Towards low-cost and sustainable activated carbon production: influence of microwave activation time on yield and CO2 uptake of PET-derived adsorbents

Microwave (MW) heating is proposed as a method to transform polyethylene terephthalate (PET) into porous adsorbents. The yield, textural properties, and CO 2 uptake of the PET-derived adsorbents irradiated at different durations (3 – 35 min) at 400 ◦ C were assessed. MW activation time influenced both the physical properties and CO 2 uptake capacities of the resulting adsorbents. The yield decreased with activation time, but the surface area, total pore volume, micropore volume, and CO 2 uptake capacities all increased with MW activation time before declining. The optimal sample (produced with 5 min of MW activation time) showed improved textural properties as well as higher equilibrium and dynamic CO 2 uptakes than the commercial activated carbon used as reference. This adsorbent also possesses good selectivity for CO 2 in the binary 10:90%vol/vol CO 2 : N 2 mixture. Additionally, an excellent recyclability over 20 cycles, (Regeneration Efficiency > 97%) was observed, and the CO 2 adsorption kinetics best fits Lagergren ’ s pseudo-second-order model. This study has shown that a low activation temperature (400 ◦ C), a short MW activation time (5 min), and a low amount of chemical agent (KOH, 0.72 M) could produce CO 2 adsorbents from a cheap and abundant material (PET-waste) with better CO 2 uptake to that of a commercial activated carbon.


Introduction
The continuous buildup of CO 2 concentration in the atmosphere has led to significant changes in the Earth's climate.In 2022, the global concentration of CO 2 emission reached over 36 billion metric tonnes [1], and this value is expected to increase considering the current scenario.The current mitigation strategies, though effective to some extent, are plagued by a wide array of setbacks, most of them the high cost occurred by the increased energy requirement, among other factors.Adsorption using carbon-based adsorbents (CBAs), such as activated carbons (ACs), is effective and materials are inexpensive as they are produced from low-cost and abundant precursors [2].
ACs have been developed from different carbon precursors such as cotton stalk, date stones [3], nut-shell [4], jujun grass and Camellia japonica [5], lotus leaf [6], vine-shoot [7], polyethylene terephthalate [8], bagasse and rice husk [9], melamine-formaldehyde [10], water chestnut shell [11], and urea/petroleum coke [12] for CO 2 capture application.This list has continued to surge, all linked to the inherent advantages associated with preparation and application.ACs possess exceptional textural properties such as a large surface area, high pore volume, thermal stability over a wide range of temperatures, tunable surface chemistry, and the ability to adsorb CO 2 over a wide range of pressures [2].These characteristics, in addition to their ease of preparation and excellent recyclability over multiple adsorption/desorption cycles, make them more promising materials for CO 2 capture than any other solid adsorbents [13].
Activated carbons are produced by carbonization of carbonaceous precursors at temperatures typically above 400 • C to produce char and subsequent activation of the char to obtain AC using thermal and/or thermochemical treatments.The activation step usually requires a higher energy than the carbonization step, typically 700-800 • C (for chemical activation) [14] and above 800 • C [15] for physical activation to produce ACs of any substantial CO 2 uptake.Aside from the energy input, a typical activation of char to produce AC requires a long duration of more than 4 h [16] for the physical activation and more than 2 h for the chemical activation [17].The high energy requirement combined with the long activation duration contributes to the increased cost of AC production, in addition to being laborious [18].A total energy of 14, 000 kJ (4 kWh) is required for the activation of char using a conventional furnace for 2 hours at a power of 2000 W to produce AC [19].
MW-assisted chemical activation (MCA) of char to produce AC is a way to decrease energy requirements [19] and lower the overall cost of producing ACs for carbon capture.Once the char has been obtained from their precursors through carbonisation using a conventional furnace, the MW is then used for the activation using lower energy and a shorter duration.This is because MCA is usually fast and effective at a low bulk temperature due to volumetric heating produced by MW radiation.In MW heating, the thermal gradients are typically less pronounced than those produced by conventional heating [20], with a significant thermal gradient from the inside of the char particle to its cold surface, resulting in energy savings and a shorter processing time [21].MW temperature distribution during the processing of various materials is demonstrated in different studies [22,23,24].MCA has been employed in the production of ACs for CO 2 capture from different precursors such as coconut husk [25], tobacco stem [26], pecan nutshell [18], pinewood and wheat straw [19], and oil palm shell [27].The outcomes of these studies reveal that MCA improved the CO 2 uptake of ACs, reduced production time, and utilised less energy for activation than conventional activation [27].These demonstrate that MCA produces ACs of comparable or higher CO 2 uptake using a lower temperature (generally around 300-400 • C), shorter activation time (not above 30 mins), and over 10-fold lower energy input than conventional activation [19] The interaction of MW radiation with materials and the resulting properties varies with several treatment process factors, such as activation time, temperature, and MW power applied.Material response under MW radiation is governed by some factors, such as the dielectric constant (εʹ), which refers to the energy storage capacity of a material, and the dielectric loss factor (εʹʹ), which describes the ability of a material to convert MW to heat.The combination of these factors describes the total relative permittivity of a material, also called the MW absorptive capacity [28].Being that MW heating of materials is complicated and the mechanism is yet to be fully understood [29], the relationship between some treatment process factors used in the activation of char as well as their effects on the properties of the ACs is still not well established.For example, different optimal MW activation times: short (7 min) [30], moderate (15 min) [27], and long (30 min) [19] -have been reported during the preparation of ACs from different precursors.There is still no clear trend of what duration is optimal to produce AC with a specific textural property, yield, or CO 2 uptake due to the difference in the behaviour of MW radiation with materials or the relative permittivity of materials.
Based on the above, there is a research need for sustainable alternative production routes for AC, involving a deep investigation of the parameters of the MW-assisted production treatments that would result in ACs with better properties, such as higher CO 2 uptake and the least energy input given that the process activation conditions impact the adsorption capacity of produced ACs [31].Optimal production conditions significantly enhance CO 2 absorption in activated carbons (ACs).Coconut shell-based AC, activated at 750 • C with a KOH to biochar ratio of 0.5 in a one-step preparation method, achieved the highest CO 2 uptake-5.92mmol/g at 0 • C and 4.15 mmol/g at 25 • C [32].Altering the activator to thiosulphate, with a 700 • C activation temperature and a 2:1 thiosulphate to biochar ratio, resulted in uptakes of 5.31 mmol/g at 0 • C and 3.59 mmol/g at 25 • C [33].Sulfur-doped ACs, synthesized from polyphenyl sulfide resin using a two-step method, showed optimal CO 2 uptakes of 5.13 mmol/g at 0 • C and 3.64 mmol/g at 25 • C, with a 650 • C activation temperature and a KOH to resin-char ratio of 2 [34].Similarly, nitrogen-doped ACs from a macadamia nutshell and melamine reached peak uptakes using a 700 • C activation temperature and a 2:1 KOH to macadamia/melamine composite ratio [35].These studies highlight the impact of process parameters on ACs' CO 2 adsorption performance.Incidentally, plastic waste is an attractive alternative to biomass for use as precursors for AC production for CO 2 capture.
Accordingly, using a cheap, abundant, and sustainable waste plastic material, such as polyethylene terephthalate (PET), we have assessed the influence of MW activation time on the yield and CO 2 uptake capacity of an alkaline-impregnated, thermally generated polyethylene terephthalate (PET) char.To the best of our knowledge, there is no report yet regarding the production of ACs from PET waste for CO 2 capture using MW.Neither there are studies focused on the investigation of the influence of MW activation time on the yield and CO 2 uptake capacity of the resultant ACs.

