Hydrothermal liquefaction (HTL) is a promising technology with huge potential for the chemical recycling of plastic waste. Herein, a two-stage sequential HTL process is proposed for the conversion of acrylonitrile-butadiene-styrene (ABS) plastic to produce energy-dense oils with minimal nitrogen content. In the first step, ABS is denitrogenated using subcritical water (250-350 °C) and an alkaline potassium hydroxide (KOH) catalyst to produce solid product 1 (SP1) and aqueous phase 1 (AP1). A 95% denitrogenation efficiency of ABS is achieved at 325 °C in 20 min, with a catalyst load of 0.30 grams of KOH per gram of ABS. Under these conditions, the nitrile bonds in ABS are hydrolyzed to form a potassium-ABS salt complex that migrates to AP1. This AP1 is then neutralized with hydrochloric acid (HCl) to produce the denitrogenated solid product 2 (SP2). The nitrogen from ABS is easily removed as ammonia (NH3) by evaporation. SP1 and SP2 from stage 1 are then liquefied at the second hydrothermal step (375-425 °C, 250 bar) to produce low-nitrogen-containing oils with high heating values up to 41.6 MJ·kg-1. The oils primarily consist of mono-, di-, and poly-aromatic compounds, with boiling ranges similar to those of kerosene and vacuum gas oil. The gases from the supercritical HTL contain C2-C5 hydrocarbons. The results indicate that the two-stage sequential HTL is successful in removing nitrogen from the ABS and producing nitrogen-free aromatics, with much greater efficiency than the single-stage supercritical HTL of ABS.
Subramanian Harisankar, Juliano Souza dos Passos, Sofie Klara Gissel Skibsted, Esben D amgaard, Patrick Biller.
Sequential Denitrogenation and Liquefaction of Acrylonitrile-Butadiene-Styrene via Two-Stage Hydrothermal Liquefaction Using Homogeneous Catalysts.
Engineering, 2026, 58 (3) : 114-124 DOI:10.1016/j.eng.2025.12.037
Plastics are among the most extensively produced and discarded materials in human history, with substantial environmental leakage stemming from poorly designed waste management systems [1-3]. Acrylonitrile-butadiene-styrene (ABS) is a notable synthetic polymer that represents approximately 2% of the global plastics market [4]. It is widely used in the automotive industry, consumer electronics, and household appliances, owing to its high impact strength, low weight, and chemical resistance. Although ABS has a slightly higher recycling rate than other polymers, the development of more efficient recycling strategies is still essential to mitigate landfill waste and increase the sustainable use of ABS [5]. Mechanical recycling is viable for pure ABS streams, but most waste lacks sufficient purity, making chemical recycling a promising alternative [4,6]. However, ABS chemical recycling via depolymerization poses challenges, primarily due to the polymer’s nitrogen content from its acrylonitrile component. For instance, thermochemical processing via pyrolysis results in oil products with 6.5-7.1 weight percent (wt%) nitrogen [7]. This relatively high nitrogen content presents significant challenges upon the use of the obtained oil, with complex and costly hydrotreatment operations required to eliminate the nitrogen. Additionally, nitrogen-rich polymers emit toxic gases during pyrolysis, such as hydrogen cyanide (HCN), posing further risk to human health [8].
Hydrothermal liquefaction (HTL) has emerged as a promising chemical recycling method for plastics [9,10]. Subcritical HTL, conducted below the critical point of water (373 °C and 220 bar, 1 bar = 105 Pa), has successfully depolymerized condensation polymers into their original monomers, including final post-user products such as textiles [11] and polyethylene terephthalate [12]. Addition polymers such as polyethylene (PE), polypropylene (PP), and ABS require higher temperatures to break their backbone carbon-carbon chains, necessitating supercritical HTL [13, 14]. The process conditions and catalyst significantly influence HTL efficiency and the composition of its products. Depending on the type of polymer, HTL may yield a synthetic oil rich in a mixture of chemicals—rather than pure monomers—that can be further refined for circular material recovery [15]. For instance, HTL of polycarbonate plastic at 350 °C leads to an aromatic-rich synthetic oil, whereas treating polycarbonate at 150 °C with a potassium hydroxide (KOH) catalyst in water can lead to the production of its monomer, bisphenol-A, with > 90% recovery [16, 17]. Generally, a one-stage HTL process is subpar for treating ABS, since the resulting oil retains the nitrogen heteroatom, reducing the oil quality. Recently, a two-stage HTL strategy for polyvinyl chloride (PVC) resulted in the removal of 99% of chlorine in the first stage, while approximately 94% of the carbon (C) was retained [18]. As such, a similar strategy is hypothesized to be effective in removing nitrogen from ABS, which is a heteroatom-containing polymer like PVC.
This experimental study presents an innovative approach to ABS chemical recycling via a two-step hydrothermal process strategy, where a homogeneous catalyst (KOH) in subcritical water is used in the first step to hydrolyze the nitrile groups of ABS and effectively eliminate nitrogen from the ABS structure. An alternative and straightforward way to simultaneously denitrogenate and decompose ABS to crude oil is provided, instead of the conventional approach of producing crude oil from plastics first, followed by hydrotreatment with heterogeneous catalysts and hydrogen (H2). The effects of the temperature and catalyst load on the denitrogenation efficiency are also investigated. The solid materials from the first step are used as the feedstock for the second step, in which supercritical water is used to convert the resulting solid material into a valuable oil phase containing several aromatic platform molecules with very low nitrogen content. In addition, a thorough compositional analysis of the oils is carried out to illustrate the potential to produce aromatics from ABS via the two-stage HTL process. This approach eliminates the requirement of drying the ABS as well as the denitrogenated solids, as supercritical HTL benefits from the hygroscopic nature of these solids. The results from the two-stage sequential HTL process are also compared with those from a single-stage supercritical HTL of ABS to provide a holistic understanding. The two-stage HTL is intended to simplify the production of nitrogen-free aromatic chemicals from nitrogen-containing plastics such as ABS; it also offers a sustainable pathway for converting plastic waste into valuable chemical resources, thereby advancing circular economy initiatives.
