Effect of Relative Humidity on Adsorption Breakthrough of CO2 on Activated Carbon Fibers

Microporous activated carbon fibers (ACFs) were developed for CO2 capture based on potassium hydroxide (KOH) activation and tetraethylenepentamine (TEPA) amination. The material properties of the modified ACFs were characterized using several techniques. The adsorption breakthrough curves of CO2 were measured and the effect of relative humidity in the carrier gas was determined. The KOH activation at high temperature generated additional pore networks and the intercalation of metallic K into the carbon matrix, leading to the production of mesopore and micropore volumes and providing access to the active sites in the micropores. However, this treatment also resulted in the loss of nitrogen functionalities. The TEPA amination has successfully introduced nitrogen functionalities onto the fiber surface, but its long-chain structure blocked parts of the micropores and, thus, made the available surface area and pore volume limited. Introduction of the power of time into the Wheeler equation was required to fit the data well. The relative humidity within the studied range had almost no effects on the breakthrough curves. It was expected that the concentration of CO2 was high enough so that the impact on CO2 adsorption capacity lessened due to increased relative humidity.


Introduction
As interests over the impact of the emissions of greenhouse gases on global warming continue to increase, especially the rising concentration of carbon dioxide (CO 2 ) primarily from the use of fossil fuels, it is imperative to alleviate CO 2 emissions by development of pollution prevention technologies. The CO 2 capture and sequestration (CCS) is a set of technologies that can greatly reduce CO 2 emissions and is considered an effective approach in mitigating global warming [1][2][3][4]. A typical untreated flue gas composition from a power plant burning low sulfur eastern bituminous coal consists of approximately 15-16% CO 2 , 70-75% N 2 , 5-7% H 2 O, 3-4% O 2 , and several trace amounts of constituents [5]. Various CO 2 capture technologies, such as absorption, adsorption, cryogenics, and membranes have been widely investigated [4,6]. Although the absorption-regeneration technologies are recognized as the most effective methods and are widely implemented, the high energy density of the absorption processes has restricted their applications. On the other hand, adsorption has been considered as one of the most cost-effective options for CO 2 separation because of the low energy requirement, cost advantage, and ease of applicability over a relatively wide range of temperatures and pressures [7].
The key point to success of adsorption technology depends on the development of the adsorbents with a high CO 2 selectivity and adsorption capacity, easy regeneration, and high stability. Several possible adsorbents, including activated carbon fibers (ACFs) [4], activated carbon [8], zeolite [9], silica adsorbents [10], metal organic frameworks (MOFs) [11], or carbon nanotubes (CNTs) [12], have Their adsorption breakthrough curves for carbon dioxide (CO 2 ) were measured and the effect of relative humidity in the gas stream was discussed.

Modification of ACFs
One commercial polyacrylonitrile (PAN)-based ACFs cloth sample was used as the starting material (AW1107, denoted as ACFC), which was provided by Taiwan Carbon Technology Co. (Taichung, Taiwan). In order to increase the micropore volume, the surface of ACFC was activated using KOH (30603, Sigma-Aldrich, Seelze, Germany). The KOH activation process was initiated by mixing ACFC with KOH in deionized water with a weight ratio of KOH to ACF = 2. The mixing was sonicated for 10 min and then dried for 24 h at 100 • C. Next, the mixture was transferred in a ceramic boat into a tubular furnace for thermal treatment. Activation conditions were 800 • C for 1 h at a heating rate of 10 • C/min under flowing N 2 of 100 sccm. The activated product was washed using hydrochloric acid (HCl, 1 M, Sigma-Aldrich, St. Louis, MO, USA) and deionized water to remove the residual KOH and then dried at 100 • C for 24 h. This sample was denoted as KOH-ACFC.
