CHAPTER 1: Carbon-based CO₂ Adsorbents

Carbon dioxide (CO₂) has been recognized to be the biggest driver of global warming, which is one of the most serious problems that our world is facing.^1–3  Furthermore, CO₂ is an important source of C1 chemical engineering, and could be converted into high-value chemical products via chemical,^2,3  photochemical,^4,5  or electrochemical processes.^6,7  However, the efficient capture of CO₂ is essential for these processes. So, there is an urgent need to develop CO₂ capture and storage (CCS) technologies. The basic concept of CCS is to capture CO₂ from emissions without releasing it into the atmosphere. CCS can be classified as post-combustion, pre-combustion, and oxy-fuel combustion technologies. Among the current CCS technologies, post-combustion capture, a technology for capturing CO₂ from post-combustion emission gases (e.g., flue gas from power plants) is the most easily applied technology for existing emission sources.

In general, post-combustion capture technologies include chemical absorption, dry adsorption, membrane-based technologies, and cryogenic technologies. Currently, chemical absorption is the most applicable technology for CO₂ capture in power plants, but this technology suffers from several drawbacks. The biggest challenge in applying a chemical absorption process for post-combustion is how to reduce the heat of regeneration. Another problem is the release of hazardous byproducts. For these reasons, dry adsorption using solid adsorbents is considered to be promising for the capture of post-combustion CO₂.

The dry adsorption technique is a process of selective adsorption of CO₂ from post-combustion gases using solid adsorbents, which has advantages such as a simple device, easy operation, it is environmentally friendly and has a high energy efficiency. When evaluating solid adsorbents, it is important to consider their surface area, apparent density, pore size and volume, feasibility of regeneration, stability, abundance and sustainability. CO₂ capture by solid adsorbents mainly relies on the mechanism of physical adsorption that is also interfered with by some weak interactions between CO₂ and the adsorbent's surface (i.e. hydrogen bonding or electric quadrupole interactions). Due to the main contribution of van der Waals forces to the physical adsorption, materials that possess a developed microporous texture are preferred for CO₂ capture. Nowadays, many kinds of solid adsorbent materials with porous textures, such as porous carbonaceous materials,^8,9  zeolites,¹⁰  zeolitic imidazolate frameworks (ZIFs),¹¹  metal–organic frameworks (MOFs),¹²  covalent organic frameworks (COFs),¹³  and porous coordination polymers (PCPs),¹⁴  have been investigated, and show excellent CO₂ capture performances.

Among these solid adsorbents, porous carbonaceous materials have been studied intensively because of their desirable physical and chemical properties, such as low cost, variety of form (powder, fibers, aerogels, composites, sheets, monoliths, tubes, etc.), ease of processability, controllable porosity (adjustable pore size and its distribution, high specific surface area and pore volume), and tailored surface chemistry (O, N, S, P, F or other heteroatom doping). They also possess some other advantages, particularly for adsorption applications: (1) carbon materials have excellent stability especially in hot and humid environments; (2) gas sorption on carbon materials is not moisture-sensitive because the surface is usually hydrophobic; (3) the energy consumption of regeneration is low due to the nature of physical adsorption; (4) the adsorption/desorption temperatures are always below 373 K; (5) these materials can be used at atmospheric pressure.

In this chapter, we summarize the recent research progress made in developing carbon-based sorbents for post-combustion CO₂ capture. Specifically, this chapter will provide overviews of (1) porous carbons, (2) graphene-based porous materials, (3) carbon nanotubes, (4) carbon-based hybrid sorbents, and (5) important factors influencing CO₂ uptake over carbon adsorbents.

Porous carbons have been extensively studied in the field of CO₂ capture. In order to control the pore structure in carbon materials, a variety of preparation methods have been developed and certain successes have been achieved. Herein, we summarize typical preparation methods for porous carbon adsorbents, including chemical activation, physical activation, metal ion activation, templating methods, and the combined method of templating and activation. In each section, synthesis principles, carbon precursors, pore structures, as well as their CO₂ adsorption performance, are discussed.

Activated carbon is the oldest and most widely used carbon material. Generally, the production routes of activated carbons are divided into physical activation and chemical activation. In chemical activation, the carbon precursor is mixed/impregnated with an activating agent (such as KOH, H₃PO₄, ZnCl₂, K₂CO₃, etc.), then the precursor is simultaneously carbonized and activated at an elevated temperature (from 400 to 900 °C) and under an inert atmosphere (usually N₂ or Ar). In physical activation, the carbon precursor is usually pre-carbonized at temperatures over 500 °C in an inert atmosphere to remove non-carbon species, followed by etching by an oxidizing gas (such as CO₂, steam, and air) at a higher temperature (from 700 to 1200 °C). Comparatively, chemical activation needs a lower temperature and shorter activation time, and generally results in a higher specific surface area and more uniform pore size distribution (PSD), while physical activation is simple and does not require chemical agents and repeated washing procedures to remove the inorganic residues after activation. The structure of activated carbons, containing the surface area, pore size and its distribution, and surface chemistry, etc., strongly depends on the activation conditions, activating agents as well as the carbon precursors used.

KOH is the most common activating agent in chemical activation. Jaroniec et al. treated a commercial carbon sorbent (Ambersorb 563) with five of the most commonly-used activating agents, including CO₂, H₂O, NH₃, KOH, and ZnCl₂, and compared their activating power for the evolution of microporosity responsible for CO₂ adsorption.¹⁵  N₂ adsorption analysis showed that the investigated activating agents enlarged microporosity and consequently surface area and pore volume of the carbons in the following order: KOH > CO₂ > NH₃ > H₂O > ZnCl₂. It was shown that KOH activation yielded the highest volume of micropores and small micropores. Besides, the CO₂ uptake for the KOH-activated sample was the highest, indicating that KOH activation appears to be the most effective to obtain carbon adsorbents for CO₂ capture.

However, the mechanism of KOH activation has not been totally understood due to the complexity of this process. In a previous review about KOH activated carbon materials for energy storage, Wang et al. suggested that KOH activation is a synergistic process of chemical activation, physical activation, and carbon lattice expansion by metallic K intercalation.¹⁶  Firstly, the potassium species serving as chemical activating reagents vigorously etch the carbon framework by the redox reactions shown in eqn (1.1)–(1.3). Secondly, H₂O (from dehydration of carbon precursors or eqn (1.4)) and CO₂ (eqn (1.5) and (1.6)) produced in situ in the activation system further develop the porosity through the gasification of carbon, namely physical activation (eqn (1.7) and (1.8)). Meanwhile, the produced metallic K intercalates into the lattices of the carbon matrix, responsible for both stabilization and widening of the interlayer spacing (Figure 1.1). After removal of the intercalated metallic K and other K compounds by washing, the expanded carbon lattices cannot return to their previous non-porous structure and thus create a narrow microporosity and large specific surface areas.

Generally, the raw materials of KOH activation could be classified into non-renewable fossil-based materials and renewable biomass resources. As shown in Table 1.1, various fossil-based precursors, such as petrol coke,^17,18  pitch,¹⁹  and synthetic polymers,^20–22  have been used as precursors for the preparation of porous carbon adsorbents. Wahby et al. prepared a series of carbon molecular sieves (CMS) from petroleum pitch using KOH as the activating agent. Depending on the type of petroleum residue and the conforming step applied, the prepared CMS possessed a well-defined pore size (0.35–0.7 nm), together with a very large surface area up to 3100 m² g⁻¹, thus exhibited a high CO₂ adsorption capacity up to 4.09 mmol g⁻¹ at 1 bar and 25 °C.²³  After further optimizing the process parameters, including the nature of the petroleum residue, the KOH–pitch ratio, the mesophase content, the temperature and time of activation, the CO₂ uptake could increase to 5.23 mmol g⁻¹ at 1 bar and 25 °C.²⁴  The activated carbons by the KOH activation of petroleum coke possessed high surface areas, over 3000 m² g⁻¹, and a high CO₂ capacity, over 15 wt% at 1 bar.²⁵  High-resolution analysis of N₂ sorption isotherms concluded that the micropores smaller than 1 nm played a critical role in CO₂ capture under ambient conditions due to the high-density filling of CO₂ in these small pores.²⁵  Using petroleum coke as the precursor, Yang et al. prepared N-doped porous carbons by combining ammoxidation with KOH activation.¹⁸  The sample prepared under mild conditions (a low temperature of 650 °C and a low KOH–precursor ratio of 2) showed the highest CO₂ uptake of 4.57 mmol g⁻¹ at 25 °C and 1 bar, while the CO₂/N₂ selectivity and CO₂ heats of adsorption of the sorbent were 22 and 37 kJ mol⁻¹, respectively. The high CO₂ capture capacity was attributed to the synergetic effect of N-doping and high narrow microporosity, while the latter was suggested to contribute more.¹⁸ 

Besides petroleum coke and pitch, synthetic polymers, such as phenolic resins,²⁷  poly(vinylidene chloride),²⁹  styrene-divinylbenzene resin,³⁰  polypyrrole,²⁰  polyaniline,²¹  polyurethane,³²  urea furfural resins,³³  polyacrylonitrile,²²  and polyimine,^37,38  have also been widely used as precursors for the preparation of porous carbon adsorbents. For instance, Jaroniec et al. prepared activated carbon spheres by direct KOH activation of phenolic resin spheres obtained by a modified Stöber method.²⁷  Due to the small micropore (<0.8 nm) and large specific surface areas, the prepared porous carbon spheres exhibited superior CO₂ uptakes reaching 4.6 and 8.9 mmol g⁻¹ at 23 and 0 °C under 1 bar, respectively.

Many studies have reported that N-doping could significantly improve the CO₂ capture performance of porous carbons, especial CO₂ uptake at low partial pressure and the selectivity of CO₂-to-N₂ (Table 1.1). Activation/carbonization of nitrogen-containing synthetic polymers is an important approach to preparing N-doped porous carbons. In 2010, Lu et al. reported a facile and rapid preparation of N-doped porous carbon monoliths using a basic amino acid as both the catalyst and nitrogen source.³⁹  This monolithic carbon directly pyrolyzed at 500 °C exhibited a CO₂ adsorption capacity of 3.13 mmol g⁻¹ at room temperature. KOH activation of nitrogen-containing polymers could result in highly developed microporosity at the same time maintaining the N-doping. Sevilla et al. prepared highly porous N-doped carbon by using polypyrrole (PPy) as the carbon precursor and KOH as the activating agent.²⁰  The mildly activated carbon (KOH–PPy = 2) showed a larger CO₂ uptake than the severely activated ones (KOH–PPy = 4) as a result of two important characteristics for the mildly activated samples: (a) a high-level N-doping (up to 10.1 wt% N) identified as main pyridonic-N and a small proportion of pyridinic-N groups, and (b) a narrower micropore size. Moreover, these polypyrrole-based carbons showed a high adsorption rate for the capture of CO₂, more than 95% of the CO₂ being adsorbed in 2 min with a high kinetic CO₂–N₂ selectivity of 12. Other nitrogen-containing polymers, such as polyacrylonitrile,²²  polyaniline,²¹  polyurethane,³²  urea furfural resin,³³  and polyimine,^35,36  have also been used as precursors and the corresponding N-doped porous carbons showed CO₂ uptakes of 4.50, 4.30, 4.33, 4.70, and 3.10 mmol g⁻¹ at 25 °C and 1 bar, respectively.

From Table 1.1, we find that the CO₂ uptakes for most of the activated carbons are lower than 5.0 mmol g⁻¹. Recently, Sethia and Syari prepared a series of strictly microporous N-doped activated porous carbons by using a nitrogen-containing ionic liquid of 1,3-bis(cyanomethyl imidazolium) chloride as a precursor and KOH as the activating agent.³⁴  The optimized material exhibited an extraordinary CO₂ uptake of 5.39 mmol g⁻¹ at 25 °C and 1 bar, which is the highest uptake reported so far for activated carbons. They concluded that both nitrogen content and ultramicropores played important roles in the CO₂ capture, with the latter being predominant.³⁴ 

Besides the N-doped porous carbons, S-doped porous carbons have also been prepared by the KOH activation of sulfur-containing synthetic polymers and are applied in CO₂ capture. For instance, S-doped microporous carbon materials can be prepared by the KOH activation of polythiophene–graphene composite.³⁷  This material displayed a high CO₂ uptake of 4.5 mmol g⁻¹ at 25 °C and 1 bar, as well as an impressive CO₂ adsorption selectivity over N₂ (51), CH₄ (12), and H₂ (214) based on the initial slopes method of Henry's law.^40–42 

Fossil-derived materials, such as petroleum coke, pitch, coal, and synthesized polymers are non-renewable and expensive. Considering the demand for a huge amount of solid CO₂ sorbents, the development of low-cost carbon materials from renewable and sustainable sources is urgent.⁴³  Biomass materials are renewable, easily available and very cheap, which are promising raw materials for the preparation of carbon adsorbents. As shown in Table 1.2, there have been over 40 kinds of biomass employed as precursors for porous carbon adsorbents, and most of them are obtained via chemical or physical activation, especially KOH activation. The specific surface areas and pore volumes of the biomass-based carbons were tuned with controlled carbonization and/or activation processes. The biomass used previously could be roughly divided into three categories: (1) polysaccharides, such as sucrose,⁴⁴  cellulose,⁴⁵  chitosan,⁴⁶  starch,^47,48  etc., (2) raw biomass or waste biomass, such as nutshells,^49,50  wood residues,⁵¹  Jujun grass,⁵²  Enteromorpha prolifera,⁵³  sugar cane bagasse,⁵⁴  granular bamboo,⁵⁵  leaves,⁵⁶  and others, and (3) microorganisms, including fungi⁵⁷  and yeast.⁵⁸ 

Polysaccharides could be converted into carbonaceous materials via a hydrothermal treatment approach using water as the carbonization medium under a self-generated pressure. In 2011, Sevilla and Fuertes reported a series of porous carbons prepared by the KOH activation of hydrothermally carbonized polysaccharides (starch and cellulose) and biomass (sawdust).⁴⁸  The prepared porous carbons were used as CO₂ adsorbents. Although the porous carbons prepared at a ratio of KOH–precursor = 2 have a considerably low-level pore development than those obtained at a ratio of KOH–precursor = 4, they exhibited significantly better CO₂ capture capacities. This is mainly due to the presence of a larger number of narrow micropores (<1 nm) in the mildly activated carbons. The sawdust-based carbon activated at 600 °C using KOH–precursor = 2 gave the highest CO₂ uptake of 4.8 mmol g⁻¹ (21.2 wt%) at 25 °C and 1 atm, which is still among the highest ever reported for activated carbons now.

