Kinetic Modeling of the Polychloroprene Pyrolysis

Under Nitrogen Atmosphere

Yannick Soudais, Cristina Serbanescu, Florent Lemont, Jean-Claude Poussin,
Gheorghe Soare, Grigore Bozga

To cite this version:

Yannick Soudais, Cristina Serbanescu, Florent Lemont, Jean-Claude Poussin, Gheorghe Soare, et al..
Kinetic Modeling of the Polychloroprene Pyrolysis Under Nitrogen Atmosphere. Waste and Biomass
Valorization, 2011, 2 (1), pp.65-76. �10.1007/s12649-010-9054-2�. �hal-01713901�

HAL Id: hal-01713901

https://hal.science/hal-01713901v1
Submitted on 13 Dec 2018

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
Kinetic Modeling of the Polychloroprene Pyrolysis
Under Nitrogen Atmosphere
Yannick Soudais • Cristina Şerbănescu •
Florent Lemont • Jean-Claude Poussin •
Gheorghe Soare • Grigore Bozga

Abstract In the present work it was performed the was found that for the TG-DTA experiments both models
pyrolysis of polychloroprene (produced by Aldrich) in two are appropriate while for the ‘‘Aubry’’ furnace it would be
types of experimental apparatus differentiated by the necessary to use a modified kinetic model which takes into
sample size they could work with. The first experimental account the heat transfer.
device was a TG-DTA analyzer and the second one was a
higher capacity thermo-balance placed inside a furnace (so- Keywords Polychloroprene ! Pyrolysis ! Kinetic
called ‘‘Aubry’’). The experimental conditions were the
following: temperatures up to 1,000!C and atmospheric
pressure (1 bar), atmosphere of flowing nitrogen, three Introduction
different heating rates (5, 10 and 15 K/min for the ‘‘Au-
bry’’ furnace and 10, 20 and 30 K/min for the TG-DTA) The plastic materials from wastes represent a danger for the
and two different quantities of samples (10 g for the first environment because they are not biodegradable. For
set-up and 10 mg for the second one). The mass loss of example, the nature takes 100–1000 years to degrade a
solid samples showed similar time evolutions for both plastic bottle. Among the waste treatment techniques used
devices. It indicated a first period of dehydrochloruration in nowadays (recycling, material valorization, energetic val-
the temperature range 300 - 400!C followed by the pyro- orization, composting) the incineration is the most impor-
lysis of intermediary products in the temperature range tant one for wastes with polymeric structure in Europe.
400–550!C yielding gaseous and liquid volatiles fraction This one is considered more and more inadequate due to
and a black carbonaceous char. Two kinetic models from the toxic nature of the compound released i.e. dioxins or
literature were tested in order to describe the weight loss polycyclic aromatic hydrocarbons. Thus, the pyrolysis, as a
kinetics of polychloroprene during the pyrolysis process: solid waste processing technique started to be in the recent
an empirical one which considered three organic fractions years a real alternative to the incineration of solid wastes
decomposing separately and a more detailed one involving [1]. It is preferred because of its environmental friendlier
191 reactions and 38 species (molecules and radicals). It character and its potential of material and energy savings
(monomers for the chemical industry, combustible gases
and oils, high calorific power char).
Y. Soudais ! C. Şerbănescu (&) ! J.-C. Poussin
Ecole des Mines d’Albi-Carmaux, ‘‘RAPSODEE’’ Center,
Albi, France
Polychloroprene
e-mail: [email protected]

