Fuel 285 (2021) 119143

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Catalytic co-pyrolysis of polycarbonate and polyethylene/polypropylene

mixtures: Promotion of oil deoxygenation and aromatic
hydrocarbon formation
Kai Sun a, Wanli Wang a, Nickolas J. Themelis b, A.C. Thanos Bourtsalas b, Qunxing Huang a, *
a
State Key Laboratory of Clean Energy Utilization, Zhejiang University, Hangzhou 310027, People’s Republic of China
b
Earth Engineering Center, Columbia University, New York, NY 10027, USA

A R T I C L E I N F O A B S T R A C T

Keywords: The co-pyrolysis of polycarbonate (PC) and polyethylene (PE)/polypropylene (PP) in the presence of HZSM-5
Co-pyrolysis catalyst was studied in a two-stage tube reactor with the objective of promoting simultaneous deoxygenation
Polycarbonate and aromatic hydrocarbon formation, at atmospheric pressure and without external hydrogen addition. As the
Polyethylene
blending ratio of PE/PP increased, the yield of oil phase product, coke, and solid decreased, and the yield of gas
Polypropylene
phase increased significantly. At the blending ratio of 75% PP, the concentration of oxygenates in the oil was
Aromatic hydrocarbon
Deoxygenation reduced to 2.9% and that of aromatic hydrocarbons increased to 97.1%. Increasing the reaction temperature
from 500 ◦ C to 700 ◦ C increased the yield of aromatic hydrocarbons further to 98.1%. Catalyst regeneration
decreased the acidity of the HZSM-5 catalyst, resulting in a slight decrease in the deoxygenation and aromati­
zation effects. The results showed that PP is an effective hydrogen donor in the catalytic deoxygenation and
aromatization process; the hydrogen provided enhanced the adsorption of phenols produced by the pyrolysis of
PC on HZSM-5 and their direct deoxygenation to form new aromatic hydrocarbons.

1. Introduction the main constituents of the oil were 32% phenols, 16% bisphenol A, and
5.6% p-isopropylphenol. The presence of various oxygenates such as
Polycarbonate (PC) is a thermoplastic resin widely used in electrical phenol leads to an increase in the viscosity, corrosivity, and thermal
and electronic appliances, packaging, manufacturing, etc. [1,2]. The instability of the pyrolysis oil and a decrease in its calorific value [8].
total global PC demand was up to 4.1 million tons in 2017 [3], which This is particularly critical when the pyrolysis oil is used for fuel or
made it one of the main oxygen-containing plastics. Most plastic prod­ chemical production. Therefore, the deoxygenation of oxygenate-
ucts are non-degradable and readily discarded after use, and usually, enriched oil is one of the grand challenges in oil upgrading at present
their lifetimes are no more than one month [4,5], which leads to the [9].
accumulation of post-recycling plastic wastes. It is neither economical Hydrodeoxygenation (HDO) is the main deoxygenation pathway for
nor environmentally friendly to dispose waste PC products by traditional oxygenates such as phenols, mainly including hydrogenation-
incineration methods due to its high oxygen content and low calorific dehydration (HYD) route and direct deoxygenation (DDO) routes [10].
value. Compared with HYD, DDO may convert phenols to aromatic hydrocar­
“Plastic-to-oil”, as a cost-effective valorization technology, has bons by direct cleavage of aromatic C-O without hydrogenating the ar­
gained continuously increasing attention in recent years. This technol­ omatic ring [11]. Thus, DDO is a better deoxygenation route, with low
ogy usually uses pyrolysis to realize the conversion of plastic wastes to hydrogen consumption and maximized carbon retention in target
fuel oil like jet fuel and diesel-like oil or high-value chemicals, which products [12]. Also, the resulting products, aromatic hydrocarbons like
could alleviate the problem of fossil oil depletion at relatively low cost benzenes and naphthalenes, are in high demand, as crucial additives or
[6]. Some researchers have summarized the pyrolysis behavior of PC in raw materials in the fuel, chemical, pharmaceutical, printing, and
previous studies. Achilias et al. [7] found that pyrolysis of PC at 550 ◦ C dyeing industries [13,14]. Especially in the field of fuel production,
yielded 6.6% pyrolysis gas, 63% oil products, and 30.4% solid residues; aromatics could act as important octane enhancers and are significant

* Corresponding author.
E-mail address: hqx@zju.edu.cn (Q. Huang).

https://doi.org/10.1016/j.fuel.2020.119143
Received 11 June 2020; Received in revised form 5 August 2020; Accepted 31 August 2020
Available online 21 September 2020
0016-2361/© 2020 Elsevier Ltd. All rights reserved.
K. Sun et al. Fuel 285 (2021) 119143