Materials
For this investigation, used plastic mineral bottles were collected from local dumpsites around the city of Aberdeen, United Kingdom.A commercial AC (Norit PK1-3) was purchased from Sigma Aldrich in the UK to be used as both a reference material and microwave susceptor.An analytical grade of potassium hydroxide, KOH (ACS reagent, > 85% pellet), was also procured from Sigma Aldrich.Adsorption gases such as carbon dioxide and nitrogen (99.995% purity) were all acquired from BOC Limited, United Kingdom.

Adsorbent preparation
About 10 g of previously washed pieces of plastic mineral bottles (particle size < 5 mm) were loaded into a quartz jacket (320 mm long with external and internal outlet diameters of 16 and 40 mm, respectively) and placed in a tubular furnace (Nabertherm GmBH RT 50-250/ 11).The carbonization process was initiated by subjecting the sample to a heating ramp rate of 150 • C/h under N 2 (100 ml/min) up to 600 • C and holding it at 600 • C for 1 h.After carbonization, the sample was allowed to cool down to room temperature under N 2 (100 ml/min).The char was ground using a laboratory mortar to a obtained a powder of particle size less than 1 mm for easy dispersion during chemical activation.The produced char was then chemically activated using KOH.For the activation, 1 g of char was placed in a 0.72 M solution of KOH (prepared by dissolving 1 g of KOH in 25 ml distilled water) and stirred continuously for 8 h at 60 • C on a hot plate.The obtained mixture was filtered, and the residue was dried in the oven for 12 h at 105 • C to drive off moisture.For the MW-assisted activation, a mixture of the dried KOH-impregnated char with a 20 g Norit PK (particle size 1-3 mm), a commercial activated carbon used as MW susceptor, was placed in a quartz jacket, and activated at a constant temperature of 400 • C under an N 2 (200 ml/min) in the Flexiwave microwave instrument (MA186-001, from Milestone).The lengths of the MW activation were short (3-5 min), moderate (15 min), and long (25-35 min).Once the activation was completed, the ACs were separated from the MW susceptors using a mesh with an aperture less than 1 mm.There was no observable change in the Norit PK before and after microwave activation.The resulting ACs after MW activation were washed with a 1 M HCl solution followed by distilled water until the wastewater became neutral before they were dried in the oven at 105 • C for 12 h.The resultant ACs were named as PET-Y, where 'Y' is the MW activation duration in minutes.The yield was taken as the weight of the AC obtained after MW activation divided by the weight of the char used for activation [25].As a precautionary measure, the as-purchased Norit PK should be dried overnight prior to activation to drive off moisture to prevent uneven heating profiles or thermal runaway.

Elemental analysis
The elemental composition analysis for the determination of carbon (C), hydrogen (H), nitrogen (N), sulphur (S), and oxygen (O) contents was carried out using a Thermo Scientific FlashSmart analyser at the University of Birmingham.Before the experiment, the samples were predried in the oven at 90 • C for approximately 14 days and stored in airtight sample containers for analysis.