2. Materials and methods
2.1. Materials
Clear ABS pellets (trade name: Terluran GP-22) were acquired from BASF (Germany). Colored ABS samples (blue, orange, brown, light green, and dark green) were sourced from LEGO bricks through Aarhus University’s material sample database (Denmark) [19]. KOH pellets were used as received from Sigma-Aldrich (Denmark). The HCl acid used in the neutralization experiments was procured from Merck (Denmark).
2.2. Two-stage hydrothermal processing of ABS
In a typical stage 1 (subcritical) experiment, 1 g of ABS, 10 g of water, and 0.05-0.40 g of KOH were loaded into 20 mL bomb-type custom-made high-pressure reactors made with 1″ Hy-Lok Fittings. After loading, the reactors were submerged in a pre-heated sand bath, and the reaction temperature setpoints (250-350 °C) were achieved within 4 min. The retention (reaction) times for the reactions were calculated from the moment the reactors were immersed in the fluidized sand bath. The retention time used in all the subcritical reactions was 20 min, after which the reactors were quenched in a water bath. Gas was vented and the gas yields were determined gravimetrically. Liquids were decanted to a 15 mL centrifuge tube and centrifuged, with a 1 mL aliquot of the supernatant recovered for analysis. The supernatant was filtered together with the solids recovered directly from the reactor. In this step, solid product 1 (SP1) was recovered and dried overnight at 105 °C in a forced ventilation oven. The liquid obtained from this filtration, aqueous phase 1 (AP1), was left at 30 °C overnight to evaporate the ammonia (NH3) generated during the nitrile hydrolysis. The now ammonia-free AP1 was acidified and neutralized using diluted HCl; this precipitated the soluble polymer, yielding the solid product 2 (SP2). The precipitated SP2 was then filtered and washed with deionized water. Overall, SP1, SP2, and aqueous phase 2 (AP2) are the product phases from subcritical (stage 1) HTL.
SP1 and SP2 from the subcritical HTL of the clear ABS at 325 °C for 20 min were subsequently liquefied under supercritical conditions (stage 2). At first, SP1 and SP2 were liquefied individually to determine their decomposition characteristics. Smaller 10 mL reactors were used to keep the solid consumption per batch at practical levels. A very similar procedure to the subcritical HTL step was used for the supercritical experiments. Temperatures and pressures were set at supercritical conditions (375-425 °C in 25 °C steps) with water loading proportional to 250 bar, estimated using the National Institute of Standards and Technology (NIST)’s "Thermophysical Properties of Fluid Systems" database [20]. Plastic loading was kept at 30 wt% for most supercritical experiments, while the retention time was varied at 30, 60, and 90 min. For some supercritical experiments with SP2, the plastic loading was maintained at 45 and 60 wt%. Supercritical HTL of ABS led to the formation of an oil phase, gases, and aqueous phase 3 (AP3). Finally, SP1 and SP2 were also simultaneously co-liquefied under supercritical conditions, with and without the neutralization of KOH step, to determine the optimum operating methodology while maintaining high oil yields and recovery. For select experiments, the reactor heads were connected to a long, stainless-steel tube attached to a valve, which was used to collect gases for analysis. To assess the efficiency of the two-stage HTL, a control experiment in which clear ABS was directly liquefied under supercritical conditions was also performed. Fig. 1 illustrates the procedure of the two-stage HTL. The mass balances and yields of the product phases were calculated as follows:
where mABS, mSP1, mSP2, mAP2, mAP3, mgases, moil, and mprod are the mass of ABS, SP1, SP2, AP2, AP3, gases, oil phase, and product phases, respectively; Yprod is the yield of all product phases.
As shown in Eqs. (1), (2), (3), (4), the initial mass of the ABS plastic was always used as the basis to calculate the yield of all product phases from both the single-stage and two-stage HTL experiments.
2.3. Characterization techniques
Elemental analysis of the solids and oils was conducted using the combustion method with an Elementar Vario Macro Cube (Langenselbold, Germany). Then, the high heating values (HHVs) of the solids and oils were determined using the Channiwala-Parikh correlation [21]. The ABS samples contained 83-86 wt% carbon, 8-9 wt% hydrogen, 4-6 wt% nitrogen, < 0.1 wt% sulfur, < 3 wt% oxygen, and < 2 wt% ash content, with an average HHV of 39.7 MJ·kg-1 (Table S1 in Appendix A). After stage 1, the degree of denitrogenation (DN) was calculated using Eq. (5). The energy recovery (Ere) from the oils from stage 2 was calculated using Eq. (6). The sensible heat (Qsensible) required for the HTL process was calculated using Eq. (7).
where [N]i and [N]f represent the initial and final nitrogen concentrations in the samples, respectively; HHVoil and HHVABS are the HHVs of oil and ABS, respectively; Yoil is the yield of oil phase; Cp,ABS is the specific heat of ABS; ΔT is the temperature difference; mwater is the mass of water; and ΔHwater is the enthalpy change of water.
For aqueous phases, the total organic carbon (TOC) and total nitrogen (TN) were quantified using a Skalar FORMACS HT-I TOC/TN analyzer (Breda, the Netherlands). Fourier-transform infrared (FTIR) spectroscopy was performed using a Bruker Alpha FTIR Platinum diamond with attenuated total reflectance (ATR) accessory. Simulated distillation (SimDis) was carried out using an Agilent 6890 gas chromatograph (GC), equipped with a high-temperature programmable temperature vaporizer (HT PTV), a DB-HT SimDis capillary column (55 m (length) × 530 μm (inner diameter) × 0.15 μm (thickness)), and a flame ionization detector (FID), according to American Society for Testing and Materials (ASTM) D2887 standards, incorporating Polywax 655 as a reference for high carbon numbers (up to C100). Proton nuclear magnetic resonance (1H NMR) and heteronuclear single quantum coherence (HSQC) spectroscopy were conducted on a Bruker Avance NEO 600 MHz spectrometer (Bruker Biospin GmbH, Germany) with a 5 mm 1H-optimized double-resonance broadband probe. The organic compositions of the oils and gases were determined using a gas chromatograph-mass spectrometer (GC/MS; Agilent Technologies 7890A) equipped with a VF-5ms column. The oil samples were diluted using dichloromethane and then injected into the GC/MS at an injection temperature of 280 °C, with helium (He) as the carrier gas. The column was initially maintained at 40 °C for 5 min; it was then ramped up to 300 °C at a rate of 5 °C·min-1 and subsequently maintained at this condition for 15 min. The same GC/MS equipment was used to identify the gas composition. For this purpose, the injection temperature was set at 150 °C. The oven was initially maintained at 30 °C for 8 min, with no solvent delay, followed by a ramp to 240 °C at 20 °C·min-1; it was then maintained at this condition for 1 min. The organic compounds from all the oils and gas samples were semi-quantitatively analyzed using their peak area percentages from their total ion chromatograms.
