Nitrogen Atom-Doped Layered Graphene for High-Performance CO₂/N₂ Adsorption and Separation

Abstract

The development of high-performance CO₂ capture and separation adsorbents is critical to alleviate the deteriorating environmental issues. Herein, N atom-doped layered graphene (N-MGN) was introduced to form triazine and pyridine as potential CO₂ capture and separation adsorbents via regulation of interlayer spacings. Structural analyses showed that accessible surface area of the N-MGN is 2521.72 m² g⁻¹, the porosity increased from 9.43% to 84.86%. At ultra-low pressure, N-MGN_6.8 have exhibited a high CO₂ adsorption capacity of 10.59 mmol/g at 298 K and 0.4 bar. At high pressure, the absolute adsorption capacities of CO₂ in N-MGN_17.0 (40.16 mmol g⁻¹) at 7.0 MPa and 298 K are much larger than that of N-doping slit pore. At 298 K and 1.0 bar, the highest selectivity of CO₂ over N₂ reached up to ~133 in N-MGN_6.8. The research shows that N doping can effectively improve the adsorption and separation capacity of CO₂ and N₂ in layered graphene, and the interlayer spacing has an important influence on the adsorption capacity of CO₂/N₂. The adsorption heat and relative concentration curves further confirmed that the layered graphene with an interlayer spacing of 6.8 Å has the best adsorption and separation ability of CO₂ and N₂ under low pressure. Under high pressure, the layered graphene with the interlayer spacing of 17.0 Å has the best adsorption and separation ability of CO₂ and N₂.

1. Introduction

With the rapid developing of economic and world population [1], global CO₂ emissions from flue gases [2] have exceed 32 Gt of CO₂ per year. The greenhouse effect, triggered by the fast accumulation of carbon dioxide (CO₂) [3] has caused a lot of harm to society, environment and healthy of human-beings [4]. Carbon capture and storage (CCS) technology is an effective method to decrease the concentration of CO₂ and alleviate the greenhouse effect [5]. For CCS, the efficiency of capture and separation of CO₂ is mainly determined by the performance of adsorbent [6]. Aqueous amine solutions [7] are widely used as practical materials for CCS. However, they suffer from high parasitic energy consumption, adverse environmental impact, solvent losses and corrosion issues [8]. Comparing to amine solutions, solid adsorbents have high working capacity and lower regeneration energy when applied to gas capture. Therefore, they have become promising candidates as a CO₂ adsorbent for CCS [9]. Hitherto, various materials have been developed as the solid adsorbent, such as carbonaceous porous materials, porous organic materials, covalent organic frameworks, metal organic frameworks, and molecular sieves. Among them, graphene with split pore, as a typical carbonaceous material, has been considered as an ideal candidate because of its low density, high chemical stability, excellent processability, and eco-friendly behavior [10].

Many factors can affect graphene structures, such as electromagnetic radiation [11], external electric field [12], laser-induction [13], and chemical doping [14], and so on. To further improve the adsorption behavior of CO₂, nitrogen (N) doping was widely investigated experimentally and theoretically due to its ability to change the packing pattern of the adsorbate, as well as to enhance the interaction between gas and frameworks [15]. To et al. [16] synthesized N-doped hierarchical porous graphitic carbons with large surface area and pore size, which showed good working capacity and separation performance. This performance is ten times larger than most previously reported mesoporous carbon adsorption materials. Psarras and coworkers [17] evaluated the uptake of CO₂ and N₂ in N-doped carbon with slit pore, and discovered that N atoms doping have a positive effect on CO₂ loading and selectivity of CO₂ over N₂. Apart from N doping, the slit pore frameworks also play a decisive character on CO₂ capture and separation performance. Kumar et al. [18] adopted molecular simulations to investigate the carbon capture behavior in N doped graphitic carbons and uncovered that the structure with slit pores is superior to disordered ones on CO₂ adsorption. In spite of these findings, the influence of pore sizes and N doping on the CO₂ capture behavior in graphite slit pore remains unelucidated.

