Acid-catalyzed conversion of chlorinated plastic waste into valuable

hydrocarbons over post-use commercial FCC catalysts

Y.-H. Lin
a,
*, M.-H. Yang
a
, T.-T. Wei
a
, C.-T. Hsu
a
, K.-J. Wu
a
, S.-L. Lee
b
a
Department of Chemical and Biochemical Engineering, Kao Yuan University, 821 Kaohsiung, Taiwan, ROC
b
Department of Chemistry, Chinese Military Academy, 830 Kaohsiung, Taiwan, ROC
1. Introduction
Although plastic waste can be regarded as cheap source of
chemicals and energy, the recycling of mixed plastic wastes
containing polyvinyl chloride not only results in the formation of
chloro-organic compounds in volatile products but also causes
serious emission pollution in their applications [1]. It is undesirable
to dispose waste plastics by landll due to high costs and poor
biodegradability. Also, the destruction of plastic wastes by
incineration is prevalent, but is expensive and often generates
problems with unacceptable emissions. Therefore, the recycling of
plastic waste is important in the conservation of resources and the
environment [2]. The production of liquid hydrocarbons from
polymer degradation would be benecial in that liquids are easily
stored, handled and transported. However, these aims are not easy
to achieve. An alternative strategy is that of chemical recycling,
known as feedstock recycling or tertiary recycling, which has
attracted much interest recently with the aim of converting waste
plastics into basic petrochemicals to be used as chemical feedstock
or fuels for a variety of downstream processes. Two main chemical
recycling routes are the thermal and catalytic degradations of waste
plastics [3]. Thermal crackingor pyrolysis is awell-knowntechnique
andis oftenusedinpetrochemical processing. The pyrolysis of waste
plastics is thethermal decompositionintheabsence of oxygenandis
carried out in vessels, shaft kilns, autoclaves, rotary kilns, screw
conveyors or uidized beds [46]. However, the thermal degrada-
tion of polymers to low molecular weight materials has a major
drawback inthat a verybroadproduct range is obtained. Inaddition,
these processes require high temperatures typically more than
500 8C and even up to 900 8C. These facts strongly limit their
applicability and especially increase the higher cost of feedstock
recycling for waste plastic treatment. Therefore, catalytic degrada-
tion provides a means to address these problems. The use of catalyst
is expected to reduce reaction temperature, to promote decom-
position speed, and to modify the products [79].
The catalytic degradation of polymeric materials has been
reported for a range of model catalysts centered on the active
components in a range of different catalysts, including zeolite-
based such as ZSM-5, BEA, USY, MOR, and modied nanocrystalline
of Y, and ZSM-5 [1015], amorphous silicaalumina (SAHA) and
the family of mesoporous MCM materials [1620]. However, these
catalysts have been used that even if performing well, they can be
J. Anal. Appl. Pyrolysis 87 (2010) 154162
A R T I C L E I N F O
Article history:
Received 20 March 2009
Accepted 6 November 2009
Available online 13 November 2009
Keywords:
Plastic waste
Pyrolysis
Catalyst
Selectivity
A B S T R A C T
Plastic waste can be regarded as potentially cheap source of chemicals and energy. A mixture of post-
consumer plastic waste (HDPE/LDPE/PP/PVC) was pyrolyzed over cracking catalysts using a uidizing
reaction system similar to the FCC process operating isothermally at ambient pressure. Experiments
carried out with various catalysts gave good yields of valuable hydrocarbons with differing selectivity in
the nal products depending on reaction conditions. A newly developed model is presented that
combines kinetic and mechanistic considerations which take into account chemical reactions and
catalyst deactivation in the modeling of the catalytic degradation of polymers. Though acid-catalyzed
hydrocarbon cracking reactions involve a large number of compounds, reactions and catalyst
deactivation and are very complex, the model gives a good representation of experimental results
from the degradation of post-consumer plastic waste. This model provides the benets of lumping
product selectivity, in each reaction step, in relation to the performance of the catalyst used and particle
size selected as well as the effect of operation conditions, such as rate of uidizing gas and reaction
temperature. The results of this study are useful for the utilization of post-use commercial FCC catalysts
for the production of chemicals and hydrocarbons and for determining both the product yield and
product distribution from PVC-containing plastic degradation, and specically for leading to a cheaper
process with more valuable products.
2009 Elsevier B.V. All rights reserved.
* Corresponding author. Tel.: +886 7 6077777; fax: +886 7 6077788.
E-mail address: lin@cc.kyu.edu.tw (Y.-H. Lin).
Contents lists available at ScienceDirect
Journal of Analytical and Applied Pyrolysis
j our nal homepage: www. el sevi er . com/ l ocat e/ j aap
0165-2370/$ see front matter 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.jaap.2009.11.006
unfeasible from the point of viewof practical use due to the cost of
manufacturing and the high sensitivity of the process to the cost of
the catalyst. An economical improvement of processing the
recycling via catalytic cracking would operate in mixing the
plastic waste with uid catalytic cracking (FCC) commercial
catalysts. In typical commercial application for renery, the ZSM-5
zeolite crystals are mixed with an amorphous matrix material such
as silicaalumina, which confers on the nished utilization of the
physical properties required for an FCC catalyst. Ultrastable Y
(USY) with lower Y framework alumina contents and hence lower
unit cell sizes, tend to produce more light alkenes, mainly at the
expense of gasoline. A more interesting is that of adding plastic
waste into the FCC process, under suitable process conditions with
the use of zero value of spent FCC catalysts [2125] performed on a
lab-scale cracking, a large number of polymers can be economic-
ally converted into valuable hydrocarbons. Potential concepts have
been investigated in our group using a catalytic uidized-bed
reactor to study the product distribution and selectivity of catalytic
degradation of several different textiles of post-consumer plastic
wastes without PVC mixture previously [2628]. In these studies,
the catalysts increase signicantly the commercial potential of a
recycling process based on catalytic degradation, as cracking
catalysts could cope with the conversion of plastic waste co-fed
into a renery FCC unit.
In addition, the type of reactor plays a key role, because of the
lowthermal conductivity and high viscosity of plastics, which may
lead to the appearance of mass- and heat-transfer constraints. This
factor inuences the product distribution along with the operation
conditions. For this purpose, a laboratory uidized-bed reactor has
been used to study catalytic cracking of polymers by limiting the
contact between primary volatile products and the catalyst/plastic
mixture. Though the results of these studies provide the
improvement of the catalyst utilization and reactor performance,
there is still a need to develop kinetic models to describe the
experimental results and to facilitate the further development of a
process to industrial scale. Typically, kinetic and mechanistic
considerations of catalytic degradation of plastics have largely
been focused on non-catalytic degradation (thermal cracking)
rather than on catalytic cracking reactions. The literatures
available are shared with the study of other complex reaction
schemes such as catalytic cracking and reforming [29]. Studies on
plastic pyrolysis by ionic and free radical mechanisms using
microporosity of porous solid acids have been proposed [3033].
Reactions occurred by carbonium ion chemistry include different
steps, such as H-transfer, chain-beta-scission, oligomerization,
isomerization, aromatization and alkylation, are inuenced by
acid-site strength, density and distribution [30]. Kinetic studies of
catalytic degradation of polyolens using TGA as a potential
method for screening catalysts have been reported previously [31].
The catalytic degradation of polymers over cracking catalysts
follows complex routes that cannot be described by one or more
