Enhanced Adsorption of Carbon Dioxide from Simulated Biogas on PEI/MEA-Functionalized Silica

A series of efficient adsorbents were prepared by a wet-impregnation method for CO2 separation from simulated biogas. A type of commercially available silica, named as FNG-II silica (FS), was selected as supports. FS was modified with a mixture of polyethyleneimine (PEI) and ethanolamine (MEA) to improve the initial CO2 adsorption capacity and thermal stability of the adsorbents. The influence of different adsorbents on CO2 adsorption performance was investigated by breakthrough experiments. Scanning electron microscopy (SEM), fourier transform infrared spectroscopy (FTIR), and N2 adsorption–desorption isotherm were used to characterize the silica before and after impregnating amine. Additionally, the thermal stability of adsorbents was measured by differential thermal analysis (TDA). Silica impregnated with mixtures of MEA and PEI showed increased CO2 adsorption performance and high thermal stability compared with those obtained from silica impregnated solely with MEA or PEI. With a simulated biogas flow rate of 100 mL/min at 0.2 MPa and 25 °C, FS-10%MEA-10%PEI exhibited a CO2 adsorption capacity of ca. 64.68 mg/g which increased by 81 % in comparison to FS-20%PEI. The thermal stability of FS-10%MEA-10%PEI was evidently higher than that of FS-20%MEA, and a further improvement of thermal stability was achieved with the increasing value of PEI/MEA weight ratio. It was showed that MEA was able to impose a synergistic effect on the dispersion of PEI in the support, reduce the CO2 diffusion resistance and thus increase CO2 adsorption performance. Additionally, if the total percentage of amine was the same, FS impregnated by different ratios of PEI to MEA did not exhibit an obvious difference in CO2 adsorption performance. FS-15%PEI-5%MEA could be regenerated under mild conditions without obvious loss of CO2 adsorption activity.


Introduction
Carbon dioxide (CO 2 ) released from fossil fuel combustion is considered to be the main cause of the greenhouse effect. Therefore, the use of renewable clean energy to substitute fossil energy has become one of the main approaches in CO 2 emission reduction policies [1][2][3][4][5]. As a clean energy source, biogas consists of 40-75 vol% CH 4 , 25-60 vol% CO 2 , and other compounds (H 2 O, H 2 S, N 2 , O 2 , NH 3 , and siloxanes) [6][7][8]. The presence of a large amount of CO 2 will reduce the calorific value of biogas [8] and increase the cost of transportation and storage, thereby limiting the application of biogas as a renewable energy source. The removal of CO 2 improves the fuel rating [9] and enhances the cost-effectiveness of the biogas. In addition, the separated CO 2 can be used as an alternative raw Int. J. Environ. Res. Public Health 2020, 17, 1452 3 of 12 (Ar) were purchased from Liming Gas Co., Ltd. (Harbin, China). PEI (99%) with a molecular weight (M W ) of 1800 and methanol (99.5%) were purchased from Aladdin Industrial Inc. (Shanghai, China). MEA (≥99.0%) were from Tianjin Fuchen Inc. (Tianjin, China). All chemicals were used without additional treatment.

Preparation of Amine-Functionalized Samples
Before used as the support, the FS was dried at 105 • C until no weight loss was observed. A wet impregnation method was used to prepare the PEI-functionalized samples [31,32]. Amine (pure or mixture of two) were dissolved in 60 mL of methanol and then the dry silica was added into the solution. The amine/methanol/silica mixture was continuously soaked for 12 h, followed by the removal of methanol by rotary evaporation from 50 • C to 70 • C gradually under a vacuum of 150 mmHg. The ratio mass/volume of adsorbents reached ca. 480 g/L. The prepared samples were labeled as FS-X% AMINE, where X% represented the weight percentage of the amine in the samples and AMINE is referred to the name of the used amine.

