Chemical Recycling of Plastic Marine Litter: First Analytical Characterization of The Pyrolysis Oil and of Its Fractions and Comparison with a Commercial Marine Gasoil

: A detailed molecular ﬁngerprint of raw pyrolysis oil from plastic wastes is a new research area. The present study focuses for the ﬁrst time on the chemical recycling of plastic marine litter; we aim to chemically characterize the obtained raw pyrolysis oil and its distillates (virgin naphtha and marine gasoil) via GC-MS and FT-IR. For all samples, more than 30% of the detected compounds were identiﬁed. 2,4-dimethyl-1-heptene, a marker of PP pyrolysis, is the most represented peak in the chemical signature of all the marine litter pyrolysis samples, and it differentiates commercial and pyrolysis marine gasoil. The presence of naphthalenes is stronger in commercial gasoil, compared to its pyrolysis analog, while the opposite holds for oleﬁns. The overlap between the two molecular ﬁngerprints is impressive, even if saturated hydrocarbons are more common in commercial gasoil, and unsaturated compounds are more common in the gasoil derived from pyrolysis. A technical comparison between the commercial marine gasoil and the one obtained from the marine litter pyrolysis is also attempted. Gasoil derived from marine litter fully complies with the ISO8217 standards for distillate marine fuel. On the other hand, the virgin naphtha is particularly rich in BTX, ethylbenzene, styrene, and alpha oleﬁns, which are all important recoverable platform chemicals for industrial upcycling. reactions, and recombination


Introduction
Plastic waste is a long-lasting and worldwide ecological problem. Plastic has becomethe emblem of waste (disposable objects), pollution, and ecotoxicity. It is estimated that in 2018, an estimated 14.5 million tons of plastics entered the ocean [1], building up each year a stock of plastic waste undergoing progressive fragmentation and releasing microparticles that enter the food chain with unknown consequences. Personal protective and single-use products associated with the current COVID-19 pandemic will probably increase the quantities of plastic entering the oceans [2]. Landfilling, incineration with or without energy recovery, and recycling are common ways to manage plastic waste. However, safe landfilling is expensive and does not fit the sustainable circular economy vision; incineration may lead to the emission of harmful greenhouse gases (NOx, SOx, COx, etc.) and other toxic volatiles. Noteworthy is that not all plastics can be mechanically recycled, considering their degradation states, complex compositions, or mixtures with other substances [3]; and sorting mixed plastic waste is a labor intensive and expensive prerequisite of the process since most plastics are not compatible with each other and can't be processed together during recycling [4]. This is particularly true for plastic marine litter, which is mainly landfilled or incinerated [5]. from the pyrolysis of mixed plastic waste from a recycling facility; aliphatic compounds were not detailed. The reaction temperature had a positive effect on the BTX yield [27].
A high resemblance of pyrolysis oil from municipal plastic waste with standard diesel was demonstrated on the basis of thorough GC-MS analyses, even if the sampling conditions were not detailed, and compounds lighter than C8 alkanes and C7 aromatics were not detected [28].
From the review of the literature, it follows that, to date, only a few studies describe a detailed molecular fingerprint of pyrolysis oil from plastics wastes, and just two studies deal also with the characterization of its distillates. Singh et al. indicated 13 major compounds in pyrolysis oil from mixed plastics and compared its composition to that of HDPE, PP, and PS pyrolysis oil; the most abundant hydrocarbon fraction in pyrolysis oil based on carbon number is the C5-C12 fraction. The distilled fraction with a boiling point up to 240 • C was found to contain 13.85% 2,4-dimethyl-1-heptene, 8.67% 1,1,3,4-tetramethyl cyclopentane, 4.35% 2-Octen-4-one, 4.27% 1,3,5-trimethyl cyclohexane, and higher quantity of the C9 and C10 compounds and branched alkanes and alkenes [29]. The rejected fraction of the packaging waste was pyrolyzed. GC-MS results for the pyrolysis oil and its distillate revealed high levels of aromatics (recoverable building blocks for industrial use), alkanes, cycloalkanes, alkenes, cycloalkenes, and oxygenated compounds for a total of 33 analytes. However, the authors did not describe how the samples were processed. Styrene was the principal aromatic compound, followed by ethylbenzene and toluene [3].
Noteworthy is that no investigation (neither exhaustive nor general) on the chemical composition of the raw pyrolysis oil (and its distillates) from marine litter is available in the scientific literature.