3. Results and analysis
3.1. Subcritical liquefaction of colored ABS
3.1.1. Overview of the subcritical liquefaction experiments
Following the work-up procedure illustrated in Fig. S1 in Appendix A, the subcritical liquefaction experiments yielded SP1, SP2, and AP2. The physical and chemical properties of ABS vary depending on the monomeric unit lengths and additives used. For example, high-impact ABS contains rubber modifiers, fillers, and bromine-based flame retardants, while high-flow ABS has increased styrene content to increase fluidity and moldability [22]. These variations in the structure and property of ABS influence both the yield and composition of products from the HTL of ABS. As an illustration, six different colors of ABS were subjected to subcritical HTL at 325 °C, 20 min, and 0.30 grams of KOH per gram of ABS (gKOH∙gABS-1) (Fig. S2 in Appendix A). Across all the samples, SP2 was the major product fraction, followed by SP1 (Fig. S2). SP2 is composed of alkali-soluble polymer chains and had to be neutralized with acid to precipitate it to the solid phase. A major fraction of the nitrogen from all the ABS samples was removed as NH3 or migrated to AP2 (Fig. S3 in Appendix A). The DN from SP1 and SP2 was greatest for the clear ABS (∼95.9%) and lowest for the orange ABS (∼63.3%). The AP2s from the ABS samples showed visual differences, suggesting that many ABS additives were also partitioned into the aqueous phase. However, a detailed analysis was not performed on the AP2s from differently colored ABS plastics, as that was beyond the scope of the present study. The remaining discussion focuses only on the clear ABS sample and the products derived from its liquefaction.
As shown in Figs. 2(a) and (b), the conversion of ABS improved with an increase in temperature and catalyst load. Under the best conditions, approximately 70 wt% of the carbon input was recovered as either SP1 or SP2, with SP2 making up about two-thirds of the total solids obtained. Regarding nitrogen recoveries, much of the nitrogen was found in AP2, irrespective of KOH concentration (Fig. 2(c)). Nevertheless, higher KOH concentrations resulted in minimal nitrogen retention in SP1 and SP2. From the H/C and N/C ratios, it was estimated that the average molar ratio of the monomeric groups in the clear ABS—that is, acrylonitrile:butadiene:styrene for the ABS polymer used in this study—was approximately 2:6:1. The stoichiometric ratio of KOH to nitrogen in the clear ABS sample, assuming each nitrile group formed a carboxylic salt upon denitrogenation, was approximately 0.28 gKOH⋅gABS-1. This is corroborated by Fig. 2(c), which shows that, at concentrations of 0.30 and 0.40 gKOH⋅gABS-1, the nitrogen levels in the SP1 and SP2 were minimal, indicating that stoichiometric amounts of KOH are advisable for the near-complete elimination of nitrogen. Notably, approximately 65 wt% of nitrogen was eliminated from the solids using only 0.05 gKOH⋅gABS-1, suggesting that the nitrogen—which is released as ammonia—initiates an autocatalytic denitrogenation process. In this process, the alkaline medium created by the ammonia further facilitates the denitrogenation of the ABS. Fig. 2(d) demonstrates that 325 and 350 °C are the optimal temperatures for nitrogen removal from the solid phases.
3.1.2. Compositions of SP1 and SP2
Figs. 3(a) and (b) demonstrate a clear trend: Higher KOH concentrations and temperatures result in a greater DN. No denitrogenation occurs in the absence of KOH. Additionally, SP1 and SP2 exhibit distinct denitrogenation behaviors. It is clear that SP2 is more easily denitrogenated. As SP2 is soluble in the alkali water, the nitrile groups are easily accessible and thus react faster than they would in a solid state, as is the case in SP1. Even a low KOH concentration of 0.05 gKOH⋅gABS-1 resulted in over 50% denitrogenation of SP2, underscoring the significance of the autocatalytic denitrogenation process. Nevertheless, the highest denitrogenation levels were obtained with stoichiometric KOH amounts at 325 °C.
It is clear that the alkali-water-soluble part of the solids (SP2) is more easily denitrogenated. This is illustrated well in Fig. 3(b), which shows that SP1 has lower denitrogenation than SP2 at all temperatures between 250 and 300 °C. Because SP2 is soluble in the alkali water, the nitrile groups are easily accessible and thus react faster than they would in a solid state (as in the case for SP1). In comparison with the KOH- and polyethylene glycol (PEG)-catalyzed denitrogenation of ABS, the hydrothermal procedure shown here requires higher temperatures but much lower residence times [23]. Moreover, the hydrothermal method excludes the use of solvents other than water itself.
For example, the denitrogenation approach reported by Zhou et al. [24] using PEG required a temperature of 160 °C for 2 h, achieving only 52.7% denitrogenation. The same conditions and setup were reported by Du et al. [23], this time using both KOH and NaOH; while these researchers achieved nearly 100% denitrogenation, their approach involved a temperature of 200 °C, high loads of NaOH, and a retention time of 8 h. In comparison, using the proposed hydrothermal method, consistent denitrogenation of over 95% is achieved within 20 min of residence time under moderate KOH loads. The hydrolysis of acetonitrile and related nitrile compounds to carboxylic acids under alkaline hydrothermal conditions has also been well documented in the literature [25].
As shown in the modified van Krevelen diagram in Fig. 4, SP1 and SP2 undergo different elemental changes with increasing KOH load. The rate of change of H/C with respect to N/C has a slope of -5.4 H/N for SP1, while it is nearly 0 H/N for SP2. This result indicates that, for every N atom eliminated from ABS, SP1 incorporates at least five H atoms, whereas SP2 maintains the same number of H atoms. Moreover, SP1 demonstrated a higher H/C ratio under all conditions in comparison with SP2. This result implies that the polymers in SP1 contained more butadiene units, which inherently possess a higher H/C ratio than styrene units; in addition, butadiene’s non-polar behavior makes it water-insoluble. On the other hand, SP2 consisted of more styrene units, which included the carboxylic salt functional group (from the nitrile hydrolysis), making them water-soluble.