F. Lemont Structure and Properties of Polychloroprene

Comisariat d’Energie Atomique, Albi, France
The polychloroprene is a polymer that can be found in an
G. Soare ! G. Bozga
Faculty of Applied Chemistry and Materials Science, important number of solid wastes (cables insulation,
‘‘Politehnica’’ University of Bucharest, Bucharest, Romania gloves, transmission belts etc.). Nowadays it is obtained by
 Cl H2 H2 H2 neoprene, i.e. a polychloroprene with predominant 1,4-
C * * C C *
1 C n
n trans structure, the dehydrochlorination step is assumed to
1 4
* C C 4 be initiated at the structural defects of the polymer chain.
C C
H2 H Cl H These defects are represented by the 1,2 units which are
a b less stable due to the presence of the most labile chlorine
i.e. chlorine bonded to a tertiary carbon [7]. It is important
H2 Cl H H2
to mention that the hydrogen chloride evolved from this
* C C * * C C *
n n degradation step has not a catalysis role [11].
1 2 3 4
The second degradation step (temperatures higher than
CH C Cl 450!C) is due to condensation reactions and other chemical
transformations i.e. the decomposition and dealkylation of
the condensed structures.
CH2 CH 2 The quantities of products obtained expressed in weight
c
d percentage relative to the sample are usually 20% for the
‘‘char’’ (solid residue) [5–8] and around 35% for the
Fig. 1 Polychloroprene structures: a 1–4 trans, b 1–4 cis, c 1–2, d 3–4 hydrogen chloride [5]. From the FT-IR spectrum of a char
obtained from polychloroprene pyrolysis at 400!C in a TG-
DTA analyzer it was found that, on average, it has a
radical emulsion polymerization of chloroprene (2-chlor-
structure of tri-substituted phenyl rings but also 12 or more
1,3-butadiene) [2] when a stereospecific 1–4 structure is
double bonds in conjugation [12].
obtained. However, other different structures are to be
found in its final structure (see Fig. 1) [2–4]. Generally
Polychloroprene Pyrolysis Kinetic Models
speaking, polychloroprene (CR-‘‘chloroprene rubber’’,
Neoprene) is and will remain an important elastomer due to
There is a broad literature regarding the mathematical
a favorable combination of technical properties.
modeling of polymer pyrolysis, in order to predict the time
dependence of solid weight as well as the nature of solid
Polychloroprene Pyrolysis and gaseous products, see for example [13]. Usually the
mathematical models were obtained by fitting the experi-
Generally, the polychloroprene pyrolysis was studied by mental data with different kinetic expressions, more or less
the thermogravimetric technique using different types of empirical [14–16].
experiments on TG-DTA analyzers: only dynamic runs [5– The parameter estimation of a mathematical model for a
7], dynamic and isothermal runs [7] or only isothermal runs given process aims to obtain in each experimental point,
[5] under nitrogen, He [6] or a mixture of air and nitrogen values for the dependent variable (y) as near as possible the
[7]. The mass sample was very small: 4–5 mg [6, 7], 1.5 g experimental ones (yobs) [17]. In this purpose, the estimator
[8] and even ca. 10 lg [9]. The samples prepared in most commonly used in literature is the least squares
presence of air are less stable than the ones prepared in its method, for example in [6].
absence [10]. For the experiments made the authors always Due to the particularities in the working conditions, a
neglected the heat transfer. comparison between the results published by different
There are two main degradation steps during polychlo- scientists is difficult these being highly dependent on the
roprene pyrolysis: firstly a dehydrochlorination between experimental conditions used in each case.
200 and 350!C [5–9] and secondly the depolymerization.
For the commercial polychloroprene there are three mass
loss intervals during its thermal degradation, the first one Experimental study
corresponding to the additive degradation [6, 7].
The dehydrochlorination step is considered to follow a Materials
non radical intramolecular mechanism [8]. There are three
steps mentioned in literature for the dehydrochlorination The material used was pure Polychloroprene (Sigma–
stage. An explanation given is that this phenomenon is due Aldrich) with a structure of 85% trans and 10% cis. The
to different monomer structures existing within the poly- rest of 5% is of an unknown structure, as the producer
mer chain. Two mechanisms may be responsible for this affirms. The product is typically delivered as ‘‘chips’’ with
process [9]: an intramolecular reaction with re-arrange- a density of about 1.23 g/cm3. It is considered to be a pure
ments or a reaction of the active zones of the polymer polychloroprene i.e. without additives included. It is a
which don’t participate 100% to the process. For the white material with elastic properties.
Table 1 Elemental analysis of pure polychloroprene from ‘‘Sigma- a
Aldrich’’
Composition %Weight %Weight %Weight 155
measured calculated literature [18] mm
Ø210 mm
C 54.5 ± 10.1 54.2 54.0
H 6.2 ± 0.6 5.7 5.8
S 0 0 0
N 0.038 ± 0.09 0 0
Autres (O, Cl) 39.3* 40.1 39.8
*100

100
mm
Total 100 100
* By difference

230
mm
mm
Its chemical composition was determined by the com- 40

40
mm

mm
plementary technique of elemental analysis. It was deter-

85
mined by this technique the volume percentage of C, H, S
and N with a ‘‘NA 2100 Protein’’ instrument (‘‘Thermo
Finnigan’’). The results are shown in Table 1. As it can be
seen the measured percentages are in very good agreement
with calculated ones and literature [18].
The gaseous reagent was pure nitrogen (0.8 wt% O2).

