Upgrading of chlorinated oils coming from pyrolysis of plastic waste

Fuel Processing Technology 137 (2015) 229–239

Contents lists available at ScienceDirect

Fuel Processing Technology journal homepage: www.elsevier.com/locate/fuproc

Upgrading of chlorinated oils coming from pyrolysis of plastic waste A. Lopez-Urionabarrenechea ⁎, I. de Marco 1, B.M. Caballero 1, M.F. Laresgoiti 1, A. Adrados 1 Chemical and Environmental Engineering Department, Faculty of Engineering of Bilbao, University of the Basque Country (UPV/EHU), Alda, Urquijo s/n, 48013 Bilbao, Spain

a r t i c l e

i n f o

Article history: Received 4 November 2014 Received in revised form 10 April 2015 Accepted 12 April 2015 Keywords: Pyrolysis oils Upgrading Thermal cracking Catalytic cracking Dechlorination Plastic waste

a b s t r a c t The objective of this paper is the upgrading of chlorinated oils coming from the pyrolysis of mixed plastic waste, in order to use them as fuel or feedstock for refineries. Two different samples of pyrolysis oils have been thermally and catalytically cracked in a 300 mL autoclave at 325 °C and the auto-generated pressure. Thermal cracking converts the plastic pyrolysis heavy oils into light liquid fractions which are only composed of alkanes and aromatics. These light fractions present a very low quantity of chlorine compared to the initial oils and resemble gasoline and diesel-like products. Besides, a gaseous fraction rich in methane and with very high heating value is also produced, together with a fuel-like viscous product which remains in the autoclave. The relative proportions of each of these three fractions depend on the nature of the initial oils. Red Mud has proved to be a dehydrochlorination and cracking catalyst, since it gives rise to higher quantity of gases and light liquid fractions with a very low chlorine content (b0.1 wt.%). Therefore, dechlorinated light oils can be obtained by Red Mud low temperature catalytic cracking of plastic derived chlorinated pyrolysis oils. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Pyrolysis of plastic waste has been intensively studied over the last few decades with the objective of producing useful liquid and gaseous fuels or chemicals from low value polymeric waste [1–6]. Scientific literature claims that pyrolysis could be used to valorise the plastic fraction of municipal and industrial solid waste; however, many of the industrial or demonstration processes implemented in the last years no longer run nowadays [7,8]. The difficulty to find market applications for the pyrolysis derived liquid product seems to be the handicap stopping the current development of the plastic waste pyrolysis process, since the process itself is by now a well-known technology for a while and there are normally no technical problems in its implementation. The liquid fraction obtained by pyrolysis of plastic waste presents good properties for application as fuel or chemical feedstock for refineries: they are hydrocarbon oils with high calorific value and high carbon content, and free of water provided that the quantity of impurities as paper or wood is not very high in the original sample (b8 wt.%) [9,10]. On the other hand, the typical drawbacks of this liquid fraction are the broad carbon-range distribution (C5–C25 or higher) and the presence of different functional groups and families of organic compounds (mainly aromatics, olefins and alkanes) [11,12]. The balance between advantages and drawbacks of the pyrolysis oils depends on the initial waste sample composition and on the process characteristics. In fact, some authors claim that when pure polyethylene (PE) and polypropylene ⁎ Corresponding author. Tel.: +34 946018245; fax: +34 946014179. E-mail address: [email protected] (A. Lopez-Urionabarrenechea). 1 Tel.: +34 946018245; fax: +34 946014179.

http://dx.doi.org/10.1016/j.fuproc.2015.04.015 0378-3820/© 2015 Elsevier B.V. All rights reserved.

(PP) are pyrolysed under specific conditions, the obtained liquids can be directly used as feedstock for fluid catalytic cracking (FCC) units of petroleum refineries, as an alternative to vacuum gas oil [13,14]. Unfortunately, it is very difficult to obtain a liquid fraction with a direct industrial application when it comes to complex plastic mixtures such as municipal plastic waste, since such mixtures contain many materials intermingled with PE and PP, among others, polystyrene (PS), poly(ethylene terephthalate) (PET) and halogenated polymers like poly(vinyl chloride) (PVC) [9,10,15]. In such cases, the liquid fraction presents additional drawbacks for direct application, hindering the implementation of a large scale pyrolysis process for municipal plastic waste. Some of these drawbacks are mainly the presence of chlorine [16] and sometimes the waxy (semi-solid) nature of this product, which complicates its handling and processing [17]. Previous works carried out by the authors indicated that these undesirable properties can be partially overcome in the pyrolysis process itself, e.g., by using catalysts to shorten the carbon-range and/ or to avoid the formation of waxy liquids [18], or by stepwise pyrolysis (two step pyrolysis) to reduce the chlorine content [16]. These works proved that a two-step process consisting of a first step at around 300 °C and a second one around 500 °C is the best option in order to obtain pyrolysis oils with very low content of chlorine, but the obtained liquids remain their waxy appearance [16]. On the contrary, every time a single step process has been carried out in presence of a catalyst or a chlorine adsorbent (Red Mud, ZSM-5 zeolite, CaCO3, mixed with the sample), fluid pyrolysis liquids have been obtained, but the chlorine content of the oils were much higher than those obtained in the two step process, and even higher than the oils obtained in a single step pyrolysis without catalyst [18]. Additionally, a catalytic two-step process was also proved by the authors and fluid pyrolysis

230

A. Lopez-Urionabarrenechea et al. / Fuel Processing Technology 137 (2015) 229–239

oils with lower chlorine content than those obtained in the catalytic single process were produced, but their chlorine content was still a little bit higher than in the case of the two-step process without catalyst [19]. Therefore, “offline upgrading” arises as an unavoidable step in order to obtain short carbon chain dechlorinated liquids. There is here a challenge for scientific community since the final step for the utilization of plastic derived pyrolysis oils seems to be very close, but surprisingly, quite little work is found in literature concerning this topic. This fact becomes even more surprising when compared to the work carried out in the upgrading of biomass pyrolysis liquids (biooils), taking into account that, at first sight, the technical drawbacks concerning the potential application of bio-oil as a liquid fuel or feedstock for refineries are quite more difficult to overcome than the aforementioned for the plastic pyrolysis liquids, since bio-oils are instable, corrosive and high oxygen containing organic compounds, always mixed with water to a greater or lesser extent [20,21]. The majority of published papers concerning the upgrading of plastic pyrolysis liquids has been carried out with PE/PP pyrolysis derived liquids [22–27]. In such works, distillation [22,23], thermal and catalytic cracking [24–26] and the combination of both [27] have been studied. The main conclusions which can be drawn from these papers are that (1) cracking is an adequate solution to upgrade pyrolysis oils, mainly by shortening the carbon-range and converting the waxy products into lighter ones through chemical reaction and (2) distillation is an effective way to separate the pre-cracked oil, but not to obtain useful fractions if it is the unique operation used. Cracking is also the selected process not only for the upgrading of PE/ PP derived liquids, but also for pyrolysis oils coming from chlorine containing complex mixtures of plastics. With this type of sample, and as far as the authors know, only some research works carried out by Lingaiah et al. [28,29] and Miskolczi et al. [30] have been published. Miskolczi et al. [30] proposed a two stage process consisting of a first thermocatalytic cracking using acid catalysts and a second stage with nickel and cobalt-based catalysts. Quite large molecules (larger than C20) still remained in the final liquid product, but they achieved an important reduction of chlorine content. On the other hand, the work of Lingaiah et al. [28,29] was focused on the catalytic activity of iron-based catalysts in order to dehydrochlorinate the pyrolysis oils. They showed the potential of iron oxide to dechlorinate pyrolysis oils but they did not observed any considerable change in the carbon number distribution. In conclusion, it seems that an integral upgrading of chlorine containing pyrolysis oils (chlorine removing and carbon-range shortening) has not been achieved up to now. For this reason, the objective of the present work was to upgrade chlorinated pyrolysis oils in terms of both dechlorination and chemical transformation, shortening the carbon number distribution and changing the proportions of functional groups. With this objective, a complete upgrading study of chlorinated plastic pyrolysis oils was carried out comparing thermal and catalytic cracking. This research work attempts to provide a step forward towards the utilization of plastic pyrolysis oils as valuable fuels and/or chemical feedstock for refineries. 2. Material and methods

