Selectable Microporous Carbons Derived from Poplar Wood by Three Preparation Routes for CO2 Capture.

Biomass-derived porous carbons are one kind of sustainable, extensive, and flexible carbon material for CO2 capture. Here, we prepared several microporous carbons from poplar wood by three preparation routes. Especially, the residues of the poplar wood after the bioethanol process were explored as precursors to prepare activated carbon by KOH and ZnCl2 activation. By the adjustment of the preparation routes and the optimization of the activation conditions, these porous carbons exhibited diversified morphology (sponge, nanosheets, and honeycomb structure), tunable porosity (specific surface areas: 511-2153 m2/g), and narrow micropore distribution (0.55-1.2 nm). These carbons had a high CO2 uptake of up to 217 mg/g at 273 K and 1 bar, which was comparable with those of many N-doped porous carbons, and possessed moderate isosteric heat of CO2 adsorption (21.1-43.2 kJ/mol), good cyclic ability, and high CO2/N2 selectivity (Henry's law: 44.0). The results indicated that CO2 uptake of these carbons was mainly decided by their micropore volume (d < 1.0 nm) at 273 K and 1 bar. This work provides an important reference for preparing promising CO2 adsorbents with tunable structures from similar biomass resources.

Introduction

With the burning of vast amounts of fossil fuel, excessive emissions of class="Chemical">carbon dioxide (class="Chemical">pan class="Chemical">CO2) cause various environmental issues such as global warming, rising sea level, and land desertification.[1−3] Simultaneously, CO2 as a feed gas can also be converted into energy and chemicals in C-1 chemistry.[4] Therefore, it is essential to develop carbon capture and storage (CCS) technologies to reduce CO2 concentration in the atmosphere. The traditional commercial CO2 capture using aqueous ethanolic amines by chemical absorption is highly effective and mature, but it has intrinsic drawbacks such as high regeneration costs and severe corrosion.[5] Notably, the solid adsorbent bed composed of porous materials in a pressure, temperature, and vacuum swing adsorption (P/T/VSA) gas separation system could selectively adsorb CO2 from a humid flue gas; in this way, it has been attracting increasing attention owing to its low cost and environmentally friendly merits in recent years.[6−8]

class="Chemical">Carbon-based materials, such as class="Chemical">pan class="Chemical">graphene,[9] nanotubes, and nanoporous carbons,[10−13] are one of the most advanced adsorbents including porous organic polymers (POPs),[14−18] zeolites,[19] and metal–organic frameworks (MOFs)[20] because of their high specific areas, strong hydrophobicity, low density, and excellent stability. As known, porous carbons can be prepared from various precursors, such as organic small molecules, MOFs, mineral coal, synthetic polymers, and biomass.[21] Among them, biomass-derived porous carbons as renewable, eco-friendly, and cost-effective functional materials have been extensively developed and applied to many fields such as inks, fuel cells, supercapacitors, catalysts, and adsorbents.[22,23] This biomass included catkin, peanut shell, corn stalk, chitosan, and so on.[24−26] Recently, considering that they pose threats to human health (respiratory ailments and skin anaphylaxis), porous carbons derived from poplar catkins (PCs) with rich lignin and amino acids, special morphology of microtubes, and high nitrogen content (>4%) have gained considerable attention[23,27,28] and have been effectively applied in supercapacitors, adsorbents, electrocatalysts, and oil/water separation.[29−32] However, there have been few studies on porous carbons derived from poplar wood sawdust (PWS), a waste product from the wood processing industry, especially on their preparation and applications, such as in gas adsorption and separation fields. Therefore, we selected PWS as the raw material to achieve porous carbons with excellent microporosity and various morphologies via different preparation routes for outstanding CO2 capture.

At present, there are several preparation routes for biomass-derived porous class="Chemical">carbons, such as directed activation class="Chemical">pan class="Chemical">carbonization, hydrothermal treatment and activation, precarbonization and activation, and biotreatment and activation carbonization.[22,23,30,33−37] For example, Chang et al.[23] reported PC-derived hierarchical carbon microtubes by precarbonization, followed by ZnCl2 chemical activation, which showed high CO2 uptake values of 6.22 and 4.05 mmol/g at 273 and 298 K, at 1 bar, respectively. Gao et al.[30] prepared versatile biomass-derived carbon materials (surface area: 1351.4–1525.3 m2/g) by direct carbonization of catkin under ZnCl2 activation for the oxygen reduction reaction, supercapacitors, and oil/water separation. Xu et al.[37] used lignin-derived byproducts (LDBs) after bioethanol production to prepare an interconnected hierarchical porous N-doped carbon (HPNC) (surface area: 2218 m2/g) by KOH activation, and it exhibited favorable properties for supercapacitors. In ref (37), a mass of LDB was produced as the residue after the biotreatment (enzymatic hydrolysis and fermentation) of lignocellulose. On the other hand, activation methods are also critical for the formation and development of carbon materials with a well-developed porous structure. Compared with physical activation, chemical activation has several advantages including simplicity, lower activation temperature, shorter activation time, and higher yield, and it has been popular for remarkably improving the porosity of the generated carbon materials. For example, Manyà et al.[33] prepared several activated carbons (ACs) from vine shoot-derived biochar using CO2 and KOH activation, and the ACs obtained by KOH impregnation exhibited the highest CO2 adsorption capacity (6.04 mmol/g at 273 K and 1.0 bar). Rao et al.[38] prepared N-rich porous sorbents by co-hydrothermal treatment of d-glucose and urea, followed by KOH activation. GN-650-1 with a surface area of 1734 m2/g showed high CO2 uptake values of 4.26 and 6.70 mmol/g at 298 and 273 K, respectively. In the above process, the control of the activator type and dosage, temperature, and residence time was vital for CO2 capture performance of the activated carbons.

