Synthesis of grape-seed derived carbon with high specific surface area for CO2 selective adsorption

Nitrogen-doped porous carbons with BET surface area of 1068.2–3314.5 m2/g and nitrogen contents of 3.2–6.5% were prepared with solid waste grape-seed as raw material, NaNH2 as activator and nitrogen source at low activation temperature. Super The activation mechanism of NaNH2 on hydrothermal carbon precursors was first explored by thermodynamic analysis and TG-IR, which provided theoretical support for pore forming of carbon materials. Maximum CO2 and CH4 adsorption capacity at 273 K and pressure of 101 kPa was 5.42 and 1.76 mmol g−1 respectively, which are higher than those of majority of carbon derived from most solid wastes reported in literature. IAST selectivities of GS-3-450 with the largest BET surface area for CO2/CH4 (40v/60v), CO2/N2 (15v/85v), CH4/N2 (50v/50v) were found to be 20.3, 71.4, 6.0 under 101 kPa and 298 K respectively. The competitive adsorption of GS-3-450 for CO2/CH4 (40v/60v), CO2/N2 (15v/85v), CH4/N2 (50v/50v) gases mixture were examined through breakthrough experiments, and the results showed that the breakthrough time of CO2 was longer than that of CH4 and N2, which was beneficial to the separation of CO2 from gases mixture. Eight cycles of CH4 adsorption–desorption studies revealed that the material exhibited excellent recycling stability. Low temperature preparation method, excellent BET specific surface area and total pore volume, as well as excellent adsorption ability of CO2 and CH4 make it have a very great potential for the capture of CO2 and CH4.


