1. Introduction
Global plastic production has reached 413.8 million metric tons in 2024 [
1] and is forecasted to surpass 1.2 billion metric tons by 2050 [
2,
3]. Polyolefins, mainly polyethylene (PE) and polypropylene (PP), dominate single-use packaging and account for approximately 55% of global plastic waste [
2]. The chemical inertness that makes these materials desirable for commercial applications also renders them persistent in the environment [
3].
Current recycling technologies have proven to be insufficient to divert plastic waste from landfills or environmental loss due to technical limitations and poor economic incentives. Conventional mechanical recycling is a form of downcycling, in which the polymer remelting process results in products with inferior material properties and reduced market value. Pyrolysis, a thermochemical route used for the chemical recycling of plastic wastes into refinery feedstock, requires severe conditions, typically 400-600 °C in an oxygen-free environment. The high temperature drives up energy costs and produces a wide range of poorly defined products, including undesirable light gases and heavy tars that pose challenges for downstream processing [
4].
As an alternative, catalytic chemical recycling offers a promising route for converting waste polyolefins back into value-added hydrocarbons—such as fuels, lubricant base oils, and other chemical feedstocks—at temperatures typically below 300 °C. This opinion article focuses on two dominant pathways in heterogeneous catalysis that are used to cleave the C-C bonds of polyolefins: hydrogenolysis and hydrocracking [
5-
15]. Both catalytic pathways have the potential to selectively convert polyolefin waste into valuable fuels and chemical feedstocks under mild conditions [
2,
3,
5,
6]. Notably, while hydrocracking primarily yields branched hydrocarbon products, hydrogenolysis predominantly produces linear hydrocarbons. Here, we compare the mechanisms and catalyst designs for hydrocracking and hydrogenolysis, analyze critical technical challenges from catalyst stability to process engineering, and provide an outlook on how these complementary pathways can be used to repurpose plastic waste into valuable products.
2. Hydrocracking versus hydrogenolysis: A comparative overview
Hydrocracking and hydrogenolysis proceed through distinct reaction pathways and invoke different elementary steps and intermediates in C-C bond cleavage; these mechanistic differences govern the tailoring of the catalyst (metal/acid pairing versus metal-only site engineering) and predetermine the types of product slates delivered by different strategies.
2.1. Mechanism of hydrocracking
The hydrocracking of polyolefins follows a bifunctional, tandem sequence in which metal sites perform dehydrogenation and hydrogenation, while nearby Brønsted-acid sites (BASs) promote carbocation chemistry—that is, skeletal rearrangement, β-scission, and double bond migration (
Fig. 1(a)) [
5,
6,
14].
(1) Polymer dehydrogenation on metal sites. Noble- or base-metal sites (e.g., Pt, Ni, Ru, and MoSx) abstract two H atoms from a molten polymer segment to generate an adsorbed alkene/alkyl intermediate.
(2) Carbenium formation on BAS. After being released from the metal, the alkene diffuses to a BAS and is protonated to form a secondary or tertiary carbenium ion.
(3) Skeletal rearrangement. The carbenium undergoes rapid 1,2-hydride and 1,2-alkyl shifts, isomerizing the chain to create tertiary or quaternary carbon branch points and migrating the positive charge along the chain.
(4) β-scission. Once a suitably branched carbenium forms, β-scission occurs, producing two shorter fragments: a smaller carbenium ion that remains associated with the conjugate base of the acid site and an olefin.
(5) Proton transfer from the carbenium. A proton is transferred from the carbenium fragment to regenerate the BAS and form an olefin.
(6) Metal-catalyzed hydrogenation. The two fragments diffuse to nearby metal sites, re-hydrogenate, and desorb as fully saturated paraffins, regenerating the surface metal hydrides and closing the catalytic cycle via the dissociative absorption of H2 and subsequent product desorption.
This tandem metal-acid catalyzed mechanism, though effective, sets up a series of tightly coupled kinetic and stability challenges:
•Dehydrogenation on the metal can be kinetically sluggish and is easily inhibited by excess surface hydrogen [
14].
