1 Introduction

Atmospheric carbon dioxide is the primary carbon source for life on Earth, and its concentration is mainly regulated by photosynthetic organisms, as well as by other natural sources, including volcanoes, hot springs and geysers (Atsumi et al. 2009). CO2 is present in deposits of petroleum and natural gas, and large amounts are produced by the combustion of these and other fossil fuels, which, together with deforestation (removing a natural sink for CO2), has led to a rapidly increasing concentration of CO2 in the atmosphere ever since the Industrial Revolution, accelerating global warming (George et al. 2007; Li 2021). CO2 also causes ocean acidification, because it dissolves in water to form carbonic acid (Kuffner et al. 2008; Zondervan et al. 2001).

Many widely used industrial processes are feasible for the capture of CO2, such as absorption, adsorption, low-temperature distillation, and membrane separation, of which adsorption has the advantages of high removal efficiency, easy operation, and reuse of adsorbent. The pressure swing adsorption (PSA) process is a non-cryogenic gas separation and purification technique, based on the pressure dependence of the adsorption capacity of gases by an adsorbent (Chou and Chen 2004). PSA has been successfully applied to the separation of mixed gases, such as CO2–CH4–N2 (Dong et al. 1999) and HO2–CH4–N2 (Gomes and Yee 2002), as well as to CO2 capture (Rege et al. 2001).

Solid adsorbents for CO2 capture include carbonaceous materials (e.g., porous carbon-based materials such as activated carbon, graphite and biochar, and newly developed carbon-based materials such as graphene, carbon nanotubes, and carbon nanofibers), molecular sieves (e.g., natural and modified clinoptilolites), metal oxides, amino-adsorbents and metal–organic frameworks (MOFs) (Banerjee et al. 2008; D'Alessandro et al. 2010; Kim et al. 2013; Orimo et al. 2003; Titirici et al. 2015; Wilmer and Snurr 2011; Xie et al. 2013; Yang et al. 2021; Zhang et al. 2019; Zhao et al. 2018). Clinoptilolites, which are natural or synthetic crystalline aluminosilicates, are widely used as adsorbents owing to their high abrasion resistance and high thermal stability, as well as their small pore diameters, which provide a high surface area and hence high adsorption capacity. Clinoptilolites have been used to remove various gaseous pollutants (e.g., SO2, NOx, CO, and CO2), as well as radioactive contamination. Previous studies have shown that the adsorption capacity of CO2 by clinoptilolites increases at higher pressures and lower temperatures (Othman et al. 2006). Modification of natural clinoptilolites can remove impurities and effectively improve their ion exchange capacity and adsorption capacity. Roasting facilitates the release of water from clinoptilolites at high temperature, thereby increasing their adsorption capacity (Huo et al. 2013). Roasting modification of clinoptilolites at 400 ℃ was shown to give a specific surface area of 50.99 m2/g and a pore volume of 0.08 cm3/g (Othman et al. 2006). Ion exchange modification was found to increase the CO2 adsorption capacity of clinoptilolites through cationic exchange between the introduced cations (e.g., Na+, Li+, and Ca2+) and the cations present in the clinoptilolites as non-skeletal structure (Baksh et al. 1992).

The work described in the present paper focused on the application of modified clinoptilolites as adsorbents for CO2 capture by PSA. The modification methods applied to the clinoptilolites included roasting, acid pickling, acid pickling–roasting, and ion exchange. Clinoptilolites prepared under the optimal modification conditions were used for CO2 adsorption in a simulated coking exhaust. In addition, regeneration of modified clinoptilolites was investigated to evaluate their renewability and adsorption capacity for CO2.

2 Materials and methods

2.1 Natural clinoptilolite

The natural clinoptilolite used in the present study was sampled from Jinyun City, Zhejiang Province (China), with conchoidal incision, fine consolidation, and compactness. This natural clinoptilolite has a paragenetic composition of quartz, as well as trace amounts of montmorillonite, cristobalite, opal, and chlorite (Kusrini et al. 2019). The chemical composition of the clinoptilolite is shown in Table S1 (in Supplementary Material): its typical unit cell composition was Na6[(AlO2)6(SiO2)30]·24H2O, with an average size of 0.5–2 nm. Other chemicals used in the present study were of analytical pure grade and were purchased from Sinopharm Chemical Reagent Co., Ltd., China.

