Carbon capture and storage (CCS) is becoming a hot topic as the urgency to contain climate change grows1. Solutions to capture CO2 from highly concentrated sources (that is, CO2 concentrated at a 10% level or more) exist and have been around for decades. These solutions have been applied extensively to exhaust vents in industrial processes and may be categorized depending upon the stage in which they are incorporated: pre-combustion, post-combustion, or oxy combustion2 (see Fig. 1). Regardless of the technology, these systems operate following two main steps: capture and release of CO2. Put simply, CO2-containing gas is blown into a contactor that contains material with the ability to capture CO2. Then, some of the CO2 present in the gas stream is captured. Next, the CO2 is released by applying energy, vacuum, moisture, or a combination of them, to move it into further sequestration or utilization. This way, the CO2-capturing material is taken back to its original state, (or “regenerated”), so the cycle can be restarted.

Figure 1
figure 1

CCS technologies for industrial waste, listing sorbents which can be applied to capture CO2 from air.

While capturing CO2 from industrial processes is relatively mature, capturing the same molecule from air is still considered an emerging technology. It turns out that capturing CO2 from highly diluted sources (in air, CO2 oscillates approximately between 410 and 420 ppm, which represents a level of concentration of 0.041%) is a completely different problem, were point-source CO2-capture technologies are not directly applicable. This is why the scientific community has been studying a wide range of materials which could serve as sorbents (either physisorbents or chemisorbents), depending upon their stability, selectivity to CO2, surface area, porosity capture capacity, and other properties3. This work focuses on adsorption, that is, solid materials which can capture CO2 from air.

Despite the recent progress, data on the stability and longevity of adsorbents are scarcely available in the literature, as show in Table 1. On the one hand, there is no single experimental method or tools to characterize adsorbent stability. While some researchers use thermal gravimetry and test a pelletized form of a composite that includes the adsorbent, others conduct sorption and desorption cycling in a fixed bed, in what is referred to as the “capture experiment”. Among them, some build structured contactors instead of testing pelletized composite. As a result, summarizing the state of the art of CO2-adsorbing materials and their properties may involve comparing data which is not strictly comparable. Out of completeness, Table 1 includes a revision of some of the most studied adsorbing materials and their properties.

Table 1 Stability data from pelletized composites and structured adsorbents adsorption/desorption cycles.

Additionally, other porous materials such as zeolite-based molecular sieves, activated carbons (ACs), and carbon nanotubes (CNTs) have attracted attention from researchers for gas adsorption. Activated carbons generally provide greater additional capacity at pressures above atmospheric pressure, compared to zeolites. In addition, ACs are often preferred over zeolites due to their relatively moderate gas adsorption strength, which facilitates desorption23,24,25,26. Furthermore, Zhang and collaborators27 have studied the microporous n-doped carbon adsorbent, obtained using polyaniline as a precursor, denoting that pore size and quantity play a critical role in the capture of CO2 in this type of material. Other studies employing wood sawdust and transforming it into biochar by a pyrolysis method have been carried out. Remarkably, it was found that the processing temperature impacts not only the yield but also the CO2 adsorption capacity of the material28.This is why it would be interesting to examine other adsorbents derived from vegetation waste, like graphene oxide.

Graphene is an increasingly important material and its storage capacity for different gasses has been suggested in theoretical studies; CO2 adsorption capacity is demonstrated at very low temperatures (195 K), which does not have much practical implication29. Therefore, it is necessary to investigate the CO2 adsorption capacity of graphene at room temperature and moderate pressure for the practical application of graphene in carbon capture and storage (CCS) technology30. In this work, GO synthesized by the Double Thermal Decomposition (DTD) method31 at different temperatures is used in the interdisciplinary Institute of Sciences at Universidad del Quindío in cooperation with Universidad Adolfo Ibáñez.

Graphene has shown intensive and promising applications in electronic devices32, batteries33, and composites34,35. Researchers have developed many methods to prepare this promising new nanomaterial, such as mechanical exfoliation, chemical vapor deposition (CVD)36, transfer printing37, epitaxial growth38, organic synthesis39, and oxidation-dispersion-reduction. Among these methods, the chemical reduction of GO sheets can produce graphene in large quantities, employing graphite as raw material. Because graphite is cheap and readily available, this chemical approach is probably the least expensive, most effective method for the large-scale production of graphene40.