Microporous carbons were prepared from wheat flour (mainly starch) via a combined process of pre-carbonization and post-KOH activation.⁵⁹  The pore texture of the wheat-flour-based carbons was significantly improved by the post-chemical activation of KOH and varied with the KOH–carbon ratio. By increasing the KOH–carbon ratio up to 3, narrow micropores smaller than 0.8 nm were developed primarily, and the corresponding activated carbon possessed a moderate surface area but the highest volume of pores smaller than 0.8 nm, thus delivering the highest CO₂ adsorption capacities of 5.70 and 3.48 mmol g⁻¹ at 0 and 25 °C, respectively. The experimental results confirmed that CO₂ adsorption uptake at ambient conditions was significantly dependent on the volume of narrow micropores with a pore size of less than 0.8 nm rather than the total volume or specific surface area.

Biomass waste is not only abundant, and sustainably renewable, but is also more cost-effective as carbon precursors. A wide range of biomass waste has been used to prepare porous carbon materials and showed excellent CO₂ sorption capacities. Shell is one of them. As early as 2004, Tan and Ani reported carbon molecular sieves by direct carbonization of palm shell waste at 600–1000 °C.⁶⁰  When the material was pyrolyzed at 600 °C, the micropore surface area was 753 m² g⁻¹; it showed CO₂ adsorption near 2.3 mmol g⁻¹ at 25 °C. When the pyrolysis temperature increased to 1000 °C, the micropore surface area and the pore volume decreased; however, the selectivity toward CO₂ increased.⁶⁰  After further KOH activation at 600 °C, the CO₂ uptake of palm-shell-based porous carbon significantly increased from 2.7 mmol g⁻¹ to 4.4 mmol g⁻¹ at 25 °C and 1 bar, due to the much more developed microporosity of the activated carbons compared to the non-activated ones.⁶¹  Hu and his co-workers prepared two series of nitrogen-doped porous carbons from coconut shells.^50,62  One type was prepared by urea modification and KOH activation, and the other was prepared by a successive process of pre-oxidization by H₂O₂, ammoxidation, and KOH activation. These carbons were found to exhibit very high CO₂ uptakes at 1 bar, almost 5.0 mmol g⁻¹ and 4.47 mmol g⁻¹ at 25 °C, 1 bar, respectively. The pre-oxidization by H₂O₂ increased oxygen-containing surface groups, which were expected to increase the amount of nitrogen incorporated into the carbon product in the subsequent ammoxidation process. The resulting carbons possessed a much higher nitrogen content and a narrower microporosity than the control sample without H₂O₂ pre-treatment. These results made coconut shell an attractive precursor for carbon adsorbents.

Shahkarami et al. performed a comprehensive study on the effects of the nature of precursors and carbonization conditions on the physical and chemical properties of activated carbon and the influence of these factors on CO₂ adsorption capacity, selectivity, and stability of the produced activated carbon in a fixed-bed reactor and mixed feed stream of N₂/CO₂/O₂.⁶³  In that work, three types of abundant feedstocks were used: agricultural waste (wheat straw and flax straw), forest residue (sawdust and willow ring), and animal manure (poultry litter). The selected precursors were carbonized via both fast and slow pyrolysis processes and were converted to porous carbons after KOH activation. Slow-pyrolysis-based activated carbon had a lower surface area and total pore volume but higher CO₂ adsorption capacity in the presence of N₂. Sawdust-based activated carbon synthesized from the slow pyrolysis possessed the largest ultramicropore volume of 0.36 cm³ g⁻¹, and the highest CO₂ adsorption capacity (78.1 mg g⁻¹) in N₂ but low selectivity (2.8) over O₂ because of the oxygen functional groups on the surface. The ultramicropores and surface chemistry of adsorbents were found to be much more important than particle size, total pore volume, and internal surface area of the adsorbents.

Besides terrestrial raw biomass, marine raw biomass, such as microalgae⁶⁴  and Enteromorpha prolifera,⁵³  were also converted into carbons. For instance, a sugar-rich microalgae (Chlorococcum sp.) was used to prepare N-doped activated carbons through hydrothermal carbonization and a subsequent activation of KOH or NH₃.⁶⁴  Although the NH₃-activated carbon possessed a much higher nitrogen content than the KOH-activated ones, the latter exhibited higher CO₂ sorption capacities than the former, indicating the main contribution of narrow microporosity on the CO₂ capacities. Ma et al. have successfully prepared highly porous N-doped carbon monoliths by using binary H₃PO₄–HNO₃ mixed acid as a co-activating agent and sodium alginate, a marine biopolymer, as a carbon precursor.⁶⁵  The SA-2N–P carbon prepared at a volume ratio of HNO₃–H₃PO₄ = 2 showed high CO₂ adsorption capacities of 8.99 mmol g⁻¹ at 0 °C and 4.57 mmol g⁻¹ at 25 °C, along with a high CO₂ capacity of 1.51 mmol g⁻¹ at 25 °C and 0.15 bar.

Microorganisms are another promising kind of carbon precursor. Wang et al. reported a series of porous carbons with narrow microporosities by KOH activation of pre-carbonized fungi (Agaricus).⁵⁷  A moderate CO₂ uptake of 5.5 mmol g⁻¹ and a high CO₂–N₂ selectivity of 27.3 at 0 °C, 1 bar were obtained. Similarly, Shen et al. found that yeast-based carbon was a promising adsorbent.⁵⁸  Hierarchical microporous carbon with a specific surface area of 1348 m² g⁻¹ and a pore volume of 0.67 cm³ g⁻¹ was prepared by KOH activation of yeast. This type of carbon material showed a high CO₂ uptake of 4.77 mmol g⁻¹ and a fast adsorption rate with an equilibrium time less than 10 min at 25 °C.

Other activating agents, such as ZnCl₂,^90,91  NaOH,⁸⁵  K₂CO₃,^38,86  H₃PO₄,⁹²  NaNH₂,⁹⁴  etc. have also been used to prepare porous carbon for CO₂ capture. ZnCl₂ is the second most commonly used activating agent in chemical activation.^96–98  However, there are only a few reports on ZnCl₂ activated carbons for CO₂ capture, as shown in Table 1.3, which may be due to its relatively undeveloped porosity as a result of the low activating power of ZnCl₂.^90,91  For example, ZnCl₂-activated polypyrrole gave a specific surface area of 1283 m² g⁻¹ and pore volume of 0.46 cm³ g⁻¹, much lower than those of KOH-activated polypyrrole (1700 m² g⁻¹ and 0.88 cm³ g⁻¹).^20,91  The as-prepared N-doped carbon exhibited a moderate CO₂ uptake of 3.80 mmol g⁻¹ at 25 °C and 1 bar when the activation temperature was 600 °C.⁹¹  Although the ZnCl₂ activated carbons showed lower CO₂ uptakes compared to the KOH activated carbons, the ZnCl₂ activation usually gave a higher carbon yield and higher carbon density thus could exhibit some advantage in volumetric CO₂ capacities.

Using K₂CO₃ as an activating agent, Fan et al. activated chitosan into N-doped microporous carbon.⁴⁶  By changing the weight ratio of K₂CO₃–chitosan and the activation temperature, the porosity and nitrogen content of the prepared carbons could be tuned in the range of 1180–2567 m² g⁻¹ and 1.29–6.02 wt%, respectively. The sample prepared at 635 °C with a K₂CO₃-chitosan ratio of 2 showed a CO₂ uptake of 3.86 mmol g⁻¹ at 25 °C, 1 atm, five consecutive recyclabilities, and a good CO₂–N₂ selectivity of ca. 21. Silvestre-Albero prepared a series of activated carbons by pre-carbonization of PANI at different temperatures and post-activation of KOH or K₂CO₃.⁸⁶  They studied the activating effects of KOH and K₂CO₃. It was found that carbonization temperature significantly influenced the porosity of the prepared carbons when using KOH as an activating agent, while K₂CO₃ mainly produced microporosity, independent of the carbonization temperature. The highest CO₂ uptake of these carbons was 7.60 mmol g⁻¹ at 1 bar and 0 °C.

In the last decade, ammonia (NH₃) treatment has been used to introduce nitrogen surface groups into the carbon framework. Several authors have reported ammonia-treated porous carbons for CO₂ capture.^79,99  Pevida observed that nitrogen groups were successfully introduced into carbon frameworks after a NH₃ treatment, and the nitrogen content was proportional to the oxygen content of the pristine porous carbons.⁹⁹  Due to the incorporated basic nitrogen groups, the ammoxidized carbons showed enhanced CO₂ uptakes at high adsorption temperatures. However, the low reaction efficiency between the NH₃ and the carbon resulted in a relatively low N-doping level. Recently, Geng et al. reported a NH₃-assisted activation process in which NH₃ played the roles of activating agent and nitrogen source at the same time.⁷⁸  When carbonizing a corncob in a NH₃ atmosphere, the nitrogen could be easily incorporated into a carbon framework. The N content increased as the activation temperature increased, reaching a high level of ∼12 wt% at 800 °C, along with an increase in the specific surface area and pore volume. Especially, the nitrogen was mainly incorporated in the form of phenyl amine and pyridinic N groups, which were very efficient for CO₂ capture. These carbons showed a moderate CO₂ uptake of 2.81 mmol g⁻¹, but superior IAST CO₂–N₂ selectivity up to 82. Similarly, Hu et al. prepared a N-doped hierarchical porous carbon by carbonization of a cellulose aerogel under a NH₃ atmosphere.⁴⁵  This N-doped carbon aerogel exhibited a N-doping of 4.62 wt%, and a high CO₂ adsorption capacity of 4.99 mmol g⁻¹ at 25 °C and 1 atm.

Most of the other activating agents usually show a lower activating ability than KOH. For example, KOH-activated polypyrrole carbons show a much higher specific surface area than the NaOH-activated ones under the same activation conditions (e.g., 2940 m² g⁻¹ vs. 1453 m² g⁻¹, activation conditions: the weight ratio activating reagent–polypyrrole = 2, and activation temperature = 700 °C).^20,85  Interestingly, NaNH₂, a strong base commonly used in organic synthesis, was reported to exhibit a more powerful activation ability compared to NaOH and KOH. At an activating reagent–carbon weight ratio of 2 and an activation temperature of 550 °C, NaNH₂ could activate a much higher microporosity, especially more small micropores, than KOH and NaOH, resulting in a higher CO₂ uptake of 3.66 mmol g⁻¹ at 25 °C and 1 bar, verifying the superiority of NaNH₂ in the activation under relatively moderate conditions.⁹⁴ 

The so-called physical activation is the partial gasification of the carbon framework with CO₂, steam, and air, or a combination of these, at high temperatures (from 700 to 1200 °C) as shown in eqn (1.9) to (1.12). In view of porosity development, the most important variables in the gasification process are the activating agents, the final burn-off ratio, and the inorganic impurities. Rodríguez-Reinoso and Molina-Sabio comprehensively studied the evolution in the porosity of several series of activated carbons prepared by physical activation of lignocellulosic materials (uncatalyzed and iron-catalyzed) in CO₂ or in a water–nitrogen mixture.¹⁰⁰  They found that at the initial stage of CO₂ activation, the micropore and macropore volume increase coincided with the proceeding burn-off. Further burn-off with CO₂ opened and enlarged the micropores of the char with even a shift to meso- and macropores, the ablation of the exterior of the particle being very important at high burn-offs over 50%.

The final activated carbon had a well-developed micro- and macroporosity, with a relatively small portion of mesoporosity. The Fe-catalyzed CO₂ gasification was initially very fast even at a lower gasification temperature and declined at an increasing burn-off due to the deactivation of the Fe catalyst. Steam activation produced a more selective attack of the carbon framework and a more uniform widening of porosity. It developed a large number of micropores, but fewer macropores, compared to CO₂ activation.

Mietek Jaroniec et al. prepared a series of porous carbon spheres by a combination of pre-carbonization and subsequent CO₂ activation of phenolic resin spheres obtained by a modified Stöber method.¹⁰¹  The activated carbon spheres had diameters from 200 to 420 nm, a high surface area (from 730 to 2930 m² g⁻¹), abundant narrow micropores (<1 nm, pore volume of 0.28 to 1.12 cm³ g⁻¹), thus exhibited very high CO₂ adsorption capacities: 8.05 and 4.55 mmol g⁻¹ at 1 bar and two temperatures, 0 and 25 °C, respectively.¹⁰¹  They further reported a series of N-doped activated carbon spheres.¹⁰²  These materials were prepared by one-pot hydrothermal synthesis in the presence of resorcinol-formaldehyde resin as a carbon precursor and ethylenediamine (EDA) as both a base catalyst and nitrogen source, followed by carbonization in a N₂ atmosphere and CO₂ activation. The N-doping level and the particle size can be tuned by varying the EDA amount in the range of 2.80–7.20% and 50–1200 nm, respectively. These activated carbon spheres exhibited a high surface area and volume of micropores smaller than 1 nm, reaching up to 1224 m² g⁻¹ and 0.36 cm³ g⁻¹, and could capture 6.2 and 4.1 mmol g⁻¹ of CO₂ at 0 °C and 25 °C. Obviously, the CO₂ uptake of the N-doped CO₂-activated carbon sphere was lower than those of non-doped or KOH-activated ones¹⁰¹  due its relative lower surface area and ultramicropores, indicating the primary importance of small micropores.

Two biomass materials – olive stones and almond shells – have been directly activated by CO₂ to prepare porous carbon adsorbents for CO₂ capture.⁸²  The effect of holding time on overall yield, pore texture, and CO₂ uptake, was studied. As the holding time was increased, the overall yield decreased and more porosity was developed, thus resulting in an increase of CO₂ uptake. The olive-stone-based carbons showed a higher specific surface area and pore volume than the almond-shell-based ones, indicating that the olive stone was more easily activated by CO₂. These carbons showed a CO₂ uptake up to 3.10 mmol g⁻¹ at 120 kPa and 25 °C, and 0.6–1.1 mmol g⁻¹ at 15 kPa and 25–50 °C.

Steam was also used to activate both synthetic polymers and biomass into highly porous carbons. N-doped activated carbons were prepared by steam activation of melamine-modified phenol-formaldehyde resins.¹⁰³  As the activation temperature increased, the specific surface area significantly increased from 382 to 1439 m² g⁻¹. The carbon activated at 850 °C could adsorb 6.71 mmol g⁻¹ CO₂ at 0 °C and 1 atm, and released 93.6% of the CO₂ adsorbed when the temperature increased to 50 °C. Young-Jung Heo and Soo-Jin Park prepared cellulose-fiber-based ultramicroporous carbons by steam activation.¹⁰⁴  The steam activation was found to have a strong influence on the development of new pores and the expansion of pore sizes and to be effective in developing optimal micropores for CO₂ adsorption. The porous carbon prepared by pre-carbonization at 800 °C and further activation with steam exhibited a high CO₂ adsorption capacity of 3.78 mmol g⁻¹ at 25 °C and 1 bar, as well as an impressive Henry's law selectivity of 47.1 for CO₂ over N₂.