sample carried out with ZSM-5 zeolite at several different temperatures in the range 425–460 °C, atmospheric pressure, and dwell times of 30 min. These samples are called in this paper pyrolysis oil 1 (PO1) and pyrolysis oil 2 (PO2) respectively. Pyrolysis oils obtained at different temperatures (PO1 at 500 °C and PO2 at 425–460 °C) were chosen because PO1 was obtained in thermal pyrolysis experiments and PO2 in catalytic pyrolysis experiments. In the installation used for the production of such oils, the optimum temperature (the temperature that produces the highest quantity of pyrolysis oils with the best properties) of the thermal pyrolysis is 500 °C. On the contrary, when a cracking catalyst is used, 500 °C is a too high temperature and 425–475 °C are more adequate ones. Therefore, the objective of such selection was to work with pyrolysis oil samples obtained at the best experimental conditions of the installation used for their production. The characteristics of both samples are shown and discussed in Section 3. Waste Red Mud was used as catalyst for the experiments. Red Mud is a by-product of the alumina production process (Bayer process) composed mainly of Fe2O3. It has been proved by the authors in previous pyrolysis experiments that this product has a noticeable catalytic effect [17,18]; therefore, it is worthwhile further investigating its behavior as catalyst, given that it is an abundant and inexpensive industrial waste product. The composition and main characteristics of Red Mud are summarized in Table 1. More detailed information on its properties and characterization has been published elsewhere [18]. 2.2. Upgrading pilot plant The upgrading experiments were carried out in a lab-scale installation equipped with a stirred autoclave and a condensation system cooled by a coolant circulating liquid. Fig. 1 shows the flow-sheet of the upgrading pilot plant. In each experiment, 100 g of sample were placed in the autoclave, which was then closed, heated up to 325 °C at a rate of 5 °C min−1, and maintained at such temperature for a 30 min dwell time. The installation was used in batch operation (closed autoclave) in order to maximize chemical reactions, thus the experiments were carried out at the auto-generated pressure (60–100 bar). Immediately after the dwell time finished, the opening valve of the autoclave was opened in order to let the formed volatile products flow to the condensation system, where the condensable vapors were condensed into light oil; the non-condensable gases were collected in plastic bags. The yields of light oil and of the heavy fraction remaining in the autoclave were calculated as weight percentage with respect to the amount of sample introduced at the beginning of the experiment. Gas yields were calculated by difference. In the catalytic runs, 10 g of Red Mud were mixed with the pyrolysis oils at the onset of the experiment; the catalyst remained in the autoclave after the experiment mixed with the heavy fraction. 2.3. Analysis of the products The composition of the initial samples (pyrolysis oils) as well as the liquid fractions obtained after the cracking experiments (light fraction and heavy fraction) was determined by gas-chromatography coupled

2.1. Pyrolysis oil samples and catalyst The pyrolysis oils used for the experiments were two samples obtained by the authors in previous experiments of pyrolysis of municipal plastic waste, carried out in an installation consisting on a 3.5 L semibatch reactor and a condensing system equipped with three condensers and an activated carbon column, and purged with a 1 L min−1 N2 flow [6,18]. More specifically, two pyrolysis oils samples have been used; (1) pyrolysis oils coming from several thermal pyrolysis of a mixture of 40 wt.% PE/35 wt.% PP/18 wt.% PS/4 wt.% PET/3 wt.% PVC carried out at 500 °C, atmospheric pressure, and dwell times from 0 to 120 min, and (2) pyrolysis oils coming from catalytic pyrolysis runs of the same

Table 1 Main characteristics of red mud. Composition (wt.%)

BET surface area (m2 g−1) Acidity (mmol NH3 g−1) a

By difference.

Fe2O3 Al2O3 TiO2 SiO2 CaO Na2O Othersa

36.5 23.8 13.5 8.5 5.3 1.8 10.6 27.5 0.09

A. Lopez-Urionabarrenechea et al. / Fuel Processing Technology 137 (2015) 229–239

231

Fig. 1. Flow-sheet of the lab-scale installation used for the experiments.

with a mass spectrometer detector (GC-MS). The characteristics of the equipment and the method have been previously published [19]. Identification of the compounds was carried out by means of: comparison of the retention times with those of calibration samples, computer matching against the commercial library of mass spectra (Wiley7n) and also comparing with mass spectra literature data. When the match quality of the identification result provided by the MS search engine was lower than 85% of spectrum coincidence, the result was not considered valid and these compounds have been classified in this paper as “Not identified”. Besides, it must be mentioned that the GCMS composition data are quantified as % area and not wt.%, which are not exactly the same since different kinds of organic compounds have different responses in MS. The composition of the non-condensable gases was obtained by gaschromatography coupled with thermal conductivity and flame ionization detectors (GC-TCD/FID). The characteristics of the equipment and the method have been also previously described [19]. The quantification of the gas components was carried out by comparing the peak area with those obtained in the analysis of calibration samples. The high heating value (HHV) of the liquid and solid samples was determined by an automatic calorimeter, while the HHV of the gases was theoretically calculated according to their composition and to the HHV of the individual components. The elemental composition (C, H, N) of the liquids was determined using an automatic analyzer and chlorine quantification was carried out by Method 5050 of the Environmental Protection Agency (EPA) of the United States. 3. Results and discussion 3.1. Pyrolysis oils Table 2 shows the detailed composition of the pyrolysis oils used for the upgrading experiments. Both pyrolysis oil samples are complex mixtures, containing about 30 different chemical species. Styrene is by

far the main component in both samples, with 42.9% area in PO1 and 27.9% area in PO2. Toluene (10.1 and 13.1% area respectively) and ethyl-benzene (8.2 and 10.5% area respectively) are also present in relatively high quantities. In view of this composition, one alternative to valorise this pyrolysis oils could be the separation of these valuable aromatic chemicals by distillation. However, these compounds are mixed with other close-boiling point hydrocarbons, some of which form azeotropes with the above mentioned aromatics, making it impossible to obtain separately such aromatics from this kind of sample by classical distillation, which has been experimentally proved by the authors. In order to make such separation extractive distillation with selective solvents must be used in [31], which makes the process quite complicated. The composition of the pyrolysis oils is summarized in Table 3. Concerning the carbon number range, C5–C10 is the main fraction of both samples (82.2% area PO1 and 84.5% area PO2) while the proportion of heavy compounds, identified as N C16, is very low (0.7% area) or inexistent in both of them. Although the two samples seem to be almost identical, there is an important difference between them, which lies on their chemical nature. Chemically, PO1 is a mixture of aromatic (79.9% area) and aliphatic (15.4% area) compounds, while PO2 is almost totally composed of aromatics (97.7% area), which is due to the catalytic effect of the ZSM-5 zeolite used in the pyrolysis experiments carried out to obtain this sample. The high aromatic content is not a critical problem as far as the potential use of pyrolysis oils as chemical feedstock for refineries is concerned, since blending operations of highly aromatic fractions are normally carried out in such installations. However, the presence of unsaturated aliphatic compounds (olefins) is generally undesired in finished products, given their high reactivity (due to the double bound) and tendency to oxidize and to polymerize into rubbery substances [32]. Additionally, the presence of long chain aliphatic products seems to be the responsible of giving the waxy consistency to pyrolysis liquids, since PO1 is a wax-like product which remains in solid phase at room temperature while PO2 is a medium-viscosity liquid.

232

A. Lopez-Urionabarrenechea et al. / Fuel Processing Technology 137 (2015) 229–239

Table 2 Detailed composition of the pyrolysis oils determined by GC–MS (% area).

Table 4 Elemental analysis (wt.%) and HHV (MJ kg−1) of the pyrolysis oils.

