Engineering insights into thermal plasma processing for plastic waste management: a review

2.1 Plasma pyrolysis

During the pyrolysis of plastic, the material undergoes a series of thermal and chemical degradation steps, breaking down into simpler chemical species such as H₂, C, or their compounds, depending on the operating conditions. Typically, homolytic cleavage initiates the process, forming free radicals that propagate further degradation. In halogenated plastics like PVC, heterolytic cleavage occurs, producing HCl via dehydrochlorination (Zhang et al. 2022). These radicals undergo secondary reactions such as β-scission, hydrogen abstraction, and radical recombination, with product distribution influenced by temperature, heating rate, and residence time (Gracida-Alvarez et al. 2018). The pyrolysis reaction mechanism, using polypropylene as an illustrative example, is described in Equations (1)–(6).

Pyrolysis is classified into slow, fast, and flash processes based on the heating rate and residence time, with each focusing on producing either solid, liquid or gaseous products. Pyrolysis often produces complex product mixtures with limited value, typically used as furnace oil or boiler fuel. Catalytic pyrolysis using zeolites improves product selectivity by controlling residence time and secondary reactions, but mixed plastic feedstocks pose challenges like varied decomposition kinetics, catalyst coking, and PAH formation (Qureshi et al. 2020). In contrast, plasma pyrolysis better handles feedstock inconsistency, offering improved process stability.

Thermal plasma pyrolysis process utilizes plasma-generated heat to break down plastic waste into gas and solid products. This occurs at high temperatures, typically between 700 and 900 °C. A typical plasma pyrolysis setup, as illustrated in Figure 2, involves introducing shredded plastic waste into the reactor via a hopper or screw feeder. Once inside, the waste particles are subjected to rapid heating rates, around 10⁶ K/s, powered by an AC or a DC arc plasma torch (Tang et al. 2003). The heated particles vaporize and exit into a cyclone, separating entrained solids, followed by gas cleaning in an alkali scrubber and bag filter. The resulting gas is either purified or released through a flame tower chimney. Plasma pyrolysis operates at lower temperatures compared to other plasma processes, providing the possibility to produce gas, liquid, and solid products. Various research studies highlight the process’s potential are summarized in Table 2, which can broadly be categorized into three primary focuses: syngas and energy recovery, monomer and intermediate chemical recovery, and the production of solid carbonaceous materials.

Figure 2:

Schematic of plasma pyrolysis process. Reproduced with permission from ref. (Punčochář et al. 2012). Copyright 2012, Elsevier.

Tang et al. (2003), investigated the plasma pyrolysis of polypropylene (PP) in a DC arc N₂ plasma reactor and reported consistently higher conversion efficiencies compared to thermal pyrolysis and catalytic pyrolysis across all feed rates. Punčochář et al. (2012), conducted a techno-economic assessment of a 1 TPD plastic waste plasma pyrolysis system, estimating energy recovery of 2.4 kW_e/kg of plastic converted. Bhatt et al. (2024) explored the pyrolysis of PET, PP, LDPE, and HDPE using a graphite electrode arc plasma. They observed that HDPE yielded the highest cold gas efficiency (66.12 %) due to greater H₂ and C₂H₂ production at 1,000 °C, while PET generated more CO (42 mol %), leading to a lower H₂/CO ratio. Further, Zhang et al. (2018), reported similar findings for PET pyrolysis in DC arc H₂ plasma. Several studies have demonstrated that plasma pyrolysis can selectively recover olefins and monomers under specific conditions. Guddeti et al. (2000a), used an RF-induced Ar plasma to depolymerize PE into a gas mixture rich in C₂–C₄ olefins, primarily ethylene and propylene. The forming gas flow rate inversely influenced propylene content by altering plasma temperature. In follow-up experiments, it was observed that higher plate power expanded the plasma jet volume, increasing residence time and thus enhancing propylene yield (Guddeti et al. 2000b). Zhang et al. (2017), applied a rotating DC arc H₂ plasma for pyrolysis of PE and PP, achieving peak plastic-to-gas conversion and acetylene selectivity of 95 % and 77.43 %, respectively. However, both declined beyond optimal input power due to excessive dilution of plasma energy. Plasma pyrolysis also allows for the production of valuable solid carbonaceous materials. Guo and Kim (2010), designed a graphite-lined double-chambered reactor to produce ultrafine carbon black from PS and HDPE. Carbon black yield increased with temperature (500–900 °C), and BET surface areas of 142.5 and 141.3 m²/g were obtained for PS and HDPE, respectively. Fazekas et al. (2016) used RF-induced Ar/H₂/O₂ plasma to treat PVC, producing approximately 40 wt% soot and a gas rich in CO, CO₂, HCl, and C₂H₂. Toluene-extracted soot contained chlorinated PAHs, and higher O₂ content suppressed PAHs while increasing polychlorinated derivatives.

These studies underline the versatility of plasma pyrolysis in tailoring product outputs by controlling process parameters and feedstock. Though promising for single polymers, mixed plastic waste remains challenging due to poor selectivity and purity. Future work should optimize reactor design and plasma conditions to enhance scalability.

2.2 Plasma gasification

Gasification is the process of converting hydrocarbonaceous liquid or solid materials into gaseous products through sub-stoichiometric oxidation at high temperatures ranging from 900 to 1,200 °C. This technology is increasingly recognized as a sustainable method for managing solid waste (Al-Ghouti et al. 2021; Mukherjee et al. 2020). The limited availability of oxygen in the process prevents the formation of harmful oxygen-derived pollutants such as NO _x and SO _x . Additionally, high temperatures, along with a reducing environment, help suppress the formation and promote decomposition of hazardous compounds like dioxins and furans (Safavi et al. 2021; Tang et al. 2024). On comparison, gasification of waste produces 2 × e⁻¹¹ g/m³ dioxins and incineration produces 1.8 × e⁻⁷ g/m³ (Yepes Maya et al. 2016). Commonly, air, pure O₂, or steam is used as the oxidizing agent, either alone or in specific ratios. In some cases, CO₂ is also employed as an oxygen source. Multiple endothermic and exothermic reactions, as shown in Equations (7)–(16), occur simultaneously within the gasifier.

Initiated by drying and propagated by pyrolysis, these reactions continue through gasification stages wherein pyrolysis products react with gasifying agents Equations (8)–(14). The process predominantly yields syngas, with minor amounts of tar typically present as impurities. While gasification of plastic waste shows great potential to produce H₂ and energy, the operational issues of generating up to 160 g/m³ of tar as a byproduct acts as a major operational issue (Cortazar et al. 2023; Lopez et al. 2018; Mastellone et al. 2010; Wilk and Hofbauer 2013; Yousef et al. 2024). The US Department of Energy defines tar as a complex mixture of hydrocarbons having a higher molecular weight than benzene (Rabou et al. 2009). Tar formation occurs due to the random cleavage of polymeric chains, influenced by varying temperatures and residence times within the reactor. The presence of polyolefinic waste promotes the formation of light olefins, which serve as precursors for tar formation at lower temperatures. Additionally, polymers containing aromatic rings, such as PS and PET, contribute to the production of PAHs, one of the key components of tar. On an industrial scale, tar buildup poses several operational challenges, including reduced heat transfer, reactor wall corrosion, increased maintenance costs due to frequent cleaning, lower carbon conversion efficiency and energy recovery. Tar management has been practiced for decades, using both primary and secondary measures. Primary measures, such as optimized reactor design, catalytic bed materials (Han et al. 2022), and two-stage gasification (Park et al. 2016) help in tar minimization. Secondary measures, including catalytic cracking, thermal cracking, oil scrubbing, and adsorption, are then applied to achieve syngas purity suitable for downstream applications (Anis and Zainal 2011). While conventional methods remain dominant in industry, recent research explores advanced techniques such as in-line pyrolysis-steam reforming (Razak et al. 2024) to further improve efficiency. Although these methods reduce tar formation and enhance syngas quality, they often require complex setups, additional catalysts, and lower operating temperatures, which can limit the degree of plastic waste conversion. In contrast, thermal plasma gasification achieves process intensification by operating at extremely high temperatures, ensuring complete decomposition of plastic polymers, minimizing tar and toxic by-products, with greater flexibility in handling diverse waste streams. However, these outcomes are highly dependent on the reactor design, which plays an important role in determining the extent of conversion.

Throughout the reviewed literature plasma gasification has three main material inputs – the feed, oxidizing gas, and plasma-forming gas. In a few cases the plasma-forming gas acts as the oxidizing agent. However, the scenario changes with the scale, an additional input in the form coke bed is added in large scale reactors which not only acts as a mechanical support for the waste bed and a heat source or heat trap but also provides a strong reducing environment. Such conditions facilitate the reduction of refractory metal oxides (e.g., Fe₂O₃ to FeO), thereby enhancing the activity of fluxing agents and promoting slag formation (Gollapalli et al. 2024; Horton et al. 2016). The oxidizing agent can be introduced at various points along the reactor walls or via peripheral injection through the plasma torch. Injecting oxidizing gas through multiple ports or nozzles at different heights in a reactor allows for uniform gas distribution, which enhances mixing and provides control over gasification zones (pyrolysis, oxidation, reduction). Peripheral injection allows for regulated distribution around the plasma arc, leading to a higher rate of heat transfer to the oxidizing gas near the feed which can increase the rate of reaction by increasing the rate of volatilization. However, temperature distribution and electrode erosion must be carefully analyzed based on the chosen injection technique to optimize reactor performance.

