Thermogravimetric analysis during the decomposition of cotton fabrics in an inert and air environment

Article (PDF Available)inJournal of Analytical and Applied Pyrolysis 76(1-2):124-131 · June 2006with30 Reads
DOI: 10.1016/j.jaap.2005.09.001 · Source: OAI
Abstract
The thermal degradation of samples of used cotton fabrics has been investigated using thermogravimetric analysis (TGA) between room temperature and 700 °C. Experiments were carried out with about 5 mg of sample in three different atmospheres: helium, 20% oxygen in helium and 10% oxygen in helium. Three different heating rates were used at each atmosphere condition. A kinetic model for the decomposition of used cotton fabrics explaining the behavior of all the runs performed has been proposed and tested. For the pyrolysis of the cotton, the model comprises two parallel reactions. For the combustion process, one competitive reaction was added to each parallel reaction of the pyrolysis model and four combustion reactions of the different solid fractions to obtain volatiles. One single set of parameters can explain all the experiments (pyrolysis, oxidative pyrolysis and combustion) at the three different heating rates used.
UNCORRECTED PROOF
+ Models
JAAP 1877 1–8
Thermogravimetric analysis during the decomposition of
cotton fabrics in an inert and air environment
Julia Molto
´
*
, Rafael Font, Juan A. Conesa, Ignacio Martı
´
n-Gullo
´
n
Chemical Engineering Department, University of Alicante, P.O. Box 99, E-03080 Alicante, Spain
Received 29 November 2004; accepted 4 September 2005
Abstract
The thermal degradation of samples of used cotton fabrics has been investigated using thermogravimetric analysis (TGA) between room
temperature and 700 8C. Experiments were carried out with about 5 mg of sample in three different atmospheres: helium, 20% oxygen in helium
and 10% oxygen in helium. Three different heating rates were used at each atmosphere condition. A kinetic model for the decomposition of used
cotton fabrics explaining the behavior of all the runs performed has been proposed and tested. For the pyrolysis of the cotton, the model comprises
two parallel reactions. For the combustion process, one competitive reaction was added to each parallel reaction of the pyrolysis model and four
combustion reactions of the different solid fractions to obtain volatiles. One single set of parameters can explain all the experiments (pyrolysis,
oxidative pyrolysis and combustion) at the three different heating rates used.
# 2005 Published by Elsevier B.V.
Keywords: Pyrolysis; Combustion; Cotton fabrics; Kinetics; Thermogravimetry
1. Introduction
The growing interest in renewable energies is accompanied
by intensified research and development of technical processes
for the thermal conversion of biomass and wastes. Used cotton
fabrics could be used as biomass, and in this way, offer a valid
alternative to disposal in landfills. Although cotton fabrics are
usually recycled in other ways, thermal decomposition only of
the wastes is interesting for some industries that mainly focus
on obtaining the potential energy by combustion.
The development of pyrolysis–combustion processes
requires an optimization of the operating conditions in order
to assure both acceptable gas outlet composition and an energy
recovery, which makes the process economically satisfactory
[1]. A good knowledge of the kinetics of the process is
fundamental for the plant design and scale-up bases on process
simulation.
Cotton is mainly comprised of cellulose. The pyrolytic
degradation of cellulose has been the subject of extensive study,
although in many instances, knowledge of the exact nature of
degradation and decomposition remains incomplete. Chatterjee
and Conrad [2] studied the kinetics of cellulose decomposition
in the temperature range of 270–310 8C with absorbent cotton,
and proposed two series reactions for the pyrolysis process.
Dollimore and Hoath [3] used differential thermal analysis
(DTA) to follow the degradation of cellulose in air products and
obtained two and sometimes three exothermic peaks. Antal and
Va
´
rhegyi [4] reviewed the literature of cellulose pyrolysis and
concluded that the pyrolysis of a small sample of pure cellulose
is characterized by an endothermic reaction governed by a first
order rate law with a high activation energy. Vo
¨
lker and
Rieckmann [5] investigated the influence of the final mass on
modelling results and evaluated the applicability of established
kinetic models for engineering purposes.
Cotton fabrics, which have a major share of the textile
market, are highly inflammable and the development of
successful flame retardant systems for cotton is of major
interest. For this reason, many authors have studied the
mechanism of pyrolysis of untreated and flame retardant treated
cotton fabrics. For instance, Faroq et al. [6] carried out the
thermogravimetric analysis of the mechanism of pyrolysis of
untreated cotton fabrics and cotton fabrics finished with various
flame retardant, considering the fraction decomposed as
between 0.1 and 0.9. These authors evaluated the activation
www.elsevier.com/locate/jaap
J. Anal. Appl. Pyrolysis xxx (2005) xxx–xxx1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
* Corresponding author. Tel.: +34 96 590 34 00x3003; fax: +34 96 590 38 26.
E-mail address: julia.molto@ua.es (J. Molto
´
).
0165-2370/$ see front matter # 2005 Published by Elsevier B.V.
doi:10.1016/j.jaap.2005.09.001
UNCORRECTED PROOF
energies obtaining values of 155–158 kJ mol
1
for the thermal
decomposition of untreated cotton. In addition, they proposed a
second order decomposition mechanism as the most suitable.