Taking into account this scarcity of available data and the fact that pyrolysis oil from plastic waste can lead to alternative fuels and valuable chemicals for a more sustainable industry, we believe that it is crucial to provide the detailed chemical composition and obtain a molecular fingerprint of the fuels that can be obtained via the chemical recycling of the marine litter for both scientific and commercial reasons. In the following, we describe the pyrolysis process of the marine litter and its characterization, especially at the molecular level via the GC-MS analysis of the raw pyrolysis oil (RPO ml-py) and its distilled fractions, namely, virgin naphtha (VN ml-py) and marine gasoil (MGO ml-py). The FTIR analysis provides an additional chemical signature of the samples. A comparison between the molecular signature of the MGO obtained from marine litter pyrolysis and the commercial MGO is also carried out along with their physicochemical characterization. This comparison may pave the way to the removal of regulatory barriers within a chemical recycling regulatory framework.

Plastic Wastes Feedstock
Within the EU project "Mapping and recycling of marine litter and Ghost nets on the sea floor" (marGnet), more than 200 kg of actual plastic marine litter was collected from the sea bottom of the Venice Lagoon and north Adriatic Sea by scuba divers with the goal to pyrolyze it "as it is" without any sorting or pretreatment and upgrade the pyrolysis oil to ISO-compliant marine fuel by means of distillation. The focus on sunk litter is justified because it is estimated that as much as 70% of marine debris sinks to the seabed, and even low-density polymers can lose buoyancy under the weight of fouling [30]. Although a precise characterization of the plastic litter was not possible due to the high amount of dirt and organic incrustation or the lack of identifying markings, two thirds of the collected plastic litter were general plastic containers, such as hard boxes and packaging films, plastic bottles, plastic nets from mussel farming, and a few boat components, and one third was fishing nets. This is confirmed in literature where it is reported that plastic items, bags, and fishing equipment are the largest part of the debris at sea [31], and benthic marine litter in the central Mediterranean Sea is composed of 68% general plastic objects and 32% fishing gear [32]. A small contribution of floating litter mainly composed of styrofoam and floating containers was also included in the feedstock. The great majority of postconsumer plastic waste is made from just three polymer families: PE, PP, and PET [33], while the fishing nets are mainly made of nylons (6 and 6.6) and PE, and mussel nets are made of PP, we can assume that the composition of the plastic marine litter was essentially made of polyolefins and polyamides with contributions of polystyrene and rubber. Organic matter, such as seaweed, mussel residue, etc., was the main recognizable contaminant along with sand. Nonplastic items easily recognizable, such as glass bottles and metals cages, were manually segregated and removed during collection. Scrap tires were also present as litter in the sea, but their pyrolysis products are not discussed in this work.

Pyrolysis Process
For the pyrolysis experiments, a cylindrical 100 L volume reactor, heat insulated, AISI 304 stainless steel, internal diameter 320 mm, 1:3 diameter to height ratio, electrically heated was employed (provider SINTOL, Torino, Italy). The reactor wall was maintained at 400 +/− 50 • C temperature and autogenic pressure by electric heaters (total rated power 15 KW). A shell and tube AISI 304 stainless steel 0.42 m 2 condenser, water-cooled, located just at the outlet of the pyrolysis reactor collected the vapors of RPO (ml-py) in a receiving tank. Noncondensable gas was flared after conditioning. The operation was batchwise, and 14 pyrolysis cycles were performed with loads of dry plastic marine litter ranging from a minimum of 4 kg to a maximum of 11.5 kg for a total of 100 kg. Each pyrolysis cycle lasted 4 h from start to end. Pyrolysis oil collected from each cycle was homogenized in order to provide the overall mean composition. After every pyrolysis cycle, the reactor was cleaned and solid residues were removed. A total of 1 kg of calcium oxide-based additive (content of CaO > 57 wt%; CaO + MgO > 92 wt%; CalcePiasco, Italy) was added to the feedstock in the reactor for each cycle to avoid the formation of acidic compounds, such as hydrochloric acid and other organic acids, as also described elsewhere [34]. Mass balance was 45 wt% RPO (ml-py), 26 wt% noncondensed gas, and 26 wt% solid residues (net of additive).

Pyrolysis Oil Distillation
Simple distillation of the RPO (ml-py) was performed with a 50 L AISI 304 stainless steel distillation flask. The first cut temperature was 180 • C collecting the light fraction, namely the VN (ml-py); the second cut temperature was set to 320 • C collecting the targeted product, namely, the MGO (ml-py). Residual product distilling above 320 • C was recovered as residue from the distillation flask. Since simple distillation cannot guarantee sharp splits, some degree of overlapping components in the distillates was expected. Mass balance was 23 vol% VN (ml-py), 52 vol% MGO (ml-py), 6 vol% residue, 10 vol% water, and 9 vol% loss as gas due to cracking reactions.
