Enhancing Supercapacitive Swing Adsorption of CO2 with Advanced Activated Carbon Electrodes

Global warming due to anthropogenic CO2 emissions argues for the rapid development of efficient carbon capture technologies. Supercapacitive swing adsorption (SSA) is a gas separation technology that relies on the reversible charge and discharge of supercapacitor electrodes to absorb and desorb CO2 highly selectively and reversibly. However, thus far SSA only showed low sorption capacity, and slow sorption kinetics. Herein, it is shown that the sorption capacity can be substantially increased via the use of carbons with a higher specific capacitance. The highest gravimetric sorption capacity is measured with electrodes made from garlic‐roots derived activated carbon valuing 273 mmol kg−1 having a specific capacitance of 257 F g−1. In addition, the overall adsorption rate and productivity are improved. Cycling the electrodes for over 100 h showed highly reproducible, reversible CO2 adsorption and desorption behavior. A preliminary technoeconomic and sensitivity analysis is provided to demonstrate the potential of SSA for commercial applications.


Introduction
The escalating levels of anthropogenic carbon dioxide (CO 2 ) in the atmosphere has caused the earth's temperature to rise by ≈1.2 °C during the last five decades. [1]Currently, human activities release ≈40 billion tons of CO 2 every year. [2]According to the intergovernmental panel on climate change (IPPC), the CO 2 emissions would need to be reduced to 20 billion tons per year to limit warming to 1.5 °C compared to the preindustrial levels. [3]ndustrial point sources such as coal, natural gas, and cement industries which contain CO 2 in the range of 4%-20% account for more than half of the anthropogenic CO 2 emissions. [4,5]6] The general requirements for an efficient carbon capture technology include high sorption capacity, low energy consumption, high selectivity for CO 2 over other gases in the mixture, facile reversibility of adsorption, fast sorption kinetics, and the use of robust, long-living, and inexpensive materials. [7,8]To date, the most widely used technology for CO 2 capture is amine scrubbing, which requires the use of corrosive, volatile, and toxic solvents. [9,10]Alternative technologies for carbon capture include pressure swing adsorption (PSA) and temperature swing adsorption (TSA) on porous sorbents such as zeolites and porous carbons.PSA and TSA tend to have relatively low CO 2 /N 2 selectivity, and frequently show high sensitivity to moisture in the gas mixture. [11,12]Solidsorbents with grafted amines have also been reported.They exhibit high selectivity for CO 2 , but have high thermal energy requirements for regeneration, and limited life-time due to the oxygen-sensitivity of amines. [13,14]][17][18][19] They avoid the energy penalty associated with thermal and pressure cycling.Among them, supercapacitive swing adsorption (SSA) is a promising emerging technology capable of reversibly capturing and releasing CO 2 by charging and discharging supercapacitor electrodes. [20][23] SSA is significantly simpler than other electrochemical sorption techniques, only requiring inexpensive, robust, and environmentally benign porous activated carbons and aqueous electrolytes (e.g., NaCl, NaHCO 3 ).Supercapacitors allow for fast charge-discharge rates, have a high round-trip energy efficiency, and a long lifetime (> 10 5 cycles).This is an advantage over electrochemical techniques that rely on the use of redox-active sorbents or solvents (e.g., quinones, redox-active amines). [15,20,21]Redox-techniques tend to have a more limited cycle-life, and frequently involve high cost materials such as carbon nanotube-based electrodes, redoxactive specialty polymers, copper, special amines, and ionic liquids.[28][29] Electrochemical methods based on pH-swings, e.g.membranecapacitive deionization (MCDI), [30] electrolysis, [31] and bipolar membrane electrodialysis (BPMED), [32] have also been reported.However, MCDI require expensive materials such as cation and anion-exchange membranes. [30]Electrolysis and BPMED capture methods also require cation and anion exchange membranes and have high energy consumption (>300 kJ mol −1 ). [5,15][35] In PCET, the increase or decrease in pH is related to reduction or oxidation of organics, or H + intercalation/deintercalation on metal oxides.However, PCET of organics is oxygen sensitive and requires the use of ion-selective membranes to maintain a pH gradient.PCET of Metal oxides has not been investigated for long term stability.Moreover, metal oxides have poor electrical conductivity and smaller charge/discharge cycle life compared to activated carbons. [36]Table S1 (Supporting Information) provides a comparison of different electrochemical carbon capture methods in terms of energy consumption, current density, and operation mode.
In 2014, our group first demonstrated the concept of SSA by separating CO 2 from 15% CO 2 /85% N 2 gas mixture, with extremely high selectivity of CO 2 over N 2 . [37][40][41] The highest adsorption capacity achieved for SSA modules with radial gas flow was ≈70 mmol kg −1 at an energy consumption of ≈100 kJ mol −1 . [41]ecently, Forse et.al. reported that the adsorption capacity of SSA can be increased to 112 mmol kg −1 by cycling the electrodes between −1 V and +1 V instead of −1 V and 0 V but at a very high energy penalty (751 kJ mol −1 ). [42]Further improvement in adsorption capacity without compromising on the energy consumption is highly desirable.
SSA is believed to function according to three mechanisms.One, the ionic liquid-solid mechanism, assumes that CO 2 hydrolyses in the electrolyte to form protons (H + ), bicarbonate (HCO 3 − ), and carbonate (CO 3 2− ) ions that can adsorb to the double-layers of the negative (as H + ) and positive (as HCO 3 − /CO 3 2− ) electrodes, respectively.Due to their high chargedensity, small radius, and high mobility, protons are preferentially adsorbed to the negative electrode, regardless of the type of ions in the electrolyte. [41]The proton adsorption leads to an increase of the pH in the bulk electrolyte that allows for more CO 2 dissolution (and subsequent hydrolysis) into the bulk electrolyte, until the chemical equilibrium is re-established.The two other mechanisms assume the adsorption of molecular CO 2 to the double-layer due to a solubility difference of CO 2 in the bulk solution and the double-layer (molecular liquid-solid mechanism), and the physisorption of molecular CO 2 to hydrophobic pores that are not infiltrated by the electrolyte (molecular gassolid mechanism).The ionic liquid-solid mechanism of SSA suggests that increasing the specific capacitance of the electrodes could lead to enhanced adsorption capacity, and that there may be a relationship between the adsorption capacity for CO 2 and the specific capacitance of the electrodes.The previously investigated BPL carbon only shows a moderate specific capacitance of ≈80 F g −1 , indicating substantial room for improvement through the use of carbons with higher specific capacitances.
Herein, we investigate biomass-(garlic skin, garlic roots, garlic powder), coal-, coke-, and carbide-derived activated carbons as electrode materials for SSA.Garlic roots, skin, and powder were selected because of the very high specific capacitance of activated carbons reported for garlic skin-derived carbons. [43,44]or comparison, we also investigated commercially available activated carbons from different non-biomass sources, namely Y carbon (derived from silicon carbide), [45] and AX-31M Supersorb carbon (derived from coke). [46]These materials exhibit varying porosity, pore size distributions, and surface functionalities that allows for studying the influence of these variables on the SSA performance.Our previously used BPL carbon (derived from bituminous coal) was used as benchmark material. [47]Two different electrode sizes (4 and 49 cm 2 ) were used to study the dependence of electrode size on the performance of SSA.