Introduction
Carbon materials possess well-developed pore structures and high specific surface area, excellent mechanical and thermal stability, and resistance to acid and alkali corrosion [1][2][3][4][5]. These significant characteristics make them as very potential adsorption materials widely used in the storage of hydrogen, separation and purification of gas mixtures, especially in the separation of greenhouse gases, such as CO 2 and CH 4 from coal-based stationary power plants (CO 2 as the major component) and bio-gases (CH 4 as the major component) [5][6][7]. In addition, it also has been used as electrode material for supercapacitor [8].
At present, the synthesizing carbon materials mainly focuses on physical activation and chemical activation methods [9][10][11]. For physical activation methods, CO 2 , water vapor, air, etc. are usually used as activators, but the specific surface area and pore volume of prepared carbon materials are not outstanding, especially the activation temperature is relatively higher. Tang and Li applied CO 2 to activate rice to obtain granular carbon adsorbents at a temperature of 800-900 °C, but their specific surface area was less than 1000 m 2 /g [12]. Similar results were also observed in Guo and Zhao who utilized air and CO 2 as activators respectively to activate waste sugarcane to synthesize carbons at a temperature of 850 °C [10]. By comparison, chemical activation methods, have been widely used to produce porous carbon materials due to its high specific surface area and relatively low activation temperature, where KOH, H 3 PO 4 , ZnCl 2 , etc. are frequently used as an activators, especially for KOH. For example, Zhang and Deng prepared ultramicroporous carbons CGUCs with uniform and narrow pore size distributions using KOH as activator and cactus and glucose as carbon sources. The optimal carbon material CGUC-1-8 showed the largest CO 2 (6.26 mmol g −1 ) and CH 4 (1.94 mmol g −1 ) adsorption amounts at 273 K and 1.0 bar, and its surface area and micropore volume were 1070.6 m 2 /g and 0.369 cm 3 /g, respectively [13]. Song and Fan applied KOH as an activator to activate anthers and the optimal carbon PA-500-KOH-4-800 possessed an excellent specific surface area of 3322 m 2 /g and the adsorption capacity of CO 2 was as high as 51.3 mmol g −1 at 50 bar and 25 °C [14]. Liu and Pan synthesized glucose-derived microporous carbon nanospheres via KOH activation, and its CH 4 gas uptake was up to 3.0 mmol g −1 at 273 K and 1.0 bar [15].
However, with further study of KOH activation, corrosion of KOH and comparatively higher activation temperature has been the most controversial and hottest topic both in practical and theoretical areas. It has been reported that there is a certain positive correlation between the corrosivity of KOH and temperature, namely the corrosivity of KOH will be enhanced at high temperatures, which undoubtedly intensify the corrosion and damage to instruments and equipment, thus causing incalculable economic losses [16,17]. These significant drawbacks will greatly restrict the wide application of KOH in the future. Nowadays, the emergence of sodium amide (NaNH 2 ) activator is gradually attracting wide attention from many researchers. Compared to traditional KOH activator, the corrosiveness and activation temperature of NaNH 2 activator is relatively weaker and lower [18,19]. However, the activation mechanism of NaNH 2 is minimal and there are few reports from relevant researchers. Additionally, with a view of the requirement of low fabricating costs and sustainability in large-scale applications, biomass material is the most promising carbon precursor, which is of wide availability, large scale, low cost and eco-friendly [20][21][22][23][24]. If the grape seed is utilized as a carbon source to synthesize porous carbons for CO 2 , CH 4 capture and electrode materials, high economic and social value could be added to this biomass waste.
In this paper, NaNH 2 and grape seeds were selected as activators and carbon sources to synthesize porous carbon materials by combination of hydrothermal carbonization and NaNH 2 low-temperature activation method. Water is the medium of hydrothermal reaction, no other chemicals need to be added, and the reaction process can be carried out in a closed conditions without secondary pollution. Additionally, the reaction conditions are mild and short, with few degradation products, and the reaction can be easily controlled. Not affected by the moisture content of raw materials, which can save the huge cost of drying materials. As the dehydration and decarboxylation in the process of hydrothermal carbonization is an exothermic reaction, it can provide some energy, thus reducing the energy consumption of the hydrothermal reaction. The water medium atmosphere of hydrothermal carbonization is conducive to the formation of oxygen-containing functional groups on the surface of the material, so the biomass carbon has abundant surface functional groups and good chemical reactivity. The activation equation of NaNH 2 with carbon precursor and the pore formation mechanism of the carbon materials were projected through thermodynamic analysis and TG-IR. The static adsorption amounts of pure component CO 2 , CH 4 , N 2 and the competitive adsorption of CO 2 /N 2 , CO 2 /CH 4 , CH 4 / N 2 mixtures were respectively measured by the ASAP2020 and BSD-MAB instrument, which is used to evaluate the gas separation performance of carbon materials.
The crystal structure of grape-seed derived carbons was characterized by XRD and the data are shown in Fig. 3a. It can be seen that there are two low and broad diffraction peaks at 2θ ≈ 25° (002) and 42° (100) respectively, indicating that all synthesized materials are amorphous graphic carbon. Raman images in Fig. 3b show that the samples have D-band (~ 1347 cm −1 ) and G-band (~ 1582 cm −1 ). D-band corresponds to the degree of defects in the carbon materials, mainly due to the presence of disordered carbons, while the G-band corresponds to the degree of graphitization in the carbon materials, mainly attributed to the vibration of sp 2 hybridized carbon atoms in the graphite layer [25]. Additionally, the I D /I G values of GS-3-400, GS-3-450, GS-3-500 and GS-3-500 are 1.021, 1.035, 1.042 and 1.057 respectively, which indicates that more defects would be generated with the increase of activation temperature. Noteworthy, a 2D band (~ 2816 cm −1 ) with a relatively low width also could be observed in Fig. 3b, which further proves the presence of graphene with a laminar structure in the carbons.

Synthesis of GS precursor
The grape seed carbon materials were purchased and collected from Xinjiang region in China. Granular grape seeds were washed with deionized water, dried in an oven at 80 °C for 48 h, and then crushed to 60 mesh powder by crusher. 8 g of dried grape seed powder was poured into 60 mL of deionized water and stirred thoroughly to obtain mixture. The mixture was poured into the inner liner of the reactor and hydrothermal carbonization was carried out in the homogeneous reactor. The reaction temperature was set at 200 °C and the reaction time was 10 h. At the end of the reaction, the heating was stopped and the reactor device was removed and allowed to cool down to room temperature naturally. The hydrothermal carbon was filtered with the assembled vacuum filter device and washed several times with a large amount of deionized water and anhydrous ethanol until the filtrate appeared light yellow. The moist sample was dried overnight in an oven at 80 °C. The obtained hydrothermal carbon was named as GS.

