Discussions on the Adsorption Behaviors Affected by the Differences Between Graphene Oxide and Graphene Grafted by Chitosan

Abstract

There are limited studies in the literature on the surface characterization of modified graphene and graphene oxide and the impact of these modified adsorbents on adsorption performance. In addition, the amine group essentially has a promising affinity for carbon dioxide (CO₂). Therefore, chitosan was used in this study to be grafted onto graphene and graphene oxide respectively. This study examines the effects of graphene, graphene oxide, and chitosan-modified graphene oxide thin films on the removal of carbon dioxide (CO₂). Thin films of graphene, graphene oxide, and their chitosan-modified counterparts were prepared via the methods of precipitation and grafting. The differences in the chemical structure, surface properties, and surface morphology of the films were evaluated, and their effect on the adsorption performance of CO₂ is discussed herein. The micrographs from a scanning electron microscope (SEM) show that the surface of graphene oxide appeared to be more porous than graphene, and the amount of grafted chitosan on graphene oxide is higher than that on graphene. An analysis of atomic force microscope (AFM) finds that the surface of chitosan-modified graphene oxide is rougher than that of chitosan-modified graphene. The results of energy-dispersive X-ray spectroscopy (EDS) spectra reveal that the composition of oxygen in graphene oxide is greater than that in graphene and confirm that the oxygen and nitrogen contents of chitosan-modified adsorbents are greater than those of the pristine materials. An analysis of Fourier-transform infrared spectroscopy (FTIR) shows that most of the oxygen-containing groups are reacted or covered by amide or amine groups due to modification with chitosan. The adsorption isotherms for CO₂ adsorbed by the prepared graphene and graphene oxide presented as type I, indicating great adsorption performance under low pressure. The appropriate amount of chitosan for modifying graphene oxide could be found based on the change in surface area. Although the breakthrough times and the thicknesses of the mass transfer regions for graphene oxide modified with 0.9% and 1.2% chitosan were similar, the modification of graphene oxide with 0.9% chitosan was appropriate in this study due to a significant decrease in surface area with 1.2% chitosan dosage. The adsorption uptake difference between chitosan-modified graphene oxide and graphene was greater than that without modification with chitosan due to more chitosan grafted on graphene oxide. The Toth adsorption isotherm model was used to fit the adsorption uptake, and the average deviation was about 1.36%.

1. Introduction

The methods for capturing CO₂ include absorption, adsorption, low-temperature condensation, and membrane separation. The absorption method was commonly used to capture low-concentration substances released by combustion and was divided into chemical absorption and physical absorption. For chemical absorption, an alkaline absorbent solution could be used to react with CO₂. High-temperature steam must be used to recycle CO₂ due to the stronger molecular bonds of chemical absorption technology. For example, soda lime was modified by silicon dioxide, and the results confirmed that, as the addition of silicon dioxide increased, not only the crushing load strength was improved, but also the CO₂ absorption performance increased [1]. To overcome the high volatility of the ammonia solution and inhibit ammonia desorption, silica particles, including SiO₂ microparticles and colloidal SiO₂ particles, were used as additives to enhance the absorption of CO₂ by the ammonia solutions [2]. The experimental results showed that the absorption performance was significantly enhanced, and the NH₃ desorption was decreased for colloidal SiO₂ particles. On the contrary, the phenomenon that the absorbent solution absorbs carbon dioxide without any chemical reaction is called physical absorption. For example, the solubility of CO₂ in methanol can be increased by adding acetone, resulting in higher absorption performance [3].

To separate and recover CO₂ during the production of syngas, carbon monoxide and water vapor produced by gasification were converted into hydrogen and carbon dioxide, and then, the adsorption method was used to separate CO₂ from the gas phase with solid adsorbents, such as molecular sieve and activated carbon [4,5]. The adsorption method could be divided into pressure swing adsorption and temperature swing adsorption. For pressure swing adsorption (PSA), the adsorption uptake at higher pressure was greater than that at lower pressure, and periodic pressure change could be controlled to achieve gas separation. For instance, type A and type X zeolites were selected to purify 95% of the oxygen in the PSA process [6]. In contrast, the principle of periodically controlling lower temperature adsorption and higher temperature desorption to separate a mixture was called temperature swing adsorption (TSA). To investigate alternatives to CO₂ capture adsorption technologies, a moving bed temperature swing adsorption (MBTSA) system, concerning energy, mass, and momentum balances, was modeled and simulated to discuss the system behavior affected by operating conditions and design parameters [7].

The cryogenic condensation method, already a commercial technology, uses two-stage compression condensation to liquefy or solidify CO₂ into dry ice. Since the compression of the gas must be performed during cryogenic condensation, this method is extremely energy-intensive. Membrane separation methods can also be used for CO₂ capture. To obtain higher purity with less energy during membrane separation, the effects of the operating variables on the membrane separation characteristics were discussed to obtain the optimal operating conditions [8]. In addition, lime (CaO) could be used to absorb CO₂ to form limestone (CaCO₃) during the exothermic carbonation process. The heat released when limestone is burned could serve as a heat source to decompose into CO₂ and lime. One of the studies found that an increase in CaO size was observed upon the sorption of CO₂ and that the improvement in the carbonation reaction was accompanied by the adsorption of water in the pores [9].