Experimental Apparatus

The thermogravimetric experiments were performed in two

apparatus: (a) a TG-DTA analyzer (‘‘Setaram’’, model
‘‘ATG-ATD 92’’) coupled, in some of the runs, to a FT-IR b
(‘‘Perkin Elmer’’, model ‘‘System 2000’’) with role in
monitoring the gases evolved and (b) a larger scale set-up
based on a thermobalance having the plateau mounted inside
an electrically heated furnace called ‘‘Aubry furnace’’. Both
devices were equipped with a thermobalance as an essential
element for this study. In neither cases the material of the Fig. 2 The ‘‘Aubry’’ furnace’s chart: a schema of elements of the
furnace: 1-electrical resistance, 2-platinum thermocouple, 3-charger
crucible didn’t catalyze nor react with the sample.
of the balance, 4-chimney, 5-balance, 6-crucible, b furnace nitrogen
The TG-DTA analyzer was a laboratory-scale furnace entries: 1-furnace, 2-balance
provided with a graphite resistance. The instrument had the
following characteristics: a cylindrical vertical furnace of limit of detection) and a capacity of 8 l. The thermobalance
high temperature (1,750!C), 30 K/min maximum heating was placed in the lower part of this device. The nitrogen
rate, an electronic microbalance (20 mg maximum weight entered the furnace by two openings, one feeding the gas
and 1 lg the limit of detection), possibility of coupling inside the furnace and the other keeping an inert atmo-
with an FT-IR analyzer by a heated line (200!C) and all the sphere around the arm of the balance. Unfortunately, even
logistic material necessary for registering the experimental with these precautions, there was a small amount of O2
results. The furnace was cooled by a water circuit. The (*4% vol) present in the gaseous mixture inside the
nitrogen entered firstly into the microbalance and after- reaction chamber, due to imperfect sealing of the opening
wards into the furnace in order to assure an inert atmo- around balance arm. For this reason further oxidation of the
sphere necessary for the balance. The crucibles used were residue is to be expected with this device.
platinum made ones of small capacity (0.1 ml).
The ‘‘Aubry’’ furnace was a pilot-scale furnace provided Experimental Conditions
with Super Kanthal electrical resistances. The instrument
had the following characteristics (see Fig. 2): a cylindrical The experimental conditions were differentiated by the
vertical furnace of high temperature (1,600!C), 34 K/min device’s scale as can be seen in Table 2. The experiments
maximum heating rate at low temperatures (1,100!C), a carried out were dynamic ones from room’s temperature to
thermobalance (6,100 g maximum weight and 0.1 g the 1,000!C in both devices. The sample was previously cut
Table 2 Experimental conditions for polychloroprene pyrolysis in expected, an increase in the heating rates produces a shift
our work of the decomposition curve toward higher temperatures
Experimental conditions Aubry furnace TG-TDA [19] and a decrease of the quantity of char obtained [8].
This observation is true for the results obtained on both
Initial mass of sample 10 g 10 mg
devices. It can be also observed that the experimental curve
Temperature interval (!C) 25–1000 25–1000 obtained in the ‘‘Aubry’’ furnace is rather different from
Nitrogen flow 10 l/min 50 ml/min the one obtained in the TG-DTA analyzer. A possible
Heating rate (K/min) 5, 10, 15 10, 20, 30 explanation may be the heat transfer limitation and the
higher error in the measurement of working temperature.
into powder. This represents a dimension small enough to Thus the experimental results in the ‘‘Aubry’’ furnace
assume that any thermal effect could be neglected in the were affected by noise. Figures 5 show the results of fil-
case of the TG-DTA analyzer experiments. tering the experimental results of the mass loss and of
The temperature was not the theoretical one which was temperature by the Savitzky-Golay algorithm (the filtered
assigned to the thermobalance but a temperature recorded value, at any time, represents a fitting of certain number of
in different manners for the two cases considered: (a) for points placed in the vicinity of a given point).
the TG-DTA it was measured just under the crucible with The TG analysis of the pure polychloroprene pyrolysis
sample and (b) for the ‘‘Aubry furnace’’ it was measured gives a curve with three main mass loss intervals for the
with a thermocouple placed at about 80 mm above the results obtained in the TG-DTA analyzer and with four
crucible with sample. main mass loss intervals for the ‘‘Aubry’’ furnace (see
The buoyancy effect of the carrier gas in each thermo- Fig. 4a, b). This is in agreement with literature which
balance was eliminated by subtracting the blanks for each mentions a mass loss of two stages [5–9] and three stages
run. The experimental data presented in this paper for the only for the dehydrochlorination [9]. The derivatives of
‘‘Aubry’’ furnace are usually the mean values for runs experimentally measured thermogravimetric (DTG) curves
performed twice and the results obtained for the duplicates are presented in Fig. 4a. As it can be seen, an increase in
were very similar in all the cases (see Fig. 3). A maximum the heating rate produces a modification in the DTG curve.
error of 2% was estimated in the measurements performed This is the reason why at 10 K/min the last two processes
in the case of highest heating rates used in the ‘‘Aubry’’ (B and C) are well differentiated, at 20 K/min there is a
furnace. For the TG-DTA analyzer the reproducibility [14] partial superposition of these decomposition steps and
was very good so only one run was performed each time. finally at 30 K/min there is nearly only one decomposition
process that appears. Also it can be seen that the first
Results and Discussions degradation step doesn’t give a DTG peak. Table 3 gives
all the information read on the TG and DTG curves. From
Figure 4 show the variation of experimental weight loss the DTG ones it can be seen that the mass loss that occurs
with temperature. The curves are obtained at different during the second decomposition step (Process B) is larger
heating rates for runs performed in both devices. As and faster than the ones observed during the other steps