Scanning electron microscope (SEM) images of clinoptilolite at different magnifications are shown in Fig. S1 (in Supplementary Material). Coarse and irregular framework structures were found, which formed cavities approximately 5 μm in size (Figs. S1c and d: 3000 × and 6000 ×). These aluminum–silicon tetrahedral cavities provided active sites for adsorption and ion exchange, endowing the clinoptilolite with high adsorption capacity for pollutants.

2.2 Modification of clinoptilolite

Natural clinoptilolite of particle size in the range of 0.300–0.391 mm was used for roasting modification. The clinoptilolite was washed several times and placed in an oven at 100 °C for drying. Afterwards, the dried clinoptilolite was transferred into a muffle furnace to be roasted for 1 h, after which the roasting-modified clinoptilolite was cooled to room temperature for further use. The furnace temperature was set variously at 200 °C, 300 °C, 400 °C, 500 °C, and 600 °C to determine the optimal roasting temperature.

Acid pickling modification of natural clinoptilolite (again in the particle size range of 0.300–0.391 mm) was conducted by immersion in aqueous HCl for 40 h. Excessive acid was washed off by distilled water, and the clinoptilolite was dried at 100 °C for further use. Various concentrations of HCl (1 mol/L, 2 mol/L, 4 mol/L, and 6 mol/L) were tested to determine the optimal pickling concentration.

Acid pickling–roasting modification of clinoptilolite was conducted at the optimal roasting temperature determined in the roasting test and with the HCl concentration determined in the acid pickling test. Natural clinoptilolite was ground into different particle size ranges of < 0.150 mm, 0.150–0.187 mm, 0.187–0.300 mm, 0.300–0.391 mm, and 0.391–0.800 mm. Samples of the natural clinoptilolite of each particle size range were immersed in HCl for 40 h, washed to remove excessive acid, and then dried at 100 °C and roasted in the muffle furnace for 1 h. The air-cooled acid pickling–roasting modified clinoptilolite was stored for further use.

The acid pickling–roasting modified clinoptilolite was further modified via ion exchange reaction. It was added to nitrate solutions of four cations (Na+, Cu2+, Mg2+, and Ca2+), and the mixture was placed in a water bath at 85 °C for 24 h. Afterwards, the clinoptilolite was washed, dried at 100 °C, and air-cooled. Various concentrations of each cation were tested (0.1 mol/L, 0.2 mol/L, 0.4 mol/L, and 0.6 mol/L) to determine the optimal combination of concentrations.

2.3 Determination of cation exchange capacity

The cation exchange capacity (CEC) was determined via the ammonium chloride (3%)–ammonium hydroxide (3%) method. Natural or modified clinoptilolite was mixed with 25 mL ammonium chloride–ammonium hydroxide and stood at room temperature for 30 min. Subsequently, 10 drops of polyoxyethylene were added to the mixture, which was then filtered. Excessive ammonium chloride was washed off by ammonium hydroxide until no Cl− could be detected. Then, 25 mL of calcium chloride (12.5%, weight/volume)–formaldehyde (12.5%, volume/volume) mixture was added, and the resulting condensation reaction between ammonium ion and formaldehyde led to the formation of hydrochloric acid, which was titrated by sodium hydroxide using phenolphthalein as indicator. The CEC (mmol/100 g) was calculated as follows:

$${\text{CEC}} = \frac{100C \times V}{m}$$
(1)

where C (mol/L) is the concentration of sodium hydroxide, V (mL) is the volume of sodium hydroxide consumed in the titration, and m (g) is the weight of clinoptilolite.

2.4 Design of PSA equipment for CO2

A schematic of the PSA equipment for CO2 adsorption is shown in Fig. 1. Mixed N2–CO2 gas (58.1%, volume/volume) from the gas tank flows into the adsorption column (20 cm in height and 2 cm in diameter) at a flow rate of 30 mL/min. The CO2 concentration in the outlet gas is determined by the CO2 analyzer (Cyes2) every 2 min, until the outlet CO2 concentration is the same as the inlet concentration.