Carbon dioxide captured by any capture technology can potentially have applications in dry ice manufacturing [10], food storage [11], carbonated drinks [12,13], fire extinguishers [14], biocides [15], propellants [16], and solvents [17,18]. In addition, CO₂ could also be converted into harmless substances through recombination [19], hydrogenation [20,21], chemical synthesis [22,23], copolymerization [24,25], and photocatalytic reactions [26,27,28]. Some studies have also reported that storage of CO₂ in the ocean was helpful for restoring near-depleted natural gas or oil fields [29,30,31].

Graphene is a single layer of two-dimensional nanomaterial composed of carbon atoms. Graphene was considered to be an adsorption and separation material with good development potential due to the higher specific surface area and stronger surface chemical activity. Therefore, the adsorption of metal ions and organic pollutants by graphene, graphene oxide, and their composites has become an important issue in the open literature. In the case of CO₂ adsorption, graphene prepared by the hydrogen-induced exfoliation of graphitic oxide has been proven to have CO₂ adsorption ability [32]. For liquid-phase adsorption, graphene was used as an adsorbent to remove Rhodamine B (RB) and Malachite Green (MG) from aqueous solutions. The results of adsorption kinetics and thermodynamics for the adsorption of RB and MG by graphene showed that the adsorption isotherms fitted well with the pseudo-second-order model, and the spontaneous adsorption process of these two dyes was discovered [33]. Graphene not only provides a larger surface area but also is used as a photocatalyst to convert CO₂ into hydrocarbon or alcohol molecules. Therefore, the adsorption of CO₂ by a single-layer graphene coating on the SiC surface was studied by temperature-programming desorption and X-ray photoelectron spectroscopy [34]. Since the adsorption or loading of polymer nanoparticles on the graphene surface could improve the consistency, durability, and mechanical resistance, a molecular dynamics simulation of CO₂ adsorption by three-dimensional graphene-polymer composites was performed [35]. In view of the advantages of implanting transition metal compounds to enhance adsorbing capacity and selectivity, graphene sheets embedded with transition metals, such as iron, nickel, and zinc, were used as adsorbents to enhance the adsorption performance of CO₂ [36].

In addition, CTS/GO/ZnO composite materials were developed to improve the surface heterogeneity and increase the surface area of the chitosan carrier, and the results showed that the adsorption capacity of CO₂ by the chitosan carrier was increased [37]. Graphene oxide (GO) is the most important derivative of graphene. Generally speaking, the surface of GO is rich in hydroxyl groups, carboxyl groups, epoxy groups, and carbonyl groups [38], which causes GO to have good dispersion in water and can serve as active sites for adsorbing and modification.

Although there is existing literature discussing the modification and application of graphene and graphene oxide, research related to the simultaneous modification of graphene and graphene oxide using chitosan with an emphasis on surface characterization and CO₂ adsorption performance of the pristine and modified materials is limited. In this study, chitosan was grafted onto graphene and graphene oxide, respectively, due to the excellent affinity between amine groups and CO₂. This work aims to characterize the surface property differences between graphene and graphene oxide thin films and compare those to the chitosan grafted films. Furthermore, the CO₂ adsorption performance affected by the changes in surface properties and morphology of the samples due to grafting was also examined in this study.

2. Experimental Section

2.1. Chemicals

Graphite powders, chitosan powders with a minimum of 90% (w/w) deacetylation, and phosphoric acid 85% (w/w) were purchased from Echo Chemical Co., Ltd., Miaoli, Taiwan. The concentrations of sulfuric acid, H₂O₂ aqueous solution, and HCl were 98% (w/w), 35% (w/w), and 32% (v/v), respectively, and were purchased from Emperor Chemical Co., Ltd., Taipei, Taiwan. Both 99.30% (w/w) sodium hydroxide (NaOH) and 99% (w/w) potassium permanganate (KMnO₄) were reagent grade and purchased from Echo Chemical Co., Ltd. Acetic acid 99.5% (w/w) and hydrazine hydrate 99.9% (w/w) were purchased from Uni-Onward Corp., New Taipei, Taiwan.