Compound name

Pyrolysis oil 1 (PO1)a

Pyrolysis oil 2 (PO2)b

Benzene Toluene 2,4-Dimethyl-heptene Ethyl-benzene p-Xylene 1-Nonene Styrene 1-Methyl-ethyl-benzene Propyl-benzene 1-Ethyl-2-methyl-benzene α-Methyl-styrene 1-Decene 1,2,4-Trimethyl-benzene 1-Propenyl-benzene Indene 2-Methyl-indene 1-Undecene 1-Dodecene Naphthalene 1-Tridecene 2-Methyl-naphthalene 1-Tetradecene 1-Pentadecene 1-Hexadecene 3-Phenyl-propyl-benzene 1-Octadecene 2-Phenyl-naphthalene Other identified compounds Unidentified compounds

1.7 10.1 5.1 8.2 2.6 0.8 42.9 0.7 n.d.c n.d.c 4.4 1.6 n.d.c 0.8 1.0 n.d.c 1.6 1.1 1.6 1.2 1.0 0.9 0.9 0.6 1.2 0.7 1.5 2.3 5.7

4.5 13.1 1.6 10.5 9.8 n.d.c 27.9 1.0 1.1 3.0 3.9 n.d.c 2.0 1.0 1.8 2.9 n.d.c n.d.c 3.2 n.d.c 4.8 n.d.c n.d.c n.d.c 1.0 n.d.c 1.6 4.6 0.7

a b c

C H N Cl Othersa H/C ratio HHV (MJ kg−1) a

85.4 10.2 0.3 0.9 3.2 1.4 41.4

By difference.

3.2. Upgrading experiments yields

Pyrolysis oils obtained in thermal pyrolysis experiments at 500 °C. Pyrolysis oils obtained in catalytic pyrolysis experiments at 425–460 °C. Not detected.

Table 3 Composition of the pyrolysis oils (initial samples) determined by GC–MS including only identified compounds (% area). Fraction/sample

PO1

PO2

C5–C10 C11–C16 NC16 Total aromatic compounds Single-ring aromatic compounds Indenes Naphthalenes Other fused polyaromatic compoundsb Non fused polyaromatic compoundsc Total aliphatic compounds Unsaturated aliphatic compounds Saturated aliphatic compounds

82.2 11.4 0.7 79.9 73.0 1.0 4.0 n.d.a 1.9 15.4 14.5 0.9

84.5 14.8 n.d.a 97.7 77.9 6.3 11.1 n.d.a 2.4 1.6 1.6 n.d.a

Not detected. Polyaromatic compounds whose aromatic rings share carbon atoms, e.g., phenanthrene. c Polyaromatic compounds whose aromatic rings do not share any carbon atom, e.g., phenyl-propyl-benzene. b

PO2

85.9 11.2 0.1 0.8 2.0 1.6 42.4

since they are at most hydrocarbon samples with few heteroatoms. Nevertheless, it must be pointed out that the chlorine content of both samples is quite high (0.8 wt.% and 0.9 wt.%), being the major disadvantage for the utilization of such pyrolysis oils. The H/C ratio of PO2 is lower than that of PO1, which is a consequence of the higher content of aromatics of PO2. HHV is very high for both samples, which clearly indicates their potential to be used as fuels provided that the rest of their properties allow it.

It is noteworthy that these pyrolysis oils, which are obtained from the pyrolysis of PE/PP rich samples, are mainly composed of aromatics and do not show paraffins in their composition. The fact of the matter is that the formation of aromatics strongly depends on the reactor design and the operating conditions used, and the generation of aromatic compounds even in the pyrolysis of pure polyolefins has been reported in the literature (e.g., [33–35]). The authors have also obtained oils with high contents of aromatics in the last years with the installation used for the pyrolysis of plastic mixtures (e.g., [6,17,18]), the same as the used for the production of the PO1 and PO2 samples. The installation used for the pyrolysis of plastic waste is not the same as the used for the experiments of the present paper, which is described in Section 2.2. Table 4 shows the elemental composition and the HHV of the pyrolysis oils. Both samples are mainly composed of carbon and hydrogen,

a

PO1

The upgrading experiments yields can be seen in Fig. 2; the two bar diagrams on the left correspond to the thermal cracking experiments and the two on the right to the catalytic cracking ones. As it can be seen, three fractions are obtained after the upgrading experiments. The light fraction, which consists of the volatile products generated during the experiment and subsequently condensed into liquid; the heavy fraction, which is the fraction remaining in the autoclave after the experiment; and the gases, the most light products which do not condense in the condensation system. Although the initial samples are totally converted in the upgrading experiments, the sum of light fraction and gases yields has been considered the conversion of the process, in order to give an idea of the conversion of the initial samples to light products. For this reason, this sum has been named “conversion” in the figure. Fig. 2 shows that the light fraction is the main product in the thermal experiments, reaching 71.2 wt.% with PO1 and 58.8 wt.% with PO2. These are quite good results, taking into account that the purpose of the upgrading process is to obtain a light and dechlorinated liquid fraction usable in petrochemical processes. With respect to the gas and heavy fractions yields, there are significant differences between PO1 and PO2. PO1 yields 15.8 wt.% of gases and 13.0 wt.% of heavy fraction while PO2 yields 8.7 wt.% of gases and 32.5 wt.% of heavy fraction. This fact indicates a clear different behavior between both samples, probably linked to their different chemical nature. PO1 contains 14.5% area of olefins, conferring this sample more reactivity than PO2, which is almost exclusively composed of aromatics, which in general terms are more stable compounds than olefins. In the case of PO1, this gives as a result a high conversion into light products (87.0 wt.%). On the contrary, PO2 yields a high quantity of heavy fraction, which may be because the temperature used is not high enough to crack completely the compounds present in this sample, and/or because such compounds have a great tendency to polymerize or condensate, yielding heavy liquid products. The yields of both samples are quite different when catalyst is used. PO1 yields more gases (46.8 wt.%) than light fraction (45.5 wt.%) and at the same time the quantity of heavy fraction (7.7 wt.%) is lower than that obtained without catalyst (13.0 wt.%). The high proportion of gases produced in this experiment indicates that Red Mud clearly favors cracking reactions, shortening the carbon chains of molecules and consequently generating higher quantities of compounds which remain in gas phase after the condensation process. The cracking performance of Red Mud may be attributed to the zeolitic nature that Al2O3 and SiO2 (in form of alumina-silicate) confer to this catalyst.

A. Lopez-Urionabarrenechea et al. / Fuel Processing Technology 137 (2015) 229–239 Light fraction

Heavy fraction

Gases

233

Conversion

100 92.3 90 87.0

86.5

80 71.2

67.5

70

66.1 58.8

wt %

60

50

46.8

45.5

40 32.5 30 20.4 20

15.8 13.5

13.0 8.7

10

7.7

0 PO1 Thermal run

PO2 Thermal run

PO1 Catalytic run

PO2 Catalytic run

Fig. 2. Yields and conversion to light products (wt.%) obtained in the four upgrading experiments.

There is some discussion within scientific community at this point. On the one hand, some publications do not attribute any catalytic effect to Red Mud in terms of cracking enhancement [28,29]. On the other hand, the authors [17,18], as well as other researchers [36,37], have demonstrated that Red Mud possesses catalytic effect. The lack of agreement concerning the catalytic effect of Red Mud can be attributed to the variation of the composition of this product depending on its origin, due to the fact that it is a by-product coming from an industrial process which uses mineral ore as raw material. For instance, Na2O is known to act as a catalytic poison in the decomposition of some hydrocarbons and it has been reported that its proportion in Red Mud can be up to 6 wt.% [38]. On the contrary, Table 1 shows that the Na2O content of the Red Mud used for these experiments is much lower (1.8 wt.%). Looking at the results obtained in the catalytic run of PO2 and comparing them to the results of the thermal run, it can be seen that the catalytic effect of Red Mud is similar to that observed with PO1: the cracking ratio of PO2 in presence of Red Mud is higher than in the case of the thermal run (from 67.5 to 86.5 wt.%), which means that the yield of the heavy fraction decreases (13.5 wt.%). Therefore cracking has been enhanced by Red Mud. At the same time, the light fraction and gases yields significantly increase (66.1 and 20.4 wt.% respectively). At first sight, the increase of the liquid fraction in the catalytic experiment of PO2 (compared to the thermal one) could seem to be contrary to the behavior of PO1 sample, which showed a decrease in the liquid fraction under catalytic conditions. However, this is also a consequence of the higher cracking obtained with Red Mud and it can explained at follows: when PO2 is cracked in presence of Red Mud, the cracked molecules seem not to be small enough to be part of the gas fraction, and most of them condense to be part of the liquid fraction. The overall result, compared to the thermal experiment, is that the liquid fraction increases, but this also means more cracking, as in the case of the PO1. This is probably due to the different chemical nature of the oils, since PO2 is an almost completely aromatic sample and, as it has been mentioned before, the aromatic chemicals have in general higher thermal stability than the aliphatics.