Adsorbent characterisation and CO 2 adsorption
The adsorbent's textural properties were assessed from nitrogen adsorption isotherms collected using a volumetric sorption analyser (ASAP 2020, Micromeritics) at − 196 • C after the samples have been degassed at 300 • C for 6 h under N 2 flow.The surface area was determined using the Brunauer-Emmett-Teller (BET) method.For a typical determination of the adsorbent's surface area using the BET method, the N 2 adsorption data was transformed into a linear plot in the partial pressure region 0.05 < P/P 0 < 0.3, with the slope and intercept of the linear plot used for the determination of the surface area [36].The total pore volume was calculated at P/P 0 = 0.49 from the Horvath-Kawazoe Equation [37], and the micropore volume was calculated using the Dubinin-Radushklevich (D-R) Equation applied to the N 2 adsorption isotherm data.
The narrow micropore volume, average narrow micropore width, and the characteristic energy were calculated using the D-R, Stoeckli-Ballerini, and Dubinin-Astakhov Equations, respectively, from CO 2 adsorption isotherms measured at 0 • C using the same ASAP 2020 sorption analyser.Due to decreased diffusion and kinetic diameter restrictions, nitrogen molecules utilised in adsorption studies at − 196 • C may not be able to reach the smallest micropores (<0.7 nm) at this temperature.In contrast, CO 2 molecules have a smaller kinetic diameter and greater diffusion rates at higher temperatures (around 0 • C), which allows them to enter these tiny pores more easily.N 2 and CO 2 isotherms were measured after the samples were degassed at 300 • C for 6 hours under N 2 .Thermal gravimetry analyses were performed in a thermogravimetric analyser (TGA/DSC 3+, from Mettler Toledo).
The CO 2 adsorption capacities of the prepared adsorbents were assessed under dynamic conditions using a thermogravimetric analyser TGA/DSC 3+ (from Mettler Toledo), while the volumetric sorption analyser (ASAP 2020, from Micromeritics) was used for the determination of the equilibrium adsorption isotherms.The initial dynamic adsorption tests were conducted in 100% CO 2 .Future studies are planned under conditions that more closely resemble post-combustion flue gases from natural gas/coal-fired plants, which contain 5/10% vol/vol CO 2 , using the optimal sample produced in this study.In a typical dynamic isothermal CO 2 capture experiment, the sample was first dried at 100 • C for 60 minutes under N 2 (50 ml/min) and then maintained at this temperature for 60 minutes to drive off moisture.After drying, the temperature was reduced to 25 • C, and then the feeding gas was changed to CO 2 (50 ml/min) for the adsorption step to commence.The latter was held for 120 minutes to ensure saturation was reached.After the adsorption stage was completed, the regeneration step commenced with a change in the feeding gas to N 2 , and with the temperature increasing to 100 • C. Afterward, the sample was cooled to 25 • C under N 2 , preparing the sample for the next adsorption-regeneration cycle to begin.This was repeated 20 times for the adsorbent's cyclability test.The quantity of CO 2 adsorbed or desorbed by the sample was calculated from the mass-time data obtained from either the adsorption or desorption steps using the gravimetric method described in detail elsewhere [38].The total amount of CO 2 adsorbed and desorbed by the adsorbents was evaluated using Eqs. 1 and 2, respectively.The stable working capacity defined as the adsorbent's effective capacity of adsorption/desorption in the adsorption and desorption cycles [39] was also elucidated from the adsorption/desorption data.

Total amount of CO 2 adsorbed
where C 1 and C 2 ….Cn represents the amounts of CO 2 adsorbed in cycles 1, 2, …n in mmol/g.
where D 1 and D 2 ….D n represents the amounts of CO 2 desorbed in cycles 1, 2, …n in mmol/g.On the other hand, the maximum CO 2 uptake of the adsorbents under equilibrium conditions was elucidated from the CO 2 isotherm measured at different temperatures (0, 25, 50 • C), and up to 1 bar.The isosteric heat of adsorption (Q st ) of CO 2 for the adsorbent was calculated from the CO 2 isotherms at different temperatures using the Clausius-Clapeyron equation [40].To assess the CO 2 selectivity of the prepared adsorbent over N 2 , the Ideal Adsorption Solution Theory (IAST) was applied to the sample N 2 isotherms at different temperatures based on Henry's law (see Eq. 3) [41].
where q i, and q j and p i andp j represent the adsorption uptakes and partial pressures of components i and j.