Figure 3 presents a schematic of plasma gasification plant, typically the process consists of waste pre-processing units, plasma torches, gasifier, power supply, and compressors, which are then followed by the down streaming processes of char separation, gas cleaning, and purification just like other thermochemical processes. The studies on treatment of plastic waste through plasma gasification have been summarized in Table 3. The energy efficiency of the process depends on the quality of syngas produced, which is greatly affected by the oxidizing gas used. The theoretical composition of syngas produced by gasification of waste plastics using different oxidizing agents can be estimated from Equations (8)–(11) (Razak et al. 2024). Fathi et al. (2024), detailed the gasification of PP using H₂O/Ar plasma and CO₂ as the oxidizing medium; the syngas produced contained high amount of H₂ and CO, occupying 85.4 vol% of the total mixture. The authors also explored the effects of super-stoichiometric and sub-stoichiometric additions of CO₂ and reported a highest CO₂ decomposition rate of 99 % with sub-stoichiometric levels. The overall energy efficiency, considering the syngas and solid carbon, was 54 %. Hlina et al. (2014), also used CO₂ oxidizer to convert a mixture of HDPE (89 %), PP (10 %), and PET (1 %) into syngas with a LHV of 10.77 MJ/m³ with 80 % carbon conversion to gas. This level of conversion is consistent with the use of stoichiometric amounts of CO₂, which limits complete gasification of carbon. Higher carbon conversion efficiency can be achieved under super-stoichiometric CO₂ conditions, as demonstrated by Fathi et al. (2024). Whereas Hrabovsky et al. (2010), worked with a mixture of CO₂ and O₂ as the oxidizing media, the syngas produced was reported to have an LHV of 8.5 MJ/m³ and 100 % conversion. The difference in conversion performances in these two cases is due to the presence of O₂, and even though the amount of H₂ that can be produced theoretically is the same for both the oxidizing agents, the LHV of syngas produced through the CO₂ oxidized system is higher due to a higher occurrence of Boudouard reactions. The choice of gasifying agent – CO₂, H₂O, O₂, or air – depends on whether the goal is to maximize CO or H₂ production, optimize energy consumption, or prioritize cost-effectiveness for waste-to-energy applications. Supplementary Table S2 compares different oxidizing agents and the objectives that they serve.

Figure 3:

Plasma gasification process. (a) Block diagram representation of the plasma gasification process with upstream and downstream processes. (b) Schematic representation of 10 TPD plasma gasification plant at GS Plastech, Cheongsong, South Korea. Reproduced with kind permission from ref. (Byun et al. 2012). Copyright 2012, IntechOpen.

3.1 Arc plasma

Thermal plasma torches used in waste processing are classified into transferred and non-transferred types, based on the location of formation of plasma arc as shown in Figure 4. In transferred torches, the arc forms between the cathode and the conductive work piece (anode), enabling efficient heat transfer ideal for high-energy tasks like welding and metal processing. Non-transferred torches form the arc within the torch itself, projecting a high-temperature plasma jet (2000–15,000 K), making them suitable for pyrolysis and gasification of non-conductive carbon-based wastes like plastics. These torches offer greater flexibility in configuration and reactor integration. Further classification based on the power source used for plasma generation is as follows:

Figure 4:

Schematic representation of transferred and non-transferred torches.

3.2 RF induced plasma

Radio frequency (RF)-induced plasma is generated using electromagnetic induction, typically at frequencies between 13.56 MHz and 40 MHz (Gabbar et al. 2021). AC flows through an induction coil, to create oscillating magnetic fields that induces eddy currents in the forming gas. These currents ionize the gas, sustaining the plasma with electron densities of the order 10¹⁶–10²³ m⁻³ and temperatures between 6,000 and 10,000 K. Generally low enthalpy gases like Ar and N₂ are used in RF plasma torches and oxidizing gases are added in small concentrations. Their design typically features concentric quartz tubes, with the central tube supplying the plasma-forming gas and a sheath gas for cooling (Guddeti et al. 2000a,b). However, RF plasma torches are less efficient than DC systems and face challenges in gas flow rate control and power output. The complexity of ignition and maintaining stability under varying conditions necessitates precise optimization of system parameters.

3.3 MW induced plasma

Microwave-induced plasma is generated using high-frequency electromagnetic waves, typically between 300 MHz and 10 GHz, with 2.45 GHz being the most common frequency due to its availability and effectiveness in ionizing gases (Gabbar et al. 2021). These plasmas utilize electron cyclotron resonance, where electrons spiral in circular paths under a magnetic field, enabling efficient energy transfer and minimal thermal losses (Sanlisoy and Ozdinc Carpinlioglu 2019). MW plasmas are characterized by larger plasma zones and higher densities compared to RF plasmas, making them ideal for applications requiring high energy distribution over a wide area (Ho et al. 2017). However, MW plasma torches face challenges, including lower efficiency compared to DC systems and difficulties in scaling for industrial use as impedance matching and plasma stability require precise engineering (Zamri et al. 2021). Despite these limitations, MW plasma technology continues to show potential for scale-up. Supplementary Table S3 provides a comparative summary of non-transferred arc and microwave plasma torches, covering both technical and non-technical aspects in the context of waste processing.

4.1 Effect of feedstock on product

In thermal plasma waste processing, the relationship between the feed material and the resulting product is central to understand the efficiency and outcomes of the process. The feed, typically consists of waste materials such as MSW or industrial residues, undergoes high-temperature transformations in the plasma reactor. This section explores various factors of feed such as composition, inorganic content, feed rate, particle size, and moisture that significantly influence the overall performance of the plasma system.

Sorting and segregation of plastic waste from other waste is difficult in countries like India (Kibria et al. 2023), leading researchers to explore plasma co-gasification of mixed waste streams as an efficient solution for managing waste while minimizing the need for complex pre-sorting steps. Park et al. (2016), studied the effect of varying blend ratios of HDPE and wood pellets on the syngas composition. The study found that increasing biomass content led to higher CO₂ concentrations, attributed to the greater oxygen content in wood. Conversely, higher HDPE content resulted in increased CH₄ production. In another plasma co-gasification study for sawdust and PVC mixtures by Chu et al. (2023), it is reported that chlorine release distribution is mainly dominated by HCl in the gas product, KCl in liquid, and KCaCl₃ in solid product. The proportion of total chlorine released into the gas phase increased significantly by 17 % as plasma torch power rose from 16 to 24 kW, and by approximately 6.3 % with an increase in the steam-to-carbon (S/C) ratio from 0 to 2. These findings suggest that input power (and thereby temperature) exerts a stronger influence on chlorine volatilization than the S/C ratio.

While the type of feed significantly affects the efficiency and outcomes of plasma gasification, the feed rate also directly affects residence time and thermal exposure of particles, thereby influencing gas yield and conversion efficiency. Tang et al. (2003), observed that increasing the PP feed rate from 20 g/min to 40 g/min at a fixed power input of 35.2 kW reduced the H₂ concentration in the gas from 68.8 % to 62 % and increased the yield of C₂ gases, indicating incomplete thermal cracking due to shorter residence time. Increasing the feed rate reduces the mean temperature available to heat each particle thus leading to lower conversion of solid to gas and a higher solid yield (Guddeti et al. 2000b). Zhang et al. (2017), demonstrated that while the overall carbon conversion was similar at feed rates of 12 g/min and 17.65 g/min, the higher feed rate resulted in greater acetylene yield and lower specific energy consumption, thus enhancing energy efficiency.

Volatilization of feed particles is the first stage of reaction inside the plasma processing reactor and plays an integral role in determining the overall reaction rate and efficiency of the reactor. The size of the particle controls the rate of heat transfer from plasma gas to the particle surface where the reaction initiates thus also impacting the rate of reaction. Hrabovsky et al. (2022), explains that with decreasing particle size the surface temperature of the particles increases and therefore increases the rate of gasification as illustrated in Supplementary Figures S2 and S3.

Moisture content in the feed is another critical parameter that influences plasma processing efficiency. Controlling the moisture is also important to maintain consistency in feed quality as moisture content can vary seasonally and regionally (Singhal et al. 2022). Syarif et al. (2019), modeled the temperature profile of a plasma electric furnace used for processing MSW (see Supplementary Figure S4), and reported the impact of varying feed moisture content on the temperature distribution along the reactor. As the wet feed moves through the reactor, moisture evaporates and generates steam, which acts as a sheath gas around the feed. This increases the power demand on the plasma torch to heat the feed sufficiently for gasification. Additionally, more energy is required to break down water molecules before gasification can proceed. Higher moisture content increases the steam inside the reactor which on mixing with the hot gases increases the total pressure and flowrate of gas inside the reactor that affects the residence time and overall conversion. The increase in feed moisture inversely affects the ratio output gas LHV to total reaction enthalpy, which means decreasing energy efficiency of the conversion process. In another study, it was shown that incorporating a resistively heated drying zone in the reactor reduced the specific power required for gasification by the plasma torch to 0.4 kWh/kg (Aliferov et al. 2016). Therefore, pre-drying the feed, whether through resistive heating or by recycling heat from the product gas can significantly improve process efficiency.