Wanna and Powell [7] studied the thermal decomposition of
untreated and treated cotton fabric with selected salts in
oxidative and inert atmospheres using TGA-FTIR. On the other
hand, no papers considering the overall decomposition kinetics
of cotton cellulose or cellulose have been found.
It is clear that the n-th order kinetic model can be correct for
homogeneous gas phase kinetics, in accordance with the
collision theory and the transition state theory developed for
elemental reactions and normally with n = 1, 2 or 3. The
pyrolysis of polymers it is a non-homogeneous reaction and
consequently the n-th model should not be valid. In literature,
we can find mechanisms for the pyrolysis of different polymers
with parallel and series reactions. Considering only one
elemental reaction for the decomposition of a solid, the
proposal of a kinetic expression based on the elemental reaction
is not easy because the decomposition takes place in a solid
phase. Font and Garcı
´
a [8] proposed the application of the
transition state theory to the pyrolysis of biomass considering
the similarity between the pyrolitic reaction that take places in
the outer surface and the first order uni-molecular catalytic
surface reactions. In accordance with the model developed, it is
possible to obtain n-th models as a consequence of the increase
or decrease of the surface with active centres where the
decomposition can take place, and with fractional values of the
reaction order n.
The n-th order reaction has been used extensively by
different researchers when studying the mechanisms of
decomposition of several polymers, no matter the mechanisms
of reactions comprises parallel or series reactions. Never-
theless, in spite of the proposal of a model that could explain
fractional reaction orders, the models obtained from the TG
runs must be considered as correlation ones, and from the
analysis of the activation energy, it can be deduced if the model
is related satisfactorily to a controlling decomposition
elemental step or must be only considered valid for correlation.
The present work, which is included in a wider project
whose objective is to study the combustion of different
industrial and municipal wastes, studies the thermal decom-
position of used and waste cotton fabrics from the thermo-
gravimetric point of view, including the kinetic analysis in a
thermobalance in inert atmosphere and with different amounts
of oxygen, proposing a kinetic model.
2. Experimental
2.1. Raw material
Used cotton fabrics were simulated by using a used blue T-
shirt made of 100% cotton. Prior to the runs, the T-shirt was cut
into small pieces with an average size of 1 cm 1 cm.
Table 1 shows some characteristics of the material studied.
Elemental analysis of the major components was carried out in
a Perkin-Elmer 2400. The moisture was determined by the
weight loss at 105 8C for 12 h. The calorific value was obtained
in an AC-350 calorimetric bomb from Leco Corporation.
Chlorine was measured using an automatic sequential spectro-
meter X-ray Fluorescence model TW 1480. Ash residue was
obtained by calcination at 850 8C.
2.2. Thermobalance
The thermogravimetric experiments were carried out in a
Setaram thermobalance model DSC92 controlled by a PC
system. The atmosphere used for pyrolysis was helium with a
flow rate of 60 ml min
1
(STP), according to the specifications of
the equipment. In the combustion runs, two mixtures of helium
and oxygen: 9:1 (10% oxygen) and 4:1 (20% oxygen) were used,
with the same total flow rate. The sample temperature was
measured with a thermocouple directly at the crucible, i.e., very
close to the sample. Because a water-cooled microfurnace was
used, the temperature could be lowered rapidly.
Before the runs with used cotton fabrics, an experiment with
a heating rate of 5 8C min
1
using Avicel PH-105 microcrystal-
line cellulose was done to check the good performance of the
equipment. The results obtained showed good agreement with
the kinetic evaluation of Avicel Cellulose TG curves at this
heating rate presented by Grønli et al. [9] in their round-robin
study of cellulose pyrolysis kinetics by thermogravimetry.
The experiments were carried out with heating rates of 5, 10
and 20 8C min
1
over a variety of temperatures that included
the entire range of solid decomposition, 80–700 8C. Experi-
ments without a sample were carried out, and used as
background in order to subtract the buoyancy effect. The mass
of the samples used was approximately 5 mg, and under these
conditions the heat transfer limitations can be neglected.
3. Results and discussion
3.1. Thermogravimetric study
Figs. 1–3 show in detail the experimental curves for used
cotton fabrics pyrolysis (helium) and combustion (helium:oxy-
gen, 4:1 and 9:1) at different heating rates. The calculated
curves of the kinetic models are also shown. In the figures, w is
defined as the residual mass fraction of the solid (including
residue formed and non-reacted initial solid), i.e., the ratio
between the solid mass at any time (m) and the initial solid mass
(m
0
). In all the processes, we can be observed the general shift
to higher temperatures when the heating rate is increased.
J. Molto
´
et al. / J. Anal. Appl. Pyrolysis xxx (2005) xxx–xxx2
+ Models
JAAP 1877 1–8
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
Table 1
Characteristics of the material used
C (wt.%) 45.5
H (wt.%) 6.6
N (wt.%) 0.3
S (wt.%) <0.1
O% by difference (wt.%) 47.5
Cl (wt.%) 0.12
Ash content (wt.%) <0.1
Net calorific value (kJ kg
1
) 17100
Moisture (wt.%) 5.0
UNCORRECTED PROOF
Fig. 4 compares the results obtained in the different
atmospheres at a constant heating rate. As can be seen, the
presence of oxygen accelerates the decomposition (taking place
at lower temperatures) and a final combustion process is
observed. On the other hand, there is not a great difference
between the runs performed under 10% oxygen and those at
20% oxygen.