•The alkene intermediate must diffuse from the metal site through a highly viscous polymer melt and reach an acid site to avoid being re-hydrogenated [
14].
•The acid sites must be strong enough for rapid β-scission but not so strong or dense that they cause over-cracking and coke formation [
7].
•Large metal ensembles could cleave the terminal C-C bond and generate low-value methane (CH
4) [
7].
•Any imbalance or physical separation between the metal and acid sites can erode the activity and selectivity, shortening the catalyst life [
16].
2.2. Hydrogenolysis of polyolefins: A monofunctional metal-site pathway
The hydrogenolysis of polyolefins is a structure-sensitive, metal-only pathway, in which C-C bonds are cleaved on metal sites without the need for BASs (
Fig. 1(b)).
(1) H2 activation on the metal surface. In what is typically considered a reversible quasi-equilibrated step, H2 adsorbs to the metal surface and dissociates to form adsorbed surface H* (M-H), creating a reactive hydrogen pool.
(2) Initial C-H activation. A molten polymer segment physisorbs to the metal through multiple C-H surface interactions; the physisorbed chain then undergoes quasi-equilibrated dehydrogenation via C-H bond activation. The number of H atoms removed from a chain depends on the metal center and the hydrocarbon structure and can be greater than four.
(3) C-C bond scission. The dehydrogenated hydrocarbon chain restructures on the metal surface, with the former C-C bond rearranging into two new M-C bonds.
(4) Hydrogenation and desorption of the fragments. The two surface hydrocarbyl fragments produced from the C-C bond scission step are rapidly hydrogenated by nearby M-H species, yielding saturated alkanes that desorb from the metal surface. Desorption leaves vacant metal sites for the next catalytic cycle.
Despite being conceptually cleaner than the hydrocracking described earlier, this metal-site-only hydrogenolysis pathway imposes its own tightly interlocked constraints:
•Surface H atom coverage must remain high enough to hydrogenate and desorb scission fragments, yet low enough to keep the first C-H activation kinetically accessible [
9,
17].
•Molten polyolefin chains blanket the surface, yielding zero-order kinetics that can obscure the other elementary steps and throttle H
2 activation [
18,
19].
•Extended chain adsorption on metal surfaces could favor terminal cleavage and facilitate cascade hydrogenolysis of end-group bonded chains, leading to undesirable CH
4 formation [
15,
20].
•Hydrogen-shuttling supports (CeO
2 and TiO
2) can buffer hydrogen coverage only if the spill-over outpaces the chain re-adsorption [
9,
17].
•A second pathway for C-C bond cleavage, involving β-alkyl-elimination and typically observed with isolated sites, leads to distinct selectivity from the metal-surface-catalyzed cleavage. Catalysts with mixtures of single-atom sites and small nanoparticles could follow both pathways [
21,
22].
Hydrocracking and hydrogenolysis operate via fundamentally distinct principles, encompassing divergent chemistries, catalyst architectures, product profiles, and control mechanisms. These foundational differences, summarized in
Table 1, necessitate the distinct catalyst design and performance improvement strategies explored in the subsequent sections.
3. Catalytic performance improvement
While both polyolefin hydrocracking and hydrogenolysis have advanced rapidly, they still present challenges that must be addressed for industrial practice. Catalyst design in this field is anchored in active-site engineering and structure-activity relationships, but the governing levers differ by pathway: In bifunctional hydrocracking, performance hinges on an optimized synergy between metal hydrogenation/dehydrogenation and Brønsted-acid β-scission/isomerization; in monofunctional hydrogenolysis, on the other hand, it is dictated by the precise control of metal-site geometry and electronic state. Building on these site-level principles, we now consider how they translate into measurable gains across the full catalysis system.
Below, we outline distinct strategies for performance improvement in hydrocracking and hydrogenolysis that extend beyond catalyst design to the entire catalysis system. The following sections illustrate these principles through advancements in product selectivity (Section 3.1) and catalyst stability (Section 3.2) at the catalyst level, as well as process efficiency (Section 3.3) and analytical rigor (Section 3.4) at the broader system level.
