The CMs-N/P-n composite materials were synthesized utilizing ammonium fluoride (NH₄F), rice hull powder, and melamine polyphosphate in a 4:1:1 weight ratio. NH₄F functioned as a pore-forming agent, while rice hull powder served as the carbon source and melamine polyphosphate provided N and P components. Initially, the mixed precursors were subjected to calcination under an argon atmosphere. The thermal treatment commenced with a heating rate of 2°C/min until reaching 400°C, at which point the materials were maintained for 30 min. Subsequently, the temperature was elevated to predetermined targets at a ramp rate adjusted to 5°C/min, followed by a 2 h isothermal treatment. The resulting carbonized products were then chemically activated through immersion in 0.1 M KOH under a vigorous agitation for 2 h. Final processing involved extensive washing with deionized water until neutral pH was achieved, followed by drying at 100°C for 24 h. The nomenclature CMs-N/P-n reflects the synthesis method, where n denotes the final annealing temperature (650, 750, 850, 950 or 1050). The weights of CMs-N/P-650, CMs-N/P-750, CMs-N/P-850, CMs-N/P-950 and CMs-N/P-1050 are about 0.331 g, 0.309 g, 0.282 g, 0.246 g and 0.202 g. In addition, the average particle size of them are about 34.8 μm, 31.2 μm, 25.7 μm, 29.6 μm and 33.4 μm, respectively.

The average particle size of CMs-N/P-n was obtained using a laser particle size analyzer (Malvern Mastersizer 2000, Britain). The phase composition was characterized by X-ray diffraction (XRD, Panalytical X’Pert3 Powder, Netherlands, 10 ~ 80°). Bonding configurations were analyzed through Infrared spectroscopy (IR, Thermo Fisher Scientific Nicolet iS20, America, 400–4000 cm⁻¹), Raman spectroscopy (RS, WITec alpha300R, Germany) and X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha, America). Porosity parameters were quantified via N₂ adsorption-desorption at −196°C (Tristar II 3020, America). Morphological and microstructural features were resolved by scanning electron microscope (SEM, ZEISS Sigma 300, Germany) and transmission electron microscope (TEM, JEM-F200, Japan), respectively, with elemental composition determined via energy dispersive spectroscopy (EDS, JEM-F200, Japan) and cross-validated by XPS. Notably, XPS served dual roles in chemical state deconvolution and elemental quantification.

In this investigation, CO₂ adsorption measurements were done over a penetration test apparatus which was verified by lots of previous reports [2,18,19]. First, about 10 mm of CMs-N/P-n powers (~ 0.2 g) were filled into the sample chamber (~ 30 mL). The impurities were cleaned out by placing the adsorbents to pure argon (60 mL/min). Then, CO₂ adsorption was completed by changing Ar to be 20% CO₂. Finally, CMs-N/P-n materials regeneration was achieved through purging them in Ar flow (60 mL/min) at 100°C. CO₂ adsorption capacities are achieved based on Eg. (1), where N stands for CO₂ uptake (mmol/g), F represents gas flow rate (mL/min), t and t_e are integral time and actual time (s), M is the mass of CMs-N/P-n materials, c₀ and c_e are CO₂ concentrations at the inlet and outlet of adsorption device.

The XRD patterns presented in Fig. 1(a) demonstrate the structural characteristics of CMs-N/P-n. Broad diffraction peaks centered at 25.5° are observed, which correspond to the (002) crystallographic planes of carbon materials. These broadened peaks are indicative of amorphous structural features in the synthesized adsorbents. Consistent with previous studies [3], weak diffraction signals at 45.1° are additionally detected and attributed to partial graphitization processes and preferential orientation growth along the (100) planes. No additional peaks appears on the diffraction patterns, indicating that all of samples are carbon materials without impurities. Notably, enhanced diffraction intensities of the (100) plane are observed in CMs-N/P-850, CMs-N/P-950 and CMs-N/P-1050 compared with CMs-N/P-650 and CMs-N/P-750, demonstrating improved graphitic ordering at elevated synthesis temperatures [20]. The XRD analysis reveals that the majority of carbon atoms in CMs-N/P-n exist in long-range disordered configurations. However, localized graphitic domains are confirmed to be present, characterized by well-ordered atomic arrangements in specific crystalline regions.