The O/C ratio of SP2 was consistently higher than those of the SP1 solids (Fig. S4 in Appendix A), indicating the presence of carboxylic chains in SP2. The presence of carboxylic groups in SP2 was also validated by its FTIR spectra, which displays a weak shoulder peak around 1700 cm-1 and a broad O-H stretch between 2700 and 3000 cm-1, typical of carboxylic groups (Fig. S5 in Appendix A). By evaluating the denitrogenation trends for SP1 and SP2 together with their elemental composition, a likely reaction mechanism for the denitrogenation of ABS was determined and is presented in Fig. 5.
The subcritical conditions hydrolyze the C≡N bond in the acetonitrile fraction in the ABS to form an amide functional group. The nitrogen in the amide functional group is then replaced by the potassium in KOH under these conditions to form a K-ABS carboxylic salt complex, while the nitrogen is removed as NH3. NH3 is soluble in water and is removed via evaporation overnight, while the K-ABS salt complex is soluble in water and forms the AP1. The SP1 contains polymer chains with fewer carboxylic salt chains (derived from the hydrolysis of the C≡N bond in the acetonitrile fraction); that is, it has a greater number of non-polar butadiene and styrene chains that are insoluble in water and thus exists as a precipitated solid in the reactor after the reaction. However, the SP2 has a greater number of carboxylic salt chains, making it water-soluble and leading to the formation of the AP1. The AP1 is then neutralized using HCl acid to precipitate out the SP2.
Plausible polymeric structures that may exist in the SP1 and SP2, along with their H/C and N/C elemental ratios, are provided in Table S2 in Appendix A. SP1 tends to contain more aliphatic structures, such as structures 9 and 13 in Table S2 or their saturated versions, whereas SP2 has less saturation and is likely to be a mixture containing more aromatic units. The structural inclinations of SP1 and SP2 were validated by the FTIR spectra for the SP2 samples, which exhibit more intense FTIR bands near 1600 and 750 cm-1, corresponding to C=C stretching and out-of-plane C-H aromatic bending vibrations, respectively (Fig. S5). These pronounced features indicate a higher concentration of aromatic moieties—such as that found in styrene—in the SP2 samples. In contrast, the SP1 spectra display weaker aromatic bands and relatively broader absorptions in the 2850-2950 cm-1 region, which are typical of aliphatic C-H stretching, suggesting the presence of more aliphatic moieties.
3.2. Supercritical conversion of SP2
3.2.1. Products from SP2 in the supercritical experiments
The major product fractions detected in the supercritical experiments were oil, AP3, and gases. Increasing the operating temperature and residence time generally favored secondary decomposition reactions that formed more AP3 and gases. Hence, the maximum amount of oil from SP2 (38.2 wt% yield with respect to ABS) was detected at 375 °C and 30 min (Fig. 6). The gas and AP3 yields varied in the ranges of 2-11 wt% and 11-27 wt%, respectively. The conversion of SP2 was nearly 100% and left very little solid residue after stage 2. However, it is important to note that the yields of the products were calculated with respect to the initial mass of ABS, in order to maintain coherence. Hence, according to the mass balances in Eqs. (1), (2), (3), (4), the total yield of the products from SP2 will be less than 100%. The total yield of products will be approximately equal to 53.2 wt%, which is the yield of SP2 from ABS at 325 °C and 20 min (stage 1). Increasing the plastic loading to 45 and 60 wt% promoted the formation of more AP3, rather than oil, with a yield varying between 18 and 26 wt% of the ABS at 400 and 425 °C (Fig. S6 in Appendix A). In this study, the masses of both water and plastic were varied to keep the final operating pressure constant at 250 bar. Faeth and Savage [26] have argued that, with a low water loading, some water evaporates when the reactor is heated, resulting in an even lower amount of liquid water during HTL and leading to a high effective concentration of the feedstock. The high effective concentration of ABS at 45-60 wt% loading in the present study had a profound effect resulting in lower oil yields.
3.2.2. Supercritical oil composition from SP2
GC/ MS analysis of the oils revealed that they mainly consisted of alkyl benzenes and poly-aromatics. Regions I, II, and III in the GC/MS spectra shown in Fig. 7 represent the number of benzene rings in the compounds detected. Region I consists of single-benzene-ring compounds, such as ethylbenzene (C8H10) and styrene (C8H8). Region II consists of compounds with two benzene rings, such as naphthalene (C10H8) and other biphenyl compounds. Region III consists of compounds with three or more benzene rings, such as 2-phenylnaphthalene (C16H12) and p-terphenyl (C18H14). These poly-aromatics are contributed by the styrene fraction in the ABS. In fact, it has been reported that the formation of poly-aromatic compounds from polystyrene and ABS increases with an increase in the operating temperature, possibly due to the initiation of repolymerization reactions by the benzene fragments under severe conditions [27,28]. For instance, an increase in the operating temperature from 350 to 450 °C led to an increase in selectivity to poly-aromatics by about 30.5% and about 9.6% for polystyrene and ABS, respectively.
The GC/MS peak area was slightly skewed toward region II in comparison with regions I and III, indicating a higher selectivity toward dimer compounds from SP2 under supercritical conditions. The results were in agreement with the 1H NMR of the oils, which indicated that most of the protons in the oils belonged to aromatic and aliphatic moieties (Fig. S7 in Appendix A). The protons from the benzene rings constitute aromatic moieties, while the protons from alkyl groups attached to the benzene rings constitute the aliphatic moieties in the oils. HSQC NMR spectroscopy of the oils also indicated that approximately 80% of the moieties belonged to the aromatic group (Table S3 in Appendix A). The FTIR spectra for the oils displayed peaks between 600 and 700 cm-1 corresponding to aromatic C-H bending, which also indicated the predominance of aromatic compounds in the oil (Fig. S8 in Appendix A).