Broadly speaking, different preparation routes and activation conditions would largely affect the physicochemical structure, yields, etc. of the produced porous class="Chemical">carbons and thus decide their class="Chemical">performance in class="Chemical">practical aclass="Chemical">pclass="Chemical">plications. For class="Chemical">pan class="Chemical">CO2 capture and storage, excellent microporosity and rich heteroatom (N, O, S) doping are very important for improving their CO2 adsorption capacity and CO2/N2 selectivity.[39] Especially, N-containing groups such as pyridine nitrogen and pyrrole nitrogen were usually introduced in carbon frameworks to provide enough interaction sites for CO2 capture.[9,24] Meanwhile, O doping also inevitably occurred on the carbon surface during activation, and the generated O-containing groups included −C=O, C–OH, and C–O–C; they could form hydrogen bonds between −OH groups and CO2 molecules.

Here, we tried to prepare selectable porous class="Chemical">carbons from class="Chemical">pan class="Disease">PWS by different preparation routes and careful adjustment of the activation conditions. Three preparation routes including direct activation carbonization, hydrothermal treatment and activation carbonization, and biotreatment and activation carbonization were selected based on the research of our groups and the universal use of these routes. Our aim was to investigate the effects of these preparation routes and the corresponding activation conditions on the structural features and CO2 adsorption of PWS-derived porous carbons by a comparative study. Especially, the porosity and microstructure of these porous carbons were well tuned by changing the activation conditions, and they exhibited diversified morphology, tunable porosity, narrow micropore distribution, and high CO2 uptake. This work will provide fundamental guidance for the preparation of porous carbons from renewable lignocellulose resources.

Results and Discussion

Macroscopic Shape, Surface Morphology, and Phase Structure Analysis

class="Disease">PWS as the raw material was converted into class="Chemical">porous class="Chemical">pan class="Chemical">carbons by three routes, which produced different physicochemical structures. The digital photographs of all precursors and porous carbons are shown in Figure S1. The white PWS changed into brown BPWS powder and black HPWS sawdust. After activated carbonization, we found that all KOH-activated porous carbons became an irregular powder, while ZnCl2-activated porous carbons changed into a complete monolith with definite strength; especially, BZC-600-2 exhibited stability in 2 mol/L HCl solution and had low density like aerogels. Figure shows the typical morphology and microstructure of these carbons. DKC-600-2 was composed of irregular carbon blocks and particles with a rough surface (Figure a1,a2); BKC-600-2 had a spongelike network structure with interconnected pores, and its rough surface also accumulated some small carbon particles. Interestingly, a large proportion of the surface of HKC-600-2 possessed vast wormhole-like structures with macropores of 50–100 nm, which was very beneficial for the fast transmission and diffusion of CO2 molecules.

Figure 1

Scanning electron microscopy (SEM) images of DKC-600-2 (a1, a2), BKC-600-2 (b1, b2), HKC-600-2 (c1, c2), HKC-700-2 (d1, d2), HKC-800-2 (e1, e2), and HKC-800-1 (f1, f2).

Scanning electron microscopy (SEM) images of DKC-600-2 (a1, a2), class="Chemical">BKC-600-2 (b1, b2), class="Chemical">pan class="Chemical">HKC-600-2 (c1, c2), HKC-700-2 (d1, d2), HKC-800-2 (e1, e2), and HKC-800-1 (f1, f2).

When the activation temperature increased, many stacked nanosheets appeared in class="Chemical">HKC-700-2, and some irregular class="Chemical">particles and blocks could be seen in class="Chemical">pan class="Chemical">HKC-800-2. When the mass ratio of KOH to HPWS was 1:1, a honeycomb structure appeared in the carbon skeleton of HKC-800-1, and the open pores with thin pore walls were interconnected in hierarchical porous carbon. These results indicated that morphologies of PWS-derived porous carbons could be tuned to a specific shape by changing activation conditions in the preset routes.

As representatives, the transmission electron microscopy (TEM) images of class="Chemical">HKC-600-2 and class="Chemical">pan class="Chemical">HKC-800-1 were investigated. HKC-800-1 shows an interlaced network with a spherical mesh (Figure a1), and an amorphous structure was formed in the porous carbon networks, which was consistent with the X-ray diffraction (XRD) results. Its high-resolution FE-TEM image (Figure a3) exhibits alternately dark and bright microstructures with extensive sizes, indicating plentiful micropores and a few mesopores. The TEM images of HKC-600-2 show thin sheets and particles (Figure b1), which may be a mixture of graphene-like layers and amorphous structure, and the high-resolution FE-TEM image (Figure b3) revealed more serried wormhole-like microstructures and a few stripe structures.

Figure 2

TEM images of HKC-800-1 (a1–a3) and HKC-600-2 (b1–b3).

TEM images of pan class="Chemical">HKC-800-1 (a1–a3) and class="Chemical">pan class="Chemical">HKC-600-2 (b1–b3).

The Raman spectra of class="Chemical">carbon materials are shown in Figure a. Two dominant characteristic class="Chemical">peaks are attributed to the symmetrical vibration of the sclass="Chemical">p2 class="Chemical">phase at 1583 cm–1 and the defective structure at 1332 cm–1, resclass="Chemical">pectively, corresclass="Chemical">ponding to the G-band and D-bands.[41] The intensity ratio of IG/ID reflected the degree of graclass="Chemical">phitization of class="Chemical">porous class="Chemical">pan class="Chemical">carbons, and the value of IG/ID of HKC-600-2 (1.00) was higher than that of HKC-800-1 (0.97), indicating that HKC-600-2 had a more ordered structure and a higher graphitization degree. Figure b exhibits the XRD pattern of porous carbons. For HKC-600-2, two broad diffraction peaks appeared at 2θ = 24.1 and 43.4°, suggesting that the amorphous phase was dominant; correspondingly, the two peaks can be attributed to the reflection (002) and (100) of the stacking graphitized carbon structure.[42] The diffraction peaks of HKC-800-1 showed a slight shift (2θ = 20.8 and 43.4°), and its XRD spectra showed a sharp diffraction peak (2θ = 28.1°) with low intensity, which can be ascribed to the slight impurities of siliceous compound crystals from the weak reaction between the activator KOH and the porcelain boat composed of SiO2. Additionally, there appeared a rapid increase in the intensity of the diffraction peaks at the low-angle region (2θ < 10°), which indicated the high porosity of the two carbon materials.[43]

Figure 3

Raman spectra (a) and XRD spectra (b) of HKC-800-1 and HKC-600-2.

Raman spectra (a) and XRD spectra (b) of pan class="Chemical">HKC-800-1 and class="Chemical">pan class="Chemical">HKC-600-2.

