Plan

Article
Experiments and DFT study on modified CaO-based adsorbents for enhanced CO${}_{2}$ capture
Comptes Rendus. Chimie, 2021, 24, no. 2, p. 177-187

Résumé

CaO-based adsorbents for carbon capture represent a promising technology for reducing carbon emission. In this study, we prepare metal oxide-doped multifarious CaO-based adsorbents using the hydration method. We investigate the effect of various working conditions, such as temperature and carbonation time, on different adsorbents in a fixed-bed reactor under multiple carbonation–calcination cycles. We examine the behavior of different metal oxides-doped synthetic adsorbents using density functional theory calculation based on experiments. The results prove that 5 wt% ZrO2-doped adsorbents show excellent CO2 adsorption efficiency, which reaches up to 38.4% after 20 carbonation–calcination cycles at 700 °C with 15 vol% CO2. The adsorbents doped with other metal oxides are also useful for CO2 capture to varying degrees. The adsorption energy of CO2 molecules on modifiequationed adsorbents is higher than that on pure CaO, especially for Zr, where the adsorption energy reached 2.37 eV. The calculation results are in good agreement with the experimental data.

Reçu le :
Révisé le :
Accepté le :
Publié le :
DOI : https://doi.org/10.5802/crchim.79
Mots clés : Density functional theory, CaO-based adsorbents, Carbon capture, Carbonation–calcination cycles, Adsorption energy
@article{CRCHIM_2021__24_2_177_0,
author = {Zhixin Li and Qinhui Wang and Yi Feng and Mengxiang Fang},
title = {Experiments and {DFT} study on modified {CaO-based} adsorbents for enhanced {CO$_{{2}}$} capture},
journal = {Comptes Rendus. Chimie},
pages = {177--187},
publisher = {Acad\'emie des sciences, Paris},
volume = {24},
number = {2},
year = {2021},
doi = {10.5802/crchim.79},
language = {en},
}
Zhixin Li; Qinhui Wang; Yi Feng; Mengxiang Fang. Experiments and DFT study on modified CaO-based adsorbents for enhanced CO$_{{2}}$ capture. Comptes Rendus. Chimie, Tome 24 (2021) no. 2, pp. 177-187. doi : 10.5802/crchim.79. https://comptes-rendus.academie-sciences.fr/chimie/articles/10.5802/crchim.79/

Texte intégral

## 1. Introduction

Fossil fuels are essential to modern society, as they are one of the most important contributors to world’s energy consumption. However, their excessive use is leading to a sharp increase in CO2 emissions, which creates many environmental concerns, such as the “greenhouse effect”. Statistical data show that carbon emissions of China increased by around 6 billion tons from 2001 to 2017. In 2016, China became the world’s largest emitter of greenhouse gases [1], so reducing carbon dioxide emissions is a challenging problem for China. A CaO-based carbon capture technology is considered promising for carbon emission reduction, and hence it is studied here in detail. The circulatory system contains a reactor to capture CO2 and a calcination furnace for regenerating CaO-based adsorbents [2, 3, 4]. However, calcination may also bring about deformation and sintering of particles of the adsorbent [5, 6]. After multiple carbonation–calcination cycles, the adsorption efficiency of calcium-based adsorbents decreased significantly [7, 8].

We have examined CaO-based sorbents containing different metal elements using a hydration method. The performance of the CaO-based adsorbent was tested in a fixed-bed reactor under different working conditions. DFT calculations were used to determine the key parameters of different adsorbents, which can help explain the experimental results.

## 2. Experimental part

### 2.1. Materials

Modified CaO-based adsorbents were prepared using the hydration method. Analytical reagent (AR)-grade CaO particles were used for the experiment. Other materials included ZrO2 (99%), 200–300-mesh Al2O3, and CuO (99%). Five adsorbent samples doped with metal oxides were prepared: (1) 95 wt% CaO + 5 wt% ZrO2; (2) 95 wt% CaO + 5 wt% Al2O3; (3) 95 wt% CaO + 2.5 wt% ZrO2 + 2.5 wt% Al2O3; (4) 95 wt% CaO + 5 wt% CuO; (5) 100 wt% CaO. All samples were evenly stirred in deionized water. After allowing samples to stand for 30 min, a specimen was put into an oven and dried for 12 h at 100 °C, and particles with diameter of 200 μm were chosen for experiment. X-ray diffraction (XRD) shows essentially the same main Ca(OH)2 and secondary CaCO3 peaks for all 5 samples (Figure 1 shows a typical spectrum obtained with the Al2O3-doped sample). Due to moisture remaining after drying, CaO particles reacted with H2O to form Ca(OH)2; then a small amount of Ca(OH)2 and CO2 reacted to form CaCO3. Therefore, samples were subsequently calcined for 3 h at 900 °C to remove Ca(OH)2 and CaCO3.