Fig. 1
figure 1

Schematic diagram of PSA equipment for CO2 adsorption. 1-gas tank with mixture of N2 and CO2; 2-gas flowmeter; 3-adsorption column with modified clinoptilolite; 4-valves; 5-vacuum pump; 6-CO2 analyzer (Cyes2)

Full size image

The CO2 adsorption capacity was used to evaluate the performance of the modified clinoptilolite. First, the volume of CO2 adsorbed, V (m3), was calculated from the breakthrough curve as follows:

$$V=Q{\int }_{{t}_{1}}^{{t}_{2}}\left({C}_{2}-{C}_{1}\right) \text{d}t$$
(2)

where Q (mL/min) is the flow rate of inlet gas, C2 (% volume/v) is the concentration of CO2 at time point t2, and C1 (% v/v) is the concentration of CO2 at time point t1. The CO2 adsorption capacity q (mol/g) under standard conditions was then calculated as follows:

$$q=\frac{1000V}{22.4\times m}$$
(3)

where m is the weight of modified clinoptilolite in the adsorption bed (g) and 22.4 L/mol is the molar volume of gas under standard conditions.

2.5 CO2 desorption and clinoptilolite regeneration

To desorb the CO2, the modified clinoptilolite was first vacuum-treated (with an oilless vacuum pump, hp-01D) at − 0.02 MPa for 50 min, then heated to 300 °C for 1 h, and finally cooled to room temperature. The CO2 re-adsorption capacity was determined after desorption, in order to evaluate the performance of the modified clinoptilolite after regeneration.

2.6 CO2 adsorption in coking exhaust

Acid pickling–roasting–ion (Na+) exchange modified clinoptilolite was used for CO2 adsorption from a gas mixture that simulated the coking exhaust from a coking plant in Hebei Province, China that manufactures metallurgical coke. In this plant, coke oven gas is used as coking fuel, and the combusted coking exhaust is released into the atmosphere. The compositions of the coke oven gas and coking exhaust were determined by gas chromatography (Table 1).

Table 1 Compositions of coke oven gas and coking exhaust
Full size table

The inlet gas simulating the main composition of the coking exhaust consisted of CO2 (12% volume/volume), oxygen (4% volume/volume), and N2 (84% volume/volume). The adsorption pressure was set at 0.4 MPa, and the gas flow rate was 60 mL/min. The adsorption column was filled with 150 g clinoptilolite as the adsorbent. The clinoptilolite was regenerated as described in Sect. 2.5.

2.7 BET test of modified clinoptilolite

The specific surface and pore distribution of the modified clinoptilolite were determined using the Micromeritics Surface Area and Porosimetry System (ASAP2010, USA), with N2 as the adsorbate. The clinoptilolite was dried at 120 °C and then vacuum-treated at 300 °C for 3 h, after which the adsorption capacity was determined at 77 K. The specific surface of the clinoptilolite was calculated according to the Brunauer–Emmett–Teller (BET) equation (Chai et al. 2019).

3 Results

3.1 Optimization of clinoptilolite modification methods

The CEC was used to evaluate the potential of modified clinoptilolite for CO2 adsorption, since it reflects the capabilities of ion exchange and porosity in providing active sites for adsorption reactions (Jayarathne et al. 2019; Wang and Wang 2019). The effects of the different clinoptilolite modification methods on the CEC are illustrated in Fig. 2. Increasing the roasting temperature to 300 °C led to a significant increase in the CEC to 69.1 mmol/L, as water and volatile materials in the clinoptilolite escaped, leading to increased porosity (Fig. 2a). Increasing the temperature further, however, led to a rapid decrease in the CEC, to just 49.4 mmol/L at 600 °C. Therefore, the optimal roasting temperature was set at 300 °C. The roasting time also influenced the CEC (Fig. 2b), and it was found that a time of 1.5 h was optimal in terms of achieving a high CEC (75.9 mmol/L) without destroying the structure of the clinoptilolite.

Fig. 2
figure 2

Influence of different parameters on CEC of modified clinoptilolite a roasting temperature b roasting time c acid concentration d size of clinoptilolite

Full size image

Acid treatment dissolves impurities that block the micropores of clinoptilolite, and pickling in 2 mol/L HCl significantly increased the CEC to 81.9 mmol/L (Fig. 2c). However, higher HCl concentrations led to excessive loss of alumina and destruction of the clinoptilolite structure, with a consequent decrease in the CEC.