Thermostability and Carbon Yields

The thermogravimetric (TG) and derivative thermogravimetric (DTG) analysis curves of three precursors (class="Disease">PWS, Bclass="Chemical">pan class="Disease">PWS, HPWS) were measured to study their thermostability and decide the activation temperature (Figure ). Three precursors exhibited similar TG curves (Figure a), but the initial decomposition temperature was different, which was about 211 °C (PWS), 156 °C (BPWS), and 232 °C (HPWS). The lowest decomposition temperature of BPWS could be ascribed to the remaining microbial protein and polysaccharides adhering to it. HPWS showed the highest decomposition temperature due to its enhanced robust structure after the hydrothermal process.[44] The weight of the three precursors mainly kept constant above 600 °C; the residues of PWS, BPWS, and HPWS were about 19.1, 30.6, and 43.6 wt %, respectively. Hence, the temperature of 600 °C was used as the initial activation temperature. The TGA curves (Figure b) more clearly revealed the weight loss process; the loss below 100 °C was mainly from the adsorbed moisture, and the strongest weight loss peaks were 341, 321, and 359 °C corresponding to PWS, BPWS, and HPWS. The peak shape of HPWS had an evident difference compared with the other two, which could be due to its more robust structure formed by the intense condensation polymerization and aromatization during the hydrothermal process.[44]

Figure 4

TG curves (a) and DTG curves (b) of PWS, BPWS, and HPWS.

TG curves (a) and DTG curves (b) of class="Disease">PWS, Bclass="Chemical">pan class="Disease">PWS, and HPWS.

The class="Chemical">DTA curves of these samclass="Chemical">ples suggested that the total class="Chemical">process was mainly endothermic (Figure S2), and Bclass="Chemical">pan class="Disease">PWS and HPWS had weak exothermic peaks at about 517, 556, and 678 °C; the reason could be that these generated small-molecule gases and condensational volatiles were formed under exothermic reactions. The yields of these samples are summarized in Table S1. As for the yields of the porous carbons before and after activation, ZnCl2-activated carbons had much higher yields (34.5–58.2%) than KOH-activated carbons (4.81–32.9%). Additionally, with the activation temperature or activator dosage increased, the yields of porous carbons were expectedly decreased. The total yields for the three routes originating from PWS were 2.52–16.44% for KOH activation and 15.92–34.50% for ZnCl2 activation. Among them, DZC-600-2 had the highest total yield of 34.50%, which could be ascribed to the carbonization of the most original components with the help of ZnCl2 activators. In short, the yields of these carbons were acceptable for practical production.[45]

Porous Texture Studies

According to the results of class="Chemical">CO2 caclass="Chemical">pture class="Chemical">performance, we selected class="Chemical">partial class="Chemical">porous class="Chemical">pan class="Chemical">carbons as representatives and measured the N2 adsorption–desorption isotherms (Figure ).

Figure 5

N2 adsorption–desorption isotherms at 77 K (a) and the pore size distribution by the NLDFT model (b) of porous carbons.

pan class="Chemical">N2 adsorclass="Chemical">ption–desorclass="Chemical">ption isotherms at 77 K (a) and the class="Chemical">pore size distribution by the NLDFT model (b) of class="Chemical">porous class="Chemical">pan class="Chemical">carbons.

According to the International Union of class="Chemical">Pure and Aclass="Chemical">pclass="Chemical">plied Chemistry (IUclass="Chemical">pan class="Chemical">PAC) classification,[46] all isotherms could be identified as the Type-I isotherms except for HKC-800-2, which belonged to a combination of Type-I and Type-IV. At P/P0 < 0.005, the N2 uptake of all carbons has a rapid increase, indicating abundant micropores, and then, there appeared a mild rise at 0.05 < P/P0 < 0.3 for HKC-700-2, HKC-800-2, and HKC-800-1, implying the presence of some mesopores, while the isotherms of DKC-600-2, BKC-600-2, and HKC-600-2 all kept constant at P/P0 > 0.01, suggesting almost no mesopores or macropores, and finally, all isotherms leveled off in the high P/P0 region. Notably, the N2 adsorption–desorption isotherms of HKC-800-2 revealed a prominent Type-H4 hysteresis loop due to the capillary condensation from some mesoporous structures.[46] The structural parameters of these carbons are summarized in Table . DKC-600-2, BKC-600-2, and HKC-600-2 prepared under the same activation conditions had different porosities; for example, their SBET was 511, 535, and 893 m2/g and their Vmicro was 0.17, 0.22, and 0.33 cm3/g, respectively. The results demonstrated that the two-step routes, especially the hydrothermally combined activation routes, could endow the porous carbons with higher SBET and Vmicro. Thus, the following study mainly focused on the HKC materials from the third route. For HKC-600-2, HKC-700-2, and HKC-800-2, with the increase in activation temperature, SBET, Vtotal, Vmicro, and DA increased, and SBET and Vtotal of HKC-800-2 reached 1587 m2/g and 1.00 cm3/g, respectively. Interestingly, when the activator dosage was adjusted to 1:1, the obtained HKC-800-1 had the largest SBET (2153 m2/g), Vtotal (1.13 cm3/g), and Vmicro (0.85 cm3/g). These results suggested that a higher activation temperature could result in larger porosity, but the increase in activators had not necessarily enhanced the porosity, especially microporosity. The pore size distribution (PSD) in Figure b revealed the dominant pore size in these carbons. DKC-600-2, BKC-600-2, HKC-700-2, and HKC-800-2 had similar micropore distributions concentrated at 1.1 and 1.2 nm, while HKC-600-2 and HKC-800-1 had two characteristic peaks focused at 0.85 and 1.1 nm and 0.55 and 0.97 nm, respectively, and HKC-800-1 had a smaller pore size with a higher intensity of 5.4 cm3 (g/nm). In addition, the PSD curves of individual samples obtained by the NLDFT model or the BJH model (Figures S3 and S4) showed the presence of mesoporosity in some samples, such as HKC-700-2, HKC-800-2, and HKC-800-1. They possessed mesopores with the pore size of 2–5 nm. The results were in accordance with the N2 adsorption isotherms. The curves of the cumulative pore volume vs pore size are also shown in Figure S5. The pore volume of different pore size regions could be found; for example, the pore volumes of d < 1 nm were 0.0010, 0.026, 0.24, 0.0051, 0.071, and 0.52 cm3/g corresponding to DKC-600-2, BKC-600-2, HKC-600-2, HKC-700-2, HKC-800-2, and HKC-800-1. The above results implied that the porous structure of PWS-derived carbons could be regularly adjusted by the preparation routes, especially the activation conditions.