parts of all the fuels, the content of which in gasoline, diesel, and jet fuel collect the pyrolytic vapor. After the experiment, dichloromethane was
could be up to 35%, 30%, and 25%, respectively [15]. In the traditional used to wash the pipes and other equipment. The solution was then dried
commodity chemical industry, the cost of raw materials usually ac­ at 40 ◦ C for 10 h to remove the solvent, and the remaining liquid was
counts for 60–70% of the total cost [16]. Therefore, recovering aromatic considered as the oil product. The weight gain of the catalyst was
hydrocarbons from wastes would bring remarkable environmental and considered to be due to coke formation. Pyrolytic char in the first stage
economic benefits. Nelson et al. [17] investigated the DDO of phenol was considered to be the solid residue. The gas yield was calculated by
over Ru/TiO2 catalyst, in the presence of H2O at 573 K and 550 psig H2, mass difference. All pyrolysis experiments were conducted at least twice
and the conversion rate reached 30%. Yoosuk et al. [18] compared the and the average value was used for analysis. Predicted values were ob­
DDO and HYD routes of phenol, catalyzed by MoS2 and CoMoS2, and a tained by calculating the linear superposition of individual composition
maximum conversion rate of 83.4% was reached at 350 ◦ C and 2.8 MPa without considering any synergistic effects.
H2. Zhu et al. [19] used bi-functional Pt/HZSM-5 in the DDO of m-Cresol An Agilent 7890B-5977A gas chromatography-mass spectrometry
at 400 ◦ C with an H2/m-cresol molar ratio of 50, and the yield of aro­ (GC–MS) equipped with an HP-5 column (50 ◦ C (4 min) → 10 ◦ C/min →
matic hydrocarbons was up to 99%. It should be noted that these studies 270 ◦ C (10 min)) was used for the analysis of oil. Thirty oil components
on catalytic deoxygenation of phenols used external H2 as the hydrogen with the highest content were used for analysis. The content of oil
source and rare metals as catalysts. components was determined by the semi-quantitative analysis method
In more recent studies, hydrogen-rich materials instead of H2 were and the proportion of each peak area (area%) in the chromatogram was
used as hydrogen donors, and DDO was carried out in catalytic co- considered as the content of the corresponding component in the oil,
pyrolysis, using less costly catalysts. Xue et al. [20] carried out the which was calculated by the following formula:
catalytic co-pyrolysis of lignin and tetralin, using HZSM-5 and HY
Ai
zeolite catalysts, and the yield of aromatic hydrocarbons was increased Content(area% ) = × 100% (1)
AT
from 48.8% to 66.2%. Ding et al. [21] studied the co-pyrolysis of corn
stove and polyethylene (PE) catalyzed by CeO2 and HZSM-5, and the where Ai is the peak area of the ith oil composition; AT is the total
deoxygenation efficiency increased by around 13%. However, there has chromatographic peak area.
been no systematic analysis on the low-cost and high-yield catalytic A TA-Q500 thermalgravimetric analyzer was used to analyze the
hydrodeoxygenation of PC co-pyrolysis products. thermal decomposition characteristics of the plastic samples and the
In this study, PE and polypropylene (PP) were used as hydrogen coke on the catalyst. For plastic samples, 8 mg of the sample was heated
donors, and HZSM-5 was used in the catalytic hydrodeoxygenation of to 850 ◦ C at a ramp rate of 10 ◦ C/min under 60 ml/min of nitrogen flow.
the PC oil produced by the co-pyrolysis of PC with PE and PP, two major The thermal decomposition characteristic of coke formed was deter­
components of waste plastics, used as cheap hydrogen sources [22]. mined by thermogravimetric analysis at the same temperature program,
HZSM-5 is a commercial catalyst with high catalytic aromatization under 60 ml/min of air flow. A VG ESCALAB MKII X-ray photoelectron
performance and high deoxygenation activity [23]. The co-pyrolysis spectrophotometer was used to analyze the surface chemical composi­
behavior and the possible hydrodeoxygenation mechanism of PC and tion of the coke. The binding energy of C 1 s peak was calibrated to
PE/PP were explored in detail. The results of this study can provide new 284.6 eV before analysis.
ideas for the upgrading of oils with high oxygen content produced by the
pyrolysis of non-recycled plastics. 3. Results and discussion

2. Materials and methods 3.1. Thermal decomposition characteristics of various blends of PC and
PE/PP
2.1. Materials and catalyst
The TG and DTG curves of PC and PE/PP are shown in Fig. 1, and the
PC was purchased from Sigma-Aldrich, US. PE and PP were obtained comparison between experimental and predicted curves of various
from Guanbu Electromechanical Technology Company, China. The as- blends, in Fig. S1. The DTG curves showed that as the percentage of PE/
received plastics were ground and sieved to a particle size less than PP increased from 0 to 100%, the maximum weight loss rate increased,
0.18 mm (80 mesh). as well as the peak temperature, from 423.2 ◦ C to 462.1/441.4 ◦ C,
HZSM-5 powder (100–200 mesh, Si/Al = 28), provided by Tianjin respectively. This indicated that PE and PP have higher thermal
Yuanli Chemical Company, China, was used as the catalyst in this study. decomposition temperatures and narrower weight loss temperature
The as-received HZSM-5 was activated at 550 ◦ C for 4 h in a tubular ranges than PC. The TG data showed that the extrapolated onset and
furnace in air. offset temperatures of PC, PE, and PP are 401.7–454.9 ◦ C,
449.8–471.9 ◦ C, 425.1–459.3 ◦ C, respectively. Compared with PE, PP
2.2. Experimental setup and analysis methods had a higher coincidence degree with PC in the weight loss temperature
range, which was more conducive to mixing, and interaction of their
A two-staged tubular furnace was used to conduct the catalytic co- volatiles. As can be seen from the TGdifference (experimental value –
pyrolysis of PC and PE/PP, in atmospheric pressure. For each experi­ predicted value) curves for the blend of PC and PE, there was a positive
ment run, the blend of PC and PE/PP with a total weight of 4 g was peak at 425.9 ◦ C when the percentage of PE was 25%, which was close to
placed in the first stage (pyrolysis section). The weight ratio of PE/PP in the peak temperature of the PC, thus indicating that the presence of PE
the mixture was set at 0, 25, 50, 75, and 100%, respectively. 6 g of hindered the pyrolysis of PC. This was probably because the relatively
HZSM-5 was placed in the second staged (catalytic section) and fixed by low pyrolysis rate of PE at this temperature led to the melting of PE
quartz wool. As control groups, only quartz wool was placed in the before decomposition and this hindered the release of volatiles from PC.
catalytic section, and some experiments were performed under the same The negative peak near 456 ◦ C indicated that the interaction between PC
experimental settings as above, which were considered as thermal and PE advanced the decomposition of PE; this effect became stronger at
cracking groups. Before each experiment, the reaction tube was pre- higher PE content. For the co-pyrolysis of PC and PP, the difference
filled with nitrogen, and then the nitrogen flow rate was adjusted to between the experimental and predicted curves was much smaller
80 ml/min. The second stage was heated and maintained at the catalytic (maximum 1.19%), than that of PC and PE (maximum 13.7%). The
temperature; then the first stage was heated at a ramp rate of 20 ◦ C/min negative peak around 420 ◦ C indicated that the interaction between PC
to a final temperature of 550 ◦ C and held there for 30 min. and PP decreased the decomposition temperature. This was most
A series of ice water-cooled condensing devices were used to trap and