Modelling of CO 2 adsorption isotherm
Langmuir, Freundlich, and Sip adsorption models were used to fit the experimental adsorption isotherm results obtained for the adsorbent using the Excel solver function.These adsorption isotherm models are well-known, with their theories and assumptions already explained in detail elsewhere [42].The mathematical equations for Freundlich, Langmuir, and Sip used in this study, with an explanation of each parameter, are presented in Eqs.4-6, respectively.
Q 0 represents the CO 2 concentration (mmol/g), C e is the CO 2 pressure concentration (kPa), K F is the Freundlich constant (mmol/(g⋅kPa − 1/ m ) and 1/n is the heterogeneity factor which is a dimensionless constant.
Q 0 represents the theoretical monolayer capacity (mmol/g), K L is the Langmuir isotherm constant (kPa/mmol) and depends on the adsorbentadsorbate interaction, Q e is the amount of an CO 2 on the surface in equilibrium.
where q mS is a constant reflecting maximum adsorption capacity (mmol/ g), b is the Sips constant (bar − 1 ), and n is the heterogeneity factor (a dimensionless constant).
The goodness of fit of these kinetic models with the experimental data was assessed based on the nonlinear coefficient of determination (R 2 ), which was evaluated using Eq. 7. The difference between the experimental data and predicted values from the models (i.e., the residual sum of squares) denoted as Δq was computed using Eq. 8.
where q t, exp is the experimental value, q t, model is the predicted value E. Dan et al. from kinetic models, q t,exp − is the average value of the experimental data, n is the number of experimental data points fitted in the model, and p represents the number of parameters of the model.
where q t, exp is the experimental value, q t, model is the predicted value from adsorption isotherm model, and n is the number of experimental data point fits in the adsorption model.

CO 2 adsorption kinetics
Three empirical kinetic models, namely Lagergren's pseudo-first order, pseudo-second order, and Avrami's model, were used to assess the kinetics of CO 2 adsorption by the prepared adsorbent.These models have been widely used to predict the adsorption kinetics of CO 2 on different carbon-based materials, including activated carbons [43,44].In Lagergren's pseudo-first order, the adsorption system is viewed as a system where the adsorption rate is proportional to the number of available active sites [45].The model is useful in predicting the physical adsorption of CO 2 on solid adsorbents, and it signifies a reversible interaction between the adsorbent and adsorbate.The integrated form of the model used in this study is shown in Eq. ( 9) [46].
where q t is the CO 2 adsorbed (mmol/g) at time t (min), q e is the CO 2 adsorbed at equilibrium (mmol/g) and k 1 (min − 1 ) is the first order rate constant.
Lagergren's pseudo-second order assumes there is a strong bond between the adsorbate and the adsorbent.Unlike the pseudo-first-order model, this model represents a system whose adsorption rate is proportional to the square of the adsorption sites present on the surface of the adsorbents.The equation for the model is presented in Eq. 10.
where q t is the CO 2 adsorbed (mmol/g) at time t (min), q e is the CO 2 adsorbed at equilibrium and k 2 (mmol/g/min) is the second order rate constant.
Avrami's fractional-order kinetic model was also used to assess the behaviour of CO 2 adsorption by the prepared adsorbent.Although this model was originally developed to mimic crystal formation and a material's change of state, it has found usefulness in the study of gas adsorption on solids [45].The mathematical form of Avrami's model is represented in Eq. 11.
where q t is the CO 2 adsorbed (mmol/g) at time t (min), k (g/(mmol min)) is the Avrami's kinetic constant, and n is the Avrami's exponent.The goodness of fit of the models with experimentally obtained data was also assessed based on the nonlinear coefficient of determination (R 2 ), as shown in Eq. 7.

Results and discussion
After the MW activation of PET char with the aid of Norit PK1-3 as the MW absorber, the physical appearance of the resulting material (microwave PET-ACs) was a shiny granulated substance of particle size < 1 mm.