Table 1

Porous Properties and Elemental Content of Porous Carbons Derived from Poplar Wood Sawdust

samples	SBETa	Vtotalb	Vmicroc	V(d < 1 nm)d	V(d < 0.7 nm)d	DAe	C (%)	N (%)	H (%)	Of (%)
DKC-600-2	511	0.22	0.17(77.3)	0.0010	0.00	1.77	73.44	<0.01	2.06	24.50
BKC-600-2	535	0.23	0.22(95.7)	0.026	0.00	1.74	57.18	<0.01	1.35	41.46
HKC-600-2	893	0.38	0.33(86.8)	0.24	0.0064	1.70	70.87	0.00	1.37	27.75
HKC-700-2	1339	0.68	0.52(76.5)	0.0051	0.00	2.02	85.84	<0.01	1.14	13.02
HKC-800-2	1587	1.00	0.55(55.0)	0.071	0.017	2.53	48.27	0.00	1.24	50.49
HKC-800-1	2153	1.13	0.85(75.2)	0.52	0.088	2.09	85.11	0.00	0.66	14.23

Calculated using the BET model with the unit of m2/g.

Calculated at P/P0 = 0.99 with the unit of cm3/g.

Cumulative micropore volume with pore size <2.0 nm using the NLDFT model with the unit of cm3/g.

Cumulative narrow micropore volume with pore size <1.0 nm or <0.7 nm using the NLDFT model with the unit of cm3/g.

The average pore width was calculated from DA = 4Vtotal /SBET.

The O content was calculated by the difference value method, O% = 100%-C%-H%-N%, and the uncertainty of all elemental contents was in the range 0.007–0.02 wt %.

Chemical Structure and Composition

The chemical structure of these porous class="Chemical">carbons was measured by elemental analysis (EA) and Fourier transform infrared (FT-IR) sclass="Chemical">pectroscoclass="Chemical">py. The C, N, O, and H contents of the samclass="Chemical">ples are class="Chemical">provided in Table . The N content of all class="Chemical">porous class="Chemical">pan class="Chemical">carbons was almost zero (<0.01 wt %). DKC-600-2 and HKC-600-2 had similar elemental compositions, and C content of both samples reached more than 70 wt %, while BKC-600-2 had a lower C content (57.18 wt %) and higher O content (41.46 wt %), which could be due to the adhered polysaccharides on BPWS.[37] With increase of activation temperature, the O content of HKC-800-2 increased to about 50 wt %, but the O content (14.23 wt %) of HKC-800-1 was low. The above results suggested that the use of a larger amount of KOH at high temperatures could lead to more O atoms being doped into the carbon skeletons, which could be because of the enhanced surface reoxidation between reactive carbon sites and excess KOH.[71]

The FT-IR spectra of Hclass="Disease">PWS and class="Chemical">pan class="Chemical">HKCs show the change of functional groups in the third route (Figure ). HPWS exhibited strong absorption peaks at 2929 cm–1 for −CH2, 1697 cm–1 for −C=O–, 1645 and 1540 cm–1 for aromatic ketone, and 1460 cm–1 from benzene rings. After activation carbonization, most of these characteristic peaks became weak, even disappeared, and these porous carbons retained some characteristic peaks from the aromatic ring. Notably, the bands of these carbons at 1087 cm–1 from the stretching vibration of −C–O–C became broader and stronger, which could be because of enhancement during KOH etching of the carbon framework, and HKC-800-2 showed the strongest absorption for −C–O–C, which was consistent with EA results.

Figure 6

FT-IR spectra of HPWS and porous carbons.

FT-IR spectra of Hpan class="Disease">PWS and class="Chemical">porous class="Chemical">pan class="Chemical">carbons.

CO2 Capture Capacity

The pan class="Chemical">CO2 adsorclass="Chemical">ption isotherms of all class="Chemical">porous class="Chemical">pan class="Chemical">carbons at 273 and 298 K were measured in the pressure range of 0–1 bar (Figures and S6).

Figure 7

CO2 adsorption isotherms of all porous carbons at 273 K (a, c) and 298 K (b, d).

pan class="Chemical">CO2 adsorclass="Chemical">ption isotherms of all class="Chemical">porous class="Chemical">pan class="Chemical">carbons at 273 K (a, c) and 298 K (b, d).

First, Figure a,b shows class="Chemical">CO2 adsorclass="Chemical">ption isotherms of these class="Chemical">porous class="Chemical">pan class="Chemical">carbons prepared from three different routes. Their CO2 uptake changed from 80.0 to 161.1 mg/g at 273 K and 1 bar and revealed that KOH-activated porous carbons had higher CO2 uptake compared with the corresponding ZnCl2-activated porous carbons except for DKC-600-2 and DZC-600-2, which can be attributed to the different activation mechanism and precursor features.[23,30] Among KOH-activated porous carbons, BKC-600-2 had a CO2 uptake of 116.0 mg/g at 273 K and 1 bar, which showed an increase of 31.1% relative to DKC-600-2 (88.5 mg/g). Similarly, the CO2 uptake of HKC-600-2 was 54.6 and 161.1 mg/g at 0.15 and 1 bar, at 273 K, respectively, with an increase of 114.1 and 82.0% compared to DKC-600-2. This clearly demonstrated that the bioethanol process and hydrothermal treatment were effective steps for improving CO2 adsorption of porous carbons.