The relative abundance of styrene—one of the monomers of ABS—in the oil was marginal, compared with other aromatic compounds. Supercritical water does not provide a favorable environment for unsaturated monocyclic aromatics such as styrene; instead, it favors the formation of more stable compounds such as ethylbenzene, which was evident from the GC/MS spectra of the oils [27]. Although elemental analysis revealed the presence of nitrogen in the oil, the nitrogenous compounds were below GC/MS-detectable levels in the present study, as the bulk of the nitrogen had been removed at the subcritical step. Elemental analysis of the oils indicated an average H/C to O/C ratio of > 1, with a nitrogen content of < 1 wt% for most of the experiments, leading to HHVs of up to 39.9 MJ·kg-1 (Table S4 in Appendix A).
The SimDis of the oils indicated that the major fraction of the compounds fell within the kerosene and vacuum gas oil boiling range (Fig. 8). Compounds within the kerosene boiling range included mono-aromatic compounds such as ethylbenzene (boiling point (BP) 136 °C) and butylbenzene (BP 183 °C), as well as di-aromatics such as naphthalene (BP 218 °C). The vacuum gas oil fraction consisted of heavier di-aromatics such as 4-methylbiphenyl (BP 267 °C) and 1,3-diphenylpropane (BP 300 °C). Tri-aromatic compounds such as p- terphenyl (BP 389 °C) constituted the bottom residue. The abundance of certain compounds such as ethylbenzene and naphthalene was evident from the BP distribution curves, where the curves experienced a sudden jump in their cumulative signal area when the temperature reached the BP of these compounds (Fig. S9 in Appendix A).
3.3. Supercritical conversion of SP1
The SP1 did not undergo any noticeable conversion at 375 °C, with the oil yield being nearly zero. It required at least 400 °C and 60 min to achieve a noticeable conversion to oil. This requirement was higher than that of SP2, elucidating the presence of more butadiene structures, as predicted previously from the H/C ratios. Straight-chain aliphatic carbon bonds such as those present in PP, PE, and the butadiene structure in the ABS are more difficult to cleave, compared with aromatic C-C linkages, resulting in a higher temperature requirement for SP1. Unlike that of SP2, the conversion of SP1 improved with increasing reaction severity; that is, a higher temperature and longer retention time favored the formation of oil. The maximum yield of oil from SP1 was obtained at 425 °C and 90 min (38.7 wt%).
The oil from SP1 contained more mono-aromatic compounds, such as benzene (C6H6, ∼0.3%), toluene (C7H8, ∼17.1%), and ethylbenzene (C8H10, ∼27.2%), compared with SP2 (Fig. 9). In fact, GC/MS analysis revealed that the total peak area percentage of the mono-aromatic content in the supercritical oil from SP1 was nearly five times higher than that of SP2, at the expense of the di-aromatic and poly-aromatic peak area percentage. Conversely, the di-aromatic and poly-aromatic peak area percentage in SP1 was nearly three times lower than that in SP2 and was particularly low in the oil from SP1. The predominance of mono-aromatics in the oil was also evident from the SimDis analysis, which showed that the oil from SP1 contained more compounds in the kerosene boiling range compared with that from SP2 (Fig. 8). HSQC spectroscopy revealed that the oil from SP1 contained 73.7% aromatic -CH- linkages and 20.2% aliphatic -CH- linkages, whereas SP2 contained 79.4% and 11.8%, respectively, at 425 °C and 60 min (Table S2). The predominance of mono-aromatic benzene rings containing alkyl functional groups (e.g., toluene and ethylbenzene) in SP1 led to the higher percentage of aliphatic linkage in SP1 than in SP2. The oil from SP1 at 425 °C and 60 min contained 90.1 wt% C, 8.8 wt% H, and 0.4 wt% N, with an HHV of 41.6 MJ·kg-1.
3.4. Co-liquefaction of SP1 and SP2
The simultaneous supercritical HTL of SP1 and SP2 was also carried out under select conditions. At 375 °C, 250 bar, and 30 min, the yield of oil was only 36.2 wt%, which was very close to the yield of oil obtained from SP2 alone under these conditions. This occurred because, under these conditions, only SP2 decomposes to form the oil, while SP1 remains unconverted. Near-complete conversion of both SP1 and SP2 takes place at 425 °C, 250 bar, and 60 min, with and without including the neutralization of KOH with HCl. The yields of oils from the co-liquefaction of SP1 and SP2 were 60.4 and 66.3 wt%, with and without neutralization under these conditions, respectively. However, the gas yield from the co-liquefaction of SP1 and SP2 was double (8.6 wt%) with neutralization compared with the gas yield without neutralization (4.3 wt%). This result was reasonable, since neutralizing the carboxylic salts leaves behind carboxylic acid chains in the polymer; these chains are highly susceptible to decarboxylation under supercritical conditions and release carbon dioxide (CO2) to the gas phase. The yields of AP3 were 31.0 and 29.4 wt% with and without neutralization, respectively, leaving very little solid residue behind. The organic compositions of the oils were similar irrespective of the neutralization step and consisted mainly of mono-aromatic compounds (∼64.5%), with a mixture of di-aromatics (∼18.6%) and poly-aromatic compounds (∼8.7%) (Fig. 9). The predominance of mono-aromatic compounds such as toluene and ethylbenzene was also evident in these oils. The SimDis results revealed a similar boiling range composition for the oils with and without neutralization, wherein kerosene and vacuum gas oil fractions were predominant (Fig. 8). The oils from the co-HTL of SP1 and SP2 evidently contained more fractions of naphtha and kerosene than those from SP2, which aligns with the GC/MS results of these oils. The HHVs of the oils with and without neutralization were 35.2 and 38.0 MJ·kg-1, respectively.