In short, researchers have synthesized a large number of porous carbon materials from a wide range of raw materials and activating agents for CO₂ capture. The porosity and chemical surface properties varied with factors including raw materials, the pre-treatment process, activating agents, and activation temperature, etc. Particularly, many biomass materials, classified as polysaccharides, raw biomass and microorganisms here, are employed to successfully fabricate porous carbon adsorbents. Biomass-based carbon materials have shown high performances in CO₂ capture and will play a more important role in post-combustion CO₂ capture in the future. However, considering the huge emissions of carbon dioxide, there will be an enormous demand for solid CO₂ sorbents, and biomass materials that are easy to mass-produce and are widely-collected should be the focus of studies. KOH is still the most common activating agent, but it is highly corrosive. Additionally, the KOH dosage in the current literature is too high (e.g., 200 wt% of the carbon precursor). Therefore, the development of new activating methods is also needed.

It has been widely accepted that the small micropores in carbon frameworks, especially ultramicropores smaller than 0.8 nm, are the most efficient for CO₂ capture.^105–107  Chemical or physical activation methods are still widely used in the preparation of highly microporous carbon materials. In order to get a high capture capacity of CO₂, severe activation conditions, like excess activating agents and high burn-offs, are usually employed to increase the ultramicroporosity. In this way, an inhomogeneous activation is inevitable, resulting in a wide pore size distribution (PSD), low carbon yield and enhanced cost. Most activated carbon materials show a wide pore size distribution (PSD) and contain substantial supermicropores (larger than 1.0 nm), resulting in low CO₂ uptake, especially at a low CO₂ pressure (e.g., CO₂ capture from flue gas). For instance, Zhao et al. have reported that porous carbon prepared by the direct carbonization of poly(vinylidene chloride) gave a slightly higher CO₂ uptake than the further KOH-activated ones at 25 °C and 1 bar, because KOH activation generated some large micropores or small mesopores in the carbon framework, and then resulted in the reduction of interactions between CO₂ molecules and the pore walls.²⁹ 

Ultramicroporosity is desired for a high CO₂ capture performance. Recently, we reported a new activating strategy for strictly ultramicroporous carbons by using phenolic resins as platforms, that could be called “single-ion activation”.^49,50,108  In this process, mono-dispersed alkali ions are introduced via an acid–base reaction between an alkali hydroxide and the acidic groups (–OH or –COOH) of phenolic resins (Figure 1.2). After a direct carbonization of the prepared phenolic salts of resin, strictly ultramicroporous carbons are obtained. The mono-dispersed ions in the form of –O⁻M⁺ or –COO⁻M⁺ serve as activating agents, and produce a homogeneous activation effect, resulting in highly developed and narrowly-distributed ultramicropores.

This strategy was first used to prepare N-doped microporous carbons (NMCs).⁸⁸  As shown in Figure 1.2, potassium phenolate of meta-aminophenol-formaldehyde resin was firstly prepared by the aqueous poly-condensation of meta-aminophenol and formaldehyde using equimolar KOH as both a catalyst precursor and a base. During the polycondensation reaction, the K⁺ ion can exchange with H⁺ in the hydroxyl groups, leading to the single dispersion of K⁺ ions in the bulk of the phenolic resin. After a simple carbonization process, N-doped ultramicroporous carbons were obtained.

From Figure 1.3, it could be seen that all the NMC samples exhibited a standard sorption isotherm of type I with a very narrow knee at a very low relative pressure (p/p⁰ < 0.02) and an absence of apparent adsorption increment in the relative pressures over 0.02, indicating that the porous carbons had a narrow microporosity and that no mesoporosity was present. The specific surface area of NMCs gradually increased from 272 m² g⁻¹ to 664 m² g⁻¹ as the activation temperature increased from 600 °C to 900 °C. Dubinin–Radushkevich (D–R) plots of CO₂ sorption on NMCs exhibited well-defined linear shapes, showing that the porosity of NMCs was strictly made up of uniform ultramicropres.^107,109  The average sizes of these ultramicropores calculated according to the slope of the D–R plots gradually increased from 0.50 to 0.58 nm as the carbonization temperature increased from 600 °C to 900 °C.

The carbon prepared at 600 °C (NMC-600) exhibited a high CO₂ uptake of 3.90 mmol g⁻¹ at 25 °C and 1 bar, although this carbon possessed a very low specific surface area of 272 m² g⁻¹. Besides, this carbon gave an outstanding CO₂ uptake of 1.67 mmol g⁻¹ mmol g⁻¹ at 25 °C and 0.15 bar, and an impressive CO₂-over-N₂ selectivity of 50 based on Henry's law. At 75 °C and 1 bar, the carbon prepared at 800 °C (NMC-800) still showed a very high CO₂ uptake of 2.0 mmol g⁻¹. The excellent performance of NMCs for CO₂ capture indicated that these carbons preferably interact with CO₂ molecules based on the combined contributions of the extremely small micropores ca. 0.5 nm and the polar surface caused by the N, O-doping.

The strategy of single-ion activation could be applied to finely tune the sizes of ultramicropores by using different alkali metal ions (Li⁺, Na⁺, K⁺, Rb⁺, Cs⁺) as activating agents (Figure 1.4).⁸⁹  All the samples presented a standard type I N₂ adsorption isotherm, showing a narrow micropore size distribution. The pore volume and the apparent surface areas gradually increased from 0.07 to 0.70 cm³ g⁻¹ and 111 to 1312 m² g⁻¹, respectively, while varying the activating ion from Li⁺ to Cs⁺, thereby revealing that the activation power significantly increased from Li⁺ to Cs⁺. All the samples, except Li⁺-activated carbon (LiAC), exhibited well-defined linear D–R plots (correlation coefficient R² > 0.995), indicating the existence of strictly uniform ultramicropores (Figure 1.5). Interestingly, a systematic widening of micropores took place; the pore size gradually increased from 0.60 nm up to 0.76 nm as the activation ions varied from Li⁺ to Cs⁺. In other words, the ultramicropore size was finely tuned at sub-angstrom level by simply varying the activating ions.

When applying for CO₂ capture, the Cs⁺-activated carbon (CsAC) showed a superior CO₂ adsorption capacity of 5.20 mmol g⁻¹ at 25 °C and 1.0 bar, high CO₂-to-N₂ selectivity and full regenerability for four consecutive cycles (Figure 1.6). Furthermore, the resulting carbons present ultrahigh capacitances of up to 223 F g⁻¹ or 205 F cm⁻³ in an ionic liquid electrolyte. The outstanding performance of these carbons is due to their uniform pore size, which is finely tuned to adapt to the dimensions of the CO₂ molecule and the ions of ionic liquids. The CO₂ capacities and specific capacitances reported here were higher than for most traditional activated carbons. More importantly, the dosage of activation agents used herein was only 30 wt% of the carbon precursors as counted in KOH, much lower than that used in traditional chemical activation,⁸⁹  confirming the high efficiency of metal ion activation.

The mechanism of metal ion activation was systematically studied by thermogravimetry-mass spectrometry (TG-MS) analysis and in-situ X-ray diffraction (XRD) (Figure 1.7). The pyrolysis process of the potassium salt of carboxylic phenolic resin (PR-COOK) was much more complicated than that of the phenolic resin carboxylic acid (PR-COOH) due to the activation of metal ions. Three gases, including H₂O (m/z 18), CO (m/z 28), and CO₂ (m/z 44), were detected by mass spectrometry. In-situ XRD patterns of PR-COOK indicated that the PR-COOK began to decompose into K₂CO₃ (PDF card 71-1466) at around 200 °C, and then transformed into K₂O at 600 °C, along with two steps of weight loss and the release of CO₂ and CO. When the carbonization temperature increased to 800 °C, new peaks at 10.4° and 16.5° appeared, which is attributed to K (100) (PDF card 01-0500) and KC₈ (100) (PDF card 04-0221), and large amounts of CO were detected (Figure 1.7). These facts are strong evidence that metallic potassium was produced by the reduction of K₂O and was further intercalated into the carbon layers to form graphite intercalation-like compounds.

The mechanism of single-ion activation was suggested as: first, alkali salts of phenolic resins decomposed into alkali carbonate (M₂CO₃), H₂O and CO₂ below 400 °C. Second, alkali oxides (M₂O) were produced by the thermal decomposition of M₂CO₃ or the redox reaction between M₂CO₃ and C. Third, framework carbon atoms were etched by M₂O via the vigorous redox reaction of M₂O + C → 2M + CO, then the produced metallic alkali intercalated into the lattices of the carbon matrix. The interlayer spacing will be primarily determined by the size of the metal intercalate, and this spacing (i.e., pore size) will be systematically widened with increasing metal ion size from Li⁺ to Cs⁺. In a word, the activation is similar to traditional KOH activation, but much more efficient.

This strategy was further applied for the synthesis of microporous carbon spheres for the following reasons: (1) the developed ultramicroporosity offered a large space for CO₂ adsorption; (2) the small size of the carbon spheres ensured a short diffusion distance of the CO₂ molecule; (3) the gaps between carbon spheres allowed a rapid flow of flue gas.¹¹⁰  In that work, potassium salts of resorcinol-formaldehyde (RF) resin spheres were prepared by the reaction of KOH and pristine RF resin spheres, and followed by carbonization into microporous carbon materials (denoted as CS-x, where x stands for the weight ratio of KOH–RF) (Figure 1.8). When the prepared resins were mixed with KOH, the –OH groups of the RF resins immediately reacted with KOH via a fast acid–base reaction and were changed into mono-dispersed K⁺ in the form of –OK groups. Based on a simplified molecular structure of RF resins, the KOH–RF resins' weight ratios of 0.50, 0.75, 1, 1.5 and 2 are roughly equivalent to the KOH–OH molar ratios of 0.50, 0.75, 1, 1.5 and 2. In the cases of CS-0.5, CS-0.75 and CS-0.75, the activating agents exclusively existed as –OK groups, while in the cases of CS-1.5 and CS-2, the KOH was in excess (Figure 1.8).

After activation, the majority of the carbon particles retained, roughly, their spherical shape even when the weight ratio of KOH to RF resins was up to 2 (Figure 1.9). The detailed microscopy morphology and the porosity of the prepared carbon materials were closely related to KOH dosage (Figures 1.9 and 1.10). The CS-0.5, CS-0.75 and CS-1 samples showed regular spherical shapes, a smooth carbon surface, and an exclusively narrow microporosity (<0.8 nm), indicating that the K⁺ activation almost occurred in the interiors of the carbon mono-spheres. In comparison, some defects and roughness were observed on the surface of the CS-1.5 and CS-2 carbon particles, and a widening of PSD took place, indicating a non-homogeneous activation. As the KOH dosage increases, the surface areas and volumes of the micropores first increase and then decrease, and CS-1 possesses the largest specific surface area, micropore surface area and volume, up to 1235 m² g⁻¹, 1084 m² g⁻¹ and 0.57 cm³ g⁻¹, respectively.

The as-prepared carbon spheres exhibited excellent CO₂ capture capacities ranging from 6.55–7.34 mmol g⁻¹ at 0 °C and 4.27–4.83 mmol g⁻¹ at 25 °C and 1 bar, and the highest values belonged to CS-1, which is attributed to its largest micropore volume with narrow micropore size. The value of 4.83 mmol g⁻¹ is higher than those for lots of phenolic-resin-based porous carbon materials under the same conditions, including porous carbon spheres prepared by traditional chemical or physical activation, such as CO₂-activated carbon spheres,^101,107  N-rich microporous carbon spheres,¹⁰²  and KOH-activated carbon spheres.²⁷ 

Overall, single-ion activation has been used to develop and adjust ultramicroporosity. This strategy has exhibited many merits, such as a low demand for activating agents, a high activating efficiency, easy operation, and it reaches high CO₂ sorption capacities; it is worthy of further study.

Historically, the templating method was first reported by Knox and co-workers in 1986, who demonstrated the synthesis of graphitic porous carbons for liquid chromatography separation by impregnation of spherical porous silica gel particles with phenolic resin and subsequent carbonization and removal of silica.¹¹¹  Since then, this method has received extensive attention and various types of template carbons are synthesized. The resulting carbon synthesized by the templating method, called a templated carbon, possesses a relatively narrow PSD and controlled architecture. The templated carbonization method permits one to control the carbon structure in terms of various aspects, such as pore structure, specific surface area, microscopic morphology and graphitizability, which makes this method very attractive.^112,113 

Hard templates and soft templates are the main kinds of templates used as molds to form templated carbons. The hard template method includes the following steps: (a) synthesis of a suitable porous template; (b) introduction of a suitable carbon precursor into the template pores using the method of wet impregnation, chemical vapor deposition or a combination of both methods; (c) polymerization and carbonization of the carbon precursor; and (d) removal of the inorganic template.^114,115  Following these steps, porous carbon with a specific pore structure is formed. Compared with the hard template, the soft template is a kind of surfactant, which has a strong interaction with the carbon source, and mesoporous carbons with different structures can be obtained through the soft template method. This method possesses good controllability and operability; as a result, it has very good application prospects. The mechanisms of the soft template method include a liquid crystal template mechanism, a synergistic assembly mechanism, a “rod micellar” mechanism and so on; these mechanisms have been widely recognized.¹¹⁶ 

Since the porous carbon replicates the morphology of the template, selecting a suitable template to synthesize the carbon with a specific porous structure is the most important step. Template carbons with different pore structures, such as ordered porous carbon, disordered porous carbon and hierarchical porous carbons, can be prepared by different templates, such as porous silica,^117–120  zeolites,¹²¹  metal–organic frameworks (MOFs),¹²²  nanoparticles (e.g., MgO¹²³  and SiO₂ ¹²⁴ ), and surfactant micelles.^125–131  These templates will be described in detail in the following sections.

