Circularizing PET-G Multimaterials: Life Cycle Assessment and Techno-Economic Analysis

The recycling of multimaterials such as payment or access cards poses significant challenges. Building on previous experimental work demonstrating the feasibility of chemically recyclable payment cards made from glycol-modified poly(ethylene terephthalate) (PET-G), we use life cycle assessment and techno-economic analysis to investigate two chemical recycling scenarios and evaluate their potential environmental and economic benefits. Recovering all components from the depolymerized products (Scenario 1) achieves substantial environmental benefits across most categories, reducing global warming by up to 67% compared to only recovering major components (Scenario 2). However, the environmental benefits in Scenario 1 incur 69% higher total annualized costs, causing its profitability to be dependent on a minimum selling price of £13.4/kg for cyclohexanedimethanol and less than a 10% discount rate. In contrast, Scenario 2 is less sensitive to discount rate variation and thus a lower risk and more economically feasible option, albeit less environmentally sustainable.


Composition of PET-G payment cards
Materials: PET-G cards with a dimension of 85.57 mm × 53.97 mm × 0.76 mm were provided by Mastercard.PET-G pellets were supplied by Push Plastic (USA).To remove any moisture, the pellets were dried for 4 h in a vacuum oven at 60 °C and then stored in a desiccator until further use.
Material Characterization: Thermogravimetric analyses (TGA) of the plastic component of the card and PET-G pellets were performed on a TA SDT650 thermogravimetric analyzer.The specimens were heated from ambient temperature to 600 °C under a nitrogen atmosphere (gas purge rate of 50 cm 3 /min), at a heating rate of 10 k/min.After which, the sample was cooled down to 300 °C and held for 2 min to equilibrate.Then, the atmosphere was switched to air with the same purge rate of 50 cm 3 /min.The sample was finally heated again from 300 to 800 °C, at a heating rate of 10 k/min.

Figure S1
. TGA thermograms of PET-G card and PET-G pellets, with an inset showing the mass loss under nitrogen atmosphere during stage 1.After stage 2 the samples were cooled, the gas was switched to air, and the samples were heated again (stage 3).
To determine the composition (e.g., polymer, additives, fillers) of the plastic component of the PET-G cards, TGA experiments were conducted under both inert and reactive atmospheres (Figure S1).Under N 2 atmosphere, the card exhibited a mass loss of 1.6% in the temperature range of 30 -370 °C, which was attributed to the evaporation of low boiling-point additives (stage 1).This was followed by a mass loss of 82.6%, attributed to the decomposition of the PET-G polymer (stage 2).The subsequent mass loss of 10.5% under air atmosphere was attributed to the combustion of the carbonized polymer (stage 3), and the remaining 5.3% was attributed to the inorganic fillers present in the card.PET-G pellets were also analyzed under identical conditions for comparison.The differences between the card and the pellets were more clearly visible from the derivative TGA (DTGA) curves (Figure S2); the virgin PET-G pellets only showed the later two stages of degradation, presumably due to the absence of additives.

Life cycle assessment
Table S1.LCA basis of the study

Goal
The goal is to assess and measure the environmental impacts resulting from the process of depolymerizing waste plastic cards, separating and purifying the depolymerized products, and reclaiming solvents.

Functional unit
The functional unit is defined as depolymerizing one tonne of PET-G payment cards per day.

Scope
The scope of this investigation begins with the depolymerization process and concludes with the recovery of the recycled products from the system.Any materials that cannot be recovered from the processing steps, such as fillers and additives, are considered to be non-recoverable and will be disposed of as wastewater.

Target audience
This study aims to provide valuable insights and information to various stakeholders including waste management professionals in academia, local government officials, policymakers, and industrial sectors.The primary focus is on effective waste management strategies for plastic cards.
The selection of a 1-tonne/day treatment capacity for the plant is influenced by several considerations.
While larger plant sizes indeed offer improved economies of scale, the existing market share of chemically recyclable PET-G cards remains relatively modest in comparison to non-recyclable PVC cards.As pilot schemes are already in progress in the UK to facilitate secure and sustainable disposal of expired payment cards, we must account for the transition phase from PVC to PET-G and the establishment of a pragmatic recycling system.This chosen capacity is in alignment with current market dynamics and the focus on conducting proof-of-concept studies.
At a 1-tonne/day capacity, the plant would be equipped to recycle ~200,000 cards daily.Should a scenario of complete collection and recycling of the annual production of 6 billion cards be achieved, approximately 80-90 plants of similar scale could effectively manage the recycling on a global scale.
As the recycling infrastructure matures and the circulation of PET-G cards expands, the feasibility of larger treatment capacities is expected to grow from an economic standpoint.

