J. Anal. Appl.

Pyrolysis 79 (2007) 415–423

www.elsevier.com/locate/jaap

Feedstock recycling of polyethylene in a two-step thermo-catalytic

reaction system
J. Aguado *, D.P. Serrano, G. San Miguel, M.C. Castro, S. Madrid
ESCET, Rey Juan Carlos University, c/ Tulipán s/n, Móstoles 28933, Spain
Received 23 June 2006; accepted 15 November 2006
Available online 8 December 2006

Abstract
The conversion of low density polyethylene (LDPE) into high value hydrocarbons has been investigated using a two-step reaction system
consisting of an initial pyrolytic furnace followed by an independent reactor containing nanosized n-HZSM-5 zeolite or Al-MCM-41
mesostructured material where the catalytic reforming of the pyrolytic vapours took place. The system was run at temperatures between 425
and 475 8C and the results compared with those obtained in the absence of catalyst. Temperatures of 450 8C and above were required to reach
conversion values in excess of 90 wt%. At that temperature and in the absence of catalysts, thermal cracking of LDPE generated almost exclusively
a-olefins and n-paraffins over a wide range of molecular weights, most of which (74.7 wt%) collected as liquid products. Catalytic reforming over
n-HZSM-5 was effective at 425 8C and caused a significant increase in the proportion of gaseous hydrocarbons (73.5 wt% selectivity at 450 8C)
that consisted primarily of olefins. The remaining liquid products contained a high proportion of valuable aromatic and branched species in the
gasoline size range (C5–C12) making it suitable for blending with commercial gasolines. Owing to its weaker acid properties, catalytic reforming
over Al-MCM-41 required the use of higher temperatures (450 8C and above) to produce a lower proportion of gas products (54–58 wt%) but a
consequently higher amount of liquid (34–42 wt%) hydrocarbons. Besides, formation of aromatics was less significant than with n-HZSM5, but the
resulting oils contained a good combination of olefins and iso-paraffins. This product composition is comparable to commercial transportation fuel.
# 2006 Elsevier B.V. All rights reserved.

Keywords: Plastic recycling; Thermal cracking; Polyethylene; Catalytic reforming; Pyrolysis; Zeolite; Al-MCM-41

1. Introduction polymers [4] showing that the nature of the resulting products
depends primarily on polymer type but also on process
Consumption of plastic products in industrialised countries conditions, mainly temperature.
currently exceeds 100 kg/person [1]. This practice generates a In the case of polyethylene (PE), thermal cracking proceeds
vast waste stream that needs to be effective managed to reduce its through a random scission mechanism that generates a mixture
environmental impact and optimise its potential as a chemical of linear paraffins and a-olefins over a wide range of molecular
resource [2]. An alternative receiving increasing attention is weights [4]. Pyrolytic processes for the treatment of pure and
feedstock recycling, a term that encompasses a range of residual PE have been developed and tested at laboratory and
processes aimed at transforming the polymers into hydrocarbon pilot plant scale using fluidised beds [5,6], batch reactors [7,8]
products that may be used directly as fuels or as a feed for the and screw kiln reactors [9]. Furthermore, a number of
petrochemical industry [2,3]. Thermal cracking, also referred to demonstration plants have been in operation worldwide,
as pyrolysis, is one of such technologies, involving the heating of although their commercial success was never proven owing
the plastic at temperatures usually between 500 and 800 8C under primarily to the low market price of the resulting products
inert atmospheric conditions to cause the cleavage of the original [3,10,11].
molecules. Cracking reactions have been studied for most The use of catalysts represents a key advantage over
conventional pyrolytic procedures for plastic recycling. A wide
range of heterogeneous catalysts have been tested, the most
* Corresponding author. common being conventional zeolites like the ones employed in
E-mail address: jose.aguado@urjc.es (J. Aguado). the cracking of heavy petroleum fractions [8,9,12–14] and less
0165-2370/$ – see front matter # 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jaap.2006.11.008
416 J. Aguado et al. / J. Anal. Appl. Pyrolysis 79 (2007) 415–423