The following study focused on the performance optimization of class="Chemical">HKCs. With an increase of the activation temclass="Chemical">perature, the class="Chemical">pan class="Chemical">CO2 uptake at 273 K and 1.0 bar slightly decreased, and the values were 161.1, 124.5, and 151.6 mg/g corresponding to HKC-600-2, HKC-700-2, and HKC-800-2. HKC-800-2 had the highest SBET of 1587 m2/g, but the CO2 uptake was not the largest, implying that SBET was not the dominating factor. To achieve higher CO2 uptake, we selected HKC-600-2 and HKC-800-2 for further optimization, and the previous references indicated that excess activators could lead to severe etching and the collapse of the micropore structure.[38,47] Hence, HKC-600-1 and HKC-800-1 were prepared by reducing the amount of activators, and the CO2 uptake of both reached 146.5 and 217 mg/g, respectively, suggesting that the appropriate activation dosage was very important at a high temperature. The CO2 uptake of these porous carbons is summarized in Table . For ZnCl2-activated porous carbons, with an increase of the activation temperature (Table S2 and Figure S6), the CO2 uptake first increased and then slightly decreased, and the values were 90.3, 120.2, and 113.8 mg/g at 273 K and 1 bar. At 298 K and 1 bar, the CO2 uptake values of all porous carbons were in the range of 48.6–126.1 mg/g.

Table 2

CO2 Capture Performance of Porous Carbons

	CO2 uptakea		CO2 uptakeb
samples	1.0 bar	0.15 bar	1.0 bar	N2 uptakea	Henry’s law SCO2/N2c	IAST SCO2/N2d	Qste
DZC-600-2	104.7	28.6	55.7	2.8	26.16	319.0	28.3
BZC-600-2	80.0	21.8	47.7	8.0	8.66	NA	22.7
HZC-600-2	90.3	24.8	56.9	NA	NA	NA	21.1
DKC-600-2	88.5	25.5	48.6	15.7	3.84	27.8	26.4
BKC-600-2	116.0	39.7	67.9	2.0	44.00	442.0	28.1
HKC-600-2	161.1	54.6	100.4	5.3	23.30	274.5	37.0
HKC-700-2	124.5	37.8	77.8	10.0	7.91	68.2	22.2
HKC-800-2	151.6	37.2	90.5	7.9	12.60	99.5	21.5
HKC-600-1	146.5	63.9	96.6	9.0	16.75	NA	43.2
HKC-800-1	217.0	59.1	126.1	9.6	27.93	NA	24.9

Gas adsorption in mg/g at 273 K/1.0 and 0.15 bar, and the standard uncertainties, u, of P, T, CO2 uptake, and N2 uptake are u(P) = 0.002 bar, u(T) = 0.1 K, u(CO2 uptake) = 0.001 mmol/g, and u(N2 uptake) = 0.0012 mmol/g; NA means unknown.

Gas adsorption in mg/g at 298 K/1.0 bar, and the standard uncertainties, u, of P, T, and CO2 uptake are u(P) = 0.002 bar, u(T) = 0.1 K, u(CO2 uptake) = 0.0012 mmol/g.

Henry’slaw SCO at 273 K.

IAST SCO at 273 K for the mixture including 85% of N2 and 15% of CO2 at 1.0 bar.

Qst of CO2 in kJ/mol calculated by the Clausius–Clapeyron equation at a low CO2 loading.

class="Gene">Gas adsorclass="Chemical">ption in mg/g at 273 K/1.0 and 0.15 bar, and the standard uncertainties, u, of class="Chemical">pan class="Chemical">P, T, CO2 uptake, and N2 uptake are u(P) = 0.002 bar, u(T) = 0.1 K, u(CO2 uptake) = 0.001 mmol/g, and u(N2 uptake) = 0.0012 mmol/g; NA means unknown.

class="Gene">Gas adsorclass="Chemical">ption in mg/g at 298 K/1.0 bar, and the standard uncertainties, u, of class="Chemical">pan class="Chemical">P, T, and CO2 uptake are u(P) = 0.002 bar, u(T) = 0.1 K, u(CO2 uptake) = 0.0012 mmol/g.

IAST SCO at 273 K for the mixture including 85% of pan class="Chemical">N2 and 15% of class="Chemical">pan class="Chemical">CO2 at 1.0 bar.

Qst of pan class="Chemical">CO2 in kJ/mol calculated by the Clausius–Claclass="Chemical">peyron equation at a low class="Chemical">pan class="Chemical">CO2 loading.

The class="Chemical">CO2 uclass="Chemical">ptake and textural class="Chemical">proclass="Chemical">perties of these class="Chemical">porous class="Chemical">pan class="Chemical">carbons were compared with various carbon materials (Tables and S3). Among free N-doped porous carbons, HKC-800-1 shows superior CO2 adsorption of 217 mg/g relative to commercial activated carbon (123.2 mg/g),[48] OMC (132 mg/g),[49] OM-CNS (175.1 mg/g),[50] CA-HC200 (198.4 mg/g),[51] and PC500 (190.5 mg/g),[52] and it was also comparable with PMMC-800 (237.6 mg/g),[53] NET2-2-700-2 (228.8 mg/g),[54] and L2600 (233.2 mg/g).[36] Compared with these N-doped porous carbons, such as salt-templated carbons with arginine (147.8 mg/g),[56] FC4 (178.2 mg/g),[57] OTSS-1-550 (191.4 mg/g),[58] N-PHCS-900 (194.5 mg/g),[59] microporous carbon from fern leaves (198.9 mg/g),[60] and PDA0.3/MA0.7-2 (202.4 mg/g),[62] the CO2 uptake of our porous carbons was also decent. HKC-800-1 was also comparable with c-CBAP-1N (223.5 mg/g),[66] H150-800 (228.1 mg/g),[67] NPC500 (235.8 mg/g),[52] Bamboo-1-973 (233.2 mg/g),[33] and AC-KOH-W-2-700 (237.6 mg/g).[68] Of course, the CO2 uptake of HKC-800-1 was inferior to those of some advanced carbon materials including ACDS-800-2 (264 mg/g),[70] CMS-K3 (286.4 mg/g),[71] and CSC-650 (295.7 mg/g).[72] To deeply understand the CO2 adsorption behavior, two adsorption models (Langmuir and Freundlich) were used to simulate the CO2 adsorption isotherms (Figure S7), and Table S4 summarizes these parameters such as Qm, KL, KF, and the correlation coefficient (R2). At 273 K, the Langmuir and Freundlich models both could well fit CO2 adsorption isotherms with R2 > 0.99, and the Freundlich model exhibited better fitting with a higher R2. Qmax of HKC-800-2 and HKC-800-1 reached 425.3 and 450.7 mg/g at 273 K, respectively, and this was because their higher SBET would provide more adsorption sites under high pressure. In addition, HKC-600-1 had higher KL and KF at 273 K, suggesting stronger affinity for CO2 molecules, and this parameter with n > 2 implied preferential adsorption. To further investigate the relationship between CO2 uptake and structural properties, SBET, V(d<1.0 nm), Vtotal, and O content were plotted with the CO2 uptake, respectively (Figures and S8).