If the results obtained with both samples (in presence of Red Mud) are compared, the same tendency as that formerly found in the thermal runs can be observed; the conversion of PO2 (86.5 wt.%) is lower than that of PO1 (92.3 wt.%). However, it is worth mentioning that the cracking ratio of PO1 increases in around 5 points (from 87.0 to 92.3 wt.%) in presence of Red Mud, while the cracking ratio of PO2 is 19 points higher (from 67.5 to 86.5 wt.%) in this case, both of them compared to the thermal runs. This fact could indicate that 325 °C is not a high enough temperature to completely crack the PO2 sample by thermal cracking (PO2 is thermally more stable than PO1), and a cracking catalyst is needed to increase the conversion to cracked products. Therefore, the cracking effect and activity of Red Mud is proved, since the conversion ratio of PO2 increases significantly when it is used. 3.3. Light fraction characteristics This section is only devoted to the description of the light fractions obtained in the four upgrading experiments. The discussion about the possible mechanisms and pathways in the chemical transformations occurring during the experiments is not included here. This discussion is presented in Section 3.6, by means of a comparison between the initial oils and the total composition of the liquid products. Table 5 shows the summarized composition of the light fractions obtained in the cracking experiments attending to, on the one hand, the carbon-range distribution and on the other hand, the functional groups. In both cases, only the identified compounds have been included. The detailed composition of these oils can be seen in Table S1 of the supplementary data. First, if the summarized composition of the light fraction obtained from PO1 is compared to the summarized composition of the original oil sample (PO1) shown in Table 3, it can be seen that the light fraction obtained in the thermal run is lighter than the original oil sample, since greater amounts of C5–C10 compounds (86.6 compared to 82.2% area) and lower yields of C11–C16 products (from 11.4 to 9.1% area) and of NC16 fraction (from 0.7% area to disappear) are obtained. This is the

234

A. Lopez-Urionabarrenechea et al. / Fuel Processing Technology 137 (2015) 229–239

Table 5 Composition of the light fractions obtained in the upgrading experiments of pyrolysis oils (% area).

C5–C10 C11–C16 NC16 Total aromatic compounds Single-ring aromatic compounds Indenes Naphthalenes Other fused polyaromatic compounds Non fused polyaromatic compounds Total aliphatic compounds Unsaturated aliphatic compounds Saturated aliphatic compounds Unidentified a

Thermal run

Catalytic run

PO1

PO2

PO1

PO2

86.6 9.1 n.d.a 73.8 67.4 n.d.a 6.4 n.d.a n.d.a 21.9 n.d.a 21.9 4.3

87.7 7.1 n.d.a 84.1 74.3 n.d.a 9.8 n.d.a n.d.a 12.0 n.d.a 12.0 3.9

78.7 17.6 n.d.a 78.8 64.8 n.d.a 10.4 n.d.a 3.6 17.5 0.6 16.9 3.7

92.1 6.0 n.d.a 89.8 75.2 n.d.a 12.7 1.9 n.d.a 8.3 n.d.a 8.3 1.9

Not detected.

result of cracking reactions which shorten the carbon chains. With respect to the distribution of compounds by functional groups, Table 5 shows that the light fraction of the thermal experiment of PO1 contains less aromatic compounds than the original oil and no unsaturated aliphatic compounds (i.e. olefins), which were in a significant proportion in the original oil (14.5% area). This is a very attractive characteristic for the application of these upgraded light fractions as petrochemical feedstock. In its stead, a significant quantity of saturated aliphatic compounds (i.e. alkanes) is found (21.9% area, when they were only 0.9% area in the original sample). When the catalyst is used, the proportion of C5–C10 compounds of the PO1 light fraction (78.7% area) is lower than that of the C5–C10 fraction of the original oil (82.2% area) and also than that of the C5–C10 fraction of the light fraction obtained from the thermal run (86.6% area). This result is quite surprising since stronger cracking than that produced in the thermal experiment might have been expected in this experiment. The explanation to this fact lies on the high generation of gases in this experiment (46.8 wt.% compared to 15.8 wt.% in the thermal run). These

gaseous compounds are in the C1–C5 range, therefore the total conversion to light products has been greater in the catalytic experiment than in the thermal run, as it has been explain in Subsection 3.2. The chemical nature of this light fraction reveals the presence of significant quantities of naphthalene-derived compounds (10.4% area) and non fused polyaromatic compounds (3.6% area). These chemicals are in the range C11–C16, explaining the increase of such fraction in this liquid. This result suggests that Red Mud promotes the formation of double-ring aromatic chemicals, which are not found in the light fraction of the thermal experiment. The promotion of double-ring aromatic compounds is in accordance with the higher content of aromatics of this light fraction (78.8% area) compared to the light fraction of the thermal run (73.8% area). The aromatization effect of Red Mud has been also reported by the authors in previous works and is attributed to the acidic nature that Al 2O 3 and SiO2 confer to this catalyst [17,18]. As in the case of the light fraction obtained in the thermal run, almost all the unsaturated aliphatic compounds present in the PO1 original oil reacted to give saturated aliphatic compounds mainly, so improving the characteristics of the oil. Concerning the light fractions obtained from PO2, the thermal run also produced a light fraction which was clearly lighter than the original oil; the C5–C10 fraction is 3 points higher in percentage (from 84.5 to 87.7% area) while the C11–C16 fraction is 7 points lower (from 14.8 to 7.1% area). Looking at the functional groups distribution, the PO2 light fraction is more aromatic (84.1% area) than that obtained in the thermal run of PO1 (73.8% area); this is directly related to the aromatic content of both original oil samples, which is much higher in PO2 (97.7 vs 79.9% area). The aromatic content of PO2 thermal light fraction is significantly lower than that of the original PO2, and on the other hand 10.7% area of saturated aliphatic compounds has been obtained even though the olefin content of PO2 is quite low (1.6% area). The catalytic run of PO2 generated a light fraction composed of 92.1% area C5–C10 fraction, which is a very interesting result and it supposes a shift of 8 points in percentage with respect to the PO2 original oil. In fact, the detailed composition of this fraction showed that only three compounds in the range C11–C16 were detected

Fig. 3. Light fractions obtained in the pyrolysis oils upgrading experiments.

A. Lopez-Urionabarrenechea et al. / Fuel Processing Technology 137 (2015) 229–239

(one C11 -methylnaphthalene- and two C16 -ethenylanthracene and phenylnaphthalene-). The results demonstrate that Red Mud has a clear potential to enhance cracking reactions and thus to shorten carbon chains. Theamountofaromatics(89.8%area)ishigherthaninthe case of the thermal run (84.1% area), as it was observed for PO1 sample, which indicates that Red Mud also promotes aromatization reactions. In all cases the four light samples obtained were yellow-brown clear products which resembled gasoline and diesel-like fractions. The physical appearance of these light samples can be observed in Fig. 3. The elemental composition and HHV of the light fractions can be seen in Table 6. There are not important differences among the four light fractions either in the elemental composition or in the HHV's. All of them are mainly composed of carbon (86.3–88.3 wt.%) and hydrogen (10.2–11.4 wt.%). It must be highlighted the low quantity of chlorine of these fractions (0.1–0.2 wt.%) compared to the chlorine content of the original oils (0.8–0.9 wt.%), which corroborates the convenience of this upgrading treatment in terms of pyrolysis oil utilization. The rate of dechlorination itself is quite high in the thermal experiments (from 0.8–0.9 to 0.2 wt.%). This fact can be explained as follows: chlorine usually remains in plastic pyrolysis oils in form of chlorinated organic compounds which come from the reaction between the HCl formed during PVC decomposition and the organic compounds evolved in the decomposition of other polymers, mainly PP and PS [39,40]; consequently, the halogenated chemicals usually found in plastic pyrolysis oils are chlorinated branched aliphatic and chlorinated aromatic species, mainly in the C6–C11 range [28,29]. When plastic pyrolysis oils are heated up, chlorine is again released as gaseous HCl at moderate temperatures due to the lower C\\Cl bond energy compared to C-C and C-H [41], and therefore the condensable organic compounds (light fractions) turn out to be dechlorinated after the experiment. Comparing the chlorine content of the light fractions obtained in the thermal and catalytic runs, the dechlorination ratio is even higher in the catalytic experiments, even though the chlorine content of the thermal runs fractions is very low, which hardens achieving further dechlorination. The explanation lies on the fact that Fe2O3 is the main constituent of Red Mud, and it is known that metal oxides like Fe2O3 act as dechlorination catalysts by attracting chlorine and weakening the C\\Cl bonds [29,42]. Taking into account that the chlorine content is a key factor for utilization of pyrolysis oils, this result highlights the interest of using Red Mud in pyrolysis oils upgrading processes. 3.4. Heavy fraction characteristics As Section 3.3., the present section is also dedicated to the description of the heavy fractions obtained in the four upgrading experiments. The discussion about the possible mechanisms and pathways in the chemical transformations occurring during the experiments is presented in Section 3.6. The characteristics of the heavy fractions are presented in Tables 7 and 8. Two types of heavy fractions were obtained depending on the experimental conditions. In the thermal runs, the heavy fractions were Table 6 Elemental composition (wt.%) and HHV (MJ kg−1) of the light fractions obtained in the upgrading experiments of pyrolysis oils. Thermal run