2.2. Preparation of Graphene Oxide Thin Film

The improved Hummer method was used to prepare graphite oxide first. Graphite 1.5 g was added into a 200 mL beaker containing 35 mL of 98% (w/w) sulfuric acid. After mixing uniformly, the mixture solution was placed in an ice bath and stirred for 30 min, and 4 g of KMnO₄ were added into the beaker and stirred for 1 h. The mixed solution was then moved to a warm water bath maintained at 40 °C and stirred continuously for 30 min. Distilled water was added into the beaker until the total volume reached 100 mL, and then, 20 mL of an H₂O₂ aqueous solution (10% w/w) were added into the mixture. Before reaching room temperature, the solid and liquid phases were separated by centrifugation. The solid phase was washed with 5% (v/v) HCl and distilled water until nearly neutral. Finally, graphite oxide could be obtained through the processes of filtration and drying at 60 °C. The graphite oxide was ground with a ball mill PM 100 (Retsch Inc., Dusseldorf, Germany) and added into the NaOH solution of pH 11 to form a graphite oxide suspension with a concentration of 500 mg/100 mL. The suspension was vibrated and filtered with an ultrasonic oscillator and centrifuge, respectively, to remove some impurities. Finally, a uniform graphene oxide colloidal suspension could be obtained.

Furthermore, various solutions of chitosan in a 5 M acetic acid aqueous solution were prepared at concentrations of 0.0%, 0.3%, 0.6%, 0.9%, and 1.2% by weight. The solutions were adequately stirred for 1 h to ensure proper mixing and full dissolution of the chitosan. The graphene oxide colloidal suspension was added to this mixture and stirred for another 2 h. Since the active sites, such as carboxyl and carbonyl groups, were distributed on the surface of graphene oxide, the COOH groups were bonded tightly with the surface of graphene oxide. Therefore, the NH₂ groups on the chitosan could be grafted with the COOH groups under an acidic environment. A microporous fiber membrane (pore size: 0.22 μm) was used to filter the colloidal suspension. The alumina carriers with dimensions of 10 cm × 2 cm × 0.3 cm were immersed in the filtered samples at room temperature and then withdrawn at a speed of 1 mm/s. The carriers covered with the filtered samples were calcined at a heating rate of 1 °C/min up to 60 °C. The carriers were continuously heated at 60 °C for another 1 h and then placed at normal condition for 1 h to return to room temperature. These coating steps were repeated 3–5 times to obtain the graphene oxide thin film with a thickness of approximately 0.2 cm. In addition, to conduct the tests in the microbalance system, a blade was used to separate the carrier and the thin film.

2.3. Preparation of Graphene Thin Film

Hydrazine hydrate 0.5 mL was added into the above graphene oxide colloidal suspension, and the temperature was controlled at 90 °C for 10 h of reaction to obtain a graphene colloidal suspension. Similarly, a microporous fiber membrane was used to filter the colloidal suspension. The alumina carriers with dimensions of 10 cm × 2 cm × 0.3 cm were immersed in the filtered samples at room temperature and then withdrawn at a speed of 1 mm/s. The carriers covered with the filtered samples were calcined at a heating rate of 1 °C/min up to 60 °C. The carriers were heated at 60 °C for another 1 h and then placed at normal condition for 1 h to return to room temperature. These coating steps were repeated 3–5 times to obtain the graphene thin film with a thickness of approximately 0.2 cm. Similar to graphene oxide, a blade was used to separate the carrier and the thin film to conduct the tests in the microbalance system.

2.4. Equipment for Thin Film Characterization

A gas chromatograph (GC) equipped with a Porapak-Q column and a thermal conductivity detector (TCD) was used to measure the concentrations of CO₂. A GC-TCD of model HP 5890 was purchased from Mingyu Technol. Com., New Taipei, Taiwan (Manufacturer: Agilent Scientific Ins., Santa Clara, CA, USA). The instruments used for characterization include a surface area analyzer Gemini VII 2390 (Micromeritics Ins. Co., Norcross, GA, USA), a field emission scanning electron microscope, SEM, Phenom Pharos (Thermo Fisher Scientific Inc., Waltham, MA, USA), energy-dispersive X-ray spectroscopy, EDS, X- act 10 mm SDD (Oxford Ins. Plc., Oxford, UK), Fourier-transform infrared spectroscopy, FTIR, Spectrum two (PerkinElmer, Waltham, MA, USA), and an atomic force microscope, AFM, Dimension ICON (Bruker Co., Billerica, MA, USA). The surface area analyzer was used to obtain the pore volume and specific surface area for the prepared adsorbates. Based on the Langmuir isothermal adsorption model, the model was extended to establish the Brunauer–Emmett–Teller (BET) adsorption theory, which assumed that there is a multilayer adsorption phenomenon, and the number of adsorption layers increased with the increase in gas pressure to measure surface and pore properties of the sample. The pretreatment of the sample was to vacuum the sample at a temperature of 150 °C for 2 h to remove moisture and impurities. The specific surface area, pore size, and pore size distribution of the sample could be obtained through nitrogen adsorption experiments on the adsorbent materials. Field emission SEM was used to observe the micromorphology and compare the differences between chitosan-modified graphene and graphene oxide. The pretreatment steps were to stick the carbon glue on the carrier, and then place the sample on the conductive carbon glue. The carrier containing the test sample was vacuumed for 1 h and dried at 120 °C for 8 h to remove moisture and impurities. The dried samples were placed on a vacuum-sputtering machine, plated with gold to a thickness of about 3 nm, and then placed in the SEM for photography. The operating conditions were controlled, with a sputtering time of 90 s, an operating voltage of 10~15 kv, and a magnification of 5000~20,000 times. Elements on the surfaces of materials could be analyzed by EDS, which uses high-energy electromagnetic radiation, such as X-rays or gamma rays, to attack materials, causing electrons inside atoms to dissociate to generate electrons and holes. If electrons in the outer orbit of an atom jump back to the inner orbit, a release of energy will result from a combination of electrons and holes. The energy released could be utilized to perform elemental analysis. The pretreatment steps of EDS are the same as for SEM. The operating conditions were controlled, with a sputtering time of 90 s, an operating voltage of 15 kv, and a magnification of 7500–20,000 times. All the sample surfaces were scanned by the electron beam, and several regions were selected to perform elemental analysis. FTIR can be used to analyze the composition and structure of the prepared samples, where infrared light radiation penetrated the sample, and the sample absorbed radiation of specific frequencies, causing internal vibrations and rotations. Since each functional group has its own specific absorption frequency, the absorption spectrum can provide information about the chemical structure of the prepared samples. The pretreatment steps for FTIR were to uniformly mix 0.3 g KBr and 0.03 g of the prepared sample in a mortar and then compress the mixture into a cake shape in a compactor. The cake was measured and analyzed using the attenuated total reflectance method. The operating conditions were controlled, with a compression pressure of 2000–3500 psi, a compression time of 5 min, a scanning wavelength of 4000–450 cm⁻¹, a resolution of 16 cm⁻¹, and a data interval of 2 cm⁻¹. The atomic interaction between the scanning probe and the test material was used by AFM, causing a tiny displacement of the cantilever beam to measure the surface morphology of the material. AFM was used to observe the surface roughness of the prepared thin film in this study. The prepared sample was examined under a microscope to observe if the sample was clean and smooth. If the sample surface was contaminated or the entire film appeared uneven, the sample must be prepared again. The scan rate and cantilever beam fixing force were set to 1 Hz and 30 nN, respectively.