conventional microporous materials specially designed to and the hydrocarbon products were characterised for their
exhibit a higher share of external surface area [15] or improved carbon atom number and hydrocarbon type using GC-PIONA
thermal stability [13]. Other acid solids like alumina, silica- and simulated distillation.
alumina, aluminium pillared clays and mesostructured materi-
als (Al-MCM-41 and Al-SBA-15) have also been investigated 2. Experimental
for this purpose [1,16,17].
One evident benefit of using catalysts relates to the lower 2.1. Plastic polymer and catalysts
temperature at which the cracking reaction takes place, which
implies lower energy consumptions or higher conversions. The plastic used in this work was pure low density
However, an even more important aspect relates to the product polyethylene (LDPE) commercially produced by Repsol-YPF
selectivity exhibited by some catalysts, which allows the process (Spain). The plastic was supplied in the form of pellets and was
to be directed towards a narrower distribution of hydrocarbon used as received.
products with higher market values. This selectivity towards The acid catalysts employed, nanocrystalline n-HZSM-5
gaseous and aromatic products was originally described for the zeolite [23] and Al-MCM-41 [24], were synthesised in our
catalytic conversion of petroleum fractions over zeolitic laboratories according to previously published procedures. The
materials [12,18] and comparable results were observed by calcined materials were characterised for their Si/Al atomic
several authors when using polyolefins as the substrate [9,17,19]. ratios by inductively coupled plasma (ICP) using a Varian
A commercial process based on direct catalytic conversion of VISTA AX CCD spectrophotometer. Their textural properties
polyolefins has been in operation in Zabrze (Poland) since 1997 were analysed in a Micrometrics ASAP 2010 nitrogen gas
for the production of diesel and gasoline [3]. adsorption analyser at 77 K, with total surface areas being
The economic benefit of increasing the proportion of calculated by application of the BET equation. Surface areas
hydrocarbon gases, at the expense of the liquid fraction, is not associated to micropores (SEXT+MESOPOR) and total
evidenced by comparing current spot price market values of micropore volumes were determined using the t-plot method,
different hydrocarbon products [20]. Thus, market values for and total pore volumes from the nitrogen adsorbed at p/
unsaturated hydrocarbon gases like ethylene and propylene are po = 0.90. Acid properties were determined by ammonia
reported to range between US$ 1.36 and 1.51 kg1, compared temperature programmed desorption (TPD) in a Micrometrics
to US$ 0.77–1.20 kg1 for aromatics like toluene, benzene or 2910 TPD/TPR apparatus. Crystal size in the nanosize n-
styrene or US$ 0.67–0.70 kg1 for liquid fuels like premium HZSM-5 was measured by transmission electron microscopy.
unleaded gasoline or MTBE. Unsaturated hydrocarbon gases
are used as feedstock in the petrochemical industry for a wide 2.2. Thermogravimetric analyses
range of applications including the production of plastic
polymers, the synthesis of MTBE or as raw chemicals in In order to evaluate the activity of the catalyst, thermo-
alkylation processes. gravimetric experiments were carried out in a TA Instruments
However, direct catalytic cracking of plastics suffers from a SDT 2960 simultaneous TGA-DSC analyser for a total mass of
number of drawbacks which has prevented its commercial 10.0  0.2 mg. Catalytic experiments required adding of a
success. The first one relates to the difficulty to recover the known amount (1.0  0.1 mg) of n-HZMS-5 or Al-MCM-41
catalyst after use, which increases the process costs and into the micro-crucible sample holder prior to the addition of
discourages the use of specially developed catalysts with higher the polymer. The samples were then heated under a nitrogen
production costs. Furthermore, as catalysts are in intimate flow (100 ml min1) from 50 to 700 8C, at 10 8C min1.
contact with the plastic, they are rapidly deactivated due to the Temperatures marking maximum degradation rates (Tmax) were
deposition of carbonaceous matter and the poisoning effect of determined from the derivative thermogravimetric (DTG) plots.
extraneous elements and impurities (chlorine, sulphur and Duplicate analyses conducted for selected samples produced
nitrogen containing species) that may be present in the residual deviations below 1 8C.
plastics [1,21]. In addition, owing to molecular size restrictions,
the diffusion of bulky polymeric molecules through the porosity 2.3. Catalytic conversion in the two-stage reactor
of certain microporous catalysts (zeolites) is partially impeded,
thus reducing their catalytic activity and potential use [21]. As illustrated in Fig. 1, the reaction system employed in this
Some of these problems may be overcome by separating the work consisted of two independent electrically heated quartz
pyrolytic stage from the catalytic reforming reaction, as it was vessels connected in series one on top of the other. The bottom
first tested by Bagri and Williams [22] for the conversion of one, which consisted of a cylindrical (140 mm high and 20 mm
polyethylene. Our work describes the performance of a i.d.) batch reactor continuously purged with 39.0 ml min1 of
purposely designed two-stage batch/fixed bed reactor for the nitrogen, contained the plastic polymer for its thermal cracking.
thermal-catalytic conversion of LDPE into higher value The vapours so generated were carried over the second step, a
hydrocarbons. The experiments were carried out at tempera- fixed bed reactor (160 mm long and 13 mm i.d.) reactor where
tures between 425 and 475 8C using nanosized n-HZSM-5 reforming takes place.
zeolite and mesostructured Al-MCM-41 in the catalytic step. In a typical experiment 5.0 g of plastic pellets were loaded
Conversions and product yields were determined for each run into the thermal reactor and 0.5 g of catalyst (n-HZSM-5 or
 J. Aguado et al. / J. Anal. Appl. Pyrolysis 79 (2007) 415–423 417