3.2. Kinetics of the process
3.2.1. Pyrolysis model
Most of the papers published corresponding to the
pyrolysis of cellulose and cotton consider only one fraction
when correlating the experimental data of the primary
decomposition, in spite of the parallel and series mechanisms
proposed in literature [4]. In the cotton fabrics used in this
work, and considering that we have extended the tempe-
rature range until high temperatures, a second decompo-
sition process has been considered in order to improve the
correlation.
The kinetic model proposed for the decomposition of the
used cotton fabrics in an inert atmosphere could be interpreted
considering this waste formed by two independent parts, each
one following an independent reaction, as follows:
c
1
C
1
!
1
v
1
V
1
þ s
1
S
1
c
2
C
2
!
2
v
2
V
2
þ s
2
S
2
where C
1
and C
2
refer to different parts of the solid material to
be decomposed (cotton in this case). V
i
are the gases + vo-
latiles evolved and S
i
are the solid residue formed in the
decomposition. The uncapitalized variables c
n
are the
amount of each material that is in the sample.
It is very useful to introduce the concept of the conversion
degree for each reaction:
a
i
¼ 1
C
n
C
no
¼
V
i
V
i1
; i ¼ n ¼ 1; 2
(Two different subscripts have been used, because in the
combustion model presented also in this model there are
competitive reactions for the same solid.)
In the previous equations, V
i1
represents the maximum
obtainable amount of volatiles via reaction i at time infinity,
J. Molto
´
et al. / J. Anal. Appl. Pyrolysis xxx (2005) xxx–xxx 3
+ Models
JAAP 1877 1–8
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
180
181
182183
184185
186
187
188
189
190
191
192193
194
195
196
197
198
Fig. 1. Cotton pyrolysis at several heating rates: 5, 10 and 20 8C min
1
.
Experimental and calculated curves.
Fig. 2. Cotton combustion at several heating rates and 20% oxygen. Experi-
mental and calculated curves.
Fig. 3. Cotton combustion at several heating rates and 10% oxygen. Experi-
mental and calculated curves.
Fig. 4. Experimental TG plots for pyrolysis and combustion with 20 and 10%
oxygen, all carried out at 20 8C min
1
heating rate.
UNCORRECTED PROOF
C
n
refers to the non-decomposed material at each time, and C
no
is the initial contribution of fraction n to the total weight. Note
that at t = 0 the value of a
i
is zero, and that V
11
equals the yield
coefficient v
1
and that V
21
equals the yield coefficient v
2
.
The kinetic equations associated with the parallel reaction
for the pyrolysis runs, taking into account the mass balance
between products and reactants and the degree conversions, can
be expressed as:
By integration of these reactions it is possible to calculate a
1
and a
2
at each time; the relationship between these two values
and the weight fraction measured in the thermobalance ðwÞ is:
w ¼ 1 V ¼ 1 ðV
1
þ V
2
Þ¼1 ðV
11
a
1
þ V
21
a
2
Þ
The values of V
11
and V
21
are related with the total
volatiles at time infinity (V
1
), that is a known amount:
V
1
¼ V
11
þ V
21
The pyrolysis data and the combustion data have been
correlated together, in order to obtain a single set of parameters
for the thermal decomposition of used cotton fabrics in different
atmospheres: pyrolysis, oxidative pyrolysis and combustion.
3.2.2. Combustion model
Similarly to the kinetic models found in literature to explain
pyrolytic processes, different authors propose different models
to explain decomposition mechanisms under oxidative atmo-
spheres. Obviously, if the model proposed satisfactorily fits
experimental data generated under a wide selection of
conditions, the model can be considered representative of
the process analyzed [10]. However, literature that includes
detailed kinetic studies, fitting experimental curves at different
heating rates and oxygen content is extremely sparse. Different
models have been found in literature, concerning the
decomposition of tannery waste under oxygenated atmosphere
[11], polycoated materials such as milk cartons [12],
polytetrafluoroethylene [13] and also of tire wastes [14].No
models for combustion of cotton or used cotton fabrics have
been found, as commented previously.
The kinetic model proposed for the combustion runs could
be interpreted considering, that the presence of the oxygen
introduces a new competitive process for the decomposition of
each fraction and in this way this model explains the fact
observed in Fig. 4, where the first weight loss in combustion
runs reaches a lower value than in pyrolysis runs.
Figs. 5 and 6 show the TG and DSC plots for pyrolysis
and combustion at 5 8Cmin
1
heating rate. As seen in Fig. 5,
an endothermic peak appears for the pyrolysis process
and in the combustion process besides two exothermic peaks
appear.
Taking into account the behavior commented above, the
oxygen to be included in the decomposition law has been
considered, as has been done with other materials [15–16].
The following scheme represents the combustion model
proposed to explain the behaviour obtained at two different
helium: oxygen atmospheres and three heating rates:
c
1
C
1
!