Evidence, to date, has determined that graphene is a sp2-bonded planar carbon material. Due to its great potential in electronic applications, it has attracted much attention since it was first isolated in 2004. Driven by a fundamental interest and potential applications, but also as an example of chemical functionalization, graphene oxidation has been intensively studied41,42,43. However, due to the amorphous nature of GO generated by the chemical manufacturing method, understanding the atomic structure and its effects on the oxidation process remains a major challenge44,45,46,47,48,49,50,51,52,53.

Some authors report working with GO using hybrid materials and postulate them as potential materials for CO2 capture31,54,55,56,57,58. Other authors impregnated materials, like Zeolite and Silica gel, with amines. Amine-functionalized porous materials outperform all others in terms of CO2 adsorption capacity and regeneration efficiency59,60. Moreover, thermodynamic changes in systems where the GO is found can help us to look for desorption points, whether at high or low temperatures61,62,63,64,65,66,67,68,69.

Thus, studies of CO2 adsorption in GO structures in foams (GO-Foams) obtained through a carbonization process (873.15–1053.15 K) of organic waste material were carried out and additional tests on two materials derived from coffee as non-adsorbing reference materials can be found in Supplementary Information. In addition, adsorption calculations for a CO2 molecule on the surface of graphene and GO were also estimated. Therefore, the performance comparison between non-carbon (Zeolite and silica gel) material and the GO-Foams derived from vegetation waste is reported here. Furthermore, this work presents a functional application for this material in highly contaminated urban environments.

Materials and methods

Characterization method

Synthesis of GO foam was carried out by employing an efficient and environmentally friendly method, so-called the double thermal decomposition method (DTD), as reported70 and presented in a flowchart in Fig. 2. The method consists of treating a waste product of commercial bamboo—Guadua angustifolia Kunth—at different carbonization temperatures. In step 1, biomass from bamboo gets passivated, cleaned, and cut to move forward to step 2, where the first pyrolysis is carried out. The tar resulting from this step is taken to a second pyrolysis in which the GO foam is obtained, as noted in Fig. 4a–c. The authors confirm that all methods in experimental research and field studies on plants, as a waste product of the commercial bamboo-Guadua, were performed in accordance with the relevant regulations. Furthermore, the oxidation degree of graphene oxide was previously correlated to the carbonization temperature through XPS analyses, which were reported before53. The material was also characterized using TEM, XRD and Raman spectroscopy, as shown in Fig. 4.

Figure 2
figure 2

Flowchart for the DTD method to obtain the GO foams used in this research.

Here, GO foams were synthesized at 873 K (9.00% oxidation), 973 K (5.25% oxidation), and 1053 K (3.00% oxidation), which oxidation rate was determined via XPS analyses, as reported before71. Table 2 shows the three oxidation rates of GO with their respective formation temperatures, the time elapsed, superficial area and porosity. The authors confirm that all methods in experimental research and field studies on plants were performed adhering to relevant regulations70,72,73.

Table 2 Types of graphene oxide used for the experiments herein.

CO2 adsorption characterization

The characterization of CO2 adsorption of GO foams was carried out in an isolated CO2 measurement system based on the use of the MHZ-19B reference CO2 sensor in parts per million (ppm)74, which presents an optical measurement mechanism, allowing accurate measurements to be obtained in a wide range, from 0 to 5000 ppm ± 50 ppm. For this, a reaction for CO2 generation was introduced at the bottom of a closed system; this reaction is based on the reaction given by Eq. (1).

$${\text{NaHCO}}_{{{3}({\text{solid}})}} + {\text{ CH}}_{{3}} {\text{COOH}}_{{({\text{liquid}})}} \to {\text{CH}}_{{3}} {\text{COONa}}_{{({\text{liquid}})}} + {\text{ H}}_{{2}} {\text{O}}_{{({\text{liquid}})}} + {\text{ CO}}_{{{2}({\text{gas}})}}$$

To guarantee controlled CO2 production within the system, two compounds were used: acetic acid and sodium bicarbonate, which give as product three other compounds: sodium acetate, water (H2O), and carbon dioxide (CO2), making it an efficient and low-cost CO2 production. An MH-Z19B CO2 sensor is located above of the GO foam to ensure better reading of the CO2 adsorption (Fig. 3). In the first part of Fig. 3, the CO2 source (NaHCO3 (solid) + CH3COOH (liquid) reaction) is located in the lower part of the experiment, and in the second part of Fig. 3, our sample holder is located in said CO2 source, followed by the third part; for this, the material under study is located on the sample holder and, thus, said material is located in our gas source. Finally, as a fourth part, the system is sealed with the upper cover (which has the sensors) that will prevent the gas from leaking into the system.