3.5. Single-stage supercritical HTL of ABS
To assess the difference between the single-stage supercritical HTL and two-stage sequential HTL of ABS, a control experiment was performed with ABS in which the ABS was directly liquefied at 425 °C for 60 min. The single-stage HTL resulted in an oil yield of 59.7 wt%, which aligned with the results from previous studies [28]. The oils contained similar organic compounds as the oils from the two-stage experiments, including benzene, toluene, styrene, ethylbenzene, di-aromatics, and poly-aromatics. However, the selectivities to certain mono-aromatic structures such as benzene (∼0.5%), toluene (∼20.3%), ethylbenzene (∼29.1%), and cumene (∼13.1%) were noticeably higher in the oil from the single-stage HTL compared with that from the two-stage HTL (Table S5 in Appendix A). The GC/MS also detected the presence of nitrogenous compounds such as isobutyronitrile (C4H7N, ∼0.3%), benzenebutanenitrile (C10H11N, ∼3%), 1-cyanonaphthalene (C11H7N, ∼0.4%), pentadecanenitrile (C15H29N, ∼0.3%), and heptadecanenitrile (C17H33N, ∼0.4%). Previous studies have reported that, under supercritical conditions, acrylonitrile is hydrogenated to form the more stable propanenitrile in the oil [29]. The higher nitrogen content of this oil was also evident from its elemental analysis, which revealed that the oil contained 80.7 wt% C, 7.8 wt% H, and 1.42 wt% N, with an HHV of 36.1 MJ·kg-1. In general, the nitrile bonds in the ABS undergo hydrolysis under subcritical conditions to form amines, amides, and carboxylic acids, which are water-soluble [30]. Hence, a noticeable portion of the nitrogen migrates to the aqueous phase, even in the absence of the KOH catalyst in a single-stage HTL process. However, the presence of KOH in the two-stage HTL catalyzes this hydrolysis reaction, leading to a significant decrease in nitrogen in the resulting oil [31]. Overall, a single-stage HTL is sufficient to liquefy ABS to produce high yields of energy-dense oils. However, the presence of nitrogenous compounds in the oils greatly affects the oil quality.
3.6. Gas composition
GC/MS of gases from select experiments revealed that the gases contained C1-C5 gases, mainly CO2, propene (C3H6), butane (C4H10), 1-butene (C4H8), isobutane (C4H10), isobutylene (C4H8), pentane (C5H12), 2-pentene (C5H10), and 1,3-pentadiene (C5H8) (Fig. S10 in Appendix A). It is clear that the presence of CO2 is due to the decarboxylation of the carboxylic acids formed during stage 1; it is also formed via oxidation from oxygen in the head space of the reactors. The remaining C2-C5 gases arise from the butadiene and aliphatic fraction of the ABS through free-radical mechanisms. In fact, butadiene in its pure form is a gas by itself at standard temperature and pressure (STP), so the butadiene fractions are more inclined to form gas products after HTL. This also explains the very low presence of butadiene fractions and aliphatic compounds in the oils from the two-stage HTL of ABS, as the aliphatic component of the ABS is converted to the gas phase. Apart from the C2-C5 hydrocarbons, traces of aromatic oil compounds such as benzene, toluene, and styrene were also detected in the gas phase, possibly due to their high volatility at room temperature.
4. Discussion and perspectives
Recovery of KOH is an important aspect of HTL in terms of scalability, sustainability, and economic feasibility. In this study, most of the carbon is recovered in the oil phase after stage 2, leaving minimal organic content in the aqueous phase; that is, the aqueous phase closely resembles the KOH solution initially introduced in the subcritical stage, especially when no neutralization step is involved. Thus, if the two-stage process is performed without neutralization with HCl, then either the KOH can be recovered via crystallization from AP2 or the AP2 itself can potentially be recirculated back, improving the process economics. However, KOH recovery only works with the two-stage HTL without the neutralization step. Neutralizing the carboxylic salts with HCl leads to the formation of KCl, which is likely to have no effect on the HTL of ABS and hence, probably have no effect when recirculated back into the system. The denitrogenation of ABS with KOH appears to be autocatalytic, with diminishing returns at higher KOH concentrations. For instance, approximately 65% denitrogenation is achieved at very low concentrations of KOH (0.05 gKOH⋅gABS-1), whereas the KOH concentration must be increased 6-8 fold (0.3-0.4 gKOH⋅gABS-1) in order to achieve about 95% denitrogenation (an increase of 30%). Hence, an optimal balance is recommended between the required DN and KOH usage. Additionally, potassium salts can corrode the reactor walls under HTL conditions. For example, the addition of 0.5 mol·L-1 K2CO3 catalyst doubled the corrosion rate of Alloy 33 to 3.6 µm·a-1 at 310 °C and to 4.2 µm·a-1 at 365 °C compared with the values exhibited by pure water under these conditions [32]. Performing the two-stage HTL without neutralization increases the risk of alkali-induced corrosion in the supercritical reactor as well. Therefore, it is better for the reactor’s operating lifetime to include a neutralization step, although doing so will result in slightly lower oil yields.
The final oil product from ABS contained a mixture of mono-, di-, and tri-cyclic aromatic compounds, with very little selectivity to monomers such as styrene and 1,4-butadiene. Hence, supercritical HTL of ABS is not ideal for monomer recycling. Alternatively, pyrolysis of ABS offers a better styrene yield (∼33 wt%) compared with HTL, but it operates at 515 °C, consuming more energy due to drying and worse heat transfer [7]. HTL is competitive, exhibiting energy recoveries of 54.6% and 63.8% from the oils for the single-stage and two-stage HTL (without neutralization), respectively. The pyrolysis oil from ABS had an HHV of approximately 39.4 MJ·kg-1 and led to an energy recovery of about 82.5%; however, it contained 6.5-7.1 wt% N, which made this oil unfit to use as is, without further downstream processing [7].
A comparison of the energy consumed by pyrolysis versus HTL can be estimated from the literature. The pyrolysis of plastic waste at 500 °C was modeled at the pilot scale and demonstrated to consume 1.6-2.4 MJ·kg-1 [33]. Energy consumption data for supercritical HTL is sparse but can be estimated from the sensible heat required to heat water to reaction conditions, which is 2.7 MJ·kg-1 at 250 bar [20]. Hence, the single-stage HTL of ABS at 425 °C and 250 bar requires 10.8 MJ·kg-1 if a 20% plastic loading is used. Heating the ABS is estimated to consume 0.7 MJ·kg-1, based on an average heat capacity (Cp) of 1800 kJ·(kg·°C)-1, resulting in a total energy consumption of 11.4 MJ·kg-1. When employing the two-step process, 6.5 MJ·kg-1 of energy is used in the 20% ABS subcritical step (stage 1), and an additional 2.5 MJ·kg-1 is used in the 60% ABS supercritical step (stage 2), resulting in a total energy consumption of 9 MJ·kg-1. These calculations highlight the importance of optimizing the ABS-to-water ratio in the HTL of plastics. In fact, a single-stage supercritical HTL at 425 °C with 60% ABS loading requires only 2.5 MJ·kg-1. This would also mean that the two-step HTL process with both stages operating at 60% ABS loading would result in approximately 4 MJ·kg-1 (Table S6 in Appendix A). In addition, it is important to consider heat recovery in HTL via heat exchange, which is routinely employed and achieves greater than 75% heat recovery [34]. A life-cycle assessment of the HydroPRS process, which is similar to the single-stage supercritical process investigated here, indicated an estimated energy consumption of 2.8 MJ·kg-1 [35], which is broadly in agreement with our estimations.