Table 3

Textural Properties and the CO2 Uptake of Various N-Doped Carbon-Based Adsorbents

				CO2 uptake (mg/g)		selectivity
samples	SBET (m2/g)	Vmicro (cm3/g)	N content (%)	273 K	298 K	Henry’s law	IAST	refs
template carbon	857	NA	2.94	147.8	NA	NA	NA	(56)
FC4	941	0.31	NA	178.2	126.3	NA	14.2b	(57)
OTSS-1-550	777.7	0.27	0.73	191.4	136.4	NA	47.77	(58)
N-PHCS-900	775	0.32	8.39	194.5	130.2	35	NA	(59)
fern carbons	1593	0.54	NA	198.9	181.3	NA	5.6b	(60)
EAZn2-2d-C	829	NA	1.23	202.4	NA	NA	NA	(61)
PDA0.3/MA0.7-2	866	0.382	20.9a	202.4	160.2	NA	115	(62)
ANCs-3-800	3138	1.33	2.73a	202.4	140.8	NA	NA	(63)
NPC-4-600	1518	0.54	9.71	207.0	128.0	16.6	34.5	(64)
MPC-750	1881	0.78	0.36	216.5	125.8	NA	NA	(65)
HKC-800-1	2153	0.85	0.00	217.0	126.1	27.93	NA	this work
c-CBAP-1N	1063	NA	2.93	223.5	154.0	25.7	NA	(66)
H150-800	1322	0.23	2.4	228.1	149.6	NA	NA	(67)
NPC500	1082	0.441	9.44a	235.8	166.3	NA	NA	(52)
Bamboo-1-973	930	NA	NA	233.2	176	11	NA	(68)
AC-K-W-2-700	1671	0.587	NA	237.6	162.8	NA	NA	(33)
NHPCT-4-7	1361	0.46	1.89	243.3	156.2	22	54	(23)
HCP2a-K700	1964	0.92	0.04	251.0	134.0	8.7	10.8	(69)
ACDS-800-2	1634	0.560	0.82	264.0	182.2	NA	NA	(70)
CMS-K3	1354	0.539	0.81a	286.4	179.1	16	15	(71)
CSC-650	1182	0.522	0.06	295.7	208.6	26.7	21.35	(72)

The N content was measured by XPS analysis.

The results were calculated at 298 K. NA means unknown.

Figure 8

Relationships between SBET and CO2 uptake (a) and Vmicro (d < 1.0 nm) and CO2 uptake (b).

Relationships between SBET and pan class="Chemical">CO2 uclass="Chemical">ptake (a) and Vmicro (d < 1.0 nm) and class="Chemical">pan class="Chemical">CO2 uptake (b).

The class="Chemical">CO2 uclass="Chemical">ptake of these class="Chemical">porous class="Chemical">pan class="Chemical">carbons at 273 K and 1.0 bar, 273 K and 0.15 bar, and 298 K and 1.0 bar exhibited the same trend with the increase of SBET, respectively, which indicated that SBET had a similar effect on the CO2 uptake under the above adsorption conditions. It has been reported that Vmicro is an important factor that determines CO2 uptake; especially, the micropores with 2–3 times the pore size of the diameter of CO2 molecules (0.33 nm) would maximally enhance the adsorption potential.[34,47,54,60] Then, Vmicro (d < 1.0 nm) was linearly fitted with the CO2 uptake, and the correlation coefficient (R2) was 0.8352 (at 273 K and 1 bar), 0.7239 (at 273 K and 0.15 bar), and 0.7580 (at 298 K and 1 bar), implying that Vmicro (d < 1.0 nm) had an evident positive effect on CO2 uptake, and the results were also in accordance with some previous reports.[47,54,60] Meanwhile, the importance of the ultramicropores with d < 0.7 nm has been pointed out in some recent references. The ultramicropore volumes (Vultra) of these carbons are summarized in Table . DKC-600-2, BKC-600-2, and HKC-700-2 did not have ultramicropores, but Vultra of HKC-600-2, HKC-800-2, and HKC-800-1 was 0.0064, 0.017, and 0.088 cm3/g, respectively. The relationship between Vultra and CO2 uptake (at 273 or 298 K and 1 bar) is plotted in Figure S9, and we found that HKC-800-2 with the largest Vultra had the highest CO2 uptake, and the CO2 uptake of these carbons without ultramicropores was lower than those of other carbons with ultramicropores. Hence, the importance of ultramicropores was verified again. In addition, Vtotal was also used to plot the CO2 uptake; the curves of Vtotal vs CO2 uptake were similar to SBET vs CO2 uptake. Generally, the O content of these porous carbons would inherently affect the acidity/basicity of carbon materials, and basic groups containing pyrone, chromene, and diketone/quinone type of structures could be beneficial for CO2 adsorption.[73] Here, the O content of these porous carbons seemingly did not have a regular relationship with the CO2 uptake (Figure S8), and HKC-800-2 with the highest O content did not appear to have a large enhancement of the CO2 uptake relative to HKC-700-2 with the lowest O content. The results suggested that the O doping may not be effective enough at improving CO2 capture under the dominant effect of microporosity, and previous references also showed similar results.[52,54]

Sensibility of Adsorption Temperature, Isosteric Heat of CO2 Adsorption (Qst), and Recycling

Considering the higher operating temperature in the adsorption of class="Chemical">CO2 from industrial flue class="Chemical">pan class="Gene">gases, the CO2 adsorption isotherms of HKC-800-1 at five adsorption temperatures (273–343 K) were measured (Figure a). The change of CO2 uptake represented its sensibility of adsorption temperature. It can be seen that the adsorption isotherms moved down with increasing adsorption temperature and basically kept unchanged at 323 and 343 K, and the values of CO2 uptake at 1.0 and 0.15 bar under different adsorption temperatures clearly showed the change (Figure b). The results indicated that HKC-800-1 had good stability of CO2 adsorption at a high adsorption temperature. We think that the larger number of ultramicropores of HKC-800-1 can capture a constant number of CO2 molecules due to its strong adsorption even at high temperatures, and thus its adsorption capacity almost has no loss in a certain temperature range.