In this work, we shed lights on how the layer distance and N doping affect CO₂ capture and CO₂/N₂ separation selectivity in multilayer graphene sheets (N-MGNs) at room temperature. First, by density functional theory (DFT) simulations, unit cell structure optimization and atomic partial charge were calculated by which was the prior preparation for CO₂ capture and separation. Then, the physical structure of the N-MGNs were explored. Afterwards, the single-component adsorption of CO₂ or N₂ is calculated under different pressures. Next, essential mechanism of N-doping was explored by the radial distribution functions, isosteric heats, and gas distribution. In the end, the separation performance of CO₂ from mixture gases (CO₂:N₂ = 1:1) was calculated to investigate the influence of different compositions. Our results highlighted the positive effect of N doping on graphene sheets, and provide the high-performance adsorbent for CCS technology.

2. Model and Computing Methodology

The supercells of different N-doped multilayer graphene nanostructures (N-MGNs) were used in this work. As depicted in Figure 1, N-doped graphene was built with two different N atoms in the form of triazine and pyrazine as doping sites. This structure can be synthesized via a well-developed ionothermal technique. The doping model and synthetic method was chosen referring to the latest work of Zhu et al. [19]. After building the N-doped graphene, we adopted DFT simulations to optimize the single-layer N-MGN structure and calculated the atomic partial charge of N-MGNs. B3LYP/6-31+g(d,p) basis set in Gaussian 09 [20] package was applied, this choice is a combination of efficiency and accuracy of the calculations. The atomic partial charges of N-graphenes were the critical parameters in coulomb law to describe the electrostatic interaction between gas and frameworks. The layer distance (H), described as the length of the center of the upper and lower N-graphene plane was introduced to describe the morphology of the N-MGNs. Finally, five-layered N-MGNs, with the H varying from 3.4 to 17.0 nm in 3.4 Å increments, were designed to study the CO₂ adsorption capacity and separation properties. These structures were named as N-MGN_3.4, N-MGN_6.8, MGN_10.2, MGN_13.6 and MGN_17.0 according to H. For the N-MGNs frameworks, the universal force field (UFF) parameters were used to describe the non-coulomb interaction [21], which was successfully applied to carbon materials, zeolites and other similar systems. For CO₂ and N₂ molecules, three rigid linear models were used. The parameters of the molecules were adopted by TraPPE model [22]. Information of force field parameters and atomic partial charge of N-MGNs frameworks were listed in

Table 1. Lennard-Jones parameters and atomic partial charges for CO₂, N₂, and N-MGNs.

Atom	Gas Molecule Models				MGNs Frameworks
	C(CO2)	O(CO2)	N(N2)	COM(N2)	C	N
σ (Å)	2.08	3.05	3.31	0.00	3.43	3.26
ε (K)	27.00	79.00	36.00	0.00	52.84	34.72
q (e)	0.748	−0.374	−0.482	0.964	-	-

.

GCMC calculations were utilized to calculate the adsorptions together with the selectivity of CO₂ over N₂ [23]. This method is derived from the physical interaction between gases and frameworks. In GCMC simulations, a series of configurations would be generated randomly, including rotation, movement, and deletion [24]. In case of every state point, 1 × 10⁷ steps were completed during GCMC calculation. An object–oriented multipurpose simulation code (MuSiC) was employed to complete CO₂ capture and separation [25].

3. Results and Discussions

3.1. Pore Topology and Morphology

Physical structure characteristics of pore are the major factors to influence capture and separation capacity of the adsorbent. To explore the physical properties of five N-MGNs, porosity (Φ) is calculated by V_P/V_Total, where V_P refers to pore size and V_Total is total volume of the N-MGNs. The parameters of the gas-accessible frameworks in

Table 2. Pore physical characteristics of N-MGNs, gas probe molecule = He.