Mesoporous molecular sieve SBA-15 has a large surface area, uniform pore diameter distribution, adjustable pore size and wall thickness, and high hydrothermal stability. Therefore, SBA-15 is an excellent template to prepare ordered porous carbons.^120,132  Wang et al. designed a new method of using a one-step “surfactant-assisted” nanocasting route to synthesize highly ordered mesoporous graphitic carbon. They selected ordered mesoporous silica SBA-15 maintaining its triblock copolymer surfactant P123 (EO₂₀PO₇₀EO₂₀) as the hard template and natural organic soybean oil (SBO) as a carbon precursor. Finally, the ordered mesoporous graphitic carbon was synthesized and named as MGC-1 (Figure 1.11).¹³³  The main advantage of this method is that the carbon precursor SBO can easily infiltrate the mesopore channels of the silica template with the help of P123 to reach a high filling level, resulting in an enhanced yield of graphitic carbon materials. Because of the improved structural ordering, the mesoporous carbon, after amine modification, could adsorb more CO₂ (57.2 mg g⁻¹) compared with amine-functionalized carbon prepared without the assistance of surfactant (31.3 mg g⁻¹).¹³³ 

In order to clarify the influence of porosity and surface chemistry on carbon dioxide capture, Sanchez–Sanchez et al. synthesized ordered mesoporous carbons with high surface areas and pore volumes by the hard template method, in which 3-aminobenzoic acid acted as both carbon precursor and nitrogen source, and SBA-15 acted as the hard template. The synthesis was accomplished in the presence of different H₃PO₄ concentrations. The surface chemistry and the porous texture of the carbons could be easily modulated by varying the H₃PO₄ concentration and carbonization temperature. Finally, they found that with H₃PO₄ concentrations higher than 50 wt%, the pyrolysis mechanism of the thermally polymerized precursor changed and both the structural order and mesopore arrangement declined. On the contrary, low H₃PO₄ concentrations led to mesoporous carbon materials with a very narrow PSD and favored nitrogen retention at the carbon surface. By comparing the amount of CO₂ adsorption of different samples under different conditions, they also proved that the CO₂ adsorption capacity depends on the narrow micropore volume and is almost unaffected by the surface chemistry at 0 °C and 1 bar; the largest CO₂ uptake of 5.04 mmol g⁻¹ was obtained for the carbon with the most developed ultramicroporous structure. On the other hand, the nitrogen functional groups exerted a beneficial influence on CO₂ capture, while oxygen and phosphorus functionalities exerted a negative influence at 25 or 50 °C and 1 bar. Moreover, among the N-containing groups, the pyrrolic N groups provoked the largest improvements in CO₂ adsorption.¹³⁴ 

In addition to SBA-15, there are also other porous silica templates used to prepare ordered mesoporous carbons. For instance, Tiwari prepared oxygen enriched carbon adsorbents with epoxy resin as the precursor and MCM-41 as a template.¹³⁵  The sample prepared at 700 °C shows the highest capacity of CO₂ adsorption of 0.65 mmol g⁻¹ at 30 °C at a CO₂ concentration of 12.5%. Zhao et al. reported nitrogen-doped mesoporous carbons, which were prepared by a nanocasting route using tri-continuous mesoporous silica IBN-9 as a hard template and p-diaminobenzene as a nitrogen-containing precursor.¹³⁶  One of the carbons possessed a very high nitrogen doping concentration (13 wt%) and exhibited an excellent performance for CO₂ adsorption at about 4.50 mmol g⁻¹ at 298 K and 1 bar.

Zeolites are highly crystalline aluminosilicate materials that possess uniform sub-nanometer-sized pores. The pore channel apertures vary between 0.3 and 1.5 nm. As a result, zeolite-templated carbons (ZTCs) possess highly ordered microporosity and a large surface area, which is helpful in enhancing the capabilities of CO₂ capture. With these advantages, ZTCs are especially attractive and many groups have prepared different types of ordered microporous carbons using various zeolites as templates to study the capabilities of CO₂ capture.^137,138 

Xia et al. synthesized zeolite-templated, high-surface-area, microporous, N-doped carbons, which exhibit CO₂ uptake capacities of up to 6.9 mmol g⁻¹ at 273 K at 1 bar and 4.4 mmol g⁻¹ at 298 K at 1 bar, along with CO₂ adsorption energies of 36 kJ mol⁻¹ to 20 kJ mol⁻¹.¹³⁹  Combined with their ease of preparation, regeneration stability, excellent recyclability and high selectivity for CO₂, the N-doped zeolite-templated carbons are suitable for CO₂ capture. They further successfully synthesized structurally well-ordered sulfur-doped microporous carbon materials for the first time.¹²¹  The sulfur-modified ordered microporous carbons were obtained by infiltration of the carbon precursors (2-thiophenemethanol) into the pores of zeolite EMC-2 via impregnation combined with a CVD method. They found that S-doped microporous carbons with well-resolved XRD patterns arising from a high level of zeolite-type structural ordering can be nanocasted from zeolite EMC-2. The carbon materials possessed a high surface area with varied sulfur content depending on the preparation conditions. They found that both the presence of S functional groups and the textural properties play important roles in the capacity of CO₂ adsorption. These carbons also showed a very high CO₂ adsorption energy up to 59 kJ mol⁻¹.¹²¹ 

Porous organic frameworks (POFs), such as metal–organic frameworks (MOFs), zeolite imidazole frameworks (ZIFs) and covalent organic frameworks (COFs), have been intensively investigated for CO₂ capture application recently.^42,140,141  MOFs generally consist of two main components: organic ligands and metal ions. The structures of MOFs change with the types of inorganic metal ions and organic ligands and can be functionally modified by a variety of methods. ZIFs are a newly emerging class of porous crystals with extended three-dimensional structures constructed from tetrahedral metal ions (e.g., Zn, Co) bridged by imidazolate. Another kind of porous material that has emerged recently, COFs possess a long-range ordered network composed of covalent connections by the light elements (C, N, O, B). Recently, these porous organic frameworks have been used to prepare porous carbons and related nanostructured functional materials.¹⁴²  Those carbons have also been quickly developed and used in CO₂ capture.

Generally, there are two routes to fabricate porous carbons from these porous organic frameworks. One is using the porous organic frameworks as sacrificial templates with the incorporation of an additional carbon source. Xu and his co-workers used the classical MOF-5 as a sacrificial template and precursor with furfuryl alcohol as the additional carbon source to prepare porous carbons, which is the first report of MOF-derived porous carbons.¹⁴³  The resultant porous carbon obtained at 1000 °C exhibited a high specific surface area up to 2872 m² g⁻¹. The as-synthesized porous carbon exhibited a high H₂ uptake, reaching 2.6 wt% at 77 K and 1 bar, much higher than that of raw MOF-5 under the same conditions. Moreover, this material gave a good electrochemical performance as an electrode material for supercapacitors with a capacitance of 258 F g⁻¹ at a current density of 0.05 A g⁻¹.¹⁴³  They further investigated the effect of carbonization temperature on the porosity of the resultant porous carbons.¹⁴⁴  The specific surface area for MOF-5 templated porous carbons obtained at carbonization temperatures from 530 to 1000 °C is in the range from 1140 to 3040 m² g⁻¹, and the dependence of surface area on carbonization temperature shows a “V” shape.

Wang prepared N-doped porous carbons using ZIF-70 with a high porosity and large pore size as the model system and polyethyleneimine (PEI) molecules as the nitrogen source. Many PEIs can be readily incorporated into the framework pores of ZIF-70. Moreover, the ZIF-70 framework has many N elements, which can be partly doped into a carbon network during the carbonization process. Finally, a series of N-rich porous carbon materials was successfully fabricated by carbonization of polyamine-incorporated MOFs, which had an excellent CO₂ capture capacity of 4.86 mmol g⁻¹ at 1 bar and 25 °C.¹⁴⁵  Pachfule et al. reported a series of porous carbon materials by using isoreticular zeolitic imidazolate frameworks (IRZIFs).¹²²  A series of carbon materials named as C-68, C-69 and C-70 was prepared using ZIF-68, ZIF-69 and ZIF-70 as templates and furfuryl alcohol as the additional carbon source under a carbonization temperature of 1000 °C. Among these carbons, the resultant C-70 synthesized using ZIF-70 as the template shows the most developed porosity with the highest specific surface area of 1510 m² g⁻¹ due to the more porous ZIF-70 template. The porous carbons C-70, C-68 and C-69 show CO₂ uptake capacities (273 K and 1 atm) of 5.45, 4.98 and 4.54 mmol g⁻¹ for C-70, C-68 and C-69, respectively. Also, the C-70 carbon shows the highest H₂ uptake capacity (77 K and 1 atm) of 2.37 wt% among these carbons. Zhang et al. reported a new strategy of pre-introducing an extra carbon source (furfuryl alcohol) into a PAF (porous aromatic framework) followed by carbonization, which afforded a new microporous carbon material with a small pore size of 0.54 nm, thus facilitating a high CO₂ uptake capacity of about 4.1 mmol g⁻¹ at 295 K and 1 bar.¹⁴⁶ 

The other route is direct carbonization in which the POFs serve as both self-templates and carbon precursors at the same time without any additives. Srinivas et al. reported a new type of hierarchical porous carbon from the controlled carbonization of Zn-MOF-5, Zn-MOF-74, and Al-MIL-53 (Figure 1.12).¹⁴⁷  The authors found that an increase in the size of the MOF-5 precursor could effectively enhance the specific surface area of the resultant porous carbons. The simultaneously enhanced surface area and pore volume in HPC5b2 were attributed to the large, millimeter-sized crystallites of the MOF precursor, that delivered more void space due to the extended pores within the crystals after carbonization, compared with the smaller crystals in HPC5b1. HPC5b2-1100, obtained by carbonization of millimeter-sized MOF-5 at 1100 °C showed a very high specific surface area and total pore volume, up to 2734 m² g⁻¹ and 5.23 cm³ g⁻¹, respectively. Through simultaneous thermogravimetric analysis and mass spectroscopy (H₂, H₂O, CO, O₂, CO₂, and carbonaceous gases), the carbonization process and mechanism for all of the MOFs were investigated. Taking Zn-MOF-74 as an example, as shown in Figure 1.12, the first mass-loss, below 150 °C, was due to the evaporation of adsorbed/terminal water molecules within the pores and the second mass-loss between 400 °C and 600 °C was attributed to framework decomposition leading to a major release of carbon containing gaseous products (mostly CO₂, CO, C₆H₆ and a small amount of H₂ and C_xH_y hydrocarbon mixtures) and formation of ZnO occurs. The third mass-loss starting at 900 °C was due to a further release of CO₂ and CO with Zn through the reduction of ZnO by carbon via ZnO + C → Zn(g) + CO. In most cases, the CO₂ uptakes in these MOFs-derived carbons were higher than in their MOF precursor. A high CO₂ uptake, over 27 mmol g⁻¹ (119 wt%) at 30 bar and 27 °C, is obtained, which is one of the largest reported in the literature for porous carbons. However, the CO₂ capacity of these carbons at ambient pressure was only about 3 mmol g⁻¹ due to their relatively low ultramicroporosity.

Ding et al. reported a series of N-doped porous carbon monoliths by direct carbonization of IRMOF-3, an isoreticular MOF-3.¹⁴⁸  The ligand (2-aminobenzene-1,4-dicarboxylic acid) served as the nitrogen source. Meanwhile, the ZnO nanoclusters formed during the carbonization process acted as the in-situ activator and self-template, which was the key to the transformation of the microporous structure to the meso-macroporous structure. With an increase of the carbonization temperature, the specific surface area, micropore surface area and volume calculated using the t-plot method simultaneously increased. Compared to the pristine IRMOF-3, IRMOF-3/800 carbonized at 800 °C exhibited a much higher CO₂ capacity (3.99 vs. 2.32 mmol g⁻¹ at 1 bar and 1.64 vs. 0.70 mmol g⁻¹ at 0.15 bar) and CO₂–N₂ selectivity (22.7 vs. 7.4 at 0 °C and 38.8 vs. 8.6 at 25 °C).

Ma et al. synthesized a N-doped porous carbon using zeolitic imidazolate framework-8 (ZIF-8) both as a solid template and precursor. The resultant carbons were further modified with NH₃. It was confirmed that absorption of CO₂ was enhanced by ammonia treatment. This is attributed to the presence of increased N-containing functional groups and a specific surface area during modifications.¹⁴⁹  Similar to that, Zhao reported the design and synthesis of a series of new B, N-containing cross-linked polymers with well-defined chemical structures and high thermal stabilities. Further, they were efficiently converted into B/N co-doped porous carbons in high yields under pyrolysis treatment. The prepared carbons showed excellent CO₂ uptakes (3.25 mmol g⁻¹ at 273 K and 1 atm) with respect to their relatively low surface areas, and high selectivity of CO₂–CH₄ with a ratio of more than 5 : 1 at 298 K.¹⁵⁰ 

Unambiguously, MOFs, ZIFs and COFs have been emerging as very promising templates and/or precursors to prepare highly porous carbons that show great potential in gas storage and electrical applications. However, the CO₂ capacities of MOFs (ZIF, COFs)-derived carbons are at a moderate level among the porous carbons. Further exploration is needed, such as how to introduce high microporosity and heteroatom-doping, the development of low-cost MOFs (ZIFs, COFs) and their large-scale synthesis methods, etc.

Carbide-derived carbons (CDCs) are synthesized by selective extraction of a metal and metalloid from carbide precursors. High-temperature chlorination is the most widely used method for CDC synthesis (eqn (1.13), Figure 1.13).^151,152  In general, the starting carbide precursor templates the initial pore size of the CDCs, and then any range of average pore sizes from less than 1 nm to greater than 10 nm can be achieved by changing the chlorination temperature.¹⁵³  From this perspective, this preparation could be considered as a crystal lattice templating process. The most well-known application of CDCs is as a supercapacitor, which was led by Prof. Gogotsi.^154,155  These carbons have also shown great promise in H₂ storage,^156–158  and CH₄ storage,¹⁵⁹  as well as CO₂ capture.^{106,159–163} 

The TiC-CDC produced at a chlorination temperature of 800 °C showed a good H₂ capacity of 2.55 wt%, which was further improved to 3.0 wt% after a post-treatment of H₂ annealing.¹⁵⁷  Gogotsi et al. prepared a large number of CDCs with various pore sizes by using different starting carbides (B4C, ZrC, TiC and SiC) and chlorination temperatures to explore the effects of pore size and volume on the H₂ sorption and heat of adsorption.^156,157  They found that H₂ storage was dominated by small pores, and thus not directly connected to total specific surface area. A clear dependence on pore size was also suggested – the smaller pores increased both the heat of adsorption and the H₂ uptake.¹⁵⁶ 

As far as we know, the first study of CDCs for CO₂ adsorption was reported by Presser and his co-workers in 2011.¹⁰⁶  The Micro-TiC-CDC sample prepared under the combined conditions of chlorination at 700 °C and post-H₂ annealing at 600 °C showed a mean pore size of 0.66 nm and exhibited a very high sorption capacity for CO₂ of up to 7.1 mmol g⁻¹ at 0 °C and 1 bar. The highest CO₂ capacity at 0.1 bar and 0 °C is 1.62 mmol g⁻¹ delivered by the sample prepared at 600 °C. More importantly, they comprehensively studied the effect of pore size on CO₂ capacities at different sorption pressures. This will be detailed in Section 1.6.1. Bhatia and Nguyen studied the potential of carbon derived from SiC (SiC-CDC) for CO₂ capture from flue gas.¹⁶¹  They found that N₂ was very weakly adsorbed onto SiC-CDC, and H₂O adsorption on SiC-CDC was highly kinetically restricted due to the strong hydrophobicity of this material inherited from the diamond-like structure (sp³ bonding) of the SiC precursor. This feature is very promising for the selective adsorption of CO₂ from real flue gas since the co-adsorption of H₂O can significantly degrade the performance of any adsorbents for CO₂ capture.¹⁶¹ 