Fig. 3 Duplicated runs in the 1

experience 1
‘‘Aubry’’ furnace for 10 K/min experience 2
0.9

0.8

0.7
Mass fraction (-)

0.6

0.5

0.4

0.3

0.2

0.1

0
0 100 200 300 400 500 600 700 800 900 1000
Sample temperature (°C)
Fig. 4 Experimental results a 1
(our work) for polychloroprene 10 K/min A
DTA
pyrolysis: a TG-DTA analyzer 0.9 20 K/min 40
at 10, 20 and 30 K/min , 30 K/min B
30
b ‘‘Aubry’’ furnace at 5, 10 and 0.8

DTA signal (µV)

15 K/min, c comparison 20
B exo
0.7
between experimental results 10
C

Mass fraction (-)

obtained in the two types of 0.6 A
0
devices
DTG
-10
0.5 0

Mass derivative (%/min)

A 100 200 300 400 500 600
Temperature (°C)
0.4 -5
C
C
0.3 -10

-15
0.2
B
-20
0.1 100 200 300 400 500 600 700 800
Temperature (°C)

0 100 200 300 400 500 600 700 800

Temperature (°C)

b
1 5 K/min
A 10 K/min
15 K/min
0.9

0.8

0.7
B
Mass fraction (-)

0.6

0.5

0.4

0.3 C

0.2

0.1 D

0 200 400 600 800 1000 1200

Temperature (°C)

c 1
10 K/min, Aubry furnace
10 K/min, TG-ATD
0.9
A
0.8

0.7
Mass fraction (-)

0.6

0.5

0.4
B
0.3

0.2
C
0.1
D

100 200 300 400 500 600 700 800 900 1000
Temperature (°C)
Fig. 5 Results of filtration of
the experimental results for the
a 0
experimental curve
runs effectuated in the ‘‘Aubry’’
-0.05 filtered curve
furnace at 10 K/min, the first
mass loss for process A:
a filtration of experimental data -0.1
for the mass loss, b filtration of
experimental data for the -0.15

Mass loss (g)

temperature
-0.2

-0.25

-0.3

-0.35

-0.4
0 5 10 15 20 25 30 35
Time (min)

b 350
experimental curve

300 filtered curve

250
Furnace temperature (°C)

200

150

100

50

-50
0 5 10 15 20 25 30 35
Time (min)