3.1. Controlling product selectivity: Achieving narrow distributions
Controlling selectivity to produce a narrow range of valuable products while suppressing the formation of low-value light gases is a key challenge. Strategies to achieve this control are distinct for each pathway.
3.1.1. Product selectivity in hydrocracking
In hydrocracking, selectivity control is dictated by the nanoscale proximity between hydrogenation metal sites and BASs for β-scission. A simple physical blend such as Pt/WO
3-ZrO
2 with HY zeolite [
5] yields approximately 85 weight percent (wt%) liquids and low C
1-C
4 at 225-250 °C. Examples of more advanced controls on metal and acid sites in hydrocracking catalysts are briefly discussed below.
(1)
Nanoscale metal-acid integration. Advanced catalysts go further by fusing the two functions on the nanoscale: Platinum (Pt) atoms or nanoparticles embedded directly inside the zeolite channel convert low-density polyethylene (LDPE) almost exclusively to a C
5-C
9 naphtha cut (> 96% selectivity) [
23], whereas the trace-cerium (Ce) doping of Pt/HY fine-tunes local acidity around the metal clusters and pushes liquid formation rates above 3000 g per gram of noble metal per hour without broadening the distribution (
Fig. 2(a)) [
24].
(2)
2D metal-acid architectures. Atom-level Pt anchored on two-dimensional tungsten trioxide (WO
3) nanosheets enables high liquid yields (C
5-C
18) at 200-250 °C while maintaining a tight product chain length distribution, because every surface WO
3 site is within a few nanometers of a Pt site for hydrogen activation (
Fig. 2(b)) [
14].
(3)
Hierarchical/mesoporous zeolites. Creating mesopores in H-USY (USY: ultra-stable Y), Beta, or ZSM-5 zeolites shortens the residence time of reactive olefins, suppresses secondary over-cracking to gases, and can increase liquid selectivity by 10%-20% compared with purely microporous analogues (
Fig. 2(c)) [
25,
26].
(4)
Earth-abundant analogues. Noble-metal-free MoS
x-H-Beta was used to achieve 80-95 wt% liquids (C
5-C
20) from mixed waste plastics; judicious sulfide-acid proximity compensates for the lower intrinsic hydrogenation rate of molybdenum (Mo) while avoiding the formation of excessive C
1-C
4 [
6].
3.1.2. Product selectivity in hydrogenolysis
Metal- only hydrogenolysis relies on the isolation and electronic property adjustment of metal ensembles so that C-C cleavage occurs preferentially at internal C-C bonds to minimize methane formation. Recent structure-function innovations are briefly discussed below.
(1)
Confinement-induced "processive" scission. Encapsulating Pt nanoparticles at the interface of a mesoporous silica (SiO
2) shell and a solid spherical silica core (mSiO
2/Pt/SiO
2) ensures that the polymer chain threads into the mesopores and cleaves at Pt surface sites located at the end of the pore. The polymer chain remains adsorbed in the mesopore to undergo many C-C cleavage events, which converts the whole PE chain into alkanes in a manner analogous to natural processive enzymes. The preferential adsorption of long-chain polymer over short-chain alkane products minimizes further hydrogenolysis of the products to light gases, especially CH
4 [
27]. Switching the solid silica core to a mesoporous core (MCM-48) increases the reaction rate, narrows product distribution, and further reduces CH
4 formation (
Fig. 3(a)) [
28].
(2)
Confinement on earth-abundant oxides. Through the pore-guided delivery of polymer chains to interfacial ZrO
x(OH)
4-2x sites, ultra-small amorphous zirconium dioxide (ZrO
2) nanoparticles confined between fused mesoporous silica platelets (L-ZrO
2@mSiO
2) enable selective C-C scission to a narrow C
16-C
18 centered range of alkane products [
29].