1
v
1
V
1
þ s
1
S
1
c
1
C
1
þ O
2
!
3
v
3
V
3
þ s
3
S
3
c
2
C
2
!
2
v
2
V
2
þ s
2
S
2
c
2
C
2
þ O
2
!
4
v
4
V
4
þ s
4
S
4
J. Molto
´
et al. / J. Anal. Appl. Pyrolysis xxx (2005) xxx–xxx4
+ Models
JAAP 1877 1–8
198
199
200
201
202
203
204
205
206
206
207
208
209210
211
212
213214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
229
230
231
232
233
234
235
236
237
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252253
254255
256257
258259
Fig. 6. Experimental TG and DSC plots for combustion with 20% oxygen at
5 8C min
1
heating rate.
Fig. 5. Experimental TG and DSC plots for pyrolysis at 5 8C min
1
heating
rate.
d
C
n
C
no

dt
¼
d
V
i
V
i1

dt
¼
da
i
dt
¼ k
i
C
n
C
no

n
i
¼ k
i
ð1 a
i
Þ
n
i
k
i
¼ k
io
exp
E
i
RT

9
>
>
>
>
>
=
>
>
>
>
>
;
; n; i ¼ 1; 2
UNCORRECTED PROOF
Reactions 1 and 2 are the same as in pyrolysis. Furthermore,
as a final new process appears in combustion runs, the
combustion reactions to the chars formed has been considered:
s
i
S
i
þ O
2
!
5
s
i
V
ci
; i ¼ 1; 2; 3; 4
V
ci
refers to the volatiles produced through the combustion
reaction with solid S
i
.
Conversion degrees are considered for each reaction:
a
i
¼ 1
ðC
n
Þ
reacted by reaction i
C
no
¼
V
i
V
i1
;
i ¼ 1; 2; 3; 4; n ¼ 1 ðfor i ¼ 1; 3Þ; n ¼ 2 ðfor i ¼ 2; 4Þ
a
ci
¼
ðS
i
Þ
burnt
s
i
¼
V
ci
V
ci1
;
i ¼ 1; 2; 3; 4; n ¼ 1 ðfor i ¼ 1; 3Þ; n ¼ 2 ðfor i ¼ 2; 4Þ
where V
i1
and V
ci1
represent the maximum amounts of
volatiles evolved if the whole solid fraction decomposes only
through the reactions that lead to the corresponding volatiles,
without competitive reaction. Note that in this case V
i1
equals
the yield coefficient v
i
and that V
ci1
equals the yield coefficient
s
i
.
Reactions 1 and 3 are competitive with respect to the same
solid C
1
, so it is possible that none of the two values of degree
conversion a
1
and a
2
can reach the value 1 at time infinity,
although the sum a
1
and a
2
logically must be equal to 1, when
the reactant C
1
is exhausted. In this case, the ratio between the
non-reacted solid C
1
and the initial solid C
1o
, taking into
account the degree conversions a
1
and a
3
, can be expressed as:
C
1
C
1o
¼ 1 a
1
a
3
and consequently the kinetic laws for the decomposition of the
solid fraction C
1
can be written as
d
C
1
C
1o
by reaction i
dt
¼
da
i
dt
¼ k
i
C
1
C
1o
¼ k
i
ð1 a
1
a
3
Þ
n
i
;
i ¼ 1; 3
Similarly for reactions 2 and 4, the following expression can
be obtained:
d
C
2
C
2o
by reaction i
dt
¼
da
i
dt
¼ k
i
C
2
C
2o
¼ k
i
ð1 a
2
a
4
Þ
n
i
;
i ¼ 2; 4
For the combustion of the residue S
1
(formed by the first
reaction) in accordance with the scheme:
s
1
S
1
þ O
2
!
5
s
1
V
c1
The degree conversion a
c1
is
a
c1
¼
ðS
1
Þ
burnt
s
1
¼
V
c1
V
c11
The kinetic law for the combustion of S
1
can be written as:
dðS
1
=s
1
Þ
burnt
dt
¼ k
5
S
1
s
1
n
5
In this way, the kinetic constant of this second decomposi-
tion does not depend on the initial mass fraction, and
consequently the kinetic constant k
5
with distinct mass
fractions.
On the other hand, it can be deduced that:
S
1
s
1
at any time
¼
S
1
s
1
formed reaction 1
S
1
s
1
burnt
¼ a
1
a
c1
Consequently, it can be written that:
da
c1
dt
¼ k
5
ða
1
a
c1
Þ
n
5
Another way of obtaining similar expressions can be found
elsewhere [17]. This procedure can be applied to the other three
combustion reactions. Consequently, the reaction model can be
solved considering the following equations:
da
i
dt
¼ k
i
ð1 a
1
a
3
Þ
n
i
; i ¼ 1; 3
da
i
dt
¼ k
i
ð1 a
2
a
4
Þ
n
i
; i ¼ 2; 4
da
1i
dt
¼ k
5
ða
i
a
1i
Þ
n
5
; i ¼ 1; 2; 3; 4
v
1
¼ V
11
; v
2
¼ V
21
; v
3
¼ V
31
; v
4
¼ V
41
s
1
¼ V
c11
; s
2
¼ V
c21
; s
3
¼ V
c31
; s
4
¼ V
c41
v
1
þ s
1
¼ v
3
þ s
3
v
2
þ s
2
¼ v
4
þ s
4
Note that the same kinetic constant and reaction order are
considered for the decomposition of S
i
.