chemical reactions, but still rather imperfectly described by either
empirical formulas featuring fractional stoichiometric coefcients
or compressive systems of elementary reaction [32]. Solid acid
catalysts such as zeolites favor hydrogen-transfer reactions due to
the presence of many acid sites. During degradation, primary and
secondary alkyl radicals are formed, and by hydrogen abstractions
and recombination of radical units, olens and monomers are
produced [33]. A more difcult task is recycling of commingled
post-consumer plastic waste since it consists of not only
hydrocarbons but also chloride-containing mixed plastics as well
as some modied materials. The catalytic recycling of post-
consumer plastics into useful hydrocarbons is important for both
environmental and economic reasons; however, the processing of
highly commingled plastics with different compositions within the
waste matrix would meet signicant technical challenges [34]. In
specically, tertiary recycling of mixed plastic wastes containing
polyvinyl chloride releases hydrogen chloride which causes not
only corrosion of the equipment but also the formation of chloro-
organic compounds in hydrocarbons. Technology using thermal/
catalytic pyrolysis toward valuable hydrocarbons shows a
promising future in the forthcoming years, specically considering
the chlorine fromresidual PVC in the waste [35]. Therefore, it is the
objective of this work to investigate a similar FCC reaction system
using a recycled FCC catalyst (FCC-R1) and a series of reaction
conditions for the study of the product distributions in the
degradation of plastic waste with PVC mixture (HDPE/LDPE/PP/
PVC) and especially for the development of a suitable reaction
model for achieving post-consumer plastic recycling.
2. Experimental
2.1. Materials and reaction preparation
The plastic mixture used in this study was obtained from post-
consumer plastic waste stream in South-Taiwan with the compo-
nent of polyethylene (63 wt% PE = 33 wt% HDPE + 30 wt%
LDPE), polypropylene (34 wt% PP), with about 3 wt% polyvinyl
chloride (PVC) mixtures. Prior to plastic waste fed into the reactor,
the polymer waste was cleaned, washed and grinded in the particle
range 75250 mm to enhance adequately uidizing throughout the
course of each reaction. Typically, the content of waste plastic
sample tested by element analysis was about 85.35% C, 12.13% H,
2.21%Cl, 0.13%O, and0.18%N. The catalysts employedare described
inTable 1. Surface area, pore volume and pore size distributionwere
measured fromthe adsorptiondesorption isotherms of nitrogen at
77 K on a Micromeritics ASAP 2020 apparatus. Total surface area of
the catalysts was estimated by application of the BET equation and
total pore volumes from the nitrogen adsorbed at p/p
0
= 0.99. Pore
size distributionwas obtainedfollowingthe BJHmodel, whereas the
micropore specic surface area and the micropore volume were
calculatedbythe t-plot method. Prior touseandcharacterization, all
the catalysts were crushed and sieved to give particle sizes ranging
from 75 to 180 mm. The catalyst (0.25 g) was subjected to nish
Table 1
Catalysts used in the catalytic degradation of post-consumer plastic mixture.
Catalyst Si/Al Surface area (cm
2
/g) Metal (ppm) Commercial name
BET
a
Micropore External V Ni
FCC-R1
b
2.1 147 103 44 2560 870 Equilibrium catalysts
c
HUSY 5.7 472 375 118 Ultrastabilised Y zeolite
c
ZSM-5 17.5 375 257 118 ZSM-5 zeolite
d
SAHA 2.6 268 21 247 Amorphous silicaalumina
c
a
Total surface area (BET).
b
FCC-R1 obtained from a commercial FCC unit with different levels of rare earth oxide (1.3wt% REO), a mixture of Y zeolite, a silicaalumina matrix (32.5wt% Al
2
O
3
) and
binder. FCC-S1 represents a spent commercial equilibrium catalyst from renery FCC units without regeneration.
c
Chinese Petroleum Corp., CPC, Taiwan, ROC.
d
Laporte, Warrington, UK.
Y.-H. Lin et al. / J. Anal. Appl. Pyrolysis 87 (2010) 154162 155
drying and calcination in the uidized-bed reactor in each cracking
run by heating in owing nitrogen to 120 8C at 60 8C h
1
. A suitable
ow rate of nitrogen uidizing gas was then used in all runs to
improve the external heat transfer betweenthe moltenpolymer and
the outsideof the catalyst particles after well purgingthe analyzer to
inert the atmosphere inside. After 2 h the temperature was
increased to 520 8C at a rate of 120 8C h
1
to activate the catalyst
for 5 h. Additionally, the deactivatedcatalyst after catalytic reaction,
it normally holds a certain amount of coke, and air is needed to burn
it off at high temperature before further application. High purity
nitrogen was used as the uidizing gas and the ow was controlled
bya needle valve andpre-heatedinthe bottomsectionof the reactor
tube. Before catalytic pyrolysis experiments were started, several
uidization runs were performed at ambient temperature and
pressure to select: (i) suitable particle sizes (both catalyst and
polymer waste) and (ii) optimize the uidizing gas ow rates to be
used in the reaction. The particle size of both catalyst (75180 mm)
and plastic mixture (75250 mm) was chosen to be large enough to
avoid entrainment but not too large as to be inadequately
uidized. High ow rates of uidizing stream improve catalyst
polymer mixing and external heat transfer between the hot bed
and the cold catalysts. On the other hand, an excessive ow rate
could cause imperfect uidization and considerable entrainment
of nes. Hence, velocities in the range 1.54 times the value of the
minimum uidization velocity of catalyst (U
mf
) were used in the
course of this work.
2.2. Experimental procedures and product analysis
A process ow diagram of the experimental system is given
elsewhere [22] and shown schematically in Fig. 1. A three-zone
heating furnace with digital controllers was used and the
temperatures of the furnace in its upper, middle and bottom
zones were measured using three thermocouples. By these means
the temperature of the pre-heated nitrogen below the distributor
and catalyst particles in the reaction volume could be effectively
controlled to within 1 8C. The polymer feed systemwas designed to
avoid plugging the inlet tube with melted polymer and to eliminate
air in the feeder. The feed system was connected to a nitrogen supply
to evacuate polymer into the uidized catalyst bed. Thus, com-
mingled polymer particles were purged under nitrogen into the top of
the reactor and allowed to drop freely into the uidized bed at
t = 0 min. At sufcient plastic/catalyst ratios (much closer resembling
FCC conditions) the outside of the catalyst particles is not wet with
polymer, so the catalyst particles move freely.
Volatile products leaving the reactor were passed through a
glass-ber lter to capture catalyst nes, followed by an ice
acetone condenser to collect any condensable liquid product. A de-
ionized water trap was placed in series after the condenser to catch
any HCl produced by the degradation of PVC component. A three-
way valve was used after the condenser to route product either
into a sample gas bag or to an automated sample valve systemwith
16 loops. The Tedlar bags, 15 L capacity, were used to collect time-
averaged gaseous samples. The bags were replaced at intervals of
10 min throughout the course of reaction. The multiport sampling
valve allowed frequent, rapid sampling of the product stream
when required. Spot samples were collected and analyzed at
various reaction times (t = 1, 2, 3, 5, 8, 12, 15, 20 min). The rate (R
gp
,
wt% min
1
) of hydrocarbon production of gaseous products
collected by automated sample system in each run was dened
by the relationship:
Gaseous hydrocarbon products were analyzed using a gas
chromatograph equipped with a thermal conductivity detector
(TCD) tted with a 1.5 m 0.2 mm i.d. Molecular Sieve 13X
packed column and a ame ionization detector (FID) tted
with a 50 m 0.32 mm i.d. PLOT Al
2
O
3
/KCl capillary column.
A calibration cylinder containing 1% C
1
C
5
hydrocarbons was
used to help identify and quantify the gaseous products. The
solid remaining deposited on the catalyst after the catalytic
Fig. 1. Schematic diagram of a catalytic uidized-bed reactor system: (1) feeder, (2) furnace, (3) sintered distributor, (4) uidized catalyst, (5) reactor, (6) condenser, (7) de-
ionized water trap, (8) 16-loop automated sample system, (9) gas bag, (10) GC, and (11) digital controller for three-zone furnace.
R
gp