Characterization of the Adsorbents
Before the following determination, the silica was dried while the amine-functionalized silica was measured without additional pretreatment. Nitrogen adsorption and desorption isotherms were performed at 77 K on a surface area and pore size analyzer apparatus (NOVA 2000e, Quantachrome, Boynton Beach, FL, USA). Before each measurement, the sample was degassed at 453 K for 3 h under a vacuum. The specific surface area was calculated from the isotherm data using the Brunauer-Emmett-Teller (BET) and the total pore volume was determined as the volume of liquid nitrogen adsorbed at a relative pressure of 0.97. Pore diameters were evaluated from the desorption branch of the isotherm based on the Barrett-Joyner-Halenda (BJH) model. The SEM images were obtained using a SU8010 microscope (Hitachi, Tokyo, Japan) equipped with a field emission gun. The acceleration voltage was set to 5 kV. The samples were stuck on the observation platform and sprayed with gold vapor under high vacuum for about 20 s. The FTIR spectra were recorded on an ALPHA-T spectrometer (Bruker, Karlsruhe, Germany) using the KBr compression method. The silica was treated at 378 K for 10 h prior to Infrared spectra (IR) analysis. Thermogravimetric (TG) curves were obtained using a STA449F5 thermal analysis instrument (Netzsch, Selb, Germany) in an N 2 atmosphere.

Adsorption Experiments
Before breakthrough experiments, each amine-functionalized sample was dried to constant weight under 120 • C and the protection of ultrapure Ar to eliminate volatile components. Breakthrough experiments were carried out with a self-assembled fixed bed system (illustrated in Figure 1), which has been reported in our previous work [33]. The testing adsorbent was packed in the adsorption column which was a stainless steel tube with inner diameter of 11 mm and length of 250 mm. In order to desorb the volatiles, the adsorbent was pretreated in advance by heating at 120 • C with an ultrapure Ar stream at a flow rate of 100 mL/min for 30 min. Each adsorption measurement was performed at 0.2 MPa and 25 • C. Prior to adsorption, Ar was introduced into the adsorption column to clear the air in the measurement system and the pressure was increased to 0.2 MPa. After that, a simulated biogas (64.4 vol% CH 4 , 35.6 vol% CO 2 ) stream at a total flow rate of 100 mL/min was introduced and passed through the adsorbent. The outlet concentration was determined by gas chromatography (GC), until the outlet concentration reached the initial feed concentration, which indicated the adsorbent had been saturated by CO 2 . There was no appreciable pressure drop across the adsorption column was observed. where Q is CO2 adsorption capacity (mg/g or mg/cm 3 ); t is the sorption time (min); F is the flow rate (mL/min); Wa is the weight of adsorbent (g); M is the molecular weight (g/mol) of CO2; C0 and Ct are the inlet and outlet concentration (vol%) of CO2, respectively; T represents the gas temperature (K); T0 is 273 K, and Vm is the molar volume (22.4 dm 3 /mol).

Characterization of Materials
The FTIR spectra of FS is depicted in Figure 2. The peak at 3431 cm −1 is attributed to the O-H stretching vibrations of the hydrogen-bonded silanol groups and adsorbed water molecules [34]. The peak located at 1628 cm −1 was corresponding to adsorbed water [35]. Additionally, the peaks at 800 cm −1 and 1107 cm −1 are associated with Si-O and Si-O-Si vibrations [36]. The N2 adsorption and desorption isotherms of pure FS and amine-functionalized samples are given in Figure 3. FS exhibits type-IV isotherms. At the relative pressures P/P0 from 0.45 to 0.95, there is a hysteresis loop which is attributed to the mesopores. The pores distribution curve (BJH) of FS is shown in Figure 4 and its average pore is ca. 8 nm. A shift to larger pores was observed with FSamine adsorbents ( Figure 4). However, the general distribution of the pore size did not change among The adsorption capacity of CO 2 on an adsorbent was calculated as: where Q is CO 2 adsorption capacity (mg/g or mg/cm 3 ); t is the sorption time (min); F is the flow rate (mL/min); W a is the weight of adsorbent (g); M is the molecular weight (g/mol) of CO 2 ; C 0 and C t are the inlet and outlet concentration (vol%) of CO 2 , respectively; T represents the gas temperature (K); T 0 is 273 K, and V m is the molar volume (22.4 dm 3 /mol).

Characterization of Materials
The FTIR spectra of FS is depicted in Figure 2. The peak at 3431 cm −1 is attributed to the O-H stretching vibrations of the hydrogen-bonded silanol groups and adsorbed water molecules [34]. The peak located at 1628 cm −1 was corresponding to adsorbed water [35]. Additionally, the peaks at 800 cm −1 and 1107 cm −1 are associated with Si-O and Si-O-Si vibrations [36]. where Q is CO2 adsorption capacity (mg/g or mg/cm 3 ); t is the sorption time (min); F is the flow rate (mL/min); Wa is the weight of adsorbent (g); M is the molecular weight (g/mol) of CO2; C0 and Ct are the inlet and outlet concentration (vol%) of CO2, respectively; T represents the gas temperature (K); T0 is 273 K, and Vm is the molar volume (22.4 dm 3 /mol).