The total weight fraction is related to the other variables by:
w ¼ 1 V
¼ 1 ðV
1
þ V
2
þ V
3
þ V
4
þ V
11
þ V
12
þ V
13
þ V
14
Þ
w ¼ 1
ða
1
V
11
þ a
2
V
21
þ a
3
V
31
þ a
4
V
4
a
c1
V
c11
þ a
c2
V
c21
þ a
c3
V
c31
þ a
c4
V
c4a
Þ
To take into account the effect of the partial pressure of
oxygen (which equals 0.10 and 0.20 atm for helium:oxygen 9:1
and 4:1, respectively), since different behaviour is observed
when comparing corresponding thermograms, the pre-expo-
nential factors for reactions with oxygen have been considered
to consist of two terms, one a typical pre-exponential factor k
0
io
and the other one the partial pressure of oxygen P
O
2
raised to
the power of an order b
io
:
k
io
¼ k
0
io
ðP
O
2
Þ
b
i
; i ¼ 3; 4; 5
J. Molto
´
et al. / J. Anal. Appl. Pyrolysis xxx (2005) xxx–xxx 5
+ Models
JAAP 1877 1–8
258259260
261
262
263264
265
266
267268
269270
271
272
273
274
275
276
277
278
279
280
281
282
283284
285
286287
288
289
290291
292
293
294295
296
297298
297298299
300301
302
303
304
305
306
307308
309
310311
312
313
314
315
316317
318319
320321
322323
324325
326327
328329
330
331
332
333334
335336
337
338
339
340
341
342
343
344
345346
UNCORRECTED PROOF
The parameters optimized were: five orders of reaction, five
pre-exponential factors, five activation energies and v
1
, v
2
, v
3
,
v
4
, s
1
, s
2
, and the values of b
3
, b
4
, b
5
. A simplification made in
the model is to assume the same dependency of the constants
with the partial pressure of oxygen, in the combustion of the
chars.
The objective function (OF) to minimize was the sum of the
square differences between experimental and calculated weight
loss values
OF ¼
X
3
m¼1
X
N
j¼1
ðw
exp
mj
w
calc
mj
Þ
2
; m; heating rates; j; points
The model validity has been tested calculating the variation
coefficient (VC):
VC ¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
OF=ðN PÞ
p
w
exp
100
where N and P are the number of data and parameters fitted,
respectively, and
w
exp
is the average of the experimental
weights. According to the procedure suggested by Martı
´
n-
Gullo
´
n et al. [18], the great interrelation existing among the
pre-exponential factor, the apparent activation energy and the
reaction order can be decreased when optimization is per-
formed in terms of a ‘comparable kinetic constant’ K
i
instead
of optimizing k
oi
. This constant is calculated in a temperature
around the maximum decomposition rate (T
max
). Since K
i
, E
i
and n
i
are optimised the pre-exponential factor k
oi
is calculated
using the following expression:
K
i
¼ k
i
ð0:64Þ
n
i
¼ k
oi
exp
E
i
RT
max
ð0:64Þ
n
i
Conesa et al. [19,20] came to a conclusion in their study of
TG-DTG curves that at least three TG curves with different
heating rates must be adjusted simultaneously in order for a
kinetic model to be considered as potentially correct.
The kinetic parameters have been optimized in order to
minimize the differences between experimental and calculated
weight loss. Table 2 presents the values of the optimised
parameters. Note that all the data of the three atmospheres
(pyrolysis, oxidative pyrolysis, and combustion) and the three
heating rates have been simultaneously fitted.
As can be seen in Figs. 1–4, the model is able to explain
all the experimental data collected, at all the heating rates
studied and in the presence and absence of oxygen. Bear in
mind that all the runs are fitted with no variation of the
parameters.
On analysing the values of the parameters in Table 2, some
interesting conclusions can be obtained:
1. Considering the pyrolysis parameters, it can be observed that
fraction 1 has a value of v
1
, that equals a V
11
, around 0.75,
indicating that this fraction is the most important, although
there is a small but significant second fraction, with a value
v
2
around 0.11. Analysing the TG data of some papers
concerning the pyrolysis of cellulose, this second fraction is
also present [21,5].
2. Concerning the kinetic parameters of the main fraction in the
pyrolysis decomposition, activation energy around
161 kJ mol
1
and reaction order around 0.57 are obtained,
that are similar to the values proposed by Antal et al. (1980)
when considering all the runs. The activation energy E
1
obtained is also very similar to the values calculated by Faroq
et al. [6] using two different methods: the iso-conversional
technique and the most probable mechanism.
3. With respect to the minor fraction in pyrolysis, a low value of
activation energy around 65 kJ mol
1
and a reaction order
equal to 2.77 are obtained, as a consequence of the slow
decomposition of this fraction at high temperatures, and
consequently these values can be considered as correlation
parameters with no physical meaning.