Figure 3
figure 3

CO2 adsorption characterization system (GO–Foam–CO2) for GO foams developed herein.

Computational details

The VASP software was used to calculate the molecules75. To perform the geometric and energetic calculations, the GGA functional, PBE76 was used because this functional is widely employed to predict various properties of molecules and non-bond interactions77. Based on the Lerf–Klinowski model78,79 and the structure presented by Prías-Barragán et al.72 a single CO2 molecule and two structures of isolated armchair graphene flakes were modeled; the first one is graphene with hydrogen passivated edges (\({C}_{100}{H}_{26}\)), and a GO structure with 9.00% oxide coverage (\({C}_{100}{H}_{34}{O}_{9}\)), given that the GO employed in the experimental case is in the lower oxidation regimen70. After the first relaxation of every structure, the CO2 molecule was placed at a certain distance from the graphene and GO surface and the process was repeated. To obtain the adsorption energies, Eq. (2) was employed,


where \({E}_{system}\) corresponds to the energy of the graphene or GO sheet with a CO2 molecule adsorbed, and \({E}_{graphene}\) and \({E}_{C{O}_{2}}\) correspond to the energy associated with the isolated graphene and CO2 molecule, respectively.

The non-covalent interactions (NCI) and molecular electrostatic potential (MEP) were calculated to analyze theoretically the adsorption of the CO2 molecule on graphene and GO structures. The charge transfer was examined by analyzing Bader charges, obtained through the critic2 software80,81, before and after adsorption. Most of the calculations were performed in the cluster from the Virtual Materials Modeling Laboratory (VMML) group, at the Center for Nanoscience and Nanotechnology, in the “Miztli” supercomputer, with a processing capacity of 228 TFlop/s, which has 8,344 processing cores, 16 NVIDIA m2090 cards, a total RAM of 45,000 GB and a 750 TB mass storage system, property of UNAM.

Results and discussions

Figure 4a–c shows photographs of graphene oxide foams at different oxidation rates. In Fig. 4d the transmission electron microscopy of the GO is observed. These graphene foams have a close porosity as seen in Table 2, this allowing the entry and exit of CO2 gas. The Fig. 4e presents the consolidated XRD patterns of GO–Foam samples synthesized at different TCA, observing in the GO–Foam samples the characteristic peaks of hexagonal Graphite in the (002), (100), (101), and (004) directions, showing that it is a polycrystalline material. Figure 4f illustrates normalized Raman spectra of GO-Foam samples, presenting the characteristic peaks G-band peak around 1560 cm−1 associate to graphene structure and D-band peak around 1350 cm−1 attribute to the disorder-induced phonon mode; The wide 2D and D + G bands around the 2800 cm−1 value suggest the presence of multiple graphene layers with edges, defects, and sp2 regions, which are prevalent features of the GO − Foams synthetized, as previously reported70,82,83.

Figure 4
figure 4

GO foams obtained employing the DTD and characterization methods, (a) 873.15 K (GO 9.00%), (b) 973.15 K (GO 5.25%), (c) 1053.15 K (GO 3.00%), (d) GO–TEM, (e) GO–XRD patterns and (f) GO–Raman at 873 and 973 K.

CO2 generation

To calibrate and fine-tune the CO2 sensors, CO2 was produced from a reaction of NaHCO3 (as solid, sodium bicarbonate) plus CH3COOH(aqueous) (acetic acid), yielding CH3COONa (aqueous) (sodium acetate), plus H2O(liquid) (water), plus CO2(gas) (carbon dioxide).

Initially, a measurement of CO2 production was performed inside the system from the reaction of 1.5 mg of NaHCO3 (solid) (sodium bicarbonate) plus 0.5 ml of CH3COOH (aqueous) (acetic acid), yielding CH3COONa (aqueous) (acetate of sodium), plus, H2O (liquid) (water), plus, CO2 (gaseous) (carbon dioxide), as products. Figure 5a identifies the CO2 production obtained, where the concentration of this gas increases from 325 to approximately 800 ppm.