Wei et al. synthesized nitrogen-doped ordered mesoporous carbon materials by using a self-assembly process with dicyandiamide as a nitrogen source, soluble resol as a carbon source and a triblock copolymer (F127) as a soft template. Figure 1.14 shows the synthesis process of this ordered N-doped mesoporous carbon. In this synthesis, resol molecules can bridge the template F127 and dicyandiamide via hydrogen bonding and electrostatic interactions. The obtained N-doped ordered mesoporous carbons possess tunable mesostructures and pore sizes (3.1–17.6 nm), high surface areas (494–586 m² g⁻¹), and high N contents (up to 13.1 wt%). Ascribed to the unique feature of a large surface area and high N content, the carbon materials show a good CO₂ adsorption capacity of 2.8–3.2 mmol g⁻¹ at 298 K and 1.0 bar.¹³¹ 

Zhang et al. reported a facile one-pot melting-assisted solvent-free method to develop ordered mesoporous carbons (Figure 1.15).¹²⁷  This method only involved a simple thermal treatment procedure and avoided the time-consuming process of evaporation-induced self-assembly (EISA). Using this method, non-doped and N-doped mesoporous carbons can be synthesized. The non-doped carbon possessed a well-developed hierarchical pore texture with a high surface area of 748 m⁻² g⁻¹ and exhibited a CO₂ capacity of 2.73 mmol g⁻¹ at 75 °C and 1 bar. Due to the large amounts of nitrogen in the carbon matrix (11.67%), the N-doped mesoporous carbon exhibited excellent CO₂ capture properties, with a CO₂ capacity of 1.66 mmol g⁻¹ as well as an excellent IAST selectivity of CO₂-to-N₂, up to 240.7 at 75 °C and 1 bar. Wu et al. reported the consecutive incorporation of carbon nitride and nitrogen-containing functionalities into FDU-type ordered mesoporous carbon materials by using a simple post-synthetic route with melamine as a nitrogen source and F127 as the soft template. The nitrogen-enriched mesoporous carbons deliver promising properties for CO₂ capture with the high heats of adsorption and capacity (6–9 mmol g⁻¹ at 10 bar), well retained or even promoted at lower pressures (<3.5 mbar at 303 K).¹²⁸ 

In summary, the templating method allows us to effectively fabricate porous carbons with specific porous structures and surface chemical properties, which are suitable for CO₂ adsorption. However, there are also some weaknesses. For instance, hard templates are usually prepared by a complicated production process, but are sacrificed after carbonization. Although the soft template method avoids the procedure of removing the templates after carbonization, only a few soft templates are successfully developed due to the requirement of thermal stability. Moreover, the templated carbons usually lack microporosity, which is the really efficient type of porosity for CO₂ capture.

Recently, hierarchical porous carbons (HPCs) have attracted much attention. Hierarchical porous carbons possess complex network structures formed by two or more levels of pores, which include mainly micro-mesopores and micro-meso-macropores and so on. These carbons have shown a better performance in many application areas than carbons with a single porosity, on account of the improved mass transport facilitated by the large-sized pores and the high surface area and pore volume from the micro-/mesopores. For CO₂ capture, the combination of micropores and mesopores can simultaneously promote the adsorption and diffusion of CO₂ molecules.

To date, several approaches have been proposed for the synthesis of hierarchical porous carbons. One of the main techniques is a dual templating strategy, in which two templates with dimensions at different length scales are combined to generate multimodal pores. Either two hard templates¹⁶⁴  or a combination of hard and soft templates^165,166  can be employed in this technique. However, the biggest shortcoming of this strategy is its low efficiency in developing microporosity, especially ultramicroporosity, which is crucial for gas adsorption. Another commonly-used method to synthesize hierarchical porous carbons is a combination of a templating method and activation, in which mesopores and micropores are generated by the two different mechanisms respectively.¹⁶⁷  This combined technique could be further classified into a two-step process and a one-step process.

As mentioned above, a combination of a templating method and activation is one of the most commonly-used routes to prepare HPCs. This method usually contains two steps: firstly, the construction of mesopores by a template method; then, physical activation or chemical activation of the templated carbons to generate micropores. Many groups have synthesized various hierarchical porous carbons by choosing appropriate templates, carbon precursors and activation conditions.^22,168,169 

Marta Sevilla et al. designed and prepared hierarchical porous carbons through the KOH activation of two ordered mesoporous carbons of CMK-3 and CMK-8, which were respectively obtained by using hexagonal SBA-15 and cubic KIT-6 ordered mesostructured silica as hard templates.⁹⁸  The process of KOH activation was carried out at different temperatures in the 600–800 °C range. Textural characterization of these activated carbons showed that they had a dual porosity made up of mesopores derived from the templated carbons and micropores generated during the chemical activation step. As a result of the activation process, there was an increase in the surface area and pore volume from 1020 m² g⁻¹ and 0.91 cm³ g⁻¹ for the CMK-8 carbon to a maximum of 2660 m² g⁻¹ and 1.38 cm³ g⁻¹ for the sample activated at 800 °C (KOH–CMK-8 mass ratio of 4). By comparison, various samples exhibited similar CO₂ uptake capacities of around 3.2 mmol g⁻¹ at 298 K no matter what type of templated carbon was used as a precursor. Through analysis of the CO₂ adsorption quantities of different porous carbons, they found that CO₂ capture capacity depends on the presence of narrow micropores (<1 nm) rather than on the surface area or pore volume of the porous carbons. Furthermore, it was found that these porous carbons exhibited a high CO₂ adsorption rate, a good selectivity for CO₂–N₂ separation and they can be easily regenerated.⁹⁸ 

Luis et al. demonstrated a novel strategy to synthesize hierarchical porous carbons with large pore volumes, high surface areas, and more importantly, tunable micro/meso/macro porosities. The approach is based on a combination of ice templating, colloidal silica and CO₂ activation for generating interconnected macro-, meso-, and microporosity, respectively (Figure 1.16). During the synthesis, the range, size and extent of porosity can be easily controlled with different conditions. The synthesis was simple, green, reproducible, and used widely available and inexpensive starting materials, all of which made the process highly scalable. It produced a good CO₂ capture capacity of 4.2 mmol g⁻¹. More importantly, for dilute CO₂ (10% CO₂–90% N₂ gas mixture), the sorption capacity measured under dry and moist conditions was the same, which was hardly affected by the presence of moisture.¹⁶⁷ 

Many studies have proved that nitrogen-doping is an effective strategy to enhance the CO₂ adsorption capacities of porous carbon adsorbents. However, to achieve a high doping level of nitrogen and a significant porosity in the carbon material simultaneously remains a challenge. Li et al. explored a facile approach to construct three-dimensional macroporous nitrogen-doped carbons with periodic ordered macropores and further etched micropores on the frameworks by KOH activation.¹⁶⁹  These hierarchical porous nitrogen-doped carbons were partly graphitic and possessed relatively high nitrogen contents of about 14.45%. The CO₂ adsorption capacities of these samples were about 2.69 mmol g⁻¹ at 298 K and 3.82 mmol g⁻¹ at 273 K at 1 bar with an extraordinarily high CO₂–N₂ selectivity of 134 calculated from the IAST method. Such an unprecedented CO₂–N₂ selectivity was largely associated with the unusually high N content and the partially graphitic framework of this material.

The one-step method is simple and practicable, and is an important way to synthesize hierarchical porous carbons. Hao et al. designed a series of poly(benzoxazine-co-resol)-based porous carbon monoliths with multiple-length-scale porosity (macro-, meso-, and micropores), a nitrogen-containing framework (polar surface), and remarkable mechanical strength (Figure 1.17).¹⁷⁰  Interconnected mesopores (cubic Im3m symmetry) were formed by the assembly of poly(benzoxazine-co-resol) under the direction of surfactant Pluronic F127, while co-continuous macropores were formed due to polymerization-induced phase separation. Herein, organic amines were used as both catalyst and nitrogen source. The nature of the organic amines strongly influences the assembly of the mesostructure. It is reported that protic organic bases, such as DMA (dimethylamine), EDA (ethylenediamine), and DAH (1,6-diaminohexane), favored the formation of ordered mesopores in carbon monoliths. In contrast, aprotic organic bases (dimethylamine and triethylamine) resulted in the formation of microporous carbons due to the lack of hydrogen bonds between these amines and F127 molecules. HCM-DAH-1 carbon gave a CO₂ adsorption capacity of 3.3 mmol g⁻¹ at 0 °C and 780 mm Hg, a high CO₂–N₂ selectivity of 28, and high isosteric heats of adsorption of 21.1–35.9 kJ mol⁻¹, indicating that the prepared carbon preferentially interacted with CO₂. Further CO₂ activation at 900 °C significantly increased the specific surface area and newly generated micropores with a size of 0.9–1.3 nm; as a result, the CO₂ capacities increased to 4.9 mmol g⁻¹. However, the selectivities of CO₂-to-N₂ and isosteric heats of adsorption decreased after the additional activation, which may be due to the widening of the microporosity.

Liu et al. reported the one-step synthesis of nitrogen-doped graphene-like meso-macroporous carbons as highly efficient and selective adsorbents for CO₂ capture. Graphene-like meso-macroporous carbons (GMCs) with high nitrogen contents and controllable nitrogen sites were synthesized by one-step carbonization of dicyandiamide or melamine with glucose. The unique layered structure and the presence of abundant meso-macropores in the GMCs largely enhance the degree of exposure and accessibility of anchored nitrogen sites to CO₂, which provides the GMCs with an excellent performance for the selective capture of CO₂ (about 101 calculated according to the IAST).¹⁷¹ 

Tian et al. reported a method of direct carbonization of algae biomass to generate hierarchical porous carbons. Algae biomass is a fantastic precursor for fabricating porous carbon due to its natural composition. Enteromorpha prolifera, which was selected as a carbon source, was converted into carbons with evident hierarchical micro-mesoporous structures through a simple direct pyrolysis of freeze-dried algae. The formation process for hierarchical pores was preliminary discussed. The surface areas of the obtained carbons were only close to 450 m² g⁻¹ with 70% areas inherited from mesopores, resulting in high density materials with the major pore sizes at around 1.5 and 5.0 nm. Moreover, the carbons show a CO₂ uptake of 6.48 mmol g⁻¹ at 20 bar and 25 °C.¹⁷² 

In this section, we have shown that researchers have fabricated an enormous number of porous carbons as CO₂ adsorbents. As we have reviewed, the porosity (e.g., single- or multi-level, narrow-distributed or wide-distributed, ordered or disordered, microporous or mesoporous, etc.) and surface chemistries (pristine or doped) could be engineered well. In general, their synthesis strategies could be roughly categorized into three types: activation, templating, and a combination of both. The CO₂ sorption capacities of these materials could be comparable to those of other well-known porous solid adsorbents, such as MOFs, ZIFs, zeolites, and others. Porous carbons have been one of the most promising candidates for CO₂ adsorbents, and will be more and more important in the field of CO₂ capture.

Graphene, the world's first two-dimensional (2D) material, has captured the attention of scientists, researchers, and industry worldwide. In simple terms, graphene is a thin layer of an sp²-bonded carbon sheet. This material was first isolated and characterized in 2004 by Andre Geim and Konstantin Novoselov at the University of Manchester. In 2010, they received the Nobel prize in physics for their work.

As we know, graphene has many advantages, such as a high specific surface area, superior electron mobility, easy self-assembly into three-dimensional (3D) macroscopic materials with controlled microstructures, high mechanical, chemical, thermal, and electrochemical stabilities, etc. With these amazingly attractive properties, graphene already exhibits tremendous promise in many applications: nanoelectronics,¹⁷³  supercapacitors,¹⁷⁴  fuel cells,¹⁷⁵  batteries,¹⁷⁶  solar cells,¹⁷⁷  water filtration,¹⁷⁸  gas sorption, separation and storage,^179,180  and sensing,¹⁸¹  etc. In recent years, there has also been an enormous effort to explore graphene and its oxygen-containing derivative, graphene oxide (GO), for CO₂ capture applications. The latest progress with both the experimental and theoretical research into graphene-based CO₂ adsorbents has been summarized in some recent reviews.^182,183  In this section, we mainly focus on the experimental reports of graphene-based porous materials for CO₂ capture.

As gas adsorption is highly dependent on the pore texture of the adsorbents, it is highly desirable that the 2D graphene sheets are engineered further in specific ways to obtain a tailored porosity with a high surface area and pore volume. Accordingly, several techniques have been adapted to develop graphene-based porous adsorbents for CO₂ capture.

Chemical activation with KOH is an efficient technique to introduce pore texture into graphene. In 2011, Zhu et al. first reported the KOH activation of microwave exfoliated GO (MEGO).¹⁷⁴  The KOH activation generated a much developed porosity in the product carbon. The specific surface area of the activated MEGO could be controlled by the ratio of KOH–MEGO. The optimized activated MEGO had a well-defined micro-mesopore-sized distribution with a very high BET surface area and pore volume, up to 3100 m² g⁻¹ and 2.14 cm³ g⁻¹, respectively. This activated MEGO carbon could provide a large and accessible surface area for charge accommodation, and, therefore, exhibited an outstanding capacitive performance in organic and ionic electrolytes to achieve high gravimetric energy densities.

In 2012, Srinivas et al. reported the synthesis of a range of high surface area graphene-oxide-derived carbons (denoted as GODCs) and their applications toward carbon capture and methane storage.¹⁸⁴  GODCs are prepared by chemical activation with KOH from thermally exfoliated GO and solvothermally reduced GO precursors. The authors explored the effects of activation temperature and KOH concentration upon the development of porosity in the resultant carbons. At low activation temperatures and KOH concentrations, the activation mainly generated pore development in the microporous region, while increasing the activation temperature and KOH concentration resulted in a large proportion of pore development in both the microporous and mesoporous regions (>2 nm). The optimum activation conditions were found to be 800 °C and a GO–KOH ratio of 1 : 9 to obtain a porous carbon (GODCsol-800) with the maximum specific surface area and total pore volume of 1900 m² g⁻¹ and 1.65 cm³ g⁻¹, respectively. The results of CO₂ sorption displayed that the high-pressure CO₂ adsorption trend is more or less linearly dependent on the BET surface area, total pore volume, and average pore size, while the low pressure (1 bar) adsorption behavior is almost independent of the surface area, but is closely related to the narrow pore size distribution in the micropore region. The optimized sample of GODCsol-800 exhibited a high CO₂ adsorption capacity of 72.1 wt% at 300 K and 20 bar, as well as a high methane adsorption capacity of 17.5 wt% at 300 K and 35 bar.

Physical activation with steam and CO₂ has also been employed to fabricate porous graphene-based materials for CO₂ capture. Sui et al. developed porous graphene-based carbons through physical activation of graphene aerogels using steam as the activating agent.¹⁸⁵  The activation temperature plays a critical role in determining the BET surface area and pore volume of the resultant carbons, and the optimal activation temperature to obtain the highest BET surface area and pore volume is 850 °C. A low activation temperature (750 °C) was less efficient at enhancing the porosity due to the slow reaction rate between graphene aerogel particles and steam. Meanwhile, above 850 °C, the samples also showed a decreased specific surface area and pore volume due to the high burn-off of graphene sheets, thus resulting in the destruction of the porous structure and a lower yield. The steam-activated graphene aerogel exhibited a high specific surface area (830–1230 m² g⁻¹), an abundant large pore volume (2.2–3.6 cm³ g⁻¹), and excellent thermal stability. The optimized SAGA-850 showed a CO₂ adsorption capacity of 2.45 mmol g⁻¹ at 1 bar and 273 K, much higher than that of the non-activated GA sample (1.45 mmol g⁻¹) under the same conditions.