Porosity and Compositional Analysis
Biomass-derived carbons were synthesized through carbonization at 600 °C and subsequent KOH activation at 800 °C.The detailed synthesis procedure is discussed in the experimental section.The activated carbons show a significant N 2 uptake at P/P 0 < 0.05, referring to a microporous structure, [48] as seen from the type-1 N 2 sorption isotherms (Figure 1a).The activated carbon prepared from garlic roots (GR) and garlic skin (GS) continue to adsorb N 2 at P/P 0 > 0.2, indicating the presence of abundant mesopores in addition to micropores.The activated carbons obtained from garlic powder (GP), garlic roots (GR), and garlic skin (GS) exhibit high Brunauer-Emmett-Teller (BET) surface areas of 2341, 2454, and 2582 m 2 g −1 , respectively.Coke-derived supersorb carbon (S), coal-derived BPL carbon, and carbide-derived Y carbon exhibit surface areas of 2715, 1023, and 1109 m 2 g −1 , respectively.A direct relation between surface area and pore volume is observed (Figure 1b).However, supersorb (S) activated carbon and garlic roots (GR) activated carbon flip positions with regards to their pore volumes.A maximum pore volume of 1.41 cm 3 g −1 is obtained for GR activated carbon.The electrode densities of biomass-derived carbons are oppositely related with surface areas and follow the sequence: GR (0.19 g cm −3 ) < GS (0.25 g cm −3 ) < GP (0.28 g cm −3 ) (Figure 1c).Commercial activated carbons exhibit higher electrode densities than biomass-derived carbons and follow the sequence: S (0.31 g cm −3 ) < BPL (0.48 g cm −3 ) = Y (0.48 g cm −3 ).The micropore surface areas and micropore volumes of biomass-derived carbons (GP, GR, GS) are lower than commercial activated carbons (S, BPL, Y), as shown in Table S2 (Supporting Information).
Information about the surface structure and composition of activated carbons (ACs) was obtained from X-ray photoelectron spectroscopy (XPS).The survey scan of all ACs shows two main peaks at 284 and 532 eV, attributed to C1s and O1s (Figure 2a).The N1s peaks at 400 eV are not visible in the survey spectra while small peaks at ≈1000 eV are due to oxygen auger electrons.The surface chemical state of each AC was obtained by deconvoluting the high resolution C1s and O1s spectra (Figure 2b; Figure S2, Supporting Information).Three surface chemical states of carbon (C-C 3 sp 2 , C-C 4 sp 3 , sp 2 /sp 3 C-O x-C y (x = 1,2; y = 1-3)) are  identified in all the samples.The percentage (atomic %) of sp 2hybridized C-atoms is highest in GR (83.5%) and follows the sequence GR > GS > Y > BPL > GP > S. The percentage of sp 3 hybridized C-atoms follows the sequence S > GS > Y > BPL > GR > GP.The percentage of C-O follows the sequence S > BPL > GP > Y > GR > GS.Higher sp 2 character and presence of oxygen functionalities are desirable for improving the conductivity and pseudocapacitance, respectively.The percentage of OH functionality is highest in GR (4.0%) and follows the sequence GR > GP > GS > S > BPL > Y.Among the six carbons, GS, S, BPL, and Y contain very small amounts of C─N bonds (Figure 2b).The el-emental percentage of C, O, N, and other elements is given in Table S3 (Supporting Information).

Cyclic Voltammetry (CV) and Electrochemical Impedance Spectroscopy (EIS) Analysis
The cyclic voltammograms of the supercapacitors from biomassderived and commercially available carbons show typical electrochemical double-layer (EDL) behavior at scan rate of 1 mV s −1 (Figure 3).GP-AC, GR-AC, and BPL-AC maintain the EDL behavior and high charge storage ability also at scan rates of 3 and 5 mV s −1 (Figure 3a,b, and e).S-AC and Y-AC show deviation from ideal EDL behavior as the scan rate increases from 1 to 5 mV s −1 .GS-AC shows the maximum deviation from EDL behavior even at 1 mV s −1 due to high resistance and slow charge transport kinetics of the electrodes.GS-AC was carbonized at 600 °C for 1 h at a heating rate of 6.5 °C min −1 and activated at 700 °C for 1 h at a heating rate of 5.5 °C min −1 .On the other hand, GP-AC and GR-AC were carbonized at 600 °C for 2 h at 5 °C min −1 and activated at 800 °C for 2 h at 5 °C min −1 .The synthesis conditions were modified to achieve a reasonable yield for GS-AC (≈11%).GS activation at 800 °C led to no yield at all.Nevertheless, GS-AC electrodes showed high specific capacitance and high adsorption capacity, comparable to GP-AC and GR-AC during the supercapacitive swing adsorption (SSA) tests (at a constant current of 50 mA g −1 ), as will be discussed in the later section.
To quantify the kinetics of charge storage, surface-controlled, and diffusion-controlled capacitances were calculated using Trasatti's method, [49] presented in the Figure S3 and Table S4 (Supporting Information).The best kinetic behavior was shown by the electrodes with higher double layer capacitance from outer surface (C DL ) and lower diffusion-controlled capacitance from the inner surface (C Diffusion ) and follows sequence from highest C DL to lowest C DL as; BPL-AC > GP-AC > Y-AC > GR-AC > S-AC > GS-AC.The details of the Trasatti method and calculations for capacitance contributions are discussed in the Section S1 (Supporting Information).EIS analysis was carried out to gain insights into the charge transport kinetics and contribution from different resistances (Figure S4, Supporting Information).The detailed quantification and analysis of solution resistance, charge transfer resistance, diffusion resistance, real capacitance, and imaginary capacitance for six different carbons is provided in Section S2 (Supporting Information).The experimental setup for electrochemical testing is shown in Figure S5 (Supporting Information).