In addition, the surface of ACFC was immersed with TEPA to decorate the nitrogen functionalities for improving the interactions between CO 2 molecules and the ACFC. The TEPA amination was conducted as follows. About 1.5 g of ACFC sample was impregnated with 21.8 mL of TEPA solution (10 wt % in alcohol). The mixing was sonicated for 90 min at 60 • C and then dried for 24 h at 50 • C. Next, the mixture was positioned in a ceramic boat into a tubular furnace for thermal treatment at 500 • C for 1 h at a heating rate of 10 • C/min under flowing N 2 of 100 sccm. The product was washed with deionized water to remove the residual TEPA and then dried at 100 • C for 24 h. This sample was denoted as TEPA-ACFC.

Characterizations of ACFs
The properties of the samples were characterized using several techniques. The surface morphology of the samples was observed using field emission scanning electron microscopy (FESEM, Hitachi S-4800, Hitachi, Krefeld, Germany). The surface structure of the samples were probed by N 2 adsorption/desorption isotherms measured at −196 • C, carried out using an ASAP 2020 (Micromeritics, Norcross, GA, USA) accelerated surface and porosimetry analysis system. X-ray photoelectron spectroscopy (XPS) was a technique which analyzed the elements constituting the sample surface, its composition, and chemical bonding state by irradiating X-rays on the sample surface, and measuring the kinetic energy of the photoelectrons emitted from the sample surface. The XPS spectra of all samples were obtained using a spectrophotometer (PHI 5000 VersaProbe II, ULVAC-PHI, Kanagawa, Japan), where the scanning X-ray monochromator (Al Anode, hν = 1401 eV) was used and the information on elements within a few nanometers of the sample surface could be obtained. For calibration purposes, the C 1s electron binding energy that corresponds to graphitic carbon was set at 284.6 eV. A nonlinear least squares curve-fitting program (XPSPEAK software, version 4.1, The Chinese University of Hong Kong, Hong Kong, China) was used for deconvolution of the XPS spectra.

Breakthrough Curves of CO 2
The breakthrough curves of CO 2 adsorption on all samples were measured at 25 • C and different relative humidity levels (0, 45, 55, 65, or 75%) according to the ASTM D5160-95 method [33]. The adsorption temperature was set at room temperature to make the measurements available for comparison to studies in the literature. In addition, according to the study of Suzuki [34], the adsorption isotherms of water vapor on ACFs at 25 • C showed minimal adsorption at relative humidity values greater than 40%. The adsorption amounts would go up with an abrupt slope at 60-70% relative humidity and then level off. Thus, this study chose five relative humidity levels covering this interval. The main items in the adsorption system were a single packed adsorbent bed and a humidifier system. The adsorption bed was composed of a stainless steel mesh column with an effective working length of 5 cm, and an outside diameter of 1.5 cm. The ACF sample was wound around the outside of the stainless steel mesh column with two circular layers. The column was in a cylinder reactor with an internal diameter of 3 cm and a height of 15 cm. The weight of the adsorbent was about 1.0-1.5 g. A thermocouple was positioned at the adsorption column to monitor the temperature of the adsorption bed. According to the references [5,35], the typical untreated flue gas compositions from a power plant burning low sulfur coal contained about 15-16% of CO 2 . Thus the CO 2 concentration was set at 15% quantified by gas chromatography with thermal conductivity detector (GC/TCD), which was supplied by the CO 2 and N 2 cylinders under appropriate mixing. The moist gas mixture was obtained by bubbling water using N 2 and the relative humidity was measured by a temperature hygrometer (HC2-S, Rotronic AG, Zurich, Switzerland). The total flow rate of the gas stream was 100 sccm. The CO 2 concentration at the exit was measured using GC/TCD. The Wheeler equation, Equation (1) [18], was used for data fitting to determine the gravimetric capacities (W e ) and the average overall rate coefficients (k v ). Since the original one could not fit the data well, a modified equation was proposed Equation (2): where t was the breakthrough time (min), W e was the equilibrium adsorption amount (g/g), W was the weight of adsorbent (g), C o was the inlet concentration (g/cm 3 ), C was the outlet concentration at time t (g/cm 3 ), Q was the flow rate (cm 3 /min), ρ b was the bulk density of adsorption bed (g/cm 3 ), k v was the overall adsorption rate coefficient (1/min), and n was the power of time. Regeneration of the adsorbents saturated with CO 2 was measured to see if the adsorbents can be reused. After the concentration of CO 2 effluent achieved the inlet concentration, desorption was carried out counter currently under dry N 2 flow at 100 sccm by thermal regeneration at 150 • C until there was no CO 2 detected in the effluent gas. Successive breakthrough experiments were conducted.