4.2 Effect of plasma-forming gas

The plasma-forming gas, combined with a carrier gas or oxidizing agent, creates the reactive environment necessary for plasma waste processing. Plasma can serve either as a heat source alone or as both a heat source and reactive medium, depending on the specific processing technique employed, as illustrated in Figure 5. Inert gases, such as Ar and N₂, are typically used in plasma pyrolysis to prevent oxidation of the waste material. In contrast, oxidizing gases are utilized in plasma gasification, contributing additional heat through exothermic reactions, as previously discussed. When scaling up to industrial applications, the selection criteria for plasma-forming gases become broader, with additional factors such as power consumption for plasma generation, cost of gas, heat dissipation, electro thermal efficiency, and most importantly electrode erosion rate. Electrode erosion rate is an important factor to consider in the economics and maintenance of plasma gasification systems. The choice of gas will affect the frequency of electrode replacement and associated downtime and costs. RF and MW-induced plasma torches avoid the issue of electrode erosion. However, for large-scale waste plasma processing, water-cooled DC/AC arc torches with hollow copper electrodes are commonly used. Economical gases like steam and air are commonly used as plasma-forming gases. While steam enhances syngas quality by increasing hydrogen yield, air is more economical but dilutes the syngas due to its high nitrogen content. The downside, however, is that these gases accelerate cathode erosion by forming oxides and other intermediates with the copper electrodes, in addition to the standard erosion caused by Joule heating and ion–cathode interactions (Subbotin et al. 2018). The cathode erosion rate increases proportionally with plasma torch power and plasma residence time. Szente et al. (1992), provide a comprehensive database on electrode erosion rates for commonly used plasma-forming gases and their mixtures, all measured at a fixed flow rate of 20 L/min. The high-speed projection of eroded cathode particles may also damage the inner refractory lining. Additionally, the thermodynamic properties of plasma-forming gases significantly influence process efficiency. For example, H₂ plasma, which has been studied for acetylene production via reductive pyrolysis (Zhang et al. 2017, 2018), induces severe electrode erosion due to its high enthalpy, requiring higher power input for ionization. Figure 6a presents the enthalpy of various plasma-forming gases at different temperatures.

Figure 5:

Selection of plasma-forming gas and carrier gas. Reproduced with permissions from ref. (Hrabovsky and van der Walt 2018). Copyright 2018, Springer.

Figure 6:

Thermodynamic properties of various individual gases under plasma conditions: (a) Enthalpy, (b) specific heat capacity, (c) thermal conductivity. Reproduced with kind permission from ref. (Boulos et al. 2016a).

The high thermal conductivity of H₂, even at low temperatures (see Figure 6c), leads to elevated arc voltages, amplifying cathode erosion. As a result, H₂ is typically used in mixtures with Ar to mitigate these effects (Munz and Habelrih 1992). In contrast, Ar and N₂ typically exhibit lower specific enthalpy under plasma conditions compared to gases like steam and air. As a result, higher flow rates of Ar and N₂ are required to deliver equivalent energy for effective waste conversion, increasing gas volume and diluting the product syngas, raising downstream purification demands and raw material costs. Steam plasma, by comparison, can transfer the same energy with a lower mass flowrate and offers high thermal conductivity even at lower temperatures, promoting efficient heat transfer to the waste. However, steam’s high heat capacity requires greater electrical energy input to reach plasma temperatures. Thus, another criterion, the cost of the plasma-forming gas, becomes especially critical at pilot and industrial scales, where large volumes are used continuously. Inert gases like Ar and He, though chemically stable and less erosive, are significantly more expensive than air or steam, impacting long-term operational expenses. Even steam, despite being derived from water, involves costs related to generation and system integration. Therefore, cost-effectiveness, alongside thermophysical properties, often dictates the final choice of plasma gas in scaled-up systems (Hrabovsky et al. 2022). Table 4 compares various plasma-forming gases in terms of their performance and effectiveness in waste plasma processing.

4.3 Solid product and its characterization

Thermal plasma processing of carbonaceous waste yields two categories of solid products: vitrified slag and carbonaceous residues. These products differ significantly in their origin, composition, behaviour in the reactor, and post-treatment applications. Independent discussion on these outputs is covered in this section.

Vitrified slag, often referred to as plasma rock, is a dense, glassy solid formed during plasma treatment of inorganic-rich waste streams. In plasma reactors, especially fixed-bed configurations, the inorganic components in the feed – such as metals, ash, minerals, and glass – melt under extreme temperatures and exit the reactor as a molten flow. Upon rapid cooling or quenching, this flow solidifies into a vitrified structure that encapsulates hazardous elements. The stability, structure, and leaching behavior of this slag are central to its environmental safety and potential for reuse. With waste sorting and segregation being one of the toughest issues at hand in managing solid waste, it is clear that the plastic waste would be sourced from municipal and medical solid waste both of which often contain significant amounts of inorganic substances in the form of glass, metallic entities, and mineral waste (Sniadecka et al. 2016). Chu et al. (1998), conducted thermal plasma processing experiments using synthetic medical waste surrogates consisting of both combustible (e.g., pork ribs, gauzes, swabs) and non-combustible (e.g., stainless steel, glass, syringes) fractions. The study highlights that when metallic components exceed the 40 wt% limit in feed, the resulting slag could surpass regulatory leaching thresholds for heavy metals. Therefore, a careful balance in the inorganic fraction is essential. To promote proper melting and glass formation, fluxing agents – primarily CaO, SiO₂, and Al₂O₃ – are typically added to the feed. These oxides reduce the viscosity and melting point of the slag, supporting the formation of a homogenous, amorphous matrix. The recommended ranges for CaO (20–35 wt%) and SiO₂ (30–60 wt%) in the solid slag were given by Ma et al. (2021). Coke is also often included to create a reducing environment that enhances the behavior of fluxing agents and facilitates melting. Another significant factor is the cooling rate of the molten slag. Rapid quenching rates, typically above 700 °C/s, inhibit crystallization and yield a dense amorphous glass, as shown by Hu et al. (2023). This rapid solidification minimizes porosity and stabilizes toxic metals, such as Cr, reducing their leached concentrations to as low as 0.002 mg/L, well below regulatory limits.

The physical properties of the resulting slag directly influence its suitability for reuse. High concentrations of amorphous oxides contribute to low porosity and high density, making the material appropriate for construction bricks, asphalt filler, or concrete aggregates (Zhao et al. 2024). Conversely, the inclusion of carbonates or sulfates – such as CaSO₄ – can introduce foaming effects during vitrification, producing porous slag with thermal conductivity as low as 0.22 W/m K, suitable for insulation materials (Lázár et al. 2018). Additionally, the formation of metallic inclusions (e.g., spheroidal or needle-like structures) can increase slag density and impact mechanical performance (Chu et al. 1998).

Environmental compliance is gauged through leachability testing, which assesses the long-term stability of heavy metals within the slag. The leachability of slag can be tested by a three-step sequential extraction procedure developed by the Commission of the European Communities Bureau (BCR) (Tessier et al. 1979). After the extraction, the concentration of heavy metals in each fraction of the slag is determined using an inductively coupled plasma optical emission spectrometer (ICP-OES).

The properties and type of solid product formed depend on the plasma processing technique used due to the operating conditions. In plasma gasification the amount of char and volatiles found in the solid residue is less compared to plasma pyrolysis primarily due to the oxidizing environment and higher temperatures. Bhatt et al. (2024), reported that during plasma pyrolysis of plastic waste the solid residue decreases with increase in temperature. On comparison of different polymers, PET produces the highest amount of solid residue due to formation of oxygenated compounds as previously discussed. Fazekas et al. (2016), observed that the high velocity of the plasma jet can cause solids to become entrained as soot in the off-gas and to adhere to reactor walls. This phenomenon occurs when particles spend insufficient time in the plasma zone, leading to incomplete conversion. Gas chromatography-mass spectrometry (GC-MS) analysis of compounds extracted from soot using toluene revealed that aromatic and polyaromatic compounds formed during breakdown adsorb onto the soot. Thus, optimizing residence time and maintaining temperature uniformity in the reactor are essential. Plasma pyrolysis of plastics has been explored for producing high-value carbon nanostructures as solid products. Polypropylene (PP), for example, has been studied as a feedstock for synthesizing carbon nanotubes (CNTs) through plasma pyrolysis. Transmission electron microscopy (TEM) analysis confirmed that increasing arc current decreases the diameter of the resulting CNTs (Mohsenian et al. 2018). However, despite the successful formation of carbon nanostructures observed through various microscopic techniques, the yield remains limited (up to 20 wt%), and the selectivity is insufficient for scalable CNT production via plastic plasma pyrolysis (Fathi et al. 2024).

Raman spectroscopy of the solid residue provides insight into the purity of the carbon nanostructures produced. For carbon black samples from PP plasma pyrolysis, an I_D/I_G ratio of 0.96 indicates a high degree of structural defects, such as vacancies, edge irregularities, and lattice distortions. This level of disorder could be advantageous in applications needing defect-rich surfaces, like battery electrodes, but is less ideal for applications requiring structural integrity. The CNTs synthesized through this method show limited selectivity and contain various impurities. High-resolution transmission electron microscopy (HRTEM) reveals additional carbon forms, including onion-like, spherical, and non-spherical nanostructures, which act as impurities. Furthermore, energy dispersive X-ray (EDX) analysis highlights the presence of inorganic elemental species within the solid residue, further impacting the purity of the final product. Mohsenian et al. (2016), found that increasing arc current during ABS pyrolysis reduces SiO₂ and CaCO₃ intensities in XRD spectra, indicating higher-purity carbon black at elevated power levels.