Fig. 1. Schematic representation of the two-stage reaction system.

Al-MCM41) were placed in the catalytic bed of the second one. simulated distillation. PIONA GC analyses were conducted in a
The system was purged with nitrogen and the catalytic reactor VARIAN CP-3800 chromatograph with a Chrompack CP SIL
was heated to a set temperature that coincided with the final PIONA (100 m length, 0.25 mm i.d.) column and FID detector,
temperature of the thermal one (between 425 and 475 8C). using a 1.4 ml min1 gas flow and a 1/200 split ratio. Liquid
When these conditions were reached, the bottom reactor was samples were injected after dilution in carbon disulphide (CS2)
heated from room temperature up to 400 8C at 10 8C min1 and while gases where injected as collected. The chromatograms
then at 2 8C min1 to reach a final set temperature of 425, 450 were analyzed using a VARIAN Detailed Hydrocarbon
or 475 8C, that was kept constant for 120 min. Experiments Analysis (DHA) Star software for product identification and
using similar conditions but in the absence of catalyst were also quantification.
conducted for comparative purposes. Simulated distillation was used to characterize heavier
The hydrocarbon products leaving the catalytic reactor were hydrocarbons in the liquid fraction. The samples were diluted in
passed through a glass condenser cooled by an ice trap to carbon disulphide and injected through a VARIAN WCOT Ulti-
separate the liquid products from the gas fraction. Liquid yields metal (10 m length, 0.5 mm i.d.) column mounted on a
were determined by weight from the amount of hydrocarbons VARIAN 3800 with FID detector.
collected in the condenser. Wax fractions were calculated from Selectivities of individual (x) species (Sx) were calculated as
the amount of solid product deposited at the exit of the catalytic their percentage contribution to the hydrocarbon products
reactor. Hydrocarbon gas yields were determined by difference generated, according to the following expression:
between the volumes determined by the calibrated flowmeters
situated at the inlet and outlet of the reaction system. wt% of product x
Sx ¼
Transformation of hydrocarbon gas volumes into mass values wt% gas þ wt% liquid þ wt% wax
was conducted by application of the equation of state of an ideal
gas at 25 8C and 1 atm.
Conversion values were determined by addition of the gas, 3. Results and discussion
liquid and wax fractions. Mass balances were closed with an
error always below 1.0 wt%. 3.1. Characterisation of catalysts

2.4. Analysis of hydrocarbon products Conventional HZSM-5 zeolites are crystalline microporous
solids exhibiting an MFI topology characterised by a three
Liquid and gas products were independently characterised dimensional micropore system consisting of straight 5.3 Å 
using PIONA gas chromatography (GC). The liquid hydro- 5.6 Å channels intersecting at right angles with 5.1 Å  5.5 Å
carbons were also analysed for their carbon atom number using sinusoidal channels. The zeolitic sample employed in this
418 J. Aguado et al. / J. Anal. Appl. Pyrolysis 79 (2007) 415–423