2
K. Sun et al. Fuel 285 (2021) 119143

Fig. 1. TG, DTG, and TGdifference curves (differences between experimental and predicted TG curves) for various blends of PC and PE/PP.

obvious at the PP blending ratio of 75%. Also, the interaction between decarboxylation, which led to an increase in the yield of gas-phase
PC and PE/PP led to a reduction in the yield of solid residues. products such as CO2 [24]. With regard to PE and PP, the acid sites
promoted the dehydrogenation processes and resulted in an increase of
3.2. Effect of HZSM-5 on product distribution. the H2 concentration [25]. Alkanes and alkenes produced by PE or PP
would undergo secondary thermal cracking, in the absence of the
The effect of HZSM-5 on the pyrolysis of each plastic sample at catalyst in the second stage, and form aromatic hydrocarbons [26], as
600 ◦ C was studied in this section. As can be seen in Fig. 2(a), the shown in Fig. 2(b) and (c). The presence of HZSM-5 promoted aroma­
addition of HZSM-5 reduced the oil yield and increased the gas yield. For tization processes by catalyzing dehydrogenation, cyclization, and
PC, the acidity of HZSM-5 promoted deoxygenation reactions, e.g., hydrogen transfer reaction, and greatly improved the selectivity of

3
K. Sun et al. Fuel 285 (2021) 119143

Fig. 2. Effect of HZSM-5 on (a) product yield, (b) oil components, and composition of (c) aromatic hydrocarbons, (d) oxygenates of PC, PE, and PP.

bicyclic aromatic hydrocarbons by promoting condensation process 3.3. Effect of PE/PP blending on product distribution.
[27]. Phenols were the main pyrolysis products (84.4%) of PC. HZSM-5
promoted the deoxygenation of some oxygenates (mainly phenols), but In this section, the catalytic temperature was set at 600 ◦ C. It should
the conversion rate was only 24.3%. This was because the cleavage of be noted that for each of the 25%, 50% and 75% blends they include two
aromatic C-O bond was much harder than that of aliphatic C-O bonds columns, the first, “exp.”, is the actual experimental value; the second,
[28]. “pre.”, is the predicted value based on the linear superposition of 0% and
The yield of coke on the catalyst in group PC-HZSM-5 was 9.8%, 100% PE/PP experimental value. For example:
which was significantly higher than the coke yield of PE and PP (1.1%
Predicted value25%PE+75%PC = 25% × Experimental value100%PE+0%PC + 75%
and 2.2%, respectively). It was reported that although the strong acidity
of HZSM-5 was favorable for coke formation, its microporous structure × Experimental value0%PE+100%PC
can hinder the development of coke precursors by enhancing the (2)
transfer of small molecules (light alkenes and wax) to the external sur­
The objective was to determine if there is an “interaction” effect
face of the catalyst and hindering contact between the polycyclic aro­
when both PE/PP and PC are present. As the blending ratio of PE/PP
matic structure and the active sites inside the catalyst [29–31]. For PC,
increased, the yield of oil phase product, coke, and solid decreased, and
phenols were liable to form heavy ethers by acid site-catalyzed
the yield of gas phase increased significantly. This indicates that the
condensation or heavy phenolic compounds by polymerization, which
pyrolysis products were becoming lighter, due to the high H/C ratio of
were important precursors for coke formation [14,32]. While for PE and
PE/PP [33]. The interaction of PC with PE or PP slightly reduced the
PP, the formed aromatic hydrocarbons were mainly monocyclic
yield of solid residues, while the coke yield slightly increased, as can be
(52.6–77.4%), which were not as active as olefinic coke precursors and
seen from the comparison of experimental and predicted values in Fig. 3
would not block pore access, as polycyclic aromatics do.
(a) and (b). The addition of PE/PP provided a large number of hydrogen

4
K. Sun et al. Fuel 285 (2021) 119143

Fig. 3. Effect of PE or PP blending with PC on (a) and (b) product yield; (c) and (d) oil composition; (e) and (f) aromatic hydrocarbons in oil and (g) and (h)
oxygenates in oil.

