Introduction

Global warming has become more severe during this decade. Carbon dioxide (CO2) is a primary reason for global warming and extreme weather. If CO2 concentrations continue to rise, the earth’s temperature could increase by 1.5℃ by 2050. Consequently, natural disasters have become much worse. According to the International Energy Agency, burning fossil fuels is the main cause of CO2 emissions into the atmosphere (99% of the world’s annual CO2 emissions, which add up total about 32 Gt) (Lee and Park 2015). Thus, capturing CO2 from these sources is essential to prevent or delay the increase in CO2 concentration in the air.

Among the currently available CO2 capture technologies, post-combustion capture via the adsorption process is the most widely used. The adsorption process usually uses a solid adsorbent to capture CO2 molecules. Solid adsorbent materials have shown high CO2 adsorption efficiency in published research, such as activated carbon (Wickramaratne and Jaroniec 2013), zeolite 13X (Chen et al. 2014), metal–organic frameworks (Gaikwad et al. 2019), composite material (Papa et al. 2019; Jeong-Potter and Farrauto 2021), bio-based aerogel (Wang and Okubayashi 2019; Mohd et al. 2021), and cellulose-based hybrid aerogel (Jiang et al. 2018; Tang et al. 2019). Among them, cellulose-based hybrid aerogel and composite materials have recently gained more attention because these materials refer to the synergistic combination of multi-scale components that can boost mechanical strength and chemical stability (Gu et al. 2018). The composite materials displayed prominent characteristics such as a large surface area, prolonged mechanical stability, significant adsorption capacity, high selectivity, and sensitivity (Hasan et al. 2021a, b). Therefore, composite materials are potential materials for the treatment of various pollutants, such as heavy metals (Awual et al. 2019a, b; Hasan et al. 2021a, b; Salman et al. 2021), phosphates (Awual 2019), toxic nitrite (Awual et al. 2019a, b), toxic dyes (Islam et al. 2021; Kubra et al. 2021), and CO2 adsorption (Jeong-Potter and Farrauto 2021). However, composite materials still have limitations, such as the requirement of high temperatures during the synthesis process, high material cost, and complicated synthesis procedures (Hasan et al. 2021a, b).

Compared to composite materials, cellulose-based hybrid aerogels have a more straightforward preparation process and are more environmentally friendly due to the advantages of cellulose properties, such as low cost, sustainability, and biodegradability. Additionally, aerogel materials have a much higher porosity than composite materials. For example, the porosity of composite materials was approximately 35.1% for graphite composites (Wang et al. 2022), 57% for geopolymer-hydrotalcite composites (Papa et al. 2019), while the porosity of cellulose-based hybrid aerogel materials was around 90–99%, which was ideal for facilitating air flow and aiding in the capture process (Zheng et al. 2014; Zhou et al. 2019; Kiliyankil et al. 2021). Therefore, cellulose-based hybrid aerogel is a promising adsorbent for CO2 adsorption.

The cellulose-based hybrid aerogel is commonly synthesized by combining the cellulose and silica (Si) components. Cellulose aerogel is low density, high porosity, and flexible but has low thermal stability. In comparison, silica aerogel is exceptionally thermally stable. Hence, the silica component is frequently utilized in producing cellulosic-based hybrid aerogel. Moreover, the cellulose nanomaterials come in a variety of hierarchical forms. Cellulose nanocrystal stands out due to their robust mechanical properties, elasticity, biocompatibility, low density, biodegradability, non-toxic component, and excellent mechanical properties (Kaur et al. 2021). Therefore, the cellulose nanocrystal is a good template for synthesizing hybrid aerogel (Jiang et al. 2018).

Cellulose/Si hybrid aerogel was prepared via a two-step gelation and impregnation method. In this procedure, cellulose component was first dissolved in solvents such as ionic liquid (1-ethyl-3-methylimidazolium acetate)—dimethyl sulfoxide (Demilecamps et al. 2015), and sodium hydroxide-thiourea (Shi et al. 2013). The dissolved cellulose was then regenerated in a non-solvent to form a cellulose hydrogel, which was then immersed in silicon alkoxides. Finally, the impregnated cellulose alcogel was obtained under supercritical or freeze drying. However, the conventional method required a lot of preparation time, and major silica sources from silicon alkoxides are toxic and expensive. Therefore, recent research has been focusing on the one-step sol–gel approach to reduce reaction time and synthesis costs, along with the use of sodium silicate solution as a low-cost and non-hazardous source of silica. For instance, Miao et al. developed a cellulose/Si hybrid aerogel by mixing sodium silicate and cellulose using the one-step sol–gel method. The prepared hybrid aerogel displayed high CO2 adsorption efficiency and selectivity (Miao et al. 2020a, b). Zhou et al. also synthesized a cellulose whisker/Si hybrid aerogel using the one-step sol–gel method. The synthesized hybrid aerogel possessed a multiple function, in which silica content can enhance thermal stability, and cellulose whiskers work as the structural skeleton to strengthen the mechanical properties and improves CO2 capture performance (Zhou et al. 2021).