MGO (my-pl) was filtered with silica gel granules (diameter 0.5 mm, porosity 800 m 2 /g, provider Bolaseca, Murcia, Spain) to remove residual humidity and give it a clearer appearance.

GC-MS Analysis
The RPO (ml-py) and its distilled fractions, namely, VN (ml-py), MGO (ml-py), and MGO (comm), were characterized by gas chromatography to detail their chemical composition and properties. Three replicate samples were used for the chromatographic analysis. A Hewlett Packard GC-MS, G1800C GCD Series II (Palo Alto, Santa Clara, CA, USA), equipped-with an HP-5MS column 30 m × 0.25 mmI.D. × 0.25 µm film thickness (Hewlett-Packard) was used. The mass spectrometer was tuned before the analyses via a reference gas (perfluorotributylamine) across the full mass range. The injector was set at 325 • C and operated in the split mode (split ratio 1:240). The carrier gas was helium with a constant flow of 1 mL/min; the oven temperature was held isothermal at 30 • C for 15 min, then programmed from 30 • C to 100 • C at 5 • C/min, and then programmed from 100 • C to 325 • C at 10 • C/min and held isothermal at 325 • C for 10 min. Mass spectra were acquired in the electron impact mode (70 eV), using a full scan with mass analysis in the range 10 atomic mass units (amu)-450 atomic mass units (amu). The ion source and the quadrupole were heated by conduction. A total of 2.500 mL of each sample were filtered to remove organic materials larger than 5 µm and put in 5 mL headspace vial, closed by PTFE/silicone septum. Both the liquid phase and the volatiles in the headspace of the samples were analyzed, respectively, via the direct injection of 1 µl of the liquid in the injection port of the GC-MS and HS-SPME-GC-MS. SPME fibers were obtained from the Supelco Company (Bellefonte, PA, USA). The fiber (divinylbenzene/carboxen/polydimethylsiloxane 50/30 µm) was selected because it is considered the most universal [35][36][37][38]; it was conditioned before use, as recommended by the manufacturer. Before extraction, the stabilization of the headspace in the vial was reached by equilibration for 30 min at 30 • C ± 0.1 • C. Volatiles were then adsorbed for exactly 10 h at 30 • C ± 0.1 • C. After extraction, injections provided the fiber thermal desorption into the GC-MS injection port equipped with a 0.75 mm i.d. inlet liner. Before the subsequent sampling, the fiber was reconditioned for 5 min at 325 • C. Blank runs (empty injections) were done periodically during the study to reveal possible carryover. Carryover never occurred.
The identification of the analytes was based on the comparison of their retention times with those of standards obtained from Sigma-Aldrich (Milan, Italy). The purity of the volatile standards was usually 97% or higher. In the absence of a commercial standard, peak identification was carried out by (i) computer matching of mass spectral data with those of compounds contained in the Mass Spectral Library of the National Institute of Standards and Technology (NIST 1998 library); the quality of the match above 98% was needed for positive identification; (ii) comparison of their linear retention indices [37,39] relative to n-alkanes, calculated using a straight-chain alkane mixture, with the averaged values reported in the bibliography for chromatographic columns similar to that used. Only compounds with a signal-to-noise ratio higher than 5 were considered. The relative proportions of the constituents in the samples were obtained using the percentage area of chromatographic peaks, as usual [35][36][37][38][39]. The statistical analysis of the variance was performed via the data analysis tool in Microsoft ® Excel for Mac, version 16.16.26 (200914) © 2022 Microsoft.

Fourier Transform Infrared (FTIR) Spectroscopy
The pyrolysis oil, its distillates, and a commercial MGO were analyzed via FTIR (PerkinElmer; Model: Spectrum II) in the wave range of 4000-600 cm −1 via the UATR arrangement to identify functional groups.

ISO8217:2017 Analysis
The MGO (ml-py) was analyzed as per compliance with ISO8217:2017, DMA/DMX, distillate fuel classification. Therefore, all parameters were evaluated according to the ISO 8217:2017 prescribed standard methods for marine fuels (ISO, 2017), except for oxidation stability that was determined by the DIN EN 16091:2012 method.

