Fung, D. P. C.; Tsuchiya, Y.; Sumi, K. Wood Sci. 1972, 5, 38.
(29) Black, E. L. Pulp Paper Magn. Can. 1967, 68, T-165.
(30) Scheirs, J.; Camino, G.; Tumiatti, W. Eur. Polym. J. 2001, 37, 933.
BM034519+