For clinoptilolite with particle sizes in the range of 0.300–0.391 mm and modified by a combination of roasting and acid pickling, the CEC of 87.8 mmol/L (Fig. 2d) was significantly increased in comparison with clinoptilolite modified by roasting alone (75.9 mmol/L) or by acid pickling alone (81.9 mmol/L). Although smaller clinoptilolite particle sizes could provide more active reaction sites and larger specific surface area, Fig. 2d shows that further grinding of clinoptilolite to smaller sizes led to only a slight increase in the CEC, and therefore the particle size range of 0.187–0.300 mm was selected for further experimentation.

3.2 Influence of pressure and ion on CO2 adsorption by acid pickling–roasting modified clinoptilolite

The influence of operation pressure on CO2 adsorption was investigated, and the CO2 concentration in the outlet gas is shown versus time for different pressures in Fig. 3. The clinoptilolite was modified by the acid pickling–roasting method. The adsorption capacity of the modified clinoptilolite increased with increasing operation pressure, from 519 mL/g at 0.2 MPa to 730 mL/g at 0.4 MPa. In view of economic considerations, 0.4 MPa was adopted as the optimal operation pressure.

Fig. 3
figure 3

Influence of operation pressure on CO2 adsorption of acid pickling-roasting modified clinoptilolite

Full size image

Four cations, namely, Na+, Cu2+, Mg2+, and Ca2+, each at various concentrations of 0.1 mol/L, 0.2 mol/L, 0.4 mol/L, and 0.6 mol/L, were used for ion exchange modification of natural clinoptilolite. The CO2 adsorption capacity of the modified clinoptilolite is shown in Fig. 4 for the different cation concentrations. The optimal concentrations were found to be 0.4 mol/L Na+ (876.7 mL/g), 0.2 mol/L Cu2+ (753.6 mL/g), 0.4 mol/L Mg2+ (730.9 mL/g), and 0.2 mol/L Ca2+ (566.9 mL/g) (Table 2). Therefore, 0.4 mol/L Na+ was selected for ion exchange modification of clinoptilolite.

Fig. 4
figure 4

Influence of cation concentrations on CO2 adsorption on ion exchange modified clinoptilolite a Na+ b Cu2+ c Mg2+ d Ca2+

Full size image
Table 2 CO2 adsorption capacity after acid pickling-roasting modification
Full size table

3.3 CO2 adsorption by Na+–acid pickling–roasting modified clinoptilolite

A gas mixture with 4% O2 and 12% CO2 was used to simulate coking exhaust, in order to investigate the effect of the presence of O2 on CO2 adsorption (Fig. 5). Up to a time of 130 min, the O2 concentration in the outlet gas remained at 4% and the CO2 concentration at 0.9%, indicating that the Na+–acid pickling–roasting modified clinoptilolite was suitable for CO2 adsorption from coking exhaust containing O2.

Fig. 5
figure 5

CO2 adsorption by Na+-acid pickling-roasting modified clinoptilolite at the presence of O2 (4%)

Full size image

3.4 Regeneration of clinoptilolite by vacuum treatment and heating

Vacuum treatment and heating were combined to desorb the CO2 from the saturated clinoptilolite, in order to evaluate the CO2 adsorption capacity of clinoptilolite after regeneration (Fig. 6). All the modified clinoptilolites tested showed a slightly decreased CO2 adsorption capacity after regeneration. Acid pickling–roasting and Na+–acid pickling–roasting modified clinoptilolites could adsorb 649.8 mL CO2/g and 780.3 mL CO2/g, respectively, after regeneration. These results suggest that clinoptilolite as a material for CO2 adsorption is renewable, and desorption via vacuum treatment and heating is able to recover approximately 89% CO2 adsorption capacity.

Fig. 6
figure 6

CO2 adsorption capacity by regenerated clinoptilolite

Full size image

4 Discussion

This study has demonstrated that the use of modified clinoptilolite as an adsorbent offers enhanced CO2 capture performance. Roasting is a well-established approach for the modification of natural clinoptilolite to improve its adsorption performance. The optimal roasting temperature determined in the present work was 300 °C. Previous studies have shown that roasting at low temperatures (< 300 °C) may lead to incomplete cleaning of the porous surface of natural clinoptilolite, while higher temperatures over 800 °C may destroy the structure of natural clinoptilolite. Thus, roasting at 300 °C could effectively improve the adsorption performance of clinoptilolite without affecting its structure (Figini-Albisetti et al. 2010).