Galvanostatic Charging/Discharging (GCD) and CO 2 Adsorption Analysis
Figure 4 shows the CO 2 concentration change with respect to time and voltage for 4 cm 2 electrodes in supercapacitive swing adsorption (SSA) experiments over five cycles.For each cycle, a decrease in concentration is initially observed as the electrodes are charged.Similarly, an increase in CO 2 concentration is observed during discharging the electrodes.At full saturation, the CO 2 concentration returns to a value of ≈15%, indicating the completion of adsorption/desorption half cycles.An increase in concentration somewhat above 15% near the end of the holding step at −1 V is seen for GP-AC, S-AC, and BPL-AC electrodes.This may be caused by some electro-oxidation of the activated carbon surface while the electrodes are charged during the first few cycles, forming additional CO 2 .The phenomenon of carbon electro-oxidation is well-known from the related capacitive deionization technology, since the standard reduction potential for carbon oxidation is within the charging potential window. [50]The electro-oxidation effect fully disappears after twenty chargingdischarging cycles (Figure S6, Supporting Information), and the electrodes show highly reproducible adsorption-desorption behavior.The cycling of BPL carbon electrodes over 60 cycles shows consistent CO 2 adsorption-desorption behavior and stable SSA  The GCD curves and CO 2 concentration profiles in Figure 4 were used to quantitatively evaluate the energetic and adsorptive performance of the electrodes (Figure 5 and Table S6, Supporting Information).The specific capacitance of the electrodes followed the sequence GR-AC (257 F g −1 ) > GP-AC (233 F g −1 ) ≃ S-AC (230 F g −1 ) > GS-AC (196 F g −1 ) > Y-AC (135 F g −1 ) > BPL-AC (85.9 F g −1 ) (black dots in Figure 5a).The specific capacitance is inversely correlated with the density of the electrodes that follows the sequence GR-AC (0.19 g cm −1 ) < GS-AC (0.25 g cm 3 ) < GP-AC (0.28 g cm −3 ) < S-AC (0.31 g cm −3 ) < Y-AC (0.48 g cm −3 ) = BPL-AC (0.48 g cm −3 ).Lower density tends to increase gravimetric specific capacitance due to higher porosity, and more effective utilization of the carbon atoms for EDL formation (a greater fraction of the carbon atoms is on a pore surface).GR-AC has the greatest specific capacitance and the lowest density (0.19 g cm −3 ).BPL-AC and Y-AC have the highest densities and the lowest specific capacitances.An outlier is the GS-AC that has the secondlowest density, but only the fourth-highest specific capacitance.This may be explained by the significantly higher resistance in GS-AC.BPL-AC and Y-AC have the same density, but different specific capacitances, which may be attributed to their different raw materials (bitumen coal and carbide, respectively), and the different activation methods.
The gravimetric adsorption capacity of the electrodes increased significantly from 70 mmol kg −1 for BPL-AC electrodes to 273 mmol Kg −1 for GR-AC electrodes and followed the sequence GR-AC > GP-AC > GS-AC > S-AC > Y-AC > BPL-AC (blue dots in Figure 5a).The gravimetric adsorption capacity was calculated based on the mass of the top electrode without considering the mass of other cell components.The trend of increase in adsorption capacity is consistent with increasing specific capacitance of the electrodes, except for GS-AC and S-AC, which show the third and fourth highest adsorption capacity but flip positions for the trend in specific capacitance.The higher adsorption capacity of GS-AC (247 mmol Kg −1 ) as compared to S-AC (201 mmol Kg −1 ) can be attributed to its higher mesoporous surface area compared to S-AC, which ease the transport of ions and electrons for more effective utilization of surface area.GP-AC shows the second highest adsorption capacity (255 mmol Kg −1 ) while Y-AC shows the second lowest sorption capacity (84.5 mmol Kg −1 ), consistent with the specific capacitance trends.Activation treatment of carbons was necessary to achieve higher capacitance and higher adsorption capacity values.Without the activation treatment, biomass derived carbon showed negligible capacitance and no CO 2 sorption during charging/discharging.The electrodes prepared by mixing carbon black, gluten, and PTFE without the addition of activated carbon also showed very low CO 2 adsorption with large noise in the data (Figure S8, Supporting Information), arguing that high surface area activated carbons are indispensable for SSA.[53] Table S8 (Supporting Information) shows a comparison of capacitances of some recently reported biomass-derived porous carbon electrodes.
The volumetric performance metrics of the electrodes are also practically relevant because they influence land use, the capital cost for housing the modules, and the ratio between nonadsorbing components (current collector, separator, gas diffusion layer), and the adsorbing electrodes.The volumetric capacitance (C V ) of the electrodes follows the sequence: S-AC > Y-AC > GP-AC > GS-AC > GR-AC > BPL-AC (Figure 5b and Table S6, Supporting Information).However, the volumetric adsorption capacities follow the sequence: GP-AC > GS-AC > S-AC > GR-AC > Y-AC > BPL-AC.Capacitance and adsorption capacity are not as strongly correlated on a volumetric basis, unlike gravimetric capacitance/adsorption capacity.GP-AC, GS-AC, and S-AC show the volumetric adsorption capacities of 65.3, 52.9, and 47.3 mol m −3 and exhibit volumetric capacitances of 62.3 F m −3 , 49.0 F cm −3 , and 71.3 F cm −3 , respectively.The biggest discrepancy between capacitive and adsorptive performance exists for Y-AC that ranks second for capacitive, but fifth for adsorptive performance.The results indicate that volumetric adsorption capacity does not solely depend on the volumetric capacitance but also on factors such as the carbon source, surface functionalities, and pore size distribution.Nonetheless, the approximately four-fold improvement in gravimetric adsorption capacity (GR-AC) and two-fold improvement in volumetric adsorption capacity (GP-AC) as compared to BPL carbon electrodes show that biomass derived carbons can be an inexpensive and efficient source for capturing CO 2 using SSA, and there might be room for further improvement.Differences between gravimetric and volumetric adsorption performances can be explained by the different densities of the materials.GR-AC has the lowest density, and therefore this material is able to use the carbon mass most effectively.This does not translate to the highest volumetric capacity because the low density creates "wasted" pore space that cannot be used for adsorption.GP-AC has a higher density and is not as effective on a mass basis because less many carbon atoms are accessible for adsorption.On the other hand, the higher density leads to less "wasted" pore space, which explains the high volumetric performance.