Characterization of Materials
The FTIR spectra of FS is depicted in Figure 2. The peak at 3431 cm −1 is attributed to the O-H stretching vibrations of the hydrogen-bonded silanol groups and adsorbed water molecules [34]. The peak located at 1628 cm −1 was corresponding to adsorbed water [35]. Additionally, the peaks at 800 cm −1 and 1107 cm −1 are associated with Si-O and Si-O-Si vibrations [36]. The N2 adsorption and desorption isotherms of pure FS and amine-functionalized samples are given in Figure 3. FS exhibits type-IV isotherms. At the relative pressures P/P0 from 0.45 to 0.95, there is a hysteresis loop which is attributed to the mesopores. The pores distribution curve (BJH) of FS is shown in Figure 4 and its average pore is ca. 8 nm. A shift to larger pores was observed with FS- The N 2 adsorption and desorption isotherms of pure FS and amine-functionalized samples are given in Figure 3. FS exhibits type-IV isotherms. At the relative pressures P/P 0 from 0.45 to 0.95, there is a hysteresis loop which is attributed to the mesopores. The pores distribution curve (BJH) of FS is shown in Figure 4 and its average pore is ca. 8 nm. A shift to larger pores was observed with FS-amine adsorbents ( Figure 4). However, the general distribution of the pore size did not change among different FS-amine adsorbents. The presence of these relatively large pores in the adsorbent is beneficial for CO 2 to access the active amino sites.    Textural properties of FS before and after loading amine were listed in Table 1. It can be seen that the BET surface and total pore volumes of amine-samples are all lower than those of raw FS, which indicated amine had been introduced into the silica. Compared with raw FS, the pore volume losses of FS-20%PEI, FS-20%MEA, and FS-10%MEA-10%PEI were 0.254076, 0.235357, and 0.227252 cm 3 /g respectively and their BET surface accordingly decreased 132.5689, 108.355, and 126.3962 m 2 /g respectively. FS-20%PEI showed the largest pore volume and surface area losses which strongly suggests that PEI was very unevenly dispersed in FS. Table 1. Textural properties of the silica before and after impregnating organic amine.    Textural properties of FS before and after loading amine were listed in Table 1. It can be seen that the BET surface and total pore volumes of amine-samples are all lower than those of raw FS, which indicated amine had been introduced into the silica. Compared with raw FS, the pore volume losses of FS-20%PEI, FS-20%MEA, and FS-10%MEA-10%PEI were 0.254076, 0.235357, and 0.227252 cm 3 /g respectively and their BET surface accordingly decreased 132.5689, 108.355, and 126.3962 m 2 /g respectively. FS-20%PEI showed the largest pore volume and surface area losses which strongly suggests that PEI was very unevenly dispersed in FS. Table 1. Textural properties of the silica before and after impregnating organic amine.