However, the cellulose/silica hybrid aerogel material has not yet been commercialized. One limitation is the drying process. Drying is an essential step in the formation of the porous structure. The technologies commonly used include freeze-drying and supercritical drying, which can remove the hydrogel liquid phase while maintaining a highly porous solid structure. However, this drying method uses special equipment with extreme drying conditions, consuming great amounts of energy. Therefore, developing aerogel materials under atmospheric drying conditions could be a viable method to commercialize aerogel materials for CO2 adsorption. However, there are few studies on preparing cellulose/silica hybrid aerogel materials under normal drying conditions because this method creates a capillary pressure that collapses the material pore structure. More researches are needed to develop hybrid aerogel materials with improved characteristics under air drying (Li et al. 2019).

Furthermore, improving CO2 adsorption performance is also essential for developing a new adsorbent. Amine-modified cellulose/Si aerogel is a priority strategy to enhance the CO2 capture capacity. There are various amine types that can modify the hybrid aerogel surface. There are two amine kinds, including amine group I and amine group II. Amine group I is loaded onto the adsorbent surface by impregnation through physical binding. In contrast, amine group II will form a covalent bond with the adsorbent (Keshavarz et al. 2021). The impregnation method is commonly used for the preparation of amine-embedded porous materials due to its simplicity and low cost. Linear tetraethylenepentamine (TEPA) and branched polyethyleneimines (PEI) are frequently utilized for physical impregnation because of their high amine content and inexpensive cost. However, using branched PEI with a high percentage of tertiary amines has the significant practical benefit of requiring less energy for CO2 desorption than primary or secondary amines in TEPA (Choi et al. 2021). However, the surface modification of cellulose nanocrystal/silica hybrid aerogels with PEI and the kinetic analysis of CO2 adsorption on this material have yet to be extensively studied. The adsorption kinetic analysis is useful for providing insight about the adsorption rate and adsorption mechanism for the solid adsorbent. Serna-Guerrero and Sayari (2010) investigated kinetic models to predict the CO2 molecule adsorption mechanism onto PEI-impregnated mesoporous silica under various adsorption temperatures. They indicated that multiple adsorption pathways controlled the entire adsorption process (Serna-Guerrero and Sayari 2010). Wang and Okubayashi (2019) applied different adsorption kinetic models to analyze the CO2 adsorption behavior of PEI-cross-linked cellulose aerogel, indicating that the chemical adsorption governed the CO2 adsorption process (Wang and Okubayashi 2019). These studies contributed helpful knowledge of the adsorption mechanism and provided a practical method for forecasting the adsorbent and adsorbate adsorption behavior. Therefore, investigating the adsorption kinetics is vital to elucidate the CO2 adsorption mechanism of PEI-modified cellulose nanocrystal/silica hybrid aerogel.

This study synthesizes and characterizes the cellulose nanocrystal (CNC)/Si hybrid aerogel properties using the one-step sol–gel preparation method along with atmospheric drying. The obtained hybrid aerogel was modified with PEI to investigate the amine modification effect on CO2 adsorption capacity and kinetics. The PEI-modified CNC/Si hybrid aerogel adsorption behavior and diffusion mechanism were analyzed using pseudo-first-order, pseudo-second-order, Avrami fractional-order models, and rate-limiting kinetic models.

Materials and methods

Materials

Microcrystalline cellulose (MCC) was prepared from cotton cloth waste in our previous study using hydrochloric hydrolysis acid (Doan and Chiang 2022), sodium silicate solution with density of 1.39 g mL−1, composition of sodium oxide of 10.6% and silicon dioxide of 26.5% (Sigma-Aldrich), sulfuric acid (H2SO4, 96%) (Aencore company-Australia), hydrochloric acid (HCl, 37%) (Fisher chemical-Canada), trimethylchlorosilane (TMCS, 98%) (Themo scientific), branched polyethyleneimine (PEI, Mw = 800) (Sigma-Aldrich), hexane (99%) (Sigma-Aldrich), and methanol (99.9%) (Sigma-Aldrich).

Cellulose nanocrystal preparation

CNC was prepared from MCC using H2SO4 acid hydrolysis. The MCC was added into 61 wt.% sulfuric acid at a ratio of MCC to acid solution of 1:30 mg L−1. The mixture was heated for 55 min at 50 °C in a water bath circulator. After hydrolysis, the solution was cooled by adding pure water and centrifuged at 5000 rpm until the supernatant became turbid. The obtained precipitate underwent dialysis using dialysis tubing cellulose membrane (molecular weight cutoff of 12–14 kDa) against pure water to remove the residual acid. The dialysis process was completed when the filtrate achieved a neutral pH. The obtained CNC was then homogenized using an ultrasonic bath for 15 min. The CNC suspension was dried at − 80℃ for 24 h using a freeze-dryer (Model FDU-2110, Tokyo Rikakikai Co., Ltd).