(Processes A and C) [9]. At higher heating rates, the second ðC4 H5 ClÞn ! 0:1ðC4 H4 ClÞn þ 0:9nHCl þ 0:9ðC4 H4 Þn
mass loss step is slower and the third one becomes almost ð1Þ
invisible. The DTA curves confirm the existence of double
bonds in the material’s structure which are responsible for This equation considers that 90% of the chlorine atoms
the exothermic effects observed (Processes B and C) [6]. found in the initial sample are lost as HCl during this step
The second mass loss corresponds also to the chlorine [8, 10, 11].
loss (dehydrochlorination). As already mentioned above, This dehydrochlorination process corresponds to a mass
the polychloroprene has a good thermal stability, beginning loss calculated by the following relation: Dmð%Þ ¼
to decompose around 200–300!C. This result is evidenced 100M0:9monomere
MHCl
¼ 38: The result is in agreement with the with
from the FT-IR spectrums presented in Fig. 6 which give
literature (35–37%) [5] and slightly inferior with our
the intensities of evolved gases during polychloroprene
results (43% in the TG-DTA for 10 K/min). The temper-
pyrolysis. The HCl spectrum is delimited between 2,500
ature ranges of the dehydrochlorination step specified in
and 3,100 cm-1 which are representative frequency values
literature are the same as those in our work [5, 7–9].
for this gaseous product. The process is delimited between
The amount of char obtained in the TG-DTA analyzer is
300 and 400!C but it continues until a temperature of
higher than the one obtained in the ‘‘Aubry’’ furnace (see
500!C in parallel with the evolution of other gases.
Table 3). This phenomenon is due to the presence of
The stoichiometry of the first degradation step can be
oxygen in the furnace which consumes the carbonaceous
described by the Eq. 1 below:
Table 3 Polychloroprene pyrolysis results
Instrument Heating rate Degradation Ti (!C) Tmax (!C) Tf (!C) DTGmax Dm (wt %) Dmchar
(K/min) step (%/min) (wt %)

TG-DTA 10 Process A 190 – 304 – 5.0 18.0

Process B 304 370 404 -8.0 43.0
Process C 404 455 574 -3.9 33.0
20 Process A 245 – 325 – 5.0 20.0
Process B 325 380 426 -13.3 48.0
Process C 426 467 588 -6.1 27.0
30 Process A 265 – 345 – 5.0 20.0
Process B 345 389 436 -16.9 49.0
Process C 436 483 611 -6.9 26.0
‘‘Aubry’’ 5 Process A 182 – 318 – 5.0 0
furnace* Process B 318 – 365 – 50.0
Process C 432 – 622 – 24.0
Process D 622 – 1000 – 14.9
10 Process A 194 – 354 – 5.0 4.5
Process B 354 – 380 – 44.0
Process C 458 – 700 – 24.0
Process D 700 – 1000 – 7.5
15 Process A 302 – 359 – 3.0 9.0
Process B 359 – 409 – 51.0
Process C 450 – 678 – 21.0
Process D 678 – 1000 – 10.4
* Total degradation for this device

residue formed during the pyrolysis process. Also there is a this model are considered to be due to the
difference between the amounts of chars obtained at dif- dehydrochlorination.
ferent heating rates for the same experimental device: for The model represents the time dependence of the mass
both devices an increase in the heating rate produces a loss by the power law expressions:
higher amount of carbonaceous char. dFi
¼ &ki ! Fini i ¼ 1; 2; 3 ð2Þ
dt