hydrocarbon production rate of gaseous products in each spot rung=min
total hydrocarbon production of gaseous products over the whole spot runs g
100
Y.-H. Lin et al. / J. Anal. Appl. Pyrolysis 87 (2010) 154162 156
degradation of the polymer were deemed residues and
contained involatile products and coke. The amount and nature
of the residues was determined by thermogravimetric analysis
(TGA). A number of runs were repeated in order to check their
reproducibility. It was found that the experimental error was
within 5%.
2.3. Kinetic modeling
A general mechanistic reaction scheme involving the discus-
sion of the carbenium ion of catalytic cracking chemistry for the
degradation of hydrocarbon has been proposed. This representa-
tion is simplied regarding the formation of carbenium ions in
that it concentrates on reaction paths rather than on surface
species. Theoretical studies suggest that, for the conversion of
hydrocarbons on active zeolites, reactions proceed via carbenium
ions as transition states (rather than as intermediates) and
product distributions are also generally in agreement with
carbenium ion studies, although there is still debate about the
actual mode of scission. In this paper, a kinetic/mechanistic
model including the main mechanistic features and kinetic
reaction schemes for plastic degradation over cracking catalysts
is investigated (Fig. 2). The model uses the following assump-
tions.
1. The liquid-phase plastic. Initially, solid polymer is freely dropped
into the reactor and immediately melts to disperse around the
catalysts. The molten plastic, in contact with the catalyst
particle forms a polymer/catalyst complex, reaction commen-
cing at the surface. Plastic melting and spread times are
negligible.
2. Evolution of intermediates. Scission reactions generate inter-
mediates that include long-chain olens and intermediate
precursors for carbeniumions. The carbeniumions rapidly reach
a steady-state concentration. Alkanes may be generated, via
hydrogen transfer, and initially will be largely long-chain
products. In general, the number of active sites limits the
number of carbenium ion precursors.
3. Evolution of products. Once the intermediates are produced,
further reactions could be expected to produce smaller chain
olens in equilibrium with surface carbenium ions, as well as
alkanes, BTX (benzene, toluene and xylene) and coke. The
equilibriummixture of olens and carbeniumions subsequently
reacts further to produce the nal products.
On the basis of the reaction pathway shown in Fig. 2, the rate of
formation, r
i,j
, of the product i from reactant r, through reaction j
can be written as
r
i; j
k
i; j
W
r
n
j
h
j
(1)
where k
i,j
is the rate constant of the product i from the jth reaction,
W
r
is the weight fraction of the reactant r present on the acid sites,
h
j
is the catalyst activity decay of the jth reaction, and n
j
is the
reaction order of the jth reaction. Polymer cracking is known to
proceed over acidic catalysts by carbocation mechanisms, where
the initially formed ions undergo chain reactions via processes,
such as scission or b-scission and isomerization and hydrogen-
transfer alkylation and oligomerization, to yield typically smaller
cracked products. For plastic waste degradation over a spent
uidizing catalytic cracking (FCC) catalyst, an exponential decay
function with activity decaying as function of coke on catalyst with
the apparent reaction order of 1 was reported previously [30].
h expa Cc (2)
where C(c) is coke content deposited on the catalyst. For the model
we consider, as an approximation, that a, in the deactivation
process, can be taken as constant for both polymers and that the
active sites are deactivated at the same rate for the acid catalysts
studied. The activity decay, h
j
, was assumed to be the same for all
reaction steps with the activity proportional to the number of
remaining sites.
h
j
h expa W
r
(3)
where W
r
is coke content deposited on the individual catalyst.
Eqs. (4)(9) were obtained assuming that reaction rates can be
represented by a simple rst-order process and catalyst deactiva-
tion involving six simultaneous equations. Initially, surface
reaction forms a complex of molten polymer/catalyst. Thus, the
rate of the consumption of the lumped liquid-phase polymer
species (W
c
) can be written as