ISO Compliance of MGO (ml-py)
The marGnet project had the general goal to tackle marine litter from a multilevel approach with a specific focus on what to do with marine litter once recovered from the sea [40]. The answer was to convert marine litter into marine fuels, being a sort of payback for fishermen and other stakeholders involved in the collection activity. Giving back the fuel obtained from the litter the fishermen delivered was a great nonmonetary instrument to attain greater involvement of stakeholders and thus an efficient sea clean up. However, to do so, the produced fuel must be fully compatible with commercial marine gasoil so that no technical issue could arise upon its usage. Compliance with conventional fuel is also crucial to avoid skepticism among stakeholders. Therefore, among the marine fuels classified according to ISO 8217, the highest quality ones, DMX and DMA, were targeted. DMX and DMA are normally commercially referred to MGO.   Table 1 confirms that MGO (ml-py) is compliant with ISO 8217, and therefore this fuel can be used both pure and blended with MGO (comm) in existing engines without any technical modification or adjustment. MGO (comm) is assumed to be compliant with ISO 8217. Noteworthy is the pour point value of −6 • C, which permits the MGO (ml-py) to meet both winter and summer quality requirements, and the sulfur content which is approximately 80% lower than the prescribed value, making the MGO (ml-py) also compliant with the latest requirements of the Emission Control Areas (ECAs) [42]. This is a remarkable result if we consider that MGO (ml-py) was obtained from marine litter once polluting the Adriatic Sea. Figure 1 shows the visual appearance of the VN (ml-py), MGO (ml-py), and RPO (ml-py) samples.

GC-MS Analysis of the Liquid Samples
The composition of VN (ml-py), MGO (ml-py), and RPO (ml-py), along with the composition of MGO (comm) are shown in Table 2, where the percent areas of all the identified compounds (organized by chromatographic retention) are summarized, and different colors are used for the average percent area below 1% (red), between 1% and 4% (green), and

GC-MS Analysis of the Liquid Samples
The composition of VN (ml-py), MGO (ml-py), and RPO (ml-py), along with the composition of MGO (comm) are shown in Table 2, where the percent areas of all the identified compounds (organized by chromatographic retention) are summarized, and different colors are used for the average percent area below 1% (red), between 1% and 4% (green), and above 4% (yellow). Standard deviations were always below 10%. Table 2. Literature and calculated retention indexes (RIlit, RIcalc), CAS number, and average percent area (triplicate injections) of analytes from MGO (comm), VN (ml-py), MGO (ml-py), and RPO (ml-py) samples. Different colors are used for average percent area below 1% (red), between 1% and 4% (green), and above 4% (yellow). Standard deviations were always below 10%.   For MGO (comm), VN (ml-py), MGO (ml-py), and RPO (ml-py), 120, 117, 120, and 131 compounds were identified, respectively, and they represent 34%, 43%, 38%, and 32% of the detected compounds, respectively. Figure 2 details the cumulative average percent area of the identified compounds. Noteworthy, the BTX, ethylbenzene, styrene, and alpha-methylstyrene contents in VN (ml-py) are ca. twofold compared to MGO (ml-py). This is important because they are potentially recoverable platform chemicals for industrial application. BTX separation from petroleum naphtha is state-of-the-art technology, and it is generally performed by liquidliquid extraction with sulfolane [43,44]. Removing BTX from VN (ml-py) would attain a double result: obtaining valuable aromatics for the chemical industry and aromaticfree naphtha, the ideal feedstock for the steam cracking process and the synthesis of polyolefins, and hence, new virgin PE and PP polymers [43]. Besides, building blocks from the pyrolysis are often the monomers of the processed polymers (e.g., styrene) that can be subsequently polymerized again [3]. This latter development is a work in progress at present. Similarly, the content of light aliphatic compounds ranging between C5 and C11 is higher in VN (ml-py) compared to MGO (ml-py); in particular, the contents of the C5-C7 compounds and the C8-C9 compounds in VN (ml-py) are at least fivefold and threefold, respectively, compared to MGO (ml-py). In VN (ml-py), aliphatic compounds with a number of carbon atoms higher than 22 were not detected. On the contrary, the content of aliphatic compounds ranging between C12 and C28 is higher in MGO (ml-py) compared to VN (ml-py), as expected. Many specific compounds detailed in Table 2 were already found in pyrolysis oils from specific plastic materials or mixed plastic waste. The general aspects of the pyrolytic degradation mechanism of polymers involve three kinds of chain reactions viz. polymeric chain scissions, side group reactions, and