2.5. Adsorption Tests

2.5.1. Microbalance Adsorption System

The microbalance adsorption system was composed of an electronic microbalance CAHN C-33 (Thermo Fisher Scientific Inc., Massachusetts, United States), gas chromatograph HP 5890 (Agilent Scientific Ins., California, United States), vacuum pump Rocker 300 (Rocker Scientific Co., Ltd., New Taipei, Taiwan), pressure gauge U-68700 (Cole-Parmer Ins. Co., Illinois, United States), and gas supplier (Zhenghong Gas Co., Ltd., Ilan, Taiwan), as shown in Figure 1. The microbalance could be used to detect mass change after adsorption. The accuracy of the microbalance is ±0.1 μg. The pressure was measured by a pressure gauge U-68700 (Cole-Parmer Ins. Co., Illinois, United States) with an accuracy of 0.1333 Pa, and the system pressures were controlled between 0 and 1.445 MPa. The system temperatures were maintained between 25 and 28 °C. The adsorption capacities (mg/g adsorbent) could be calculated from mass change to depict the adsorption isotherms. The steps for measuring the adsorption capacity of carbon dioxide on the adsorbent using the microbalance adsorption system are as follows. After completing the regeneration process, the stripped film of approximately 50 mg was placed on a pan of the microbalance. Carbon dioxide was supplied from a gas cylinder, as shown in Figure 1. The degassing procedure was performed three times. A certain amount of carbon dioxide was released into the system. When the adsorption process reached equilibrium, the mass change and pressure of the system were recorded. Open the valve to allow carbon dioxide to enter the adsorption system again and increase the system pressure. When the adsorption equilibrium was reached again, the second adsorption capacity was obtained. Repeat the above processes to obtain several adsorption capacities until the adsorption capacities reach equilibrium. The adsorption isotherm could be composed of these adsorption capacities.

2.5.2. Thin Film Adsorption System

The advantage of the microbalance adsorption system is that the adsorption isotherm can be obtained with less time and energy consumption compared to the fixed-bed adsorption system. However, the breakthrough curve is difficult to obtain with the microbalance adsorption system. To obtain the breakthrough curves to dynamically discuss adsorption behaviors for carbon dioxide adsorbed by the prepared adsorbents, including graphene, graphene oxide, and the modification of them by chitosan, the adsorption system installed with the staggered alumina carriers coated with graphene, graphene oxide, and modification of them by chitosan was used to adsorb carbon dioxide in this study. The fixed-bed gas separation system was composed of a staggered thin film adsorption unit, an air compressor, a CO₂ cylinder, a gas mixer, and a mass flow controller, as shown in Figure 2. Since the larger pressure drop would be produced from the traditional fixed-bed adsorption system, the alumina carriers coated with graphene and graphene oxide, and the modification of them by chitosan were installed by the staggered method in the adsorption system to adsorb carbon dioxide. The dimensions of the alumina carriers were 10 cm × 2 cm × 0.3 cm in length, width, and thickness. The inlet concentration of carbon dioxide flowing into the staggered adsorption unit was controlled at 5000 ppm, and the volume flow rate for the gas phase was 0.06 L/min. Similar to the tests in the microbalance adsorption system, the temperature was maintained between 25 and 28 °C. The gas phase concentrations flowing into and out of the adsorption system were dynamically recorded to obtain the breakthrough curves. The adsorption capacities for carbon dioxide adsorbed by the alumina carriers coated with graphene and graphene oxide and the modification of them by chitosan were compared in this study. In addition, the removal performance of carbon dioxide affected by characteristic differences resulting from the modification of graphene and graphene oxide by chitosan were also discussed.