Synthesis of GS-X-T
The solid NaNH 2 and GS carbon precursor were ground and mixed in different mass ratios (g/g) and transferred to porcelain boat and calcined in a tube furnace with an N 2 flow rate of 200 mL/min, a heating rate of 5 °C/min, activation temperature below 600 °C (T = 400, 450, 500, 550 °C) and activation time of 1 hour. When the temperature dropped to room temperature, take out the samples and magnetically stirred it with a proper amount of 1 M dilute hydrochloric acid solution for 5 h. The samples were washed with plenty of deionized water until the filtrate appeared neutral, collected and dried in an oven at 80 °C overnight. The synthesized grape seed-based activated carbon was named GS-X-T, where X was the mass ratio of NaNH 2 to GS and T was the activation temperature.

Characterization of GS-X-T
The pore structure of the GS-X-T was recorded on a Micrometrics ASAP 2020 instrument. Specific surface areas (S BET ) and pore volumes (V t ) were calculated from the N 2 adsorption isotherms using a Brunauer-Emmet-Teller equation and a t-plot method, respectively. The pore size distribution curves were estimated by means of DFT (density functional theory) model. X-ray diffraction (XRD) were recorded on a Small-angle X-ray diffractometer (D8 advance) in the range of 2θ from 10° to 70°. X-ray photoelectron spectroscopy (XPS) analysis was conducted on a K-Alpha plus X-ray photoelectron spectrometer using Al-Kα as the excitation source (1486.6 eV). Raman spectra were performed by a LabRAM Aramis instrument to determine the carbon composition of the adsorbents. Surface micromorphology of the GS-X-T was recorded using a Hitachi S-3400 N scanning electron microscope (SEM) and a FEI G2 F20 transmission electron microscope (TEM). The adsorption isotherms of CO 2 and CH 4 on GS-X-T were measured by ASAP 2020 physical adsorption apparatus, repectively. Before the test, about 0.1 g GS-X-T sample was taken in a tube and degassed at 160℃ for 12 h. Subsequent non-corrosive gas CO 2 , CH 4 and N 2 were injected and each point of equilibrium recorded continuously in the pressure range of 0 to 101 kPa.

Breakthrough experiment of CH 4 , N 2 and CO 2 mixture gas
The breakthrough curves of mixed gas adsorption were determined by a BSD-MAB multi-component adsorption breakthrough curve analyzer. Before performing the adsorption test, about 0.50 g sample was taken in a quartz tube and the sample was stabilized at 250 ℃ for 1 h with a He flow rate of 28 mL·min −1 . After that, He flow was switched to the desired gas mixture (CO 2 /N 2 = 15v:85v; CO 2 /CH 4 = 40v:60v; CH 4 /N 2 = 50v:50v) at a flow rate of 10 mL·min −1 , the compositions of outlet gas were determined online by a Hiden HPR-20 mass spectrometer. The carbon materials were regenerated under flowing He and target temperature for 1 h for the next breakthrough experiment. The experimental temperature was 298 K and the pressure was 1 bar. Figure 1 shows the synthesis process of grape-seed derived porous hierarchical carbons through hydrothermal carbonization and one-pot NaNH 2 modification at a lower temperature. The carbon precursor GS (Fig. 2a), obtained via the hydrothermal method at 200℃, possesses a relatively dense and rough surface with block structure agglomeration and without obvious pores. After activation by NaNH 2 at 450 °C, abundant pores and obvious cavities can be observed on the surface of carbon materials GS-3-450 (Fig. 2b, c). TEM images of this sample (Fig. 2d, f) further verified that a large number of worm-like nanopores could be clearly observed.