C H N Cl Othersa H/C ratio HHV (MJ kg−1) a

By difference.

Catalytic run

PO1

PO2

PO1

PO2

87.1 11.4 b0.1 0.2 1.2 1.6 40.7

86.3 10.6 b0.1 0.2 2.8 1.5 39.5

87.2 11.1 0.3 b0.1 1.3 1.5 42.2

88.3 10.2 0.2 0.1 1.2 1.4 39.0

235

Table 7 Composition of the heavy fractions obtained in the thermal upgrading experiments of pyrolysis oils (% area). Thermal run

C5–C10 C11–C16 NC16 Total aromatic compounds Single-ring aromatic compounds Indenes Naphthalenes Other fused polyaromatic compounds Non fused polyaromatic compounds Total aliphatic compounds Unsaturated aliphatic compounds Saturated aliphatic compounds Unidentified a

PO1

PO2

65.2 28.8 4.9 97.9 64.8 n.d.a 23.7 6.3 3.1 1.0 n.d.a 1.0 1.1

55.1 30.3 13.0 82.4 46.0 n.d.a 27.5 5.0 3.9 16.0 n.d.a 16.0 1.6

Not detected.

very viscous dark liquid products which resembled heavy petroleum fractions, while the heavy fractions obtained in the catalytic runs were carbonaceous solids, which remained mixed with the catalyst. This is the reason why only the composition of the heavy fractions from the thermal run is presented in Table 7. Table 7 shows that although these fractions are named “heavy fractions” in this study, C5–C10 is still the main fraction in both samples, ranging from 65.2% area for PO1 to 55.1% area for PO2. A significant quantity of C11–C16 compounds (28.8 and 30.3% area respectively) and small quantities of N C16 compounds (4.9 and 13.0% area respectively) are also present in both samples. Comparing with the light fractions presented in Table 5, these fractions are obviously heavier and contain heavy compounds coming from recombination reactions among the cracked products that take place during the experiment. The physical appearance of the heavy fractions can be observed in Fig. 4. Attending to the functional groups of the heavy fractions, again no olefins are found. The heavy fraction from PO1 is almost completely aromatic (97.9% area) and the heavy fraction from PO2 is mainly composed of aromatics (82.4% area), but also contains a significant proportion of alkanes (16.0% area). The elemental composition and HHV of the heavy fractions obtained in all the experiments is presented in Table 8. The elemental composition and HHV of the heavy fractions obtained in the thermal experiments are very similar. It has to be pointed out again the very low proportion of chlorine in both of them (0.1 and 0.2 wt.%), which may be explained by the dehydrochlorination mechanism presented in Subsection 3.3. The heavy fractions from thermal runs are very different to the fractions obtained in the catalytic runs, which is due to the presence of Red Mud mixed with the later heavy fractions, in this case, solid carbonaceous products. The mixtures recovered from the autoclave were black powdered products in which it could not be distinguished between the solid coming from the experiment and the Red Mud dyed and covered with coke. For this reason, this mixture was analyzed as a whole, and therefore the values of C and H contents and HHV are quite lower than those of the heavy fractions obtained in the thermal runs. The “others” term is representative of the proportion of Red Mud in the solid fraction. The different proportion of Red Mud in both products is due to the fact that more solid was produced in the experiment with PO2. However, the other wt.% does not correspond to 10 g, which was the amount of Red Mud used in the experiments. This is attributed to the heterogeneity of these fractions, which makes it difficult to withdraw perfectly representative samples for the analyses. It is noteworthy that the amount of chlorine of these heavy fractions (4.5 and 4.6 wt.%) is rather high compared to the liquid heavy fractions obtained in the thermal runs. This can be explained by the capacity of Red Mud to adsorb chlorine both physically (in the form of HCl) and chemically (trough reaction to form FeCl3) [37,43]. Consequently, the

236

A. Lopez-Urionabarrenechea et al. / Fuel Processing Technology 137 (2015) 229–239

Table 8 Elemental composition (wt.%) and HHV (MJ kg−1) of the heavy fractions obtained in the upgrading experiments of pyrolysis oils. Thermal run

C H N Cl Othersa H/C ratio HHV (MJ kg−1) a

Table 9 Composition (vol.%) and HHV (MJ Nm−3) of the gases generated in the upgrading experiments of pyrolysis oils.

Catalytic run

Thermal run

PO1

PO2

PO1

PO2

90.8 8.9 b0.1 0.1 0.1 1.2 40.8

90.6 9.0 b0.1 0.2 0.1 1.2 40.7

39.1 2.4 b0.1 4.5 53.9 0.7 15.3

51.7 3.0 b0.1 5.6 39.7 0.7 21.9

By difference.

chlorine is strongly trapped in the Red Mud which is mixed with the heavy fraction. The dechlorination capacity of Red Mud has also been proved by the authors in previous pyrolysis experiments [18]. 3.5. Gas fraction Table 9 shows the composition and HHV of the gases obtained in all the experiments. The compositions presented in the table are in a HCl free basis, since although chlorine evolves from the oils as hydrogen chloride, HCl would be easily removed from gases just by means of water absorption in an hypothetic industrial plant, obtaining chlorinefree gaseous fractions. Some chlorinated organic compounds in the range C1–C4 could also appear in such fractions but their presence or absence could not have been corroborated through GC-TCD/FID analysis. There are not significant differences between the four gaseous fractions. All of them are mainly composed of methane (51–58 vol.%), ethane (18–22 vol.%) and propane (9–12 vol.%), that is, saturated hydrocarbons. Noticeable quantities of hydrogen (4–7 vol.%) are also worth mentioning, while CO, CO2 and other hydrocarbons from C2 to C6 appear in rather low concentrations. More methane is obtained in the gases derived from the catalytic runs compared to the thermal runs, together with a slight depletion of C2, C3 and C4 hydrocarbons,

Fig. 4. Heavy fractions obtained in the pyrolysis oils thermal upgrading experiments.