R	H_3.4	H_6.8	H_10.2	H_13.6	H_17.0
VP (cm3/g)	0.04	0.461	0.882	1.307	1.727
Porosity, Φ (%)	9.43	54.72	72.71	77.49	84.86
DL (Å)	0.95	3.90	6.94	10.71	13.68
DM (Å)	0.97	3.96	7.11	10.80	13.76
Density g/cm3	2.372	1.186	0.791	0.593	0.474
Accessible surface area (m2/g)	2521.72	2521.72	2521.72	2521.72	2521.72

were obtained by Sarkisov [26] and Duren [27] methods. In case of slit pore, the limiting diameter and maximum diameter are nearly the same for each N-MGN, and the values are close to the corresponding layer distance. Accessible surface area of the five N-MGNs are exactly the same (2521.72 m² g⁻¹), because only the upper and lower surface of the five N-graphene sheets are exposed for all five N-MGNs. The accessible surface area is larger than many porous structures, such as zeolitic imidazolate frameworks (620–1730 m² g⁻¹) [28], some MOFs (282–515 m² g⁻¹) [29], some coordination polymers (503–609 m² g⁻¹) [30], and nanoporous organic frameworks (~1535 m² g⁻¹) [31], [32], which could lead to superior adsorption capacity and selectivity. Different from the surface area, the porosity differs from each other for all N-MGNs. With the increase of layer distance, the porosity increased from 9.43% to 84.86%. N-MGN_3.4 has ultra-small porosity of 9.43%, which lead to no adsorption capacity. N-MGN_6.8 have moderate porosity of 54.72% which could result in better performance at low pressure. And the rest N-MGNs have higher porosity than 70%, which could achieve excellent performance at high pressure. Therefore, the five N-MGNs have great potentials to achieve excel CO₂ adsorption capacity.

The geometric pore size distributions (PSDs) was shown in Figure 2 to investigate the pore morphology on step ahead. These data were obtained by Poreblazer v3.0 [33]. Helium was chose as the gas probe molecule. The results indicated that the PSDs peaks for the materials are narrow and independent due to the uniformly distributed pore sizes of slit pore. We found that the sum of the value of PSD peak and kinetic diameter of Helium (0.26 nm) is approximately equal to the corresponding layer distance. According to the IUPAC classification, N-NGM_3.6 and N-NGM_6.8 have ultramicropores (<7.00 Å), the other N-NGMs have micropores (7.00~20.00 Å) [34]. In summary, the pore spaces and diameters of N-MGNs were enhanced due to the increase of layer distance, particularly for N-NGMs NGM_3.4 and N-NGM_6.8. This can facilitate the gas adsorption and separation in gas-framework systems.

Figure 3 shows the density of state (DOS) for N-NGMs, which analyzes orbital compositions of N-NGM. The result shows that N-NGM presents the typical metallic properties without band gaps. The total DOS of N-NGM consists of the p orbitals of the N and C atoms, and the peaks for these orbitals overlapped evidently, which indicates a strong bonding between the carbon and nitrogen in N-NGM.