Yield, textural properties, and dynamic CO 2 uptake of microwave activated PET-ACs
The yield of the prepared ACs ranged from 81% to 96% (Table 1), with the highest yield (96%) obtained for sample PET-3, while the lowest (81%) was obtained for sample PET-35.The result shows that as activation time increases, the yield of AC decreases.Fig. 1a represents the plot of mass loss (wt%) against temperature for all the ACs, with the derivative TG plots depicted in Fig. 1b.The thermal decomposition profiles of the ACs are similar having two pronounced peaks of weight loss with similar rates, but different intensities in the temperature ranges: 25-180 • C and 625-1000 • C. A third peak of weight loss is observed in the range 180-420 • C with greater intensity for sample PET-35.Weight loss occurring in the temperature range 25-180 • C is attributed to the loss of pre-adsorbed water and/or CO 2 molecules, while the loss in the range 200-400 • C is likely that of surface carboxylic acids (strong and weak types) [47], and the loss observed in the range of 625-1000 • C is attributed to the loss of oxygenated groups such as hydroxyl (-OH) and carbonyls [47] that were formed during the impregnation of PET char with KOH and subsequent activation.The total weight loss for each AC over the temperature range of 25-1000 • C is displayed in Table 1.A trend of decreasing weight loss with increasing activation time was observed, though it was discontinued when activation time was increased from 25 to 35 minutes.The highest weight loss (6.19%) was obtained for sample PET-35, while the lowest (5.07%) was obtained for sample PET-25.However, the ACs were stable over the tested temperature range (25-1000 • C), with total weight losses lower than 7%.
The N 2 isotherms at − 196 • C for the MW-activated PET-ACs are shown in Fig. 1(c).All the prepared samples exhibit a type I isotherm suggestive of microporous adsorbents [48].A plateau was observed with no significant increase in the volume of N 2 adsorbed by the samples at higher relative pressures than 0.1, indicating that the adsorbents were entirely composed of micropores.The BET surface area, total pore volume and micropore volume first increased when the activation time was changed from 3 to 5 minutes before declining when the activation time was further prolonged from 5 to 35 minutes (see Table 1).The results show that the optimum textural properties were achieved at an activation time of 5 minutes.
The maximum CO 2 uptake of the ACs under dynamic conditions is shown in Table 1, while the variation of the AC's CO 2 uptake with the activation time is given in Fig. 1d.The CO 2 uptake increased from 1.67 to 1.80 mmol/g when the activation time was increased from 3 to 5 min, before reducing with further increases in activation times.This is because, under prolonged microwave heating, micropores are enlarged, thus increasing the average narrow micropore width (L 0 ) (see Table 1), and the local hot spots produced on further MW heating led to the destruction of the carbon surface, thereby limiting the CO 2 molecules from gaining access to the functional sites and resulting in reduced CO adsorption [25].PET-5 is the sample with the lowest average narrow micropore width (0.54 nm) and the greatest CO 2 uptake under dynamic conditions (1.80 mmol/g).For reference, the dynamic CO 2 uptake by the optimal sample prepared in this study is higher than that reported for (i) mixed plastic-derived AC (1.53 mmol/g) [49], (ii) microcrystalline cellulose-derived adsorbent (1.59 mmol/g) [50] and (iii) commercial AC Norit PK1-3 (1.65 mmol/g) (dynamic capture capacity was measured in this study under the same conditions).
A shorter MW activation duration helps in gaining extra active sites by opening previously closed pores and creating new ones through selective activation [25].As the activation duration exceeds the optimum, the AC's surface area reduces because a further increase in activation time results in uneven activation, leading to the destruction of some pores and thereby decreasing the surface area [25,26], From the CO isotherms at 0 • C (see Table 1 and Fig. 2a), the volume of narrow micropores (pores < 1 nm) was found to follow the same trend conveyed from the N 2 isotherm data, with the PET-5 sample having the highest narrow micropore volume (0.24 cm 3 /g).The average narrow micropore width (L 0 ) ranged between 0.54 and 0.68 nm for all the prepared ACs.The result reveals that the least narrow pores (0.54 nm) were obtained in sample PET-5, which was reflected in increased characteristic energy (E 0 ).

CO 2 adsorption isotherms
The CO 2 isotherm data collected at temperatures between 0 and 50 • C are depicted in Fig. 2. The maximum adsorption capacities under equilibrium are summarised in Table 2.The maximum equilibrium uptake displayed a similar trend to the one found with the dynamic CO 2 uptake for all the temperatures (Table 1).Sample PET-5 exhibited the highest CO 2 uptake, while the lowest uptake was recorded for sample PET-35.This may be attributed to the improved textural properties shown by PET-5 (Table 1).Interestingly, the maximum equilibrium uptake displayed by all the produced microwave-activated PET-ACs is comparable to or higher than some previously reported CO 2 adsorbents [38,51,52], This is of particular importance as the ACs from this work are prepared using a low concentration (0.72 M) of the chemical agent KOH.The CO 2 uptake capacities decreased with increasing temperature for all the ACs, confirming that CO 2 adsorption onto the ACs is exothermic and likely via physisorption [12].The plot of maximum equilibrium uptake with temperature for all the ACs is represented in  Fig. 2d.The observed decrease in CO 2 uptake by these ACs with increasing temperature may be explained by the fact that as temperature increases, the kinetic molecular energy of the adsorbate (CO 2 ) becomes greater, thereby allowing for the escape of the molecule from the surface of the ACs [18,46].Next, the relationship between the textural properties and the CO 2 uptake capacities of the prepared ACs at 25 • C and 100 kPa was assessed.The plots of CO 2 uptake against some textural parameters taken from Table 1, such as the BET surface area (S BET ), total pore volume (V t ), micropore volume (W 0, N2 ), narrow micropore volume (W 0, CO2 ) and the average narrow micropore width (L 0 ) are presented in Fig. 3.These results show that there is a good correlation between the CO 2 uptake by the prepared adsorbents and each of the textural properties plotted against.The CO 2 uptake shows a positive correlation with the increase of AC's surface area, total pore volume, micropore volume, and narrow micropore volumes (Fig. 3a-d).For the average narrow micropore width (L 0 ), however, it was observed an increase in the CO 2 uptake with the decrease in the L 0 values (Fig. 3e).This agrees with the reported by Martín et al [53], and that for the same L 0 value, other parameters such as the ones mentioned earlier have the greatest impact on the CO 2 uptake.
According to these correlations, the CO 2 uptake of the PET-ACs is shown to be determined by a combination of the BET surface area, V t , W 0, N2 , W 0, CO2 , and L 0 , and especially by the W 0, N2 (as samples' pores were mainly microporous) and L 0 .This was evidenced by the fact that samples PET-3 and PET-15, despite having the same total pore volume (V t ), micropore volume (W 0, N2 ), and narrow micropore volumes (W 0, CO2 ), only differing by the size of the average narrow micropore width (L 0 ), were found to display different CO 2 uptakes.A low CO 2 uptake capacity was displayed by samples with L 0 in the range of 0.6-0.63 nm and lower W 0 , N2, and V t .However, the samples irradiated for 5 minutes showed the highest S BET , V t, W 0, N2 , and W 0, CO2 , but the lowest L 0 was found to have the highest CO 2 uptake.The importance of textural properties such as the volume of narrow micropores (diameter < 1 nm) [4], micropore volume [8], surface area and pore volume [5], the average narrow micropore width [7], and combinations of two or more textural properties [6] to govern the CO 2 uptake of adsorbents at low temperatures (25 • C) and pressure (1 bar) have been reported in the literature.Under the mentioned conditions, the parameters which govern the CO 2 uptake the most are narrow micropore volume (W 0, CO2 ) and the average micropore width (L 0 ) [54].
From the results presented, samples that were irradiated with MW for a short time of 5 minutes exhibited improved textural properties and produced the optimal CO 2 uptake capacity in both dynamic and equilibrium conditions, compared to those activated for a longer time.Therefore, 5 minutes of MW activation time was required to produce AC with optimal CO 2 uptake from PET, using a less concentrated KOH  solution (0.72 M) compared to 2.84 M for chicken-manure derived AC [55], and 3.55 M for empty fruit brunch-derived AC [56].The use of less activating agent (KOH) is of great benefit, as this would contribute to reducing the cost of AC production, cut down on washing time and potentially reduce water consumption following chemical activation.As a result, the PET-5 sample was chosen as the reference sample for further analysis.