Acid pickling is usually used for the modification of natural clinoptilolite by dealumination. Previous studies have shown that acid pickling can significantly increase both the surface area of clinoptilolite and its Si/Al ratio, thereby enhancing its CO2 adsorption capacity (Lin et al. 2015; Luo et al. 2022; Strejcova et al. 2020; Wahono et al. 2020; Wojciechowska 2019).

In addition, the present study has found that ion exchange modification of clinoptilolite, particularly with Na+, can significantly improve its adsorption capacity for CO2. This is consistent with the results of other recent studies showing that ion exchange modification can achieve the highest removal efficiency for various environmental contaminants, such as fluoride (Nunez et al. 2019), As(V) (Velazquez-Pena et al. 2019), and ammonium and phosphate (Gao et al. 2019), compared with natural clinoptilolite. Ion exchange modification also improves the performance with regard to the purification and separation of valuable gases from mixtures. For example, Hao et al. reported that Na+ (0.2 mol/L)-modified clinoptilolite showed promise as an adsorbent for the separation of low concentrations of CH4 from coal bed gas (Hao et al. 2018). Kennedy et al. found that Ca2+-modified clinoptilolite had great potential for CO2/CH4 separation owing to its high selectivity, while Fe3+-modified clinoptilolite was more suitable for CH4/N2 separation over wider ranges of temperature and pressure (Kennedy et al. 2019).

To unravel the mechanism underlying the enhanced performance of clinoptilolite after acid pickling–roasting modification and Na+-acid pickling-roasting modification, the pore volume, pore size, and specific surface area of natural clinoptilolite, acid pickling–roasting modified clinoptilolite, and Na+–acid pickling–roasting modified clinoptilolite were calculated using the BET method, and the results are shown in Tables 3, 4 and 5. Pickling-roasting and Na+-acid pickling-roasting modification of clinoptilolite both significantly increased the percentage of pore volume of V2 (10–30 nm) (Table 3), compared with that of natural clinoptilolite. Similarly, although the specific surface areas of the modified clinoptilolites decreased slightly, the percentages of specific surface areas of larger pores (> 10 nm) in the modified clinoptilolites increased by 2.3 %-2.4 % compared with those of natural clinoptilolite (Table 4). These results indicate that modification had effectively removed impurities from larger pores, thereby increasing the surface area and providing more active adsorption sites.

Table 3 Distribution of pore volume
Full size table
Table 4 Distribution of specific surface area
Full size table
Table 5 Specific surface area via BET calculation
Full size table

Clinoptilolite serves as a molecular sieve for the adsorption of smaller molecules, and therefore CO2 (diameter 0.33 nm) can theoretically be adsorbed by clinoptilolite. The modified clinoptilolites had significantly increased mean pore sizes for both large pores and micropores (Table 5), which possibly contributed to their improved CO2 adsorption performance. Also, the specific surface areas in larger pores were significantly increased in the modified clinoptilolites (Table 4), indicating that it may be the larger pores that play key roles in CO2 adsorption, rather than the micropores.

X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), Fourier transform infrared (FTIR) spectroscopy, and nuclear magnetic resonance (NMR) could also be employed to investigate the structure of clinoptilolite before and after its modification, to further explore the mechanisms underlying enhanced CO2 adsorption capacity. In many previous studies, no changes have been found after modification processes such as roasting below 300 °C and acid modification (Lin et al. 2015; Nunez et al. 2019). In addition, it should be noted that exchangeable cations in clinoptilolite have a strong attraction for CO2, owing to its quadrupole moment. This is in accordance with the adsorption results obtained here, in which Na+–acid pickling–roasting modified clinoptilolite demonstrated the highest CO2 adsorption capacity.

5 Conclusions

The adsorption of CO2 by modified clinoptilolite was investigated experimentally, and the results showed that modification of clinoptilolite by a combination of acid pickling and roasting or by ion exchange led to the best CO2 adsorption capacity. Clinoptilolite desorption via vacuum treatment and heating enabled recovery of approximately 89% CO2 adsorption capacity. The results of a BET test suggested that modified clinoptilolite had significantly increased mean pore size for both large pores and micropores, possibly improving adsorption capacity. This work provides a renewable, easily obtained, and highly effective approach for CO2 capture and removal, with promise as a tool for the control of air pollution and global warming. Future work should focus on unraveling the mechanisms underlying the improved CO2 capture performance of modified clinoptilolite, using further characterization methods, such as XRD, XPS, FTIR spectroscopy, and NMR, as well as numerical simulations.