5.1 Heat transfer

Heat transfer in plasma processing of waste differs significantly from conventional processes due to the extreme temperatures, high energy densities, and unique interactions between plasma gases and waste particles, plasma systems involve complex interactions like rapid convection and potential radiative effects, requiring tailored study for accurate and efficient processing (Demchenko et al. 2019; Elaissi and Alsaif 2023; Pancholi et al. 2022).

In a plasma processing reactor, the heat required to carry out the waste conversion is provided by the plasma torch. However, the total power exerted by the torch ( P net plasma ) is the remainder value from the difference between the actual power supplied to the torch ( W ) and the power lost due to the heating of the torch ( P loss torch ). The power required to form the plasma is proportional to the enthalpy of the forming gas used and the capacity of the system, as represented in Equation (18) where H AV is the mass-averaged enthalpy of forming gas and M is the mass flowrate (Hrabovsky et al. 2022). Looking at Equation (19) the heat delivered by the torch gets utilized for heating and volatilizing the plastic waste ( P volat org ) in the feed and melting the inorganics ( P melt inorg ). Post conversion to gas, the term P react gas represents power of chemical reactions which can either be positive or negative owing to extent of endothermic or exothermic reactions. Furthermore, the power is also dissipated through the reactor walls ( P heat wall ), for reduction of which thick refractory lining is necessary. The heat is also consumed in heating the carrier gas or oxidizing gas P heat oxi / carr that enter the reactor through reactor walls. Products of the conversion along with unreacted plasma leaving the reactor also dissipates heat from the system. Hence, although the heat is provided by the torch, there are several factors such as reactant flow rates, thermal conductivity of refractory lining, residence time of gases, and thermodynamic properties of plasma that control the temperature inside the reactor.

In waste thermochemical conversion systems, the time constants for gas phase reactions are contrastingly shorter when compared to the volatilization of feed material which makes it significant for determining the rate of conversion. Therefore, it is important to understand the heat transfer routes and the factors affecting the rate of heat transfer to the feed.

Convection is the dominant mode of heat transfer to feed in waste plasma processing as the hot plasma comes into direct contact with waste particles transferring heat through forced convection due to high-velocity plasma flow. Radiation plays a role in both heating the waste particle and dissipating heat from the system. However, radiative heating of particle is considered negligible with the consideration that the plasma is optically thin, and angle of incidence is low, meaning that radiation can escape without significant reabsorption (Boulos et al. 2016a). That said, there are certain cases where plasma radiation does become significant which will be discussed in detail in the sub-Section 5.1.2. The loss of heat from particle surface through radiation is particularly notable for particles with high melting point such as metals and inorganics. In contrast materials with a prevailing organic structure tend to volatilize quickly and hence follow a different route. As a result, any heat distribution model must account for the concentration of organic and inorganic species in the mixed feed to accurately predict heat transfer behavior within the plasma reactor. Regarding conduction, this mode of heat transfer is mainly relevant for inorganic materials. It occurs as internal conduction, transferring heat from the particle surface to its center and is highly dependent on the thermal conductivity of the material. The modes of heat transfer between plasma, waste particle, and the reaction environment are schematically represented in Figure 7. The net heat received by a particle, Q net is given by Equation (20) which is a sum of the heat received through convection, Q convec , heat received through radiation from plasma, Q rad plasma , and heat dissipating through radiation from particle surface, Q rad particle , which are given by Equations (21)–(23), respectively.

Here, A p , ϵ , σ , represent the surface area of the particle, emissivity and Stefan-Boltzman constant, respectively. T ∞ is the temperature of the bulk plasma phase, T p is the surface temperature of the particle, T w is the ambient wall temperature, and h is the heat transfer coefficient which can be calculated as shown in Equation (24) by using Nusselt number (Nu) which is most commonly a function of Reynolds (Re) and Prandtl (Pr) number, but corrections to the equation exist when the composition along with thermodynamic and transport properties of plasma are accounted for (Boulos et al. 2016a).

Here, k and D p are the thermal conductivity of plasma-forming gas and the particle diameter respectively.

Figure 7:

Heat transfer routes inside a waste plasma processing reactor.

Demchenko et al. (2019), studied steam plasma gasification of carbonaceous feed (coal particles in particular) by using a modified version of Equation (20) and calculated the time required for heating the particle up to the reaction temperature. Their calculation was based on common assumptions typically applied to entrained flow gasifiers – such as uniform spherical particle shape, negligible internal temperature gradients, and dominant convective heat transfer from the gas to the particle surface. However, they also noted that in the plasma gasification environment, where gas temperatures exceed 2,000 K, radiative heat transfer becomes the dominant mechanism. This suggests that such assumptions may not hold under plasma conditions, and heat transfer dynamics in plasma reactors may differ significantly from conventional entrained flow systems. The equation is given as follows:

where ρ p , and c P represent the density and specific heat capacity of the particle respectively.

The heat transfer to carbon feed is divided into three stages, (i) heating of the particle, (ii) volatilization of particle surface and sheath gas formation, (iii) conversion to coke. Upon integrating the convection and radiation terms in Equation (26) individually we get the time required for heating the particle ( t h

The heating time of coal particles, t h is calculated using a consecutive approximation method. The convective heating temperature, T _pc, is first determined from Equation (27). Through this the radiative heating component is estimated as T ₁ = 1800 – T _pc. The t h , value is then refined by solving Equation (28). The process iterates until convergence, effectively modeling combined convective and radiative heat transfer during gasification. As previously discussed, the time required for conversion is directly proportional to the particle size and inversely proportional to temperature of the particle. The same pattern is observed from the modelled calculations depicted in Figure 8a. Smaller particles heat faster due to their higher surface area-to-volume ratio, which accelerates both convective and radiative heat exchanges. Notably, as T _g > 3,000 K, the rate of heating time reduction slows due to reduced thermal conductivity, attributed to the disassociation of water molecules. Figure 8b depicts the relationship between reactor chamber volume, V (L) and T _g for various particle sizes given by Equation (29). The graph indicates that for particles up to 50 μm, the required chamber volume does not change significantly beyond T _g = 2,000 K. Larger particles (D _p ≥ 150 μm) demand significantly higher chamber volumes at all temperatures, reflecting their longer heating and conversion times. The chamber volume decreases as T _g increases up to 3,000 K, due to elevated thermal conductivity, accelerated reaction rates and faster particle conversion. These insights are crucial for optimizing reactor design, balancing temperature, particle size, and reactor capacity.

where, m g , R , and P g are mass of gas, gas constant, and pressure inside the reactor.

Figure 8:

Effect of particle size and reactor gas temperature on (a) heating time, (b) required reaction volume at constant pressure. 1 –  D p  = 200 μm; 2 –  D p  = 150 μm; 3 –  D p  = 100 μm; 4 –  D p  = 50 μm. Reproduced with kind permission from ref (Demchenko et al. 2019). Licensed under CC BY 4.0.

The model incorporated temperature-dependent thermal conductivity of plasma-forming gas enabling detailed analysis of the prevalent mode of heat transfer in different regions of the reactor and the required time to heat the waste particle. This information is valuable in deciding the dimensions of plasma and freeboard regions of the reactor. Moreover, the reactor dimensions will vary based on the type of plasma-forming gas used, owing to the differences in their physical and thermal properties. However, the model has certain limitations, including the assumption of 1,800 K as the CBP (carbon boundary point – a temperature value beyond which carbon will not exist in solid form), the exclusion of energy exchange from gas-phase reactions which will also influence the particle heating time. Additionally, a higher valued range of particle sizes (2–50 mm) can be explored to imitate real world conditions.

5.1.2 Radiation

Radiative heat transfer is generally neglected in waste plasma processing systems due to the complexity of the calculations involved. However, radiation occurs through the plasma, the solid particles within the system, and the refractory lining which absorbs and re-radiates heat depending on its emissivity and temperature. Waste plasma processing reactions are mostly performed at atmospheric pressure and with an arc temperature range of 5,000–15000 K where the effects of radiation are subdued. Nevertheless, the thermodynamic properties of plasma change at high pressures, currents and temperatures. Diab et al. (2024), in their work explain that plasma at 1 bar can dissipate significant amounts heat at currents higher than 100 A and at higher pressures such as 20 bar it becomes optically thick as the plasma density increases and even emit and absorb radiation at lower current values. Figure 9 shows how the increasing trend of axial temperature varies with and without accounting for radiation in the model as pressure and current values are increased for a hydrogen arc plasma. Plasma arc torches generally operate in the range of 100–1,500 A as seen in the studies listed previously in Tables 2 and 3 indicating at the possibility of heat dissipation through radiation inside the arc chamber and loss of power to torch that can reduce the electro thermal efficiency. The practical ranges for significance of radiative plasma are as follows:

1. Insignificant radiation: current < 100 A, pressure < 1 bar, temperature < 10,000 K.

2. Significant radiation: current > 100 A, pressure > 10 bar, temperature > 15,000 K.

3. Dominant radiation: current > 200 A, pressure > 50 bar, temperature > 20,000 K.

Figure 9:

Axis temperature of wall stabilized hydrogen arc column with and without radiation at 1 bar and 20 bar. Reproduced with kind permission from ref. (Gueye et al. 2017). Copyright 2017, AIP Publishing.