The IR characterization of CMs-N/P-n presented in Fig. 1(b) reveals distinct structural features through vibrational mode analysis. Two broad absorption bands are observed at 3432 cm⁻¹ and 1113 cm⁻¹, accompanied by six absorption peaks at 2927 cm⁻¹, 2856 cm⁻¹, 1624 cm⁻¹, 1446 cm⁻¹, 732 cm⁻¹, 603 cm⁻¹ and along with a weak shoulder spanning 811 ~ 935 cm⁻¹. The broad vibrational feature at 3432 cm⁻¹ is attributed to overlapping N-H and O-H stretching modes [3]. The absorption band at 1113 cm⁻¹ is associated with the vibrations of P=O, aromatic P-C linkages and P-O-C structural units [21,22]. The weak peaks at 2927 cm⁻¹ and 2851 cm⁻¹ is identified as symmetric and asymmetric C-H stretching vibrations [3]. The prominent absorption features at 1624 cm⁻¹ and 1446 cm⁻¹ are correlated with conjugated C=C/C=O stretching modes and covalent C-N bonding configurations [1,3]. The characteristic peaks at 732 cm⁻¹ and 603 cm⁻¹ are assigned to P-C and P=O vibrational modes in the P-containing framework [21]. Additionally, the faint shoulder observed between 811 ~ 935 cm⁻¹ is interpreted as the evidence of P-OH bond formation [21]. The coexistence of N-H, C-N, P-C, P-O-C, and P-OH functional groups is confirmed through IR analysis, demonstrating that N and P heteroatoms derived from melamine polyphosphate are integral components of the carbon matrix. This comprehensive spectral evidence establishes the successful incorporation of both N and P species into the carbonaceous framework through high temperature annealing.

The Raman features of CMs-N/P-n, as illustrated in Fig. 1(c) ~ Fig. 1(h), are systematically analyzed through peak deconvolution. The spectral profiles are deconvoluted into four characteristic components. The prominent peaks observed between 1200–1610 cm⁻¹ (D band) and 1390–1800 cm⁻¹ (G band) are associated with disordered carbon structures (sp³ hybridization) and graphitic domains (sp² hybridization), respectively (Fig. 1(c) ~ Fig. 1(g)) [1,23]. The D band intensity is correlated with the symmetrical decomposition of graphitic C species, while the G band corresponds to in-plane vibrational motions of sp²-hybridized carbon atoms in graphitic layers [23]. The intensity ratio (I_D/I_G) is conventionally employed as a critical indicator for evaluating structural ordering in graphitic systems [23]. Calculated values of 1.13, 1.09, 1.00, 0.99 and 0.97 are obtained for CMs-N/P-650, CMs-N/P-750, CMs-N/P-850, CMs-N/P-950 and CMs-N/P-1050, respectively (Fig. 1(h)). The progressive decrease in I_D/I_G ratios with increasing pyrolysis temperature demonstrates enhanced graphitization degree and structural ordering in CMs-N/P-850, CMs-N/P-950 and CMs-N/P-1050 compared to CMs-N/P-650 and CMs-N/P-750. This thermal evolution trend is consistent with XRD observations, which is attributed to both of high-temperature activation and heteroatom incorporation [20,23]. The peaks at 1196 cm⁻¹, 1216 cm⁻¹, 1218 cm⁻¹ and 1204 cm⁻¹ are identified as vibrational modes arising from C-C, C=C and their alternating configurations in carbon frameworks [2,23]. Additionally, the low-intensity peaks observed at 1493 cm⁻¹, 1485 cm⁻¹, and 1484 cm⁻¹ are attributed to symmetry-breaking perturbations in aromatic rings induced by the doping of N and P species [23]. These spectral modifications provide direct evidence for structural distortions caused by heteroatom integration into the carbon lattice.