recombination reactions. For example, the thermal degradation of both HDPE and PP yielded alkanes, alkenes, and alkynes; double-bonded hydrocarbons compounds are generated via the single-bonded carbon (C-C) dissociation in a polyalkane structure and the stabilization of free radicals produced from the chain scissoring mechanism [29]. From the data in Table 2, it is clear that olefins (alkenes, cycloalkenes, and polyenes) constitute a very important family of pyrolysis compounds, comprising 40.45%, 23.23%, and 22.08% of the area of VN (ml-py), MGO (ml-py), and RPO (ml-py) samples, respectively; it follows that VN (ml-py) might be an interesting source of these precious building blocks for the chemical industry. Linear alpha olefins (normal alpha olefins) from 1-hexene to 1-tridecene were typical of HDPE pyrolysis, and 1-heptene and 1-octene were also found to be markers of the PP pyrolysis [11]. We found these compounds to be very expedient in the chemical industry. They are lighter in terms of the homologous series of the linear alpha olefins up to 1-octene and are used (i) as a comonomer in the production of polyethylene and (ii) as feedstock for the hydroformylation process to produce linear aldehyde via oxo synthesis and, subsequently, short-chain fatty acids (via aldehyde oxidation) or linear alcohols (via aldehyde hydrogenation) for plasticizer applications. Abundant amounts of C10-C13 linear alpha olefins can be upcycled in making surfactants: their reaction with benzene generates linear alkylbenzenes (LAB), and their subsequent sulfonation yields linear alkylbenzene sulfonates (LABS), which is a popular low-cost surfactant for household and industrial detergent applications. Alkynes were seldomly identified. Alkanes and cycloalkanes constantly represent ca. 25% of the eluted area for all samples. PP methyl side groups were reported to increase the yield of branched hydrocarbons and alkenes compared to PE [29,45]. Actually, we found, as detailed in Table 2, 2-methyl-1-pentene, 2,4-dimethyl-1heptene, and 3-methylcyclopentene related to PP, as well as trimethyl cyclohexanes [11,29] and many other branched hydrocarbons and alkenes. The largest eluted percent area in the VN (ml-py), MGO (ml-py), and RPO (ml-py) samples is constantly due to 2,4-dimethyl-1-heptene, which represents 16.51%, 4.26%, and 2.74% of the area, respectively. A major source of aromatic building blocks in the pyrolysis oil is PS breakdown, which involves the elimination of styrene moieties thereby increasing the number of aromatic compounds in pyrolysis oil; in addition, PET is a contributor as it is reported that PET pyrolysis leads to a benzene-rich pyrolysis oil, in particular, high temperatures and slow heating rates maximize the amount of aromatics and the selectivity of benzene [46]. However, in this work, pyrolysis was performed in the presence of calcium-based additives, and PET decomposes into terephthalic and benzoic acids at 350 • C, which are subsequently decarboxylated into aromatics thanks to CaO. Kumagai et al. reported an increase in benzene yield when terephthalic acid is decarboxylated in the presence of CaO [47,48].
As regards aromatics identified and detailed in Table 2, benzene, toluene, and xylenes may come from multiple plastic waste sources (HDPE, LDPE, PP, PS, and PET), while ethylbenzene was related to PP and PS pyrolysis; styrene, propyl benzene, and 1-ethyl-2methylbenzene were particularly reported to be markers of PS pyrolysis [11]. Even if [13] it was reported that no aromatic groups were identified in the pyrolysis oil from virgin PP and HDPE, it is well known that aromatics are generated during PP pyrolysis from secondary reactions via cyclization and aromatization reactions, especially at high processing temperatures and heating rates [12]; the low molecular weight olefins (C2-C5) from PE and PP may undergo the Diels-Alder intermolecular cyclization and radical intramolecular cyclization. A subsequent dehydrogenation process rationalizes the formation of aromatic compounds [49][50][51][52]. Furthermore, the interactions of different plastic wastes during the pyrolysis process were reported to increase the content of aromatic compounds [53][54][55][56]. In Table 2, the presence of trimethylbenzenes, indane, and indene can be related to LDPE pyrolysis, even if indene was also related to PP [11]. Naphthalene is generated by the LDPE and PS pyrolysis, while 1-methylnaphthalene, dimethyl naphthalenes, and trimethyl naphthalenes are markers of PVC and LDPE pyrolysis. Biphenyl was related to PET, PVC, and PP pyrolysis feedstocks [11]. Oxygen and nitrogen compounds were previously reported to be a residual fraction in pyrolysis oil samples [3,13,27,57]; actually, only benzonitrile was identified (both in distillates and in the RPO). The lack of oxygenated compounds among those identified in Table 2 is expedient since high oxygen content results in low calorific value, corrosion problems, and