H2 CO CO2 CH4 C2H6 C3H8 C2H4 C3H6 C4 fraction C5 fraction C6 fraction HHV (MJ Nm−3)

Catalytic run

PO1

PO2

PO1

PO2

4.5 3.0 0.9 53.1 21.9 11.6 0.8 1.3 2.4 0.4 0.1 50.3

6.7 3.8 1.1 51.2 19.4 12.0 0.9 1.2 3.0 0.6 0.1 49.8

4.7 1.5 2.6 56.8 20.9 9.5 0.4 0.9 2.2 0.4 0.1 48.1

6.0 1.6 3.2 57.7 17.9 9.7 0.4 0.8 2.2 0.4 0.1 46.8

which may be attributed to the stronger cracking that is produced in the presence of catalyst. This fact was clearly demonstrated by the significant increase in gas yields that is produced in the catalytic runs, as it has been previously mentioned. Finally, it is also worth noting that the HHV of the gases is in all cases even higher than that of natural gas (37–40 MJ Nm−3). For this reason this gaseous fractions are valuable streams to be used as gaseous fuels to supply the energetic demand of the process, and/or to valorize or sell them in case of surplus. For this to be possible, the absence of chlorinated organic compounds should be corroborated. 3.6. Chemicals balance in the thermal upgrading processes Tables 10 and 11 show the chemicals total balance of the thermal experiments carried out with PO1 and PO2 samples respectively. In each table, the detailed composition of the initial sample can be seen in the first column (the same shown in Table 2), while in the second column the total quantity of each compound in the “total product” (gases + light fraction + heavy fraction) is shown. It has to bear in mind that in the case of the liquid products the calculation of this second column has been made by multiplying the area percentage of each compound in the corresponding fraction by the fraction yield (Fig. 2, wt.%). For this reason, the figures of the second columns of Tables 10 and 11 do not reflect the exact concentration of the compound in the “total product”, but it is enough to get a general idea about the possible chemical pathways during the experiments. The contribution of gaseous compounds to this “total product” has been calculated by using the weight composition of the gas fractions. On the other hand, the chemicals balance of the catalytic experiments has not been included here because, as it has been commented in Section 3.4, the heavy fraction obtained in the catalytic pyrolysis experiments were carbonaceous solid materials, whose detailed chemical composition was not possible to determine in the way the liquid samples were. Table 10 confirms the tendencies that can be inferred from the comparison of Table 3 with Tables 5 and 7, i.e., there are a net dearomatization and a net hydrogenation of alkenes in the upgrading process. It seems at first sight that it is quite difficult that these two reactions types happen at the same time since both of them require hydrogen to take place and no hydrogen was added to the reaction medium in these experiments. However, a deep evaluation of the chemical pathways could explain this fact. Looking at the transformations of aromatic compounds, Table 10 shows that styrene, α-methyl-styrene, propenyl-benzene and indene completely disappear during the process, increasing the percentages of other aromatics like ethyl-benzene (preferentially), toluene, xylenes and methyl-ethyl-benzene, and generating new alkane-substituted mono-ring aromatic compounds (alkyl-benzenes) such as trimethylbenzenes or butyl-benzenes. However, the increasing of already existing monoaromatic compounds plus the generation of new

A. Lopez-Urionabarrenechea et al. / Fuel Processing Technology 137 (2015) 229–239 Table 10 Chemicals balance in the thermal upgrading of PO1.

237

Table 11 Chemicals balance in the thermal upgrading of PO2.

Compound name

Original sample

“Total product”

Compound name

Original sample

“Total product”

Hydrogen Carbon monoxide Carbon dioxide Methane Ethane Ethylene Propane Propylene C4 compounds C5 gaseous compounds C6 gaseous compounds Pentane 2-Methyl-pentane Hexane Heptane 4-Methyl-heptane Octane 2,4-Dimethyl-heptane 2,4-Dimethyl-heptene Propyl-cyclohexane Benzene Toluene Ethyl-benzene Styrene α-Methyl-styrene Xylenes 1-Methyl-ethyl-benzene Propyl-benzene 1-Propenyl-benzene 1-Ethyl-2-methyl-benzene Trimethyl-benzenes 1-Methyl-propyl-benzene Butyl-benzene 1-Phenyl-1-butene Indene 1-Nonene 1-Decene Undecane 1-Undecene Dodecane 1-Dodecene Tridecane 1-Tridecene 1-Tetradecene 1-Pentadecene 1-Hexadecene 1-Octadecene Naphthalene 2-Methyl-naphthalene Biphenyl 3-Phenyl-propyl-benzene 2-Methyl-1,1-diphenyl-2-propene Anthracene 2-Phenyl-naphthalene 2-Phenyl-methyl-naphthalene Terphenyl Fluoranthene Unidentified compounds

– – – – – – – – – – – n.d.a n.d.a n.d.a n.d.a n.d.a n.d.a n.d.a 5.1 0.9 1.7 10.1 8.2 42.9 4.4 2.6 0.7 n.d.a 0.8 n.d.a n.d.a n.d.a n.d.a n.d.a 1.0 0.8 1.6 n.d.a 1.6 n.d.a 1.1 n.d.a 1.0 0.9 0.9 0.6 0.7 1.6 1.0 0.7 1.2 0.7 n.d.a 1.5 n.d.a n.d.a n.d.a 5.7

b0.1 0.5 0.3 5.6 4.3 0.1 3.4 0.4 0.9 0.2 b0.1 1.7 2.3 1.1 1.3 2.2 1.6 1.2 n.d.a n.d.a 1.4 14.0 21.0 n.d.a n.d.a 4.6 6.2 2.8 n.d.a 1.2 3.3 0.2 1.1 0.2 n.d.a n.d.a n.d.a 1.6 n.d.a 1.1 n.d.a 0.8 n.d.a n.d.a n.d.a n.d.a n.d.a 2.3 2.4 0.2 n.d.a n.d.a 0.1 2.8 0.2 0.1 2.0 3.4

Hydrogen Carbon monoxide Carbon dioxide Methane Ethane Ethylene Propane Propylene C4 gaseous compounds C5 gaseous compounds C6 gaseous compounds Pentane 2-Methyl-pentane Hexane 2,3-Dimethyl-pentane Heptane 4-Methyl-heptane Octane 2,4-Dimethyl-heptane 2,4-Dimethyl-heptene Benzene Toluene Ethyl-benzene Styrene α-Methyl-styrene Xylenes Propyl-benzene 1-Propenyl-benzene 1-Methyl-ethyl-benzene 1-Ethyl-2-methyl-benzene Trimethyl-benzenes 1-Ethenyl-3-methyl-benzene Butyl-benzene Indene 2-Methyl-indene Decane Undecane Dodecane Naphthalene 2-Methyl-naphthalene 1-Ethyl-naphthalene 2-Ethenyl-naphthalene Tridecane Tetradecane Pentadecane Hexadecane Heptadecane Octadecane Nonadecane Biphenyl 4-Methyl-biphenyl 3-Phenyl-propyl-benzene 1,1-Cyclopropylidenebis-benzene Anthracene Phenyl-naphthalenes 2-Phenyl-methyl-naphthalene Phenyl-indenes 3-Methyl-1-phenyl-indene Terphenyl 5-Methyl- terphenyl Unidentified compounds

– – – – – – – –– – – – n.d.a n.d.a n.d.a n.d.a n.d.a n.d.a n.d.a n.d.a 1.6 4.5 13.1 10.5 27.9 3.9 9.8 1.1 1.0 1.0 3.0 2.0 n.d.a n.d.a 1.8 2.9 n.d.a n.d.a n.d.a 3.2 4.8 0.7 0.8 n.d.a n.d.a n.d.a n.d.a n.d.a n.d.a n.d.a n.d.a n.d.a 1.0 0.6 n.d.a 2.5 n.d.a 0.8 0.8 n.d.a n.d.a 0.7

b0.1 0.4 0.2 2.9 2.1 b0.1 1.9 0.2 0.6 0.2 b0.1 0.6 1.5 0.4 0.8 0.6 1.2 0.7 1.1 n.d.a 1.6 13.8 20.5 n.d.a n.d.a 7.3 2.8 n.d.a 6.8 2.2 2.9 0.9 0.7 n.d.a n.d.a 0.8 0.7 0.6 3.2 4.5 n.d.a n.d.a 0.6 0.5 0.4 0.4 0.5 0.4 0.5 0.6 0.4 n.d.a n.d.a 0.5 4.4 1.3 n.d.a n.d.a 0.7 0.9 3.0

a

Not detected.

a

monoaromatic compounds do not match the disappearance of the above mentioned four compounds (49% disappearance vs 33% increasing + appearance). Some of these monoaromatic compounds can be the source of polyaromatic compounds (from naphthalene to fluoranthene in Table 10), since there is an increasing of these compounds of around 4%. In such transformation hydrogen is released, and consequently available for the abovementioned hydrogenation of alkanes, which will be explained later on. The remaining 12% of aromatic compounds must be the source of new non aromatic products, by cracking of the aromatic rings and stabilization of the cracked carbon chains under the process conditions.

Not detected.