The kinetics of SSA for the different electrodes was studied by calculating the adsorption rate (AR, μmol kg −1 s −1 ) and productivity (P, mmol h −1 m −2 ) of CO 2 adsorption on the electrodes (Figure 5c and Table S6, Supporting Information).The adsorption rate used here is defined as the number of moles of CO 2 adsorbed at the full adsorption capacity over the time required to reach full adsorption capacity, and the mass of the negative electrode.The adsorption rate followed the same sequence as gravimetric adsorption capacity, i.e., GR-AC > GP-AC > GS-AC > S-AC > Y-AC > BPL-AC.An approximately three-fold increase in adsorption rate was seen for GR-AC (83.6 μmol kg −1 s −1 ) compared to BPL-AC (28.7 μmol kg −1 s −1 ).This can be explained by greater porosity of GR-AC, which allows faster mass transfer, and improved kinetics.The productivity of the electrodes followed the same trend as the volumetric adsorption capacity, with GP-AC and GS-AC showing the highest productivities of 46.1 and 39.9 mmol h −1 m −2 , respectively.Since fast sorption kinetics and high productivity are directly related to the cost of CO 2 capture, a further increase in CO 2 adsorption rate and productivity would be highly desirable for SSA.Studies that quantitatively determine the factors that limit kinetics should be performed.One possibility is that slow CO 2 hydrolysis in the neutral electrolyte limits the kinetics.This could be addressed by the addition of CO 2 hydrolysis catalysts to the electrolyte.Possibly, also the pore surfaces of the activated carbon could be modified so that the carbon acts as a heterogeneous catalyst.In our XPS studies, GR-AC and GP-AC showed a higher degree of -OH functionality while GS showed both -OH and C-N functionality, and these three biomass-derived carbons exhibit the highest adsorption rates.These results argue that pore surfaces with nucleophilic, chemical functionalities may enhance the CO 2 sorption rates during capacitive charging/discharging of the electrodes.
The coulombic efficiency ( c ) of the electrodes lies in the range of 75%-90% and follows the sequence GR-AC (87%) ≃ Y-AC (87%) > S-AC (81%) ≃ GS-AC (80%) > GP-AC (76%) ≃ BPL-AC (75%) (Figure 5d).No correlation was found between  c and any other parameter, suggesting that the ratio of charge recovered during discharging to the charge stored during charging depends specifically on the type of electrode material.Similarly, the energy efficiency ( e ) did not depend on any other performance metric and followed the sequence GR-AC (54.2%) ≃ Y-AC (54.1%) > S-AC (52.8%) > BPL-AC (52.4%) > GP-AC (49.6%) > GS-AC (16.7%) (Figure 5d).The exceptionally low  e of GS-AC is due to the very high resistance (235 Ω cm 2 ) of these electrodes (Figure 5e).The higher resistance may be due to modified synthesis conditions and greater sp 3 character of GS-AC compared to other biomass-derived carbons, as shown in XPS surface chemical state analysis.Despite the high resistance, GS-AC electrodes showed high specific capacitance and high adsorption capacity, comparable to GP-AC and GR-AC.Among commercial carbons, S-AC shows the highest resistance due to its highest sp 3 character and smallest sp 2 character.BPL-AC, GP-AC, Y-AC and GR-AC exhibit reasonable resistance values of 7.2, 9.2, 13.6, and 25.5 Ω cm 2 , respectively.The energy loss is lowest for BPL-AC electrodes and highest for GS-AC electrodes and follows the sequence BPL-AC (1.84 J) ≃ GR-AC (1.88 J) < GP-AC (2.73 J) ≃ S-AC (2.74 J) < Y-AC (2.78 J) < GS-AC (5.40 J) (Figure 5e and Table S6, Supporting Information).
The charge efficiency (CE) signifies the ratio of adsorbed CO 2 over the charges stored in the electrodes.Biomass-derived activated carbons (GP-AC, GR-AC, and GS-AC) exhibit higher CE than commercial activated carbons (S-AC, BPL-AC, and Y-AC) and follow the sequence GP-AC (0.24) > GS-AC (0.17) > GR-AC (0.16) > S-AC (0.13) > BPL-AC (0.12) > Y-AC (0.10) (Figure 5f).High charge efficiency of biomass-derived electrodes is still far less than the theoretical maximum value of 1.0 (assuming the ionic liquid-solid mechanism as the only mechanism).][56] In addition, the conversion of HCO 3 − ions to form CO 3 2− that have a greater affinity to the anode due to their higher negative charge may reduce the charge efficiency.CE ranging from 0.5 to 0.7 has been achieved by using ion-selective membrane, thinner electrodes (0.25 vs. 0.70 mm used in this study), and pure DI water instead of salt solutions in a related membrane-capacitive deionization process (MCDI). [30]However, the added cost of membranes along with their mass-transfer limitations, mechanically weak electrodes with less absolute capacitance, and low ionic conductivity of the CO 2 sparged DI water are disadvantageous for capturing CO 2 , especially at low concentrations.Separate studies requiring in situ characterization are needed to gain insights about the adsorbed species at the double-layer, and possible faradaic surface reactions.These insights may then be used to improve the selectivity of the electrodes for H + /HCO 3 − over the electrolyte ions.The energy consumption (EC) signifies the amount of energy required to adsorb one mole of CO 2 and was calculated from the ratio of energy loss to absolute amount of CO 2 adsorbed.The overall sequence is as follows: GP-AC (177 kJ mol −1 ) < GR-AC (179 kJ mol −1 ) < S-AC (204 kJ mol −1 ) < BPL-AC (243 kJ mol −1 ) < Y-AC (304 kJ mol −1 ) < GS-AC (371 kJ mol −1 ), (Figure 5f).GP-AC, GR-AC, and S-AC have higher surface areas and absorb larger amounts of CO 2 relative to the energy loss, thus resulting in smaller energy consumption values.However, despite the high surface area and high adsorption capacity of GS-AC, the very high energy loss results in large energy consumption.BPL-AC and Y-AC exhibit lower surface areas and adsorb much less amounts of CO 2 relative to the energy loss, resulting in higher energy consumption.Our group reported an energy consumption of ≈100 KJ mol −1 with 78% energy efficiency using 7 cm BPL-AC electrodes in 5 M NaCl and 1 M MgBr 2 electrolyte, arguing that improved energy consumption may be achievable by use of 5 M NaCl, or 1 M MgBr 2 . [57]More recently, we reported energy consumption values of ≈70 KJ mol −1 using hot-pressed garlic-roots derived electrodes, [58] arguing that development of new electrode materials is another promising route to further minimize the energy consumption values.