CNC/silica aerogel preparation

CSA aerogel was synthesized using one-pot synthesis, as shown in Scheme 1. The sodium silica solution was diluted to a specific gravity of 1.05. Different amounts of CNC suspension at 8 wt.% were added to a given amount of sodium silica solution and vigorously stirred for 2 h under room temperature. The CNC/silica (CNC/Si) mixture was then sonicated for 2 min (sonics Vibra cell, VCX750, 20% Ampl) to form a homogeneous mixture. The weight ratio of CNC suspension to silica was investigated, including 0:1, 1:1, 1.5:1, and 2:1. Afterward, the mixture was hydrolyzed using 1N HCl with an adjusted pH of 5.5. The hydrolyzed CNC/Si mixture was poured into cylindrical molds for gelation at room temperature. After gelation, CNC/silica gels strengthened network integrity by aging for 3 h at 50℃. The gels were then immersed in pure water and kept for 24 h at 50℃. The pure water was replaced each 4 h. The CNC/Si hydrogels were exchanged with methanol for 24 h at 40℃. Furthermore, the CNC/Si gel surface was modified for 20 h at room temperature using a TMCS/n-hexane solution. The molar ratio of TMCS/n-hexane and silica was 1:7 and 4:1, respectively. CSA-(x), where x represents the weight ratio of CNC to silica, was obtained by atmospheric drying at 50℃ for 1 h and 120℃ for 2 h.

Scheme 1
scheme 1

A schematic description of CSA preparation

Full size image

PEI-modified CSA material preparation

PEI was used to modify the CSA-1 surface using the impregnation method. PEI was added into methanol and stirred for 20 min until completely dissolved. A given amount of CSA-1 was then added to the mixture and stirred for 2 h at room temperature. The slurry was dried in a vacuum oven at 80 °C for 24 h after being modified. The obtained materials were named CSA-PEI-y, where y represents the theoretical PEI load weight in the final material. The theoretical PEI load weight ranged from 40 to 60%.

Characterization of CSA and CSA-PEI materials

Dynamic rheological analysis

The dynamic rheological characteristics of CNC/silica gel were determined at 25℃ using an MCR 302 Rheometer. The measuring equipment configuration included a parallel plate measuring plate (PP25) with an angular frequency range of 0.1 to 100 rad/s. Each sample was measured two times.

Morphological structure

The CSA and CSA-PEI surface morphology was visualized using scanning electron microscopy (SEM) (SU-8200, Hitachi). The equipment was operated at 10 kV accelerating voltage and 7.9 mm working distance. High-resolution transmission electron microscopy (HR-TEM) was operated at 80 kV accelerating voltage for determining the morphology of CNC and CSA-1 samples. The elemental composition of the CSA-1 surface was analyzed via energy dispersive X-ray spectroscopy (EDX).

Pore structure analysis

The sample specific area and pore characteristics were measured at 77 K using a nitrogen adsorption–desorption isotherm (Autosorb iQ equipment, Anton Paar). The sample was first degassed at 100℃ for 24 h to remove moisture and volatile components. The Brunauer–Emmett–Teller (BET) method was used to determine the specific area in the relative P/Po pressure range of 0.05–0.3. The Barrett Joyner Halenda (BJH) model was employed to calculate pore size distribution. The total pore volume was calculated based on the amount of adsorbed N2 at a P/Po pressure ratio of 0.98.

Thermal analysis

The sample thermal properties were analyzed via a thermogravimetric analyzer (TGA) (STA 7300, High-Tech Science Corporation) under 100 mL min−1 nitrogen gas and a heating rate of 10℃ min−1. The sample weight of 5–6 g was loaded into a ceramic pan. Each sample was tested twice.

Functional groups analysis

Attenuated total reflectance—Fourier transform infrared spectrometer (ATR-FTIR) from PerkinElmer Frontier spectrometer ((PerkinElmer, Inc.) was used to determine the CSA-PE150 functional groups. The spectra were collected at room temperature with a wavenumber range of 4000 to 650 cm−1 and a spectra resolution of 4 cm−1.

CO2 adsorption

The CSA and CSA-PEI sample CO2 adsorption tests were implemented using the TGA. This method was also used in literature reviews (Jiang et al. 2018; Wang and Okubayashi 2019). About 5 g of the sample was loaded into a platinum sample pan. Before the adsorption test, the sample was heated to 100℃ for 60 min under atmospheric pressure and 100 mL min−1 nitrogen to remove any moisture and impurities content that the sample may have adsorbed from the air. The CO2 adsorption process was carried out by switching the gas flow to pure dry CO2 (100 mL/min, 1 atm) until no further weight increase was observed. CO2 uptake analysis was implemented at different CNC to silica (0:1, 1:1, 1.5:1, and 2:1) blending ratios, temperatures (30, 50, 70, 90, and 120℃), and PEI contents (40, 50, and 60 wt.%). The CO2 adsorption capacity was determined via the weight change of the adsorbent during CO2 measurement (Wang and Okubayashi 2019).

For the CO2 desorption process, the nitrogen gas flow was switched, and the adsorption temperature was heated to 105℃ for 60 min. The CO2 adsorption regeneration capacity was investigated on the CSA-PEI50 under pure dry CO2 at 70℃ with ten cycles of CO2 adsorption–desorption measurement. Amine efficiency was calculated from the CO2 adsorption capacity and nitrogen (N) content. The N content was determined using a Vario Micro cube CHONS elemental analyzer (Elemental Analysensysteme GmbH, Germany). The amine efficiency equation is shown in Eq. (1) (Bai et al. 2019):

$$\mathrm{Amine efficiency} = \frac{{\mathrm{CO}}_{2}\mathrm{ adsorption capacity}}{\mathrm{N content}}$$
(1)