Elemental analysis
The elemental composition of PET char obtained from the thermal carbonisation of PET plastics and the MW-activated PET char (PET-5) are displayed in Table 3. From the results, the carbon and hydrogen contents of the PET char were higher than those obtained for the MWactivated sample.The observable decrease in carbon and hydrogen contents may be due to the release of volatile compounds during the activation of PET char [57].The results show that PET char and MW-activated char contained no sulphur, however, the MW-activated sample was found to contain some amount (0.37 wt%) of nitrogen when compared to the composition of its parent char (no nitrogen content).The presence of nitrogen in MW-activated PET char could arise from the complex interaction of MW with species such as KOH and the surface functionalities present in the activated carbon, all in the presence of N 2 gas.This observation agrees with the reported by Biti    [58], as they also identified the presence of Nitrogen in the MW-activated microcrystalline cellulose-derived ACs despite Nitrogen not being present in the parent biochar.This was attributed to the micro plasma hot spots developed from MW heating under a nitrogen environment.This reaction is believed to occur through a free radical mechanism and has been explained in detail elsewhere [59].The higher oxygen content in the MW-activated sample than in the PET char is attributed to the KOH activation process.During the chemical activation of the char, metal oxides decompose into oxides within the carbon matrix, thereby aiding oxidation and the creation of more oxygenated moieties [60,61].

Isosteric heat of adsorption and CO 2 /N 2 IAST selectivity
The isosteric heat of adsorption (Q st ) of CO 2 on sample PET-5 was found to range from 32.9 to 39.1 kJ/mol over a CO 2 loading range of 0.15-1.20 mmol/g.The values for the Q st for all temperatures (0, 25, and 50 • C) were higher at the beginning of adsorption (lower CO 2 loading) and became smaller as loading increased (Fig. 4a), showing a possible multilayer adsorption and the strengthening of intermolecular interactions between CO 2 and the ACs.This observation, which has been reported previously, indicates that the ACs produced possess a heterogeneous energy surface with multiple adsorption sites [12].The Q st values reported in this study are comparable to or higher than those reported for polyvinylidene fluoride and rice-husk derived ACs (21.3-24.2kJ/mol) [62] and 20.8-18.55 kJ/mol [63], respectively.The relatively high Q st displayed by sample PET-5, an average of 34.59 kJ/mol for the tested temperatures, compares to the typical average value for the same type of samples, i.e., highly microporous ACs [64] which may be related to the low average micropore width and high volume of narrow micropores that PET-5 was shown to possess (see Table 1), which results in stronger interactions between the CO 2 molecules and the adsorbent's surface [18].
For an adsorbent to be considered for carbon capture application, it is highly desirable to exhibit a selective uptake of CO 2 over other gases.The selectivity of CO 2 over N 2 for sample PET-5 was calculated from the pure CO 2 (see Fig. 2c) and N 2 isotherms (Fig. 4b) measured at different temperatures (0, 25, and 50 • C), using the IAST selectivity model (Eq.3).For all temperatures, the volume of CO 2 adsorbed by the sample was higher than that of N 2 , as shown by their pure isotherms.The selectivity of CO 2 over N 2 was assessed for a typical flue gas mixture consisting of CO 2 and N 2 , 10:90% vol/vol [65].Under these conditions the selectivity of CO 2 over N 2 obtained for PET-5 were 29.3, 35.1, and 39.7 for 0, 25, and 50 • C, respectively.The increase in selectivity observed with increasing adsorption temperature (Fig. 4c) could be due to the larger difference between the amount of CO 2 and N 2 adsorbed by the sample at different temperatures.Increasing CO 2 selectivity over N 2 with increasing adsorption temperatures has also been attributed to the presence of oxygen containing functional groups like hydroxyls and carbonyls on the AC's surface [66] as evidenced by the oxygen contents revealed by the elemental analysis (see Table 3).The trend obtained for selectivity with temperature is consistent with the literature [10,66] and the values obtained are consistent with those reported for other CO 2 adsorbents for the same gas composition [11,63,65].