Experiments with Scaled Electrodes (49 cm 2 vs. 4 cm 2 )
We further investigated if the performance of the module was dependent on the size of the electrodes, by preparing additional electrodes having an increased area of 49 cm 2 .The voltage response and CO 2 concentration profiles for all the electrodes exhibit similar adsorption-desorption behavior (Figure 6).The overall change in concentration increased during charging and discharging, owing to the longer distance that CO 2 has to travel through the electrodes.Charging of 49 cm 2 GS-AC electrodes resulted in a CO 2 concentration decrease from 15% to ≈5% (Figure 6c) while only 1%-2% decrease in CO 2 concentration was observed with 4 cm 2 electrodes.The size of the electrodes was limited only by the size of the Ti current collector end plates.Further scaling of the end plates would allow for even larger electrodes and may produce a pure N 2 effluent stream upon charging the SSA module.
The energetic and the adsorptive metrics for 49 cm 2 electrodes are shown in Figure 7 and Table S7 (Supporting Information).The trend of capacitance and gravimetric adsorption capacity is the same as 4 cm 2 electrodes, with GR-AC, GS-AC, and GP-AC exhibiting the maximum gravimetric adsorption capacities of 198, 194, and 189 mmol Kg −1 , respectively.S-AC, Y-AC and BPL-AC show significantly lower values of 119, 69.2, and 62.8 mmol kg −1 (Figure 7a and Table S7, Supporting Information).The volumetric sorption capacity follows the sequence: GS-AC (36.7 mol m −3 ) ≈ GP-AC (36.3  7b).This is similar to the trend seen for 4 cm 2 electrodes.Generally, sorption capacities are lower for the 49 cm 2 electrodes on both a gravimetric and volumetric basis.Nonetheless, there is still a 3.2-fold improvement on a gravimetric basis (198 vs. 62.8 mmol kg −1 ) and a 1.6-fold increase (36.7 mol m −3 vs. 23.1 mol m 3 ) on a volumetric basis compared to BPL-AC.The productivity, and the adsorption rates of 49 cm 2 electrodes are also less compared to 4 cm 2 electrodes.The lower adsorption capacity, adsorption rate, and productivity may be because the gas flow may not be exactly radial for the 49 cm 2 electrodes.There could be preferential gas flow from the center to the exit hole at the periphery of the bigger electrodes caused by a pressure gradient within the smaller void volume around the electrodes.Increasing this void space by increasing the size of the current collector plate may eliminate the pressure gradient inside the void space.The adsorption rate and productivity follow the same trend as gravimetric adsorption capacity and volumetric adsorption capacity, respectively (Figure 7c).The coulombic efficiency of 49 cm 2 electrodes is higher than for the 4 cm 2 electrodes (Figure 7d).GS-AC electrodes show a coulombic efficiency of 94.9%, indicating that the charge-discharge process is highly reversible.The energy efficiency and energy loss follow a similar sequence to 4 cm 2 electrodes with BPL-AC showing the highest energy efficiency (64.1%) and smallest energy loss (12.1 J) while GS-AC exhibits the lowest energy efficiency (14.8%) and highest energy loss (55 J) (Figure 7d and e).Generally, higher resistance values and lower capacitance values are seen for the 49 cm 2 electrodes as compared to the 4 cm 2 electrodes.This may be explained by the fact that the same torque (20 in.lb) was used to assemble the SSA modules with different electrodes sizes, possibly resulting in higher contact resistance for the 49 cm 2 electrodes.The results agree with EIS comparison of 4 cm 2 and 49 cm 2 BPL electrodes (Figure S9, Supporting Information), where cm 2 electrodes show slightly higher resistance and slightly lower specific capacitance as compared to 4 cm 2 electrodes.Adjusting the torque and packing pressure may result in a size-independent energetic performance of the SSA modules.
The energy consumption is lowest for GR-AC electrodes and follow the sequence GR-AC (133 kJ mol −1 ) < BPL-AC (165 kJ mol −1 ) < GP-AC (245 kJ mol −1 ) < S-AC (327 kJ mol −1 ) < Y-AC (353 kJ mol −1 ) < GS-AC (439 kJ mol −1 ) (Figure 7f).Except GR-AC and BPL-AC, the 49 cm 2 electrodes exhibit higher energy consumption than 4 cm 2 electrodes.This can be attributed to lower energy loss and better energy efficiency of GR-AC and BPL-AC.Since the energy consumption is the ratio of energy loss to the moles of CO 2 absorbed, scaling GR-AC and BPL-AC from 4 to 49 cm 2 led to ≈7 times increase in energy loss and ≈9 times increase in the moles of adsorbed CO 2 .In contrast, scaling of GP-AC and S-AC lead to ≈12 times increase in energy loss and ≈7.5 times increase in the moles of CO 2 absorbed.GS-AC and Y-AC electrodes show ≈10 times increase in energy loss and 8.5 times increase in the moles of adsorbed CO 2 .The smallest increase in energy loss of GR-AC and BPL-AC explains the smallest energy consumption of these two electrodes.The charge efficiency of the electrodes follows the sequence: GR-AC (0.16) > GS-AC (0.14) > GP-AC (0.12) ≃ BPL-AC (0.12) > S-AC (0.10) > Y-AC (0.09) (Figure 7f).Similar to the energy consumption, the charge efficiency (CE) of 49 cm 2 GR-AC and BPL-AC electrodes stays the same as 4 cm 2 electrodes but significantly decreases for GP-AC, GS-AC, BPL-AC, Y-AC.These results argue that not only the energy loss but also the absolute number of moles of CO 2 adsorbed in case of GR-AC and BPL-AC is highest among other electrodes.Overall, 49 cm 2 electrodes results in improved purity of the effluent gas but requires further optimization to avoid scaling effects that influence adsorptive and energetic performance metrics.