4. The oxidative pyrolysis of the main fraction has activation
energy (159 kJ mol
1
) similar to that of pyrolysis
(161 kJ mol
1
), and the reaction order (0.29) is somewhat
lower than in pyrolysis (0.57). An order reaction less than
unity can be due to an increase of the active surface with the
reaction extension [8], so it is possible that in the oxidative
process, the active surface increases more intensively than in
pyrolysis. The value of v
3
(0.569), that equals V
31
, is less
than the value (V
11
= 0.755) corresponding to the main
fraction in pyrolysis. This fact could be explained as a
J. Molto
´
et al. / J. Anal. Appl. Pyrolysis xxx (2005) xxx–xxx6
+ Models
JAAP 1877 1–8
345346347
348
349
350
351
352
353
354
355
356357
358
359
360361
362
363
364
365
366
367
368
369
370
371
372373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391392392
393393
394
395
396
397
398
399
400400401
402
403
404
405
406
407
408
409409410
411
412
413
414
415
416416417
418
419
420
421
422
423
424
425
426
427
427
Table 2
Kinetic parameters obtained for the thermal decomposition of used cotton
fabrics (k
0
io
in s
1
atm
b
i
, E
i
in kJ mol
1
)
Pyrolysis
k
0
1o
9.76 10
10
E
1
161.1
n
1
0.57
v
1
0.749
s
1
0.103
k
0
2o
1.32 10
3
E
2
65.2
n
2
2.62
v
2
0.124
s
2
0.024
Oxidative pyrolysis
k
0
3o
2.70 10
11
b
3
0.39
E
3
159.4
n
3
0.29
v
3
0.589
s
3
0.263
k
0
4o
1.33 10
23
b
4
0.57
E
4
306.4
n
4
1.49
v
4
0.148
s
4
0
Combustion of char
k
0
5o
4.47 10
6
b
5
0.31
E
5
126.8
n
5
0.64
VC (%) 2.628
UNCORRECTED PROOF
consequence of the partial oxidation of the molecular organic
chains to ketones, aldehydes, carboxylic acids, etc., causing a
weight loss lesser than in the case without oxygen. Therefore,
all these parameters can be considered as acceptable and
logical from a physical–chemical point of view.
5. In the oxidative pyrolysis of the minor fraction, the activation
energy is around 150 kJ mol
1
, the reaction order is 0.97 and
there is no formation of residue solid. These parameters must
only be accepted for correlation of the data.
6. In the combustion of the chars, assuming the same kinetics
for the three residue solids formed (S
1
, S
2
and S
3
), the
apparent activation energy is around 126 kJ mol
1
.The
apparent activation energy obtained for the combustion of
the char formed (126 kJ mol
1
) is similar to that obtained by
other researchers. Haji-Sulaiman and Aroua [22] proposed
values around 135 kJ mol
1
for the oxidation of Malaysian
coal chars. Henrich et al. [23], also obtained values of
apparent activation energy around 140 kJ mol
1
are for
oxidation in pure oxygen of soot, graphite, activated
charcoals and chars of municipal waste and electronic
scrap. Nevertheless, Walker et al. [24] suggested apparent
activation energy around 209–242 kJ mol
1
for the reaction:
C + 1/2 O
2
! CO. A reaction order (0.57) less than unity is
obtained. This latter aspect could also be explained as a
consequence of the increase of the active surface with the
reaction extension.
7. The reaction orders with respect to the oxygen for the
oxidative pyrolysis are 0.39 for the main fraction and 0.45 for
the smaller fraction, indicating a low dependence of this with
respect to the oxygen. For the combustion of char the
reaction order with respect to oxygen is 0.31, that is a low
value in comparison with respect to the reaction order
between 0.7 and 1 considered in the gasification of the
carbon [25]. The experimental data shown in Fig. 4 shows
the small dependence of the oxygen on the TG runs. On the
other hand, it is possible that the simplification of the model,
considering that the three combustion reactions follow the
same kinetic law causes the correlation parameters to be
somewhat different to those expected.
A study of the main by-products formed in the pyrolysis and
combustion of used cotton fabrics, was carried out in a very
recent paper [26]. More than 90 compounds, including carbon
oxides, light hydrocarbons and PAHs, have been identified and
quantified. In the gas phase some of the main components
obtained were methane, ethene and benzene. The main
semivolatile compounds detected were styrene, phenol,
naphthalene, acenaphthylene and phenanthrene.
The evolution of the different fractions throughout the
heating process can be observed in Figs. 7 and 8, that show the
experimental and calculated mass fraction of volatiles v and
conversion factors a at 5 8C min
1
in He:O
2
4:1. The kinetics
parameters for reaction 2 show that there is a second order
reaction with a low activation energy, and this could explain
that the decomposition of the fraction C
2
occurs in a wide range
of temperatures, as seen in the big separation between the
curves of a
2
and a
4
(Fig. 8).
It must be emphasised that the parameters obtained must be
considered as correlation ones, some of them with physical–
chemical significance. There is a great interrelation between
them, so other sets of parameters could be also valid for the
correlation of the data with a very small increase of the
objective function. In spite of this great interrelation of the
parameters, it must be indicated that the objective function has
been minimized reducing the great interrelation that exits
between the pre-exponential factor, the activation energy and
the reaction order.