instability; the presence of PS can drastically abate the oxygen content, thereby mimicking the diesel standard [57]. Some putative sulfur, nitrogen, and oxygen compounds were detected but not positively identified. involves the elimination of styrene moieties thereby increasing the number of aromatic compounds in pyrolysis oil; in addition, PET is a contributor as it is reported that PET pyrolysis leads to a benzene-rich pyrolysis oil, in particular, high temperatures and slow heating rates maximize the amount of aromatics and the selectivity of benzene [46]. However, in this work, pyrolysis was performed in the presence of calcium-based additives, and PET decomposes into terephthalic and benzoic acids at 350 °C, which are subsequently decarboxylated into aromatics thanks to CaO. Kumagai et al. reported an increase in benzene yield when terephthalic acid is decarboxylated in the presence of CaO [47,48]. As regards aromatics identified and detailed in Table 2, benzene, toluene, and xylenes may come from multiple plastic waste sources (HDPE, LDPE, PP, PS, and PET), while ethylbenzene was related to PP and PS pyrolysis; styrene, propyl benzene, and 1ethyl-2-methylbenzene were particularly reported to be markers of PS pyrolysis [11]. Even if [13] it was reported that no aromatic groups were identified in the pyrolysis oil from virgin PP and HDPE, it is well known that aromatics are generated during PP pyrolysis from secondary reactions via cyclization and aromatization reactions, especially at high processing temperatures and heating rates [12]; the low molecular weight olefins (C2-C5) from PE and PP may undergo the Diels-Alder intermolecular cyclization and radical intramolecular cyclization. A subsequent dehydrogenation process rationalizes the formation of aromatic compounds [49][50][51][52]. Furthermore, the interactions of different plastic wastes during the pyrolysis process were reported to increase the content of aromatic compounds [53][54][55][56]. In Table 2, the presence of trimethylbenzenes, indane, and indene can be related to LDPE pyrolysis, even if indene was also related to PP [11]. Naphthalene is generated by the LDPE and PS pyrolysis, while 1-methylnaphthalene, dimethyl naphthalenes, and trimethyl naphthalenes are markers of PVC and LDPE pyrolysis. Biphenyl was related to PET, PVC, and PP pyrolysis feedstocks [11]. Oxygen and nitrogen compounds were previously reported to be a residual fraction in pyrolysis oil samples [3,13,27,57]; actually, only benzonitrile was identified (both in distillates and in the RPO). The lack of oxygenated compounds among those identified in Table 2 is expedient since high oxygen content results in low calorific value, corrosion problems, and instability; the presence of PS can drastically abate the oxygen content, thereby mimicking the diesel standard [57]. Some putative sulfur, nitrogen, and oxygen compounds were detected but not positively identified. The top three compounds in the eluted area for the VN (ml-py) are 2,4-dimethyl-1heptene (16.51%), 1-hexene (2.87%), and 4-methylheptane (2.30%) followed by short-chain alpha-olefins. Those for MGO (ml-py) are 2,4-dimethyl-1-heptene (4.26%), hexadecane (1.83%), and 1-tridecene (1.71%), followed by linear alkanes and alpha-olefins. For RPO (ml-py) we have 2,4-dimethyl-1-heptene (2.74%), styrene (2.28%), and ethylbenzene (2.1%), mainly followed by linear alkanes with a carbon number higher than 13. For MGO (comm), the most represented compounds are tridecane (2.65%), undecane (2.40%), and dodecane (2.36%), followed by linear alkanes with a carbon number higher than 10. It can be confirmed that 2,4-dimethyl-1-heptene, detected in all samples obtained from the pyrolysis but in MGO (comm), is a good putative marker of the plastic origin of the fuel.
From Figure 3 it is clear that the olefin content is significantly higher in samples from pyrolysis, while paraffins are characteristic of MGO (comm). VN is the richest sample as regards olefins.
In an attempt to go more in depth with the comparison and to confirm or disconfirm similarities, Figure 4 illustrates the averaged percent area of identified analyte in MGO (comm) and MGO (ml-py). The presence of massive amounts of 2,4-dimethyl-1-heptene in the MGO from marine litter pyrolysis oil (ml-py), due to the PP presence [43], differentiate the GC-MS fingerprints of these samples. The presence of naphthalenes is stronger in commercial MGO, compared to its pyrolysis analog. Notwithstanding this, MGO (ml-py) is compliant with ISO8217 specifications as regards marine gasoil, and the incomplete overlap between the molecular fingerprints of MGO (comm) and MGO (ml-py), as evident from Table 2, does not underscore an a priori better or worse performance of the fuel based on marine litter. Further testing of MGO (ml-py) as regards emissions and other parameters is a work in progress since pyrolyzed oil-based fuels could lead to reduced emissions [9,19].