The XPS survey of CMs-N/P-n is presented in Fig. 1(i) ~ Fig. 1(n). In the full spectrum (Fig. 1(i)), C, O, N, P and O elements are detected. Quantitative elemental compositions are summarized in Fig. 1(j), where C is identified as the predominant constituent in all samples. A progressive increase in C content is observed with elevated annealing temperatures, correlated with enhanced carbonization degree during thermal treatment. Concurrently, reductions in O, N and P concentrations are detected, which could be explained by the thermal decomposition of melamine polyphosphate precursors and elimination of volatile impurities. Deconvolution of the C 1s core-level spectrum (Fig. 1(k)) reveals four distinct components at 284.7 eV (C-C/C=C), 286.3 eV (C-O/C-N/C-P), 288.6 eV (C=O/C=N), and 292.6 eV (π→π* transitions) [21,24,25]. The characteristic π→π* satellite peak at 292.6 eV provides direct evidence for sp²-hybridized carbon domains within the material [24,25]. N speciation analysis demonstrates four chemical states (Fig. 1(l)): pyridinic N (398.6 eV) and pyrrolic N (400.6 eV) configurations where N atoms adopt sp² hybridization within hexagonal and pentagonal C rings, respectively, along with graphitized N (401.6 eV) and oxidized N (403.4 eV) species [26]. The P 2p spectrum (Fig. 4(m)) is resolved into two components corresponding to P-C (132.9 eV) and P-O/P=O (134.1 eV) bonds [21]. The O 1s spectral profile is deconvoluted into four contributions (Fig. 4(n)): C=O/P=O (530.8 eV), C-O/P-O-C (532.3 eV), C-OH/P-OH (533.9 eV) and O-H (535.1 eV) [21,26]. These spectral features collectively confirm effective heteroatom incorporation, with N and P species being successfully integrated into the carbon matrix through covalent bonding and functional groups. The comprehensive XPS analysis provides conclusive evidence for the structural modification of carbon frameworks through simultaneous N and P doping.

The SEM and TEM results of CMs-N/P-n are presented in Fig. 2(a) ~ Fig. 2(d). In SEM imaging (Fig. 2(a)), numerous voids are observed, confirming the porous nature of CMs-N/P-n. The TEM analysis (Fig. 2(b) ~ Fig. 2(d)) reveals characteristic diffraction rings along with seven parallel lattice stripes, which are associated with partial graphitization in the carbon framework. These structural features are found to align with XRD and Raman data, indicating a predominantly amorphous carbon matrix with localized graphitized domains.

The N₂ adsorption/desorption characteristics of CMs-N/P-n are systematically analyzed in Fig. 2(e), Fig. 2(f) and Table 1. A sharp increase in adsorption capacity at P/P₀ values of 0 ~ 0.1, combined with pronounced hysteresis loops in the 0.45 ~ 1.0 range, is observed for all samples (Fig. 2(e)). These type II isotherms with distinct hysteresis loops are identified as characteristic of porous materials, where capillary condensation is confirmed to occur in the narrow pore channels. These findings are further supported by aperture distribution analysis displayed in Fig. 2(f) and Table 1. The pore structure is determined to be dominated by micropores and mesopores. For CMs-N/P-850, the pore diameters are within 1.5 ~ 42 nm (Fig. 2(f)) and the ratios of micropores and mesopores are 18.3% and 81.7% (Table 1). The CMs-N/P-850 sample possesses a remarkable specific surface area (S_BET) of 877 m²/g and a pore volume (V_pore) of 0.554 cm³/g (Table 1). The phenomenon is similar to the results of the present work^[26], which is attributed to the formation of carbon skeleton and the collapse of pores. When the annealing temperature is lower than 850°C, more and more rice husk powders are decomposed. In this process, carbon atoms are recombined and form porous frameworks. Hence, the S_BET is obviously enhanced. When the temperature is beyond 850°C, partial pore structure collapses, leading to the blockage and disappearance of pores. Therefore, the S_BET is decreased. This hierarchical micro-mesoporous architecture is demonstrated to facilitate gas molecule accommodation within nanoscale voids. The synergistic effects of high surface area and multimodal pore distribution are proposed to underlie CO₂ adsorption capabilities.