CO2 adsorption kinetic

Adsorption kinetics can provide valuable information on the adsorption rate, establishing the time needed for the adsorption process to reach equilibrium (Keshavarz et al. 2021). Kinetic models can explain details about the adsorption pathways and potential underlying mechanisms. Many kinetic models are applied to explain the adsorption process characteristics. The pseudo-first-order, pseudo-second-order, Avrami, Boyd’s film-diffusion, and intraparticle diffusion kinetic models were applied in this study to describe the dynamic adsorption process. Root mean square error (RMSE) (Eq. (2)) is used to assess the model regression precision. The different kinetic equations are shown in Table 1 as follows.

$$\mathrm{RMSE} = \sqrt{\frac{{\sum }_{1}^{\mathrm{N}}{\left({\mathrm{q}}_{\mathrm{exp}}-{\mathrm{q}}_{\mathrm{cal}}\right)}^{2}}{N}}\times 100\%$$
(2)

where qexp and qcal are the CO2 adsorption capacity of the observed and predicted values using kinetic models, respectively (mmol g−1), and N is the number of observations.

Table 1 Kinetic models for CO2 adsorption process
Full size table

Results and discussions

CNC/Si gel, CSA, and CSA-PEI50 sample characterizations

Rheological property of CNC/Si gel

The CNC amount influence on the mechanical properties of the CNC/Si gel structure was investigated using rheological analysis. Figure 1a shows that the CNC/silica was successfully formed during the aging process. Figure 1b displays that the storage modulus (G’) was higher than the loss modulus (G”) in all samples, indicating that the samples possessed a solid-like structure.

Fig. 1
figure 1

Photographs of CNC/Si mixture and CNC/Si gel after aging process (a), G’ and G” of CNC and CNC/Si gel under various blending ratio between CNC and sodium silicate

Full size image

As shown in Fig. 1b, the G’ of CNC/silica gel increased substantially when the blending amount of CNC was increased from CNC/Si-0 to CNC/Si-1, meaning that the CNC component played a significant role in strengthening the CNC/Si gel structure. CNC contributed to the strong primary skeleton in the gel network. At the same time, the condensed colloidal silica works as a bridge to interconnect the CNC with the gel matrix (Jiang et al. 2018). However, as the amount of CNC components increased, the G’ of CNC/Si-1.5 and CNC/Si-2 decreased gradually, indicating that adding too many CNC components can disrupt the matrix structure and reduce the gel structure mechanical strength. Based on the analysis of gel strength, it can be identified that the weight ratio of CNC to Si of 1:1 (CNC/Si-1) is an optimum concentration for enhancing the skeleton structure in hybrid aerogel materials.

Morphological structure of CSA and CSA-PEI50 samples

SEM results for CSA and CSA-PEI50 are presented in Fig. 2. As shown in Fig. 2a, the CSA-0 possessed a porous three-dimensional structure under atmospheric drying. This finding indicated that the solvent exchange and surface modification overcame the capillary tension and produced a hydrophobic aerogel surface (Khedkar et al. 2019). The hydrophobic property of the CSA sample provides a stable structure against humidity when dried under ambient pressure. Moreover, the CSA structure from CSA-1 to CSA-2 remained a porous network as the CNC content increased. It was not easy to recognize the CNC component in Fig. 2 because the CNC is enclosed in aerogel and embedded deep within the CSA structure (Zhou et al. 2021).

Fig. 2
figure 2

SEM images of CSA-0 (a), CSA-1 (b), CSA-1.5 (c), CSA-2 (d), and CSA-PEI50 (e)

Full size image

These CSA samples contained secondary silica particles that were larger and more evident than CSA-0. These findings can be explained by the rod-shaped CNC (Fig. S1) providing a strong framework for the silica particles to easily attach, resulting in the robust silica aerogel skeleton. The elemental composition analysis of CSA-1 (Fig. S2) using energy dispersive X-ray illustrates that the carbon (C), oxygen (O), and silicon (Si) components were dominant in the sample. On the other hand, the amounts of chloride (Cl) and sodium (Na) were negligible and not detected. These findings indicated that the impure components of Cl and Na were removed entirely during the CSA preparation process. From Fig. 2e, after modifying CSA-1 with 50 wt.% PEI, the CSA-PEI50 surface was entirely flat. Its pore structure was filled by PEI molecules, causing the adsorbent to lose its porous feature (Zhou et al. 2021).

Thermostability analysis

The thermostability properties of CNC, CSA, and CSA-PEI are shown in Fig. 3. For CSA samples, CSA-0 presented an insignificant weight loss about 9% at 800℃. The CSA-0 main degradation temperature occurred at 500–600℃, with a minor weight loss assigned to methyl groups degradation from TMCS surface modification. This result indicated that silica aerogel possessed high thermal stability (Jiang et al. 2018). The thermal behavior was changed when CNC was added to the silica aerogel matrix. CSA-1, CSA-1.5, and CSA-2 samples showed peak degradation in the 325 to 340℃ range, corresponding to the decomposition of cellulose. The sample weight loss increased as the blending ratio between CNC and Si increased from 1 to 2.

Fig. 3
figure 3

CNC, CSA, and CSA-PEI sample thermal properties

Full size image

Although the CSA sample thermal stability minorly reduced with the increase in CNC amount, the thermal behavior of these materials improved significantly compared to the original CNC materials. This result indicated that CNC penetrates the silica matrix and binds tightly to the silica particles (Zhou et al. 2021). After modification, the thermal properties of CSA-PEI were divided into three stages. The first stage was below 180 °C. The sample mass degraded considerably at a temperature around 55℃ due to water evaporation and adsorbed gas in the air by adsorbent materials. As the temperature was raised to 110℃, the CSA-PEI weight loss corresponded to the evaporation of absorbed water or low molecular impurities (Mazrouaa et al. 2019).