Results and Discussion
The FESEM images of the adsorbents are shown in Figure 1. The diameter of a single fiber of ACFC as-received was approximately 6.0 µm and the surface was smooth with typically well-defined striations almost parallel to the fiber axis ( Figure 1a). After KOH activation process, there was no significant change in morphology and the diameters of the fibers were narrowed slightly ( Figure 1b). This could be due to the etching of the carbon framework by the redox reactions between various K-containing compounds with carbonaceous materials, which was responsible for further generating pore network [36]. The TEPA amination also did not make a visible change in the apparent morphology of the fibers (Figure 1c). TEPA was featured with higher amino-groups content and less viscous nature compared to several amines [13,15]. and a humidifier system. The adsorption bed was composed of a stainless steel mesh column with an effective working length of 5 cm, and an outside diameter of 1.5 cm. The ACF sample was wound around the outside of the stainless steel mesh column with two circular layers. The column was in a cylinder reactor with an internal diameter of 3 cm and a height of 15 cm. The weight of the adsorbent was about 1.0-1.5 g. A thermocouple was positioned at the adsorption column to monitor the temperature of the adsorption bed. According to the references [5,35], the typical untreated flue gas compositions from a power plant burning low sulfur coal contained about 15-16% of CO2. Thus the CO2 concentration was set at 15% quantified by gas chromatography with thermal conductivity detector (GC/TCD), which was supplied by the CO2 and N2 cylinders under appropriate mixing. The moist gas mixture was obtained by bubbling water using N2 and the relative humidity was measured by a temperature hygrometer (HC2-S, Rotronic AG, Zurich, Switzerland). The total flow rate of the gas stream was 100 sccm. The CO2 concentration at the exit was measured using GC/TCD. The Wheeler equation, Equation (1) [18], was used for data fitting to determine the gravimetric capacities (We) and the average overall rate coefficients (kv). Since the original one could not fit the data well, a modified equation was proposed (Equation) (2): where t was the breakthrough time (min), We was the equilibrium adsorption amount (g/g), W was the weight of adsorbent (g), C0 was the inlet concentration (g/cm 3 ), C was the outlet concentration at time t (g/cm 3 ), Q was the flow rate (cm 3 /min), ρb was the bulk density of adsorption bed (g/cm 3 ), kv was the overall adsorption rate coefficient (1/min), and n was the power of time. Regeneration of the adsorbents saturated with CO2 was measured to see if the adsorbents can be reused. After the concentration of CO2 effluent achieved the inlet concentration, desorption was carried out counter currently under dry N2 flow at 100 sccm by thermal regeneration at 150 °C until there was no CO2 detected in the effluent gas. Successive breakthrough experiments were conducted.