Overall, it can be assumed that the energy consumption for pyrolysis and HTL are comparable, although considerable uncertainties are evident. Considering the quality of the oils obtained from pyrolysis and HTL, the former requires more extensive catalytic upgrading, which takes additional energy, H2, and heterogeneous catalysts; this can contribute to energy and material consumption for the overall ABS processing by pyrolysis. A complete mass balance of the product categories from the single-stage and two-stage HTL of ABS is illustrated in Fig. 10. It is possible that the oils from the two-stage HTL of ABS could be directly used as an energy-dense fuel in boilers and furnaces, since they have a very high calorific value. These oils can also potentially be co-processed with petroleum crude oil in refineries to recover conventional fuels and aromatic chemicals. Researchers have shown that distilled HTL oils can be directly blended with petroleum fractions such as diesel and used in commercial internal combustion engines, resulting in satisfactory performance and emission characteristics from the blended oil [36,37]. Hydrocracking the aromatic-rich oil could also increase the presence of monocyclic aromatics, at the expense of di- and tri-cyclic aromatics, allowing the recovery of valuable platform chemicals such as benzene, toluene, and other alkylbenzenes.
Based on the results from numerous experiments, the optimal method to produce nitrogen-free, aromatic-rich oil involves first denitrogenating the ABS under subcritical conditions (350 °C, 20 min, 0.3 gKOH⋅gABS-1) and then removing the nitrogen as NH3 via overnight evaporation, followed by supercritical HTL (at 425 °C, 60 min) without any neutralization step. While the two-stage HTL shows promise at lab scales to produce nitrogen-free aromatics, its implementation will increase the capital and operating costs at commercial scales. Although commercial-scale single-stage supercritical HTL plants are being developed [35], there are process engineering considerations that need to be accounted for when it comes to a two-stage HTL setup. The stage 1 product mixture will have to be cooled down to allow the removal of NH3, which then requires re-pressurization and re-heating to stage 2 supercritical conditions. Typically, HTL reactors converting biomass are plug-flow-type reactors with heat exchangers that pump the biomass-water slurry as a mixture to reach high pressures. Due to the inability of plastics to form homogeneous water slurries for pumping, different pumping solutions such as extruders must be employed to introduce molten ABS to a pressurized reactor containing the hot water. As mentioned above, the ratio of water to ABS also needs to be optimized in order to minimize energy consumption. The heat exchanger design of a two-step process is equally important and should be optimized by modeling in future techno-economic assessment (TEA) studies. Thus far, no literature study has addressed these issues. Therefore, a detailed techno-economic feasibility and life-cycle analysis of the two-stage HTL is required to further justify its use over the conventional single-stage HTL at commercial scales. Future studies could also explore the coupling of subcritical HTL with pyrolysis as a second stage, which could potentially combine the benefits of both approaches: higher monomer recovery and lower nitrogen content in the oil.
5. Conclusions
The two-step sequential HTL of ABS exhibited remarkable efficiency in denitrogenating the ABS at the subcritical step and subsequently producing high-quality oil at the supercritical step. Operating conditions of 325 °C and a 20 min retention time, with a KOH load above stoichiometric requirements, were found to be the optimum to achieve 95% ABS denitrogenation efficiency. The nitrogen was mainly removed as NH3 via evaporation from the aqueous phase. Subsequent supercritical HTL led to the production of energy-dense oils, rich in mono-, di-, and tri-cyclic aromatics. The oils contained very little nitrogen content, indicating that sufficient denitrogenation was achieved using KOH under subcritical conditions. Compared with the two-stage methodology, the single-stage direct HTL of ABS under supercritical conditions led to an oil with higher nitrogen content. The butadiene fractions in the ABS mainly contributed to the formation of gases. Overall, the optimum methodology was found to be the two-stage HTL without the neutralization step, which exhibited the highest energy recovery from the oils (63.8%). Using this methodology, for every gram of clear ABS, 0.049 g out of the total 0.051 g of nitrogen can be separated as NH3, and 0.663 g of aromatic-rich energy-dense oil with very little nitrogen content and an HHV of 38 MJ·kg-1 is produced.
GeyerR, JambeckJR, LawKL. Production, use, and fate of all plastics ever made. Sci Adv2017; 3(7):1700782.
[2]
JambeckJR, GeyerR, WilcoxC, SieglerTR, PerrymanM, AndradyA, et al. Plastic waste inputs from land into the ocean. Science2015; 347(6223):768-71.
[3]
MacLeodM, ArpHPH, TekmanMB, JahnkeA. The global threat from plastic pollution. Science2021; 373(6550):61-5.
[4]
DeshmukhD, KulkarniH, SrivatsDS, BhanushaliS, MoreAP. Recycling of acrylonitrile butadiene styrene (ABS): a review. Polym Bull2024; 81(13):1-38.
[5]
QinY, HuangZ, MaK, DongL, LiJ, WangQC, et al. Life cycle assessment of recycling high impact polystyrene and acrylonitrile butadiene styrene plastic from waste refrigerators. J Clean Prod2024; 435:140294.
[6]
RomaniA, LeviM, PearceJM. Recycled polycarbonate and polycarbonate/acrylonitrile butadiene styrene feedstocks for circular economy product applications with fused granular fabrication-based additive manufacturing. Sustain Mater Technol2023; 38:e00730.
[7]
JungSH, KimSJ, KimJS. The influence of reaction parameters on characteristics of pyrolysis oils from waste high impact polystyrene and acrylonitrile-butadiene-styrene using a fluidized bed reactor. Fuel Process Technol2013; 116:123-9.
[8]
KaminskyW. Chemical recycling of plastics by fluidized bed pyrolysis. Fuel Commun2021; 8:100023.