3. Results and Discussion

3.1. Adsorption of CO₂

Figure 3 shows the adsorption isotherms for graphene oxide modified by adding different amounts of chitosan. The added amounts of chitosan in this study were 0.0%, 0.3%, 0.6%, 0.9%, and 1.2% (w/w), respectively. As shown in Figure 3, the adsorption isotherms confirmed that the adsorption of carbon dioxide by graphene oxide modified by chitosan belonged to type I. The adsorption capacity was increased with an increase in the amount of chitosan added to the acetic acid solution. However, the adsorption isotherm for the addition of 1.2% chitosan was close to that for the addition of 0.9% chitosan. Therefore, the addition of 0.9% chitosan was chosen to modify graphene oxide in this study.

The adsorption isotherms and breakthrough curves could be obtained through the adsorption operation of the microbalance adsorption system (Figure 1) and the fixed-bed gas separation system (Figure 2), respectively. Figure 4 shows a comparison of the adsorption capacities in this study with data obtained in previous studies. In a prior article, Cavenati et al. used zeolite 13x to absorb CO₂ at room temperature, and the pressures were controlled from 0 to 3.2 MPa [39]. A Rubotherm Magnetic Suspension Balance adsorption measurement system was used as the adsorption system, and the balance can handle a sample mass from 0 to 25 g. The adsorption capacities under pressure from 0 to 1.445 MPa were selected for comparison with this study. Figure 4 shows that the adsorption uptakes in this study are higher than those of reference [39]. Since the data fitted well with Toth and the multisite Langmuir isotherm models were proven, the adsorption of carbon dioxide by zeolite 13Xx and graphene oxide belonged to the Type I adsorption system [40]. In a more recent study by Giraldo et al., the activated carbon prepared from African palm shell was modified with HNO₃ and NH₄OH, and the modified activated carbon was used to investigate the relationship between the chemical properties of activated carbons and CO₂ adsorption capacity [40]. A commercial semi-automatic sortometer, Autosorb IQ2 (Quantachrome Ins., Boynton Beach, United States), was used as the adsorption system. A dosage of 100 mg of activated carbon was used, and the samples were degassed at 423 k for 24 h until the system reached a pressure between 0.1 and 1 Pa. The adsorption uptakes in this research are higher than those of reference [40] due to the higher surface area and effective modification. Moreover, Keramati et al. grafted chitosan and triethylenetetramine onto activated carbon to improve the capacity of CO₂ adsorption [41]. The adsorption system comprises two high-pressure stainless-steel vessels, including pressure and adsorption cells, and their effective volumes were 150 and 108 cm³, respectively. The adsorption data of chitosan-modified activated carbon under a pressure of 0–2000 kPa and a temperature of 303 K were selected for comparison with this study. The adsorption uptakes in this study are higher than those of the literature study [41] due to the higher surface area. Since chitosan-grafted multi-walled carbon nanotubes showed significant catalytic activity towards the chemical fixation of CO₂ with epoxides, multi-walled carbon nanotubes grafted by chitosan were used by Hsan et al. in their study to improve CO₂ capture [42]. A commercial Quantachrome autosorb iQ2 analyzer was used as the adsorption system. The adsorbents 100−125 mg were degassed under vacuum at 150 °C by taking the sample in a 9 mm large cell before the adsorption–desorption measurements. The adsorption data under a pressure of 0–100 kPa and a temperature of 298 K were selected for comparison with this study. The adsorption uptakes in this study are higher due to the lower pressures and smaller surface area for the cited literature. The above not only shows that activated carbon has a strong adsorption effect on CO₂ but also confirms that the design of the adsorption system and the preparation of the adsorbents in this study were successful.

3.2. Adsorbent Characterization

The adsorbents, including graphene and graphene oxide and the modification of them by chitosan, were characterized with a BET sorptometer Gemini VII 2390 (Micromeritics Ins. Co., Norcross, GA, USA) to obtain the surface properties, such as specific surface area, pore volume, and average pore diameter, and the results are shown in

Table 1. Surface and porous properties of the adsorbents.