Synthesis of BHET
The BHET used in this study can be prepared from terephthalic acid (TPA) and ethylene glycol (EG), as shown in Scheme S1.The conditions described below were used to evaluate the environmental impact of the BHET preparation (a background process in this study).
The BHET synthesis was conducted following a previously published method. 1 The esterification reaction of TPA and EG was performed with a TPA : EG molar ratio of 1:3, at 210 ℃ for 610 min until a transparent solution was obtained.

Synthesis of CHDM
The CHDM used in this study can be prepared through the hydrogenation of BHET, as shown in Scheme S2.The conditions described below were used to evaluate the environmental impact of the CHDM preparation (a background process in this study).The solid sample was dried in an oven at 393 K for 12 h and subsequently calcined at 723 K for 4 h to obtain the oxide.Prior to use, the as-prepared oxide was reduced by H 2 at 573 K.

Hydrogenation of BHET to CHDM:
The hydrogenation reactions were performed in a stainless-steel high-pressure reactor with magnetic stirring.The hydrogenation of BHET to CHDM was conducted in two steps (Scheme S2).In the first step, 2.0 g of BHET and 0.1 g of 10% Pd/C catalyst were charged into the reactor, which was then purged with H 2 three times to remove air.The reactor was pressurized to the desired H 2 pressure (5.5 -7 MPa) and heated to the desired temperature (428 K) at a stirring rate of 400 rpm.After the reaction, the intermediate products, mainly consisting of bis(2-hydroxyethyl) cyclohexane-1,4-dicarboxylate (BHCD), were obtained and used as the raw material for the next step.
In the second step, the intermediate products (  glycolysis and the conventional BHET synthesis process from dimethyl terephthalate (DMT) and EG. 3 In their conventional BHET synthesis, a GWP of 4.37 kgCO 2 eq/kg BHET was observed, a value 2.5 times higher than our approach utilizing TPA and EG.Notably, if the GWP implications of the BHET production from DMT were considered, the environmental benefits identified in our study would be significantly amplified (2.5 times) beyond our current estimations.

Synthesis of DBU
The organocatalyst used for depolymerization in this study (DBU) can be prepared from caprolactam, which is in turn derived from cyclohexanone as shown in Schemes S3 and S4.The conditions described below were used to evaluate the environmental impact of the DBU preparation (a background process in this study).
A 500 mL flask was charged with 70 mL tert-butyl alcohol, 100 g caprolactam, and 0.     Figure S12.Global warming potential of two scenarios.S1recovery of all components; S2recovery of metals and BHET.Note: LU = land use; IR = ionizing radiation; HNCT = human non-carcinogenic toxicity; FRS = fossil resource scarcity; GWP = global warming potential; HCT = human carcinogenic toxicity; TE = terrestrial ecotoxicity.Metals, specifically copper, were represented as solid components within the model.Chemicals such as DBU, EG, BHET, and CHDM were modelled as conventional components.