The correlation of the three pyrolysis runs and six
combustion runs have been done simultaneously, so 26
parameters were optimized for minimizing the objective
function. No papers have been found in literature where
pyrolysis and combustion runs were optimized simultaneously,
so the correlation work presented in this paper can be
considered as original. As a result, an acceptable set of
parameters is obtained, that correlates the nine TG runs
J. Molto
´
et al. / J. Anal. Appl. Pyrolysis xxx (2005) xxx–xxx 7
+ Models
JAAP 1877 1–8
427
428
429
430
431
432432433
434
435
436
437437438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455455456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
505506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
Fig. 7. Calculated conversion factor (a
i
, a
ij
)at58C min
1
in He:O
2
4:1.
Fig. 8. Calculated mass fraction of volatiles (v
ij
)at58C min
1
in He:O
2
4:1.
UNCORRECTED PROOF
satisfactorily, as can be seen in Figs. 1–3 that show the
conversion degrees versus temperature for a run, observing a
logical variation in accordance with the parameters deduced.
Acknowledgments
Support for this work was provided by the Ministerio de
Educacio
´
n y Ciencia of Spain, research projects PPQ2002-
00567 and PPQ2002-10548-E.
References
[1] J.A. Caballero, R. Font, M.M. Esperanza, J. Anal. Appl. Pyrol. 47 (1998)
165.
[2] P.K. Chatterjee, C.M. Conrad, Text. Res. J. 36 (1966) 487.
[3] D. Dollimore, J.M. Hoath, Thermochim. Acta 45 (1981) 87.
[4] M.J. Antal, G. Va
´
rhegyi, Ind. Eng. Chem. Res. 34 (1995) 703.
[5] S. Vo
¨
lker, Th. Rieckmann, J. Anal. Appl. Pyrol. 62 (2002) 165.
[6] A.A. Faroq, D. Price, G.J. Milnes, Polym. Degrad. Stab. 44 (1994) 323.
[7] J.T. Wanna, J.E. Powell, Thermochim. Acta 226 (1993) 257.
[8] R. Font, A.N. Garcı
´
a, J. Anal. Appl. Pyrol. 35 (1995) 249.
[9] M. Grønli, M.J. Antal, G. Va
´
rhegyi, Ind. Eng. Chem. Res. 38 (1999) 2238.
[10] J.A. Conesa, R. Font, A. Marcilla, J.A. Caballero, J. Anal. Appl. Pyrol. 58–
59 (2001) 619.
[11] J.A. Caballero, R. Font, M.M. Esperanza, J. Anal. Appl. Pyrol. 47 (1998)
165.
[12] J.A. Reyes, J.A. Conesa, A. Marcilla, J. Anal. Appl. Pyrol. 58–59 (2001)
747.
[13] J.A. Conesa, R. Font, Polym. Eng. Sci. 41 (2001) 2137.
[14] J.A. Conesa, R. Font, A. Fullana, J.A. Caballero, Fuel 77 (1998) 1469.
[15] S. Vossoughi, Y.M. El-Shoubary, Thermochim. Acta 157 (1995) 37.
[16] I. Martı
´
n-Gullo
´
n, M. Esperanza, R. Font, J. Anal. Appl. Pyrol. 58–59
(2001) 635.
[17] R. Font, I. Martı
´
n-Gullo
´
n, M. Esperanza, A. Fullana, J. Anal. Appl. Pyrol.
58–59 (2001) 703.
[18] I. Martı
´
n-Gullo
´
n, M.F. Go
´
mez-Rico, A. Fullana, R. Font, J. Anal. Appl.
Pyrol. 68–69 (2003) 645.
[19] J.A. Conesa, J.A. Caballero, A. Marcilla, R. Font, Thermochim. Acta 254
(1995) 175.
[20] J.A. Conesa, A. Marcilla, R. Font, J.A. Caballero, J. Anal. Appl. Pyrol. 36
(1996) 159.
[21] M.J. Antal, G. Va
´
rhegyi, E. Jakab, Ind. Eng. Chem. Res. 37 (1998) 1267.
[22] M.Z. Haji-Sulaiman, M.K. Aroua, J. Inst. Energy 70 (1997) 52.
[23] E. Henrich, S. Bu
¨
rkle, Z.I. Meza-Renken, S. Rumpel, J. Anal. Appl. Pyrol.
9 (1999) 133.
[24] P.L. Walker, F. Rusinko, L.G. Austin, Adv. Catal. 9 (1959) 133.
[25] Z. Du, A.F. Sarofim, J.P. Longwell, in: J. Lahay, P. Ehrburger (Eds.),
Fundamental Issues in Control of Carbon Gasification, 1992, p. 91.
[26] J. Molto
´
, J.A. Conesa, R. Font, I. Martı
´
n-Gullo
´
n, Environ. Sci. Technol. 39
(14) (2005) 5141.