dW
c

dt
h k
i;c
k
p;c
k
h;c
k
r=a;c
W
c
(4)
Therefore, the formation of intermediates of olens and carbenium
ions can be expressed as
dW
i
dt
h k
i;c
W
c
k
o;i
k
p;i
k
h;i
k
r=a;i
W
i
(5)
Fig. 2. Kinetic and mechanistic reaction schemes for the mixture of PE/PP/PS/PVC plastic waste degraded over various cracking catalysts.
Y.-H. Lin et al. / J. Anal. Appl. Pyrolysis 87 (2010) 154162 157
Similarly, the evolution of the olenic lump (W
o
), the parafnic
lump (W
p
), the HCl lump (W
h
), and the coke/BTX lump (W
r/a
) are
described as follows
dW
o
dt
h k
o;i
W
i
(6)
dW
p
dt
h k
p;i
W
i
k
p;c
W
c
(7)
dW
h
dt
h k
h;i
W
i
k
h;c
W
c
(8)
dW
r=a
dt
h k
r=a;i
W
i
k
r=a;c
W
c
(9)
The mass balance can be written as

dW
c

dt

dW
i
dt

dW
o
dt

dW
p
dt

dW
h
dt

dW
r=a
dt
(10)
Eqs. (4)(10) were numerically integrated by a fourth-order
RungeKutta algorithm with Matlab to very the individual rate
constants by minimizing the sum of the squared deviations
between calculated and experimental results. This gave values for
the apparent rate constants.
3. Results and discussion
3.1. Degradation of commingled plastic waste (PE/PP/PVC) over
various catalysts
Both the carbon number distribution of the products of post-
consumer plastic waste (HDPE/LDPE/PP/PVC) cracking at 390 8C
over various catalysts and the nature of the product distribution
were found to vary with the catalyst used. As shown in Table 2, the
yield of volatile hydrocarbons for zeolitic catalysts (ZSM-
5 > HUSY) gave higher yield than non-zeolitic catalysts (SAHA)
and zeolite-based equilibrium FCC catalyst (FCC-R1 FCC-S1),
and the highest was obtained for ZSM-5 (nearly 88 wt%). Overall,
the bulk of the products observed with these acidic cracking
catalysts (FCC-R1, HUSY, ZSM-5 and SAHA) were in the gas phase
with less than 5 wt%liquid collected. The differences in the product
distributions between those catalysts can be seen with ZSM-5
producing a much more C
1
C
4
hydrocarbon gases (53.7 wt%) than
FCC-R1, HUSY and SAHA catalysts. Some similarities were
observed between FCC-R1 and SAHA with C
1
C
4
and C
5
C
9
yields,
which were approximately 2529 wt% and 5354 wt%, respec-
tively. The highest level of unconverted plastic was observed with
FCC-R1 and SAHA about 1011 wt%, while the highest coke yields
were observed with HUSY (4.4 wt%).
A large amount of solid residue, presumably unconverted
commingled polymer and high molecular weight degradation
products, remained on the spent catalyst was observed at 390 8C
for the mixture of commingled plastic degradation over FCC-S1
(deactivated spent catalyst from renery FCC units without
reactive pretreatment). As also can be seen in Table 2, the
conversion was only 16.4 wt% for FCC-S1 compared with 81.7 wt%
when regenerated and reactive FCC commercial equilibrium
catalyst (FCC-R1) was used. Typically thermal degradation
productions were observed with FCC-S1 showing primary cracking
products (HCl and styrene) and an even spread of carbon numbers
(C
1
C
4
= 8.3 wt% and C
5
C
9
= 7.9 wt%) with some isomerization of
BTX (0.3 wt%). However, product distributions with FCC-R1
catalyst at 390 8C contained more gasoline materials in the range
of C
5
C
9
(54.8 wt%) with 25.3 wt% light gases. For comparison,
Kaminsky et al. [21] tested the uidized bed using a FCC catalyst
with PE at temperatures in the range of 370515 8C. Similarly, the
presence of the catalyst changes the product distributions, when
compared to thermal degradation. In this case, gases (4852 wt%)
and low boiling oils (3840 wt%) were the main products, instead
of the waxes that were yielded in the non-catalytic runs. However,
Cardona and Corma [22] using a semi-batch stirred reactor for PP
degradation obtained lower yields of gases: 710 wt% (variation
with reaction time) at 380 8C compared with 25.3 wt% (at 390 8C)
in this work. It seems, although insufcient data are available for
full comparison olen yields were much lower than those in the
present work.
The chlorine was chemically separated from the PVC compo-
nent and as a hydrochloric acid (HCl) in de-ionized water system.
The major products of PVC cracking over various catalysts were
hydrogen chloride at about 12 wt% with light aromatics and
smaller chain olens and parafns, and with some amount of
residue products (involatile residue and coke formation over the
reaction) deposited on the catalyst. The results indicate that
although the initial cracking of plastic waste must be conned to
the external surface and pore mouths of the cracking catalysts, the
resultant initial cracked products are then degraded further within
the catalyst. The rate of gaseous hydrocarbon evolution further
highlights the slower rate of degradation over SAHA and FCC-R1 as
shown in Fig. 3 when comparing all catalysts under identical
Table 2
Summary of the main products of post-consumer plastic mixture (HDPE/LDPE/PP/
PVC) degraded at reaction temperature of 3908C over various catalysts
(uidizing N
2
rate=570ml min
1
, catalyst to plastic ratio=30wt% and catalyst
particle size=125180mm).
Catalyst type
FCC-R1 FCC-S1 HUSY ZSM-5 SAHA
Yield (wt% feed)
Gaseous 81.7 16.4 85.1 87.9 83.3
Liquid
a
4.9 1.5 4.7 4.2 4.6
Residue
b
11.7 82.1 8.6 6.4 10.5
Involatile residue 9.1 80.4 4.2 5.1 7.8
Coke 2.6 1.7 4.4 1.3 2.7
HCl 1.7 1.5 1.6 1.5 1.6
Distribution of gaseous products (wt% feed)
Light gases (C
1
C
4
) 25.3 8.3 34.7 53.7 29.2
Gasoline (C
5
C
9
) 54.8 7.9 49.4 33.7 53.4
BTX
c
1.6 0.3 1.0 0.5 0.7
a
Condensate in condenser and captured in lter, unidentied.
b
Coke and involatile products.
c
Benzene, toluene and xylene.
Fig. 3. Comparison of hydrocarbon yields as a function of time for the catalytic
degradation of post-consumer plastic mixture (HDPE/LDPE/PP/PVC) at 390 8C over
different catalysts (catalyst to plastic ratio = 30 wt%, rate of uidization
gas = 600 ml min
1
and catalyst particle size = 125180 mm).
Y.-H. Lin et al. / J. Anal. Appl. Pyrolysis 87 (2010) 154162 158
conditions at 390 8C. For FCC-S1, the yield and initial rate of
gaseous production were signicantly decreased and lower during
the course of reaction and the time for plastic waste to be degraded
lengthened until the end of total collection time. The maximum
rate of generation was observed after 2 min with the zeolite
catalysts (ZSM-5 and HUSY), whereas the maximum was observed
after 3 min with the zeolite-based FCC-R1 catalyst and SAHA
catalysts. The systematic experiments discussed in this work
indicate that catalyst deactivation is being produced by active-site
coverage, and consequently decrease the activity of the catalyst,
giving the reason of decreasing reaction rate with reaction time.