Results and Discussion
The FESEM images of the adsorbents are shown in Figure 1. The diameter of a single fiber of ACFC as-received was approximately 6.0 μm and the surface was smooth with typically well-defined striations almost parallel to the fiber axis ( Figure 1a). After KOH activation process, there was no significant change in morphology and the diameters of the fibers were narrowed slightly ( Figure 1b). This could be due to the etching of the carbon framework by the redox reactions between various K-containing compounds with carbonaceous materials, which was responsible for further generating pore network [36]. The TEPA amination also did not make a visible change in the apparent morphology of the fibers (Figure 1c). TEPA was featured with higher amino-groups content and less viscous nature compared to several amines [13,15].   Table 1 shows the surface characteristics of the adsorbents, which were determined from the N 2 adsorption/desorption isotherms of the samples at -196 • C. The adsorption isotherms for all samples were essentially type I according to the classification proposed by Brunauer, Emmett, and Teller [37]. Since the c value in Brunauer-Emmett-Teller (BET) surface area report was negative, it implied the BET surface area was invalid. The c value is the BET constant, which is defined as 1 + A/I, where A and I are the slope and the y-intercept in a BET plot. Therefore, the Langmuir model was used to find the specific surface areas in this study. The Langmuir specific surface areas of ACFC, KOH-ACFC and TEPA-ACFC were 1385, 2304, and 1051 m 2 /g, respectively. The total pore volume and micropore volume for all samples have the similar trend. Furthermore, the ratios of micropore volume to total pore volume were 0.80 (ACFC), 0.79 (KOH-ACFC), and 0.73 (TEPA-ACFC). As seen from the data, KOH activation of ACFs generated a lot of mesopore and micropore volumes. Moreover, KOH-ACFC exhibited the highest specific surface area, and its microporosity still retained. Table 1 also recorded the volume of micropores with diameter less than 1 nm (fine pores), which followed the order KOH-ACFC > ACFC > TEPA-ACFC. The fact that KOH-ACFC had the largest fine pore volumes could be due to a series of processes [36,38]: The chemical activation resulted in the generation of pore network, the physical activation contributed to the further development of porosity through carbon gasification, and the intercalation of metallic K into the carbon matrix. The metallic K was very active and mobile at the activation temperature so that the lattices of carbon matrix expanded and leading to microporosity. It was expected that the long chain structure of TEPA molecules could block the original micropores and further decrease the available surface area and pore volume. The ratios of fine pore volume to micropore volume were 0.51 (ACFC), 0.55 (KOH-ACFC), and 0.47 (TEPA-ACFC), respectively. It is believed that KOH activation generated a high fraction of fine pores (<1 nm) [39]. The mean equivalent pore widths from Dubinin-Astakhov (DA) method were approximately 1.6 nm, corresponding to the optimum pore width for CO 2 adsorption in literature [40], though Thiruvenkatachari et al. [41] reported them as 1.8 nm. α Micropore area was determined by Dubinin-Astakhov (DA) method. β Total pore volume (V t ) represents the single point total pore volume at P/P o ∼ = 0.99. γ Micropore volume (V mi ) was determined by DA method. η Mesopore volume (V me ) was found by Barrett-Joyner-Halenda (BJH) method. ϕ Macropore volume (V ma ) was found by subtracting V mi and V me from V t . ξ Micropore volume (<1 nm) was determined by non-local density functional theory (NLDFT) method. ζ Mean equivalent pore width was determined by DA method.
The XPS survey scan spectra of the samples revealed that the major peaks in the scan spectra were due to the C 1s , O 1s , and N 1s photoelectrons. Table 2 summarized the surface atomic contents and atomic ratios. The as-received ACFC contained about 2.43 at.% of N 1s because of the PAN precursor and 8.20 at.