Mathematical Modeling of the Process t ¼ 0; Fi ¼ wi0

w10 þ w20 þ w30 þ I ¼ 1 ð3Þ
Empirical Kinetic Model
The temperature dependences of reaction rate constants,
A simplified scheme of the pyrolysis process was proposed ki, are considered of Arrhenius type:
! " # ! "$
by Caballero et al. [6] considering three organic fractions Ei Ei 1 1
ki ðTÞ ¼ k0i exp & ¼ kr;i exp & & ;
in the structure of the material, with different kinetic par- RT R T Tref
ticularities, as well as an inert fraction (I). Each organic kr;i ¼ ki ðTref Þ
fraction (F) decomposes into volatiles (G) as it can be seen ð4Þ
in Scheme 1.
In a thermogravimetric experiment it is impossible to The last expression is often used in order to diminish the
differentiate between the non-decomposed fraction of the correlation that exists between the estimated values of Ai
initial product and the solid residue (‘‘char’’) formed. Such and Ei. In this expression, Tref is usually taken as the
an experiment gives information only about the global average of the working temperature interval.
weight loss. For this reason we used the same approxi- The mathematical calculations have been done using the
mation as Caballero and al i.e. we assumed that the inert, software Matlab 7.0 (the functions lsqcurvefit and ode).
‘‘I’’, designates the inorganic fraction of the initial polymer The pyrolysis simulation results obtained with the kinetic
and the solid residue. The first two mass losses assumed by parameters given by Caballero et al., do not agree
Fig. 6 The FTIR spectrums of
HCl in gaseous form: a evolved
gas spectrum in our work
showing the beginning of the
formation of HCl for a pyrolysis
at 30 K/min (316!C), b evolved
gas spectrum in our work
showing the end of the
formation of HCl for a pyrolysis
at 30 K/min (411!C),
c spectrum of HCl
Scheme 1 F organic fraction of
the initial polymer, G gases
Fi Gi , i=1,2,3 the atomic balance calculations involve all these 38 spe-
cies. It is important to mention, as the authors also do [5],
formed, I inert fraction of the
initial polymer [6] that the polymer rest called ‘‘P’’ has no significance for the
I I atomic balance. This rest has only the purpose of an
indication of the phase; more precisely the presence of
‘‘P’’ indicates melt phase specie while its absence indi-
satisfactorily with our experimental data. This can be cates gas species. The initial specie is the monomer
explained by the differences existing in the working con- formed in the first reaction (R1), the chloroprene radical
ditions and the initial material composition. Consequently _ ¼ CClCÞ & P: Other hypothesis for this model are as
ðCC
we fitted the parameters of the same model to our experi- follows: the melting mass decreases continuously while
mental results. gases are formed, the solid residue (‘‘char’’) and the inert
There were three parameters to be optimized for each fractions remain in the reaction mixture, the reactions are
fraction of solid: the pre-exponential factor, the activation considered elementary (the reaction orders equal to
energy and the reaction order (9 parameters in total). molecularities) and the reaction rates are expressed by
Besides the nine parameters model presented above, we power law expressions. The time dependencies of con-
also tested simplified versions, considering a lower centrations (mass fractions) in the melt mixture are given
number of solid fractions or, correspondingly, a smaller by the mass balance equations:
number of parameters. The results for the data obtained dwj Mj
using TG-DTA analyzer at 10 K/min are the best obtained ¼ ! vRj ; ½s&1 (; j ¼ 1. . .38 ð5Þ
dt qam
and are shown in Fig. 7. As it can be seen the nine
m ðsÞ
parameter model with three fractions decomposing where: wj ¼ mj 0j ; the mass fraction [adim]; mj (s) = the
simultaneously is in good agreement with the experi- mass of the specie j at moment s and m0j = the initial mass
mental data. The results, presented in Table 4 show that of the specie j; Mj = the molecular mass of the specie j;
for the case of 10 K/min the correlation with literature qam = the density of liquid polychloroprene (considered
values is good for the case of nine parameters, or three 1,340 g/ml).
P
fractions decomposing simultaneously. For the case of vRj ¼ 191i¼1 ðtij ! VRi Þ; the reaction velocity [mole/l s]
two fractions, the results are not good, seen the value of for every specie j = 1…38 of the system of i = 1…191
the reaction order (1.6 9 10-14). For the case of one reactions; tij = the stoechiometric matrix of the reaction
fraction the value founded is not very good, seen the system.
value of the reaction order which is twice bigger than the Eatt
i n
0

literature. VRi ¼ k0i T bi e& R T wj i ; the rate of reaction i, [mole/l s];