Property	G 1	GO 2	GO Grafted by Adding 0.3% CTS 3	GO Grafted by Adding 0.6% CTS	GO Grafted by Adding 0.9% CTS	GO Grafted by Adding 1.2% CTS	G Grafted by Adding 0.9% CTS
surface area (m2/g)	856	883	861	843	824	781	801
Total volume (cm3/g)	0.68	0.71	0.69	0.66	0.64	0.58	0.62
Vmicro	0.20	0.21	0.21	0.20	0.19	0.17	0.18
Vmeso	0.38	0.39	0.38	0.36	0.35	0.32	0.34
Vmacro	0.10	0.11	0.10	0.10	0.10	0.09	0.10
Pore diameter (nm)	2.98	3.02	2.45	2.43	2.41	2.37	2.39

¹ G: graphene; ² GO: graphene oxide; ³ CTS: chitosan; confidence limit: 0.5%.

. Use nitrogen to detect pore size, specific surface area, and pore structure. The software ASAP 2024 V.4.00 and DFT V.1.03 were used for control and analysis. All the adsorbents were degassed under a vacuum at 473 ± 0.1 K for 24 h, and then subjected to nitrogen adsorption at 77 ± 0.1 K. All surface areas, total pore volumes, and pore diameters were reduced by chitosan modification, as shown in Table 1. The changes in pore diameter and total volume were decreased with increases in the amount of chitosan added. Although the changes in pore diameter and total volume were decreased moderately as the chitosan addition increased from 0.3% to 0.9%, the changes were more significant when 1.2% chitosan was added, resulting in a significant decrease in surface area. Therefore, the specific surface area decreased with increases in the amount of chitosan added, and the surface area decreased significantly when the added chitosan amount approached 1.2%. When graphene and graphene oxide were grafted with chitosan, the surface area and total volume were decreased with increases in the amount of chitosan added. Although the surface area and total volume of modified graphene oxide are still larger than those of graphene, the difference between them seems to be smaller than that between unmodified graphene and graphene oxide.

The surface morphologies of the prepared graphene and graphene oxide and the modification of them by chitosan were observed and compared using a scanning electron microscope, SEM (Thermo Scientific/Phenom Pharos). Figure 5a,b show SEM micrographs of graphene and graphene oxide, respectively. The surface of graphene oxide appeared to be more porous than graphene due to bonding with oxygen. Figure 5c,d show SEM micrographs of chitosan-modified graphene and graphene oxide. Since the bonded oxygen could be considered an active site, the number of grafted chitosan on graphene oxide seemed to be higher than that on graphene. In addition, the surface carbon atoms were oxidized, resulting in a rougher surface for graphene oxide, as shown in Figure 6a,b. The more active sites, the more chitosan could be grafted. Therefore, Figure 6c,d show that the surface of chitosan-modified graphene oxide is rougher than that of chitosan-modified graphene. The roughnesses in Figure 6a–d are 28.8, 30.2, 33.5, and 36.8 nm, respectively. Elemental analysis was performed by EDS in this study. EDS spectra could be used to compare the elemental compositions for graphene and graphene oxide and the modification of them by chitosan. Figure 7a,b show the EDS spectra of graphene and graphene oxide, and the results reveal that the composition of oxygen in graphene oxide is more than that in graphene. Since hydrazine hydrate was added into a graphene oxide colloidal suspension to obtain graphene, the composition of nitrogen could be found in the graphene. Figure 7c and d show the EDS spectra of chitosan-modified graphene and graphene oxide, respectively. A comparison of Figure 7c,d with Figure 7a,b confirms that the contents of oxygen and nitrogen in Figure 7c,d are higher than those in Figure 7a,b due to the modification of graphene and graphene oxide by chitosan. The results also proved that chitosan was loaded on graphene and graphene oxide successfully.

FTIR could be used to characterize the functional groups of the prepared adsorbents, as shown in Figure 8. Due to the oxidation of graphene, some functional groups were generated in the graphene oxide thin films, resulting in a peak corresponding to the –OH stretching vibrations appearing at about 3422 cm⁻¹, a peak corresponding to the carboxyl C=O stretching vibrations appearing at about 1725 cm⁻¹, a peak corresponding to the C=C group at 1622 cm⁻¹, the peaks corresponding to the C–OH group at 1402 cm⁻¹ and 1322 cm⁻¹, a peak corresponding to the alkoxy C–O–C group at 1037 cm⁻¹, and a peak corresponding to the hydroxy –OH group at 587 cm⁻¹, as shown in Figure 8a. When graphene oxide was reduced by hydrazine hydrate to obtain graphene, most of the peaks weakened or disappeared, except for the peak around 1628 cm⁻¹ corresponding to the C=C groups. Therefore, the main functional groups of graphene included C=C, C-O, and -OH groups after the reduction process. The FTIR spectra of chitosan-modified graphene and graphene oxide are shown in Figure 8a. Since the –OH group and carboxyl group resulted from acid treatment, in graphene oxide and graphene, the peaks of graphene/chitosan and graphene oxide/chitosan at about 3433 cm⁻¹ and 3440 cm⁻¹ could be attributed to stretching vibration and intermolecular hydrogen bonding between the -NH₂ and -OH groups. In addition, a peak of the spectrum of graphene oxide/chitosan corresponding to the N-H stretching vibration of the amino group was located at about 3455 cm⁻¹. Since graphene oxide was modified with chitosan, most of the oxygen-containing groups were reacted or covered by amide or amine groups. Therefore, the dominant peaks of the GO–chitosan spectrum include the peaks at 1597 cm⁻¹ and 1035 cm⁻¹, corresponding to the stretching vibrations from C=O of -NHCO- and N-H of NH₂, and the peaks at 1655 cm⁻¹ and 1378 cm⁻¹, corresponding to the C=O stretching vibrations of Amide I and the CH₂ wagging vibrations (Amide III), as shown in Figure 8b. Compared with the peak at 587 cm⁻¹ in Figure 8a, the peak at 593 cm⁻¹ in Figure 8b was weaker, which could be attributed to that part of the -OH groups that were reacted with chitosan. Although chitosan was 90% deacetylated, there are probably still enough NHCOCH₃ groups on the molecule that the FTIR could pick up. Therefore, a peak forming between 2850 and 3000 cm⁻¹ corresponds to sp3 C-H groups.