Material synthesis and characterization
The crystal structure of grape-seed derived carbons was characterized by XRD and the data is shown in Fig. 3a. It can be seen that there are two low and broad diffraction peaks at 2θ ≈ 25° (002) and 42° (100) respectively, indicating that all synthesized materials are amorphous graphic carbon. Raman images in Fig. 3b show that the samples have D-band (~ 1347 cm −1 ) and G-band (~ 1582 cm −1 ). D-band corresponds to the degree of defects in the carbon materials, mainly due to the presence of disordered carbons, while the G-band corresponds to the degree of graphitization in the carbon materials, mainly attributed to the vibration of sp 2 hybridized carbon atoms in the graphite layer [25]. Additionally, the I D /I G values of GS-3-400, GS-3-450, GS-3-500 and GS-3-500 are 1.021, 1.035, 1.042 and 1.057 respectively, which indicates that more defects would be generated with the increase of activation temperature. Noteworthy, a 2D band (~ 2816 cm −1 ) with relatively low width also could be observed in Fig. 3b, which further proves the presence of graphene with a laminar structure in the carbons. Figure 4a displays the FT-IR spectra of the samples GS and GS-3-T (T = 400, 450, 500, 550). It can be seen that three obvious peaks are observed in the samples. The strongest peak at 3436 cm −1 is the stretching vibration of -OH/N-H group [26,27], and the peak centered at 1618 cm −1 and 1053 cm −1 are the vibration of N-H/C=C and C-N [28,29], respectively, which indicates that N element was successfully doped into carbon precursor after NaNH 2 activation. The elemental composition of N-doped hierarchical carbons was investigated by XPS, as shown in Fig. 4b and Table 1 Figure 4d shows that N1s can be deconvoluted into three peaks, representing graphitic-N group at 401.7 eV, pyrrole or pyridone-N group at 400.2 eV, and pyridine-N group at 398.9 eV [33,34]. It could be observed that the peak area at around 400.2 eV of pyrrole or pyridone-N is the largest. It indicates that pyrrole or pyridone-N group was major nitrogen-containing group on the surface of GS-3-450. Likewise, O1s peak can also be deconvoluted into two peaks at 533.7 eV and 532.2 eV, respectively (Fig. 4e), corresponding to O-C=O and C-O groups [12,35]. The element contents of C, N and O of these samples analyzed by XPS and Element analysis are listed in Table 1. Compared to prestine carbon precursor GS, N content of GS-3-T significantly increase, especially for the sample GS-3-500. It indicates that N was integrated into the carbon materials successfully via NaNH 2 activation. Figure 5 shows TG and TG-IR images of the mixtures of NaNH 2 and GS, where the mass ratio of NaNH 2 /GS is 3. As shown in Fig. 5a, mass fraction of the mixtures decreases by a total of 55% when temperature is increased to 600 °C. However, there is only 4.41% mass loss when the temperature is increased from 600 to 800 °C. Obviously, mass loss of mixtures mainly happens in the temperature range of 300-600 °C. TG-IR of Fig. 5b verified that lots of gases such as H 2 O, NH 3 , CH 4 , CO 2 , etc. can be observed at the same temperature range. It indicates that carbon materials can be obtained at a temperature below 600 °C.

Activation mechanism analysis
(1) In order to explore the activation mechanism of NaNH 2 , the reaction equation between NaNH 2 and GS was calculated by HSC thermodynamic analysis software, and the results were shown in Fig. 5c, which provides some theoretical support for the pore-forming reason of carbon materials. It can be seen that when temperature is below 270 °C, NaNH 2 reacts with H 2 O adsorbed on the surface of GS to form NaOH and NH 3 (Eq. 1). When temperature is higher than or equal to 270 °C, NaNH 2 can react with C matrix to form CH 4 and N 2 (Eq. 2). When temperature is larger than 450 °C, a part of NaNH 2 begins to decompose (Eq. 3), but the reaction between NaNH 2 and the C matrix still happened simultaneously in this process. When the temperature is higher than 600 °C, NaOH produced by the reaction (Eq. 2) can react with the C matrix (Eq. 4). Combined with the TG image, there is only 4.41% weight loss when the temperature is increased from 600 to 800 °C, thus the formation of pores in carbon materials is ascribe to the reaction of Eq. 2 and Eq. 3 between 300 and 600 °C. Therefore, in this paper, we only systematically explore the reaction below 600 °C because the lower temperatures can reduce energy consumption and avoid the decomposition and loss of N loaded in the carbon. Figure 6 shows N 2 adsorption-desorption isotherms of GS-X-T (X = 2, 3, and 4; T = 400, 450, 500 and 550 °C) at 77 K and their pore size distributions, where X is the mass ratio of NaNH 2 and GS, T is activation temperature. It can be clearly seen that all the adsorption isotherms belong to type