Table 1
Acid and textural properties of the catalysts
Catalyst Si/Ala Acidityb Surface area c Porosityc
Tmax (8C) NH3 desorption (meq g1) SBET (m2 g1) SEXT+MESOPOR (m2 g1) Vmic (cm3 g1) Vtotal (cm3 g1)
n-HZSM-5 36 366 0.502 418 74 0.169 0.236
Al-MCM-41 38 271 0.331 1123 1123 0 0.901
a
ICP-AES measurements.
b
Ammonia TPD.
c
Nitrogen adsorption at 77 K.

work, n-HZSM-5, was synthesized to exhibit a nanocrystalline 3.3. Conversions and products yields in the
structure aimed at providing a higher share of external surface thermo/catalytic conversion of LDPE
area and a readier accessibility of reacting molecules to the
catalytic sites. In contrast, Al-MCM-41 is an acid non- Fig. 3 illustrates the conversions and product yields resulting
crystalline mesostructured aluminosilicate synthesized using a from the thermal and subsequent catalytic reforming of LDPE
micelle template method. at temperatures between 425 and 475 8C. As expected, overall
The results in Table 1 show that both catalysts exhibited plastic conversions were controlled solely by the temperature in
comparable aluminium contents with Si/Al ratios between 36 the thermal reactor but were not affected by the presence of
and 38. However, owing to its crystalline nature, the zeolite catalysts in the second stage. When the reaction was conducted
presented a stronger acidity and a larger number of acid sites at 425 8C, only 53–55 wt% of the original plastic was collected
than the mesostructured solid, as evidenced by a higher as hydrocarbon products in the condenser or gas sampling bag.
desorption temperature (Tmax = 366 8C compared to 271 8C) The remaining 45–47 wt% was unable to pass through the
and a larger evolution of ammonia (0.502 meq g1 compared to reaction system and stayed in the thermal reactor as liquid oil. A
0.331 meq g1) desorbed in the TPD experiments. marked increase in the conversion values (between 90 and
Nitrogen gas adsorption analyses in Table 1 show that Al- 95 wt%) was observed at 450 8C and the reaction was complete
MCM-41 exhibited an extensive BET surface area
(1123 m2 g1) and total pore volume (0.901 cm3 g1), while
no contribution of micropores was detected in this sample. In
contrast, nanocrystalline zeolite n-HZSM-5 was essentially
microporous with comparatively lower surface area
(418 m2 g1) and total pore volume (0.236 cm3 g1). Com-
pared to conventional HZSM-5 materials, where the contribu-
tion of non-microporous structures to their textural properties is
usually negligible, the nanocrystalline sample used in this work
exhibited a significant external surface area (74 m2 g1) that is
associated with its small crystal size (60 nm).

3.2. Catalytic activities determined by TG analysis

The catalytic activity of the acid solids employed in this

work has been evaluated using thermogravimetric (TG)
analysis. This activity has been related to the capacity of each
catalyst to shift the degradation of LDPE to lower temperatures.
TG analyses in Fig. 2 illustrate that, in the absence of catalysts,
the polymer started to degrade at 419 8C and showed a
maximum in the derivative plot (DTG) at 477 8C.
Zeolite n-HZSM-5 exhibited a strong catalytic activity
reducing the degradation temperature by 81 8C down to 396 8C.
Despite its weaker acid properties, Al-MCM-41 also exhibited
a notable catalytic activity and reduced the degradation
temperature to 419 8C. It has been described in the literature
that the comparatively weaker acid properties of mesostruc-
tured solids are partly compensated by their larger pore
dimensions, which pose fewer impediments for the diffusion of
the reacting molecules to their internal catalytically active sites
[21,24]. Fig. 2. TG analyses describing the thermal and catalytic cracking of LDPE.
 J. Aguado et al. / J. Anal. Appl. Pyrolysis 79 (2007) 415–423 419

