Catalytic Depolymerization of Waste Polyolefins by Induction Heating: Selective Alkane/Alkene Production

Abstract

Low- and high-density polyethylene (LDPE/HDPE) have been selectively depolymerized, without added H₂, to C2-C20 + alkanes/alkenes via energy-efficient radio frequency induction heating, coupled with dual-functional heterogeneous Fe₃O₄ and Ni- or Pt-based catalysts. Fe₃O₄ was used to locally generate heat when exposed to magnetic fields. Initial results indicate that zeolite-based Ni catalysts are more selective to light olefins, while Ni supported on ceria catalysts are more selective to C7-C14 alkanes/alkenes. LDPE conversions up to 94% were obtained with minimal aromatic, coke, or methane formation which are typically observed with thermal heating. Two depolymerization mechanisms, a reverse Cossee-Arlman mechanism or a random cleavage process, were proposed to account for the different selectivities. The depolymerization process was also tested on commercial LDPE (grocery bags), polystyrene, and virgin HDPE using the Ni on Fe₃O₄ catalyst, with the LDPE resulting in similar product conversion (∼48%) and selectivity as for virgin LDPE.

Figure 1

Weight loss and rate variation curves: (a) HDPE wt loss curves over modified ZSM-5 catalysts heated to 350 °C as a function of time, (b) temporal rate variation in TGA/DSC analyses for the zeolites catalysts, and (c) temporal rate variation in TGA/DSC analyses for the metal oxide catalysts.

Figure 2

PL response of the Fe₃O₄/YVO₄/Eu³⁺ mixture under applied rf fields. The inset highlights the linear response of the normalized intensity at high applied fields (200–400 °C).

Figure 3

rf (64 mT field)-initiated LDPE depolymerization for various zeolite (left) and non-zeolite (right) catalysts. The different colors are just an aid to the eye. The “Fe” in all but Fe–Ni denotes that 50 wt % of the catalyst is Fe₃O₄ nanoparticles. For Fe–Ni, there are 97.6 wt % Fe₃O₄ nanoparticles; 115 mg of total catalyst.

Figure 4

Coke and acid site analysis. TPO weight derivatives for used, extracted catalysts after LDPE depolymerization for (a) catalysts containing CeO₂, (b) Fe₃O₄ and Ni-supported on Fe₃O₄, and (c) zeolite catalysts. The presence of coke is seen from the peaks at above 420 °C. (d) Differential thermal analysis of amine desorption of the three ZSM-5-based catalysts. Peaks A and B arise from the desorption of weakly adsorbed 1-PA not on Brønsted sites, peak C from the Hofmann elimination of 1-PA to propene and NH₃ on Brønsted sites, and peak D from dehydrogenation chemistry on strong Lewis sites, normally associated with extra-framework Al³⁺. The Si/Al molar ratio obtained by magic angle spinning-NMR spectroscopy for H-ZSM-5 is 20, while the ratio computed from these data is 21. The small “C” peak for Ni–ZSM-5 corresponds to <10% residual H⁺.

Figure 5

rf-initiated commercial LDPE reaction. Product distribution of commercial LDPE (WP-LDPE) and polystyrene (WP-PS) over Fe–Ni catalysts and virgin HDPE over the Fe catalyst exposed to 64 mT rf field for 2 h.