Preliminary Technoeconomic Analysis
Garlic root-derived activated carbon has the highest gravimetric sorption capacity, and fastest adsorption kinetics.We therefore carried out a preliminary techno-economic analysis for the garlic roots derived carbon to estimate the cost of CO 2 capture using supercapacitive swing adsorption (SSA).The preliminary technoeconomic analysis uses estimated component costs retrieved from the literature and the performance data from experiments.The current annual world garlic production is ≈28 million tons per year. [59]There is ≈6% mass percent of garlic roots per mass unit of garlic. [60]Assuming a garlic root to activated carbon conversion rate of 20%, [61] a maximum amount of 0.34 million tons of garlic root-derived activated carbon could be made per year.Based on the SSA sorption capacity of 0.20 mol kg −1 and 1 h cycle times, this amount could capture ≈26 million tons (Mt) CO 2 per year.Assuming a cycle life of 100000 cycles, the sorbent would live 11.42 years, hence on an annualized basis 299 Mt CO 2 /year could be captured, i.e., SSA with garlic-root derived carbons could be scaled up to this amount (see Section S4 for calculations, Supporting Information).Assuming a typical activated carbon production cost of $3.00 per kg, [62] a cost of $10.00 kg −1 for the PTFE binder, [63] $6.00 kg −1 for the carbon black, [64] and $1.35 kg −1 for the gluten, [65] the cost of an electrode is $3.38 kg −1 at a AC:PTFE:CB:Gluten ratio of 80:5:5:10.At an electrode thickness of 0.7 mm, and electrode density of 0.20 g cm 3 , this translates to an electrode cost of $0.47 m 2 .Application of a Lang factor of 3.63 gives $1.71 m 2 (Calculation S6, Supporting Information).The productivity at 0.20 mol kg −1 and 1 h cycle times is 0.028 mol h −1 m 2 or 0.0108 tons m −2 year −1 .Assuming an interest rate of 3% and 11.42 years electrode lifetime gives a capital recovery factor of 0.10.This results in an annualized electrode cost of $0.18 m 2 and carbon capture cost of $16.82/ton CO 2 .This does not yet consider the counter-electrode, the separator, and the electrolyte.The symmetric counter-electrode does not further increase adsorption; hence it doubles the cost to $34.64/ton.The cost of electrolyte is negligible, but the separator increases the cost to $75.91/ton CO 2 assuming a separator cost of $1.2 m 2 . [64] further consideration is the expanded graphite current collectors ($5 m 2 ), [63] and the carbon cloth as the gas diffusion layer ($9 m 2 ). [66]Figure 8(a) shows the cost breakdown of these com-ponents.The number and thus cost of current collectors can be minimized by scaling the technique using bipolar electrodes in series similar to fuel cell stacks because then only two current collectors at the end of the stacks are needed.If, for instance, there was a stack of hundred 1 m 2 electrodes, one would only need two 1 m 2 current collectors, and hence they would not significantly add to the cost (to $79/ton in this example case).The ionic current within a bipolar electrode must be interrupted to avoid the short-circuiting of the ionic circuit.This could be accomplished by "gluing" the positive and the negative electrode together using graphite cement to give the bipolar electrode, or by using carbon black containing polyethylene foils. [67]The graphite cement would clog the pores at the interface of the two electrodes interrupting the ionic circuit.At an estimated graphite cement cost of ≈$1 m 2 the carbon capture cost would increase to $115/ton (Calculations S8 to S12, Supporting Information).The carbon cloth as the gas diffusion layer is the most expensive component ($9 m 2 ), would increase cost greatly to $432/ton.We are currently exploring cell designs without carbon cloth that use electrodes with imprinted gas flow channels.This possibility arises from the large thickness of our electrodes (≈0.7 mm).
Sensitivity analysis shows how much room for cost reduction there is by the improvement of individual parameters (Figure 8(b)), here investigated for the sorption capacity, the electrode thickness, the electrode density, and the activated carbon cost based on a cell design without carbon cloth as gas diffusion layer at a cost base of $115/ton as stated above.At a sorption capacity of 50, 100, 400, 600, 800, and 1000 mmol kg −1 , the cost changes to $459, $229, $57, $38, $29, and $23 ton −1 , respectively.At an electrode thickness of 0.2, 0.4, 1.0, 1.3, 1.7, and 2.0 mm, the cost changes to $317, $176, $90, $77, $67, and $62.This is because the ratio the of adsorbing components (electrodes) relative to the non-adsorbing components (separator, carbon cement, current collector) increases with decreasing thickness.A change of electrode density to 0.05, 0.1, 0.3, 0.4, 0.6, and 0.8 g cm −3 (at constant gravimetric sorption capacity), changes the cost to $359, $196, $88, $74, $60, and $53, respectively for the same reasons.Reduced activated carbon cost ($1.5 kg −1 , $0.75 kg −1 ) changes the capture cost to $102, and $97 ton −1 .Increasing the activated carbon cost to $6 kg −1 , and $12 kg −1 , increases the capture cost to $138, and $185 ton −1 , respectively.It can be seen that the greatest sensitivity exists for the sorption capacity, and that a moderate further improvement to values of 600 mmol kg −1 can reduce the cost to < $50 ton −1 (Figure 8 (b)).The cost of sorbent is least sensitive to CO 2 capture cost because the cost of the nonadsorbing components is higher.Increase of electrode density could be achieved by greater compression.A simple and effective way to reduce cost is to increase the thickness of the electrodes, as long as this can be done without significantly decreasing the adsorption rates and increasing energy consumption due to increased resistance in the supercapacitor.The energy cost at an energy consumption of 179 kJ mol −1 is $56 ton −1 at an electricity production cost of 5c kWh −1 (Section S13, Supporting Information).The contribution of energy cost would be reduced to $31 at 100 kJ mol −1 , and $16 ton −1 at 50 kJ mol −1 , respectively.
The cost may be further reduced by crushing electrode stacks at the end of their lifetime, and using them as soil conditioner, similar to biochar.Biochar is currently sold at a price of $2500/ton indicating significant reductions in cost through the use of crushed, "spent" activated carbon electrode stacks as soil conditioner. [68]Regeneration of the activated carbon is another potential way to reduce cost.
The DOE 2025 cost target for post-combustion capture is $40/ton CO 2 . [69]Although the above techno-economic analysis is not comprehensive, it indicates that the SSA has the potential to reach the DOE cost targets with moderate improvements.For example, the target would be met at a sorption capacity of 0.6 mol kg −1 , an electrode thickness of 1 mm, an electrode density of 0.4 g cm 3 , and an energy consumption of 50 kJ mol −1 without considering the end-of-lifetime value of the electrodes.The SSA technology does not only appeal for separation of CO 2 from point sources (4%-20% CO 2 ) but also from air (0.04%) due to the combination of chemical robustness, moisture-compatibility, se-lectivity, reversibility, and low cost of materials.An investigation on the applicability of SSA to DAC is currently on the way in our laboratory.