gas fraction represented only to 16.3 wt% of the total. The use
of higher temperatures (475 8C) favoured the formation of
gases (27.8 wt%) and reduced the proportion of liquid products
which was associated with a more intense cracking effect on the
plastic products.
Catalytic reforming over zeolite n-HZSM-5 led to a
significant increase of the gas fraction at the three temperatures.
At the optimum reaction temperature of 450 8C, the gas yield
amounted to 73.5 wt% of the original plastic mass and a similar
proportion was generated at 475 8C, although the proportion of
liquid hydrocarbons was slightly higher (22 wt%) at the latter
temperature.
Al-MCM-41 also exhibited a notable catalytic activity,
although this was less intense than n-HZSM-5 zeolite, which
was attributed to its weaker acid properties. Thus, when the
reaction was conducted at low temperatures (425 8C), the yields
by hydrocarbon fractions were comparable to those obtained in
the absence of catalytic reforming. At temperatures of 450 8C
and above, a marked increase in the formation of gas products
was observed (between 55 and 58 wt%), although these values
were still below those obtained when using the acid zeolite.

3.4. Characterisation of hydrocarbon products

Hydrocarbon products generated from the thermal-catalytic

conversion of LDPE were characterised for both their carbon

Fig. 3. Conversions and products yields from the thermo-catalytic conversion

of LDPE in the two-stage reaction system.

at 475 8C (over 95 wt% conversion). It was also observed that at

450 8C, the evolution of vapours from the thermal reactor
proceeded over the totality of the reaction experiment
(120 min) while at 475 8C this was only evident during the
first 60 min, after which the plastic polymer in the pyrolytic
reactor was essentially consumed.
With respect to product yields, it was observed that the
amount of solid waxes was very limited in all cases (less than
1.5 wt%). This has been attributed to the characteristics of the
reaction system, which does not allow large molecular weight
species to proceed through the system and remain in the
pyrolytic reactor until they have been sufficiently cracked.
In the absence of catalyst, conversion of LDPE at 450 8C
resulted in the formation of a high proportion of liquid Fig. 4. Selectivity by carbon atom number of products generated from the
hydrocarbons (74.7 wt% of the original plastic mass), while the thermo-catalytic reforming of LDPE at different temperatures.
420 J. Aguado et al. / J. Anal. Appl. Pyrolysis 79 (2007) 415–423