In the FTIR spectrum of graphene/chitosan, the peak at 3433 cm⁻¹ corresponds to the stretching vibrations of N-H, which also overlaps with the unreacted hydroxyl group. Similar to the modification of graphene oxide by chitosan, the peaks also presented for the functional groups derived from amines, including the C=O stretching vibration of amide, the stretching vibration from C=O of -NHCO-, the CH₂ wagging vibrations from Amide III, and the N-H bending of NH₂, located at 1660 cm⁻¹, 1598 cm⁻¹, 1378 cm⁻¹, and 1040 cm⁻¹, respectively. However, the intensities of the peaks were slightly weaker than those of chitosan-modified graphene oxide.

3.3. Adsorption Performance for the Prepared Adsorbents

To examine the adsorption behavior of CO₂, the alumina carriers coated with the prepared adsorbents, including graphene oxide and graphene and the modification of them by chitosan, were stalled in a staggered manner in the adsorption system to obtain the breakthrough curves. These adsorbents were used to adsorb CO₂ in this study. Generally speaking, the longer the breakthrough time and the steeper the breakthrough curve in the mass transfer region, the better the adsorption effect between the adsorbent and the adsorbate. Figure 9 shows the breakthrough curves of CO₂ adsorbed by graphene oxide modified by adding 0.0%, 0.3%, 0.6%, 0.9%, and 1.2% chitosan to an acetic acid solution. Since chitosan was helpful for the adsorption of CO₂, the adsorption performance for graphene oxide modified by chitosan would be better than that without modification. In addition, the adsorption performance was increased with increases in the added amount of chitosan. However, the breakthrough times and the thicknesses of the mass transfer region were similar for graphene oxide modified by adding 0.9% and 1.2% chitosan. This result indicated that the dosage of 0.9% chitosan used in this study to modify graphene oxide was optimal due to a significant decrease in surface area for the dosage of 1.2% chitosan, as shown in Table 1. The results were also similar to those of the adsorption isotherms in Figure 3.

The adsorption isotherms for the adsorption of CO₂ by graphene and graphene oxide and the modification of them by chitosan could be obtained from conducting experimental runs through the microbalance adsorption system, as shown in Figure 1, and the results are shown in Figure 10. Due to the excellent affinity between the amine groups and the CO₂, Figure 9 showed that the adsorption uptakes of chitosan-modified graphene and graphene oxide were higher than those of graphene and graphene oxide. The difference in adsorption uptake between graphene and graphene oxide could be attributed to the larger surface area of the graphene oxide, as shown in Table 1. The difference in adsorption uptake between graphene oxide and graphene modified with chitosan was greater than that without modification with chitosan. Table 1 shows that the surface area of oxide and graphene without modification was similar to that with modification by adding 0.9% chitosan. The results implied that the larger difference in adsorption uptake resulted from chitosan grafted onto the graphene oxide, and the results could also be confirmed by FTIR and EDS tests.

Figure 10 shows the breakthrough curves for the adsorption of CO₂ by graphene and graphene oxide and modification of them by chitosan, respectively, in the microbalance system. Since the breakthrough times and the thicknesses of the mass transfer regions for chitosan-modified graphene oxide and graphene were longer and narrower than those without modification with chitosan, the adsorption performance was better for chitosan-modified graphene oxide and graphene. The breakthrough curves could also confirm that the adsorption performance of chitosan-modified graphene oxide was slightly better than that of chitosan-modified graphene. Similar to the results discussed in Figure 10, the larger difference in adsorption performance could also be found between chitosan-modified graphene and the graphene in Figure 11 due to more chitosan being grafted onto the graphene oxide.