The second stage was in the 200 to 430℃ range, which is the main degradation of CSA-PEI samples. The weight loss came mainly from the PEI and CNC component decomposition. The CSA-PEI sample weight loss increased from 40.30 to 44.21% when the PEI concentration was loaded into the CSA-1 sample increased from 40 to 60 wt.%, respectively. The final stage ranged from 430 to 800℃, and the modified sample weight gradually decreased. Furthermore, the more PEI concentration was modified on the CSA surface, the less residual weight it had. These findings indicated that the thermal stability of the PEI content was weak and easy to decompose at high temperatures (Li et al. 2017).

Pore structure analysis

The nitrogen adsorption–desorption isotherms and pore size distribution of CSA samples and CSA-PEI40 are displayed in Fig. 4. As shown in Fig. 4a, all CSA curves correspond to IV isotherms with a type H3 hysteresis loop, which correlated to mesopores and the slit-shape pores (Li et al. 2017). The pore size distribution curves of all CSA samples exhibited one dominant peak (Fig. 4c). The peak position of the CSA-1 sample is distinct from other CSA samples. The curve for CSA-1 displayed a maximum pore size of 6.5 nm, while the different samples had the same peak positions with a maximum pore size of 20 nm. The specific surface area (SBET) of CSA samples (Table 2) indicated that the CNC-added samples decreased the SBET and pore volume. This can be explained by CNC taking up much space and the cross-linking between CNC and secondary silica particles blocking the pore interior (Li et al. 2017).

Fig. 4
figure 4

N2 adsorption–desorption curve of CSA samples (a) and CSA-PEI40 (b). Pore size distribution of CSA samples (c) and CSA-PEI40 (d)

Full size image
Table 2 Textural properties and nitrogen (N) content of CSA and CSA-PEI samples
Full size table

The isotherm analysis of PEI-modified CSA material is also investigated and presented in Fig. 4b. CSA-PEI40 illustrated the type IV isotherm with an H2-type hysteresis, corresponding to a mesoporous material (Salman et al. 2023a, b). However, the volume absorbed by CSA-PEI40 was much lower than that of CSA samples. The type H2 hysteresis loop also demonstrated the multifaceted pore structure with typical ink bottle-shaped pores (Zhou et al. 2021). Previous studies have shown that the ink bottle-shaped pore structure had a higher adsorption capacity than the slit-shaped pore structure (Liu et al. 2020). Therefore, transforming the pore structure from slit-shaped to ink bottle-shaped after modification with PEI may positively enhance the CO2 adsorption capacity of the CSA-PEI materials. However, modification with PEI significantly reduces the specific surface area of the CSA-PEI40 material. This is because the PEI component could fill the pores of CSA material (Bai et al. 2019). The phenomenon was similar to reducing the specific area in composite adsorbents preparation (Hasan et al. 2023; Salman et al. 2023a, b). In addition, the pore structure surface gradually becomes rougher and denser because PEI can occupy the pores (Sanz et al. 2010). Besides, amination on CSA material was confirmed using nitrogen (N) content analysis. As shown in Table 2, N content increased when the PEI concentration increased. A higher N content could establish more covalent bonds with CO2, anticipating greater CO2 adsorption capacity.

CO2 adsorption performance

CNC to silica ratio effect

The CSA sample CO2 adsorption capability with varying proportions of CNC to silica is depicted in Fig. 5. The increase in CNC ratio from CSA-0 to CSA-1 resulted in an increase in the CO2 system capacity. However, once the CNC quantity was increased even further, a tendency toward decreased adsorption capability was seen. The maximum CO2 capacity was found in CSA-1, which measured 0.25 mmol/g. The CSA samples had a high surface area and pore volume, which facilitated CO2 capture and may be associated with physical adsorption (Miao et al. 2020a, b). Moreover, the CSA-1 three-dimensional network structure was more robust and stable than the other materials; the CO2 molecules quickly entered the material network and increased the adsorption capacity (Kiliyankil et al. 2021). In addition, CSA-1 was selected for PEI modification because of its high adsorption capacity and stable material properties. The CO2 adsorption kinetics and adsorption mechanisms of CSA-PEI were further investigated in the following.

Fig. 5
figure 5

CSA-x sample CO2 adsorption capacity at room temperature and atmospheric pressure

Full size image

Temperature adsorption and PEI content effects

The CO2 adsorption capacity on CSA-PEI samples was investigated under adsorption conditions, including different adsorption temperatures and PEI concentrations (Fig. 6). Figure 6a shows the CO2 adsorption capacity on CSA-PEI50 under different adsorption temperatures from 30 to 120℃. The CO2 adsorption at equilibrium increased from 0.9 to 2.61 mmol g−1 when the temperature increased from 30 to 90℃, respectively. This phenomenon was similar to the CO2 capture study using PEI-impregnation resin adsorbent (Bai et al. 2019). It was observed that the CO2 capture performance of CSA-PEI50 samples was substantially enhanced compared with CSA-1 material. This is because the CSA-PEI50 material can adsorb CO2 through physical interactions and chemical interactions between amino groups on CSA-PEI50 and CO2 molecules (Lee and Park 2015). The reaction was illustrated as follows:

Fig. 6
figure 6

CSA-PEI50 CO2 adsorption capacity under various adsorption temperature (a), CSA-and PEI50 sample CO2 adsorption capacity under different PEI contents at 70℃ (b)

Full size image
$${\mathrm{CO}}_{2}\mathrm{+}{\mathrm{2RNH}}_{2}\to {\mathrm{RNHCOO}}^{-}+{\mathrm{RNH}}_{3}^{+}$$
(10)
$${CO}_{2}+{2R}_{2}NH\to {R}_{2}{NCOO}^{-}+{R}_{2}{NH}_{2}^{+}$$
(11)

FTIR spectra were also examined to confirm that the amino groups successfully interacted with CO2 molecules, as shown in Fig. S3. The wavenumber at 1550 cm−1 was considered to be (N)COO-asymmetric stretching (Sepahvand et al. 2020), corresponding to the formation of ammonium carbamates. Moreover, the temperature increase can lead to an increase in the CO2 molecule diffusion rate, which can easily access the amine groups on CSA-PEI50. However, The adsorption capacity displayed a declining trend when the adsorption temperature increased to 120 °C due to the naturally CO2 adsorption exothermic process (Bai et al. 2019). The CSA-PEI50 displayed a high adsorption capacity in the temperature range of 70–90℃, suggesting a prospective adsorbent for CO2 capture from flue gas. Moreover, when the adsorption temperature was increased from 50 to 70℃, the CO2 adsorption capacity increased by 34%. In contrast, the efficiency of CO2 adsorption only increased by 11% when the temperature was raised from 70 to 90℃. Although the maximum adsorption capacity was achieved at 90℃, the adsorption efficiency increase was not impressive. It is essential to consider both the adsorbent material’s cost-effective energy consumption and durability when selecting the optimal adsorption temperature. Therefore, this study identified the optimal adsorption temperature at 70℃ for CO2 capture.

Figure 6b shows the effect of PEI concentrations on CO2 adsorption capacity at 70℃. The CO2 adsorption capacity dramatically rose when the PEI concentrations were increased from 40 to 50 wt.%. The findings can be associated with increasing N content when the PEI concentration increases (as shown in Table 2). It can be confirmed that a higher N content can facilitate the creation of more covalent bonds with CO2, resulting in an increased capacity for CO2 adsorption (Wang and Okubayashi 2019). However, the CO2 adsorption capacity exhibited a decreasing trend when the PEI content was increased to 60 wt.%. This is because the PEI-loaded material pore volume can be blocked with a further increase in PEI concentration, hindering the intraparticle diffusion of CO2 molecules and reducing the access between CO2 and the inner amino groups (Song et al. 2016). Therefore, the PEI concentration of 50 wt% was suitable for impregnation and improving the CO2 adsorption capacity. Therefore, the PEI concentration of 50 wt% was suitable for impregnation and improving the CO2 adsorption capacity.

Furthermore, the CO2 adsorption performance also exhibited by parameter of amine efficiency. The mentioned equation illustrated a maximum theoretical amine efficiency of 0.5 mmol CO2 per mmol N under anhydrous conditions (Sanz et al. 2010). Besides, CSA-PEI50 had the highest amine efficiency of 0.24 mmol CO2/mmol N at 90 °C (Table 3). The result cannot achieve its theoretical value as expected because not all amino functional groups can react with CO2 in practice. The branched PEI contains about 25% of tertiary amines, which cannot react with CO2 under anhydrous conditions (Sanz et al. 2010).

Table 3 CO2 adsorption capacity and amine efficiency of CSA-PEI samples
Full size table

In addition, the CSA-PEI50 displayed a higher CO2 adsorption performance than previous studies in Table 4. Besides, the CSA-PEI50 was easily prepared and produced by a low material cost. In contrast, almost all the cited research used silicon alkoxides as the silica source, which are too expensive and toxic. These aerogel materials were made using a complicated synthesis process along with supercritical or freeze-drying, which consumes more energy and is not feasible for practical applications. Therefore, the CSA-PEI50 in this study may be considered as a cost-effective material and is a potential adsorbent for CO2 adsorption in scale-up studies.

Table 4 CSA-PEI50 CO2 adsorption capacities’ comparison with various aerogel types
Full size table

CO2 adsorption kinetic

Apparent kinetic analysis

Figure 7 shows the CO2 adsorption kinetics of CSA-PEI50 at different adsorption temperatures. At all temperature ranges, the pseudo-first-order and pseudo-second-order models fitted well with the experimental data within the initial 5 min. After this time, these models underestimated the equilibrium amount of CO2 adsorbed. A previous study indicated that the CO2 adsorption process followed pseudo-first-order model, which corresponds to the physical adsorption. The pseudo-second-order model relates to the rate-determining step and is considered a chemical reaction between CO2 molecules and the adsorbent (Song et al. 2016). However, the CO2 adsorption over the amine-modified adsorbent can involve both physical and chemical adsorption (Yang et al. 2022). Hence, it is challenging to predict the CO2 molecule adsorption mechanism on CSA-PEI using the pseudo-first-order and pseudo-second-order kinetic models. The Avrami model provided a good fit for the experimental CO2 uptake during the whole adsorption process with low percent standard deviation values (Table 5).