In the case of the ‘‘Aubry’’ furnace the results are not k0i = the pre exponential factor [s-1], Eatti = the activa-
very good even after optimizing the model. They were not tion energy [J/mole]; n0i = the reaction order [adim];
given because they are very wrong. An explanation may be bi = a temperature coefficient [adim];
due to the heat transfer that cannot be neglected for this T ¼ T0 þ b ! t; the temperature [!C]; b = the heating
experimental device. rate [K/s], t = the time [s].
The values of the pre exponential factor Ai, the activa-
tion energy Eatti and of the coefficient bi are those given in
Radical Kinetic Model
the article of Faravelli, Appendix B [5].
Nevertheless, the kinetic model of Faravelli was mod-
The true reaction mechanism of the pyrolysis process is
ified in order to obtain correct mass balances. The coef-
involving formation and consumption of organic radical
ficients for all the 191 equations of the proposed model
species. Therefore, a more realistic process model should
were recalculated in order to fulfill the atomic balances on
explain the transformation of the initial polymeric struc-
each reaction. Also there was a specie that could not be
ture to the final products, by the intermediate of radical
found in the equations of Appendix B and that was
reactions. A kinetic model of this type was proposed by
replaced by another which was present in these equations
Faravelli et al. [5]. Accordingly, as seen in the Appendixes
but absent in the list of Appendix A. It is the case of the
A and B of this article [5], the thermal decomposition of
the polychloroprene takes place by 191 elemental reac- specie P & ðC_ 10 H8 ClÞ & P from Appendix A that was
tions that involve 38 species (molecules and radicals). The replaced by the specie P & ðC_ 8 H8 ClÞ & P found in
chemical species are polymer segments, placed between Appendix B. The mathematical treatment consisted in
the brackets, and also reaction products having well fitting the experimental data to the model by the Matlab
known chemical structures. The reactions considered in function ‘lsqcurvefit’. The results for TG-DTA analyzer
Fig. 7 Results of the
mathematical modeling with the
a 1

empirical kinetic model for runs 0.9

in the TG-DTA analyzer at 10
K/min: a considering one 0.8
fraction and three optimization
parameters, b considering two 0.7

Mass fraction (-)

fractions and six optimization
0.6
parameters, c considering three experimental curve

fractions and nine optimization 0.5 simulated curve

parameters
0.4

0.3

0.2

0.1

0 100 200 300 400 500 600

Temperature (°C)

b 1

0.9

0.8

0.7
Mass fraction (-)

0.6 experimental curve

simulated curve
0.5

0.4

0.3

0.2

0.1

100 200 300 400 500 600

Temperature (°C)

c 1

0.9

0.8

0.7
Mass fraction (-)

0.6

0.5 experimental curve

simulated curve
0.4

0.3

0.2

0.1

100 200 300 400 500 600

Temperature (°C)
Table 4 Optimized values for the kinetic parameters of the empirical model at 10 K/min (the three fractions model)

No Paramètre cinétique F.O.

fractions
A1 (min-1) E1 n1 A 2 (min-1) E2 n2 A 3 (min-1) E3 n3
(kJ.mol-1) (kJ.mol-1) (kJ.mol-1)
1 2,4 × 1013 160,0 4,3 - - - - - - 0,309
2 1,7 × 107 97,2 1,6× 10-14 3,9 × 1012 152,4 4,8 - - - 3,315 × 10-2
0,286
3 3,4 × 1020 99,1 3,8 3,2 × 107 37,0 0,3 2,7 × 1010 153,3 0,9 3,52 × 10-3
0,309
0,792

experiments at 10 K/min are presented in the Fig. 8. As it Conclusions

can be seen this model offers a good correlation with
experimental values but not better than the three fractions Pure neoprene pyrolysis under nitrogen atmosphere was
one presented above. realized in two experimental devices differentiated by the
capacity of sample that they could work with. The present
study approached only the weight loss kinetics of the solid
Comparison Between the Two Models Considered during polychloroprene pyrolysis. For the TG-DTA device
there were two literature kinetic models tested. It was
The three fraction model is a very versatile one which found that the model with three fractions which decompose
reproduces well the runs performed at different heating independently has the advantage of simplicity but the dis-
rates and even isothermal runs [5]. It gives a good repre- advantage of a great generality while the second one offers
sentation for the mass loss. better qualitative results but his major disadvantage is his
The radical model is also good and versatile in great complexity. An optimal kinetic model will be a
describing the char and gases formation. It offers a more combination of the two-ones tested in this work. For the
detailed model for the process. Thus offers a better case of the ‘‘Aubry’’ furnace it was found that the kinetic
explanation of the insights of this pyrolysis process. models where not adequate. For this device it is necessary
The first kinetic model has the advantage of simplicity to take into account also the heat transfer phenomenon and
but the disadvantage of great generality. The second one the presence of oxygen. Further studies are necessary in
offers better qualitative results but his major disadvantage order to identify the physico-chemical mechanism of the
is his great complexity. pyrolysis process and to improve the kinetic model struc-
An optimal kinetic model will be a combination ture. More detailed experimental runs could help in
between the two-ones proposed. This more appropriate founding a more precise kinetic model especially if there
model was not an objective of the present work. could be possible to monitor the evolved gas evolution.