3.2. Single Component Adsorption of CO₂/N₂

One of the criterion factors to estimate adsorption capacity of the adsorbents is the absolute adsorption of gases, which were calculated by GCMC simulations in this work. Figure 4a,b shows the single component adsorption isotherms of CO₂ for N-MGNs in the range of 0–7.0 MPa, at 298 K. For N-MGN_3.4, the absolute adsorption capacity of CO₂ is 0, which suggested that no CO₂ could be adsorbed, because of the steric effect presented by the narrow interlayer spacing. At ultra-low pressure (p < 0.04 MPa), the adsorption capacity of CO₂ in the N-MGNs follows the sequence of 6.8 > 10.2 > 13.6 > 17.0 > 3.4. The phenomenon can be explained that smaller pore sizes could lead to stronger CO₂-frameworks interactions. Interestingly, the CO₂ uptake of N-MGNs_6.8 can rapidly reach to 10.59 mmol g⁻¹ at 0.04 MPa and 298 K, which is around 91.2% of the saturated capacity. The saturated pressure of N-MGNs_6.8 is smaller than that of SIFSIX-3-Cu [35], and larger than that of nanoporous carbons [36], which reflects a strong interaction between frameworks and gas molecules. At the pressure of 0.04–0.1 MPa, the CO₂ adsorption is in an order of 10.2 > 6.8 > 13.6 > 17.0 > 3.4. This phenomenon is mainly because of the different pore sizes that cause the adsorption capacity of N-MGN_10.2 exceeding the saturation adsorption N-MGN_6.8 at low pressure. With continued growth of the pressure, the sequence of adsorption capacity for N-MGNs has changed. Under 0.1 MPa, the single component capacity of CO₂ in the N-MGN_6.8 is 10.70 mmol g⁻¹, which is greater than phosphorene slit pores (~1.5–5 mmol g⁻¹) [37], kgm−1–F (1.13 mmol/g) [38], C₃N pore (6.27 mmol/g) [39], S-graphite split pore (36.48–51.00 mmol/mol) [40], graphene with different embedded surface functional groups (~1.0–9.0 mmol/g) [41], and carbon slit pores (~5 and 0.8 mmol g⁻¹) [41] at the similar conditions. When the pressure is over 4 MPa, CO₂ adsorption of N-MGNs follows the sequence of 17.0 > 13.6 > 10.2 > 6.8 > 3.4 in Figure 4b, which is well match with the order of pore size. The situation clarified the fact that the adsorption capacity correlated with interaction strength at low pressure, while it correlated with free volumes at high pressure [42]. To evaluate the structural advantage, the N-MGNs were compared with N-doping materials with slit pore in Kumar’s work [18]. The result shows that, at 0.1 MPa and 298 K, the absolute adsorption capacities of CO₂ in the N-MGNs_6.8 (11.87 mmol g⁻¹) is much larger than that of N-doping slit pore with pore width of 0.8 nm (4.6 mmol g⁻¹). Moreover, at 7.0 MPa, the single component adsorption capacities of CO₂ in N-MGN_17.0 (40.16 mmol g⁻¹) is much greater than N-doping slit pore with pore width of 2.0 nm (~16 mmol g⁻¹). This result confirms the adsorption advantage of our frameworks at both low and high pressure.

Figure 4c,d delineate the single component adsorption isotherms of N₂ in the N-MGNs with different layer distance at 298 K. Moreover, the N-MGN_3.4 has no adsorption capacity in the pressure range of 0–7.0 MPa, because N₂ molecules can’t get into the pore with relatively small pore size which is similar to CO₂ adsorption capacity. For other N-MGNs, the single component adsorption of N₂ follows the order of 6.8 > 10.2 > 13.6 > 17.0 at low pressure (0–1 MPa). This is because that the smaller pore size leads to the stronger interaction, which plays a critical role at low pressure. However, the adsorption changes upon the increase of pressure and it follow the consequence of 17.0 > 13.6 > 10.2 > 6.8 > 3.4 at 7 MPa. the type-I Langmuir adsorption behavior of CO₂/N₂ in N-MGNs demonstrate that the pore size of N-MGNs are typical micropores [43]. The result is consistent with the above-mentioned topology and morphology analysis of N-MGNs. Comparing with other materials, the N-MGNs could achieve preferable adsorption capacity of CO₂ at declined pressure, indicating the ability of N-MGNs to separate CO₂ from N₂.

3.3. Selectivity of CO₂ over N₂ with Equal Molar Fraction

The selectivity of CO₂ over impurity gases is of equal significance to evaluate adsorbent materials in CCS technology compared with the ability of CO₂ adsorption. Therefore, the selectivities of CO₂ over N₂ for all the N-NMGs were studied and defined as [36]:

$$S_{{\mathit{CO}}_{\mathit{2}}\mathit{/}{\mathit{N}}_{\mathit{2}}}{\mathit{=}\mathit{(}x}_{{\mathit{CO}}_{\mathit{2}}}\mathit{/}{x}_{N_{\mathit{2}}}{\mathit{)}\mathit{/}\mathit{(}y}_{{\mathit{CO}}_{\mathit{2}}}\mathit{/}{y}_{N_{\mathit{2}}}\mathit{)}$$