CO 2 adsorption-desorption cycle under dynamic conditions
Although PET-5 exhibited excellent dynamic isothermal CO 2 capture capability, which is a significant factor in the selection of good adsorbents [62], the ability for adsorbents to be easily regenerated with high cyclic stability over repetitive cycles (each cycle comprises adsorption and desorption steps) is also crucial.This is of great importance as it would provide useful insight into potential stability under dynamic process conditions, desorption behaviour of the ACs, energy requirement for regeneration and by extension the associated operational cost in a typical carbon capture process.The CO 2 cyclic performance for PET-5 was measured over 20 cycles in the TGA as described in the experimental section.Fig. 5a depicts the 20 CO 2 adsorption-desorption cycles, while Fig. 5b displays the maximum dynamic CO 2 uptake for each cycle against the cycle number.The total amount of CO 2 adsorbed over the 20 cycles was 30.94 ± 0.01 mmol/g, with a stable working capacity of 1.55 mmol/g, whereas a total of 28.65 ± 0.00 mmol/g CO 2 was desorbed, with 1.43 mmol/g CO 2 desorbed during each single desorption cycle.
Using the adsorption and desorption data from the TGA, we quantified the adsorbent's desorption and regeneration efficiencies.Desorption Efficiency (DE) was calculated using Eq.11, while the Regeneration Efficiency (RE) was calculated using Eq. ( 12) [67].DE = q des /q ads × 100 (12) where q des is the amount of CO 2 desorbed per cycle, and q ads is the amount adsorbed in each cycle.
where q i is the amount of CO 2 adsorbed in other cycles, and q o is the amount adsorbed in the first cycle.
The CO 2 desorption and regeneration efficiencies of sample PET-5 against cycle number are displayed in Fig. 5c.The sample's desorption efficiency over the whole 20 cycles ranged from 90.36% to 92.93%, with individual cycles recording DE of above 92% except the first (90.36%).The difference between the DE of the first cycle and the rest of the cycles is due to the higher CO 2 adsorbed and desorbed in the first cycle, when the sample was fresh, compared to those of the other cycles.The regeneration efficiency (RE) ranged from 99.24% to 100% (with the exclusion of cycle 1).With the inclusion of cycle 1, RE ranged from 97.84% to 100%, with an average RE of 97.99 ± 0.19%, and each cycle exhibiting RE not below 97%.If cycle 1 is ignored, RE ranges from 99.24% to 100%, and an average of 99.37 ± 0.12% RE is reported.The slight changes in RE (< 1% when cycle 1 is excluded) within the entire 20 cycles may be due to the experimental procedures, which may not have been optimal [67].

CO 2 adsorption kinetics
To assess the CO 2 adsorption dynamics of the prepared adsorbent and its suitability for a real-world application, the dynamic isothermal CO 2 uptake data was fitted with three kinetic models as described previously.
Fig. 6 shows the predicted CO 2 capacities from the three kinetic models for sample PET-5 at 25 • C fitted with experimental data obtained from the TGA.The values for the kinetic parameters from the models investigated, the error values from different error models adopted, as well as the nonlinear correlation coefficients are summarised in Table 4.For the uptake, it was observed that most of the adsorption took place within the initial 20 minutes, during which 95% of the total CO 2 maximum uptake at equilibrium (i.e., the end of the adsorption) was achieved as indicated in Fig. 6.
The experimental data fitted better with the pseudo-second-order model at all stages of adsorption compared to the fit obtained for both the pseudo-first-order and Avrami models (see Fig. 6).The pseudosecond order model showed the lowest sum of the square error (SSE) and average relative error (ARE) values, the greatest R 2 , and predicted a CO 2 uptake at equilibrium which was identical to the experimental value (1.80 mmol/g) (see Table 1).To further ascertain if the pseudosecond-order model was applicable in the adsorption of CO 2 on the prepared AC, a plot of t/qt against time resulted in a straight-line graph from the origin with a coefficient of linear regression (R 2 ) of 1 (see supplementary information), confirming the adsorption process followed a pseudo-second order model [68].This implies that CO 2 adsorption on PET-5 is governed by a physisorption process.Both Avrami and the pseudo-first-order models displayed similar trends, with both models fitting well with the experimental data at the initial phase of adsorption but overestimating and underestimating at the later and equilibrium stages, respectively.The rate constants, K, for the kinetic models, were 0.42 min − 1 for the pseudo-first-order, 0.44 g/mmol/min for the pseudo-second-order, and 0.58 min − 1 for Avrami's model.The values of K for the pseudo-first and pseudo-second-order models were similar.For the best fitting kinetic model (i.e., pseudo-second-order), the value of K (0.44 min − 1 ) obtained for the adsorbent in this study is 11-fold higher than 0.04 min − 1 for palm solid waste-derived AC [69] and 12-fold higher than 0.036 min − 1 for pinecone-derived AC [70] reported for the adsorption of CO 2 .This indicates that, compared to palm solid waste or pinecone-derived ACs, the adsorption of CO 2 by MW-activated PET-derived AC is faster, which is quite outstanding as a smaller quantity of this adsorbent would be required to capture a given volume per unit time [39].To be emphasized, the adsorption of CO 2 on the prepared AC was observed to be significantly faster-11 times more than phenols on bagasse fly ash [71], and 31 times more than methane on palm kernel shell-derived ACs [72]-even though those ACs were specifically developed for adsorbing those molecules.This indicates a notably slower adsorption rate for phenols and methane despite their targeted design.