Based on the above information and prior researches^[21], it is proposed that N atoms are bound to the carbon framework through four primary bonding configurations: C-NH, C=N, N⁺, and C-N-O⁻. Concurrently, P species are incorporated into the carbon matrix via P-C, P=O, P-OH, and P-O-C bonds. The synthesis pathway for CMs-N/P-n is schematically illustrated in Scheme 1.

In the present work, all of the capacities are obtained by eliminating the impact of Ar. The adsorption results shown in Fig. 3(a) tell us that CO₂ and Ar exhibit different adsorption behaviors on the CMs-N/P-n adsorbents. For Ar adsorption, Ar penetrates immediately when the gas molecules contact with the adsorbents and then reaches the saturation adsorption quickly. Therefore, the breakthrough and saturation points are zero and 70 s, respectively. However, all of CO₂ molecules are caught within 180 s and then penetrate through the CMs-N/P-850 material, leading to the achievement of large breakthrough time (180 s) and saturation point (400 s). The phenomenon has something to do with different adsorption mechanisms on CMs-N/P-n materials. Ar, a monatomic noble gas, is difficult to be adsorbed on the porous carbon. However, CO₂ is an active gas with quadrupole moment, which is easy to form electron delocalization. Hence, CO₂ could be easily caught through the reactions between gas molecules and N and P defects, leading to the better adsorption feature on CMs-N/P-n.

Compared with CMs-N/P-850, other adsorbents exhibit a poor CO₂ adsorption feature (Fig. 3(b) and Fig. 3(c)). The low values of S_BET and V_pore of CMs-N/P-750 declare that the adsorbent could only provide a limited sites and space for the accommodation and storage of adsorbates [27]. Moreover, the existence of plentiful micropores displays a huge hindrance for the diffusion of adsorbates and the contact with heteroatoms [27]. For CMs-N/P-950, the decrease of V_pore and heteroatoms number is the main reason for the reduced adsorption performance [27,28]. Therefore, both CMs-N/P-750 and CMs-N/P-950 show a low penetration time, saturation time and CO₂ adsorption capacity.

In order to further confirm the adsorption behavior on the CMs-N/P-n materials, situ IR tests on the CMs-N/P-850 sample were performed (Fig. 3(d)). Obviously, the peaks intensity increases significantly with the prolongation of contact time. Moreover, the signals: 1) at 749 cm⁻¹ are connected with P-C bonds [21]; 2) at 1456 cm⁻¹, 1521 cm⁻¹ and 1728 cm⁻¹ are related to C-N, COO- and C=O stretching vibrations in NCOO⁻ skeleton [21]; 3) at 2257 cm⁻¹ and 2426 cm⁻¹ correspond to $N{H}_2^{+}$ species; Moreover, $N{H}_2^{+}$ deformation peaks are also detected at 1640 cm⁻¹ [2]; 4) located at 2342 cm⁻¹ is relevant to the stretching of CO₂ molecules exist in the pore channels of CMs-N/P-n materials. Based on the in situ results and the previous reports [3,21,29,30], CO₂ could be adsorptiond on the surface of CMs-N/P-n materials through zwitterionic or base catalyzed hydration mechanisms (Scheme 2): 1) CO₂ molecules react with N defects to form zwitterionic intermediates and then the products react with other amine groups, resulting in the generation of carbamates; 2) Adsorbates bind to P defects directly through electrostatic attraction and van der Walls forces, leading to the formatin of P-C bonds. Consequently, N and P defects show a co-adsorption phenomenon on CMs-N/P-n materials.