Surface textural properties
2C + 6NaOH → 2Na + 2Na 2 CO 3 + 3H 2 ΔG 4 < 0(T > 600 • C).   This conclusion can be verified by the pore size distribution of these samples with pore size mainly distributed in the range of 0.3-3 nm (Fig. 6b, d). Micropores of the carbons are mainly served as the adsorption sites for small molecule gases CO 2 , CH 4 and N 2 . Textural parameters of GS-X-T are listed in Table 2. The results show that GS-3-450 possesses excellent specific surface area, V Total and V Micro , with values of 3314.5 m 2 /g, 1.45 cm 3 /g and 1.30 cm 3 /g, respectively. However, it is interesting to note that when the activation temperature was increased from 400 to 450 °C, the specific surface area of the carbon material increased sharply from 1068.2 to 3314.5 m 2 /g, which is most likely attributed to the violent reaction of sodium amide with the moisture adsorbed on the carbon material and also accompanied by the redox reaction of sodium amide with carbon and the decomposition of sodium amide. The superposition of the three thermodynamic reactions contributed to the formation of an ultra-high specific surface area of the carbon material GS-3-450. However, with the continuous increase of temperature,  the specific surface area of the carbon material decreases significantly from 3314.5 to 1074.1 m 2 /g, it is likely that the collapse of the carbon skeleton was caused by the violent activation reaction, most of the pore structure was destroyed, and the pores are linked to form larger mesopores, resulting in a sharp drop in the specific surface area. Figure 7 shows the adsorption isotherms of CO 2 , CH 4 and N 2 on GS-3-T (T = 400, 450, 500, 550) at 273 and 298 K. It is obvious that the adsorption amounts of CO 2 and CH 4 are higher than that of N 2 under the same conditions, especially for CO 2 . This is attributed to the difference in their  polarization ratio (CO 2 > CH 4 > N 2 ). Additionally, the adsorption capacity of these pure component gases at 237 K is higher than that of 298 K, which indicates that the adsorption of micromolecules CO 2 , CH 4 and N 2 on the materials is physical adsorption [12]. The adsorption capacities of GS-3-T for pure components CO 2 , CH 4 and N 2 are summarized in Table 3. Obviously, the gases adsorption capacity of GS-3-T presents an increasing and then decreasing trend with the increasing of activation temperature and GS-3-450 has the largest adsorption capacity, which is consistent with their specific surface area (S BET ) and pore volume (V Total ). It indicates that the gases adsorption capacity of GS-3-T is limited to their pore structure parameters.

Gas adsorption and separation performance
To explore the effect of pore volumes of these samples with different pore sizes (0-1 nm, 1-2 nm as well as 2-3 nm) on the adsorption ability of CO 2 and CH 4 , the correlation between pore volumes and gas adsorption amounts under different pore sizes was analyzed linearly. It can be seen that when the pore size is in the range of 0-1 nm, the corresponding ultramicro-pore volume presents good linear relationships with their CO 2 and CH 4 uptakes, and their linear fitting parameters R 2 value is larger than 0.9 (Fig. 8a-c). However, it can be seen from Fig. 8d-f and g-i that correlation between pore volumes (1-2 nm, 2-3 nm) and gas adsorption amounts of CO 2 and CH 4 is weak, and their R 2 value is much less than 0.9. This means that gas adsorption amounts of carbon materials are largely attributed to the influence of their ultramicro-pore volume (0-1 nm). This conclusion is further confirmed by Everett's ideal slit pore model theory. It shows that when the pore size is closer to the kinetic diameter of gas molecules, the adsorption potential is stronger. That means the adsorption force is also stronger. As a result, ultramicro-pores (less than 1 nm) exhibit very strong adsorption force on small gas molecules and result in rapid filling of pores. Obviously, gases storage capacity of carbons are largely limited by the ultramicro-pore volume (0-1 nm).
where q 1 , q 2 , a, b, c, d are parameters of DSLF model. Among, q 1 and q 2 respectively represent saturated adsorption capacity(mmol g −1 ) of two different adsorption sites. a and c respectively represent the trend of adsorption. b and d respectively represent the deviation from the ideal plane.

