(79a) Plastic Waste Contaminants: A Deal-Breaker for Steam Cracking?

Pyrolysis of plastic packaging waste yields a liquid product that can be processed in steam crackers producing light olefins and hence closing the loop towards new virgin plastics. However, the huge variety of plastic waste in terms of polymer composition and contamination with residues and additives makes a prediction of the pyrolysis oil quality difficult. In fact, pyrolysis oils from real post-consumer waste fractions contain more and different contaminants compared to fossil-based feedstocks. The associated uncertainty explains why thermochemical recycling of contaminated plastic waste is not yet industrially established. Metals (i.e. aluminum, silicon, zinc and others) and heteroatoms (i.e. chlorine, nitrogen and oxygen) will require special attention. In this work, a comprehensive experimental assessment of plastic waste pyrolysis oil quality was performed. Based on the pyrolysis oil compositions, potential decontamination techniques are discussed and evaluated.

Post-consumer plastic waste fractions, namely mixed polyolefins (MPO), polyethylene (PE), and polypropylene (PP) were processed in a pilot-scale pyrolysis unit and the pyrolysis oils were characterized using advanced analytical techniques. Substantial amounts of olefins were detected in all pyrolysis oils. In terms of steam cracking, olefins are typically less wanted due to a substantial impact on coke formation and fouling of heat exchanger surfaces [1, 2]. Considering an olefin tolerance in steam cracking of 2 wt.% [3], the olefin concentrations in the respective pyrolysis oils need to be substantially reduced either by high dilution factors with (olefin-free) fossil feedstocks or with hydrotreatment as an established technology [4, 5]. The total olefin concentration including linear and branched olefins as well as diolefins of PP pyrolysis oil is ~88 %. PE pyrolysis oil contains a total of ~44 % olefins indicating that PE pyrolysis oil requires a lower upgrading effort and mixing ratio compared to PP pyrolysis oil from an olefin point of view.

The most crucial heteroatoms for plastic waste pyrolysis oils are nitrogen, oxygen and chlorine. Nitrogen in plastic waste stems from nitrogen containing polymers (i.e. PUR, PA) or from N-containing compounds which were in contact with the plastic packaging material (i.e. amino acids or N-containing detergents). The maximum allowable concentrations of nitrogen in industrial steam crackers are 100 ppmw for light feedstocks and 2000 ppmw for heavy feedstocks [3] due to issues such as explosive gum formation as well as the formation of NO_x species. All studied pyrolysis oils contained nitrogen. However, PP (~29 ppm) and PE (~40 ppm) pyrolysis oils comply with both thresholds while the MPO pyrolysis oil (~1144 ppm) only complies with the nitrogen threshold for heavy feedstocks. An established upgrading technology for nitrogen is hydrodenitrogenation [4].

Oxygen in plastic waste pyrolysis oils stems from additives, organic residues or unwanted plastic fractions (i.e. PET) [6]. Elemental oxygen as well as oxygen containing compounds in steam crackers lead to fouling and corrosion issues [7]. Oxygen in form of CO can also be a poison to a number of downstream catalysts. Thresholds for oxygen in industrial steam crackers are typically around 100 ppm [3]. This value is exceeded substantially by PE pyrolysis oil (~2100 ppm) while PP and MPO pyrolysis oil contained no detectable oxygen. To fully remove oxygen, hydrodeoxygenation has been an active field in the research area of renewable transportation fuels [5]. Studies show that full removal of oxygen and nitrogen from liquid hydrocarbon samples is possible [5, 8].

Halogens and especially chlorine belong to the most problematic contaminants in steam cracking due to the formation of corrosive compounds. Chlorine mostly stems from PVC impurities in the sorted waste fraction. The maximum allowable concentration of Cl in industrial steam cracker feedstocks has been reported as 3 ppmw [3] and is exceeded substantially by all three pyrolysis oils (PP ~137 ppm; MPO ~474 ppm; PE ~143 ppm). Hydrotreatment is a viable technique to reduce the chlorine to sufficiently low concentrations. Furthermore, in an industrial context, Ca salts can be added in the pyrolysis process in order to bind HCl [10]. Thermal dehalogenation is another option to remove halogens from the plastic waste feedstock prior to pyrolysis [11]. However, detailed analyses of individual halogenated compounds remains a challenge. First results of GC × GC-AED analysis as the most promising analytical technique for halogen detection and quantification will be presented.