Upon plasma – solid interaction inside the reactor the heat loss due to radiation becomes more prominent with increasing surface temperature, surface area, and the melting point of the solid. When then particle size of materials such as metals increases the temperature of plasma required to melt the particle also increases as it continues to lose heat through radiation (Boulos et al. (2016a). The temperature increase is however subjective to the thermal conductivity of the forming gas. The inclusion of radiation heat transfer in numerical models of plasma reactors significantly influences the accuracy of predicted outcomes, such as temperature distribution, soot formation, and gas concentrations. Liavonchyk et al. (2021), using an entrained-flow plasma gasifier, compared the simulated and experimental results for plasma gasification of rubber pyrolysis vapors and demonstrated that accounting for radiation improved temperature predictions, reducing discrepancies between modeled and experimental values. Without radiation, temperatures at various points in the reactor were substantially overestimated, with deviations of up to 470 K. Incorporating radiation reduced these errors to a range of 100–180 K. The errors arose due to the inaccuracy in assigning the radiation model to the fibrous insulation refractory material in the reactor. In a similar previous study, Liavonchyk et al. (2017) stated that the inaccuracies were a result of defining the properties of refractory material as independent of temperature and also due re-radiation from walls.

Soot formation also varied significantly with the inclusion of radiation effects. In Liavonchyk et al. (2021), models without radiation over-predicted soot content at 6.7 wt%, whereas models with radiation reduced this to 2.9 wt%, aligning better with experimental trends. This reduction is attributed to improved heat distribution, which enhances soot oxidation into syngas. It is further emphasized that accurate modeling of soot kinetics remains challenging due to thermodynamic equilibrium simply not allowing the formation of soot at high temperatures and the employed models being more accurate for complete combustion reaction mechanisms while only being nearly approximate for partial oxidation reactions (Liavonchyk et al. 2017). These findings underscore the critical role of radiation in enhancing model reliability but also point to areas needing further refinement.

5.2 Theoretical studies

Thermal plasma processing of waste demands extremely high temperatures, resulting in significant electricity consumption and a large supply of plasma-forming gas at elevated pressures. Additionally, frequent maintenance of plasma torches further increases the cost of collecting data through experiments. The process is also highly time-consuming due to the need for optimizing multiple variables, such as feed flow rate, gas flow rate, temperature, pressure, plasma torch configuration, and the type of forming gas used. Therefore, to increase the pace of attaining technological maturity mathematical models are developed (Munir et al. 2019). Mathematical models can be of three types namely – equilibrium, numerical, and kinetic. Equilibrium models also known as systematic models are a simpler approach to predict the outcomes of a process based on minimization of Gibb’s free energy and by assuming a thermodynamic equilibrium (Janajreh et al. 2013; Montiel-Bohórquez et al. 2022; Pitrez et al. 2023; Sharma 2008). Accurate prediction of output is necessary to design the down streaming processes. Equilibrium models do not take into account the effects of reactor geometry, heat and mass transfer patterns, reaction kinetics, and assume infinite residence time which results in lack of validation from experimental results. Most studies published on non-stoichiometric Gibb’s free energy minimization approach provide data with the assumption of an idealized behavior, which is impractical and insufficient in reactor designing. On the other hand, Numerical models consider flow equations, reactive species, and reactor geometry to solve complex equations of heat, mass, and momentum transfer by using computational techniques. Finite element analysis (Sedej and Mbonimpa 2021), finite volume analysis (Qing et al. 2022), Monte carlo simulations, and density function theory (Huang et al. 2013) are some of the commonly reported computational techniques. Kinetic models observe the chemical reactions in a system and determine their kinetic rates. These high-fidelity models prove to be very useful in estimating the harmful pollutants such acid gases, and dioxins and furans produced during the thermal treatment of waste as they follow slow kinetics (Grandesso et al. 2008; Bei et al. 2022). These different models are discussed in depth in the sub-sections ahead.

5.2.1 Numerical models (CFD)

Computational fluid dynamics studies group together conservation of mass, energy, and momentum in a defined volume to provide a theoretical description of the process. For multiphase reactions such as plasma pyrolysis and plasma gasification, CFD is important to visualize the plasma – material interaction, heat distribution, flow and concentration of gases inside the reactor, heat transfer to the material, progression of the reaction through volatilization and describe the kinetics of homogeneous chemical reactions. Two general modelling approaches, Euler–Euler and Euler–Lagrangian have been followed by researchers to simulate multiphase plasma reactors. The Euler–Euler method uses volumetric representation to show the presence of both continuous and discrete phases in definite volumes, here both phases are treated as interpenetrating continuous phases. On the other hand, Euler–Lagrangian method analyses continuous and discrete phases separately. This approach is widely used even in conventional fluidized bed gasification to track the trajectory and interaction of individual particles. However, the high conversion rates of waste particles such as plastic in a plasma reactor raise the question as to which method will be more efficient. Popularly Euler–Euler approach has been followed mainly due to the quick conversion of solids into gas and also because fewer boundary conditions (particle size boundaries eliminated) and lesser computational power is required compared to Euler–Lagrangian. A more detailed explanation on the application of these CFD modelling approaches for thermochemical waste processing technologies in the turbulent regime is given by (Ramos et al. 2019a,b). CFD models are an economical and time efficient option for optimization of operating conditions and structure of the plasma reactor, Table 5 features CFD simulation studies carried out for waste plasma processing. Ismail et al. (2019), simulated a plasma fixed bed gasifier for processing MSW by describing kinetic reactions using Arrhenius correlation in a two-step model of primary pyrolysis to produce gases followed by tar cracking. Homogeneous reactions were solved using finite-rate/eddy dissipation method and heterogeneous reactions were modelled using kinetic/diffusion surface reaction model to include the reactant diffusion resistances and their behavior in a turbulent flow. The standard k-ε model was used to simulate flow of materials and energy in the turbulent regime. Performance of the reactor was improved by optimizing parameters like equivalence ratio, steam-fuel ratio, and input power to plasma torch. Zhang et al. (2013), optimized same parameters within a different numerical range and validated the predicted results with experimental runs with good agreement showcased by an error percentage of mere −4.6 %. From the gas distribution contours presented in the study it can be said that formation of CO is much faster than H₂ as high concentration of CO is found in the middle part of the reactor close to interaction point between waste and plasma. Qing et al. (2022), took inspiration for the design of downdraft plasma gasifier incorporating a circumferential injection of plasma-forming gas to increase residence time from the work of Ibrahimoglu et al. (2017), and studied the impact of three different MSW sub-components – food/paper/yard waste on syngas quality. Highest yield of H₂ of 55 mol% was observed with yard waste due to high moisture. On comparison of H₂ concentration results from Ibrahimoglu et al.’s work a deviation of +20 % is observed, potentially due to implementation of species transport model. Another contributing factor could be the inclusion of radiative heat transport in their study, where heat dissipation through radiation might limit heat available for endothermic homogeneous reactions. Although complex, as previously discussed radiative heat transfer becomes significant in large-scale plasma torches utilizing high current arcs and under these conditions radiative heat loss becomes dominant (Diab et al. 2024). Therefore, integrating radiation heat transfer model into plasma processing models is crucial for achieving more accurate results closer to real – world conditions. Models like discrete ordinates (DO), and P-1 radiation can be used to solve radiative transfer equation has they are suitable for high temperature application such as waste plasma processing and provide high fidelity responses for complex geometries (Thynell 1998).

Table 5:

Summary of CFD simulation studies carried out for waste plasma processing.

Feed	Reactor	Forming/oxidizing gas	Dimension/modelling approach	Study focus	Outcome	References
MSW	Plasma FB	Air/steam	2D/E-E	Parametric optimization: ER – 0.1–0.5; SFR – 0–0.8; input power – 40–100 kW	Optimal conditions – ER = 0.3; SFR = 0.5; input power = 60 kW	Ismail et al. (2019)
Coal	MW plasma DD	Air/steam	2D, 3D/E-L	Structural optimization: circumferential injection of plasma-forming gas	Increased residence time; syngas – 18.4 % (H2), 37.2 % (CO); CGE – 55.3 %	Ibrahimoglu et al. (2017)
MSW	Plasma UD	Argon/air	2D/E-E	Temperature and mass fraction distribution in updraft gasifiers	–	Sakhraji et al. (2022)
MSW	FB-UD	Air/steam	2D/E-E	Parametric optimization: ER – 0.043–0.077; SFR – 0–0.25; DPER – 0.108–0.128	Good agreement between calculated and experimental values and optimal conditions for syngas production at ER = 0.6; SFR ≤ 0.208	Zhang et al. (2013)
Hazardous waste	Plasma UD	Air	2D, 3D/E-E	Structural optimization: Effect of varying feed inlet positions and plasma torch inclination angles	Maximum temperature increased from 2,340 °C to 2,440 °C due to tangential feed flow; better temperature distribution with angular projected torches instead of horizontal	Zhang et al. (2024)
Wood particles	HEF	Air	2D/E-L	Design optimization: effect of varying plasma outlet nozzle and feed inlet position	Improved heat distribution and residence time with increased nozzle size and reduced distance between feed inlet and plasma jet	Bastek et al. (2023)
Food/yard/paper	FB-DD	Air	2D/E-L	Comparison of feedstock type based on syngas quality produced	Yard waste produces more H2 (∼55 mol%) due to presence of high moisture content	Qing et al. (2022)
MSW	FB-UD	Air/air, steam	0D, 2D, 3D/E-L	Parametric and structural optimization: ER – 0.1–1; SFR – 0.2–1.5; temperature – 500–2,500 K; plasma torch configuration	Required reaction volume for conversion of MSW and optimal operating conditions	Rojas-Perez et al. (2018)
MSW	FB-UD	Air	3D/E-L, GBR, PSO	Structural optimization and machine learning: optimizing reactor dimensions for a fixed design by creating a dataset within the range	Plasma gasifier designed using CFD and ML, validated with an experimental set up optimized for ER, reaction temperature and plasma jet mode	Shi et al. (2024)