atom numbers and hydrocarbon type in order to evaluate the PIONA GC analysis was used to determine the proportion of
extent of the thermal-catalytic reactions and also to determine paraffins, iso-paraffins, olefins, naphthenes and aromatics in the
their potential application as refinery feedstock or fuel. hydrocarbon products generated from the two-step thermo-
catalytic conversion of LDPE at temperatures between 425 and
3.4.1. Characterisation by carbon atom number 475 8C. The results are represented in terms of product
Fig. 4 illustrates the selectivities by carbon atom number selectivity in three different categories: light (C1–C4) gases,
exhibited by the hydrocarbon products generated at different gasolines (C5–C12) and diesel type (C13–C22) hydrocarbons.
temperatures (425–475 8C) in the absence of catalyst and also Fig. 5 shows that the hydrocarbons generated from the
after reforming over n-HZSM-5 and Al-MCM-41. When no thermal cracking of LDPE correspond, almost exclusively, to
catalyst was placed in the second stage, the results show a linear paraffins and a-olefins over a wide range of molecular
considerable amount of light species (C1–C4) being formed, as weights. The formation of these species is well referenced in
well as heavier hydrocarbons in the gasoline (C5–C12) and the literature and has been associated with the random scission
diesel (C13–C22) ranges. The selectivities of these three mechanism involved in the thermal cracking of this polymer. A
fractions at 450 8C were 16, 41 and 40 wt%, respectively. At close inspection of these results show a comparable proportion
higher reaction temperatures (475 8C), the proportion of light of paraffins and olefins in the gas (C1–C4) and gasoline (C5–
products increased significantly to 37 wt% due to a more C12) fractions, but a relatively higher selectivity towards
intense cracking effect. A small proportion (around 3 wt%) of paraffins in the diesel (C13–C22) products. A small proportion
heavier hydrocarbon products (C23–C25) was also generated at of iso-paraffins was also generated, particularly when higher
all temperatures. (475 8C) cracking temperatures were employed, although their
Catalytic reforming over n-HZSM-5 caused a marked total selectivity was always below 3 wt%.
modification in the carbon number profile, with a significant
increase in the light (C1–C4) fraction and complete elimination
of diesel type (C13–C22) products. Thus, for a reaction
temperature of 450 8C, the total selectivity of C1–C4 products
reached 77 wt% while the gasoline (C5–C12) fraction amounted
to the remaining 23 wt%. A very similar product distribution by
carbon atom number was observed when the reaction was
conducted at lower temperatures (425 8C), which has been
attributed to the strong acid properties of this material. Similar
conversions but comparatively lower gas yields (50 wt%) had
been reported for direct catalytic cracking of LDPE over a
similar nanosized n-HZSM-5 zeolite when using a conventional
batch reaction system [14].
In contrast, owing to the weaker acid characteristics of Al-
MCM-41, catalytic reforming over this mesostructured solid
required the use of temperatures of 450 8C and above in order to
show a clear modification in the products profile. At lower
temperatures (425 8C), the selectivity by carbon atom number
was very similar to that obtained when the process was
conducted in the absence of catalyst, although the hydrocarbon
types were significantly different, as discussed below. LDPE
conversion over Al-MCM-41 at higher temperatures (450–
475 8C) led to a notable increase in the proportion of light
hydrocarbon products, particularly in the C2–C6 range, as well
as a marked reduction in the amount of diesel type
hydrocarbons. Still, compared to n-HZSM-5, conversion over
Al-MCM-41 presented a lower selectivity towards gas (C1–C4)
products (41 wt%) but a higher proportion of gasoline type (C5–
C12) hydrocarbons (56 wt%). The reforming activity of a
similar Al-MCM-41 was reported to be significantly lower
when it was used in direct catalytic cracking experiments using
a conventional batch reactor, with gas yields below 15 wt% in
all cases [14].

3.4.2. Characterisation by hydrocarbon types

Analysis of products by hydrocarbon type is essential to Fig. 5. Characterisation of products by hydrocarbon type for the thermo-
ascertain their potential value as fuels or refinery feedstock. catalytic reforming of LDPE at different temperatures.
 J. Aguado et al. / J. Anal. Appl. Pyrolysis 79 (2007) 415–423 421

For a reaction temperature of 450 8C, the most abundant As discussed above, conversion over n-HZSM-5 caused
paraffin in the C1–C4 fraction was n-butane, with a selectivity of complete elimination of the diesel (C13–C22) fraction and a
3.8 wt% while the most common olefins were propylene significant increase in the light (C1–C4) fraction. Analysis of
(2.5 wt%) and 1-butene (2.9 wt%). In the liquid fractions, the results obtained at 450 8C show that the highest selectivity
individual paraffins and a-olefins in the size range C5–C12 corresponds to C1–C4 olefins, particularly propylene
averaged selectivities around 1.5–3.0 wt% and lower values (32.9 wt%) but also ethylene (8.1 wt%) and 1-butene
were observed for heavier hydrocarbons. Without any further (6.4 wt%). A smaller proportion of gaseous paraffins (n-butane
treatment, the hydrocarbons described have a very limited 7.0%) and iso-paraffins (iso-butane 9.8%) was also detected in
application as transportation fuels or in the synthesis of organic that fraction.
products, but could be employed as refinery feedstock for In the gasoline fraction (C5–C12), the most abundant
further cracking and reforming. hydrocarbon types were aromatics (9.1 wt% of which
Results in Fig. 6 evidence that catalytic reforming over 4.0 wt% correspond to toluene), olefins (6.5 wt% of which
zeolite n-HZSM-5 generated a similar range of products 1.9 wt% were cis-2-pentene) and iso-paraffins (3.9 wt%).
regardless of the reaction temperature employed. This effect Traces of the unreformed paraffins, generated directly from
has been related to the strong acid character of the catalyst that the thermal cracking of LDPE, were also identified with a total
provided a catalytic activity sufficiently high to ensure a selectivity determined at 2.5 wt%.
complete reforming of the hydrocarbon products even at low Considering that catalytic reforming over n-HZSM-5 led to
(425 8C) temperatures. 23 wt% selectivity of gasoline type hydrocarbons and complete