(comm), the most represented compounds are tridecane (2.65%), undecane (2.40%), and dodecane (2.36%), followed by linear alkanes with a carbon number higher than 10. It can be confirmed that 2,4-dimethyl-1-heptene, detected in all samples obtained from the pyrolysis but in MGO (comm), is a good putative marker of the plastic origin of the fuel.
From Figure 3 it is clear that the olefin content is significantly higher in samples from pyrolysis, while paraffins are characteristic of MGO (comm). VN is the richest sample as regards olefins. In an attempt to go more in depth with the comparison and to confirm or disconfirm similarities, Figure 4 illustrates the averaged percent area of identified analyte in MGO (comm) and MGO (ml-py). The presence of massive amounts of 2,4-dimethyl-1-heptene in the MGO from marine litter pyrolysis oil (ml-py), due to the PP presence [43], differentiate the GC-MS fingerprints of these samples. The presence of naphthalenes is stronger in commercial MGO, compared to its pyrolysis analog. Notwithstanding this, MGO (mlpy) is compliant with ISO8217 specifications as regards marine gasoil, and the incomplete overlap between the molecular fingerprints of MGO (comm) and MGO (ml-py), as evident from Table 2, does not underscore an a priori better or worse performance of the fuel based on marine litter. Further testing of MGO (ml-py) as regards emissions and other parameters is a work in progress since pyrolyzed oil-based fuels could lead to reduced emissions [9,19].  Since Figure 3 takes into account only identified compounds, we decided to compare the chromatographic fingerprints of these samples displaying the total percent of the eluted area as a function of time, as shown in Figure 5, in order to avoid artefacts due to the lack of identification of some analytes. The overlap between the two series is impressive, even if we know that saturated hydrocarbons are more common in MGO (comm), and unsaturated compounds are more common in MGO (ml-py). Again, for the MGO (mlpy) series, the massive presence of 2,4-dimethyl-1-heptene and 1-nonene eluted at 16 and 20minutes, respectively, explains the appearance of two putative outliers. The majority of the eluted area resides between 28 min and 44 min, corresponding, to an oven temperature of 95 °C and 250 °C, respectively. Since Figure 3 takes into account only identified compounds, we decided to compare the chromatographic fingerprints of these samples displaying the total percent of the eluted area as a function of time, as shown in Figure 5, in order to avoid artefacts due to the lack of identification of some analytes. The overlap between the two series is impressive, even if we know that saturated hydrocarbons are more common in MGO (comm), and unsaturated compounds are more common in MGO (ml-py). Again, for the MGO (ml-py) series, the massive presence of 2,4-dimethyl-1-heptene and 1-nonene eluted at 16 and 20minutes, respectively, explains the appearance of two putative outliers. The majority of the eluted area resides between 28 min and 44 min, corresponding, to an oven temperature of 95 • C and 250 • C, respectively. the lack of identification of some analytes. The overlap between the two series is impressive, even if we know that saturated hydrocarbons are more common in MGO (comm), and unsaturated compounds are more common in MGO (ml-py). Again, for the MGO (mlpy) series, the massive presence of 2,4-dimethyl-1-heptene and 1-nonene eluted at 16 and 20minutes, respectively, explains the appearance of two putative outliers. The majority of the eluted area resides between 28 min and 44 min, corresponding, to an oven temperature of 95 °C and 250 °C, respectively.

FTIR Analysis
Pyrolysis oil, its distillates, and commercial MGO are complex mixtures of alkanes, alkenes, aromatic and other compounds hence their FTIR spectra are complicated. FTIR analysis in a wave range of 4000-600 cm −1 using an ATR arrangement can highlight the presence of functional groups but also provide a chemical signature of the samples [11]. The spectra are shown in Figure 6.