In order to improve the efficiency of plasma processing reactors and achieve maximum conversion, uniform temperature distribution and mixing of gases must be ensured by optimizing the design and structure of the reactor. In an interesting study, Zhang et al. (2024), studied the hydrodynamics of a plasma gasification reactor with varying waste feed inlet orientations and plasma torch configurations. The initial design of oblique incident feed ports was replaced with three different orientations – tangential, diametrically opposite, and perpendicular inlets. Figure 10 shows the nephogram of plasma the gasifier with difference feed inlet orientations. From Figure 10a–c it can be observed that velocity of the particles is high near the walls when entering the reactor tangentially slows down as it moves closer to the center, thus creating a vortex allowing the particles to revolve around and experience more time inside the reactor. The temperature contours seen in Figure 10d–f show that due to the tangential feed flow, the hot gases swirling in the center are allowed to rise and a higher temperature sufficient for hetero and homogenous reactions is obtained in convergence and freeboard zones respectively. On the other hand, the particles coming in through diametrically opposite and perpendicular inlets collide in the center, thus reducing the temperature in the convergence zone and acting as a barrier for the rise of hot gases which creates undesired hot zones in the bottom part of the reactor. With this optimization it was possible to raise the maximum temperature up to 2,440 °C. Furthermore, the studies performed for the configuration of plasma torches showed that an angular injection of 23° provides slightly better temperature and velocity distribution along the reactor length.

Figure 10:

Comparison of the impact of tangential, diametrically opposite, and perpendicular waste feed inlet orientations on (a–c) velocity profile, (d–f) temperature distribution nephogram, (g–i) temperature distribution profile of inside the reactor, respectively. Reproduced with kind permission from, ref. (Zhang et al. 2024). Copyright 2023, Taylor and Francis.

Numerous variables come into the picture when designing a plasma reactor including geometrical parameters like reactor and plasma torch dimensions, operating conditions, special materials for refractory, electrodes, and nozzle lining and incorporating their properties (thermal conductivity, emissivity, etc.). The sheer number of permutations makes it a onerous task to explore the combinations through CFD due to computational limitations. By integrating machine learning (ML) with CFD models. ML models can efficiently learn patterns relationships between these variables and their corresponding reactor performance metrics through datasets of experimental results and CDF simulations to predict the optimum conditions for a desired outcome. In a recent study, Shi et al. (2024), explored the CFD-ML approach to optimize the geometrical configuration of a thermal plasma gasifier. For optimization, dimensions including heights, radii, and transition angles of the core reaction area and the free-board zone were considered. A dataset of 150 reactor combinations was thus created to train ML models. Gradient Booster Regression (GBR) model emerged as the most suitable model as it could incorporate the non-linear interactions between the operating parameters and analyzing the contribution of each parameter to reduce prediction errors. The predicted parameters were then refined using particle swarm optimization (PSO) and compared against CFD results an R ² value of 0.997 and 0.954 for training and testing respectively were recorded with a low prediction error of 1.3 %. The predicted temperature in the core reaction zone (3621 K) was close to the CFD simulated value of 3668 K. Experimental validation further verified these findings. Here equivalence ratio and reaction temperature, were optimized, an ER of 0.18 provided a balance between energy efficiency and syngas calorific value, while system performance was maximized at a reaction temperature of 1700 K. Additionally, laminar plasma jet mode demonstrated superior efficiency, achieving 16.4 % higher energy efficiency and 12.4 % greater syngas calorific value compared to turbulent mode.

5.2.2 Kinetic studies

Based on the complexity of reactions and responses to be recorded kinetic studies are categorized into global, lumped and intrinsic kinetics. The most simplistic approach for pyrolysis and gasification reactions is calculating global kinetics. Here the degree of conversion of the feed as a whole into the final products over time is observed rather than calculating the rate of consumption and formation of every intermediate chemical species. This method is preferred for studying plastic waste degradation due to the varying nature of the feed. Thermogravimetric analysis-based kinetic models follow similar methodology (Bozkurt et al. 2022; Martínez-Narro et al. 2023; Perez et al. 2023), but the same experimental data cannot be used for plasma processing given the stark contrast in heating rates. Additionally, apparent kinetics are subjective to the reactor design as they are affected by the heat and mass transfer patterns inside the reactor. Another representation of multiple complex reactions occurring simultaneously is executed by “lumping” chemical species with similar physical, chemical, and thermodynamic properties, which progress through the reaction as a single entity. This approach is practical for complex feedstocks but the assumed homogeneity of the input stream creates scalability challenges. Moreover, ensuring robustness in such a model requires extensive pilot-scale data with various feed compositions. Supplementary Figure S5 shows the general route followed to study lumped kinetics of plastic waste. In the context of plasma processing the rate constants of conversion reactions from bulk hydrocarbon solid to wax, oil, and eventually to gas are significantly higher. As a result, it is reasonable to assume a direct conversion from solid to gas, allowing for a simplified modelling approach with fewer lumped species. Intrinsic kinetics include the fundamental molecular level kinetics that are unaffected by macrolevel extrinsic factors, meaning the resistances offered due mass and heat transfer are minimized.

Estimating kinetic parameters of thermal plasma processing of plastic waste is an onerous task due to two reasons. First, the thermal conversion process is a result of rapid pyrolysis and gasification reactions, that means a high conversion rate of the hydrocarbonaceous feed making it difficult to collect data in practical time. Second, lack of kinetic studies and data for validation and selection of model. Regarding thermal plasma pyrolysis or gasification experimental kinetic studies have been conducted for propane pyrolysis under hydrogen plasma and methane pyrolysis under microwave Ar plasma by Ma et al. (2016) and Dors et al. (2014) respectively, but these works fail to provide clarity over model selection and development process.

Limited work has been published in exploring the kinetics of hydrocarbonaceous solid waste processing using plasma resulting in lack of data for validation of existing models and further development to create a set of robust equations which can accurately predict the outcome of the process. Shuangning et al. (2005), studied the plasma pyrolysis of wheat straw and corn stalk particles in a polydisperse range of 117–173 μm in a plasma heated entrained flow reactor using Ar plasma in the range of 750–950 K aiming to estimate kinetic parameters using the volumetric model, which is the most simplified technique that assumes uniform reaction rates throughout the particle volume. The first order rate of reaction is given by Equation (30), where W and W ∞ are present and final percentages of pulverized biomass.

(30) d W d t = A · W ∞ − W   e − E a RT

The reported frequency factors and activation energies for wheat straw and corn stalk were 1,028 s⁻¹, 1,013 s⁻¹ and 31.51 kJ mol⁻¹, 33.74 kJ mol⁻¹ respectively. In another study Tu et al. (2009), fitted the kinetic study data for rice straw plasma pyrolysis in an integrated core model given in Equation (31):

(31) d M STF d t − t ST = − k · M STF n

where k , is the reaction rate constant, n is the reaction order, M STF denotes the residual mass fraction after reaching t ST , which is the time taken to reach the set reaction temperature. The experiments were performed in a tubular reactor with a top mounted RF induced plasma torch operated between 0.35 and 0.88 kW to provide a temperature range of 740–900 K. Conversion of rice straw during plasma pyrolysis was constrained to a maximum conversion of 60 %, by operating the reactor in a differential mode to calculate the system’s average reaction rate. This was done to calculate the true rate of reaction which is dependent on concentration, which drops as the reaction progresses, therefore in order to state that the rate of reaction is constant throughout the reactor, an average is calculated by maintaining low conversion values (Levenspiel 1999). Values of Arrhenius constant and activation energy were obtained as 5,759.5 s⁻¹ and 74.29 kJ mol⁻¹, respectively.

These models follow the general misconception that the thermal degradation reactions occurring below 1,000 °C are kinetically controlled and thus the heat and mass transfer resistances are ignored (Mahinpey and Gomez 2016). Moreover, the values of kinetic parameters calculated through these models are subjective to the ultimate composition of the feed, this questions the strength of the model over its lack of consistency. Therefore, a model governing the rate of reaction of either C or H in the feed (as they are the largest gravimetric and molar constituent of the hydrocarbonaceous solid feed) can prove to be more robust.

For gathering preliminary experimental data for kinetic model development, operating with simpler systems such as coke, individual plastics, and coal as feed, rather than complex systems like biomass, and mixed waste can yield more accurate results as compositional variable are eliminated. Kholiavchenko et al. (2019), proposed a model to determine the kinetic constants for steam plasma gasification of wet coal. By including the thermally varying mutual diffusion coefficient of the reactant and product gases the model as presented in Equation (32) accounts for the resistance to mass transfer.