Figure 5
figure 5

(a) CO2 generation characterization, CO2 adsorption in (b) Zeolite, (c) Silica gel, and (d) CO2 adsorption using GO-9.00% at 294.15 K.

CO2 adsorption on zeolite, silica gel, and graphene oxide foam

The graph in Fig. 5b, where zeolite was used as an absorbent material, shows CO2 concentration vs. time in seconds, starting with a minimum CO2 concentration of 280 ppm; CO2 production was observed with an approximate maximum of 700 ppm at 3,500 s after starting the CO2 production reaction. After this time, absorption of the zeolite is evident with adsorption reaching 575 ppm at 7000 s and, thereafter, it is observed that it does not contain CO2 within for a long time, again showing CO2 release, increasing to 650 ppm. In Fig. 5c, in the presence of silica gel as adsorbent material, the graph shows CO2 concentration vs. time in seconds, starting with a minimum CO2 concentration of 370 ppm and CO2 production with an approximate maximum observed, from 600 ppm at 1200 s after starting the CO2 production reaction. After this time, the absorption of the silica gel is evident with adsorption reaching 300 ppm at 4000 s and later it is observed that it does not contain CO2 inside for a long time, again showing CO2 release, increasing to 450 ppm. This indicates its low retention capacity inside its structure. Figure 5d, presents the evolution of the CO2 production in GO-9.00% foam at 294.15 K (room temperature) in a closed system shown in Fig. 3. After approximately 2500 s, a clear slow absorption of CO2 in the system is noted, thus, revealing a slow decrease in CO2 gas, with slow adsorption over time until stable departure levels are reached. When comparing these three adsorbent materials, the superiority of the oxidized graphene foam is identified concerning zeolite and silica gel, given that they contain more CO2 gas and maintain it over time due to their high efficiency and performance.

Temperature effects on the saturation of CO2 adsorbed on GO foam

In the experiment using GO-9.00% for CO2 adsorption, this gas was produced using 1.5 mg NaHCO3 (solid) (sodium bicarbonate) and 0.5 ml CH3COOH (aqueous) (acetic acid). The graphene oxide used in the experiment was heated to 423.15 K for 48 h and, subsequently, it was measured if it had already released CO2 from its interior, as seen in Fig. 6a. A constant trend of stability in CO2 production is determined, starting from 300 ppm within the system up to 700 ppm of production, this last value identified as constant, from 1500 s on; this is attributed to the fact that this CO2 gas was not released due to a stationary regime. CO2 Gas in the GO at 423.15 K cannot be retained on its walls, and therefore could not adsorb more CO2 gas; since it requires more energy to desorb and thus be able to be ready for a new adsorption. Therefore, this GO foam continued to heat up further. This also occurred with the other two temperatures explored before knowing the ideal desorption temperature of the GO foam using heating temperatures of 523.15 K, as seen in Fig. 6b, starting from 100 ppm and obtaining a maximum CO2 production of 650 ppm, remaining stable at this value. When heated to 573.15 K, it was again exposed to a CO2 reaction, where at 573.15 K (Fig. 6c) it is observed that it departs from a CO2 concentration of 200 ppm, reaching a maximum of 650 ppm and then remaining stable in a valley evidencing that there is no adsorption of the gas due to a stationary regime. It is important to note that the exact temperature dependence of CO2 adsorption on GO foams will depend on the specific properties of the foam, such as pore size, surface area, and functional groups. Therefore, experimental studies are needed to determine the temperature dependence of CO2 adsorption on a particular GO foam.

Figure 6
figure 6

(a) GO-9.00% with a temperature of 423.15 K, (b) GO-9.00% with a temperature of 523.15 K, and (c) GO-9.00% with a temperature of 573.15 K.