Chowdhury and Balasubramanian prepared graphene-based porous carbons though CO₂ activation using reduced graphene oxide as a precursor (Figure 1.18).¹⁸⁶  By increasing the activation temperature of CO₂, the specific surface area, micropore pore volume and total pore volume increase, resulting in an improvement of CO₂ uptake. Specifically, the adsorbent material obtained at an activation temperature of 950 °C (i.e., a-RGO-950) exhibited the largest specific surface area (above 1300 m² g⁻¹), the highest pore volume (over 1 cm³ g⁻¹), and a well-defined bimodal micro-mesoporous structure. This adsorbent material displayed a good gravimetric CO₂ uptake (3.36 and 2.45 mmol g⁻¹ at 0 °C, 1 bar and 25 °C, 1 bar, respectively), rapid adsorption kinetics, as well as stable and readily reversible adsorption–desorption cycling behavior at room temperature. Moreover, a-RGO-950 exhibited excellent Henry law CO₂–N₂ selectivities of 162 and 253 under conditions pertinent to CO₂ capture from the dry flue gas steam of a coal-fired (75% N₂ and 15% CO₂) and natural-gas-fired (80% N₂ and 5% CO₂) power plant, respectively.

Similarly, Xia et al. prepared a series of hierarchical porous graphene-based carbons (HPGCs) by CO₂ activation of graphite oxide.¹⁸⁷  HPGC-850, which was prepared by 2 hours of CO₂ activation at 850 °C, possessed the highest specific surface area and micropore volume, thus exhibited the highest CO₂ sorption capacity of 1.76 mmol g⁻¹ at 274 K and 1 bar, as well as the highest H₂ sorption capacity of 3.76 mmol g⁻¹ at 77 K and 1 bar.

Other techniques, like hydrogen-induced exfoliation,¹⁸⁸  and covalent functionalization,¹⁸⁹  are also used to introduce pore structures into graphene materials for CO₂ capture. For instance, porous graphene materials prepared via hydrogen-induced exfoliation of graphite oxide exhibited a maximum sorption capacity of 21.6 mmol g⁻¹ at 11 bar and 25 °C.¹⁸⁸  The physical adsorption nature of CO₂ in the prepared graphene material was confirmed using a Fourier transform infrared spectroscopy (FTIR) study. Beyond the normal hydroxyl (3435 cm⁻¹), carboxyl (1726 cm⁻¹) and carbonyl (1173 cm⁻¹) peaks, a new peak was observed at 2324 cm⁻¹ in the IR spectra. This peak corresponds to the asymmetric stretching of CO₂, implying physisorption of CO₂ onto graphene sheets.

Chowdhury reported the CO₂ capture performance of thermally treated graphene oxides.¹⁹⁰  As the thermal treatment temperature increased, the pore texture of the graphene sheets became more developed. GPN-800 treated at 800 °C processed the highest specific surface area (484 m² g⁻¹) and micropore volume (0.094 cm³ g⁻¹), resulting in the highest CO₂ uptake of 2.9 mmol g⁻¹ at 0 °C and 1 bar. Kumar designed two pillared porous graphene frameworks (PGFs) by linking reduced graphene oxide layers with 1,4-diethynylbenzene (PGF-1) and 4,4′-diethynylbiphenyl (PGF-2) via a C–C coupling reaction.¹⁸⁹  Both frameworks show high CO₂ uptakes of 112 wt% for PGF-1 and 60 wt% for PGF-2 at 195 K and 0.85 atm.

Besides, graphene has been widely used to fabricate composite materials as a carrier support due to its 2D structure with a high surface area. Lu's group reported the preparation of porous carbon nanosheets (PCNs) with precisely tunable thicknesses in which GO played the role of shape-directing agent. The resorcinol-formaldehyde resins grew in situ on the GO sheets due to the bridging effect of asparagine and were converted into porous carbon sheets. The thickness of the carbon sheets was tuned from 20 to 200 nm according to the mass ratio of the resin–GO. At 25 °C and 1 bar, the maximum CO₂ uptakes of PCN-9.9, PCN-17, and PCN-71 with thicknesses of 9.9, 17, and 71 nm, respectively, were 2.02, 2.36, and 2.88 mmol g⁻¹. Moreover, these porous carbon sheets showed a good ability to separate CO₂ from simulated flue gas (a water-saturated CO₂/N₂ stream) under dynamic conditions; the CO₂ capacity reached 0.28 mmol g⁻¹ at a CO₂ concentration of 4 v%. PCN-17 could stably work for 200 cycles in total under a CO₂–N₂ gas stream of 14 : 86 v% (Figure 1.19). Kim and colleagues reported various N- or S-doped porous carbons by using graphene/polypyrrole,³¹  rGO/polyaniline,¹⁹²  rGO/polyindole,¹⁹³  and rGO/polythiophene³⁷  as carbon precursors and KOH as an activating agent. The graphene in the carbon precursor is believed to increase the contact area between the KOH activator and carbon precursors. As a result, the synthesized carbons featured a high surface area, a large pore volume, and developed microporosity, and captured a large amount of CO₂ under ambient conditions (>4 mmol g⁻¹), a value much higher than those of the porous graphene materials discussed earlier.

Hybrid CO₂ adsorbents, like PEI/graphene,¹⁹⁴  polyaniline/graphene,¹⁹⁵  LDH/graphene,¹⁹⁶  etc., have also been reported to exhibit excellent CO₂ capture performances, in which the supporting effect of the graphene ensures the high loading content of the active materials, and improves the stability and CO₂ capacities. These hybrid materials will be briefly discussed in Section 1.5. Clearly, various graphene-based materials have been investigated in the CO₂ capture field. The present studies have demonstrated the promise of graphene-based materials for selectively capturing and isolating CO₂ from flue gas. However, this investigation is still in its infancy, and further research is needed.

Carbon nanotubes (CNTs) are allotropes of carbon with a cylindrical nanostructure. These materials were first discovered by Ijima in 1991 as minority by-products of fullerene synthesis. According to the number of carbon walls, CNTs are categorized as single-walled nanotubes (SWNTs) and multi-walled nanotubes (MWNTs). The inner diameter of CNTs can vary from approximately 1 nm for SWCNTs to over 10 nm for MWCNTs. Considering the uniform inner diameter and almost defect-free wall of CNTs, the interaction of gas molecules with the inner pores of the CNTs can be described by a smooth potential energy surface. Recent computational simulations have proved that the smoothness of the inner pores make gas molecules diffuse and transport rapidly through the CNTs. For example, the transport rates of light gases, such as H₂, CH₄, and N₂, in CNTs are orders of magnitude faster than in microporous materials with comparable pore sizes.^197,198  This feature makes CNTs ideal candidates for the selective sorption and separation of gases, such as CO₂ capture from flue gas.

Lu et al. carried out a comparative study of CO₂ capture by CNTs with inner diameters <10 nm, activated carbons, and zeolites.¹⁹⁹  Raw CNTs showed a CO₂ sorption amount of 22.7 mg g⁻¹ under an influent CO₂ concentration of 10% at 25 °C, which is slightly less than activated carbons (24.9 mg g⁻¹) but larger than zeolites (19.0 mg g⁻¹) under the same conditions. After modification with 3-aminopropyltriethoxysilane, the modified CNTs showed an enhanced adsorption capacity of CO₂ of up to 40.9 mg g⁻¹ and 96.3 mg g⁻¹ at CO₂ concentrations of 10% and 50%, respectively. The mechanism of CO₂ adsorption on these adsorbents appears to be mainly attributable to physical force.

Most studies of pure CNTs for CO₂ capture are simulations, including quantum calculations (i.e., density functional theory (DFT) and ab initio simulations),^200–202  grand canonical Monte Carlo (GCMC) simulations,^203,204  and molecular dynamics (MD) simulations.²⁰⁵  DFT calculations predicted that zigzag SWNTs possessed stronger bonds with CO₂ molecules compared to armchair and chiral ones, and N-doping could increase the binding energies between zigzag SWNTs and CO₂, suggesting that zigzag SWNTs show greater promise as a means of CO₂ capture.²⁰⁰  Based on Monte Carlo simulations, Liu and Bhatia found that increasing the nanotube diameter from 1.36 nm (10, 10) to 2.03 (15, 15) led to enhanced CO₂ capacity at 5 bar due to the increase of accessible volume, while the change in chirality had a negligible effect.²⁰⁵  Furthermore, the efficiency of CO₂ capture from flue gas can be significantly improved by pre-adsorbing small clusters of water in carbon nanotubes.²⁰⁵  Rahimi et al. simulated adding a positive charge to the carbon nanotubes, which caused a significant increase in adsorption (up to 35% at a pressure of 1.88 bar), while the gas adsorption decreased by up to 15% on negatively charged carbon nanotubes.²⁰⁶  The increase or decrease of adsorption upon charging was attributed to the change in potential energy for the interactions between the individual CO₂ molecules and the nanotubes.

Owing to their extraordinary thermal conductivity, mechanical, and electrical properties, carbon nanotubes find applications as additives to various structural materials. Jin et al. fabricated carbon composite monoliths by CO₂ activation of a composite of a commercial phenolic resin and carbon nanotubes.²⁰⁷  CNTs in the composite monoliths played an important role in improving pore structures and the CO₂ capacity of CNT-modified CCMs. CNT-modified carbon monolith samples were much more reactive with CO₂ during activation; they exhibited significant burn-off within a much shorter activation time, thus achieving a well-developed porosity, especially a narrow microporosity. CPD-30 with a burn-off of 25.5 wt% gave the highest CO₂ uptake of 3.5 mmol g⁻¹ at 298 K and 1 atm, while CPD-15 with a burn-off of 16.7 wt% had the highest CO₂ uptake of at 1.1 mmol g⁻¹ at 298 K and 0.15 atm, more than twice that of the resin alone, Res-60 (only 0.55 mmol g⁻¹). Mixed matrix membranes (MMMs) based on CNTs dispersed inside a polymer matrix have been well-developed and characterized for CO₂ separation measurements. For instance, Ahmad et al. fabricated MMMs from cellulose acetate with β-cyclodextrin-functionalized MWCNTs (MWCNTs-F) using a wet phase inversion technique.²⁰⁸  The results showed that MMMs with 0.1 wt% loadings of functionalized MWCNTs demonstrated enhanced permeance and selectivity towards the separation of CO₂/nitrogen (N₂), as well as outstanding mechanical properties with a high tensile strength of 16 MPa. This behavior results from the enhanced compatibility of MWCNTs-F within the polymer chain segments, which increased the sufficient interlayer spacing.

As discussed, CNTs have attracted much attention in the field of CO₂ capture. These materials could directly serve as adsorbents, as well as supporters/additives for the fabrication of hybrid adsorbents. However, the studies reported have mainly focused on simulations; more experimental studies should be performed in the future.

CO₂ sorption on pristine carbon materials is essentially “physisorption”, which is very sensitive to temperature and pressure, and is usually low-selective. Generally, the sorption capacities of CO₂ on carbon materials will significantly decrease as the sorption temperature increases or the sorption pressure decreases. For example, the CO₂ uptakes of the previously reported porous carbons at 0.15 bar and 25 °C (typical CO₂ partial pressure in flue gas) are usually lower than 1.5 mmol g⁻¹. Reducing the pore size to extremely small micropores (<0.5 nm) and introducing polar surface chemistry can elevate the CO₂ uptake to about 2 mmol g⁻¹ at low pressure or high temperature.⁸⁸  However, this will lead to a serious obstruction to gas diffusion kinetics. Synthesis of carbon-based hybrid sorbents is an effective strategy to enhance CO₂ capture performance. Mesoporous carbons, carbon nanotubes, and graphene are commonly used because these materials possess larger pore sizes, an opened pore texture and a high surface area, ensuring a high mass loading and good dispersion of the incorporated components. Herein, carbon-based hybrid sorbents are classified into carbon–organic hybrid sorbents and carbon–inorganic ones according to the nature of the incorporated component.

Organic amines are the most commonly used functional components for carbon–organic hybrid sorbents. Various amines, such as monoamines, diamines, triamines, polyamines and amine-containing polymers, have been studied for this purpose. In the last few years, several excellent review papers have been published and provide a good perspective towards amine-functionalized solid adsorbents for CO₂ capture and separation.^209–211  Generally, there are two ways to incorporate the amines: (1) impregnation, and (2) grafting.

With impregnation, the amine compounds are dispersed inside the pores of the carbon support, and thereby, they produce an enhanced CO₂ capture performance relative to the mass loading of the bulk amines. In theory, the more amines that are loaded, the higher the CO₂ capture capacities of the amine-impregnated carbon materials. The dispersion of the incorporated amines is also important. The excess loading of amines may result in a decline of the amine utilization ratio due to a lengthening of the diffusion distance and the agglomeration of the amine molecules.^212,213 

Polyethyleneimine (PEI) is most commonly used due to its high CO₂ uptake/release capacity and good stability. For example, Wang et al. developed highly efficient CO₂ sorbents by loading PEI on mesoporous carbons.²¹²  As the PEI loading increased, the utilization ratio of the PEI first increased and then decreased, and the optimal PEI loading is determined to be 65 wt% with a CO₂ sorption capacity of 4.82 mmol g⁻¹ in 15% CO₂/N₂ at 75 °C.²¹²  These sorbents could work over a wide CO₂ concentration range from 5% to 80%. Furthermore, moisture was found to have a promoting effect on the sorption separation of CO₂. This group also reported a sorbent of PEI-impregnated millimeter-sized mesoporous carbon spheres for post-combustion CO₂ capture (Figure 1.20).²¹⁴  Under the optimal conditions of using a high molecular weight PEI and polyethylene glycol (PEG) loading of 20 wt%, the adsorbent exhibited a high equilibrium adsorption capacity of 187.5 mg g⁻¹ (4.26 mmol g⁻¹) for 15% CO₂ at 75 °C at a relative humidity of 60%. Noticeably, these adsorbents could be regenerated by a novel electric swing adsorption (ESA) operation due to the good electrical conductivity of the carbon support.

With grafting, the amine groups are chemically bound to the carbon supports. In 2011, Houshmand et al. reviewed this strategy well.²⁰⁹  In their review, the methods used for grafting are roughly categorized into amination, silylation with aminosilanes, nitration followed by reduction, anchoring diamines/polyamines, anchoring halogenated amines, surface-initiated polymerization of ethyleneimine and its derivatives, and plasma treatment etc. To compare, the interactions between the grafted amines and carbon supports are strong covalent bonds, while the interactions between the impregnated amines and carbon supports are mainly weak interactions, such as dipole–dipole interactions, van der Waals forces, or hydrogen bonding. Therefore, amine groups incorporated by grafting are more stable and do not desorb during regeneration.²⁰⁹  Nevertheless, grafting suffers the shortcoming of a complicated synthesis process. In addition, with both strategies, the porosities of the carbon supports should be tailored to be suitable for the amine compounds selected.