Fig. 7
figure 7

Kinetic fitting using pseudo-first-order, pseudo-second-order, and Avrami model for CO2 adsorption on CSA-PEI50 and CSA-PEI-y samples under different adsorption temperature (a–e) and different PEI contents (f), respectively

Full size image
Table 5 CSA-PEI sample kinetic parameters under different adsorption kinetic models
Full size table

As shown in Table 5, the RSME of the Avrami model was much smaller than that of the pseudo-first-order and pseudo-second-order models, which was consistent with the relevant data presented in Fig. 7. The CO2 adsorption process can be accurately described by the Avrami model, which represents complex CO2 adsorption pathways (Guo et al. 2019). This result was in line with the reported study, which the authors mentioned that the CO2 adsorption kinetic of the composite aerogel material was well predicted via the Avrami model (Zhou et al. 2021). Furthermore, when the temperature was raised from 30 to 70℃, the kinetic rate constant of the Avrami model increased. The kinetic rate constant remained stable at 90℃ with a further temperature increase. However, when the temperature was elevated to 120℃, the kinetic rate constant increased significantly. The findings indicated that the high temperature had a significant effect on the CO2 transfer rate into adsorbents. However, the desorption process can coincide when the temperature is raised too high, decreasing CO2 adsorption capacity. The adsorption temperature at 70 °C was therefore an excellent choice for the CSA-PEI50 material.

The Avarmi model was also applied to investigate the PEI content effect on CO2 adsorption kinetic at 70 °C (Fig. 7f). From Table 5, the Avrami kinetic constant dramatically increased and then decreased with increasing PEI content. These findings indicated that increasing the PEI content can increase the CO2 molecule accessibility to active amine sites. However, when the PEI concentration increased to 60 wt.%, the kinetic constant decreased significantly because the pore volume was blocked, hindering CO2 diffusion into the amino group positions.

Diffusional mass transfer analysis

The kinetic analysis results indicated that the CO2 adsorption process occurred through multiple pathways. Therefore, the rate-limiting models were investigated to elucidate the CO2 adsorption mechanism. The CO2 molecule adsorption mechanism into a porous adsorbent is as follows: (i) bulk diffusion, where the CO2 molecules are moved from the bulk gas phase into the outer gas film around the porous adsorbent material; (ii) film diffusion, where the CO2 molecules diffuse across the gas film; (iii) interparticle diffusion, where the CO2 molecules diffuse into the inter-particle space; (iv) surface diffusion, where the CO2 molecules interact with the surface sites; and (v) intraparticle diffusion, where the CO2 molecules diffuse from the gas film into the internal surface of the adsorbent (Yoro et al. 2020). According to previous research, the bulk and surface diffusion steps are relatively rapid. Their resistance level in the adsorption process is assumed to be inconsequential (Song et al. 2016). Therefore, the current study focused only on the film, interparticle, and intraparticle diffusion models for CO2 adsorption onto CSA-PEI materials.

Film diffusion is the most crucial stage in any adsorption process because it governs the diffusion of gases into most adsorbents (Yoro et al. 2020). Boyd’s film diffusion plots under different adsorption temperatures and PEI contents are shown in Fig. 8. If the Bt plot versus t is a straight line through the origin, it indicates that only pore diffusion controls the adsorption process (Yoro et al. 2020). However, if the plot is non-linear or linear and does not pass through the origin, the adsorption process can be controlled by film diffusion (Song et al. 2016). As shown in Fig. 8a, b, the Boyd plots illustrate a non-linear trend, indicating that the CO2 adsorption onto CSA-PEI samples is governed by film diffusion.

Fig. 8
figure 8

Boyd’s film model plots, interparticle diffusion and intraparticle diffusion for CO2 adsorption on CSA-PEI50 under different adsorption temperatures (a, c, e), and on CSA-PEI samples under different PEI contents (b, d, f)

Full size image

The interparticle diffusion curves under various adsorption temperatures and PEI contents are shown in Fig. 8c, d. Based on Eq. (8), a plot of ln (1-qt/qe) against time is assumed to be a linear function with an intercept with an intercept that should be equal to ln (6/π2). However, Fig. 8c, d, displays that the plots of ln (1-qt/qe) against time were non-linear, and the intercept of these plots is not similar to ln (6/π2) (as shown in Table 6). These findings indicated violations when assuming the interparticle diffusion model governed the rate-limiting step in CO2 adsorption onto CSA-PEI materials under different adsorption conditions. Consequently, we can assume that there must be another rate-limiting step regulating the CO2 adsorption onto the CSA-PEI materials (Song et al. 2016).

Table 6 Interparticle and intraparticle diffusion model parameters for CSA-PEI under various adsorption temperatures and PEI contents
Full size table

The intraparticle diffusion was examined for the CO2 adsorption rate-limiting process analysis. Figures 8e and 8f illustrate the intraparticle diffusion model via the qt versus square root (sqrt) time plot. The curves obtained under different adsorption temperatures and PEI contents showed multi-linearity. Each linear slope correlates to a natural rate-limiting step. The lowest sloping linear plot can account for the intraparticle diffusion barrier (Song et al. 2016). These findings confirmed that the CO2 adsorption mechanism was governed by different rate-limiting steps.