Temperature influence on the re-adsorption of CO2 adsorbed on GO foam

The GO-3.00% already saturated with CO2 was used, which was synthesized at 1053.15 K. The same graphene from the previous experiments was used, already saturated with CO2, placed in a muffle, and heated to 673.15 K for 5 h and 30 min. Desorption results were successful because the material recovered its adsorbent condition, as shown in Fig. 7a, going from a CO2 reduction from 600 to 420 ppm in 12,000 s, to again show its adsorption qualities. This results in a great quality of CO2 adsorption and desorption, called re-adsorption. Re-adsorption of CO2 on GO foams can be influenced by temperature in several ways; solubility of CO2 in a material decreases with increasing temperature. However, CO2 re-adsorption onto GO foam is a complex process involving multiple mechanisms, so the effect of temperature on re-adsorption may not be straightforward, as seen in this work. GO-5.25%, already saturated with CO2 was used, synthesized at 973.15 K. This already saturated graphene from the previous experiments was used, placed in a muffle, and heated to 673.15 K for 5 h and 30 min. The desorption results were successful because the material recovered its adsorbent condition, as shown in Fig. 7b, going from a CO2 reduction from 700 to 450 ppm in 16,000 s, to then also show its re-adsorption qualities, thus improving the results of GO-3.00%. Physisorption is a process in which CO2 gas molecules are held to a surface by weak van der Waals forces. The interaction of these forces increases as temperature decreases, thereby, lowering the temperature may increase the amount of CO2 that can be physiosorbed onto the GO foam or, conversely, if temperature is increased these forces are weakened, thus allowing the GO-Foam-CO2 to desorb. Another mechanism that can be influenced by temperature is chemisorption. Chemisorption is a chemical reaction between the adsorbate (CO2) and the adsorbent (GO foam), which can be exothermic or endothermic, depending on the specific reaction. Changes in temperature can affect the activation energy of the reaction and the energy required for the adsorption process, which—in turn—can affect the rate and extent of re-adsorption. GO-9.00% already saturated with CO2 was used, synthesized at 973.15 K. This already saturated graphene from the previous experiments was used, placed in a muffle, and heated to 673.15 K for 5 h and 30 min. The desorption results were successful because the material recovered its adsorbent condition, as shown in Fig. 7c, going from a CO2 reduction from 750 to 400 ppm in 15,000 s, to again begin to show its adsorption qualities; resulting in a great property of CO2 re-adsorption. It is evident that it did not improve the conditions of the results of the GO-5.25% but did improve those of the GO-3.00%. In summary, temperature can influence CO2 re-adsorption on GO foam through physisorption and chemisorption mechanisms. The specific effect of temperature will depend on the specific conditions and properties of the GOFs and the CO2 gas.

Figure 7
figure 7

(a) GO-3.00% with re-adsorption temperature of 673.15 K, (b) GO-5.25% with re-adsorption temperature of 673.15 K, and (c) GO-9.00% with re-adsorption temperature of 673.15 K.

Low temperatures

The graphene’s were also exposed to low temperatures (ranging from 260.15 to 253.15 K) obtaining favorable low re-adsorption results, as seen in Fig. 8a, because of less than 10% re-adsorption. This figure shows how very low graphene oxide adsorbed more CO2 from the system. After being exposed to low temperatures for several hours, it intervened in the CO2 saturation obtained from previous experiments, starting from a concentration of 100 ppm before the CO2 production reaction and with a maximum CO2 concentration of 550 ppm, at 2000 s; after this time, a decrease in concentration of approximately 450 ppm is obtained in 8500 s. As in the previous experiment, it is shown how very low graphene oxide adsorbed more CO2 from the system after exposure to low temperatures of 253.15 K for 24 h, starting from a CO2 concentration of 100 ppm and a maximum of 550 ppm of carbon dioxide at 2000 s, but after this time a decrease in concentration of approximately 480 ppm was obtained in 7500 s; where low CO2 adsorption is observed, as identified in Fig. 8b. It would be very important to continue exploring even with lower temperatures, given that if temperature is too low, the CO2 molecules can freeze and become less mobile, which could decrease the total re-absorption amount, which serves as another desorption method.

Figure 8
figure 8

(a) GO 9.00% at 260.15 K and (b) GO 9.00% at 253.15 K.

Table 3 shows the different materials used in this work for CO2 capture. The three oxidation rates of graphene are included, the zeolite and the silica gel, listed with their respective efficiencies, yields, system temperatures, and humidity. The methods to estimate efficiency and yield are briefly described in the Supplementary information.