All pyrolysis oils contained substantial amounts of metals. Metals can be transferred from the feedstock into the liquid product in the form of organometallic complexes in vapor form. However, entrainment of solid pyrolysis residue is another possibility. Filtration of the crude pyrolysis oil or distillation would remove entrained particles leaving only the chemically bound metals in the pyrolysis product [12, 13]. Strict specifications exist for industrial steam cracker feedstocks in terms of metals. These are for instance: 0.5 ppm for calcium, 50 ppm for copper, 0.001 ppm for iron; 0.005 ppm for mercury; 0.125 ppm for sodium; 100 ppm for nickel; 0.1 ppm for lead and 1 ppm for silicon. In all studied pyrolysis oils, thresholds are exceeded substantially in terms of the Fe, Na and Pb concentrations. MPO pyrolysis oil also exceeds the threshold for Ca and Hg. The absence of an industrial limit value for certain metals such as Al, Sb or Zn does not imply that they are not harmful in steam cracking but that there is a lack of knowledge about the impact of these metals. This is mostly due to the fact that some metals do not occur in fossil feedstocks. The presented metal contaminant results imply that thorough upgrading or severe dilution with metal-free fossil feedstocks is needed prior to steam cracking. Catalytic demetallization and advanced filtration are promising techniques to further improve the quality of the pyrolysis oils towards steam cracking [14, 15].

Quick removal of residual metals from plastic waste pyrolysis oils is highly important to protect hydrotreatment catalysts from deactivation due to metal deposition on the active surface. In catalytic hydrotreatment of metal complexes, metals are released and immediately attach on the catalyst surface, permanently poisoning the catalyst. Therefore, mostly cheap aluminum oxide catalysts are used and installed prior to more precious catalyst beds [14, 16]. Guard bed catalysts are for instance nickel, molybdenum or cobalt oxides on Al₂O₃. It has been reported that a total initial V, Ni and Fe concentration of 80 ppm in a petroleum feed was reduced to a total amount of 8 ppm of V, Ni and Fe [14]. Although a significant reduction, other methods of metal removal need to be applied to meet the threshold values for industrial steam crackers. However, it can be anticipated that the fractions which are interesting for steam cracking, i.e. the naphtha and middle distillate fractions are relatively free of metalloorganic compounds since metalloporphyrins are mostly found in higher carbon number ranges [14].

It can be concluded that the produced pyrolysis oils require thorough treatment which will be a challenging factor for the economic competitiveness compared to fossil-based fuels. Potentially, the pyrolysis oil quality could be improved by using innovative concepts such as supercritical water decomposition, catalytic pyrolysis or contaminant adsorption using metal-organic frameworks.

[1] Kopinke, F. D. et al., Ind Eng Chem Res, vol. 32, no. 1, pp. 56-61, 1993, https://doi.org/10.1021/ie00013a009.

[2] Müller-Steinhagen, H. et al., Heat Transf. Eng., vol. 32, no. 3-4, pp. 189-196, 2011, https://doi.org/10.1080/01457632.2010.503108.

[3] Baumgartner, A. J. et al., in AIChE Spring National Meeting, Ethylene Producers Conference, New Orleans, Louisiana, 2004.

[4] Prado, G. H. C. et al., Energy Fuels, vol. 31, no. 1, pp. 14-36, 2017, https://doi.org/10.1021/acs.energyfuels.6b02779.

[5] Zacher, A. H. et al., Green Chem., 10.1039/C3GC41382A vol. 16, no. 2, pp. 491-515, 2014, https://doi.org/10.1039/C3GC41382A.

[6] Anuar Sharuddin, S. D. et al., Energy Convers. Manage., vol. 148, pp. 925-934, 2017, https://doi.org/10.1016/j.enconman.2017.06.046.

[7] Ristic, N. D. et al., Fuel Process. Technol., vol. 167, pp. 241-249, 2017, https://doi.org/10.1016/j.fuproc.2017.07.008.

[8] Murena, F. and Gioia, F., J. Hazard. Mater., vol. 60, no. 3, pp. 271-285, 1998, https://doi.org/10.1016/S0304-3894(98)00151-4.

[9] Miller, S. J. et al., in Feedstock Recycling and Pyrolysis of Waste Plastics, J. Scheirs and W. Kaminsky, Eds.: John Wiley & Sons, Ltd, 2006, pp. 345-362.

[10] Okuwaki, A. et al., in Feedstock Recycling and Pyrolysis of Waste Plastics, J. Scheirs and W. Kaminsky, Eds., 2006, pp. 663-708.

[11] Bockhorn, H. et al., J. Anal. Appl. Pyrolysis, vol. 49, no. 1, pp. 97-106, 1999, https://doi.org/10.1016/S0165-2370(98)00124-7.

[12] Heydariaraghi, M. et al., J. Anal. Appl. Pyrolysis, vol. 121, pp. 307-317, 2016, https://doi.org/10.1016/j.jaap.2016.08.010.

[13] Wiriyaumpaiwong, S. and Jamradloedluk, J., Energy Procedia, vol. 138, pp. 111-115, 2017, https://doi.org/10.1016/j.egypro.2017.10.071.

[14] Ali, M. F. and Abbas, S., Fuel Process. Technol., vol. 87, no. 7, pp. 573-584, 2006, https://doi.org/10.1016/j.fuproc.2006.03.001.

[15] Wilson, G. R., US Patent US3898155A, 1973.

[16] van Dongen, R. H. et al., Ind Eng Chem Process Des Dev, vol. 19, no. 4, pp. 630-635, 1980, https://doi.org/10.1021/i260076a021.