J. Molto
´
et al. / J. Anal. Appl. Pyrolysis xxx (2005) xxx–xxx8
+ Models
JAAP 1877 1–8
525526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
571
    • "Both samples exhibited a similar behavior. As the heating rate increased, there was a displacement of the TG curves to higher temperatures (Moltó et al., 2006). This thermal behavior provided information on the reliability of the thermogravimetric data for the application of the kinetic method -Model Free Kinetics. "
    [Show abstract] [Hide abstract] ABSTRACT: Thermal conversion processes that use coffee husk are an alternative to solve the environmental problems of disposal and waste burning in open places and avoid greenhouse gases emissions in the atmosphere. The present study evaluates in natura coffee husk samples and residues obtained from a combustion process in a Drop Tube Furnace (DTF). Such an evaluation consists in understanding the efficiency of the burning process, therefore the activation energies (E a) of the combustion process for both samples were determined. The isoconventional kinetic method (Model Free Kinetics) was used for the determination of the E a values of the samples. The E a values of the main stages of the combustion process (devolatilization and carbonization) for both samples were compared. Thermogravimetric (TG) and Derivative Thermogravimetric (DTG) curves at five heating rates (10, 15, 20, 25 and 30 ºC min -1) were used for the determination of hemicellulose, cellulose and lignin. SEM images and EDS analysis were applied as complementary techniques in the combustion process. The results show that for both samples the E a values were higher for the carbonization step than for devolatilization. The E a values for the stages of devolatilization and carbonization for the residues were 33 and 15% lower than those for the in natura coffee husk samples. The lower E a values in both steps for the residues are indicative of a reduction in the complexity of the reaction mechanism, which can be a parameter for the evaluation of the biomass combustion process. According to the SEM images, the residues showed exploded surfaces caused by the combustion process, whereas in the in natura samples a denser and robust structure was observed. The ash formed after the combustion process in the thermobalance was also evaluated by SEM and EDS analyses and the results showed a more homogenous structure with tiny particles in comparison with the in natura coffee husk samples. The EDS analysis confirmed the presence of precursor elements in the samples, such as potassium and other inorganic materials, which were intensified after the combustion process.
    Full-text · Conference Paper · Nov 2014 · Thermochimica Acta
    • "However, costs of this technology are still a significant barrier for its commercial application1213141516. The knowledge on the kinetic behavior of the thermal degradation of solid fuels is fundamental for the design of conversion plants with efficient operating systems [6,17]. The determination of the kinetic parameters of the thermal decomposition of the biomass is crucial for an accurate prediction of their behavior under different conditions. "
    [Show abstract] [Hide abstract] ABSTRACT: Isoconventional kinetic model (Model Free Kinetics) was used in this study to determine the activation energies (E a) of the combustion process of five different biomass samples under atmospheres of N2/O2 (conventional combustion) and CO2/O2 (typical oxy-fuel combustion). Thermogravimetric (TG) and Derivative Thermogravimetric (DTG) curves were used to obtain experimental data on the thermal degradation steps of the main constituents (hemicellulose, cellulose and lignin) of the biomasses. Five types of biomass (pine sawdust, sugarcane bagasse, coffee and rice husk, and tucumã seed) widely available in Brazil were studied. The TG experimental conditions were sample mass of 10.0 ± 0.3 mg and heating rates of 10, 15, 20, 25 and 30 °C min -1 , from room temperature up to 700 °C. Two different atmospheres (100 mL min -1) were used: N2/O2 (20% O2) and CO2/O2 (20% O2). The results show that the E a obtained for N2/O2 ranged from 68 to 236 kJ mol -1 for hemicellulose, 119 to 209 kJ mol -1 for cellulose, and 87 to 205 kJ mol -1 for lignin. Such a variation was caused by volatiles release and char formation. Under CO2/O2 atmosphere with 20% O2 , and in comparison with N2/O2 , E a showed decreases of 35% for hemicellulose and 26% for cellulose. However, a 6% increase was observed for lignin. These variations can be understood by differences between CO2 and N2 gas properties.
    Full-text · Conference Paper · Jun 2014 · Thermochimica Acta
    • "Few studies have examined simultaneously the thermal degradation of these materials in inert and oxidizing atmosphere. These works have highlighted the influence of oxygen on the mass loss [9], the gaseous products91011 and the kinetics10111213141516. However, there are still questions about the influence of oxygen on thermal degradation of cellulosic materials and how to model the different steps. "
    [Show abstract] [Hide abstract] ABSTRACT: The kinetics of thermal decomposition of cellulose wadding was investigated from TG–MS experiments. Different oxygen concentrations in the atmosphere and several heating rates were used to study the influence of oxygen concentration on the mass loss of the sample and on the emission of gases. A shift and an amplitude variation of the DTG curves as well as an increase of the emission of gases such as CO and CO2 were observed. Then, a kinetic model was proposed to predict the mass loss of the cellulose wadding. Three stages were considered: the cellulose pyrolysis, the char oxidation and the decomposition of calcium carbonate. For the pyrolysis, the kinetic parameters were expressed according to the partial pressure of oxygen. For the char oxidation, a power law was used to account for the influence of oxygen whereas the other kinetic parameters were considered constant regardless of oxygen concentration. The decomposition of calcium carbonate was modelled by a first order influenced by the pressure of CO2.
    Article · Oct 2011
Show more

  • undefined · undefined
  • undefined · undefined
  • undefined · undefined

Recommended publications

Discover more