% of O 1s . The nitrogen content on KOH-ACFC significantly decreased compared to ACFC, implying the loss of N-moieties during KOH activation at high temperature, which was consistent with the results in several literature [39,[42][43][44]. On the other hand, the TEPA amination of ACFC resulted in a large increase in N 1s contents, although it suffered a heat treatment at 500 • C. It verified that TEPA has been grafted onto the surface of ACFC. In order to understand the chemical bonding states, the deconvolution of high-resolution XPS spectra over the C 1s , O 1s , and N 1s regions for all samples was conducted. Figure 2 illustrates the optimum curve fitting of the high-resolution XPS C 1s spectra for all samples and Table 3 shows the calculated percentages of non-functional and functional carbon atoms. We found that the C 1s spectra were decomposed into at most seven identified components that represented carbon atoms in , and satellite peaks due to π-π* transitions in aromatic rings (B.E. = 291.6 eV) [45]. The C (sp 2 ) and C (sp 3 ) were predominant in C 1s region. An increase in C (sp 2 ) and a decrease in C (sp 3 ) displayed the improvement of graphite-like structure after modification. The -COOH groups were the major functional groups in ACFC. After KOH activation, the -OH groups increased significantly. On the other hand, the TEPA amination resulted in the increase of -OH groups and C=O, which indicated the grafting of TEPA molecules [46]. ACFC resulted in a large increase in N1s contents, although it suffered a heat treatment at 500 °C. It verified that TEPA has been grafted onto the surface of ACFC. In order to understand the chemical bonding states, the deconvolution of high-resolution XPS spectra over the C1s, O1s, and N1s regions for all samples was conducted. Figure 2 illustrates the optimum curve fitting of the high-resolution XPS C1s spectra for all samples and Table 3 shows the calculated percentages of non-functional and functional carbon atoms. We found that the C1s spectra were decomposed into at most seven identified components that represented carbon atoms in , and satellite peaks due to π-π* transitions in aromatic rings (B.E. = 291.6 eV) [45]. The C (sp 2 ) and C (sp 3 ) were predominant in C1s region. An increase in C (sp 2 ) and a decrease in C (sp 3 ) displayed the improvement of graphite-like structure after modification. The -COOH groups were the major functional groups in ACFC. After KOH activation, the -OH groups increased significantly. On the other hand, the TEPA amination resulted in the increase of -OH groups and C=O, which indicated the grafting of TEPA molecules [46].  Deconvolution of XPS O1s peaks gave additional information on the nature of the surface oxygen-containing groups. The optimum curve fitting of the O1s peak for all samples are shown in Figure 3 and the calculated percentages of oxygen-containing functional groups are shown in Table  4. Four different O functionalities and the contribution of chemisorbed water can be identified [45]. The peak at 531.1 eV corresponded to the carbonyl oxygen atoms; the peak at 532.3 eV to the carbonyl oxygen atoms in esters, amides and anhydrides as well as oxygen atoms in hydroxyls or  Deconvolution of XPS O 1s peaks gave additional information on the nature of the surface oxygen-containing groups. The optimum curve fitting of the O 1s peak for all samples are shown in Figure 3 and the calculated percentages of oxygen-containing functional groups are shown in Table 4. Four different O functionalities and the contribution of chemisorbed water can be identified [45]. The peak at 531.1 eV corresponded to the carbonyl oxygen atoms; the peak at 532.3 eV to the carbonyl oxygen atoms in esters, amides and anhydrides as well as oxygen atoms in hydroxyls or ethers; the peak at 533.3 eV to the ether oxygen atoms in esters and anhydrides; and the peak at 534.2 eV to the oxygen atoms in the carboxyl groups. The contribution of chemisorbed water located at 536.1 eV (H 2 O) was found in all samples. As seen from the data, the oxygen atoms in KOH-ACFC were mainly the -OH and C=O groups. However, the chemisorbed water was the primary surface oxide on TEPA-ACFC.
Materials 2017, 10, 1296 7 of 14 ethers; the peak at 533.3 eV to the ether oxygen atoms in esters and anhydrides; and the peak at 534.2 eV to the oxygen atoms in the carboxyl groups. The contribution of chemisorbed water located at 536.1 eV (H2O) was found in all samples. As seen from the data, the oxygen atoms in KOH-ACFC were mainly the -OH and C=O groups. However, the chemisorbed water was the primary surface oxide on TEPA-ACFC.   