BET Surface Area
Total Pore Volume Average Pore Textural properties of FS before and after loading amine were listed in Table 1. It can be seen that the BET surface and total pore volumes of amine-samples are all lower than those of raw FS, which indicated amine had been introduced into the silica. Compared with raw FS, the pore volume losses of FS-20%PEI, FS-20%MEA, and FS-10%MEA-10%PEI were 0.254076, 0.235357, and 0.227252 cm 3 /g respectively and their BET surface accordingly decreased 132.5689, 108.355, and 126.3962 m 2 /g respectively. FS-20%PEI showed the largest pore volume and surface area losses which strongly suggests that PEI was very unevenly dispersed in FS. The scanning electron microscope (SEM) morphology of silica before and after impregnating organic amine was shown in Figure 5. It can be seen from SEM photos magnified 10,000 times that FS-20%MEA and FS-10%MEA-10%PEI still retained the original morphology of the support. However, a partial agglomeration of PEI could be observed on the outer surface, which indicates that PEI was not homogeneously dispersed on the surface of FS. The scanning electron microscope (SEM) morphology of silica before and after impregnating organic amine was shown in Figure 5. It can be seen from SEM photos magnified 10,000 times that FS-20%MEA and FS-10%MEA-10%PEI still retained the original morphology of the support. However, a partial agglomeration of PEI could be observed on the outer surface, which indicates that PEI was not homogeneously dispersed on the surface of FS. The thermal stabilities of FS before and after loading amine were shown in Figure 6. It could be seen that there was a weight loss before 100 °C for FS (black line) caused by the desorption of adsorbed moisture and no significant weight losses appeared after 100 °C in FS. The overall weight decrease for FS appeared to be 4.3% in the range of 25-800 °C. With regard to FS-20%PEI (green line), the first weight loss stage located before 130 °C arose from the desorption of pre-adsorbed H2O and CO2. The second stage in the range 230-380 °C was originated from leaching or decomposition of PEI in the FS, which suggested that the PEI-functionalized FS was stable below these temperatures. For FS-20%MEA (blue line), a large weight decrease (ca. 12%) appeared before 130 °C mainly because of leaching of MEA in the FS, suggesting FS-20%MEA had poor thermal stability. FS-10%MEA-10%PEI (red line) exhibited two TG stages and an overall weight decrease of ca. 21% in the range 25-800 °C. Its first weight loss stage appeared before 130 °C and reached ca. 7%, which was mainly originated from leaching of MEA. The second stage in the range 230-380 °C was originated from leaching or decomposition of PEI in the FS. With regard to FS-5%MEA-15%PEI (magenta line), its first weight loss reached to ca. 5%, lower than that of FS-10%MEA-10%PEI, indicating that the thermal stability of FS-5%MEA-15%PEI was higher than that of FS-10%MEA-10%PEI. The thermal stabilities of FS before and after loading amine were shown in Figure 6. It could be seen that there was a weight loss before 100 • C for FS (black line) caused by the desorption of adsorbed moisture and no significant weight losses appeared after 100 • C in FS. The overall weight decrease for FS appeared to be 4.3% in the range of 25-800 • C. With regard to FS-20%PEI (green line), the first weight loss stage located before 130 • C arose from the desorption of pre-adsorbed H 2 O and CO 2 . The second stage in the range 230-380 • C was originated from leaching or decomposition of PEI in the FS, which suggested that the PEI-functionalized FS was stable below these temperatures. For FS-20%MEA (blue line), a large weight decrease (ca. 12%) appeared before 130 • C mainly because of leaching of MEA in the FS, suggesting FS-20%MEA had poor thermal stability. FS-10%MEA-10%PEI (red line) exhibited two TG stages and an overall weight decrease of ca. 21% in the range 25-800 • C. Its first weight loss stage appeared before 130 • C and reached ca. 7%, which was mainly originated from leaching of MEA. The second stage in the range 230-380 • C was originated from leaching or decomposition of PEI in the FS. With regard to FS-5%MEA-15%PEI (magenta line), its first weight loss reached to ca. 5%, lower than that of FS-10%MEA-10%PEI, indicating that the thermal stability of FS-5%MEA-15%PEI was higher than that of FS-10%MEA-10%PEI.

Separation of Carbon Dioxide from Simulated Biogas
The CH4/CO2 gas separation behaviors of the dynamic experiments were illustrated in Figure 7. It can be seen that FS-20%PEI, FS-20%MEA and FS-10%MEA-10%PEI were all able to separate CO2/CH4 effectively which is indicated by the large differences in the breakthrough times of CO2 and CH4, but the separation ability of the three samples was quite different. For further studies, adsorption capacity of CO2 on the samples are shown in Table 2. It can be observed that the CO2 adsorption capacity of FS-20%PEI was much smaller than that of FS-10%MEA-10%PEI and FS-20%MEA, which were mainly because PEI was very unevenly dispersed in FS and thus amino groups were not easily accessible to CO2. FS-20%PEI showed larger pore volume and surface area losses than FS-10%MEA-10%PEI and FS-20%MEA did (shown in Table 1) compared with their supports (FS), which just indicated that amine was dispersed more unevenly in FS-20%PEI and thus showed much lower CO2 adsorption capacity. The uniformity of the amine dispersion can also be seen in SEM ( Figure 5). While FS-10%MEA-10%PEI exhibited a much higher CO2 adsorption capacity than FS-20%PEI, this was because MEA mixed with PEI could boost the dispersion of PEI and thus facilitated CO2 coming into contact with amino groups. The lower decrease in pore volume and surface area losses of FS-10%MEA-10%PEI can explain this. For FS-20%MEA, its CO2 adsorption capacity was slightly higher than that of FS-10%MEA-10%PEI, but it exhibited significant signs of leaching owing to the low volatility of MEA. As for FS-10%MEA-10%PEI, it was more stable with regard to possible leaching problems. Therefore, FS-10%MEA-10%PEI can overcome the weakness of low CO2 adsorption capacity on FS-20%PEI and conquer the defect of instability of FS-20%MEA.