(32) G C = α d R T g N 3 · p 2 1 + N 3 + N 1 · p 1 + 0.5 p 5 1 + N 1

where, α d represents the mass transfer coefficient calculated using the Nusselt diffusion criterion, N i is the dimensionless ratio of kinetic rate constant and mass transfer coefficient for the particular heterogeneous reaction, p i is the partial pressure of the gaseous components, T g denotes the reaction temperature, and G c is the rate of gasification of carbon (kmol/m²s).

The assumption that the thermophysical properties of feed material remain constant throughout the heating time can be deemed appropriate for smaller particles of range 50–200 μm as considered in the study but for larger particles properties like thermal conductivity will change as the reaction progresses and products like ash and volatiles are formed. The second assumption made that the particles maintain a constant size during the heating time contradicts with general conception that a carbon based feed undergoing thermal degradation follows the mechanism of shrinking core model with changing particle size (Levenspiel 1999; Peters et al. 2012; Zhang et al. 2013). To test the theory a comparison was made in Figure 11 between the proposed model and shrinking core models for constant and changing size particles for the total time of reaction based on the data reported in the study. Equations (33–35) describe the models in discussion, here τ is the total reaction time, M c is molecular mass of carbon = 12 kg/kmol, b denotes the stoichiometric coefficient = 1 in this case, D is diffusivity (the values of mutual diffusion coefficients were taken from (Massman 1998)), C og concentration of oxidizer gas, and k ″ is the reaction rate constant.

(33) τ = ρ p · D p 2 M c G c   → Kholiavchenko et al .   2019

(34) τ = ρ p · D p 2 24 b D C og + ρ p · D p 2 b k ″ C og   →  Shrinking core model − constant particle size

(35) τ = ρ p · D p 2 b k ″ C og   →  Shrinking core model − changing particle size

Figure 11:

Comparison of total time of reaction for coal particles undergoing steam plasma gasification calculated with the proposed model (Kholiavchenko et al. 2019) and shrinking core models.

For this scenario the resistance offered by the gas–solid film is neglected considering the small size (500 μm) and high gas velocity. The plot displays constant size model has good agreement with the proposed model, whereas in the changing size model diffusional resistance does not play a role as the surface is consumed during the reaction and hence shorter reaction times are obtained due to fast reactions. This comparison attempt highlights that mass transfer resistance attributed to the volatile sheath layer, however the diffusivity coefficient values must be different for volatile sheath and solid carbon unlike in this case. Therefore, is a need of an upgraded model that can account for the radially varying mass transfer resistances. However, clarity over the robustness of the model will only be obtained after experimental validation.

To determine the highest rate of reaction it is important to study the intrinsic kinetics, which are calculated by marginalizing heat and diffusional resistances, and account only the molecular level reactions. Although, it cannot be said that mass transfer resistances are eliminated, and that is why the net rate is a combination of both intrinsic and diffusion limited rate as explained in the work of (Horton et al. 2016), where the authors have defined the net rate for O₂ plasma gasification of coke as:

(36) r net = r intrinsic 1 + D a

here, r intrinsic is the intrinsic rate of reaction [given by Equation (37)], and D a is the ratio of intrinsic and diffusion limited rate.

(37) r intrinsic = y O 2 A O 2 e − E O 2 RT

(38) D a = r instrinsic r diffusion   limited

(39) r diffusion   limited = k O 2 y O 2 SA p V p

y , O 2 , A O 2 , E O 2 , k O 2 , V p , and SA p denote the mass fraction of oxygen in gas, concentration, pre-exponential factor, activation energy, particle volume, and surface area of the particle. For this model the surface area to volume of particle ratio is a function of time and extent of carbon conversion which accounts for the changing particle size and can be deemed apt for feed with high ash content by overcoming the shortcomings of shrinking core model to account for partially reacting core. From Figure 12 it can be inferred that diffusional limitations do become significant at temperatures higher than 1400 K. The diffusion limited regions can be shifted based on the intrinsic rate of reaction which will vary according to the plasma gas and oxidizer used.

Figure 12:

Comparison of reaction rates for (a) 4 in. and (b) 6 in. coke particles undergoing O₂ plasma gasification. Reproduced with kind permission from ref. (Horton et al. 2016). Copyright 2016, ACS.

6.1 Step 1: understanding reactions and formulation of reaction kinetics

Understanding the key reactions occurring during plasma treatment and their mechanisms is important. The primary step is the film diffusion of plasma-forming and carrier gas followed by the vaporization of plastic waste surface, which is the thermal decomposition of macromolecules into smaller hydrocarbon vapors. This is followed by the diffusion of gases through the sheath layer to the core, the final step in the process is the gas phase reactions described in Equations (7)–(16). The kinetics of gas phase reaction are covered largely in the literature (Bustamante et al. 2005; Gómez-Barea and Leckner 2010; Jones and Lindstedt 1988; Westbrook and Dryer 1981), to identify the rate limiting steps and estimate the kinetic parameters for plasma processing the following strategy is proposed:

1. Reactor selection: a tubular semi-batch reactor in a vertical configuration is considered with a top-mounted plasma torch instead of the plasma plume immersed in the bed for better control of the reaction and observing the flow of the reaction, and figuring out hydrodynamic conditions.

2. Marginalizing heat transfer resistances: as the plasma plume impinges on the waste bed a temperature gradient develops along the length of the waste bed due to the changing thermal conductivity of the bed undergoing thermal degradation with time and the increasing loss of heat from plasma to reactor environment as the distance from plasma origin to the point of contact grows, this affects rate of reaction across the bed, therefore the feed quantity selected should be small enough to occupy a differential length (dx) that can be assumed to behave isothermally yet significant enough to produce measurable quantities of gas and solid products.

3. Marginalizing mass transfer resistances: increasing the gas velocity reduces the thickness of film boundary and therefore lowers the external mass transfer resistance. In context of waste plasma processing, the velocities of the plasma-forming gas and oxidizing/carrier gas at the film boundary can be considered approximately equal. Moreover, the velocity of plasma emitted from a DC torch can reach up to 15–40 m/s (Chau et al. 2011), making film diffusion resistance negligible. To identify intraparticle mass transfer limitations the waste particle size must be varied and the rate of reaction can be monitored by measuring the carbon moles in the gas output. An increase in reaction rate with decreasing particle size indicates mass transfer limitations, whereas a constant rate suggests a shift to the kinetically controlled regime. It can also be assumed that the reaction of other species (H, N, S, O) from the particle surface occurs in proportion to carbon.

4. Effect of concentration of oxidizing gas and carbon: the rate of reaction is concentration dependent. The concentration of oxidizing gas can vary due to the formation of gaseous products over time and the reactor’s hydrodynamics. Therefore, for the assumption to stand true that a constant reaction rate exists throughout, the reactor must be operated in differential mode for small conversions limited to 50 %. Experimental runs should be conducted where the concentration of the oxidizer introduced either through the forming gas or distributed injection are varied with the help of an inert gas (mostly Ar). These experiments yield an average reaction rate corresponding to the mean oxidizer concentration in the reactor. By performing a series of such runs under a constant temperature and flow rate, a set of rate-concentration data can be generated and analyzed to derive the rate equation. Similarly, this method can be put into practice for concentration of carbon inside the reactor by varying the ultimate composition of the feed or using high melting point inerts of similar size such as powdered high alumina bricks.

5. Estimation of intrinsic kinetics: to deduce the kinetic parameters effect of temperature on degree of solid conversion must be observed. The data points can then be analyzed using the logarithmic form of the Arrhenius equation. The rate of thermal degradation increases significantly with rise in temperature given the endothermicity of the overall process. In plasma processing, however, the temperatures experienced by the feed can reach >3000 K depending on torch position and reactor hydrodynamics, under such extreme conditions the exponential term in Arrhenius equation asymptotically tends toward unity, indicating that the activation energy is much lower than the thermal energy provided therefore resulting in insignificant rise in the reaction rate despite the increment in temperature (Kholiavchenko et al. 2019).

6.2 Step 2: understanding the relationship between fluid mechanics and design objective

The primary objective of involving thermal plasma for processing plastic or solid waste, is to obtain high conversion rates, and the complete destruction of waste without generating harmful gaseous emissions and corrosive tar. The secondary goal is to utilize the produced syngas for energy recovery. To achieve these objectives, it is obvious that the feed has enough residence time in the high temperature environment, which requires establishing homogeneity inside the reactor. The velocity of gases, flow patterns, and temperature gradients in waste plasma processing reactors differ significantly from conventional pyrolysis and gasification reactors.