Figure 4 shows the results of curve-fitting for the high-resolution XPS N1s spectra for all samples, which found that the N1s spectra were decomposed into at most six identified components [47]. The calculated percentages of nitrogen-containing functional groups are shown in Table 5. The peak at 395.7 eV represented the aromatic N-imines; the peak at 398.4 eV was the pyridine-type N; the peak at 400.1 eV showed the pyrrolic or amine moieties; the peak at 401.2 eV was the quarternary N; and the peak at 402.4 eV exhibited the pyridine-N oxides. The contribution of chemisorbed NO2 was located at 405 eV. Except for the chemisorbed NO2, the quarternary N and the pyridine-type N were the predominant N-functionalities on ACFC, while the grafting of TEPA would lead to the increase in pyrrolic or amine moieties which implied the linear TEPA resulted in the increase of pyridine-like structures of six-member or five-member rings. Although KOH activation resulted in the decrease in N content, the major N groups on KOH-ACFC were the pyrrolic or amine moieties and the chemisorbed NO2. It was believed that the pyrrolic or amine moieties were stable compared to the pyridine-type N and quaternary N, and that the KOH activation destroyed the pyridine N structure.     Figure 4 shows the results of curve-fitting for the high-resolution XPS N 1s spectra for all samples, which found that the N 1s spectra were decomposed into at most six identified components [47]. The calculated percentages of nitrogen-containing functional groups are shown in Table 5. The peak at 395.7 eV represented the aromatic N-imines; the peak at 398.4 eV was the pyridine-type N; the peak at 400.1 eV showed the pyrrolic or amine moieties; the peak at 401.2 eV was the quarternary N; and the peak at 402.4 eV exhibited the pyridine-N oxides. The contribution of chemisorbed NO 2 was located at 405 eV. Except for the chemisorbed NO 2 , the quarternary N and the pyridine-type N were the predominant N-functionalities on ACFC, while the grafting of TEPA would lead to the increase in pyrrolic or amine moieties which implied the linear TEPA resulted in the increase of pyridine-like structures of six-member or five-member rings. Although KOH activation resulted in the decrease in N content, the major N groups on KOH-ACFC were the pyrrolic or amine moieties and the chemisorbed NO 2 . It was believed that the pyrrolic or amine moieties were stable compared to the pyridine-type N and quaternary N, and that the KOH activation destroyed the pyridine N structure. ethers; the peak at 533.3 eV to the ether oxygen atoms in esters and anhydrides; and the peak at 534.2 eV to the oxygen atoms in the carboxyl groups. The contribution of chemisorbed water located at 536.1 eV (H2O) was found in all samples. As seen from the data, the oxygen atoms in KOH-ACFC were mainly the -OH and C=O groups. However, the chemisorbed water was the primary surface oxide on TEPA-ACFC.   Figure 4 shows the results of curve-fitting for the high-resolution XPS N1s spectra for all samples, which found that the N1s spectra were decomposed into at most six identified components [47]. The calculated percentages of nitrogen-containing functional groups are shown in Table 5. The peak at 395.7 eV represented the aromatic N-imines; the peak at 398.4 eV was the pyridine-type N; the peak at 400.1 eV showed the pyrrolic or amine moieties; the peak at 401.2 eV was the quarternary N; and the peak at 402.4 eV exhibited the pyridine-N oxides. The contribution of chemisorbed NO2 was located at 405 eV. Except for the chemisorbed NO2, the quarternary N and the pyridine-type N were the predominant N-functionalities on ACFC, while the grafting of TEPA would lead to the increase in pyrrolic or amine moieties which implied the linear TEPA resulted in the increase of pyridine-like structures of six-member or five-member rings. Although KOH activation resulted in the decrease in N content, the major N groups on KOH-ACFC were the pyrrolic or amine moieties and the chemisorbed NO2. It was believed that the pyrrolic or amine moieties were stable compared to the pyridine-type N and quaternary N, and that the KOH activation destroyed the pyridine N structure.   The breakthrough curves of CO 2 (15%) on all adsorbents at 25 • C are sketched in Figure 5. The breakthrough curves obtained with KOH-ACFC were less steep, suggesting a higher pore diffusion resistance than with the other two adsorbents [31], which could be attributed to its large micropore volume. As seen in Figure 5, it implies that the relative humidity within the studied range (45-75%) had almost no effects on the breakthrough curves, which is consistent with the study of Li et al. [48]. It was expected that the interactions between CO 2 molecules and the ACFs were stronger than those between the water vapor molecules and the ACFs such that the competitive adsorption of water vapor on ACFs was insignificant [49,50].  