3.4. Adsorption Isotherm Models

The Toth adsorption isotherm model was combined with the Langmuir equation and Henry’s law to reduce the deviation between the experimental data and the predicted adsorption uptakes. Generally speaking, a heterogeneous adsorption system, such as the adsorption of gaseous adsorbates by solid adsorbents, could be modeled well by the Toth adsorption isotherm model. In addition, the Toth adsorption isotherm model was appropriate for fitting the situation where the adsorption uptake increased with increases in the adsorbate concentration at lower concentrations and increased moderately at higher concentrations. Therefore, the Toth adsorption isotherm model was used to model the relationship between adsorption uptake and adsorbate pressure in this study. The Toth adsorption isotherm model is formulated as follows:

$$q=q_{e}KP/{[1+{\left(KP\right)}^t]}^{1/t}$$

(1)

where q means the adsorption uptake of the adsorbent in mg/g, q_e means the equilibrium adsorption uptake in mg/g, K means the adsorption equilibrium constant, P means the gas pressure in KPa, and t means the parameters related to the heterogeneity of adsorption system. The Toth adsorption isotherm model was built in Systat 10.0 software, and the parameters in Equation (1) for adsorption of CO₂ by graphene, graphene oxide, and chitosan-modified graphene and graphene oxide could be obtained by fitting the experimental data. The parameters obtained from the fitting process are summarized in

Table 2. Parameters for adsorption data fitted by the Toth adsorption isotherm models.

Parameter	G 1	GO 2	GO Grafted by 0.3% CTS 3	GO Grafted by 0.6% CTS	GO Grafted by 0.9% CTS	GO Grafted by 1.2% CTS	G Grafted by 0.9% CTS
qe (mg/g)	362	373	389	400	423	426	419
K (mmHg−1)	0.123	0.129	0.252	0.362	0.461	0.479	0.395
t	0.516	0.513	0.459	0.447	0.436	0.433	0.409

¹ G: graphene; ² GO: graphene oxide; ³ CTS: chitosan.

. The deviation between the practical adsorption uptake and the predicted adsorption uptake calculated by the fitted equation was expressed as the average deviation D, and the formula is shown as follows:

$$D=\sum [\left|q_{exp}-q_{pre}\right|/q_{exp}]\times 100\%$$

(2)

The equilibrium adsorption uptake was increased with increases in the amount of chitosan added due to the high affinity between the amine groups and the CO₂. However, the increase by adding 1.2% chitosan was limited. Similarly, the adsorption equilibrium constants were also increased, with increases in the amount of chitosan added, and the increase with the addition of 1.2% chitosan was insignificant. Generally speaking, the closer the value t approaches one, the more similar the Toth adsorption isotherm model is to the Langmuir adsorption isotherm model, indicating the higher homogeneity of the adsorption system. Since the gaseous CO₂ was adsorbed by solid graphene and graphene oxide, the properties of the adsorbent–adsorbate pair tended to be a more heterogeneous system. Therefore, all t values were smaller than one, revealing that all adsorption runs were performed in a heterogeneous system. The degree of agreement between the adsorption isotherm model and the experimental data was examined by comparing the experimental data with the results obtained from the fitted adsorption isotherm model. The parameters in Table 2 were substituted into the Toth absorption isotherm model, and the relationships between the Toth absorption isotherm model and the experimental uptakes are shown in Figure 12. The adsorption uptakes could be fitted well by the Toth absorption isotherm model, with an average deviation of about 1.36%.

4. Conclusions

Graphene and graphene oxide were modified by chitosan to enhance the CO₂ adsorption performance in the adsorption system. Microbalance and staggered adsorption systems were used to obtain adsorption isotherms and breakthrough curves, respectively. Although the breakthrough times and the thicknesses of the mass transfer regions for graphene oxide modified with 0.9% and 1.2% chitosan were similar, the modification of graphene oxide with a 0.9% chitosan dosage was appropriate in this study due to a significant decrease in surface area with 1.2% chitosan dosage. The adsorption uptakes of chitosan-modified graphene and graphene oxide were higher than those of graphene and graphene oxide due to the excellent affinity between the amine groups and the CO₂. The adsorption uptake difference between chitosan-modified graphene oxide and graphene was greater than that without modification with chitosan. This could be attributed to the higher concentration of chitosan grafted on graphene oxide, which was also confirmed by FTIR and EDS analysis. Heterogeneous adsorption systems can be modeled well by the Toth adsorption isotherm model. In this study, the gaseous adsorbate (i.e., CO₂ gas) adsorbed by the solid adsorbent (i.e., thin films) is such a system. Therefore, the Toth adsorption isotherm model was used to simulate the relationship between adsorption uptake and adsorbate pressure in this study. The degree of agreement between the adsorption isotherm model and the experimental data was examined by comparing the experimental data with the results obtained from the fitted adsorption isotherm model. The results showed that the adsorption uptake could be predicted well by the Toth absorption isotherm model, with an average deviation of about 1.36%. The results demonstrated that the CO₂ adsorption uptake can be enhanced by chitosan-modified graphene oxide significantly in the fixed-bed adsorption system. In addition, heterogeneous adsorption systems, where gaseous adsorbates are adsorbed by solid adsorbents, can be well modeled by the Toth adsorption isotherm model. The systat software provided for obtaining parameters in the Toth adsorption isotherm models can also be used to establish linear or nonlinear models or correlations.