According to Scheme 2, a chemical adsorption process originating from N and P heteroatoms occurs on the CMs-N/P-n materials. Usually, gas chemisorption will be affected by the sweeping of temperature. Therefore, CO₂ adsorption performance of CMs-N/P-850 at different temperatures was also analyzed. In Fig. 3(e), one can see that the breakthrough and the saturation points decline with the rise of system temperature within 25 ~ 100°C. Corresponding to these, CO₂ uptake declines from 3.5 mmol/g to 2.5 mmol/g (Fig. 3(f)). When the temperature is 100°C, CO₂ uptake is approximately 71% of that at 25°C. This phenomenon suggests that temperature is a propelling factor for the decrement of adsorption value. According to Scheme 2, an exothermic equilibrium is set up between CO₂ molecules and N and P defects. Compared with the reaction at low temperature, the equilibrium with a reverse movement is formed at high temperature [31,32]. Therefore, a low adsorption value is obtained at 100°C. CMs-N/P-850 has a cavity space of 0.554 cm³/g. Moreover, the pore diameters are larger than that of CO₂ (0.33 nm) molecules, which is very beneficial for the accommodation of adsorbates. Hence, the holes play a significant part in CO₂ physical adsorption. When the temperature is improved, the exothermic physical adsorption process will weaken gradually [27]. Based on the above information, the total adsorption amount declines with the rise of temperature on account of the decrement of both physical and chemical adsorption capacities.

In Fig. 3(f), adsorption rate enhances with the improvement of temperature. In general, adsorption speed before the saturation point could be expressed by the ratio of CO₂ adsorption value/spent time (N/t_e). The change of temperature will cause the variation of adsorption amount and CO₂ diffusion speed. According to Fig. 3(e) and Fig. 3(f), both of adsorption amount and the contact time between CO₂ and CMs-N/P-n decrease with the enhancement of temperature. However, the change of used time is slightly higher than that of adsorption level because of the significant accelerated diffusion rate and faster reaction speed between basic defects and adsorbates. Therefore, an improvement of adsorption rate is foreseeable.

The influence of gas flow rate on CO₂ adsorption should not be ignored, because it will cause the change of CO₂ concentration in the adsorption system (C), mass transfer efficiency of adsorbates in the pore channels (η) and the contact time between gas molecules and CMs-N/P-n particles (T’) [33]. Moreover, these factors will play a leading role in different stages of CO₂ adsorption. In Fig. 3(g), both of the breakthrough and saturation points declines rapidly as the improvement of CO₂ stream, revealing that the contact time between CMs-N/P-850 and CO₂ molecules decreases dramatically. As seen in Fig. 3(h), the adsorption amount elevates dramatically as the improvement of CO₂ stream from 10 mL/min to 50 mL/min, whereas it declines when the stream exceeds 50 mL/min. Moreover, the optimized CO₂ uptake of 5.2 mmol/g is achieved. The phenomenon is closely correlated with the variation of C, η and T’. According to the previous reports [10,33,34], the adsorption process is decided by C under low flow rate, whereas it is dominated by η and T’ at high CO₂ stream. The enhancement of gas stream promotes the elevation of C, causing a positive shift of adsorption equilibrium (Scheme 2). Therefore, more and more CO₂ are bonded to N and P defects, resulting in the enhancement of CO₂ uptake. In Fig. 3(g), T’ decreases dramatically when CO₂ stream exceeds 50 mL/min, causing the generation of incomplete adsorption. Excepting from this, the decrement of η will also lead to the identical change of adsorption feature. In general, η is characterized by mass transfer length (L) that is displayed in the Eq. (2), where L_e is the length of filled CMs-N/P-n, t_e and t_b are the time corresponding to ^{CCO₂,outlet}/_CCO2,inlet are 0.95 and 0.5, respectively [35]. The larger the L, the higher the η. According to equation (8) and Fig. 3(i), the value of L increases with the variation of CO₂ stream. Moreover, the variation amplitude within 50 ~ 70 mL/min is much higher than that range from 10 mL/min to 50 mL/min. The results suggest that the η levels at 50 ~ 90 mL/min are low in comparison with that within 10 ~ 50 mL/min. Consequently, CO₂ adsorption value decreases with the decrement of η and T’. Based on these, both CO₂ uptake and contact time decrease gradually. Moreover, the former changes less than the latter. Therefore, the adsorption speed defined as the slop of adsorption curve is positively correlated with gas stream (Fig. 3(i)).