5
K. Sun et al. Fuel 285 (2021) 119143

free radicals during the thermal cracking of PC, thus promoting the (10.8% higher than the predicted value). When the percentage of PP
conversion of intermediates to lighter products and reducing the re­ reached 75%, the deoxygenation effect of all oxygenates (phenols, fu­
actants involved in the dehydrocondensation reactions [34]; while the rans, ethers, etc.) was obvious. The content of aromatic hydrocarbons in
active alkenyl radicals from PE or PP enhanced the reactivity of the coke the oil product reached 97.1%, of which 54.5% were bicyclic aromatic
precursors and therefore led to the slight increase in the coke yield. hydrocarbons, mostly naphthalenes.
With the increase of the PE/PP blending ratio, the content of aro­ Benzenes and naphthalenes were the high value-added aromatic
matic hydrocarbons increased significantly. At a low PE/PP blending components in the oil produced by co-pyrolysis (Fig. 4). Naphthalene
ratio, the interaction between the PC and PE/PP appeared to inhibit the and methylnaphthalene were the main components in thermal cracking
formation of aromatic hydrocarbons (experimental value < predicted oils. The content of most aromatic hydrocarbons increased along with
value), but at high blending ratio, the effect was opposite. This indicated the blending ratio of PE/PP. However, toluene showed the highest
that the high PE/PP blending ratio was favorable for deoxygenation selectivity when the blending ratio of PE/PP reached 75%. When the
reactions. When the PE/PP percentage was relatively low, the amount of blending ratio of PP reached 75%, the concentrations of the two main
hydrogen produced during the dehydrogenation processes was limited. components of the oil, naphthalene, and methylnaphthalene, were close
In this case, the phenolic hydroxyl group cannot be effectively removed to the maximum. Also, dimethylnaphthalenes, trimethylbenzenes, and
because of the scarcity of hydrogen radicals [10]; HZSM-5 can effec­ trimethylnaphthalenes appeared only when the PE/PP blending ratio
tively promote the alkylation of phenols [35], which would consume reached 50%, 75%, and 100%. These phenomena indicated that there
alkyl and alkenyl radicals, thereby reducing the reactants in the were differences between the aromatic formation pathways in the cat­
aromatization process. At relatively high PE/PP ratios, there was suffi­ alytic co-pyrolysis and that in the catalytic pyrolysis of PE/PP alone. PC
cient hydrogen for the deoxygenation of phenol through the DDO tended to form less-branched aromatic hydrocarbons, while PE and PP
pathway, with PE/PP as hydrogen donors and aromatic hydrocarbons as favored the formation of branched ones.
products [10].
For the co-pyrolysis of PC and PE, since the main weight loss range of 3.4. Effect of catalytic temperature on product distribution.
PC appeared earlier than that of PE, phenol radicals cannot efficiently
contact with hydrogen radicals. As can be seen in Fig. 3(e) and (g), when From the results in the previous section, the effect of catalytic
PE percentage was 25% and 50%, the experimental values of monocyclic deoxygenation was the best when the weight ratio of PC and PP in the
aromatic hydrocarbons were 8.8% and 14.5% lower than the predicted blend was 25%: 75%. Thus, this blend was used to study the effect of
values, while the experimental values of phenol content were 9.2% and catalytic temperature on product yield and oil composition. It can be
2.4% higher than the predicted values. The interaction between PC and seen from Fig. 5(a) that the increase of the catalytic temperature from
PE slightly promotes the deoxygenation of phenols (0.3%) and the for­ 500 ◦ C to 700 ◦ C slightly reduced the yield of gas and oil products by
mation of monocyclic aromatic hydrocarbon (2.5%) only when the 1.5% and 2.0%, while the coke on catalyst increased by 3.6%. Higher
blending ratio of PE was as high as 75%. temperatures favored the conversion of coke precursors to more
For the co-pyrolysis of PC and PP, the coincidence degree of their condensed structures [36], which led to the evolution of heavy oil-phase
main weight loss ranges was much higher. Compared with those of PE, products into coke. Meanwhile, increasing temperature promoted the
pyrolysis products of PP contained more tertiary carbon, which was Diels-Alder reaction catalyzed by HZSM-5, through which aromatic
beneficial to the removal of oxygenated side chains on aromatics [34]. hydrocarbons were produced under the consumption of light hydro­
When the blending ratio of PP reached 50% and 75%, their interaction carbons [37]. This also led to a slight reduction in gas yield. Also, the
greatly improved the deoxygenation effect of oxygenates by 12.6% and higher temperature favored the production of phenyl-end groups by
15.2%, respectively, as shown in Fig. 3(d). Fig. 3(f) and (h) showed that molecular rearrangement or further decomposition of PC [38], which
when the percentage of PP reached 50%, deoxygenation was mainly due led to a further reduction in the phenol concentration (Fig. 5(d)). Since
to the dehydroxylation of phenol; branched phenols were firstly deox­ the yield of aromatic hydrocarbons at 500 ◦ C was already very high (up
ygenated and then further aromatized under the effect of acid sites, to 95.8%), the increase in the yield of aromatic hydrocarbons with the
leading to the formation of mainly bicyclic aromatic hydrocarbons temperature was less pronounced (Fig. 4(b)).

Fig. 4. Effect of PE and PP blending on the evolution of major aromatic hydrocarbons.

6
K. Sun et al. Fuel 285 (2021) 119143

Fig. 5. Effect of catalytic temperature on (a) product yield, (b) oil components, (c) concentration of aromatic hydrocarbons, and (d) oxygenates from the blend of
25% PC and 75% PP.