Besides the commonly used mesoporous carbon materials, graphene and carbon nanotubes have also been used as supports. For instance, Liu et al. prepared hybrid sorbents by covalent grafting of PEI on hydroxylated 3D graphene (HG).¹⁹⁶  The optimal HG-PEI hybrid sorbent exhibited a high CO₂ adsorption capacity up to 4.13 mmol g⁻¹ at 1 atm of dry CO₂. PEI-impregnated graphene-silica sheets also showed an excellent CO₂ adsorption capacity (171 mg g⁻¹ at 75 °C/100 kPa) and good cycle stability (>20 cycles).¹⁹⁴  Lu et al. grafted 3-aminopropyltriethoxysilane (APTS) onto MWCNTs, and studied the CO₂ capture behaviors of the amine-loaded MWCNTs.^215,216  The adsorption capacity of CNT(APTS) was significantly influenced by the presence of water vapor and reached a maximum of 2.45 mmol g⁻¹ at 2.2% water vapor.²¹⁵  In addition to the high surface area and opened pore texture, which promotes diffusion of CO₂ to the active adsorption sites, the high thermal conductivity of graphene and CNTs allowed a fast transfer of heat and avoided degradation of the organic amines.

Other organics, such as ionic liquids and quinones, are also used to fabricate carbon-based hybrid sorbents. Tamilarasan reported that polymerized ionic liquid functionalized graphene showed an adsorption capacity that was 22% higher than graphene, while ionic liquid functionalization improved it only by 2%.²¹⁷  Wang et al. developed a hydroquinone (quinone)-functionalized hierarchical porous carbon sorbent via the Friedel–Crafts reaction.²¹⁸  The hydroquinone-grafted carbon gave a CO₂ adsorption capacity of 5.41 mmol g⁻¹ at 1 bar and a CO₂–N₂ selectivity of 26.5, higher than those of pristine porous carbon.

Fabrication of carbon–inorganic hybrid materials is another approach to enhance CO₂ uptake. Various carbon materials, such as porous carbon materials,²¹⁹  carbon nanotubes,²²⁰  and graphene and its derivative,²²¹  are used as solid supports. The inorganic sorbents are mostly metal compounds, including metal coordination polymers (MOFs,^222,223  ZIFs²²⁴ ), metal oxides (MgO,²²¹  CaO,²²⁵  ZnO²²⁶ ), alkali metal carbonates,²²⁷  layered double hydroxides (LDHs),²²⁸  etc. According to a recent review by Prof. Wang, these carbon–inorganic hybrid sorbents could be classified into intermediate-temperature (200–400 °C) and high-temperature (>400 °C) solid CO₂ sorbents, and have been summarized well.²²⁹ 

In summary, a large number of carbon-based hybrid sorbents have been studied. Due to the strong affinity of incorporated components to CO₂, these sorbents have shown high CO₂ capture capacities at high sorption temperatures or low CO₂ concentrations. Carbon supports have been proved to have significant promoting effects on CO₂ capture, which could be mainly because: (1) the high surface area and pore volume of carbon supports improve the dispersion of active materials while ensuring a high mass loading; (2) the opened pore texture facilitates the diffusion of CO₂ molecules to the active adsorption sites; (3) the high dispersion also guarantees a high adsorption efficiency and good cyclic stability.

A large number of studies have shown that there are two main strategies to improve the CO₂ adsorption capacity of carbons. One simple strategy is to introduce polarity by chemical doping in the carbon framework to enhance its affinity with CO₂. Another approach that has been investigated to enhance CO₂ capture capacities is to tailor pore structures to make them suitable for CO₂ adsorption. In this section, we pay attention to studies of the effect of carbon structure on CO₂ adsorption, which are of use for the development of high-performance carbon-based adsorbents.

The IUPAC (International Union for Pure and Applied Chemistry) has proposed a classification of pores based on pore sizes. Pores are generally classified into micropores (<2 nm), mesopores (2–50 nm), and macropores (>50 nm). Micropores are further divided into ultramicropores (<0.7 nm) and supermicropores (0.7–2 nm). Since the porosity of carbons is responsible for their adsorption properties, the analysis of the porosity, especially the microporosity, is very important to foresee the behavior of these porous solids in CO₂ capture applications. Pores in carbon materials could be identified by several techniques depending on their type and size, such as gas adsorption (opened pores < 100 nm), mercury porosimetry (opened macropores), small angle scattering (SAS) (latent pores, including closed pores), transmission and scanning electron microscopy (TEM and SEM) (extrinsic pores on the surface of carbon materials), etc.

Physical adsorption of various gases is, undoubtedly, the most widely used technique. N₂ is the most widely used adsorptive. However, N₂ adsorption at 77 K is not suitable for the characterization of narrow microporosity (size < 0.7 nm), because diffusion kinetics of N₂ molecule is very slow in ultramicropores at a liquid N₂ temperature of 77 K. To overcome this problem, the use of CO₂ sorption at 273 K has been proposed. The higher temperature of CO₂ adsorption results in a larger kinetic energy of the molecules, which are able to enter the narrow pores. However, CO₂ adsorption analysis of the whole porosity should be performed at high pressures due to its high saturated vapor pressure (P⁰) at 273 K (34.7 bar). In practice, the combination of N₂ and CO₂ adsorption allows us to obtain comprehensive information about the porosity.

In order to evaluate the specific surface area and PSD, various methods have been applied to gas isotherms, including, but not limited to, the BET (Brunauer–Emmett–Teller) method, D–R (Dubinin–Radushkevich) plot, BJH (Barret–Joyner–Halender) method, t plot, HK (Horvath–Kawazoe) method, and the DFT (density functional theory) method. Porous carbon materials, especially activated carbons, typically possess wide PSDs ranging from ultramicropores, supermicropores and mesopores to macropores. The NLDFT (non-linear density functional theory) method has been widely applied and featured in a standard by ISO-15901-3.²³⁰  The standard NLDFT model assumes an idealized flat graphitic-like surface. Due to the heterogeneity in the chemistry and structure of the carbon materials, this model may not describe all the subtleties of a real carbon surface. The QSDFT (quenched solid-state density functional theory) model takes into account carbon surface roughness features. This model provides a better representation of the surface properties, including surface roughness features,^231,232  and has been recommended in recent years.^106,233  However, QSDFT models are not yet available for some of the measurement instruments (e.g., micromeritics instruments) and sorption conditions (e.g., CO₂, 0 °C). The NLDFT method is still widely used to evaluate porosity from N₂ sorption measurements.

Micropores, especially, feature in interactions between gas adsorbate molecules and adsorbents, and there are still no definitive methods to determine the PSD in the micropore region. For the analysis of microporosity, Dubinin and Radushkevitch have developed an equation:

where V (cm³ g⁻¹) is the volume filled at a temperature T and a relative pressure (P/P₀), V₀ (cm³ g⁻¹) is the micropore volume, and k and β are constants for the pore structure and the affinity coefficient (e.g., β = 0.35 for CO₂). This equation, based on the postulation that the mechanism for adsorption in micropores is that of pore filling rather than layer-by-layer surface coverage, generally applies well to adsorption systems involving only van der Waals forces and is especially useful to describe adsorption on activated carbon.^109,234  This equation has been used to calculate micropore volumes and to estimate the mean pore size widths (L) for both N₂ and CO₂ adsorption on different samples.^{88,89,235,236}  The mean pore width could been estimated using the equation of Stoeckli et al. (eqn (1.15))²³⁷  when the characteristic energy (E₀) is between 42 and 20 kJ mol⁻¹:

Its validity has been tested by direct determination of the distribution in the range 0.35–1.3 nm. A well-defined linear D–R plot over the entire log²(P₀/P) range indicates a uniform micro-PSD, while the non-linear one usually indicates a wide microporosity with various micropore sizes (Figure 1.21).²³⁸ 

In order to reveal the adsorption mechanism, isotherms are simulated using two typical adsorption models including Langmuir and Dubinin–Astakhov (D–A) models (Figure 1.22).

The well-known Langmuir isotherm model can be expressed by eqn (1.16).

where Q (cm³ g⁻¹) denotes the amount adsorbed, Q₀ (cm³ g⁻¹) denotes the saturated amount adsorbed, P (kPa) denotes the equilibrium pressure, and b/kP⁻¹ denotes the adsorption affinity.

A linear expression for the Langmuir equation is

Figure 1.22 shows the correlation of adsorption data with the Langmuir equation for a PANI-based activated carbon.¹⁰⁵  It was found that the CO₂ adsorption isotherms could not be fitted well by the Langmuir model.

The Dubinin–Astakhov (D–A) model is applied to the adsorption equilibrium of CO₂ onto the activated carbons. The D–A equation is proposed to describe a wide microporosity as a general form of the D–R equation, which can be written as

or

where W denotes the volume adsorbed, W₀ denotes the limiting micropore volume, E denotes the characteristic energy of the system, n denotes the heterogeneity parameter, R is the gas constant, T is the equilibrium temperature, and P₀ is the saturated pressure. Apparently, the experimental data can be fitted well by the D–A equation (Figure 1.22).

This simulation shows evidence that CO₂ adsorption on porous carbons can be depicted well by a micropore filling mechanism, not by monolayer adsorption. In other words, the CO₂ sorption capacity should depend on the pore volume of a certain pore size range but not the specific surface area.

The main contribution of micropores, especially ultramicropores, to CO₂ uptake, has been widely recognized. Therefore, it is necessary to finely explore the relationship between microporosity and CO₂ capture capacity, which is significant for designing high-performance carbon adsorbents for CO₂ capture. A pioneering study of the relationship between porosity and CO₂ capacity was conducted by Gogotsi and his co-workers.¹⁰⁶  They prepared a large number of different carbide-derived carbons (CDCs) derived from nano- and micrometer-sized precursors, with and without activation, and investigated their CO₂ sorption performances. Various porosity parameters, including specific surface area, pore size, and pore volume, were correlated with the resulting sorption properties. They found that there was no clear dependency between the total pore volume or the volume of pores larger than 1 nm and CO₂ capacities at 1 bar (Figure 1.23). Moreover, there is no clear trend correlating CO₂ uptake and average pore size. Hence, they concluded that the CO₂ sorption behavior at ambient conditions may be largely governed by the volume of micropores, similar to the result with H₂ sorption.

They further correlated the micropore volumes of various pore sizes with CO₂ uptake at different sorption pressures. At 1 bar, pores smaller than 0.8 nm were found to contribute the most to CO₂ uptake and at 0.1 bar, pores smaller or equal to 0.5 nm were preferred (Figure 1.23). This result is regardless of carbon particle size and activation/non-activation. Coefficients of determination between CO₂ uptake at a certain pressure (0.1–1.0 bar) and the cumulative pore volume of pores smaller or equal to a certain pore size (0.4–1.5 nm) were also obtained (Figure 1.24). With a lower total pressure, smaller pores contributed more to the measured CO₂ uptakes.

Almost all of the studies on CO₂ capture were carried out at low temperatures such as 0 °C and room temperature (Tables 1.1–1.3). However, industrial flue gas, the biggest contributor of CO₂, is at a relatively high temperature (>70 °C). The sorption temperature significantly influences gas uptake on porous sorbents, and a high temperature always leads to a low gas uptake. Apparently, a well-designed porosity for carbon adsorbents is required for CO₂ capture at a high temperature. An investigation into the relationship between CO₂ uptake and the porosity of carbons at different sorption temperatures, especially a high temperature (e.g., 75 °C), is also necessary.

In 2012, our group prepared a series of N-doped porous carbon materials by using PANI as a carbon precursor via a pre-carbonization and post-KOH activation method.¹⁰⁵  The micropore size distribution of the prepared carbons varied with the activation conditions (KOH–carbon ratio and activation temperature). These carbons showed very high CO₂ uptakes of up to 1.86 and 1.39 mmol g⁻¹ under 1 bar, 75 °C and 0.15 bar, 25 °C, respectively, which are amongst the highest of the known carbon materials for CO₂ capture.

It is well known that the smaller the micropore, the stronger the adsorption potential. The smaller micropores will be the preferred spaces for adsorption of CO₂ molecules, and could be filled to a higher degree due to their stronger adsorption potentials. Based on this view, the CO₂ sorption capacity per pore volume is used as a factor reflecting the adsorption capability of porous carbons. The correlations of CO₂ sorption capacities per pore volume with the volume fractions of micropores smaller than a critical size under different sorption conditions were investigated (Figure 1.25). At low temperatures (0 °C), the CO₂ sorption capacity per pore volume is strictly linear with the volume fraction of the pores smaller than 0.80 nm in total pore volume (high correlation coefficient of 0.97, Figure 1.25), agreeing well with the results of Gogosti.¹⁰⁶  Similar results are obtained at 25 °C and 75 °C, while this critical pore size (0.70 nm and 0.54 nm for 25 °C and 75 °C, respectively) decreases as the sorption temperature increases. Furthermore, the simulated lines plotted in Figure 1.25 are found to have an intercept of near zero, indicating that the small micropore is the most important (even exclusive) factor for CO₂ uptake. In a word, the CO₂ adsorption capacity of porous carbons is associated with the pores below a temperature-dependent size.

The correlation of CO₂ sorption capacity per pore volume at 0.15 bar, 25 °C, with the N contents of the carbons or the volume fractions of small micropores (<0.54 nm) in the total pores are also presented (Figure 1.26). Interestingly, the CO₂ sorption capacity per pore volume shows a very good linear correlation with the volume percentage of extremely small micropores (<0.54 nm). By contrast, the correlation of CO₂ sorption capacity per pore volume at 0.15 bar, 25 °C, with the N contents of the carbons obviously deviates from the linear. This fact further confirmed that the CO₂ capture capabilities of the investigated carbons are mainly contributed by the small micropores (<0.54 nm for 0.15 bar at 25 °C) but not the N-doping.

To enhance the interactions of porous carbons with CO₂ molecules, many authors have tried to introduce polarity (e.g., N-doping, S-doping, or cation-introducing, etc.) into the carbon frameworks because the CO₂ is highly quadrupolar and weakly acidic. However, the effect of heteroatom-doping, especially N-doping, on CO₂ capture is still unclear and is often controversial. Recently, there have been some experimental and simulation studies aimed at clarifying the effects of surface chemistry on the CO₂ uptake of carbons. Herein, we briefly summarize the recent relevant progress.