(3)
Electronic and ensemble tuning. Surface-ligand engineering that stabilizes Ru
δ+ (i.e., ligand-modified Ru/C [
30] or heteroatom modulation that anchors single Fe atoms and FeO
x nanoclusters adjacent to Ru on CeO
2 [
31]) biases scission toward internal C-C bonds (
Fig. 3(b)), boosting liquid yields while suppressing light gases.
(4)
Single-site confinement inside metal-organic frameworks (MOFs). Ru-dihydride sites anchored inside UiO-type MOFs yielded a size selectivity of > 80 wt% C
8-C
16 liquids at only 200 °C (2.0 MPa H
2 pressure) [
22].
3.2. Increasing catalyst stability and impurity tolerance
Beyond high selectivity, practical deployment demands catalysts with durable stability and impurity tolerance. Real polyolefin waste feeds often contain antioxidants, fillers, halogens, sulfur species, and pigments that can poison Brønsted acids or metal nanoparticles. A recent systematic study on organic stabilizers (i.e., antioxidants, amines, and lubricants) showed that even 1-2 wt% of a single nitrogen (N)-containing additive can reduce hydrocracking conversion by 40%-90% across multiple nickel (Ni)-based catalysts [
32]. For halogen-rich streams, a Mg
3AlO
4.5 trap was used to remove hydrochloric acid (HCl) generated during the reaction, which enabled Pt/WO
3-ZrO
2 to maintain approximately an 80 wt% liquid yield even in the presence of polyvinyl chloride (PVC) [
33]. Impurity-tolerant MoS
x-H-Beta operates on mixed, additive-rich plastics without noble metals [
12].
Hydrocracking also leads to rapid coke deposition. "Softening" the acid function with high-silicon/aluminum (Si/Al) or mesoporous Si-Al oxides diminishes poisoning and coke formation [
32]. The Ce-promotion of Ni/Beta both stores spill-over hydrogen and anchors Ni, cutting coke formation by > 60% and raising LDPE conversion above 85% [
34]. The addition of water modulates the metal-acid balance (MAB) of Ru/zeolite-Y catalysts, which has been shown to significantly boost liquid fuel yields and suppress coke formation, thereby increasing catalyst stability [
2].
In hydrogenolysis, it is crucial to stabilize the metal phase and control the metal-support interface. Strong metal-support interactions stabilize nanoscale Ru on cerium(IV) oxide (CeO
2) and sustain its activity [
35]. Ammonia-treated Ru/titanium dioxide (TiO
2) triples the activity while mitigating sintering [
17]. Ru/TiO
2 enables solvent-free photothermal operation with light-driven, localized heating and performance under intermittent illumination, offering a green energy route for additive-containing polymer hydrogenolysis [
36].
3.3. Overcoming engineering hurdles: Reactors and scaleup
While advancements in catalyst stability are essential, translating these material-level gains into a scalable technology requires addressing the broader catalysis system—particularly the reactor engineering hurdles imposed by the feedstock itself. Molten PE and PP with high viscosity (10
2-10
3 Pa·s) impose severe mass-/heat-transfer issues that have been sidestepped by most batch autoclave studies. Apart from short (< 24 h) batch trials, most studies offer data on hydrogen pressure drop, fouling, or catalyst attrition in melted polymer media. A recent computational fluid dynamics (CFD) and experimental study (
Fig. 4) [
37] explored the hydrogenolysis of commercial-grade HDPE and PP in a stirred autoclave and demonstrated how hydrodynamics dominate the catalysis process. Without credible data on the pressure drop, fouling, attrition, and temperature control, scaling projections remain speculative.
Recent studies in flow processes demonstrate two distinct strategies: One employs a tandem fixed-bed reactor for the continuous reforming of volatile products from hydrocracking into high-yield C
2-C
4 light olefins and benzene, toluene, and xylene (BTX) aromatics [
38], while the other utilizes a semi-batch reactor with a "reactive separation" design to suppress over-hydrogenolysis by continuously removing the volatile liquid products (mainly C
9-C
20 alkanes and olefins) from the hydrogenolysis reaction [
39]. These flow processes shed lights on future continuous chemical recycling of waste plastics.