The impact of N₂ and O₂ on CO₂ adsorption should also be performed, because they are the major component in air and both of them could be adsorbed in the porous materials. Usually, N_CO2/N_N2 and N_CO2/N_O2 proportions are used to estimate the impact of these impurities [36]. The greater the proportions, the better the selective trapping performance. In Fig. 3(j), great breakthrough and saturation points for CO₂ adsorption are obtained in comparison with that of N₂ and O₂ adsorption. Moreover, CO₂ uptake is greater than the relevant values of N₂ and O₂ (Fig. 3(k)). The vastly different adsorption feature is attributed to the distinct mechanisms. 1) According to Fig. 6, the pore diameters of the CMs-N/P-850 particles are within 1.5 ~ 42 nm. Moreover, the diameter of CO₂ is smaller than the size of N₂ and O₂. Therefore, the CMs-N/P-850 particles are more favourable to the entry of CO₂ molecules, leading to the higher CO₂ storage. 2) Both N₂ and O₂ are inactive gases that could not react with N and P defects. Hence, N₂ and O₂ could only be caught in the pores of CMs-N/P-850 through physical adsorption method. CO₂, an acidic gas with a quadrupole moment, could be easily attracted on account of the great polarity created by N and P sites [36]. As seen in Scheme 2, CO₂ molecules react with N and P heteroatoms, leading to CO₂ chemical adsorption. Different from N₂ and O₂, CO₂ can be caught by CMs-N/P-n materials through both physical and chemical adsorption. Therefore, high N_CO2/N_N2 and N_CO2/N_O2 values are achieved.

An excellent adsorbent should exhibit both good CO₂ adsorption feature at different conditions and cyclical adsorption feature. In this investigation, the CMs-N/P-850 material was used to accomplish the cyclical adsorption experiments at 25°C and 75°C. In Fig. 3(l), the fluctuations of CO₂ adsorption amount are less than 1.0% and 1.5%, suggesting that CO₂ uptake keeps stable at 25°C and 75°C. Consequently, the CMs-N/P-n materials possess a superior cyclical adsorption feature.

In order to confirm CO₂ adsorption behavior on the CMs-N/P-n materials, four models displayed in Eqs. (3) ~ (6) were used for the simulation of adsorption results, where Q_T and Q_e represent real-time adsorption value and maximum capacity, T is adsorption time, Q_{e, exp} is the maximum uptake in the experiments, k₁ and k₂ represent capture speed constants, α_E and β_E are speed constants for initial adsorption/desorption, k and z are dynamic constants, R² is correlation factor [27,31,37].