Table 3 Efficiency (η) v/s yield (Y), with their respective system *Room temperature (Rt), humidity (h) and atmospheric pressure (hPa).

Theoretical results

The optimized structures can be seen in Fig. 9, showing the positions of the functional groups: hydroxyl (–OH) and epoxy (–O–). On the surface of the final relaxed graphene structure, the CO2 molecule was placed at 3.32 Å, as shown in Fig. 9c, consistent with that reported in the literature84, suggesting weak interactions, like Van Der Waals and NCl. The CO2 molecule was positioned in three locations, the first one, GO-1, between the bottom hydroxyl groups, the second one, GO-2, at the top hydroxyl of the structures and last, GO-3, near the single hydroxyl on the right of the structure, which correspond to Fig. 9d–f, respectively.

Figure 9
figure 9

Structures studied. (a) Graphene, (b) GO with hydroxyl bridges, (c) CO2 molecule adsorbed in pristine graphene passivated by hydrogen atoms in its edges, (d) converged CO2/GO structures for the GO-1 position, (e) GO-2 position, (f) GO-3 position, (g) MEP for pristine graphene/CO2, (h) MEP for GO-1 position, (i) NCI for pristine graphene/CO2 and (j) NCI for GO-1 position.

Figure 9g and h presents MEP calculation images and reveals a high or low electron density, presenting the reactivity point of the surface of GO structures. Comparing Fig. 9g and h, the charge redistribution is noted of the CO2 molecule due to the interaction with the GO structure in which the NCI displayed in Fig. 9i and j proves the existence of the weak relation between the two structures through the van der Waals interaction for each system.

The Eads for each system is displayed in Table 4, which shows the adsorption energy for the CO2/Graphene system, (− 0.2288 eV), agreeing with the values reported by Wang et al.85 and the decreasing trend of the values from graphene to each position of CO2 in GO is visible and suggests physisorption, as the main adsorption mechanism; there is also the decreased distance between the contaminant molecule and the GO, associated with differences of the electrical dipoles84. The Bader charge of the GO-2 system (0.6349 e) compared to the other structures, exhibits the highest value and, therefore, it is feasible to assume a stronger interaction between the components, given a charge transfer from the oxygen atom to the hydrogen from the adjacent hydroxyl, making this group highly important for carbon capture, especially in the GO-2 position.

Table 4 Calculated properties and adsorption energy for the CO2/Graphene and CO2/GO systems.

These results suggest a possible physisorption mechanism between the graphene and CO2, which describes Van Der Waals interaction between the GO and CO2, making these materials excellent candidates for carbon capture and air decontamination.

Possible applications

The GO–Foam–CO2 prototype could be used as a CO2 capture, purification, and monitoring system in many places, like parks, main squares, trains, planes, airports and, overall, in cities with high concentrations of CO2, as seen in Fig. 10.

Figure 10
figure 10

Possible application of GO–Foam–CO2 for carbon removal in a traffic light. (Permissions allowed by Erica Valencia (left figure) and Humberto Franco (right figure). copyright holders).


The results of heating the GO at 9%, 5%, and 3% to 673.15 K for several hours show successful results because GOs recover their conditions as adsorbent material. Conversely, when heating these GO below 673.15 K, it was observed that the GOs do not release the CO2 gas. It was noted how at specific low temperatures, in this case, 260.15 and 253.15 K, GOs do not recover their adsorption capacity; therefore, making a better sweep of this low-temperature area would be extremely important, for example where CO2 stops being gas. The organic materials used in this work to make an ideal comparison with non-adsorbent materials, in this case, roasted and dry coffee, are identified as non-adsorbent materials. Zeolite and silica gel in this work are used as a reference to a CO2 gas adsorbing material; therefore, it was quite comfortable to make the comparison with the three different oxidation rates of GO. The GO at 873.15 K had the best performance, but the GO at 1053.15 K had the highest efficiency. The relaxed structures present adsorption values in the weak physisorption range, indicating interactions of the hydroxyl groups on the surface of GO with the CO2 molecule, which can be interpreted that GO is a promising material for carbon capture from the air and opens the possibility of developing technological devices with these types of materials. It was also of great importance to find that this material can desorb at 673.15 K. These results suggest that GO foams are a promising material for carbon capture and future development of a new clean tech, given their highest CO2 adsorption efficiency and yield.