Fig. 6. Characterisation of products by hydrocarbon type for the thermo- Fig. 7. Characterisation of products by hydrocarbon group type for the thermo-
catalytic reforming of LDPE over n-HZSM-5 at different temperatures. catalytic reforming of LDPE over Al-MCM-41 at different temperatures.
422 J. Aguado et al. / J. Anal. Appl. Pyrolysis 79 (2007) 415–423

absence of diesel products, it may be calculated that 40 wt% of Catalytic reforming over n-HZSM-5 significantly increase
the oil products generated correspond to aromatics, 29 wt% to the yield of gaseous products (over 74 wt%), that consisted
olefins and 17 % to iso-paraffins. Compared to commercial primarily of light olefins like propylene (33 wt% selectivity).
gasolines, the composition of this liquid fraction is significantly The process also resulted in complete elimination of diesel type
higher in aromatics but has a good mixture of olefins and iso- (C13–C22) hydrocarbons and the formation of a high proportion
paraffins, making it valuable for blending with other of aromatics and iso-paraffins in the gasoline (C5–C12) fraction,
commercial fuels. making this oil suitable for blending with conventional
Conversion of LDPE over Al-MCM-41 generated a different transportation fuels.
range of products to n-HZSM-5, which was attributed to the Compared to zeolite n-HZSM-5, catalytic reforming over
weaker acid character and wider pore dimensions of the former Al-MCM-41 generated a lower proportion of gaseous
catalyst. The results in Fig. 7 also evidence that the reforming hydrocarbons (58 wt% at 450 8C) but a comparatively higher
reaction was not complete at 425 8C, generating a different yield of gasoline type condensable products, which was
range of products to those observed at higher temperatures. attributed to its weaker acid properties and larger pore
Thus, catalytic conversion at 425 8C led to the formation of a dimensions. Catalytic conversion over this mesostructured
high proportion of n-paraffins, primarily concentrated in the material showed a lower selectivity towards the formation of
gasoline (C5–C12) and diesel (C13–C22) fractions, which aromatic products than n-HZSM-5 but resulted in the
corresponded to unreformed hydrocarbons from the initial formation of higher proportions of liquid iso-paraffins and
thermal cracking reaction. These results also evidence the olefins. The composition of this oil fraction is comparable to
higher reactivity of the a-olefins generated in the pyrolytic commercial gasoline fuel.
reaction which were reformed more efficiently than their
paraffinic counterparts.
Acknowledgements
When the reaction was conducted at 450 8C, the main
components in the light (C1–C4) fraction were olefins (38 wt%
The authors gratefully acknowledge financial support from
selectivity of which 15 wt% was propylene). Olefins were also
Ministerio de Ciencia y Tecnologia (project CICYT CTQ2005-
the main components in the gasoline fraction (23 wt%
09078). One of the authors, G. San Miguel, also wishes to thank
selectivity), although it should be noted that the majority of
Ministerio de Ciencia y Tecnologia for the scholarship granted
these alkenes did not correspond to the linear a-olefins
under the Ramon y Cajal programme.
generated from the thermal degradation of LDPE but to
secondary products derived from the reforming reactions. C5–
C12 hydrocarbons also contained high amounts of iso-paraffins References
(11 wt%) and paraffins (12 wt%). Comparatively, the propor-
tion of aromatics in the gasoline fraction was significantly [1] J. Aguado, D.P. Serrano, in: J.H. Clark (Ed.), Feedstock Recycling of