Concerning the transformation of alkenes, as it has been commented before, they disappear almost completely. Looking at Table 10, it can be seen that some of them (e.g., undecene, dodecene, tridecene) quantitatively give their corresponding alkanes, while other ones would be the source of alkanes of lower carbon number, such as the gaseous products or the liquid products from pentane to dimethyl-heptane. This case, the pathway would be cracking and hydrogenation. The behavior of PO2 sample, which can be inferred from Table 11, is quite similar to that of PO1. In general, there is also a disappearance of

238

A. Lopez-Urionabarrenechea et al. / Fuel Processing Technology 137 (2015) 229–239 Liquids

Solids

Gases

90 84.0 80.5

79.7

80

70

60 51.0 (wt.%)

50 43.3 40

30

20

17.8 13.0

10

7.2

5.6

7.3

8.6

1.6 0 PO1 thermal run

PO2 thermal run

PO1 catalytic run

PO2 catalytic run

Fig. 5. Chlorine distribution among the products obtained in the thermal and catalytic upgrading of PO1 and PO2 samples.

styrene-derived products and indene-derived products together with an increasing of alkyl-benzenes like ethyl-benzene or methyl-ethylbenzene. However, this sample did not generate any new monoaromatic compound as it was the case of PO1. On the other hand, the generation of poly-aromatic compounds is higher (+5.4%) compared to that of sample PO1 (+4%), which means more generation of hydrogen which can be available for hydrogenation reactions. Concerning non aromatic products, apart from the gaseous products and the short chain liquid alkanes (from pentane to dimethyl-heptane) there is a remarkable generation of C10–C19 carbon number alkanes (+ 5.1%), whose origin must be the aromatic compounds since PO2 is almost completely composed of aromatic compounds. This result is quite surprising since the formation of alkanes from aromatic compounds is not a very usual route. One possible pathway to generate the long chain alkanes could be based on ring-opening reactions followed by alkyl displacement (transposition). If necessary, the combination of more than one opened ring could happen. On the other hand, the short chain alkenes could be formed by the separation (through cracking) of the aromatic ring and the aliphatic chain of the alkyl-benzenes or by the cracking of the ring itself followed by hydrogenation. However, other chemical pathways could be also possible, since it is a very complex chemical system whose mechanisms are not easy to infer, due to the high number of chemical species involved and the high number of reaction possibilities. 3.7. Chlorine distribution The chlorine balance of the upgrading process can be seen in Fig. 5. In this figure, the percentage of the initial chlorine retained in each product is shown. Therefore, the distribution of the chlorine can be discussed. Two clear different patterns can be observed in Fig. 5. On the one hand, most of the chlorine is found in the gas fraction when the oils are upgraded through thermal pyrolysis (80.6 wt.% with PO1 and 79.7 wt.% with PO2), which corroborates the fact that, in thermal pyrolysis of chlorinated organic samples, chlorine is mainly released as HCl that is recovered in the gas fraction, together with some chlorinated organic compounds in the range C1–C4. This behavior has been

previously observed and reported by the authors [16,19]. The small differences between the chlorine distribution in the liquid and solid fractions of both experiments are attributed to the differences in the products yields of each experiment, since, as it can be seen in Tables 6 and 8, the chlorine quantity of the products obtained in the thermal experiments of PO1 and PO2 are very similar. On the other hand, when Red Mud is used as catalyst in the process, the percentage of the initial chlorine retained in the solid fraction becomes quite high, counting 43.3 wt.% with PO1 and 84.0 wt.% with PO2. The reason for such an elevated quantity of chlorine in the solid fractions is the presence of Red Mud which, as it has been explained in Section 3.4., possesses high capability to trap the chlorine. In this case, the difference between the quantity of chlorine retained in the heavy fractions of both samples cannot be attributed only to the higher heavy fraction yield of PO2 compared to that of PO1. In fact, Table 8 shows that the chlorine content of the heavy fraction of the PO2 is higher than the chlorine content of the heavy fraction obtained from the catalytic pyrolysis of PO1. The explanation could lie on the fact that the heavy fraction obtained in the catalytic pyrolysis of PO2 was a solid carbonaceous material, which could show chlorine adsorption capacity, retaining the chlorine together with Red Mud. This high retention of chlorine in the heavy fraction is the reason why, contrary to what it happens in the thermal runs, there are important differences in the percentage of chlorine presented in the gas fraction between the two oils, PO1 and PO2. While around half of the chlorine (51 wt.%) can be found in the gas fraction obtained in the catalytic pyrolysis of PO1, only the 8.7 wt.% of the initial chlorine is found in the gas fraction obtained in the catalytic pyrolysis of PO2, which is the result of the strong retention of chlorine in the solid fraction of this experiment. 4. Conclusions The upgrading of chlorinated oils derived from pyrolysis of plastic waste can be carried out by means of thermal and catalytic cracking. The upgrading process produces an abundant valuable liquid light fraction with no olefins and very low chlorine content, together with a gaseous fraction with greater high heating value than natural gas and a

A. Lopez-Urionabarrenechea et al. / Fuel Processing Technology 137 (2015) 229–239

heavy fraction which remains in the autoclave. The proportions and properties of the upgrading products depend on the characteristics of the initial oil and on whether catalyst is used or not. In this sense, Red Mud has proved to have catalytic activity, promoting cracking and dechlorination reactions. This catalyst reduces the chlorine content of the liquid light fractions to the half compared to that of the fractions obtained in thermal cracking, being closed to make chlorine disappear from these liquid fractions. Highly aromatic oils give rise to high heavy fraction yields and low gas yield by thermal cracking. This kind of oils needs the use of Red Mud in order to further crack and produce higher quantities of light oil and gas. On the contrary, oils with lesser content of aromatics are upgraded enough by thermal cracking, and the use of Red Mud makes the gaseous product be the main fraction of the process, focusing the upgrading process to the generation of gas. The results obtained in this work open the possibility of using the upgraded pyrolysis oils as a feedstock for refineries, which could close the loop of plastic waste recycling, from refinery to plastic goods and from plastic waste to refinery. Besides, the utilization of another waste as catalyst (Red Mud) increases significantly the sustainability of the process. However, there is still a long way to go from research laboratories to industrial refineries and more research concerning this topic is needed. It is expected and desired it will take place in the forthcoming years. Acknowledgments The authors thank the Spanish Ministry of Education and Science (MEC) (CTQ 2007-67070/PPQ) as well as the Basque Country Government (ETORTEK 2007IE07-207, GIC 07-09-IT-354-07 and the Researchers' formation program — 2008, 2009, 2010) for financial assistance for this work. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.fuproc.2015.04.015. References [1] A.G. Buekens, Some observations on the recycling of plastics and rubber, Conserv. Recycl. 1 (1977) 247–271. [2] W. Kaminsky, Pyrolysis of plastic waste and scrap tyres in a fluid bed reactor, Resour. Recover. Conserv. 5 (1980) 205–216. [3] E.A. Williams, P.T. Williams, The pyrolysis of individual plastics and a plastic mixture in a fixed bed reactor, J. Chem. Technol. Biotechnol. 70 (1997) 9–20. [4] J. Aguado, D.P. Serrano, Feedstock Recycling Of Plastic Wastes, The Royal Society of Chemistry, Cambridge, 1999. [5] J. Scheirs, W. Kaminsky, Feedstock Recycling And Pyrolysis Of Waste Plastics, John Wiley & Sons, Chichester, 2006. [6] A. Lopez-Urionabarrenechea, I. de Marco, B.M. Caballero, M.F. Laresgoiti, A. Adrados, Influence of time and temperature on pyrolysis of plastic wastes in a semi-batch reactor, Chem. Eng. J. 173 (2011) 62–71. [7] S.M. Al-Salem, P. Lettieri, J. Baeyens, Recycling and recovery routes of plastic solid waste (PSW): a review, Waste Manag. 29 (2009) 2625–2643. [8] A.K. Panda, R.K. Singh, D.K. Mishra, Thermolysis of waste plastics to liquid fuel. A suitable method for plastic waste management and manufacture of value added products—a world prospective, Renew. Sustain. Energy Rev. 14 (2010) 233–248. [9] A. Lopez-Urionabarrenechea, I. de Marco, B.M. Caballero, M.F. Laresgoiti, A. Adrados, Pyrolysis of municipal plastic wastes: Influence of raw material composition, Waste Manag. 30 (2010) 620–627. [10] A. Lopez-Urionabarrenechea, I. de Marco, B.M. Caballero, A. Adrados, M.F. Laresgoiti, Empiric model for the prediction of packaging waste pyrolysis yields, Appl. Energy 98 (2012) 524–532. [11] N. Kiran, E. Ekinci, C.E. Snape, Recycling of plastic wastes via pyrolysis, Resour. Conserv. Recycl. 29 (2000) 273–283. [12] K.H. Lee, Thermal and catalytic degradation of pyrolytic oil from pyrolysis of municipal plastic wastes, J. Anal. Appl. Pyrolysis 85 (2009) 372–379. [13] J.M. Arandes, M.J. Azkoiti, I. Torre, M. Olazar, P. Castaño, Effect of HZSM-5 catalyst addition on the cracking of polyolefins pyrolysis waxes under FCC conditions, Chem. Eng. J. 132 (2007) 17–26.