Fig. 5(c) shows that increasing temperature increases significantly aromatization processes, like the Diels-Alder reaction, which consumed
the condensation degree of aromatic hydrocarbons. When the temper­ light hydrocarbons. This led to a decrease in oil yield and an increase in
ature increased from 500 ◦ C to 700 ◦ C, the selectivity of monocyclic gas yield. At the same time, the condensation reaction was also sup­
aromatic hydrocarbons decreased from 33.7% to 15.6%, while the pressed, which led to a decrease in coke yield. As can be seen in Fig. 6(b)
content of tricyclic aromatic hydrocarbons increased significantly from and (d), as the efficiency of deoxygenation reactions catalyzed by HZSM-
6.8% to 20.1%. The content of tetracyclic aromatic hydrocarbons started 5 decreased [10], the content of phenols increased significantly from
to increase appreciably over 600 ◦ C. Considering both the yield and the 2.9% to 7.8%. For aromatic hydrocarbons, the content of monocyclic
composition of aromatic hydrocarbons, it is better to operate at a cata­ and bicyclic ones didn’t change significantly, while the content of tri­
lytic temperature between 500 and 600 ◦ C. cyclic aromatic hydrocarbons was reduced from 11.6% to 7.3%. Tetra­
cyclic aromatic hydrocarbons disappeared after regeneration. These also
proved the decrease in the acidity of HZSM-5. However, in general, the
3.5. Catalyst regeneration and coke characterization.
catalyst regenerated twice had no significant effect on the product yield
and the aromatic hydrocarbon selectivity, indicating HZSM-5 had the
In order to investigate the lifetime of HZSM-5 in catalytic co-
excellent activity and stability in this catalytic reaction.
pyrolysis, the effect of catalyst regeneration times on product yield
Fig. 7 showed the TG and DTG analysis of the coked HZSM-5 in the
and oil compositions was analyzed at 600 ◦ C (Fig. 6). With the increase
air atmosphere. The DTG curves of all used HZSM-5 had two peaks. The
of the regeneration time from zero to twice, the yield of oil and coke
first peak at around 75 ◦ C represented the volatilization of moisture. The
slightly decreased by 1.8% and 3.7%, respectively, and the yield of gas
second peak at around 430–620 ◦ C was associated with the coke
products increased by 5.1%. The surface acidity and the ratio of strong
oxidation. According to Bauer and Karge’s classification, these cokes all
acid sites / weak acid sites decreased because of the dealumination and
match type II, which are highly condensed polyaromatic residues with
dehydroxylation of the catalyst during regeneration [16,39]. The
an H/C ratio usually less than 0.8 [40]. Moreover, the increase of the
decrease in the acidity led to a decrease in the efficiency of the

7
K. Sun et al. Fuel 285 (2021) 119143

Fig. 6. Effect of catalyst regeneration on (a) product yield, (b) oil component, (c) concentration of aromatic hydrocarbons, and (d) oxygenates from the blend of 25%
PC and 75% PP.

maximum peak temperature resulted in a higher condensation degree C–O, C=O, and O–C– – O, respectively [42,43]. Carbon mainly existed in
and a lower H/C ratio of polyaromatic structure in coke [29,41]. Coke the form of C–C, which was derived from the aromatic structure in coke.
oxidized at higher temperatures was less active and had higher thermal As can be seen in Fig. 8(c), (d), and (e), As the temperature increased, the
stability, thus more difficult to be removed during the regeneration of content of carbon combined with oxygen gradually increased. This may
catalyst [40]. Fig. 7(a) shows that with the increase of catalytic tem­ be because the condensation reaction of phenols was accelerated at the
perature, the amount and the condensation degree of the coke also high temperature, which generated more oxygenated polyaromatic
increased. This increase became more pronounced at temperatures structures in coke than in oil. As shown in Fig. 8(b) and (d), the pro­
above 600 ◦ C. Fig. 7(b) indicates that the coke formed in the PC + PP portion of carbon combined with oxygen was higher in the PC + PE
group was more active than that in the PC + PE group. This was prob­ group than in PC + PP. This also indicated that the deoxygenation effect
ably because the interaction between PP and PC was more pronounced in the PC + PP group was better.
and introduced more hydrogen into the pyrolysis products of PC. As
shown in Fig. 7(c), with the increase of the regeneration time, there was 3.6. Catalytic co-pyrolysis mechanism
no significant change in the amount and properties of coke. However,
the increase in regeneration times resulted in a significant decrease in The possible reaction pathways of catalytic co-pyrolysis of PC and
moisture absorption, which may be due to the incomplete removal of the PE/PP are illustrated in Fig. 9. The effective conversion of highly
coke that clogged the micropores of the catalyst. oxygenated pyrolysis oil from PC to aromatic hydrocarbon-rich oil was
Fig. 8(a) shows the wide scan XPS spectra of used HZSM-5. It can be the result of the combined effect of the addition of PP/PE, with high H/C
seen that the signal of the C 1s peak, due to coke deposition gradually ratios, and the deoxygenation activity of HZSM-5. The pyrolysis of PC
increased with increasing temperature, while the signal of the O 1s peak, mainly yielded phenol and iso-propenyl phenol through thermal
mainly due to oxygen in the framework of HZSM-5, decreased. In the cracking, disproportionation, and CH-transfer [38]. Some secondary
high-resolution C 1s spectra, peaks at 282.5, 284.6, 286.1, 287.3, and reactions (such as rearrangement and cyclization reactions, etc.) also
288.4 eV could be attributed to dehydrogenated carbon (DHC), C–C, promote the formation of some aromatic ethers and benzofurans [44].

8
K. Sun et al. Fuel 285 (2021) 119143

Fig. 7. TG and DTG curves of coke deposition on HZSM-5 at (a) different catalytic temperatures (feedstock: blends of 25% PC and 75% PP), (b) different plastic
compositions (catalytic temperature: 600 ◦ C), and (c) different regeneration times (feedstock: blends of 25% PC and 75% PP, catalytic temperature: 600 ◦ C).