(1)

where S is the selectivity of CO₂ over N₂; $x_{{\mathit{CO}}_{\mathit{2}}}$ and $x_{N_{\mathit{2}}}$ are the molar fractions of CO₂ and N₂ in their adsorbed phase; and $y_{{\mathit{CO}}_{\mathit{2}}}$ and $y_{N_{\mathit{2}}}$ are the molar fractions. Figure 5 shows the selectivity of CO₂ over N₂ with 1:1 gas ratio for the N-MGNs at 298 K. Due to the non-adsorptive behavior for both CO₂ and N₂, N-MGN_3.4 is excluded in the selectivity study. In the above-mentioned range of pressures, the selectivities follow order of 6.8 > 10.2 > 13.6 > 17.0. The fact that increasing layer distance can reduce the $S_{{\mathit{CO}}_{\mathit{2}}\mathit{/}{N}_{\mathit{2}}}$, indicates that the layer distance has a significant influence on $S_{{\mathit{CO}}_{\mathit{2}}\mathit{/}{N}_{\mathit{2}}}$, especially at low pressure. The higher $S_{{\mathit{CO}}_{\mathit{2}}\mathit{/}{N}_{\mathit{2}}}$ of N-MGN_6.8 is ascribe that the accessible N atoms and suitable channel are the predominant factors for the remarkable selectivity of CO₂ over N₂. At equilibrium pressure, N-MGN_6.8 exhibits the highest selectivity (~133), which is superior to many N-doped mesoporous (~32) [44], pristine GY (~50) [45] and hierarchical (~21–67) [17] carbon adsorbents. Unlike isosteric heats of the N-MGNs, the sequence of the selectivity remains the same in case of all pressures. This finding further highlights that the N-MGNs show optimal $S_{{\mathit{CO}}_{\mathit{2}}\mathit{/}{N}_{\mathit{2}}}$ in the mixture gas, especially for the one with a layer distance of 0.68 nm.

3.4. Mechanism of CO₂/N₂ Adsorption in N-MGNs

Isosteric heats (Q_st). Q_st [46] is a benchmark to describe the interaction of adsorbing gas molecules and framework at a certain surface, which represents the affinity between gas molecules and frameworks. Q_st can be calculated from the Clausius−Clapeyron equation [47]. The Q_st of N-MGN_3.4 was not considered due to no adsorption of the gases. Figure 6a exhibits the relationship between Q_st and adsorption capacity of CO₂ at 298 K. At low pressure, the Q_st of CO₂ follows the order of 6.8 > 10.2 > 13.6 > 17.0, which is the same as the sequence of CO₂ adsorption. On the other hand, the isosteric heat profiles of N-MGNs_13.6 and N-MGNs_17.0 have three intersection points when the CO₂ uptake is lower than 12 mmol/g. This is because large pore sizes of N-MGNs_13.6 and N-MGNs_17.0 lead to weaker interactions at low pressure, thus resulting in little difference in the isosteric heat profiles. Therefore, the absolute adsorption isotherms of CO₂ for N-MGNs _1.36 and N-MGNs_1.70 nm almost overlapped at low pressure. The Q_st of CO₂ in N-MGNs_6.8 nm (~55 kJ mol⁻¹), is larger than edge-functionalized nanoporous carbons (24.63–34.25 kJ mol⁻¹), copper metal organic framework (~20 kJ mol⁻¹) [48], and SIFSIX-3-M series (~45–50 kJ mol⁻¹) [35]. It manifests the stronger interaction between CO₂ and N-MGNs, which leads to the larger adsorption capacity at low pressure than other adsorbent materials. For the N₂, the order of Q_st corresponds to the absolute adsorption capacity N₂, namely, 6.8 > 10.2 > 13.6 > 17.0 at low pressure (Figure 6b). The low Q_st of N₂ in N-MGNs_6.8 (~30 kJ mol⁻¹) indicates a typical weak physisorption of N₂. On the whole, the adsorption capacity of both CO₂ and N₂ are strongly related to the gas-framework interaction. And the stronger quadrupole moment of CO₂ indicates that CO₂ has the stronger interactions with N-MGNs than N₂.