Techno-economic analysis
The purification steps for the metal involve several unit operations to ensure effective separation and purification.Initially, metal is separated from the depolymerized products using a filter.The filtered metal is then subjected to a washing process in a Swash unit, where acetone is used as the solvent to remove any impurities and residues.
The crystallizer is represented as a combination of a cooler and a separator.BHET exhibits relatively high solubility in hot water; however, upon cooling, it precipitates from the solution.Subsequently, the white BHET crystals can be readily separated from the liquid phase.The recovery of BHET from recrystallization is estimated based on its solubility parameters in water at various temperatures. 5At a cooling temperature of 2 °C, our modeling indicates a BHET recovery of 99.4% through recrystallization.Due to the lack of available data on crystallization kinetics for BHET, the capital expenditure estimates pertaining to the crystallizer are exclusively derived from mass and energy balance considerations; we note that crystallization kinetics may impact the required equipment dimensions.
In the RadFrac model, the determination of the tray count initiates with the application of the shortcut model (DSTWU) to acquire preliminary values.These baseline figures subsequently serve as the initial inputs for the RadFrac model.By systematically tailoring the RadFrac model according to predefined design criteria, such as targeted mole recovery and purity, achieved through modulation of reflux ratio and distillate to feed ratio, a sequential process of iterative optimization ensues.This iterative approach culminates in the attainment of the desired operational outcomes.
The application of the NRTL-RK method is specifically directed towards the separation and purification of the depolymerized products, rather than the depolymerization process itself.Detailed and comprehensive experimental data concerning depolymerization, separation, and purification processes can be found in our previous study; 6 the outcomes of these smaller-scale batch experiments served as vital benchmarks for the larger-scale separation and purification process investigated here.Though these processes were simulated under standard conditions and thus an ideal gas approach would have sufficed, we decided to implement the RK equation of state (EOS) to assure the model's future applicability to elevated pressures and non-ideal conditions.The capital cost of the reactor is dependent on the processing capacity of the PET-G payment cards.A 2000 L multifunctional stainless steel reactor costs $18,000 (£14575.50,sourced from Alibaba Group Holding Limited).The depolymerization reaction will be conducted three times daily, with each batch requiring the depolymerization of 333.33 kg of payment cards.The costs associated post depolymerization processes, such as separation and purification, were calculated using Aspen Process Economic Analyzer (APEA).Given that the cost of the reactor is not factored into the estimation in APEA, both the total capital and operating costs, including the reactor, were adjusted proportionally based on the installation costs calculated from APEA.The internal rate of return (IRR) is assumed to be 42.5% in the economic analysis.
The installation cost was calculated using the following equation:  Here, C is the total plant Inside Battery Limits (ISBL) capital cost, including engineering costs; F is the Lang factor (F = 3.63 for processing mixed fluids-solids); C e is the total delivered cost of the equipment.

Figure S2 .
Figure S2.DTGA traces of PET-G card and PET-G pellets.

Figure S5 .
Figure S5.Principal component analysis plot of the 42 payment card samples used in this study.The ATR-IR spectra were pre-processed using a combination of baseline correction (automatic weighted least squares), 1 st derivative processing (SavGol, filter width = 7, polynomial order = 2), normalization, and mean-centering.Venetian blinds were used as the cross-validation method (10 splits and 1 sample/split).The 95% confidence ellipse indicates how the payment cards are grouped based on their different plastic composition.The PLA cards are majority PLA laminated with thin layers of PVC, the PET cards are majority PET laminated with thin layers of PET-G, the PVC/PET-G cards are majority PET-G laminated with PVC, and the PVC/PLA cards are majority PLA laminated with PVC.

Figure S10 .
Figure S10.Overall impact assessment of DBU synthesis.

Figure S15 .
Figure S15.Simplified process flow diagram of the chemical recycling of payment cards.

Table S2 .
Elementary flows of input and output data for the background process (BHET and CHDM) 2.0 g) and Cu-Zn/Al 2 O 3 catalysts (0.1 g) were charged into the reactor, which was again purged with H 2 to remove air.The reactor was then pressurized to the desired H 2 pressure (10.5 MPa) and heated to the desired temperature (563 K) under stirring at 400 rpm to obtain the final product, CHDM. 2 Scheme S2.Synthetic procedure for the hydrogenation of BHET to CHDM. 2T (tonne FU -1 )CHDM (tonne FU -1 )Note: BHET = bis(2-Hydroxyethyl) terephthalate; CHDM = 1,4-cyclohexanedimethanol; FU = functional unit.

Table S4 .
Resources input, output, and energy consumption for the foreground process (Scenario 1, modified to increase water use to 5 tonnes per day)

Table S5 .
Tabulate data for assessing the overall environmental impact of two end-of-life scenarios in the chemical recycling process.S1: Recover all components.S2: Recover only metals and BHET, with the remaining components disposed of as wastewater

Table S6 .
Tabulate data for the major environmental impact of Scenario 1

Table S7 .
Tabulate data for the major environmental impact of Scenario 2

Table S8 .
Input parameters for the metal recovery process

Table S9 .
Input parameters for the BHET recovery process

Table S10 .
Input parameters for the distillation columns All columns in the distillation process are equipped with a total condenser and a kettle reboiler.The key variables considered include the reflux ratio and distillate to feed ratio.

Table S12 .
Parameters used for utility modeling