Separation of Carbon Dioxide from Simulated Biogas
The CH 4 /CO 2 gas separation behaviors of the dynamic experiments were illustrated in Figure 7. It can be seen that FS-20%PEI, FS-20%MEA and FS-10%MEA-10%PEI were all able to separate CO 2 /CH 4 effectively which is indicated by the large differences in the breakthrough times of CO 2 and CH 4 , but the separation ability of the three samples was quite different. For further studies, adsorption capacity of CO 2 on the samples are shown in Table 2. It can be observed that the CO 2 adsorption capacity of FS-20%PEI was much smaller than that of FS-10%MEA-10%PEI and FS-20%MEA, which were mainly because PEI was very unevenly dispersed in FS and thus amino groups were not easily accessible to CO 2 . FS-20%PEI showed larger pore volume and surface area losses than FS-10%MEA-10%PEI and FS-20%MEA did (shown in Table 1) compared with their supports (FS), which just indicated that amine was dispersed more unevenly in FS-20%PEI and thus showed much lower CO 2 adsorption capacity. The uniformity of the amine dispersion can also be seen in SEM ( Figure 5). While FS-10%MEA-10%PEI exhibited a much higher CO 2 adsorption capacity than FS-20%PEI, this was because MEA mixed with PEI could boost the dispersion of PEI and thus facilitated CO 2 coming into contact with amino groups. The lower decrease in pore volume and surface area losses of FS-10%MEA-10%PEI can explain this. For FS-20%MEA, its CO 2 adsorption capacity was slightly higher than that of FS-10%MEA-10%PEI, but it exhibited significant signs of leaching owing to the low volatility of MEA. As for FS-10%MEA-10%PEI, it was more stable with regard to possible leaching problems. Therefore, FS-10%MEA-10%PEI can overcome the weakness of low CO 2 adsorption capacity on FS-20%PEI and conquer the defect of instability of FS-20%MEA.  [39] a T is the adsorption temperature (°C). b P is the total pressure (MPa) in the adsorption column. c qad is the CO2 adsorption capacity (mg/g) of the sample under specific conditions. The effect of the ratio of PEI to MEA on the CO2 adsorption characteristics of PEI/MEA functionalized FS was studied (Figure 8). It is very interesting to note that CO2 adsorption capacity of PEI/MEA impregnated FS wasn't enhanced with the increase of MEA/PEI weight ratio. The similar CO2 adsorption capacity was exhibited although different ratios of MEA to PEI were impregnated into FS, which also suggested that for MEA/PEI functionalized FS, it was not MEA itself that played the most significant role on adsorbing CO2 but a better dispersion of PEI promoted by MEA. This is probably because MEA facilitated PEI to disperse evenly on the surface of FS and thus allowed CO2 to access more easily to the active amino groups. So the PEI/MEA functionalized FS showed much higher CO2 adsorption capacity than FS-20%PEI, even when the weight ratio of MEA was as low as 5%.   [39] a T is the adsorption temperature ( • C). b P is the total pressure (MPa) in the adsorption column. c q ad is the CO 2 adsorption capacity (mg/g) of the sample under specific conditions.
The effect of the ratio of PEI to MEA on the CO 2 adsorption characteristics of PEI/MEA functionalized FS was studied (Figure 8). It is very interesting to note that CO 2 adsorption capacity of PEI/MEA impregnated FS wasn't enhanced with the increase of MEA/PEI weight ratio. The similar CO 2 adsorption capacity was exhibited although different ratios of MEA to PEI were impregnated into FS, which also suggested that for MEA/PEI functionalized FS, it was not MEA itself that played the most significant role on adsorbing CO 2 but a better dispersion of PEI promoted by MEA. This is probably because MEA facilitated PEI to disperse evenly on the surface of FS and thus allowed CO 2 to access more easily to the active amino groups. So the PEI/MEA functionalized FS showed much higher CO 2 adsorption capacity than FS-20%PEI, even when the weight ratio of MEA was as low as 5%.
into FS, which also suggested that for MEA/PEI functionalized FS, it was not MEA itself that played the most significant role on adsorbing CO2 but a better dispersion of PEI promoted by MEA. This is probably because MEA facilitated PEI to disperse evenly on the surface of FS and thus allowed CO2 to access more easily to the active amino groups. So the PEI/MEA functionalized FS showed much higher CO2 adsorption capacity than FS-20%PEI, even when the weight ratio of MEA was as low as 5%. As TG results showed, improvements of thermal stability would be achieved with the increasing value of PEI/MEA weight ratio, so FS-15%PEI-5%MEA was chosen to evaluate the regeneration performance of FS-PEI/MEA adsorbents. The regeneration performance of the FS-15%PEI-5%MEA adsorbent was evaluated by conducting 6 sequential adsorption-desorption cycles (see Figure 9). The adsorption was performed at 25 • C, 0.2 MPa, as well as a simulated biogas flow rate of 100 mL/min and the regeneration was performed at 130 • C, 0.2 MPa as well as an Ar (99.999%) flow rate of 200 mL/min for 15 min. The breakthrough CO 2 adsorption capacities did not significantly change on 1st, 2nd, and 3rd cycles, remaining within the narrow ranges of 59.80-64.17 mmol/g (Figure 8). This also indicates that PEI coated on FS did not degrade significantly at 130 • C in the presence of oxygen. The breakthrough CO 2 adsorption capacities slightly decreased on 4th, 5th and 6th cycles with 55.3-55.54 mmol/g due to accumulation of very small amount of residual CO 2 in adsorbent and this possibly required longer regeneration time in 4th to 5th cycles for the case of high CO 2 concentration feed. However, these results showed that the adsorption performance of FS-15%PEI-5%MEA was stable. Therefore, this adsorbent could be regenerated under mild conditions without obvious loss of CO 2 adsorption activity and so may be suitable for CO 2 removal. As TG results showed, improvements of thermal stability would be achieved with the increasing value of PEI/MEA weight ratio, so FS-15%PEI-5%MEA was chosen to evaluate the regeneration performance of FS-PEI/MEA adsorbents. The regeneration performance of the FS-15%PEI-5%MEA adsorbent was evaluated by conducting 6 sequential adsorption-desorption cycles (see Figure 9). The adsorption was performed at 25 °C, 0.2 MPa, as well as a simulated biogas flow rate of 100 mL/min and the regeneration was performed at 130 °C, 0.2 MPa as well as an Ar (99.999%) flow rate of 200 mL/min for 15 min. The breakthrough CO2 adsorption capacities did not significantly change on 1st, 2 nd , and 3rd cycles, remaining within the narrow ranges of 59.80-64.17 mmol/g (Figure 8). This also indicates that PEI coated on FS did not degrade significantly at 130 °C in the presence of oxygen. The breakthrough CO2 adsorption capacities slightly decreased on 4th, 5th and 6th cycles with 55.3-55.54 mmol/g due to accumulation of very small amount of residual CO2 in adsorbent and this possibly required longer regeneration time in 4th to 5th cycles for the case of high CO2 concentration feed. However, these results showed that the adsorption performance of FS-15%PEI-5%MEA was stable. Therefore, this adsorbent could be regenerated under mild conditions without obvious loss of CO2 adsorption activity and so may be suitable for CO2 removal.