Fig. 8 Results of simulation for 1.1

the thermal degradation of experimental curve
polychloroprene in the TG-DTA 1
simulated curve
analyzer at 10 K/min (the
0.9
radical kinetic model)
0.8
Mass fraction (-)

0.7

0.6

0.5

0.4

0.3

0.2

0.1
0 100 200 300 400 500 600 700 800
Temperature (°C)
References 10. Gardner, G.L., McNeill, I.C.: The thermal degradation of poly-
chloroprene-I. Thermal analysis studies of the stability of poly-
1. http://www.techniques-ingenieur.fr/dossier/dechets_et_risques_ chloroprene samples and measurements of the kinetics of
pour_la_sante/G2450 degradation. Eur Polym Journal 7(6), 569–591 (1971)
2. http://www.iisrp.com/WebPolymers/04FinalPolychloropreneIISRP. 11. Jaroszynska, D., Kleps, T., Gdowska-Tutak, D.: Investigation of
pdf thermal degradation of polymers containing chlorine by ther-
3. Encyclopedia of polymer science and technology. Wiley, USA mogravimetry. J. Therm. Anal 19, 69–78 (1980)
(2005) 12. Gardner, D. L., McNeill, I. C.: The thermal degradation of
4. Locatelli, Jean-Louis, ‘‘Techniques de l’ingénieur’’, traite de polychloroprene—II study of the products of degradation Eur.
«génie des procédés», J6530 Polymer J. (1971). doi:10.1016/0014-3057(71)90089-9
5. Mehl, M., Marongiu, A., Faravelli, T., Bozzano, G., Dente, M., 13. Caballero, J.A., Conesa, J.A.: Mathematical considerations for
Ranzi, E.: A kinetic modeling study of the thermal degradation of nonisothermal kinetics in thermal decomposition. J. Anal. Appl.
halogenated polymers. J. Anal. Appl. Pyrolysis 72, 253–272 Pyrolysis 73, 85–100 (2005)
(2004) 14. Blazek, J.: Study of the reaction kinetics of the thermal degra-
6. Caballero, J.A., Conesa, J.A., Martin-Gullon, I., Font, R.: Kinetic dation of polymers. Doctoral thesis, EMAC, France, Sept 2005
study of the pyrolysis of neoprene. J. Anal. Appl. Pyrolysis 74, 15. Maciejewski, M.: Computational aspects of kinetic analysis. Part
231–237 (2005) B: The ICTAC kinetics project-the decomposition kinetics of
7. Aracil, I., Font, R., Conesa, J.A., Fullana, A.: TG-MS analysis of calcium carbonate revisited, or some tips on survival in the
the thermo-oxidative decomposition of polychloroprene. J. Anal. kinetics minefield. Thermochim. Acta 355, 145–154 (2000)
Appl. Pyrolysis 79, 327–336 (2007) 16. Wendlandt Wesley, W.M.: Thermal analysis. Wiley, USA (1986)
8. Dick, C.M., Liggat, J.J., Snape, C.E.: Solid state 13C NMR study 17. Maria, Gh.: «Metode de estimare a parametrilor modelelor ma-
of the char forming processes in polychloroprene. Polym. Degrad. tematice ale proceselor chimice», course of University «Po-
Stab. 74, 397–405 (2001) litehnica» from Bucharest, Chemical Engineering Faculty (1999)
9. Lehrle, R.S., Dadvand, N., Parsons, I.W., Rollinson, M., Horn, 18. Kameda, T., Watabe, Y., Grause, G., Yoshioka, T.: Dehydro-
I.M., Skinner, A.R.: Pyrolysis-GC-MS used to study the thermal chlorination behaviour of polychloroprene during thermal deg-
degradation of polymers containing chlorine. III. Kinetics and radation. Thermochim. Acta 476, 28–32 (2008)
mechanisms of polychloroprene pyrolysis. Selected ion current 19. Wendlandt, W.W.: Thermal analysis, 3rd edn. Wiley, New York
plots used to evaluate rate constants for the evolution of HCl and (1986)
other degradation products. Polym. Degrad. Stab. 70, 395–407
(2000)