Conclusions
We evaluated three biomass-derived (garlic roots/skin/powder) and three commercially available carbons from different sources (bitumen coal, carbide, and coke) as electrode materials for supercapacitive swing adsorption (SSA).Galvanostatic charging and discharging coupled with SSA adsorption measurements showed that higher capacitance not only led to up to four times higher gravimetric adsorption capacities but also accelerated the adsorption rates and improved productivity.There is a positive correlation between specific capacitance and sorption capacity, arguing that the development of electrode materials with increased specific capacitance is a pathway to improve the sorption capacity.A quantitative overview over the relationships between capacitance and adsorption capacity shows that GR-AC (gravimetrically) and GP-AC (volumetrically) show the best CO 2 adsorption capacities among the tested electrodes.GR and GS are abundantly available biomass waste and can be used economically for activated carbon synthesis.The positive correlation between capacitance and adsorption capacity suggests that the ionic-liquid solid mechanism plays an important role in CO 2 adsorption and desorption.Scaling of electrodes resulted in improved purity of the effluent gas but requires further optimization to avoid scaling effects that influence adsorptive and energetic performance metrics.The techno-economic analysis shows that the improvements presented herein substantially increase the practical potential for SSA as carbon capture technique from a cost perspective.
Synthesis of Activated Carbons: Synthesis protocols similar to those reported for activated carbons derived from garlic skin were used. [43]The biomass precursor (e.g., garlic roots) was thoroughly rinsed, washed several times with DI water, dried in the oven at 100 °C overnight and sieved to less than or equal to 350 microns particle size.The powdered samples were carbonized in a tube furnace in N 2 atmosphere.About 26 g of pulverized garlic roots were heated to 600 °C at a heating rate of 5.0 °C min −1 in a 100 × 45 × 20 mm alumina crucible.The sample was kept at 600 °C for 2 h and then cooled slowly to room temperature.The yield of carbonization was around 30%-35% for all the samples.The carbonized garlic roots sample was mixed with KOH (mass KOH : mass carbon = 4:1) and ≈30 mL DI water to form a homogeneous slurry.The slurry was transferred to a custom-made nickel sample holder and heated to 800 °C at 5.0 °C min −1 under N 2 in the tube furnace.The temperature was maintained at 800 °C for 2 h before the sample was cooled to room temperature.The obtained sample was dissolved in 100 mL DI water and washed with 1 M HCl (20-30 mL) ≈3-4 times to remove the remaining impurities (K, K2O, KOH, and K2CO3).It was further washed with DI water until the pH reached 7. The resulting sample was dried in the oven at 120 °C overnight to obtain activated carbon powder.The final yield from garlic roots and garlic powder varied between 10%-15% while no significant yield was obtained from garlic skin.For garlic skin, the carbonization and activation time were therefore reduced to 1 h and activation temperature was reduced to 700°C to obtain ≈11% yield from garlic skin.The activated carbon samples obtained from garlic roots, garlic powder, and garlic skin were named GR-AC, GP-AC, and GS-AC, respectively.The remaining carbons were named BPL-AC, S-AC, and Y-AC, respectively.
Characterization of Activated Carbon: Surface and porosity analysis: The N 2 sorption isotherms were measured at 77 K using Micromeritics ASAP 2020 equipment.The samples were degassed under 200°C for 24 h before the analysis.Low N 2 dosing mode was carried out before the relative pressure P/P 0 = 0.000105 in the adsorption branch.Total surface areas were calculated from the adsorption data at P/P 0 = 0.001-0.051using the Brunauer-Emmett-Teller (BET) model.Total pore volumes were calculated from the adsorption data at P/P 0 = 0.99.The area and volume information of the micropore and mesopore were calculated from the adsorption data using the t-plot method (Carbon Black STSA model).The pore size distribution curves were obtained from the adsorption data using the nonlocal density functional theory (NLDFT) method, specifically with the N2-Carbon Finite Pores, AS = 6.Here, AS = 6 means a 2D model of finite slit pores having a diameter-to-width aspect ratio of 6. Slit type geometry of pores was used for simulation.
X-ray photoelectron spectroscopy: Activated carbon powders were affixed to the sample plate using copper tape.Measurements were performed using a custom-built SPECS XPS instrument.Charge neutralization was invoked during the analyses using a beam of 2 eV electrons at a current of 20 μA.All spectra were acquired using a photon energy of 1486.6 eV Al K  , with analysis performed on electrons escaping in a direction normal to the sample surface.The pass energy for all core level scans was 20 eV and the pass energy for survey scans was 70 eV.Core level spectra were fit using pseudo-Voigt profile components comprised of ≈20% Lorentzian and 80% Gaussian broadening.In C1s and O1s spectra, the widths of the components were constrained to be equal to one another.In spectra with doublet peaks (Si2p), the relative heights, widths, and energy separation of the peaks were constrained using library values from Thermo's Advantage software.The elemental percentages were estimated by calculating the area of each core level peak and normalizing by its Scofield cross section and mean free path.
Electrodes Preparation and Supercapacitor Assembly: Figure S1 (Supporting Information) shows the scheme for the synthesis and preparation of the activated carbon electrodes.Briefly, 0.083 g of PTFE was added to 10 mL of ethanol.The solution was mixed and sonicated simultaneously for 10 min to obtain a homogeneous mixture.0.8 g of activated car-bon, 0.1 g of gluten, and 0.05 g of conductive carbon black were added to the PTFE solution.The mixture was covered with Al foil and magnetically stirred at 60-65 °C for 2 h.The temperature was then increased to 80 °C for 30 min to evaporate the remaining solvent.The resulting slurry was transferred onto a glass plate, pressed, and mixed with metal spatulas for 30 min.The obtained dough was processed through a pasta machine (Atlas 150 roller, Marcato) to achieve the desired thickness of the electrodes.The electrodes were cut to the desired size (4 cm 2 ), placed in a vacuum oven at 100°C and 25 mmHg for 12 h.The same procedure was used to prepare 49 cm 2 electrodes.
The bottom electrode was soaked in 1 M NaHCO 3 solution for 2 h.For the 4 cm 2 electrodes, a 4.84 cm 2 filter paper was wetted with the electrolyte and used as a separator.A 4 cm 2 carbon cloth (AvCarb 1071 HCB) was used to cover the top electrode and served as a gas diffusion layer.The electrodes, separator, and gas diffusion layer were surrounded by an EPDM rubber gasket (9.5 cm × 9.5 cm outer area, 8 cm × 8 cm inside area, 0.15 cm thickness, Fuel cell store) and sandwiched between titanium grade 7 plates (13 cm × 13 cm, 0.95 cm thick), which served as current collectors.The resulting assembly was pressed together by evenly tightening 16 set screws connecting the Ti plates using a torque wrench to 20 in.lb of force.For 49 cm 2 electrodes, 49 cm 2 carbon cloth and 56 cm 2 filter paper were used while the dimensions of gasket and current collector stayed the same.
Electrochemical Testing of Electrodes: The electrochemical testing was performed using a Gamry 3000 potentiostat in a two-electrode configuration.Cyclic voltammetry (CV) tests were conducted at different scan rates in the potential range of 0 to -1 V to gain information about the charge storage behavior of the supercapacitors.The specific capacitance values from CV were calculated using the relation C = Q/(ΔV × m) where Q is total charge stored during potential sweep, ΔV is the voltage window, and m is the average mass of the electrodes.
Electrochemical impedance spectroscopy (EIS) was performed at −0.5 V DC in the frequency range from 10 5 to 10 −3 Hz.Information about the solution resistance, charge transfer resistance, Warburg diffusion resistance and capacitance were extracted from the EIS data.The real capacitance, imaginary capacitance, and total capacitance were calculated using the following relations: [43,70] Here, C real , C imag , and C T represent the real capacitance, imaginary capacitance, and overall capacitance of the supercapacitor.f is the frequency, Z′ is real impedance, Z′′ is the imaginary impedance, and |Z| signifies the overall impedance of the supercapacitor.
Galvanostatic charging and discharging (GCD) was performed at a constant current density of 50 mA g −1 to test the gas adsorption and desorption behavior of different supercapacitor electrodes.The specific capacitance (C s ) and equivalent series resistance (R ES ) were found using the following relations: where I is the constant current applied during the test, t g is the galvanostatic charging duration, m is the total mass of activated carbon in top and bottom electrodes, and ∆V is the potential window.Supercapacitive Swing Adsorption (SSA) Testing: A schematic illustration of the SSA experimental setup is shown in Figure S5 (Supporting Information), and was described in detail in a previous publication. [57]A 15% CO 2 /85% N 2 gas mixture flows through a bubbler filled with 1 M NaHCO 3 .The purpose of bubbler is to moisten the gas before it enters the SSA module.This would prevent the drying out of electrodes and undesired concentration changes during the SSA tests.The moistened gas passes through a mass flow controller adjusted to a flow rate of 1 sccm.The gas mixture enters the SSA module from the center inlet port (1 mm in diameter) and flows radially through the gas diffusion layer inside the module.The gas exits from the outlet port at the corner of the top Ti plate, passes through a drying tube, and is then analyzed by a CO 2 analyzer (Quantek Instruments, Model 9-6).The gas mixture was flown through the modules for 8 h without applying any bias to complete conventional CO 2 adsorption to the electrodes and conventional CO 2 absorption by the electrolyte.After that time, the concentration of the exiting gas attained a constant value of 15% showing that conventional ad/absorption was complete.The module was then charged at a rate of 50 mA g −1 to −1 V (the gas-exposed electrode being the negative electrode at −1 V relative to the bottom (counter) electrode.The potential was then held at −1 V for an additional 30 min to allow for complete saturation of the electrode with CO 2 .After that, the module was discharged to 0 V at 50 mA g −1 followed by a 30 min holding step to complete the desorption of CO 2 gas from the electrode.
ΔE = E c − E d (11)   where V pstat is the potentiostatic voltage (either 0 V or -1 V), V drop is the voltage at the start of the charging or discharging step, I is the constant current, t g is the charging time, m is the total mass of two electrodes, ΔV is the charging potential window,  is the density of electrodes, t is the average thickness of the electrode pair, C is the absolute capacitance, E d is the energy delivered during discharging, and E c is the energy consumed during charging.The capacitance calculation assumes equal capacitance of two electrodes.