[9]
Dos PassosJS, GlasiusM, BillerP. Screening of common sy nthetic polymers for depolymerization by subcritical hydrothermal liquefaction. Process Saf En viron Prot2020; 139:371-9.
[10]
PeplowM. A new answer to the plastic pollution crisis? Nature 2025; 638:22-5.
[11]
MatayevaA, MadsenAS, BillerP. Evaluation of different fiber impurities on hydrothermal liquefaction of mixed textile waste. Resour Conserv Recycl2023; 190:106833.
[12]
Jaime-AzuaraA, LemmingM, WimmerR, KohansalK, HingeM, HelmerPT. Continuous hydrothermal processing of poly(ethylene terephthalate) (PET) under subcritical water conditions: a proof-of-principle closed-loop study. Chem Eng J2024; 495:153223.
[13]
ChenWT, JinK, Linda WangNH. Use of supercritical water for the liquefaction of polypropylene into oil. ACS Sustain Chem Eng2019; 7(4):3749-58.
[14]
LuT, JanK, ChenWT. Hydrothermal liquefaction of pretreated polyethylene-based ocean-bound plastic waste in supercritical water. J Energy Inst2022; 105:282-92.
[15]
KusenbergM, EschenbacherA, DjokicMR, ZayoudA, RagaertK, De MeesterS, et al. Opportunities and challenges for the application of post-consumer plastic waste pyrolysis oils as steam cracker feedstocks: to decontaminate or not to decontaminate? Waste Manag 2022; 138:83-115.
[16]
Dos PassosJS, LorentzC, LaurentiD, RoyerSJ, ChontzoglouI, BillerP. Hydrothermal liquefaction of plastic marine debris from the North Pacific Garbage Patch. Resour Conserv Recycl2024; 209:107822.
[17]
HarisankarS, BillerP. Kinetics and recovery of bisphen ol-A from fiber-reinforced polycarbonate using subcritical water. React Chem Eng2025; 10(12):2835-48.
[18]
TitoE, dos PassosJS, RombolàAG, TorriC, BensaidS, PironeR, et al. Sequential hydrothermal dechlorination and liquefaction of PVC. Energy Convers Manage2024; 304:118228.
[19]
HenriksenML, HingeM. Re-PlastInternet. Aarhus: Aarhus University; c2025 [cited 2025 Feb 10]. Available from: https://re-plast.dk/.
[20]
Linstrom PJeditor. Thermophysical properties of fluid systems. NIST chemistry webbook [Internet]. Gaithersburg: National Institute of Standards and Technology; [cited 2025 Feb 10]. Available from: https://webbook.nist.gov/chemistry/fluid/.
[21]
ChanniwalaSA, ParikhPP. A unified correlation for estimating HHV of solid, liquid and gaseous fuels. Fuel2002; 81(8):1051-63.
[22]
LinX, FengW, XuL. Study on flame retardant ABS. Mater Sci Mater Rev2017; 1 (1):1-10.
[23]
DuAK, ZhouQ, WenZB, YangJW, van KasterenJMN, WangYZ. Denitrogenation of acrylonitrile-butadiene-styrene copolymers using polyethylene glycol/hydroxides. Polym Degrad Stabil2011; 96(5):870-4.
[24]
ZhouQ, YangJW, DuAK, WangYZ, Van KasterenJMN. Fuel oil from acrylonitrile-butadiene-styrene copolymers using a tandem PEG-enhanced denitrogenation-pyrolysis method. AIChE J2009; 55(12):3294-7.
[25]
ShenZ, YuY, WangK, YueC, MiaoJ, XiaM, et al. Hydrothermal alkaline conversion of actual acrylonitrile wastewater to organic acids. Process Saf Environ Prot2020; 144:93-9.
[26]
FaethJL, SavagePE. Effects of processing conditions on biocrude yields from fast hydrothermal liquefaction of microalgae. Bioresour Technol2016; 206:290-3.
[27]
BaiB, JinH, ZhuS, WuP, FanC, SunJ. Experimental investigation on in-situ hydrogenation induced gasification characteristics of acrylonitrile butadiene styrene (ABS) microplastics in supercritical water. Fuel Process Technol2019; 192:170-8.
[28]
HarisankarS, Pragadeshwar BabuMV, SrinivasanS, VinuR. Hydrothermal liquefaction of diverse plastics using water and seawater. Chem Eng J2024; 501:157365.
[29]
LiuY, FanC, ZhangH, ZouJ, ZhouF, JinH. The resource utilization of ABS plastic waste with subcritical and supercritical water treatment. Int J Hydrogen Energy2019; 44(30):15758-65.
[30]
IzzoB, KleinMT, LaMarcaC, ScrivnerNC. Hydrothermal reaction of saturated and unsaturated nitriles: reactivity and reaction pathway analysis. Ind Eng Chem Res1999; 38(4):1183-91.
[31]
SanliO. Homogeneous hydrolysis of polyacrylonitrile by potassium hydroxide. Eur Polym J1990; 26(1):9-13.
[32]
LiuM, ZengY. Key processing factors in hydrothermal liquefaction and their impacts on corrosion of reactor alloys. Sustainability2023; 15(12):9317.
[33]
NetschN, ZellerM, RichterF, BergfeldtB, TavakkolS, StapfD. Energy demand for pyrolysis of mixed thermoplastics and waste plastics in chemical recycling: model prediction and pilot-scale validation. ACS Sustain Resour Manage2024; 1(7):1485-92.
[34]
AnastasakisK, BillerP, MadsenRB, GlasiusM, JohannsenI. Continuous hydrothermal liquefaction of biomass in a novel pilot plant with heat recovery and hydraulic oscillation. Energies2018; 11(10):2695.
[35]
OzoemenaMC, ColesSR. Hydrothermal treatment of waste plastics: an environmental impact study. J Polym Environ2023; 31(7):3120-30.
[36]
ChenWT, ZhangY, LeeTH, WuZ, SiB, LeeCFF, et al. Renewable diesel blendstocks produced by hydrothermal liquefaction of wet biowaste. Nat Sustain 2018;1(11):702-10.
[37]
HarisankarS, AlawaB, ChakmaS, VinuR. Industrial waste to diesel engines: a novel hydrothermal co-liquefaction approach to produce diesel blendstocks. Resour Conserv Recycling2024; 205:107573.
PDF
(2224KB)