Gas distribution. To explore the gas distribution in equilibrium states for the pore of N-MGNs, the relative concentration profiles were presented as a function of the distance between layers. The relative concentration is defined as a ratio, that is calculated by dividing actual concentration by average concentration of the gas in the pore. The distribution of both CO₂ and N₂ in each layer of the N-MGNs can be reflected by relative concentration profiles. In most cases, it shows strong primary peaks close, minor secondary peaks, and bulk region [49]. T14he primary peak is resulted from the interaction between the first-layer of gas molecules adsorbed on the pore walls, that is, adsorbed phase. The secondary peak is mainly due to the equilibrating interactions between gas–gas or gas–surfaces, that is, equilibrated phase.

The relative concentration profiles of CO₂ in each layer of N-MGNs are shown in Figure 7a,d. For N-MGNs_6.8, an isolated and high primary peak appear in each layer, because the layer with narrow distance can only hold CO₂ molecules in the adsorbed phase. Moreover, the narrower layer distance leads to the overlap of two primary peaks for each pore. Differently, the two primary peaks were separated and next to each other for N-MGN_10.2. And in case of N-MGN_13.6, the equilibrated phase can be observed in between the two primary peaks. As for N-MGN_17.0, two primary peaks, two secondary peaks, and bulk region can be clearly visualized due to the large the enough space of each pore. The results explain the phenomenon that the N-MGN with larger layer distance have higher adsorption capacity at high pressure.

Figure 7e–h show that the relative concentration profiles of N₂ for the N-MGNs have similar trend as CO₂. When layer distance is narrow, only overlapped primary peaks can be seen. With the increase of layer distance, primary peaks separate and overlapped secondary peaks appear. When the layer distance reaches 1.70 nm, the integrated primary peaks, secondary peaks, and bulk region can be observed. However, the primary and secondary peaks of N₂ are inferior comparing to CO₂ due to the large kinetic diameter [50] and the stronger electric quadrupole moment of CO₂ molecule.

Overall, the relative concentration profiles visualized the distributions of gas molecules in each layer of N-MGNs. The results indicate that the layer distance is a significant factor to influence the adsorption capacity of CO₂/N₂ at high pressure, namely, the larger layer distance causes higher adsorption. This perfectly explains the change of the sequence of absolute adsorption isotherms at high pressure.

4. Conclusions

The adsorption and separation behaviors of CO₂ and N₂ for N-MGNs with different layer distances were studied by simulations. Results have shown that the N-MGNs structure provides favorable pore space for CO₂ adsorption capacity and separation performance. At ultra-low pressure, N-MGN_6.8 have shown an excellent CO₂ adsorption behavior (10.59 mmol/g) at low pressure (0.4 bar). In comparison, the absolute adsorption of CO₂ for N-MGN_17.0 (40.16 mmol/g) at 7.0 MPa is much larger than that of N-doping slit pore under high pressures. At 1.0 bar, the highest $S_{{\mathit{CO}}_{\mathit{2}}\mathit{/}{N}_{\mathit{2}}}$ reached up to ~133 in N-MGN_6.8. The adsorption heat and relative concentration curves confirm that the adsorption ability of CO₂ for the all the materials are better than that of N₂, and the change of the interlayer spacing has an important influence on the adsorption capacity of the gases. The effect has the advantage of adsorption. Under high pressure, the structure with large interlayer spacing has a strong adsorption capacity due to the large porosity. The five structures all showed strong CO₂/N₂ adsorption capacity at 0–7 MPa. This work highlights the advantages of CO₂/N₂ adsorption and separation in N-MGNs structure, as well as providing effective methods and strategies for the design and screening of high-quality adsorbate materials for carbon capture and storage technology.