SSA Adsorptive Metrics:
The SSA adsorptive metrics include the calculation of the number of moles of CO 2 adsorbed during the charging plus the holding step (n a, CO2., μmol), adsorption capacity with respect to mass (AC, mmol kg −1 ), external electrode area (AC, mmol m 2 ), and volume (AC, mol m 3 ) of the gas-exposed electrode, adsorption rate (AR, μmol kg −1 s −1 ), productivity (P, mol h −1 m −2 ), charge efficiency (CE, dimensionless), energy consumption (EC, KJ mol −1 ), and time-energy efficiency (TEE, mol KJ −1 .s−1 ).( Where P is the pressure (1 atm), T is the temperature (296 K), R is the general gas constant, f i and f e are the influent and effluent gas flow rates, m top is the mass of top electrode, A is the area of electrodes, V is the volume of electrodes, and t is the total charging time (including the 30 min holding step).

Figure 1 .
Figure 1.N 2 sorption isotherms of garlic powder (GP), garlic roots (GR), garlic skin (GS), coke derived supersorb (S), coal-derived (BPL), and carbide derived (Y) activated carbons a).The desorption isotherms overlap with adsorption isotherms.Activated carbon type versus surface area and pore volume b), and surface area and electrode density (c) show the difference in the porosities of each carbon.

Figure 2 .
Figure 2. XPS survey spectra of six activated carbons a) and percentage contribution of C-C 3 sp 2 , C-C 4 sp 3 , C-O x -C y, OH, C-N, and other chemical states in each activated carbon b).

Figure 4 .
Figure 4. Voltage response (shown in black) and CO 2 concentration changes (shown in blue) for six different electrodes, garlic powder, GP-AC a), garlic roots, GR-AC b), garlic skin, GS-AC c), supersorb, S-AC d), BPL-AC e), and Y-AC f) The blue dotted line in the center of the plot denotes the equilibrium concentration of ≈15% CO 2 .Electrodes with an area of 4 cm 2 and thickness of 0.7 mm were used for these experiments.

Figure 5 .
Figure 5. Gravimetric capacitance and gravimetric adsorption capacity a), volumetric capacitance and volumetric adsorption capacity b), adsorption rate and productivity c), coulombic efficiency and energy efficiency d), resistance and energy consumption e), and f) charge efficiency and energy consumption of biomass-derived (GP, GR, GS) and commercial (S, BPL, Y) activated carbon electrodes.The vertical lines on data points show error bars.

Figure 6 .
Figure 6.Voltage response (shown in black) and CO 2 concentration changes (shown in blue) for six different electrodes, garlic powder, GP-AC a), garlic roots, GR-AC b), garlic skin, GS-AC c), supersorb, S-AC d), BPL-AC e), and Y-AC f) The blue dotted line in the center of the plot denotes the equilibrium concentration of around 15% CO 2 .Electrodes with an area of 49 cm 2 and thickness of 0.7 mm were used for these experiments.

Figure 7 .
Figure 7. Gravimetric capacitance and gravimetric adsorption capacity a), volumetric capacitance and volumetric adsorption capacity b), adsorption rate and productivity c), coulombic efficiency and energy efficiency d), resistance and energy consumption e), and f) charge efficiency and energy consumption of biomass-derived (GP, GR, GS) and commercial (S, BPL, Y) scaled (49 cm 2 ) activated carbon electrodes.

Figure 8 .
Figure 8. Cost breakdown of major components of the SSA module a), Sensitivity analysis showing the influence of change in sorption capacity, electrode density, electrode thickness and activated carbon cost on the CO2 capture cost in $/ton ().The baseline at 115 $/ton in (b) represents the cost achieved at an adsorption capacity of 200 mmol Kg −1 , electrode density of 0.2 g cm −3 , electrode thickness of 0.7 mm, and activated carbon cost of 3$/ton.