The breakthrough curves of CO2 (15%) on all adsorbents at 25 °C are sketched in Figure 5. The breakthrough curves obtained with KOH-ACFC were less steep, suggesting a higher pore diffusion resistance than with the other two adsorbents [31], which could be attributed to its large micropore volume. As seen in Figure 5, it implies that the relative humidity within the studied range (45%-75%) had almost no effects on the breakthrough curves, which is consistent with the study of Li et al. [48]. It was expected that the interactions between CO2 molecules and the ACFs were stronger than those between the water vapor molecules and the ACFs such that the competitive adsorption of water vapor on ACFs was insignificant [49,50]. Since the original Wheeler equation (Equation (1)) could not fit the data well around the breakthrough point (i.e., C/Co = 0.1) and the We was highly overestimated, another parameter, n (the power of time), was added to the model. To determine an appropriate value of n, the adsorption isotherms of CO2 were required. Figure 6 depicts the adsorption isotherms of CO2 collected on the adsorbents at 25 °C for low CO2 pressures less than 120 kPa, measured using an ASAP 2020 (Micromeritics, Norcross, GA, USA). The temperature during the CO2 adsorption process was Since the original Wheeler equation (Equation (1)) could not fit the data well around the breakthrough point (i.e., C/Co = 0.1) and the W e was highly overestimated, another parameter, n (the power of time), was added to the model. To determine an appropriate value of n, the adsorption isotherms of CO 2 were required. Figure 6 depicts the adsorption isotherms of CO 2 collected on the adsorbents at 25 • C for low CO 2 pressures less than 120 kPa, measured using an ASAP 2020 (Micromeritics, Norcross, GA, USA). The temperature during the CO 2 adsorption process Materials 2017, 10, 1296 9 of 14 was maintained by a circular water bath thermostat. The uptakes of CO 2 had typical monotonic curves, which were increasing with increasing pressures and indicative of the favorable adsorption. The adsorption amounts of CO 2 at 25 • C and 15 kPa followed the order KOH-ACFC (0.855 mmole/g) > ACFC (0.713 mmole/g) > TEPA-ACFC (0.486 mmole/g). These equilibrium data were utilized to calibrate the values of n. After the curve fitting process, it produced n = 0.1 for ACFC, n = 0.08 for KOH-ACFC, and n = 0.001 for TEPA-ACFC, respectively. The reason that impregnating amine did not result in increase of CO 2 uptake could be attributed to the low TEPA loading [13]. Therefore, the increase in CO 2 uptake due to carbamate formation was limited.

C=O R-O-C=O, O=C-NH 2 , O=C-O-C=O, C-OH, R-O-R R-O-C=O, O=C-O-C=O -COOH H 2 O
Materials 2017, 10, 1296 9 of 14 maintained by a circular water bath thermostat. The uptakes of CO2 had typical monotonic curves, which were increasing with increasing pressures and indicative of the favorable adsorption. The adsorption amounts of CO2 at 25 °C and 15 kPa followed the order KOH-ACFC (0.855 mmole/g) > ACFC (0.713 mmole/g) > TEPA-ACFC (0.486 mmole/g). These equilibrium data were utilized to calibrate the values of n. After the curve fitting process, it produced n = 0.1 for ACFC, n = 0.08 for KOH-ACFC, and n = 0.001 for TEPA-ACFC, respectively. The reason that impregnating amine did not result in increase of CO2 uptake could be attributed to the low TEPA loading [13]. Therefore, the increase in CO2 uptake due to carbamate formation was limited.  The above fitted values of n were kept unchanged for each adsorbent. Then the modified Wheeler equation was used further to fit the breakthrough data measured at different relative The above fitted values of n were kept unchanged for each adsorbent. Then the modified Wheeler equation was used further to fit the breakthrough data measured at different relative humidity for each adsorbent. The fitting reports are shown in Table 6. The high values of k v for TEPA-ACFC were responsible for the steep breakthrough curves. The equilibrium adsorption amounts, W e , exhibited a little increase with increasing relative humidity on ACFC and KOH-ACFC. However, the values of k v , the overall adsorption rate coefficient, was decreased in the presence of water vapor. This appeared that the adsorption rate was surface adsorption limited because the active sites were occupied by water vapor molecules or it was pore diffusion limited such that the adsorption rate slowed down and the values of k v decreased [18]. In order to check the recyclability of the adsorbents, cycle tests were performed for these three adsorbents. As shown in Figure 7, it is evident that these three adsorbents can be used up to five adsorption/desorption cycles almost without any changes