Figure 9

CO2 adsorption isotherms of porous carbons at different temperatures (a) and the change of CO2 uptake with adsorption temperature at 1.00 and 0.15 bar (b).

class="Chemical">CO2 adsorclass="Chemical">ption isotherms of class="Chemical">porous class="Chemical">pan class="Chemical">carbons at different temperatures (a) and the change of CO2 uptake with adsorption temperature at 1.00 and 0.15 bar (b).

To investigate the interaction strength between these class="Chemical">carbons and class="Chemical">pan class="Chemical">CO2, the CO2 adsorption isotherms at 273 and 298 K were used to calculate Qst by the Clausius–Clapeyron equation (Figures and S10). For ZnCl2-activated carbons, DZC-600-2 had a larger Qst (28.3 kJ/mol) compared with BZC-600-2 (22.7 kJ/mol) and HZC-600-2 (21.1 kJ/mol). For KOH-activated carbons, the Qst curves of HKC-600-2 and HKC-600-1 evidently decreased with an increase of CO2 loading due to the heterogeneous adsorption sites, and thus the remaining weak sites made Qst decrease. The initial Qst of DKC-600-2, BKC-600-2, and HKC-600-2 was 26.4, 28.1, and 37.0 kJ/mol, respectively, which was less than 40 kJ/mol, suggesting a physical adsorption process. In addition, the initial Qst of all HKCs (Figure b) decreased with the increase of activation temperature; especially, Qst of HKC-600-1 reached 43.2 kJ/mol, suggesting the possibly existing weak chemical interaction. The initial Qst of the above activated carbons (21.1–43.2 kJ/mol) exceeded the heat of liquefaction of CO2 (17 kJ/mol),[67] and Qst of these carbons at low CO2 loading were comparable, even higher than many other carbon-based adsorbents, such as OM-CNS (28.4 kJ/mol),[50] PMMC-800 (∼24.5 kJ/mol),[53] NET2-2-700-2 (23.2 kJ/mol),[54] AcA5 (∼29.7 kJ/mol),[55] FC4 (24.9 kJ/mol),[57] and CSC-650 (25.8 kJ/mol).[72] Fortunately, HKC-800-1 with the highest CO2 uptake had a moderate Qst of 24.9 kJ/mol, which was beneficial to easy regeneration and good cycling of adsorbents. The reclaimed HKC-800-1 was degassed at 30 °C for 30 min before the next cycle, and the reusability was measured (Figure c). After the first cycle, the values of CO2 uptake at 273 K and 1 bar had slightly decreased due to the difficult desorption of some CO2 molecules and the influence of impurities possibly, but the CO2 adsorption isotherms were almost coincidental in the following four cycles and had no loss. This result indicated the acceptable recyclability. HKC-800-1 as a CO2 adsorbent could be used in pressure, vacuum swing adsorption (P/VSA) processes.

Figure 10

Qst of porous carbons with CO2 loading (a, b) and the cycle performance of HKC-800-1 at 273 K (c).

Qst of porous class="Chemical">carbons with class="Chemical">pan class="Chemical">CO2 loading (a, b) and the cycle performance of HKC-800-1 at 273 K (c).

CO2/N2 Selectivity

The class="Chemical">CO2/class="Chemical">pan class="Chemical">N2 selectivity is also an important index for CO2 capture. The N2 adsorption isotherms of all samples were measured at 273 K (Figure S11). The N2 uptake of these carbons was in the range of 2.8–15.7 mg/g at 273 K and 1 bar. DKC-600-2 has a higher N2 uptake of 15.7 mg/g, which was about one-sixth of CO2 uptake (88.5 mg/g), implying low selectivity. Other microporous carbons have a lower ratio of N2 uptake to CO2 uptake, implying higher selectivity. To evaluate the CO2/N2 selectivity in practical applications, the selectivity at 273 K was calculated by the initial slope method (Henry’s law) and ideal adsorption solution theory (IAST).[74] These results are summarized in Table and Figures S2,S5,S12, and S13. For Henry’s law selectivity, the values of these microporous carbons were in the range of 3.84–44.0. DZC-600-2 (26.16) had higher selectivity relative to DKC-600-2 (3.84), while BKC-600-2 had (44.0) higher selectivity compared to BZC-600-2 (8.66), implying that the synthetic route also had an important effect on the CO2/N2 selectivity apart from the activator. Among all HKC carbons, HKC-800-1 had the highest CO2/N2 selectivity of 27.93, which could benefit from its highest Vmicro and V(. Regarding IAST selectivity, the calculations of some samples failed because the fitting of these models showed a poor correlation coefficient. Other microporous carbons revealed selectivities of 17.3–442.0 at 273 K and 1 bar, and the order of IAST selectivity of these porous carbons was in accordance with Henry’s law selectivity. BKC-600-2 had an ultrahigh selectivity of 442.0, and we thought that its higher O content of 41.46% and the largest microporosity (Vmicro/Vtotal) of 95.7% could make important contributions. The Henry’s law selectivity of HKC-800-1 (27.93) surpassed those of many porous carbons, such as commercial activated carbon (17),[48] NPC-4-600 (16.6),[64] c-CBAP-1N (25.7),[66] Bamboo-1-973 (11.0),[33] NHPCT-4-7 (22),[23] HCP2a-K700 (8.7),[69] and CSC-650 (26.7).[72]

Conclusions

In this work, poplar wood sawdust-derived microporous class="Chemical">carbons were class="Chemical">preclass="Chemical">pared by direct chemical activation, bio-class="Chemical">pretreatment and activation, and hydrothermal class="Chemical">pretreatment and activation. By the adjustment of the class="Chemical">preclass="Chemical">paration routes and the oclass="Chemical">ptimization of the activation conditions, the class="Chemical">produced class="Chemical">porous class="Chemical">pan class="Chemical">carbons exhibited diverse morphologies, tunable porosity (specific surface area: 511–2153 m2/g), and a narrow micropore distribution (0.55–1.2 nm). These microporous carbons exhibited high CO2 uptake (80–217 mg/g) at 273 K and 1 bar, which was comparable with those of many N-doped porous carbons, and high CO2/N2 selectivity and good cyclic ability. These properties make them promising adsorbents for industrial CO2 capture.