3.2. Product selectivity variation with reaction conditions
The inuence of operation conditions including reaction
temperatures, ow rates of uidizing gas (300900 ml min
1
)
and ratios of plastic waste to catalyst feed has been investigated.
The rate of hydrocarbon production as a function of time at the
same ve temperatures is compared in Fig. 4 and as expected,
faster rates were observed at higher temperatures. At 450 8C, the
maximum rate of hydrocarbon production was 38 wt% after only
2 min with all the plastic degraded after approximately 8 min. As
the temperature decreased, the initial rate of hydrocarbon
production dropped and the time for the polymer to be completely
degraded lengthened. At 290 8C, the rate of hydrocarbon produc-
tion was signicantly lower throughout the whole reaction with
the polymer being degraded over 20 min. Both acidity and
diffusion constraints within individual micropores of each catalyst
may play signicant roles in the observed product distribution. The
results shown in Fig. 5 illustrate that for efcient plastic waste
degradation good mixing is required, with a dramatic drop-off in
the rate of degradation observed only at the lowest uidizing ow
used (300 ml min
1
). Furthermore, changing the uidizing ow
rate inuences the product distribution. At low ow rates (longer
contact times), secondary products are observed with increased
amounts of coke precursors (BTX) although the overall degradation
rate is slower as shown by increasing amounts of partially
depolymerized products (Table 3). The amount of FCC-R1 used in
the degradation of post-consumer plastic mixture remained
constant and, therefore, as more waste plastics were added to
the reactor then fewer catalytic sites per unit weight of catalyst
were available for cracking. The overall effect of increasing the
plastic to catalyst ratio from 1:0.1 to 1:0.6 on the rate of
hydrocarbon generation was small but predictable. As the plastic
to catalyst ratio increases, the possibility of post-consumer plastic
adhesion to the reactor wall increases as the amount of unreacted
polymer waste in the reactor rises. However, for the work carried
out in this paper no such problems were observed. The total
product yield after 20 min showed only a slight downward trend
even after a 10-fold increase in added plastic waste. This can be
attributed to the sufcient cracking ability of FCC-R1 and excellent
contact between post-consumer plastic mixture and catalyst
particles. Consequently, as more plastic mixture was added, lower
C
5
C
9
gasoline yields but higher liquid yields and involatile
products were observed (Table 4). In addition, more BTX (coke
precursor) was produced but increasing the plastic to catalyst ratio
had only virtually no effect on C
1
C
4
hydrocarbon gases and HCl
productions.
3.3. Kinetic results and discussion
The kinetic model has been used to represent product
distributions for the degradation of polyolens over acidic cracking
catalysts under the uidized-bed reaction. As shown in Fig. 6, it
shows that the calculated values using various catalysts are in good
agreement with the experimental data. The apparent rate
Fig. 4. Comparison of hydrocarbon yields as a function of time at different reaction
temperatures for the catalytic degradation of post-consumer polymer mixture
(HDPE/LDPE/PP/PVC) over FCC-R1 catalysts (rate of uidization gas = 600 ml min
1
,
catalyst to plastic ratio = 30 wt% and catalyst particle size = 125180 mm).
Fig. 5. Comparison of hydrocarbon yields as a function of time at different
uidization gases for the degradation of post-consumer polymer mixture (HDPE/
LDPE/PP/PVC) over FCC-R1 catalysts (reaction temperature = 390 8C, catalyst to
plastic ratio = 30 wt% and catalyst particle size = 125180 mm).
Table 3
Product distributions shown from FCC-R1 catalyzed pyrolysis of plastic mixture
(HDPE/LDPE/PP/PVC) at different uidizing N
2
rates (reaction temperature =3908C,
uidizing N
2
rate =570ml min
1
, catalyst to plastic ratio=30wt% and catalyst
particle size=125180mm).
Fluidizing N
2
rates (ml/min)
900 750 600 450 300
Yield (wt% feed)
Gaseous 84.1 83.1 81.7 82.2 81.5
Liquid
a
5.0 5.4 4.9 4.2 4.7
Residue
b
8.8 9.4 11.7 11.8 11.8
Involatile residue 6.6 7.0 9.1 9.5 9.4
Coke 2.2 2.4 2.6 2.3 2.4
HCl 2.1 2.1 1.7 1.8 2.0
Distribution of gaseous products (wt% feed)
Light gases (C
1
C
4
) 29.7 27.9 25.3 25.1 25.5
Gasoline (C
5
C
9
) 53.6 54.0 54.8 55.1 53.6
BTX
c
0.8 1.2 1.6 1.9 2.4
a
Condensate in condenser and captured in lter, unidentied.
b
Coke and involatile products.
c
Benzene, toluene and xylene.
Y.-H. Lin et al. / J. Anal. Appl. Pyrolysis 87 (2010) 154162 159
constants based on the model for various catalysts used in this
study are summarized in Table 5. The generation of intermediates
(long-chain olens and carbenium ions) from the liquid-phase
plastic is much faster than other reaction rates such as the
generation of the parafnic lump (k
p;c
) and the coke/BTX lump
(k
r=a;c
). The value of k
i;c
is higher for zeolites (ZSM-5 > FCC-R1)
than for SAHA. This may suggest that the nature of the catalyst
with zeolites is more effective in converting polymers into
Fig. 6. Comparison of calculated (m) and experimental (e) results for the
degradation of post-consumer plastic mixture (HDPE/LDPE/PP/PVC) over (a)
ZSM-5, (b) FCC-R1 and (c) SAHA catalysts at 390 8C (catalyst particle size = 125
180 mm, uidizing N
2
rate = 600 ml min
1
and catalyst to plastic ratio = 30 wt%).
Table 4
Product distributions shown from FCC-R1 catalyzed pyrolysis of plastic mixture
(HDPE/LDPE/PP/PVC) at different ratios of catalyst to plastic waste (reaction
temperature =3908C, uidizing N
2
rate=600ml min
1
and catalyst particle
size =125180mm).
Ratio of catalyst to plastic waste (wt%)
10 20 30 40 60
Yield (wt% feed)
Gaseous 79.1 80.5 81.7 83.5 84.7
Liquid
a
6.1 5.5 4.9 4.3 4.1
Residue
b
13.1 12.2 11.7 10.4 9.4
Involatile residue 11.0 10.1 9.1 7.9 6.8
Coke 2.0 2.1 2.6 2.5 2.6
HCl 1.7 1.8 1.7 1.8 1.8
Distribution of gaseous products (wt% feed)
Light gases (C
1
C
4
) 25.9 25.4 25.3 25.9 25.7
Gasoline (C
5
C
9
) 50.7 52.9 54.8 56.3 58.1
BTX
c
2.5 2.2 1.6 1.3 0.9
a
Condensate in condenser and captured in lter, unidentied.
b
Coke and involatile products.
c
Benzene, toluene and xylene.
T
a
b
l
e
5
C
o
m
p
a
r
i
s
o
n
o
f
a
p
p
a
r
e
n
t
r
a
t
e
c
o
n
s
t
a
n
t
s
a
n
d
p
r
o
d
u
c
t
s
e
l
e
c
t
i
v
i
t
y
f
o
r
t
h
e
d
e
g
r
a
d
a
t
i
o
n
o
f
p
o
s
t
-
c
o
n
s
u
m
e
r
c
o
m
m
i
n
g
l
e
d
p
l
a
s
t
i
c
m
i
x
t
u
r
e
(
H
D
P
E
/
L
D
P
E
/
P
P
/
P
V
C
)
o
v
e
r
d
i
f
f
e
r
e
n
t
c
a
t
a
l
y
s
t
s
(
r
e
a
c
t
i
o
n
t
e
m
p
e
r
a
t
u
r
e
=
3
9
0
8
C
,