Insights from existing literatures and available description of commercialized technologies highlight three types of plasma reactor designs with unique flow characteristics: (i) tubular reactors with an extended freeboard zone to provide enough volume for gas phase reactions and provide better control over material flow, therefore are suitable for experimental investigations as demonstrated in several studies (Astashov et al. 2018; Hu et al. 2023; Tang et al. 2003; Zhang et al. 2013); (ii) compact plasma reaction chambers that are beneficial in establishing isothermal environment through heat circulation inside the reactor (Agon et al. 2016; Chu et al. 2023; Hlina et al. 2014; Hrabovsky et al. 2010; Hu et al. 2023; Mączka et al. 2013; Punčochář et al. 2012); (iii) plasma shaft furnaces are generally preferred for handling heterogeneous waste on a commercial scale and are subjected to large temperature gradients (Alekseenko et al. 2019; Anshakov et al. 2019a,b, 2024; Byun et al. 2010, 2012). The flow of the materials is also governed by the bed arrangement inside the reactor. Tang et al. (2013) in their review discuss in detail about various bed arrangements in waste plasma processing reactors, comparing their ease of operation and effects on degree of waste conversion. Additionally, the position of plasma torch inside the reactor plays a crucial role in determining the flow regime. Reactors with a top-mounted plasma torch facilitate complete waste conversion by allowing the gas, formed during the breakdown of solid materials, to re-interact with the high-temperature plasma before exiting the reactor (Hlina et al. 2014; Hrabovsky et al. 2010; Tang and Huang 2007; Tang et al. 2003). However, the plasma loses a significant amount of heat through radiation due to the distance between feed and plasma origin point, and results in temperature drop from 5,000 – 10,000 K to 2000 K. Another widely adapted configuration is immersing the plasma jet inside the waste bed (Mączka et al. 2013; Punčochář et al. 2012), here the waste feed experiences extremely high temperatures resulting in a higher rate of reaction can, However, the high velocity of the thermal plasma can entrain unconverted waste in the upper layers, reducing overall efficiency. To overcome this issue Byun et al. (2010) designed plasma chambers with recirculation zones. In their design, the plasma torch is aligned at a 30° angle, leveraging centrifugal force to enhance heat distribution and ensure sufficient heating of the feed, thereby minimizing waste entrainment and improving conversion efficiency.

Furthermore, visualizing the movement of particles and gas flow inside the reactor system can be done with CFD modelling studies and through empirical methods such as particle image velocimetry, or high-speed imaging. Visualization of plasma arc generated through DC and AC torches can also be done with the help of charge-coupled device (CCD) cameras working at shutter speeds of up to 1/8,000 s (Rutberg et al. 2013). Videography of plasma arcs can help in determining the arc velocity to calculate residence times and arc diameter, which can aid in estimating heat distribution and accordingly optimize the reactor.

6.3 Step 3: optimization of parameters

Optimization of plasma waste processing involves managing multiple variables, as shown in Figure 13. Independent factors such as feed rate, gas flow rate, torch power, and reactor geometry influence flow dynamics and residence time, which in turn affect mass and heat transfer, and reaction rates. Additionally, adjusting the torch position and bed arrangement must be done as well to optimize flow dynamics, and factors like particle size and moisture content must be controlled to improve heat transfer efficiency.

Figure 13:

Mind-map illustrating interdependencies between parameters for optimization of waste plasma processing technology.

Dependent variables like CAPEX and OPEX are shaped by reactor volume and electrode erosion, which can be minimized by selecting suitable forming gases and electrode materials. High-quality syngas output requires careful tuning of oxidizing and forming gas compositions based on feed characteristics. Advanced tools like CFD and real-time monitoring can support system optimization, improving performance, reducing costs, and enhancing environmental sustainability.

6.4 Step 4: implementation

Before scaling up a plasma processing plant, validating the design through pilot tests is essential to confirm assumptions. Key components like plasma torches and reactors must be tested under full load. Large-scale operations also demand advanced process control and automation systems that can adjust plasma power, gas flow rates, and feed rates in real-time, ensuring stable performance and safety. Automations like, automatic shutdowns and gas venting during abnormal conditions, and helps manage fluctuations in feedstock quality (Sharma 2017). Safety remains a priority due to the high temperatures and pressures involved, so installing pressure relief valves, explosion-proof designs, and fire suppression systems is essential (Hall 2012). Furthermore, environmental controls must be in place, including emissions control systems like scrubbers and bag filters, to meet regulatory standards. Additionally, securing a reliable feedstock supply chain is crucial for continuous operation, requiring pre-processing facilities and careful logistics for sourcing, transporting, and storing plastic waste. Finally, waste management should focus on efficient slag disposal, syngas utilization, and profitability assessment through techno-economic analysis (TEA), discussed further in the next section.

8.1 Past and present installations

The primary drivers for adopting plasma technology in the waste management sector are regulatory pressures for mitigation of hazardous waste, reduction of waste volume, minimization of harmful emissions without the need of extensive sorting and cleaning of the waste, and to some part economic factors such as gate fee. These are also the reasons why plasma gasification saw widespread commercialization globally compared to plasma pyrolysis and incineration. In 1990 the Westinghouse Plasma Corporation (WPC) commissioned its first 48 TPD plasma gasification testing facility in Madison, USA where they tested over 100 feedstocks and for their potential to produce syngas. Based on their water cooled non-transferred DC plasma torch technology, WPC, as a subsidiary of the Alter NRG group established numerous waste to energy plants across Asia with the most recent ones in Wuhan and Shanghai, China. The European market share is significantly captured by companies like Europlasma and Tetronics. Europlasma has setup plasma vitrification and gasification plants with Inertam and CHO-Power respectively in Morcenx, France. Tetronics, based in the UK, has commercialized transferred DC arc graphite-electrode technology in collaboration with Advanced Plasma Power, establishing syngas production plants utilizing waste plastics and biomass. As previously mentioned, initially plasma processing facilities were primarily focused on melting hazardous wastes, including medical waste, polychlorinated biphenyls, radon-based paints, nuclear waste, and asbestos. InEnTec^®’s plasma enhanced melting (PEM) technology was instrumental in addressing these challenges throughout the 1990s (Achinas 2019). However, over time, a clear shift has occurred in the strategic intent of plasma technology deployment. Emerging installations are increasingly focused not only on safe disposal but also on resource recovery, particularly hydrogen and syngas production for energy or chemical synthesis. This trend is driven by technological maturation and the growing economic viability of synthetic fuels. In 2012, South Korea-based start-up Green Science Platech (GS Platech) commissioned a 10 TPD MSW processing plant in Cheongsong-gun city, capable of producing 20 Nm³/hr of H₂ (Byun et al. 2012; Sesotyo et al. 2019). In 2021, GS Platech also launched the world’s first microwave plasma-based gasification plant in Taebaek, South Korea, which produces 3 MW of electric power annually from plastic waste and wooden chips (Pulse 2024). Another emerging start-up, SGH2, has established a 3,800 tpy H₂ plant using its proprietary Solena Plasma Enhanced Gasification (SPEG) technology to process waste biomass (SGH2 Energy 2024). A summary of other commercialization attempts is provided in Table 7.

8.2 Challenges with commercialization

The four leading stumbling blocks for large scale commercialization of plasma processing of waste are (i) regulatory frameworks, (ii) market agreement and acceptance, (iii) technological maturity, and (iv) social acceptance. Regulations and societal aspects are linked and present the most significant challenge in establishing a plasma processing plant for waste treatment. The construction of plasma gasification or pyrolysis facilities, which include thermal reactors operating at extremely high temperatures and high-caloric hazardous waste storage units, often raises significant public concerns about potential harmful air emissions, fire risks, and toxic chemical leachate; such facilities have faced cancellation in the past due to economic issues, public mistrust, and safety concerns. Such concerns necessitate stringent regulations from institutions like the World Health Organization (WHO) and Central Pollution Control Board (CPCB) to ensure the safe operation of plasma technology (Das et al. 2020). An assessment carried out by Sampathraju and Mansuri 2019 highlighted that air pollutants from plasma disposal plant for biomedical waste were well below than conventional incineration and commit to the limits set by CPCB. Furthermore, ensuring the safety and stability of plasma plants to prevent accidents becomes the responsibility of local authorities who must conduct rigorous safety and environmental audits, and facilitate community feedback that can also help in advocacy of the technology to set up more plants. Additionally, regulatory requirements for operating permits for energy-intensive plants and agreements for the treatment of hazardous wastes are essential for plasma processing companies. These regulations not only ensure compliance with environmental standards but also impact the financial viability and profitability of the business. Thus, navigating the regulatory landscape and addressing societal concerns are fundamental for the successful implementation and growth of plasma processing technology in waste management.

Challenges in commercialization also occur when complete technology readiness is not achieved (Munir et al. 2019). Technical failures in plasma gasification and pyrolysis plants often arise due to aggressive scale-ups, when technologies that are only at a demonstration level are scaled up to 10 folds it leads to problems such as formation of cold spots. This was the situation faced during operation of the Hitachi Metals plasma gasification plant in Utashinai, Japan where the large bottom diameter led to uneven heat distribution resulting in lower conversion, this ultimately led to entrainment of particulates with the produced syn-gas and attacking the refractory lining. This case highlights the importance of prudential reactor design in terms of ensuring proper heat transfer, residence time calculations, and high temperature resistive refractory linings. Similarly, the Teesside, England plasma gasification project, which aimed to construct two 50 MW plants, was abandoned due to the technology’s failure at scale. Besides technological issues, which are solvable, strategic planning for raw material supply is crucial. The closure of the Hitachi plasma plant in Eco-valley, Utashinai, was partly due to increased recycling practices that led to a shortage of MSW feed, highlighting the need for agreements with local authorities to secure a consistent feedstock supply (Hrabovsky et al. 2022.). This makes the location of the plant an important element in the successful establishment of plasma gasification plants for waste treatment. Having a plasma gasification plant established in a location with high waste generation capacity to available landmass ratio will be beneficial as it ensures consistent feed supply. Alternatively, portable or on-site plasma processing units are being explored as flexible solutions to address the dual challenges of waste accessibility and the limited processing capacities often associated with large-scale, centralized plasma facilities.