Conclusions
PEI/MEA impregnated FS adsorbents are promising candidates for separating CO2 from biogas. They are easily prepared from widely available and industrially produced materials and are able to

Conclusions
PEI/MEA impregnated FS adsorbents are promising candidates for separating CO 2 from biogas. They are easily prepared from widely available and industrially produced materials and are able to adsorb CO 2 efficiently. The PEI/MEA impregnated FS was found to exhibit much more promoting CO 2 capacities than PEI impregnated FS adsorbents. These new adsorbents presented higher thermal stability than MEA impregnated FS adsorbents. Additionally, similar CO 2 adsorption capacity was exhibited although different ratios of MEA to PEI were impregnated into FS, indicating that MEA could facilitate PEI to disperse evenly on the surface of FS and thus reduce the CO 2 diffusion resistance. These FS-PEI/MEA adsorbents exhibited the attractive CO 2 adsorption capacity of ca. 64.68 mg/g with a simulated biogas flow rate of 100 mL/min at 0.2 MPa and 25°C. An improvement of thermal stability would be achieved with the increasing value of PEI/MEA weight ratio. FS-15%PEI-5%MEA adsorbent could be regenerated at 130 • C, 0.2 MPa and an Ar (99.999%) flow rate of 200 mL/min for 15 min for at least 3 cycles without loss of CO 2 adsorption activity. This approach of impregnating a mixture of MEA and PEI into porous supports may deliver new ideas on preparing amine impregnated porous materials, although further studies are necessary.