u
i
d
i
z
i
n
g
N
2
r
a
t
e
=
6
0
0
m
l
m
i
n

1
,
c
a
t
a
l
y
s
t
t
o
p
l
a
s
t
i
c
r
a
t
i
o
=
3
0
w
t
%
a
n
d
c
a
t
a
l
y
s
t
p
a
r
t
i
c
l
e
s
i
z
e
=
1
2
5

1
8
0
m
m
)
.
C
a
t
a
l
y
s
t
k
i
;
c

1
0

2
m
i
n

1
)
k
r
=
a
;
c

1
0

2
m
i
n

1
)
k
p
;
c

1
0

3
m
i
n

1
)
k
h
;
c

1
0

3
m
i
n

1
)
k
o
,
i
(

1
0

2
m
i
n

1
)
k
r
/
a
,
i
(

1
0

2
m
i
n

1
)
k
p
,
i
(

1
0

2
m
i
n

1
)
k
h
,
i
(

1
0

3
m
i
n

1
)
k
o
,
i
/
k
T
a
(
%
)
k
r
/
a
,
i
/
k
T
(
%
)
k
p
,
i
/
k
T
(
%
)
k
h
,
i
/
k
T
(
%
)
F
C
C
-
R
1
2
1
2
.
3
1
0
.
8
4
.
2
1
.
0
1
4
.
8
1
.
6
2
0
.
2
0
.
7
3
9
.
7
4
.
3
5
4
.
1
1
.
9
Z
S
M
-
5
2
3
9
.
1
2
.
6
1
.
3
0
.
6
2
6
.
1
0
.
8
9
.
6
0
.
5
7
0
.
5
2
.
2
2
5
.
9
1
.
4
S
A
H
A
1
8
0
.
4
1
6
.
5
2
.
3
0
.
4
2
7
.
9
2
.
2
5
.
4
0
.
6
7
7
.
3
6
.
1
1
5
.
0
1
.
7
a
K
T
=
k
o
,
i
+
k
r
/
a
,
i
+
k
p
,
i
+
k
h
,
i
.
Y.-H. Lin et al. / J. Anal. Appl. Pyrolysis 87 (2010) 154162 160
intermediates of volatile precursors than the amorphous SAHA.
The mechanism involved is related to the diffusion of bulky
fragments for the unreacted or partially reacted plastic softened by
dissolution of low molecular species deposited on the catalyst. As
can be seen in Fig. 6, higher evolution of intermediate species
predicted in the kinetic model indicating that plastic waste can be
cracked at or close to the external surface of the catalysts and
therefore, the controlling catalytic factors will be associated with
the number of accessible spaces related to the value of coke
content deposited on the catalyst and the total number of acid
sites. The subsequent reactions for the generation of the olenic
lump, coke/BTX lump, HCl lump, and parafnic lump from long-
chain olens and carbeniumions lump were found to vary with the
catalyst used. Higher values of the apparent rate constant of
parafnic lump (k
p;c
) and coke/BTX lump (k
r=a;c
) were observed in
FCC-R1 compared to in ZSM-5 and SAHA. For zeolites of ZSM-5 and
FCC-R1, the generation of olenic and parafnic lumps is much
faster than the generation of coke/BTX lump. For SAHA, the
generation of the olenic fraction (k
o,i
) is much faster than the
other values of the generation of the parafnic lump (k
p,i
) and HCl
lump (k
h,i
). The difference in the amount of olenic product (k
o,i
/k
T
)
for different catalysts is in the order SAHA (77.3%) > ZSM-5
(70.5%) FCC-R1 (39.7%). Product selectivity varied strongly
depending on catalysts type and structure. Compared with those
of zeolites (ZSM-5 and FCC-R1), SAHA with large mesopores and
lowacidity resulted in a relatively lowrate of bimolecular reaction
to further react to produce parafns. This also can be seen for ZSM-
5 (2.2%) producing a much lower selectivity to BTX/coke than for
SAHA (6.1%). The difference in product selectivity for zeolites can
be seen with FCC-R1 giving a higher selectivity for parafns than
ZSM-5. However, ZSM-5 generates a much higher selectivity for
olens compared with FCC-R1. These results may reect FCC-R1
catalyst containing potentially cracking activity with bimodal pore
structures, which is composed of both the micropore of zeolite and
the mesopore of silicaalumina used in the FCC catalyst matrix,
and can provide the unzipping ability when they are enough
accessible to the reactant and may allow bulky reactions to occur,
ultimately leading to the generation of hydrocarbon products with
substantial coke/BTX levels (4.3%). However, ZSM-5 with narrower
pore openings (0.55 nm 0.51 nm) and no supercages, the bulky
feed molecules have restricted access to the internal active
catalytic sites and special restrictions within the pore system
tend to inhibit the bimolecular processes leading to lower values of
HCl and coke/BTX production in this study.
The rate constants for the catalytic degradation of commingled
plastic waste over FCC-R1 with particle sizes of 125180 and 75
120 mm at ve different temperatures (330450 8C) are listed in
Table 6. The selectivity toward the olenic products (k
o,i
/k
T
) as a
function of reaction temperature shows that the amount of olens
decreased with increasing reaction temperature from330 to 430 8C
then increased at 450 8C. On the other hand, the results show the
parafnic products (k
p,i
/k
T
) occurred at 430 8C rather than at the
higher temperature of 450 8C. It could be the case that selectivity
products of parafnic fraction increased with increasing reaction
temperature and olenic fraction decreased with an increase in
temperature below 430 8C. But at higher temperature (450 8C), the
degradation of commingled plastic waste to volatile products may
proceed both by catalytic and thermal reactions leading to the