CO 2 adsorption isotherms modelling
Whilst it is important to assess the behaviour of the prepared AC under dynamic conditions as presented earlier, it would also be beneficial to evaluate the adsorption capacity of the AC under equilibrium conditions determined in terms of the volume of gas adsorbed and assessed using adsorption isotherm models.This would provide some information, such as the AC's surface features and the adsorption capacity at equilibrium.As stated previously, three adsorption isotherm models, namely Freundlich, Langmuir, and Sips, were adopted for the characterisation of the CO 2 adsorption isotherms measured.Fig. 7 displays the CO 2 adsorption isotherms at different temperatures (0, 25, 50 • C) fitted to different isotherm models for PET-5.Table 5 summarises the model parameters' values, the goodness of fit (based on the nonlinear correlation coefficient (R 2 ), and the residual sum of squares between the model's predicted CO 2 uptake and the experimental uptake Δq (%)) obtained from the fit of each model to the experimental data based on Eq. 8.
Sips model fitted the experimental data well, with the value for the R 2 close to unity and the lower percentage of the sum of the residual square error calculated for all temperatures.The values obtained for the Sips parameter (n) for all the temperatures were greater than 1.The Sips parameter usually defines the heterogeneity or homogeneity of an adsorbent's surface, with a value greater than unity indicating heterogeneity [73].Based on this, it can be confirmed that the AC's surface is heterogeneous.This agrees with the trend observed previously on the varying values of Q st with respect to CO 2 loadings.The values of the b constant (Sip adsorption affinity parameter) revealed a decrease with an increase in the adsorption temperature, indicating that the AC has a  (mmol/g) 1.80 q e cal (mmol/g) 1.80 q e expt obtained from TGA result, q e cal kinetic model using Excel solver, SSEsum of squares of the error, ARE-average relative error, R 2 -nonlinear correlation coefficient.
E. Dan et al. higher affinity for CO 2 at 0 • C than at 50 • C, which reveals a physisorption process and an indication of low regeneration energy.The other two models (Langmuir and Freundlich) were found to have the least goodness of fit (see Supplementary Information) when tested over the experimental values.At lower pressures of adsorption, the Sips model had previously been reported to provide the best fit for CO 2 adsorption experimental data [42,74], which agrees with the results found in this study.

Conclusion
MW activation time influences the yield, textural properties, and CO 2 uptake capacity of PET-derived adsorbents.A shorter MW activation time improves the yield and CO 2 uptake capacity of PET-derived adsorbents compared to a moderate or long activation time.The optimal MW activation time of 5 minutes resulted in an improved CO 2 uptake, which was found to be higher than that shown by Norit PK1-3, a commercial AC, as well as some of the published CO 2 adsorbents.Beside this, the optimal sample, PET-5, also displayed great properties as a CO 2 adsorbent, such as high thermal stability, moderate average heat of adsorption, good selectivity of CO 2 over N 2 (in a flue gas mixture of 10% CO 2 /90% N 2 ), excellent recyclability, and cyclic stability with high regeneration efficiency (97%) over 20 repetitive adsorption-desorption cycles.This study demonstrates that promising CO 2 adsorbents with significant CO 2 uptake can be produced from a cheap, abundant, and widely accessible waste material (i.e., PET plastic), using a less concentrated KOH solution (0.72 M), a moderate activation temperature (400 • C) and a short MW activation time of 5 minutes.The short MW activation time (lower than obtained in conventional activation) and moderate activation temperature employed to produce AC in this study have outstanding transformative potential, as adsorbent materials produced from waste using this highly effective heating source (microwaves) will reduce the energy inputs on the production routes, save processing time and make AC production for CO 2 capture applications more sustainable, economically feasible and therefore more attractive.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.b = Sips adsorption affinity constant, n = heterogeneity constant, q max, cal = from Sips model elucidated using Excel solver, q max, expt =obtained from experiment using the ASAP 2020, Δq is the residual sum of squares between the experimental uptake and that predicted by the models.

E
.Dan et al.

Fig. 5 .
Fig. 5. (a) CO 2 cyclic stability test at the constant adsorption temperature of 25 • C with N 2 purge for desorption at 100 • C; (b) maximum dynamic CO 2 adsorption for each cycle against cycle number; and (c) CO 2 desorption and regeneration efficiencies for PET-5.

Table 1
Yield, textural properties, and dynamic CO 2 uptake performance of MW-activated PET-ACs.
a BET surface area.b total pore volume.c total micropore volume.d narrow micropore volume (pore width < 1 mm).e average narrow micropore width.f characteristic energy.g maximum dynamic CO 2 uptake at 25 • C.

Table 2
Maximum equilibrium uptake of MW-activated PET-ACs at different temperatures.
a obtained at 1 bar using volumetric sorption analyser.E.Dan et al.

Table 3
Elemental composition (wt%) of PET-char and the 5-min MW-activated PET char.
a obtained by difference.E.Dan et al.

Table 4
Kinetic model parameters of CO 2 adsorption on sample PET-5.

Table 5
Sip model fitting parameters for CO 2 adsorption isotherm at different temperatures.