When the HZSM-5 catalyst participated in the pyrolysis, the acid sites on large number of hydrogen radicals. Since the weight loss temperature
HZSM-5 promoted secondary reactions, like alkylation, transalkylation, range of PC and PP were similar, hydrogen radicals can be fully con­
isomerization, and dehydration of phenol, which accelerated the for­ tacted with oxygenates from PC. Due to the acidity of HZSM-5, hydrogen
mation of phenol pools containing various (branched) phenols radicals could enhance the adsorption of phenols on the surface of
[32,35,45]. It was found that phenol would be converted to phenoxyl, HZSM-5, and promote the direct hydrodeoxygenation of adsorbed
aryl, and hydrogen radicals under the effect of zeolites [46]. Aromatic phenols by dehydration without involving decarbonylation, which
hydrocarbons could be formed both from the decarbonylation of inter­ could avoid unnecessary carbon loss [20]. Besides, light hydrocarbons
mediate products (like 1,4-benzoquinone and cyclopenta-2,4-dien-1- produced by PE/PP pyrolysis can also participate in the Diels-Alder re­
one) formed by the interaction among phenolic radicals and by the action of benzofurans, which promoted its conversion to aromatic
interaction between hydrogen radicals and aryl radicals [47]. However, hydrocarbons.
the direct conversion of phenols to aromatic hydrocarbons on HZSM-5
was very limited, and the yield of aromatic hydrocarbons was low 4. Conclusions
[14,48]. This was because the high bond energy of the phenolic C–O
bond resulted in the high stability and low reactivity of phenolic hy­ Catalytic co-pyrolysis of PC and PE/PP in the presence of HZSM-5
droxyl groups [28,49]. HZSM-5 can also promote the conversion of catalyst was conducted with the objective of promoting simultaneous
furan to aromatic hydrocarbons through Diels-Alder reaction [37], but deoxygenation and aromatic hydrocarbon formation. The effect of
the reaction was difficult to perform due to the lack of light alkenes as HZSM-5, the PE or PP to PC blending ratio, catalytic temperature, and
reactants. catalyst regeneration on the product yield and oil components were
PE and PP are considered as hydrogen-rich materials. However, the investigated. The main conclusions are as follows:
thermal cracking of PE/PP mainly produced hydrocarbons (both in the
gas phase and the oil phase) [50,51], and the number of hydrogen (1) The phenols in the oil produced by the pyrolysis of PC accounted
radicals generated was limited [52]. The addition of HZSM-5 promoted for 84.4%. HZSM-5 promoted the deoxygenation of phenols to
a series of aromatization reactions (dehydrogenation, cyclization, Diels- aromatic hydrocarbons, but the conversion rate was only 24.3%.
Alder reaction, and hydrogen transfer reaction) of the chained hydro­ (2) Blending PP with PC had a more significant contribution to the
carbons by the effect of both Brønsted and Lewis acid sites [26,27]. The deoxygenation effect than blending PE with PC. When the PE
dehydrogenation process in these aromatization reactions can provide a blending ratio was increased from 25% to 75%, the difference

9
K. Sun et al. Fuel 285 (2021) 119143

Fig. 8. Wide scan XPS spectra for (a) used HZSM-5 and high resolution C 1s spectra for used HZSM-5 under different experimental conditions: (b) 25% PC + 75% PE,
600 ◦ C; (c) 25% PC + 75% PP, 500 ◦ C; (d) 25% PC + 75% PP, 600 ◦ C; (e) 25% PC + 75% PP, 700 ◦ C.

10
K. Sun et al. Fuel 285 (2021) 119143

Fig. 9. Reaction pathways of catalytic co-pyrolysis of PC and PE/PP.

between the experimental and the predicted values of the aro­ decrease of the catalyst acidity and the ratio of strong acid sites
matic content changed from − 10.2% to 2.4%; for the same /weak acid sites.
change in the PP to PC ratio, this difference changed from − 3.3% (6) The HZSM-5 catalyst promoted the aromatization of PE/PP py­
to 15.2%. rolysis products due to the effect of acid sites which can provide a
(3) At the blending ratio of 75% PP, the concentration of aromatic large number of hydrogen radicals. Hydrogen radicals on the
hydrocarbons reached 97.1%, in which the proportion of bicyclic surface of HZSM-5 enhanced the adsorption of phenols produced
aromatic hydrocarbons reached 54.5%. by PC and promoted the direct hydrodeoxygenation of adsorbed
(4) The increase of the reaction temperature from 500 ◦ C to 700 ◦ C phenols by dehydration.
slightly increased the content of aromatic hydrocarbon in oil from
95.8% to 98.1%, at the blending ratio of 75% PP; increasing CRediT authorship contribution statement
temperature also led to the formation of more polycyclic aromatic
hydrocarbons. Kai Sun: Software, Validation, Formal analysis, Investigation,
(5) Increasing the regeneration time from zero to twice decreased the Writing - original draft. Wanli Wang: Investigation, Writing - original
deoxygenation and aromatization reactions by 4.9%, due to the draft. Nickolas J. Themelis: Writing - review & editing, Project