in breakthrough time and dynamic adsorption capacities. These revealed that all adsorbents were fairly stable in five adsorption/desorption cycles and the heel (the residual adsorbate present in the bed following regeneration) was insignificant. In addition, the stability of the N-functionalities on TEPA-ACFC during thermal regeneration was analyzed. The nitrogen contents for the new TEPA-ACFC sample and the samples after 10 cycle-tests (by thermal regeneration) were probed using XPS. Their breakthrough curves and XPS N1s spectra are compared in Figures 7f and 8, respectively. It can be observed that two patterns are similar. And the loss of nitrogen atoms on TEPA-ACFC after ten adsorption/thermal regenerations processes was estimated about 7%. It is believed that the N-functionalities were grafted and fixed strongly on ACFC. Table 7 compares the CO 2 dynamic adsorption capacities on carbonaceous adsorbents. The KOH-ACFC showed a higher CO 2 uptake compared to those published data in literature under similar operation conditions. humidity for each adsorbent. The fitting reports are shown in Table 6. The high values of kv for TEPA-ACFC were responsible for the steep breakthrough curves. The equilibrium adsorption amounts, We, exhibited a little increase with increasing relative humidity on ACFC and KOH-ACFC. However, the values of kv, the overall adsorption rate coefficient, was decreased in the presence of water vapor. This appeared that the adsorption rate was surface adsorption limited because the active sites were occupied by water vapor molecules or it was pore diffusion limited such that the adsorption rate slowed down and the values of kv decreased [18]. In order to check the recyclability of the adsorbents, cycle tests were performed for these three adsorbents. As shown in Figure 7, it is evident that these three adsorbents can be used up to five adsorption/desorption cycles almost without any changes in breakthrough time and dynamic adsorption capacities. These revealed that all adsorbents were fairly stable in five adsorption/desorption cycles and the heel (the residual adsorbate present in the bed following breakthrough curves and XPS N1s spectra are compared in Figure 7f and Figure 8, respectively. It can be observed that two patterns are similar. And the loss of nitrogen atoms on TEPA-ACFC after ten adsorption/thermal regenerations processes was estimated about 7%. It is believed that the N-functionalities were grafted and fixed strongly on ACFC. Table 7 compares the CO2 dynamic adsorption capacities on carbonaceous adsorbents. The KOH-ACFC showed a higher CO2 uptake compared to those published data in literature under similar operation conditions.  * Co: the inlet concentration of CO2, T: adsorption temperature, and RH: relative humidity.

Conclusions
Results showed that the morphology of the activated carbon fiber cloth samples almost kept intact after modification. The KOH activation at high temperature resulted in smaller diameter of fibers, the loss of nitrogen functionalities, and the increase in surface area and pore volume. The

Conclusions
Results showed that the morphology of the activated carbon fiber cloth samples almost kept intact after modification. The KOH activation at high temperature resulted in smaller diameter of fibers, the loss of nitrogen functionalities, and the increase in surface area and pore volume. The TEPA amination was an effective way to introduce nitrogen onto the surface of fibers, which has been verified using XPS analysis. However, its long chain structure could block the original micropores and further to decrease the available active sites and pore volume. The modified Wheeler equation with a new parameter, the power of time, fitted the data better than the original one, dependent on the nature of the adsorbent. The micropores controlled the steepness of the breakthrough curves, implying the mass transfer was pore diffusion limited. The equilibrium adsorption capacity and adsorption rate of carbon dioxide at a concentration of 15% on the adsorbents were not influenced by the presence of moisture. Although the pore width remains an important factor in determining CO 2 adsorption, the pore volume and surface area should be taken into account for evaluation. This study proposed the equations of breakthrough curves in ACF application for CO 2 capture and the service time could be estimated after the operation conditions are given. Further studies conducted at higher relative humidity levels or at higher adsorption temperatures will be the scope of future experimental work.