3.4. Analytical blind spots: The overlooked bottleneck
Just as scalable reactor design is essential for process viability, rigorous analytical protocols are foundational for understanding reaction mechanisms and comparing catalytic performance; yet, the latter remain an overlooked bottleneck. Across the reported studies, further development in analytical practice is still needed. It remains difficult to achieve carbon balance in polyolefin hydrogenolysis and hydrocracking, and quantification protocols for products differ among studies. Especially in vapor-rich systems, incomplete capture of volatile products can result in carbon balance deficits that occasionally surpass 50% [
40]. Continuous sweep-gas collection has proven effective, with carbon balance closures reported as 96.0% ± 9.2% (statistical range from Ref. [
40]) (
Fig. 5). Because quantification demands depend on the product slate, there is no simple protocol best suited for all cases; complementary methods must be carefully selected and optimized accordingly to account for all products.
Robust analytical protocols are also needed to untangle the respective contributions of hydrocracking and hydrogenolysis when both pathways can run concurrently on bifunctional catalysts. Multiple studies have demonstrated parallel pathways on Ru and Pt metal centers, such as Ru/MFI (25% CH
4 with isomer-rich C
5-C
20 products) [
41], as well as Ru/faujasite (FAU) and Ru/Beta (10% CH
4 with isomer-rich C
7-C
32 products) [
7]. Because the metal-ensemble size, oxidation state, and acid strength affect both the hydrocracking and hydrogenolysis pathways, tuning one often perturbs the other.
4. Conclusions and outlook
Hydrocracking and hydrogenolysis have reached a stage where they can be deliberately matched to different product targets rather than viewed as rival technologies. Hydrocracking delivers branched alkanes, whereas hydrogenolysis provides a cleaner path to linear n-alkanes. With proof-of-concept activity and selectivity having now been repeatedly demonstrated in laboratories, the grand challenge is no longer whether these chemistries work for polyolefins but how to translate them into robust, selective, continuous technologies that can digest real post-consumer polyolefin waste streams at scale. We define five interlocking research priorities in this field:
(1) Impurity-robust, low-noble-metal catalysts. Earth-abundant metal oxides, Ni/Mo carbides, nitrides, or diluted noble-metal alloys that tolerate chlorine (Cl), sulfur (S), and metal contaminants must be developed, and in situ H2/H2O or mild oxidative regeneration cycles must be engineered to eliminate costly shutdowns.
(2) Continuous operation with real feeds. There is a need to validate weeks to months of on-stream stability in flow reactors that can accommodate viscous mixed polymer melts without fouling, attrition, or runaway temperature gradients, while closing full carbon balances.
(3) Hydrogen integration and pressure minimization. Low-pressure active (≤ 1 MPa) catalysts such as dilute Ru-Pt alloys or water-promoted Ru/acid systems should be coupled with green or in situ H2 generation (aqueous phase reforming) to reduce capital costs and loss of hydrocarbon.
(4) Unified, high-resolution analytics. It is necessary to standardize the full quantification of gas, liquid, and residual solids, deploy advanced analytical techniques (e.g., two-dimensional gas chromatography with flame-ionisation and mass spectroscopic detection (GC × GC-FID/MS)) for isomer maps and time-resolved online data to discriminate primary from secondary C-C cleavage pathways, and adopt common data-reporting templates for seamless cross-lab benchmarking.
(5) Data-driven catalyst and process design. Operando spectroscopy, population-balance or micro-kinetic modeling, and machine-learning optimization should be integrated to forge quantitative links between active-site structure and product spectra, accelerating the "design-performance" loop.
If these priorities are met, hydrocracking could reliably provide branched alkanes (e.g., high-octane fuels), while hydrogenolysis could feed linear-alkane supply chains, turning plastic waste from an environmental liability into a versatile carbon resource. The convergence of atom-precise catalysis, resilient continuous reactors, and rigorous analytical protocols could give polyolefin waste a second life.
PDF
(2297KB)