As a result, a CO2 mechanism was proposed in that the CO2 molecules initially diffused into the CSA-PEI material outer surface. The adsorption rate in this stage is controlled by film diffusion. After passing through the gas film, the CO2 molecules diffused into the CSA-PEI pore network under intraparticle diffusion control and reacted with amino groups. As shown in Fig. 6b, the adsorption process occurs rapidly in the first 15–20 min. At this time, CO2 molecules transferred into the pore and reacted with the inner amino groups, accelerating the CO2 adsorption capacity. However, the CO2 adsorption rate slowed until an equilibrium value was reached. This phenomenon may be related to the CSA-PEI material blocked pore size, which caused a significant diffusion barrier and inhibited CO2 diffusion into the pores to react with amino groups. Therefore, it is proven that the change in CO2 adsorption rate is greatly impacted by the CO2 diffusion process into the pores under intraparticle diffusion.

The diffusion kinetic parameters under different adsorption conditions are presented in Table 6. The diffusion rate constant increased as the adsorption temperature increased because the CO2 molecules tended to be more active at a higher temperature. Consequently, the CO2 diffusion capability into the inner pore surface is enhanced. In contrast to previous reports, which found that the diffusion rate constant decreased with increasing temperature (Song et al. 2016), it is confirmed that CO2 adsorption onto CSA-PEI materials occurs mainly at active sites inside the pore surface. This phenomenon is also contributed to by the change in the PEI physical state from increasing temperature, in which the PEI agglomeration becomes loose and the CO2 molecules easily diffuse through the PEI phase.

CO2 adsorption–desorption cycles

The CO2 adsorption–desorption cycle performed on the hybrid aerogel material is an essential criterion for evaluating the feasibility of a hybrid aerogel-based adsorbent. Figure 9 displays the CO2 capture regeneration on CSA-PEI50 with ten cycles (15,000 min) under atmospheric pressure and temperature conditions at 70℃ for adsorption and 105℃ for desorption processes. The CO2 adsorption capacity slightly decreased after three cycles and became stable with subsequent runs that maintained 83% of the initial value. The reduction of CO2 adsorption effectiveness can cause by a part of the pore structure collapsing and the loss of amino groups during regeneration runs. A previous study also observed a similar phenomenon (Wang and Okubayashi 2019).

Fig. 9
figure 9

Adsorption–desorption curves of CO2 on CSA-PEI50

Full size image

Cyclic stability indicates that CSA-PEI50 is a promising material for CO2 adsorption from flue gas. In addition, the CO2 amount can be removed entirely from CSA-PEI50 during desorption, preventing energy penalties. It can be confirmed that lower energy consumption also leads to lower operation costs and enhances adsorption performance. Moreover, the cost of the adsorption process is substantially influenced by the choice of feedstock sources and the method used to prepare the aerogel. The qualitative comparison of energy and feedstock costs for the synthesis of aerogel materials for the CO2 adsorption process is presented in Table 7. The CSA-PEI50 was prepared using low material costs and consumed less energy. Therefore, the CSA-PEI50 can supply a cost-effective material for CO2 adsorption and generate an economic incentive to proceed with scale-up studies.

Table 7 Comparison of energy consumption and precursor costs of various aerogel materials for CO2 adsorption process
Full size table

Conclusions

A hybrid aerogel was successfully synthesized from the hybridization of CNC and sodium silica by a facile strategy using a one-pot sol–gel method along with atmospheric drying. The textural analysis indicated that all CNC-Si aerogel (CSA) samples illustrated a three-dimensional structure and possessed a nanoscale pore with a high specific surface area. The CSA-1 material showed the highest CO2 adsorption capacity at 0.25 mmol/g at 30℃, 1 atm. The adsorption capacity of CSA-1 was improved by PEI modification. The effect of adsorption temperatures, PEI concentrations, and the adsorption behavior of CSA-PEI materials was investigated. The experimental CO2 adsorption capacity of CSA-PEI50 under different temperatures (30–120℃) and PEI concentrations (40–50 wt%) was highly consistent with the Avrami kinetic model, which has been related to multiple adsorption pathways. The kinetic rate constant and CO2 adsorption capacity increased significantly as the temperature rose from 30 to 70℃. Therefore, the adsorption temperature of 70℃ is a suitable choice for the CO2 adsorption process, with a CO2 adsorption capacity of 2.36 mmol g−1 and a kinetic rate constant of 0.017. The rate-limiting kinetic model was analyzed to elucidate the CO2 adsorption mechanism. The CO2 adsorption mechanism of CSA-PEI50 was governed by film diffusion and intraparticle diffusion resistances. This study aligns with previous research showing the effectiveness of amine-modified aerogel materials for CO2 adsorption. However, the study’s unique contribution is successfully synthesizing of a hybrid aerogel with improved inherent properties by incorporating CNC and Si and elucidating the adsorption mechanism by analyzing the kinetic models. CSA-PEI50 demonstrated excellent recyclability after ten adsorption–desorption cycles, indicating that it is a promising sorbent for CO2 capture from flue gas. Moreover, the factors controlling the CO2 adsorption process on CSA-PEI50 material, such as pressure, humidity, flow rate, and CO2 concentration, need further research for practical implementation.