variation of product distributions. The external surface of the
catalysts is an important factor for controlling activity and
selectivity. From values of k
o,i
/k
T
and k
r/a,i
/k
T
, the smaller particle
FCC-R1-Scatalyst (75120 mm) shows ahigher selectivityfor olens
(39.146.2% vs. 36.643.7%) and a lower selectivity for coke/BTX
(3.14.3% vs. 3.65.4%) than the larger particle FCC-R1-L catalyst
(125180 mm). The FCC-R1 catalyst containing cracking activity
with bimodal pore structures, which is composed of both the
micropore of zeolite and the mesopore of silicaalumina used in the
FCC catalyst matrix, can provide the cracking ability when they are
enough accessible to the reactant and may allow bulky reactions to
occur, ultimately leading to the generation of hydrocarbon products
withsomecokecontent depositedonthecatalyst. Theresult appears
that the feed can be cracked at or close to the external surface of the
catalyst and therefore, the controlling catalytic parameters will be
not only the total number of acid sites but also the number of
accessible ones related to the particle size selected. The results
obtainedcanbecomparedwithsome of the results publishedfor the
thermal and/or catalytic degradations of polymer waste. This
comparison is not straight forward as reaction conditions are not
perfectly matched. Much higher yields of liquids with aromatics
(1520 wt%) in the 100 of molecular weight range were provided by
Lee et al. [23] fromthe catalytic cracking of a pure polyolen (PE, PP
and PS) over zeolitic catalysts in a stirred semi-batch at 400 8C. The
reaction time was relatively long, up to 3 h, and was presumably set
by the relative low polymer degradation rate at this temperature.
However, catalytic pyrolysis of polymer waste performed in the
uidized-bed reactor in the temperature range 330450 8C was
shown to produce valuable hydrocarbons in the range of C
3
C
7
carbonnumber witha higholenic content. Data are provided by de
la Puente et al. [24], using a riser simulator reactor for the catalytic
cracking of the LDPE dissolved into toluene, obtained gases yield
(20 wt%) and again relative high aromatics yields (25 wt%) and coke
content (nearly 10 wt%) deposited on the catalyst. Although
catalysis has been used, this often involves thermal cracking of
the polymer followed by catalytic conversion of the degradation
products. However, the conguration of the pyrolysis-reforming
reactors poses serious engineering and economics constraints.
Problems associated with blockage and limited polymer/catalyst
contact within the reactor make continuous processing difcult in
xed-bed reactors. Arandes et al. [25] performed a fast pyrolysis of
PP to obtain waxes that were subsequently degraded in a riser
simulator with a commercial FCC catalyst to improve mass- and
heat-transfer constraints. At 500550 8C, mixtures of waxes with
vacuumgasoline oils result in a synergetic effect to produce mostly
liqueed petroleum gas and gasoline (42 wt%) with about 45 wt%
coke content. In this work, the production of olens with potential
value as a chemical feedstock is potentially attractive and may offer
greater protability than production of saturated hydrocarbons and
aromatics. Therefore, thecatalyticdegradationof amixtureof plastic
Table 6
Comparison of product selectivity for the degradation of post-consumer commingled plastic mixture (PE/PP/PVC) over different particle sizes of FCC-R1 catalyst in the
temperature range 3304508C using the lumping model (uidizingN
2
rate=600ml min
1
and catalyst to polymer ratio=30wt%).
Temperature (8C) FCC-R1-L (125180mm) FCC-R1-S (75120mm)
k
o,i
/k
T
a
(%) k
r/a,i
/k
T
(%) k
p,i
/k
T
(%) k
h,i
/k
T
(%) k
o,i
/k
T
a
(%) k
r/a,i
/k
T
(%) k
p,i
/k
T
(%) k
h,i
/k
T
(%)
330 43.7 5.4 49.5 1.4 46.2 4.3 48.0 1.5
360 41.5 4.9 52.0 1.6 43.8 4.0 50.7 1.5
390 39.7 4.3 54.1 1.9 41.3 3.6 53.4 1.7
430 36.6 3.9 57.4 2.1 39.1 3.3 55.7 1.9
450 41.9 3.6 52.2 2.3 43.2 3.1 51.6 2.1
a
K
T
=k
o,i
+k
r/a,i
+k
p,i
+k
h,i
.
Y.-H. Lin et al. / J. Anal. Appl. Pyrolysis 87 (2010) 154162 161
waste (HDPE/LDPE/PP/PVC) over post-use FCC catalysts performed
in uidized-bed reactor under suitable operating conditions was
shown to be a useful method for the production of potentially
valuable hydrocarbon fuels.
4. Conclusions
A catalytic uidizing reaction system has been used to obtain a
range of volatile hydrocarbons by catalytic degradation of post-
consumer commingled plastic mixture (HDPE/LDPE/PP/PVC) in the
temperature range 330450 8C. In this work, the uidized bed has
been shown to have a number of advantages in the pyrolysis of
commingled plastic waste; it is characterized by excellent heat-