lower (4 wt%) than with n-HZSM-5, which was attributed to Plastic Wastes, The Royal Society of Chemistry, Cambridge, 1999.
[2] Association of Plastic Manufacturers in Europe (APME), Analysis of
the weaker acid character of the mesostructured solid. This
plastics consumption and recovery in Europe 2002–2003, Brussels,
weaker acidity is also responsible for the presence of relatively Belgium, 2004.
high amounts of unconverted paraffins, which appeared in all [3] A. Tukker, Plastic Waste Feedstock Recycling, Chemical Recycling and
three fractions. Incineration, Rapra Review Reports, vol. 3 (4), Report 148, Rapra
Technology Ltd., Shropshire, United Kingdom, 2002.
4. Conclusions [4] K. Pielichowski, J. Njuguna, Thermal Degradation of Polymeric
Materials, Rapra Technology Limited, Shropshire, United Kingdom,
2005.
A two-stage reaction system has been successfully tested for [5] P.T. Williams, E.A. Williams, J. Anal. Appl. Pyrolysis 51 (1999)
the conversion of LDPE into valuable hydrocarbon products, 107.
using a nanosized n-HZSM-5 and a mesostructured Al-MCM- [6] W. Kaminsky, J.-S. Kim, J. Anal. Appl. Pyrolysis 51 (1–2) (1999)
127.
41 for the catalytic reforming of the vapours generated in the
[7] A. Okuwaki, Polym Degrad. Stab. 85 (3) (2004) 981–988.
initial pyrolytic step. Conversion values were determined by the [8] F. Pinto, P. Costa, I. Gulyurtlu, I. Cabrita, J. Anal. Appl. Pyrolysis 51
temperature in the thermal reactor while the conditions in the (1999) 39.
catalytic step controlled the extent of the reforming reaction to [9] D.P. Serrano, J. Aguado, J.M. Escola, E. Garagorri, Appl. Catal. B. Env. 44
generate the final products. (2003) 95.
The best results were obtained at 450 8C where the [10] W. Kaminsky, J. Anal. Appl. Pyrolysis 8 (1985) 439–446.
[11] P.T. Williams, P.A. Horne, D.T. Taylor, J. Anal. Appl. Pyrolysis 25 (1993)
conversion of LDPE into hydrocarbon products was above 325–334.
90 wt% for a 2 h reaction period. In the absence of catalysts, [12] A. Corma, Chem. Rev. 95 (1995) 559–614.
thermal conversion of LDPE generated linear paraffins and a- [13] G. Manos, I.Y. Yusof, N. Papayannakos, N.H. Gangas, Ind. Eng. Chem.
olefins over a wide range of molecular weights (between C1– Res. 40 (10) (2001) 2220–2225.
C24) as only products. At that temperature, the main selectivity [14] D.P. Serrano, J. Aguado, J.M. Escola, Ind. Eng. Chem. Res. 39 (2000)
1177–1184.
was on liquid hydrocarbons (75 wt%), while only 16 wt% [15] D.P. Serrano, J. Aguado, J.M. Escola, J.M. Rodriguez, Stud. Surf. Sci.
corresponded to gas products, 1 wt% to solid waxes and 8 wt% Catal. 142-B (2002) 1671–1678.
to unconverted plastic. [16] P.A. Jalil, J. Anal. Appl. Pyrolysis 65 (2002) 185–195.
 J. Aguado et al. / J. Anal. Appl. Pyrolysis 79 (2007) 415–423 423

[17] J. Aguado, D.P. Serrano, J.M. Escola, J.M. Rodriguez, G. San Miguel, J. [21] J. Aguado, D.P Serrano, G. San Miguel, J.M. Escola., J.M. Rodriguez, J.
Anal. Appl. Pyrolysis 73 (2004) 79–87. Anal. Appl. Pyrolysis, in press.
[18] T.F. Degnan, J. Catal. 216 (2003) 32–46. [22] R. Bagri, P.T. Williams, J. Anal. Appl. Pyrolysis 63 (2002) 29–41.
[19] G. Tasi, I. Pálinkó, A. Molnár, I. Hannus, J. Mol. Struct. (THEOCHEM) [23] R. Van Grieken, J.L. Sotelo, J.M. Menéndez, J.A. Melero, Microporous
666–667 (2003) 69–77. Mesoporous Mater. 39 (2000) 135.
[20] ICIS, Reed Business Information Limited, Quadrant House, The Quad- [24] J. Aguado, D.P. Serrano, J.M. Escola, Microporous Mesoporous Mater. 34
rant, Sutton, Surrey. http://www.icispricing.com, April 2006. (2000) 43–54.