In Fig. 4(a), the correlation coefficient of R² for the spent formulas are 0.9777, 0.9705, 0.9114 and 0.9984, respectively. The information suggests that: 1) CO₂ adsorption process is not decided by single physisorption or chemisorption because of low R² values for Quasi 1st-order and Quasi 2nd-order kinetic formulas [27,31]; 2) CO₂ adsorption behavior could be characterized using Bangham model on account of the high R² value (0.9984). Therefore, the adsorption process including two steps: CO₂ diffusion and CO₂ adsorption. CMs-N/P-n samples are porous materials. Moreover, the pore diameters are larger enough to accommodate CO₂ molecules. The presence of N and P defects promotes the proximity of acidic molecules to the adsorbents. When adsorbates reach the outer surface of CMs-N/P-n particles, they could pass through the gas film and then diffuse rapidly. In this step, CO₂ is caught via Van der Waals force [37]. Then, some of adsorbates react with the heteroatoms, leading to the chemisorption and the formation of CO₂-N or CO₂-P species [37]. Therefore, CO₂ uptake is dramatically enhanced due to the simultaneous physical and chemisorption behaviors. After the rapid growth period, most of spaces are occupied and the majority of N and P defects are consumed. Consequently, no significant enhancement is observed.

To confirm the universality, the adsorption results achieved under different environments were also simulated with Bangham model. In Fig. 4(b), Fig. 4(c) and Table 2, the fitting curves have a high matching degree with experimental results and the R² values are higher than 99.6%, providing further evidence for the co-action of physisorption and chemisorption at different temperatures and gas streams.

In this study, isosteric heat (Q_st, KJ/mol) was evaluated for the analysis of the strength between CO₂—CMs-N/P-n, because it plays a great role in the reflection of energy distribution of porous carbon functionalized with different heteroatoms. According to the relevant investigations [37,38], isosteric heat on porous CMs-N/P-n materials could be calculated following the Clausius-Clapeyron equation expressed in Eq. (7), where N is CO₂ uptake (mmol/g), R, P and T are the ideal gas constant (J/(mmol·K)), system pressure (kPa), Kelvin temperature, respectively. In Fig. 4(d) ~ Fig. 4(h), the slopes based on lnP and 1/T could be used to afford the calculation of Q_st within 0 ~ 3.2 mmol/g. For CMs-N/P-850, the heat declines gradually as the function of CO₂ uptake. Moreover, the slop before 1.6 mmol/g is higher apparently than that beyond 1.6 mmol/g (Fig. 4(i)). The phenomenon is closely associated with the porous characteristics and N and P multiple defects. CMs-N/P-850 is a porous material with 18.3% of micropores, 81.7% mesopores, 0.554 cm³/g of void spaces, 6.2% of N and 2.6% of P species. Both pore volume and diameters are high enough for the storage of CO₂. Based on these information, CO₂ could easily diffuse in the pore channels and contact with N and P defects. Therefore, Q_st must be analyzed based upon both CO₂—CMs-N/P-850 and CO₂—CO₂ interactions [39]. In the initial stage, CO₂ diffuses easily and react with N and P defects. As the adsorption progresses, most of holes and basic sites are consumed. Hence, the adsorption amount increases rapidly and the Q_st value decreases dramatically. The Q_st value at the beginning is high (33.7 kJ/mol), which is associated with the bi-functionalization of N and P heteroatoms. When CO₂ uptake is larger than 1.6 mmol/g, most of N and P defects are consumed and the chemisorption behavior is gradually weakened. Hence, adsorbates could only interact with weak affinity sites, causing the significant decrement of Q_st value. With the further enhancement of adsorption value, CO₂—CO₂ interaction force improves gradually and it becomes the dominant factor [40]. However, its growth is not enough to offset the decrement of CO₂—CMs-N/P-850 force. Thus, a slow decrement appears at the adsorption amount arrange from 1.6 to 3.2 mmol/g.

Similar to CMs-N/P-850, other adsorbents decrease gradually with the enhancement of CO₂ adsorption value, which are also related to the change of CO₂ adsorption amount at different adsorption stages. Compared with CMs-N/P-850, other samples have a low Q_st value. CMs-N/P-650 and CMs-N/P-750 have a low void space and a large percentage of micropores, which has a great negative effect on the accommodation, diffusion and chemisorption of CO₂. CMs-N/P-950 and CMs-N/P-1050 have a low content of N and P defects, causing the weak chemisorption in the pore channels. Therefore, the Q_st level of them are less than that of CMs-N/P-850.