Experimental Section

Materials

pan class="Disease">PWS (diameter: 1–2 mm, length: 2–15 mm) was obtained from a wood class="Chemical">processing factory. Sources of other materials are shown in Table .

Table 4

Molecular Weights, Purities, Sources, and CAS-Numbers of the Chemicals

chemical name	mol. wt.	purities (%)	CAS-no.	sources
ethanol	46.07	≥99.5	64-17-5	Tianjin YongDa Chem. Technol. Co., Ltd.
KOH	56.11	≥85.0	1310-58-3	Sinopharm Chemical Reagent Co., Ltd.
ZnCl2	136.30	≥98	7646-85-7	Sinopharm Chemical Reagent Co., Ltd.
HCl	36.46	38.0	7647-01-0	Sinopharm Chemical Reagent Co., Ltd.
H2SO4	98.04	≥98.0	7664-93-9	Nanjing Chemical Reagent Co. Ltd.
CO2	44.0	≥99.99	124-38-9	Changsha XinXiang Gas Chem. Co., Ltd.
N2	28.0	99.999	7727-37-9	Changsha XinXiang Gas Chem. Co., Ltd.

Preparation of Microporous Carbons by Three Synthetic Routes

All preparation processes of microporous pan class="Chemical">carbons are shown in Scheme .

Scheme 1

Fabrication of Poplar Wood Sawdust-Derived Porous Carbons with Multiple Requirements from Three Synthetic Routes

For the first route, the one-step activation class="Chemical">carbonization, tyclass="Chemical">pically, aclass="Chemical">pclass="Chemical">proximately 1.0 g of dried class="Chemical">pan class="Disease">PWS and 2.0 g of KOH or ZnCl2 were thoroughly mixed in an agate mortar and heated to 600 °C at a rate of 5 °C/min under N2 flow and kept at this temperature for 2 h in a tubular furnace. After cooling, the black solids were thoroughly rinsed with 2 mol/L HCl solution, followed by deionized water until pH 7 was reached, and then dried at 80 °C in vacuum for 24 h. Finally, the samples were denoted as DKC-600-2 for KOH activation and DZC-600-2 for ZnCl2 activation.

For the second route, bio-pretreatment and activation, the class="Chemical">bioethanol class="Chemical">process of class="Chemical">pan class="Disease">PWS and the main components of the residues can be seen in ref (40), and the bioethanol residues were obtained from the Ministry of Forestry Bioethanol Research Center, Changsha, China. The residues containing waste lignin/polysaccharides and adhering protein were washed with deionized water and then dried at 60 °C overnight to obtain the precursor named BPWS. The BPWS went through activated carbonization by a similar process as above, and the achieved porous carbons were denoted as BKC-600-2 for KOH activation and BZC-600-2 for ZnCl2 activation.

For the third route, hydrothermal pretreatment and activation, typically, 5 g of dried class="Disease">PWS was immersed in 60 mL of deionized class="Chemical">pan class="Chemical">water with 0.3 mL of H2SO4. The mixture was transferred into a 100 mL sealed stainless-steel autoclave with Teflon lining and then heated to 160 °C for 24 h. After cooling, the produced hydrochar (HPWS) was isolated by filtration, washed with deionized water and ethanol several times, and dried at 80 °C in vacuum for 24 h. Next, the HPWS was activated through a similar process as above with different carbonization conditions. The final obtained carbons were denoted as HKC-x-y for KOH activation and HZC-x-y for ZnCl2 activation (x = 600, 700, and 800, representing the carbonization temperature; y = 1 and 2, denoting the mass ratios of activators to HPWS).

Characterization

The thermogravimetric analysis (TGA) and derivative thermogravimetric (DTG) analysis were carried out on a Q600 thermal analysis instrument (American, TA instruments, Inc.). Fourier transform infrared (FT-IR) spectra were collected on a Nicolet 510class="Chemical">P FT-IR sclass="Chemical">pectrometer. The class="Chemical">porosity of samclass="Chemical">ples was obtained by the class="Chemical">pan class="Chemical">N2 adsorption–desorption isotherms at −196 °C using a Micromeritics ASAP2020M+C sorption analyzer. Before adsorption, the sample (about 0.10 g) was degassed at 120 °C for at least 8 h. The surface areas (SBET) were calculated according to the Brunauer–Emmett–Teller (BET) equation in the range of P/P0 = 0.001–0.15, the total pore volume (Vtotal) was determined by adsorption at P/P0 = 0.990, and the pore size distribution (PSD) was estimated by the nonlocal density functional theory (NLDFT) method. The morphology of samples was observed by a field-emission scanning electron microscope (FE-SEM, S4800, Hitachi Ltd., Japan). High-resolution transmission electron microscopy (HRTEM) was conducted on a Tecnai G2 F20 microscope at 200 kV. Elemental analysis (EA) data (CHNS) of the samples were detected using Elementar (Vario EL cube, Germany). X-ray diffraction (XRD) was performed on a Bruker D8 Advance diffractometer with Cu Kα radiation (λ = 1.5418 Å (2θ = 5–80°)). The Raman spectra (LabRAM HR Evolution, France) were obtained in the 200–2100 nm spectral region to evaluate the bonding state.

Gas Adsorption Measurements

The adsorption isotherms of class="Chemical">CO2 and class="Chemical">pan class="Chemical">N2 were measured using a Kubo-X1000 sorption analyzer, China, at 273 and 298–343 K using an ice–water bath and a thermostatic water bath, respectively. Before analysis, the samples were outgassed under vacuum at 120 °C for 12 h to remove guest molecules and then cooled to room temperature, followed by introduction of CO2 or N2 in the pressure range of 0–1 bar.