FTIR Analysis
Pyrolysis oil, its distillates, and commercial MGO are complex mixtures of alkanes, alkenes, aromatic and other compounds hence their FTIR spectra are complicated. FTIR analysis in a wave range of 4000-600 cm −1 using an ATR arrangement can highlight the presence of functional groups but also provide a chemical signature of the samples [11]. The spectra are shown in Figure 6. For the most part, the spectra can be classified into 3 wavenumber ranges based on the absorption bands observed ranging between 3100 cm −1 and 2800 cm −1 , 1700 cm −1 and 1300 cm −1 , and 1000 cm −1 and 600 cm −1 . They all represent typical hydrocarbon vibrations. In all samples, the absorption bands between 3000 cm −1 and 2800 cm −1 at wave number 2923-2925 cm −1 and 2854-2856 cm −1 represent the aliphatic asymmetric (υasC--H) and symmetric (υsC---H) stretching vibrations of the -CH2 group. The absorption bands at wave number 2955-2957 cm −1 and 2872cm −1 represent the aliphatic asymmetric (υasC--H) and symmetric (υsC--H) stretching vibrations of the -CH3 group. While the former is present in all samples, the latter is clearly visible only in the VN (ml-py) sample. The set of two peaks at 1457 cm −1 and 1377 cm −1 highlight the sp3 C-H asymmetric and symmetric bending [58]. For the most part, the spectra can be classified into 3 wavenumber ranges based on the absorption bands observed ranging between 3100 cm −1 and 2800 cm −1 , 1700 cm −1 and 1300 cm −1 , and 1000 cm −1 and 600 cm −1 . They all represent typical hydrocarbon vibrations. In all samples, the absorption bands between 3000 cm −1 and 2800 cm −1 at wave number 2923-2925 cm −1 and 2854-2856 cm −1 represent the aliphatic asymmetric (υasC--H) and symmetric (υsC--H) stretching vibrations of the -CH2 group. The absorption bands at wave number 2955-2957 cm −1 and 2872cm −1 represent the aliphatic asymmetric (υasC--H) and symmetric (υsC--H) stretching vibrations of the -CH3 group. While the former is present in all samples, the latter is clearly visible only in the VN (ml-py) sample. The set of two peaks at 1457 cm −1 and 1377 cm −1 highlight the sp3 C-H asymmetric and symmetric bending [58].
The band at 722 cm −1 is due to the rocking of the -CH2-groups and the intensity of this band increases with increasing carbon chain [59]; actually, the strongest band is observed in RPO (ml-py). The massive presence of olefins in the pyrolysis sample is confirmed by their FTIR spectra. In all samples from marine litter pyrolysis, a variance with MGO (comm) the small adsorption band at 3075 cm −1 indicates alkene C-H stretching. The band at 1650 cm −1 representing the alkene C-C bond stretching confirms the higher presence of olefins [60] in the pyrolysis samples, consistent with results shown in Figure 3. This band is not detectable in MGO (comm), since, as also indicated in Figure 3, in this sample, olefins are less represented. The two strong absorption bands at 990-992 cm −1 and 908-909 cm −1 represent the C-H stretching of the monosubstituted double bond, and it is missing only in MGO (comm); the peak at 887-888 cm −1 marks the presence of a vinylidene-type double bond only in all samples from marine litter pyrolysis. The peak at 697-675 cm −1 reinforced the identification of the alkene =C-H bending vibration. The presence of the aromatic group can be inferred from the presence of the C-H stretching between 3100 cm −1 and 3000 cm −1 and the small C=C stretching aromatic bands at 1495 cm −1 [57] that are absent in MGO (comm).
Overall, the spectra obtained are in good agreement with the published literature for pyrolysis oil from mixed plastic waste [29,57,58,60].

Conclusions
Recycling difficult plastic waste streams is crucial for a successful circular economy and to reduce environmental pollution. This goal can be achieved through chemical recycling using pyrolysis and distillation. Following this route, plastic marine litter collected from the sea bottom of the northern Adriatic Sea was processed "as it is" and converted into useful ISO-compliant marine gasoil and useful material for the chemical industry. Since marine litter is probably the most difficult waste to upcycle, the same approach could be applied to all kinds of rejected plastic waste and highly nonhomogeneous plastic streams, which do not find a proper recycling solution. Compliance of waste-derived MGO (ml-pl) with ISO8217 standards is demonstrated, and its chemical composition is finally revealed, which is different from the one of conventional MGO (comm). In particular, the presence of massive amounts of 2,4-dimethyl-1-heptene can be considered the chemical signature of plastic-derived fuels and might be used as tracer for its presence for commercialization and blending. A significant amount of valuable aromatics (BTX) then alkenes, cycloalkenes, and polyenes suggest that VN (ml-py) might be an interesting source of these precious building blocks for the chemical industry and even be used as alternative feedstock for the production of virgin polymers using state-of-the-art technologies. A future outlook will be the use of distillates derived from marine litter in steam cracking equipment for the manifacturing of ethilene and propylene to be polymerized into virgin PE and PP.