11
K. Sun et al. Fuel 285 (2021) 119143

administration, Visualization. A.C. Thanos Bourtsalas: Data curation, [17] Nelson RC, Baek B, Ruiz P, Goundie B, Brooks A, Wheeler MC, et al. Experimental
and Theoretical Insights into the Hydrogen-Efficient Direct Hydrodeoxygenation
Writing - review & editing. Qunxing Huang: Conceptualization,
Mechanism of Phenol over Ru/TiO2. ACS Catal 2015;5:6509–23.
Methodology, Resources, Supervision, Funding acquisition. [18] Yoosuk B, Tumnantong D, Prasassarakich P. Unsupported MoS2 and CoMoS2
catalysts for hydrodeoxygenation of phenol. Chem Eng Sci 2012;79:1–7.
[19] Zhu X, Nie L, Lobban LL, Mallinson RG, Resasco DE. Efficient Conversion of m-
Declaration of Competing Interest Cresol to Aromatics on a Bifunctional Pt/HBeta Catalyst. Energy Fuels 2014;28:
4104–11.
[20] Xue Y, Zhou S, Bai X. Role of Hydrogen Transfer during Catalytic Copyrolysis of
The authors declare that they have no known competing financial Lignin and Tetralin over HZSM-5 and HY Zeolite Catalysts. ACS Sustainable Chem
interests or personal relationships that could have appeared to influence Eng 2016;4:4237–50.
[21] Ding K, He A, Zhong D, Fan L, Liu S, Wang Y, et al. Improving hydrocarbon yield
the work reported in this paper. via catalytic fast co-pyrolysis of biomass and plastic over ceria and HZSM-5: An
analytical pyrolyzer analysis. Bioresour Technol 2018;268:1–8.
[22] Fan L, Chen P, Zhang Y, Liu S, Liu Y, Wang Y, et al. Fast microwave-assisted
Acknowledgments
catalytic co-pyrolysis of lignin and low-density polyethylene with HZSM-5 and
MgO for improved bio-oil yield and quality. Bioresour Technol 2017;225:199–205.
The authors would like to gratefully acknowledge the National Key [23] Bayat A, Sadrameli SM, Towfighi J. Production of green aromatics via catalytic
cracking of Canola Oil Methyl Ester (CME) using HZSM-5 catalyst with different Si/
Research and Development Program of China, China
Al ratios. Fuel 2016;180:244–55.
(2018YFC1901300), the National Key Research and Development Pro­ [24] Antonakou EV, Kalogiannis KG, Stefanidis SD, Karakoulia SA, Triantafyllidis KS,
gram of China, China (2016YFE0202000), Key Research and Develop­ Lappas AA, et al. Catalytic and thermal pyrolysis of polycarbonate in a fixed-bed
ment Program of Zhejiang Province, China (2020C03084) and Key reactor: The effect of catalysts on products yields and composition. Polym Degrad
Stab 2014;110:482–91.
Research and Development Program of Shandong Province, China [25] Ates F, Miskolczi N, Borsodi N. Comparision of real waste (MSW and MPW)
(2019JZZY020806), the China Scholarship Council, China, and the ZJU pyrolysis in batch reactor over different catalysts. Part I: product yields, gas and
Scholarship for Outstanding Doctoral Candidates, China. Some of the pyrolysis oil properties. Bioresour Technol 2013;133:443–54.
[26] Mastral JF, Berrueco C, Gea M, Ceamanos J. Catalytic degradation of high density
authors are thankful for the support by the Earth Engineering Center of polyethylene over nanocrystalline HZSM-5 zeolite. Polym Degrad Stab 2006;91:
Columbia University, United States. 3330–8.
[27] Sun K, Huang Q, Meng X, Chi Y, Yan J. Catalytic Pyrolysis of Waste Polyethylene
into Aromatics by H3PO4-Activated Carbon. Energy Fuels 2018;32:9772–81.
Appendix A. Supplementary data [28] Wang H, Male J, Wang Y. Recent Advances in Hydrotreating of Pyrolysis Bio-Oil
and Its Oxygen-Containing Model Compounds. ACS Catal 2013;3:1047–70.
[29] Elordi G, Olazar M, Lopez G, Castaño P, Bilbao J. Role of pore structure in the
The differences between experimental and predicted TG-DTG curves
deactivation of zeolites (HZSM-5, Hβ and HY) by coke in the pyrolysis of
for various blends of PC and PE/PP. (Figure S1). Supplementary data to polyethylene in a conical spouted bed reactor. Appl Catal B 2011;102:224–31.
this article can be found online at https://doi.org/10.1016/j.fuel.2020 [30] Ibáñez M, Artetxe M, Lopez G, Elordi G, Bilbao J, Olazar M, et al. Identification of
.119143. the coke deposited on an HZSM-5 zeolite catalyst during the sequenced
pyrolysis–cracking of HDPE. Appl Catal B 2014;148–149:436–45.
[31] Guisnet M, Costa L, Ribeiro FR. Prevention of zeolite deactivation by coking. J Mol
References Catal A Chem 2009;305:69–83.
[32] To AT, Resasco DE. Role of a phenolic pool in the conversion of m-cresol to
aromatics over HY and HZSM-5 zeolites. Appl Catal A 2014;487:62–71.
[1] Huang J, He C, Li X, Pan G, Tong H. Theoretical studies on thermal degradation
[33] Lu P, Huang Q, Bourtsalas AC, Chi Y, Yan J. Synergistic effects on char and oil
reaction mechanism of model compound of bisphenol A polycarbonate. Waste
produced by the co-pyrolysis of pine wood, polyethylene and polyvinyl chloride.
Manage 2018;71:181–91.
Fuel 2018;230:359–67.
[2] Antonakou EV, Achilias DS. Recent Advances in Polycarbonate Recycling: A
[34] Brebu M, Spiridon I. Co-pyrolysis of LignoBoost® lignin with synthetic polymers.
Review of Degradation Methods and Their Mechanisms. Waste Biomass
Polym Degrad Stab 2012;97:2104–9.
Valorization 2012;4:9–21.
[35] Shen HY, Judeh ZMA, Ching CB, Xia QH. Comparative studies on alkylation of
[3] Wiesweg M. Polycarbonate Market Overview. IHS Market 2018:21.
phenol with tert-butyl alcohol in the presence of liquid or solid acid catalysts in