and mass transfer, much less prone to clogging with molten
polymer and gives a nearly constant temperature throughout the
reactor. The catalytic degradation of polymer waste over the spent
commercial FCC equilibrium catalyst (FCC-R1) using uidizing
cracking reactions was shown to be a useful method for the
production of potentially valuable hydrocarbons. Observed differ-
ences in product yields and product distributions can be inuenced
by the change in reaction conditions.
This paper outlines some recent results relevant to the
conversion of polymers for the production of potentially valuable
hydrocarbons and also attempts to apply a model for the
mechanism and kinetics of catalytic degradation of polymers. A
combined kinetic and mechanistic model giving chemical infor-
mation, applicable to the uidized-bed reactions, was used to
predict production rates and product selectivity for the catalytic
degradation of polymers. This model gives a good representation of
product selectivity and product distributions for catalytic plastic
recycling and also provides an improvement of the empirical
lumping techniques. It is concluded that the use of this reaction
systemcoped with a spent FCC equilibriumcatalyst can be a better
option since it may lead to a cheaper process with valuable
products and can be further used as an adequate approach for the
catalytic recycling of plastic waste.
Acknowledgements
The authors would like to thank the National Science Council
(NSC) of the Republic of China (ROC) for their kindly nancial
supports (NSC 94-2211-E-244-008 and NSC 98-2221-E-244-014).
Thanks also are due to Professor Keng-Liang Ou and Research
Center for Biomedical Implants and Microsurgery Devices for
sample measurements.
References
[1] G. Scott, Polymers and the Environment, The Royal Society of Chemistry, Thomas
Graham House, Cambridge, UK, 1999.
[2] D.D. Cornell, Plastics, Rubber, and Paper Recycling: A Pragmatic Approach, Amer-
ican Chemical Society, Washington, 1995, p. 72.
[3] J. Brandrup, M. Bittner, W. Michaeli, G. Menges, Recycling and Recovery of Plastics,
Carl Hanser Verlag, Munich, New York, 1996.
[4] S. Hardman, S.A. Leng, D.C. Wilson, Polymer Cracking, Eur. Patent Appl. 567292
(1993) to BP Chemicals Limited.
[5] W. Kaminsky, B. Schlesselmann, C. Simon, J. Anal. Appl. Pyrol. 32 (1995) 19.
[6] S.F. Sodero, F. Berruti, L.A. Behie, Chem. Eng. Sci. 51 (1996) 2805.
[7] P.N. Sharratt, Y.H. Lin, A. Garforth, J. Dwyer, Ind. Eng. Chem. Res. 36 (1997)
5118.
[8] Y.H. Lin, P.N. Sharratt, A. Garforth, J. Dwyer, Energy Fuels 12 (1998) 767.
[9] Y.H. Lin, M.H. Yang, J. Mol. Catal. A: Chem. 171 (2001) 143.
[10] Y. Uemichi, J. Nakamura, T. Itoh, A. Garforth, J. Dwyer, Ind. Eng. Chem. Res. 38
(1999) 385.
[11] J. Aguado, D.P. Serrano, J.M. Escola, E. Garagorri, J.A. Fernandez, Polym. Degrad.
Stab. 70 (2000) 97.
[12] A. Dawood, Miura, Polym. Degrad. Stab. 76 (2002) 45.
[13] K. Gobin, G. Monos, Polym. Degrad. Stab. 86 (2004) 225.
[14] Y.H. Lin, M.H. Yang, J. Mol. Catal. A: Chem. 231 (2005) 113.
[15] A. Marcilla, A. Gomez Siurana, F. Valdes, J. Anal. Appl. Pyrol. 79 (2007) 433.
[16] A. Ghanbari-Siakhalu, A. Garforth, C.S. Cundy, J. Dwyer, J. Mater. Chem. 11 (2001)
569.
[17] G. Puente, C. Klocker, U. Sedran, Appl. Catal. B: Environ. 36 (2002) 279.
[18] Z. Gao, K. Ksuyoshi, I. Amasaki, M. Nakada, Polym. Degrad. Stab. 80 (2003) 269.
[19] Y.H. Lin, M.H. Yang, Polym. Degrad. Stab. 92 (2007) 813.
[20] A. Marcilla, A. Gomez, A. Reyes-Labrta, A. Giner, Polym. Degrad. Stab. 80 (2003)
233.
[21] J. Mertinkat, A. Kirsten, M. Predel, W. Kaminsky, J. Anal. Appl. Pyrol. 49 (1999) 87.
[22] S.C. Cardona, A. Corma, Appl. Catal. B: Environ. 25 (2000) 151.
[23] K.H. Lee, D.H. Shin, Y.H. Seo, Polym. Degrad. Stab. 84 (2004) 123.
[24] G. de la Puente, U. Sedran, C. Klocker, Appl. Catal. B: Environ. 36 (2002) 279.
[25] J.M. Arandes, I. Torre, P. Castano, M. Olazar, J. Bilbao, Energy Fuels 21 (2007) 561.
[26] Y.H. Lin, M.H. Yang, Appl. Catal. B: Environ. 69 (2007) 145.
[27] Y.H. Lin, M.H. Yang, Appl. Catal. A: Gen. 328 (2007) 132.
[28] Y.H. Lin, M.H. Yang, J. Anal. Appl. Pyrol. 83 (2008) 101.
[29] B. Saha, A.K. Ghoshal, Ind. Eng. Chem. Res. 46 (2007) 5485.
[30] A. Corma, Chem. Rev. 95 (1995) 559.
[31] Y.H. Lin, M.H. Yang, Thermochim. Acta 471 (2008) 52.
[32] Y.H. Lin, M.H. Yang, Polym. Degrad. Stab. 94 (2009) 25.
[33] B. Singh, N. Sharma, Polym. Degrad. Stab. 93 (2008) 561.
[34] S.M. Al-Salem, P. Lettieri, J. Baeyans, Waste Manage. 29 (2009) 2643.
[35] J. Aguado, D.P. Serrano, J.M. Escola, Ind. Eng. Chem. Res. 47 (2008) 7892.
Y.-H. Lin et al. / J. Anal. Appl. Pyrolysis 87 (2010) 154162 162