N-doping is widely used. Many authors have suggested that N-doping enhances the CO₂ capture performance of porous carbon adsorbents. In these works, the enhancement effect of N-doping is generally assumed to be chemical sorption based on a one-on-one base–acid interaction. Based on this view, our group prepared N-doped microporous carbons by a two-step casting process using zeolite NaY as a hard template, furfuryl alcohol as a carbon precursor and acetonitrile as the CVD (chemical vapor decomposition) substrate, in which the use of acetonitrile is to introduce nitrogen functionalities to the microporous carbons.²³⁹  Based on the analysis results of X-ray photoelectron spectra, it was found that the type of nitrogen group strongly depends on the treatment temperature of the carbon samples (Figure 1.27). The nitrogen groups on the carbons prepared at 500 and 600 °C are mainly Lewis-basic species of amides/pyrrolic (400 eV) and pyridinic (398.9 eV) groups, while most of these basic groups decomposed at a higher temperature (>600 °C) and were converted into aromatic nitrogen species built into the carbon matrix (mainly quaternary N with typical Lewis acidity, ∼401 eV). Although the carbon prepared at 500 °C possessed a very low CO₂ uptake (57 mg g⁻¹), this carbon showed the highest heat of adsorption (35–40 kJ mol⁻¹), much higher than the typical heat of physical adsorption (∼20 kJ mol⁻¹), indicating a strong affinity between the surface of this carbon with CO₂ molecules. As the carbon temperature increased, the heat of adsorption significantly decreased due to a decrease in the interactions between CO₂ and carbon materials resulting from the loss of unstable basic nitrogen groups at high carbonization temperatures. These results demonstrated that the basic N groups (amides/pyrrolic and pyridinic) had a positive effect on CO₂ capture via a mechanism of base–acid interactions, especially for the N-doped carbons prepared at a low carbon temperature.²³⁹ 

However, the preparation temperature of most carbon adsorbents is usually higher than 600 °C with the purpose of developing high microporosity. The nitrogen species in these N-doped carbons frequently appear in four weakly alkaline or Lewis-acidic types: pyridinic N, pyrrolic N, quaternary N, and pyridine-N-oxide (Figure 1.28). The one-on-one base–acid interactions could only be related to the lone pair of electrons in pyridinic-N and pyrrolic-N. Further, considering the relatively low N-doping level of the carbon adsorbents reported (<5 at%), the N-doping effect of acid–base interaction may be marginal.

In 2012, our group suggested a hydrogen-bonding interaction mechanism on the N-doped activated carbons through quantum chemical calculations.²⁴⁰  A theoretical N-doped carbon surface model (NCSM) containing five different typical N-containing functional groups was developed. For comparison, a pure carbon surface model (CSM) was also devised, as is shown in Figure 1.29. Density functional theory B3LYP was employed to study the interactions between these models and CO₂ at 12 different positions, corresponding to 12 different H atoms in the NCSM model, and all the configurations were optimized with the 6-31+G* basis set for all atoms using the Gaussian-03 suite package. The optimized results at the 6-31+G* level showed that there were eight NCSM–CO₂ complexes and seven CSM–CO₂ complexes. In these complexes, hydrogen bonds between CO₂ and NCSM/CSM were formed due to the high electronegativity of the oxygen atom in the CO₂ molecule. This type of weak hydrogen bond has been widely studied in recent years, although the interaction of C–H/O was weaker than that of typical hydrogen bonds such as O–H/O and N–H/O. Computational results indicated that binding energies for such hydrogen bonds were different at various positions. The larger the binding energy ΔE (kJ mol⁻¹), the stronger the adsorption affinity. The average binding energy of eight NCSM–CO₂ complexes was 7.84 kJ mol⁻¹, and that of CSM–CO₂ complexes was only 1.26 kJ mol⁻¹, suggesting that the hydrogen bonds in the NCSM–CO₂ complexes were much stronger than those in CSM–CO₂ complexes. This demonstrated that the introduction of N into a carbon surface facilitated the hydrogen-bonding interactions between the carbon surface and CO₂ molecules. The existence of this weak hydrogen-bonding interaction was further confirmed by the FTIR analyses of SK-0.3-700 carbon (Figure 1.30), which showed that the peak due to a C–H anti-symmetric vibration was broadened and red-shifted to a low wavenumber under a CO₂ atmosphere compared with that measured under a N₂ atmosphere.

In the previous section, we discussed a temperature-dependent pore size effect on CO₂ uptake. In that work, the nitrogen content of the PANI-based carbons was also correlated with CO₂ sorption capacities at different sorption temperatures (0–75 °C). What was surprising was that there was no positive correlation between the CO₂ sorption capacity and the N content, concluded from the considerably low coefficient of determination (Figure 1.31, R² ≈ 0.20). According to the D–A theory, the CO₂ molecules fill the micropores in a liquid-like state. We evaluated the CO₂ sorption capacities at 25 °C by multiplying the density of CO₂ at 25 °C by micropore volume for pore sizes smaller than 0.7 nm. These calculated values were very close to the measured CO₂ sorption capacities although the investigated carbons possess very different N contents (from 2.8 wt% to 6.7 wt%). Based on the these facts, we suggested an innovative view that the small micropore plays a critical role in high CO₂ uptake, and the N-doping doesn't enhance CO₂ capture capacities for porous carbons.

Sevilla et al. also investigated the role of micropore size and N-doping in CO₂ capture, and confirmed our view.²³⁵  They prepared two types of activated carbons with analogous pore textures and different surface chemistry (N-free and N-doped), and comprehensively compared the CO₂ capture performances of these carbons. The CO₂ capture capacities of N-free and N-doped carbons were analogous, and no significant differences could be observed between the isotherms of the N-free and N-doped samples over the whole pressure range at three sorption temperatures: −15, 0, and 25 °C. This observation further showed that the nitrogen functionalities present in these N-doped samples do not influence CO₂ adsorption capacities.

Kumar used molecular simulations to study the effects of quaternary N-doping on CO₂ uptake and CO₂–N₂ selectivity in representative carbon pore architectures (slit and disordered carbon structures) at 298 K.²⁴¹  The simulation results showed that N-doping can only deliver a marginal enhancement on CO₂ uptake, but it can improve CO₂–N₂ selectivity in smaller micropores. Therefore, they suggested that CO₂ uptake and CO₂–N₂ selectivity were predominantly controlled by the pore architecture as well as ultramicropores. They further found that the tendency of linear CO₂ molecules to lie flat on the carbon surface favored CO₂ uptake in slit pore architectures rather than disordered carbon pore structures. It should be noted that the N-doping effect demonstrated here may be difficult to exemplify experimentally if the material has a disordered pore architecture and complex surface chemistry (such as the presence of other functional groups).²⁴¹ 

There are some other studies on the effect of N-doping. For instance, Sánchez-Sánchez et al. suggested that nitrogen functionalities acted as basic sites and oxygen and phosphorus groups as acidic ones toward the adsorption of CO₂ molecules, and among the N-containing groups, pyrrolic groups exhibited the highest influence, while the positive effect of pyridinic and quaternary functionalities was much smaller.¹³⁴  Their results agreed well with our experimental results.²³⁹  On the contrary, a DFT simulation by Lim et al. indicated that pyridine groups were most effective at enhancing affinity with CO₂ via Lewis base–acid interactions, and pyrrolic groups generated weaker interactions with CO₂.²⁴²  Babu et al. carried out experimental research on the CO₂ capture of mesoporous carbon nanotubes, and suggested that the presence of nitrogen functionalities has a beneficial influence on the CO₂ adsorption characteristics over a wide range of pressure (0–36 bar).²⁴³  In a word, the influence of nitrogen incorporation on CO₂ capture is still a controversial research topic with conflicting conclusions, therefore, we can't draw firm conclusions here.

Besides N-doping, other types of heteroatom-doping, such as S-doping,^{37,121,244–246}  O-doping,^{134,247–250}  and P-doping,²⁵¹  etc. have also been researched.

There are several reports on the effects of S-doping on CO₂ capture. For instance, Xia et al. found that S-doped microporous carbon materials templated by EMC-2 possessed a homogeneous thiophenic S-doping, and exhibited a large isosteric heat of CO₂ adsorption. They suggested that the thiophenic groups strongly interact with CO₂ molecules via strong base–acid interactions and strong pole–pole interactions between the large quadrupole moment of CO₂ and the polar carbon surface associated with S-doping.¹²¹  Seema and his co-workers suggested that the oxidized S content played a more significant role in increased CO₂ adsorption, while the plot of total S content (the sum of oxidized S and thiophenic S) vs. CO₂ uptake showed a negative correlation.³⁷  DFT calculations indicated that the oxidized S groups attracted CO₂ molecules more strongly than pyrrolic N groups (Figure 1.32). The attraction energy of oxidized S groups with CO₂ was mainly due to the attraction energy between the negative O atom of thiophene (−0.94) and the positive C atom of the CO₂ molecule (+1.07). The enhanced effect of oxidized S (sulfone group) on CO₂ uptake was also reported for MOF-based CO₂ adsorbents.²⁵²  Li et al. reported that sulfur, nitrogen, and phosphorus doping were all effective to enhance the gas adsorption capacity of carbon materials; among them, sulfur doping was the most positive.²⁵¹  However, Bandosz et al. recently reported that the S-containing groups were not stable in a CO₂ atmosphere; the thiophenes were oxidized by CO₂ molecules to sulfones and sulfonic acids, which were thermodynamically unstable and further decomposed forming SO/SO₂ and water, providing additional electrons for CO₂ reduction.^245,246,253 

Oxygen-containing functional groups inherently exist in the carbon frameworks. To some extent, O-containing groups, such as carbonyls, alcohols and ethers, can also induce Lewis base–acid interactions with CO₂ due to the weak electron-donating effect of the oxygen atom. CO₂ adsorption measurements showed that the oxidation process led to an increase in CO₂ adsorption capacity for the porous carbons.^247,248  Liu and Wilcox carried out a simulation investigation on the effect of oxygen-containing surface functionalities on the CO₂ adsorption of microporous carbons.²⁵⁰  Bader charge analysis indicated that the oxygen atom was highly electronegative and thus can serve as a basic adsorption site to attract more CO₂ molecules. GCMC simulations predicted that oxygen-containing functional groups enhanced CO₂ adsorption in microporous carbon materials in the absence of water vapor. The surface heterogeneity leads to a more condensed packing pattern of CO₂ molecules in the micropores, allowing for maximum utilization of the limited pore space (Figure 1.33). They further observed that the induced polarity of the surface functionalization could enhance the selectivity of CO₂ over CH₄ and N₂, especially in a low-pressure regime.²⁴⁹ 

Metal ion-doping was also proved to enhance the CO₂ capture performance. For example, the K-doped porous carbons reported by Han et al. exhibited a high CO₂ adsorption capacity of 1.62 mmol g⁻¹ at 0.1 bar and 25 °C and a high kinetic CO₂–N₂ selectivity of 44.²⁵⁴  The calculation suggested that the highly ionic K–O bonds led to a strong polarization and charge separation within the carbon cluster and thus significant improvement of CO₂ adsorption. Zhu et al. reported a carbon adsorbent derived from pine-cone biomass, and suggested that the well-dispersed nitrogen and calcium dopants (Ca²⁺) were positive to CO₂ capture.²⁵⁵ 

Undoubtedly, surface chemistry influences the CO₂ capture of carbon-based adsorbents, including CO₂ capacity, heat of adsorption, and adsorption selectivity of CO₂ over other gases. Many experimental and simulation investigations have proved this conclusion. Based on the results obtained so far, the interactions between the surface doping and CO₂ molecules could be generally summarized as Lewis acid–base interactions, hydrogen bonding, and electrostatic interactions. However, the exact role of heteroatom-doping, particularly N-doping, remains controversial and more studies are needed. Furthermore, the effect of heteroatom-doping should be evaluated in a practical (or simulated) condition with flue gas because the polar surface sites will enhance the affinity for the H₂O that saturates the flue gas, which will significantly degrade the CO₂ capture.

In this chapter, the recent research progress in the field of carbon-based materials for post-combustion CO₂ capture has been reviewed. The carbon-based CO₂ adsorbents touched on here are classified into porous carbons, graphene-based carbons, carbon nanotubes, and carbon-based hybrid adsorbents. In the section on porous carbon materials, activated carbons, templated carbons, hierarchical porous carbons, and MOFs (ZIFs, COFs)-based carbons, were respectively discussed. The carbon-based hybrid adsorbents were divided into carbon–organic hybrid materials and carbon–inorganic hybrid materials according to the nature of the incorporated component. Since the CO₂ capture of porous carbon adsorbents is primarily determined by the microporosity, and is strongly influenced by the surface chemistry of the carbon, the effects of pore size and surface chemistry on CO₂ capture are specifically summarized. Overall, carbon-based materials have been extensively researched, and show attractive prospects for pre-combustion CO₂ capture. Nevertheless, this research is in its infancy, and further investigations are needed in many areas.

1. Improving the CO₂ capture performances of carbon adsorbents: post-combustion CO₂ capture from flue gases possesses the characteristic of a low partial pressure of CO₂ (about 10–15% in flue gases) and selective separation of CO₂ and N₂. The capacity and selectivity of carbon materials are still low for this application. Tuning the pore size, inducing polarity by heteroatom-doping, and making hybrid composites, are efficient strategies for enhancing the CO₂ capture performances of carbons.²⁵⁶  On the subject of selectivity, three calculation methods are commonly used in the literature, including the capacity ratio of CO₂–N₂, the selectivity based on Henry's law, and the IAST selectivity, resulting in a difficulty of comparison between different carbon adsorbents. Among them, IAST selectivity is recommended because this is the standard in mixture adsorption predictions.²⁵⁷ 

2. Relationship between CO₂ capture and the carbon's structure: in this chapter, we have reviewed the research development of the effects of pore size and surface chemistry. However, controversies remain in this field. Besides, other structure factors of carbon adsorbents, such as the hierarchical pore structure, the ratio of micro/meso/macro porosity, the length of pores, the size and shape of the carbon particle, etc., also influence CO₂ capture (e.g., adsorption diffusion dynamics of CO₂), and need to be studied.

3. Evaluation of CO₂ separation under real (or simulated) adsorption conditions: the current works mostly focus on the static adsorption of a single-component gas. However, real CO₂ separations are performed as dynamic adsorption processes. Furthermore, flue gas in the field of post-combustion capture is a mixture of mostly N₂, CO₂, and some moisture. The co-existing gas components can cause a decline in the CO₂ adsorption capacity. Therefore, evaluation of CO₂ capture should be preferentially conducted on real (or simulated) flue gas under dynamic conditions.

4. Development of low-cost and sustainable carbon adsorbents: the cost and sustainability of CO₂ sorbents is crucial in practical applications due to the huge demand for them. In view of this regard, utilization of renewable precursors (typically various biomass sources), reduction/elimination of the activating agents (e.g., single-ion activation, self-template method), and lowering the energy consumption of carbonization (e.g., low carbonization temperature, microwave heating, etc.), can be considered in the manufacture of low-cost carbon adsorbents.

This work was financially supported by the Natural Science Foundation of China (NSFC21576158, 51302156, 21476264), Natural Science Foundation of Shandong Province (ZR2017JL014) and Distinguished Young Scientist Foundation of Shandong Province (JQ201215).