[4] Narancic T, O’Connor KE. Plastic waste as a global challenge: are biodegradable
ionic liquids. J Mol Catal A Chem 2004;212:301–8.
plastics the answer to the plastic waste problem? Microbiology 2019;165:129–37.
[36] Castaño P, Elordi G, Olazar M, Aguayo AT, Pawelec B, Bilbao J. Insights into the
[5] Achilias DS, Roupakias C, Megalokonomos P, Lappas AA, Antonakou EV. Chemical
coke deposited on HZSM-5, Hβ and HY zeolites during the cracking of
recycling of plastic wastes made from polyethylene (LDPE and HDPE) and
polyethylene. Appl Catal B 2011;104:91–100.
polypropylene (PP). J Hazard Mater 2007;149:536–42.
[37] Cheng YT, Huber GW. Production of targeted aromatics by using Diels-Alder
[6] Zhang Y, Duan D, Lei H, Villota E, Ruan R. Jet fuel production from waste plastics
classes of reactions with furans and olefins over ZSM-5. Green Chem 2012;14:3114.
via catalytic pyrolysis with activated carbons. Appl Energy 2019;251:113337.
[38] Puglisi C, Sturiale L, Montaudo G. Thermal Decomposition Processes in Aromatic
[7] Achilias DS, Antonakou EV, Koutsokosta E, Lappas AA. Chemical recycling of
Polycarbonates Investigated by Mass Spectrometry. Macromolecules 1999;32:
polymers from Waste Electric and Electronic Equipment. J Appl Polym Sci 2009;
2194–203.
114:212–21.
[39] Wang J, Lin BC, Huang QX, Ma ZY, Chi Y, Yan JH. Aromatic Hydrocarbon
[8] Yoosuk B, Tumnantong D, Prasassarakich P. Amorphous unsupported Ni–Mo
Production and Catalyst Regeneration in Pyrolysis of Oily Sludge Using ZSM-5
sulfide prepared by one step hydrothermal method for phenol
Zeolites as Catalysts. Energy Fuels 2017;31:11681–9.
hydrodeoxygenation. Fuel 2012;91:246–52.
[40] Bauer, F. Karge, H.G. Characterization of Coke on Zeolites; 2006: 5, 249-364.
[9] Mercader FdM, Groeneveld MJ, Kersten SRA, Venderbosch RH, Hogendoorn JA.
[41] Cerqueira HS, Caeiro G, Costa L, Ramôa Ribeiro F. Deactivation of FCC catalysts.
Pyrolysis oil upgrading by high pressure thermal treatment. Fuel 2010;89:
J Mol Catal A Chem 2008;292:1–13.
2829–37.
[42] Lin B, Wang J, Huang Q, Ali M, Chi Y. Aromatic recovery from distillate oil of oily
[10] Kay Lup AN, Abnisa F, Daud WMAW, Aroua MK. A review on reaction mechanisms
sludge through catalytic pyrolysis over Zn modified HZSM-5 zeolites. J Anal Appl
of metal-catalyzed deoxygenation process in bio-oil model compounds. Appl Catal
Pyrol 2017;128:291–303.
A 2017;541:87–106.
[43] Sun K, Huang Q, Ali M, Chi Y, Yan J. Producing Aromatic-Enriched Oil from Mixed
[11] Wang W, Tan S, Wu K, Zhu G, Liu Y, Tan L, et al. Hydrodeoxygenation of p-cresol
Plastics Using Activated Biochar as Catalyst. Energy Fuels 2018;32:5471–9.
as a model compound for bio-oil on MoS2: Effects of water and benzothiophene on
[44] Grause G, Kärrbrant R, Kameda T, Yoshioka T. Steam Hydrolysis of Poly(bisphenol
the activity and structure of catalyst. Fuel 2018;214:480–8.
A carbonate) in a Fluidized Bed Reactor. Ind Eng Chem Res 2014;53:4215–23.
[12] Asadieraghi M, Wan Daud WMA, Abbas HF. Model compound approach to design
[45] Ma Z, Troussard E, van Bokhoven JA. Controlling the selectivity to chemicals from
process and select catalysts for in-situ bio-oil upgrading. Renew Sustain Energy Rev
lignin via catalytic fast pyrolysis. Appl Catal A 2012;423–424:130–6.
2014;36:286–303.
[46] Thilakaratne R, Tessonnier JP, Brown RC. Conversion of methoxy and hydroxyl
[13] Mullen CA, Boateng AA. Production of Aromatic Hydrocarbons via Catalytic
functionalities of phenolic monomers over zeolites. Green Chem 2016;18:2231–9.
Pyrolysis of Biomass over Fe-Modified HZSM-5 Zeolites. ACS Sustainable Chem Eng
[47] Wang K. Pyrolysis and catalytic pyrolysis of protein- and lipid-rich feedstock, in
2015;3:1623–31.
Mechanical Engineering. Lowa State University; 2014. p. 160.
[14] Rezaei PS, Shafaghat H, Daud WMAW. Production of green aromatics and olefins
[48] Gayubo AG, Aguayo AT, Atutxa A, Aguado R, Bilbao J. Transformation of
by catalytic cracking of oxygenate compounds derived from biomass pyrolysis: A
Oxygenate Components of Biomass Pyrolysis Oil on a HZSM-5 Zeolite. I. Alcohols
review. Appl Catal A 2014;469:490–511.
and Phenols. Ind Eng Chem Res 2004;43:2610–8.
[15] Yeletsky PM, Kukushkin RG, Yakovlev VA, Chen BH. Recent advances in one-stage
[49] Jia X, An W, Wang Z, Zhou J. Effect of Doped Metals on Hydrodeoxygenation of
conversion of lipid-based biomass-derived oils into fuel components – aromatics
Phenol over Pt-Based Bimetallic Alloys: Caryl–OH Versus CaliphaticH–OH Bond
and isomerized alkanes. Fuel 2020;278:118255.
Scission. J Phys Chem C 2019;123:16873–82.
[16] Tagliabue M, Carati A, Flego C, Millini R, Perego C, Pollesel P, et al. Study on the
stability of a Ga/Nd/ZSM-5 aromatisation catalyst. Appl Catal A 2004;265:23–33.

12
K. Sun et al. Fuel 285 (2021) 119143

[50] Kaminsky W, Schlesselmann B, Simon C. Olefins from polyolefins and mixed [52] Al-Salem SM, Antelava A, Constantinou A, Manos G, Dutta A. A review on thermal
plastics by pyrolysis. J Anal Appl Pyrol 1995;32:19–27. and catalytic pyrolysis of plastic solid waste (PSW). J Environ Manage 2017;197:
[51] Anuar Sharuddin SD, Abnisa F, Wan Daud WMA, Aroua MK. A review on pyrolysis 177–98.
of plastic wastes. Energy Convers Manage 2016;115:308–26.

13