239

[14] J.M. Arandes, I. Torre, M.J. Azkoiti, P. Castaño, J. Bilbao, H. de Lasa, Effect of catalyst properties on the cracking of polypropylene pyrolysis waxes under FCC conditions, Catal. Today 133–135 (2008) 413–419. [15] A. Adrados, I. De Marco, A. Lopez-Urionabarrenechea, B.M. Caballero, M.F. Laresgoiti, Pyrolysis behavior of different type of materials contained in the rejects of packaging waste sorting plants, Waste Manag. 33 (2013) 52–59. [16] A. Lopez-Urionabarrenechea, I. de Marco, B.M. Caballero, M.F. Laresgoiti, A. Adrados, Dechlorination of fuels in pyrolysis of PVC containing plastic wastes, Fuel Process. Technol. 92 (2011) 253–260. [17] I. de Marco, B.M. Caballero, A. Lopez-Urionabarrenechea, M.F. Laresgoiti, A. Torres, M.J. Chomón, Pyrolysis of the rejects of a waste packaging separation and classification plant, J. Anal. Appl. Pyrolysis 85 (2009) 384–391. [18] A. Lopez-Urionabarrenechea, I. de Marco, B.M. Caballero, M.F. Laresgoiti, A. Adrados, A. Aranzabal, Catalytic pyrolysis of plastic wastes with two different types of catalysts: ZSM-5 zeolite and Red Mud, Appl. Catal. B Environ. 104 (2011) 211–219. [19] A. Lopez-Urionabarrenechea, I. de Marco, B.M. Caballero, M.F. Laresgoiti, A. Adrados, Catalytic stepwise pyrolysis of packaging plastic waste, J. Anal. Appl. Pyrolysis 96 (2012) 54–62. [20] S. Czernik, A.V. Bridgwater, Overview of applications of biomass fast pyrolysis oil, Energy Fuel 18 (2004) 590–598. [21] D. Mohan, C.U. Pittman Jr., P.H. Steele, Pyrolysis of wood/biomass for bio-oil: a critical review, Energy Fuel 20 (2006) 848–889. [22] N. Miskolczi, L. Bartha, G. Deák, B. Jóver, Thermal degradation of municipal plastic waste for production of fuel-like hydrocarbons, Polym. Degrad. Stab. 86 (2004) 357–366. [23] N. Miskolczi, A. Angyal, L. Bartha, I. Valkai, Fuels by pyrolysis of waste plastics from agricultural and packaging sectors in a pilot scale reactor, Fuel Process. Technol. 90 (2009) 1032–1040. [24] K.H. Lee, K.Y. Nam, Continuous thermal degradation of pyrolytic oil in a bench scale CSTR reaction system, Fuel Process. Technol. 91 (2010) 514–519. [25] K.H. Lee, S.C. Oh, Thermal and catalytic degradation of pyrolytic waxy oil in a plug flow reactor, J. Anal. Appl. Pyrolysis 93 (2012) 19–23. [26] K.H. Lee, Effects of the types of zeolites on catalytic upgrading of pyrolysis wax oil, J. Anal. Appl. Pyrolysis 94 (2012) 209–214. [27] H.S. Joo, J.A. Guin, Continuous upgrading of a plastic pyrolysis liquid to an environmentally favorable gasoline range product, Fuel Process. Technol. 57 (1998) 25–40. [28] N. Lingaiah, M.A. Uddin, A. Muto, T. Imai, Y. Sakata, Removal of organic chlorine compounds by catalytic dehydrochlorination for the refinement of municipal waste plastic derived oil, Fuel 80 (2001) 1901–1905. [29] N. Lingaiah, M.A. Uddin, A. Muto, Y. Sakata, T. Imai, K. Murata, Catalytic dechlorination of chloroorganic compounds from PVC-containing mixed plastic-derived oil, Appl. Catal. A Gen. 207 (2001) 79–84. [30] N. Miskolczi, F. Buyong, A. Angyal, P.T. Williams, L. Bartha, Two stages catalytic pyrolysis of refuse derived fuel: production of biofuel via syncrude, Bioresour. Technol. 101 (2010) 8881–8890. [31] F. Abushwireb, H. Elakrami, M. Emtir, Recovery of aromatics from pyrolysis gasoline by conventional and energy-integrated extractive distillation, Comput. Aided Chem. Eng. 24 (2007) 1071–1076. [32] J.H. Gary, G.E. Handwerk, Refinery feedstocks, Petroleum Refining, CRC Press, London, 2001. [33] A.G. Buekens, H. Huang, Catalytic plastics cracking for recovery of gasoline range hydrocarbons from municipal plastic wastes, Resour. Conserv. Recycl. 23 (1998) 163–181. [34] P.T. Williams, E.A. Williams, Fluidised bed pyrolysis of low density polyethylene to produce petrochemical feedstock, J. Anal. Appl. Pyrolysis 51 (1999) 107–126. [35] W. Kaminsky, I.J. Nuñez, Catalytical and thermal pyrolysis of polyolefins, J. Anal. Appl. Pyrolysis 79 (2007) 368–374. [36] S. Wang, H.M. Ang, M.O. Tadé, Novel applications of red mud as coagulant, adsorbent and catalyst for environmentally benign processes, Chemosphere 72 (2008) 1621–1635. [37] T. Karayildirim, J. Yanik, M. Yuksel, M. Saglam, C. Vasile, H. Bockhorn, The effect of some fillers on PVC degradation, J. Anal. Appl. Pyrolysis 75 (2006) 112–119. [38] S. Sushil, A.M. Alabdulrahman, M. Balakrishnan, V.S. Batra, R.A. Blackley, J. Clapp, J.S.J. Hargreaves, A. Monaghan, I.D. Pulford, J.L. Rico, W. Zhou, Carbon deposition and phase transformations in red mud on exposure to methane, J. Hazard. Mater. 180 (2010) 409–418. [39] T. Bhaskar, J. Kaneko, A. Muto, Y. Sakata, E. Jakab, T. Matsui, Md.A. Uddin, Pyrolysis studies of PP/PE/PS/PVC/HIPS-Br plastics mixed with PET and dehalogenation (Br, Cl) of the liquid products, J. Anal. Appl. Pyrolysis 72 (2004) 27–33. [40] T. Bhaskar, M. Tanabe, A. Muto, Y. Sakata, C.F. Liu, M.D. Chen, C.C. Chao, Analysis of chlorine distribution in the pyrolysis products of poly(vinylidene chloride) mixed with polyethylene, polypropylene or polystyrene, Polym. Degrad. Stab. 89 (2005) 38–42. [41] M.F. Ali, M.N. Siddiqui, Thermal and catalytic decomposition behavior of PVC mixed plastic waste with petroleum residue, J. Anal. Appl. Pyrolysis 74 (2005) 282–289. [42] M. Blazsó, E. Jakab, Effect of metals, metal oxides, and carboxylates on the thermal decomposition processes of poly (vinyl chloride), J. Anal. Appl. Pyrolysis 49 (1999) 125–143. [43] J. Yanik, Md.A. Uddin, K. Ikeuchi, Y. Sakata, The catalytic effect of Red Mud on the degradation of poly(vinyl chloride) containing polymer mixture into fuel oil, Polym. Degrad. Stab. 73 (2001) 335–346.