Dynamic breakthrough experiments
To evaluate the separation performance of grape-seed derived carbon materials for different gas mixtures in practical industry, the BSD-MAB coupled with MS was used to examine the competitive adsorption of CO 2 /N 2 (15v/85v), CO 2 /CH 4 (40v/60v), CH 4 /N 2 (50v/50v) gases mixture on the GS-3-450 at 298 K, and the data are presented in Fig. 10a-c. It is clear that CO 2 exhibits longer penetration time relative to N 2 or CH 4 , which further suggests that the adsorption force between CO 2 and GS-3-450 sample is stronger than that of N 2 and CH 4 , mainly due to the difference in polarization rates between them. In addition, Fig. 10c also presents the competitive adsorption of CH 4 /N 2 mixture on the GS-3-450. It can be seen that the penetration time of CH 4 is slightly longer than that of N 2 , which is also attributed to different polarizability of CH 4 and N 2 .
To further evaluate the cycling stability of GS-3-450, eight cycles of adsorption and desorption penetration experiments of pure CH 4 gas were performed on GS-3-450, as shown in Fig. 10d. It can be clearly observed that the penetration time of CH 4 gas in the GS-3-450 has no significant change, and it is almost kept at about 15 min, which indicates that GS-3-450 has excellent recycling stability and can meet the requirements of using adsorbent materials in practical industry.

Gases adsorption isosteric heat
Isothermal adsorption heats (Q st ) is usually used to characterize the strength of affinity between adsorbent and adsorbate. In order to better understand the interaction between grape-seed derived carbon materials and small molecule gas, isothermal adsorption heats of GS-3-T (T = 400, 450, 500, 550 °C) for CO 2 , CH 4 and N 2 are calculated via Clausius-Clapeyron equation (Eq. 6) from gas adsorption isotherms collected at 273 and 298 K of Fig. 7. As shown in Fig. 11, reduction in Q st value with increasing equilibrium uptake may be related to heterogeneity of adsorption sites, since for homogeneous sites, Q st value should be constant and independent of equilibrium uptake (based on assumption in the Langmuir isotherm model) [43]. The isosteric heat of CO 2 on GS-3-T are 20 to 36 kJ/mol. It is higher than that of CH 4 and N 2 , further indicating that the interaction between GS-3-T and CO 2 is stronger than that of CH 4 and N 2 . Which is beneficial to the separation of CO 2 from gas mixture. In addition, the Q st of GS-3-T for CO 2 , CH 4 and N 2 is lower than 40 kJ/mol, suggesting that the adsorption of CO 2 , CH 4 and N 2 on GS-3-T samples are physical adsorption [44]. Therefore, lesser amount of energy will be expected for regeneration process due to weaker interaction between adsorbate and adsorbent [45].
where Qst (kJ/mol) is the isosteric heat of adsorption, T (K) is the temperature, P (kPa) is the pressure, R is the gas constant, and q (mmol/g) is the adsorbed amount.

Conclusions
A novel strategy for low temperature activation and solvent-free preparation of N-doped porous carbon using grape seed solid waste as carbon source and NaNH 2 as activator was proposed. The activation mechanism of NaNH 2 was first explored by thermodynamic analysis and TG-IR. The prepared carbon material GS-X-T possess a large number of micropores and GS-3-450 has the excellent BET specific surface area (3314.5 m 2 /g), total pore volume (1.45 cm 3 /g) and the highest CO 2 and CH 4 adsorption amounts. Moreover, the adsorption type of CO 2 and CH 4 on the prepared carbon materials is physical adsorption, and the adsorption heat value of CO 2 is higher than that of CH 4 and N 2 . This shows that the prepared grapeseed derived carbon materials was beneficial to the separation of CO 2 from the gas mixture. The low temperature preparation method, excellent BET surface area and total pore volume, as well as excellent CO 2 and CH 4 adsorption ability make it have a very great potential for the capture of CO 2 and CH 4 .