### 2.2. Experimental apparatus

We used a fixed-bed reactor in our experiments. The outer and inner diameters of the quartz tube were 125 and 120 mm, respectively, and its length was 1400 mm. The samples were placed in a quartz boat and gently pushed to the center of the quartz tube. The heating and cooling rates were set at 15 and 20 °C/min, respectively. The carbonation reaction temperature varied from 600 to 750 °C, depending on the working condition, and the calcination temperature was set at 850 °C. The reaction time for carbonation and calcination was set at 25 and 15 min, respectively. Here, 500 mg of experimental samples was investigated for 20 carbonation–calcination cycles under various working conditions.

The CO2 and N2 gas valves were connected to the fixed-bed reactor. The total gas flow was fixed at 400 ml/min. The purge and furnace gases composed of N2 were used to effectively preserve the fixed-bed reactor. The carbonation conversion rate of the synthetic adsorbent was defined as

 $X=Mt−M0M0f×MCaOMCO2,$ (1)
where $Mt$ is the weight of the adsorbent at time t, M0 is the original mass of the adsorbent after calcination, f is the mass fraction of CaO in the adsorbent, MCO 2 and MCaO are the molar masses of CO2 and CaO, respectively, and X is the carbonation conversion after multiple carbonation–calcination cycles.

The XRD analysis was carried out to study the specific distribution of crystal phase in the sample. To determine the crystal structure, atomic coordinates, and absolute configuration, we used a Gemini XRD analyzer, with a CCD detector and a cryogenic system.

### 2.3. Preparation of DFT calculation

We established the crystal structure model using Materials Studio software to perform the calculation. The CaO crystal has a cubic structure. The lattice constant after optimization is 4.862 Å, which is in agreement with the experimental results [33]. The C–O bond length and O–C–O bond angle of the CO2 cell are 1.173 Å and 179.895°, respectively, which is consistent with the experimental results [34]. The CaO crystal cell is defined by adopting crystal indices (100). A vacuum layer with a thickness of 15 Å is chosen. Simultaneously, the CaO (110) surface is modified by adding Zr, Al, and Cu to achieve the doping of metal elements and structure optimization.

Calculations were performed using the general gradient approximation method and the Cambridge Serial Total Energy Package (CASTEP) module. The cut-off energy value of 570 eV was selected for truncation. The energy difference between adjacent ion steps was less than 1.0 × 10−5 eV/atom. The maximum atomic interaction force, maximum atomic stress, and atomic displacement were 0.3 eV/nm, 0.05 GPa, and 0.0001 nm, respectively. Here, the energy convergence standard of the electron step in the self-consistent field operation was 10−5 eV/atom. The K point was set as 3 × 4 × 1 to ensure the accuracy of calculation. The adsorption energy Ead is defined as

 $Ead=Esystem−(Eadsorb+Esubstrate),$ (2)
where $Esystem$, Eadsorb, and Esubstrate are the total electronic energy of the system after adsorption, the electronic energy of the adsorbate, and the electronic energy of the surface, respectively.

## 3. Results and discussion

### 3.1. Performance test of modified adsorbents

#### 3.1.2. Effect of carbonation temperature on the performance of modified adsorbents

We also investigated the performance of modified adsorbents at different carbonation temperatures. The adsorption efficiency of adsorbents doped with (a) 5 wt% ZrO2, (b) 2.5 wt% ZrO2 + 2.5 wt% Al2O3, and (c) 5 wt% Al2O3 increased by (a) 6.9%, (b) 7.15%, and (c) 7.43% when the carbonation temperature increased from 650 to 700 °C (20 carbonation–calcination cycles). The effect of temperature on the performance of adsorbents can also be clearly seen in Figures 46. However, adsorption efficiency decreased significantly at 750 °C (Figure 7). Figure 8 further confirmed 700 °C as the optimal temperature.

#### 3.1.3. Effect of carbonation time on the performance of modified adsorbents

The effect of carbonation time on the adsorption efficiency of adsorbents was also investigated. The 2.5 wt% Al2O3 + 2.5 wt% ZrO2 adsorbent was our main object of study. The carbonation temperature was 700 °C, and the atmosphere was N2/CO2 with 15% CO2 concentration. As shown in Figures 9 and 10, the adsorption efficiency increased by 23.2%, 8.2%, and 4.7% for 0–5 min, 5–15 min, 15–25 min, respectively. The initial steepness of the curve in Figure 10 shows that the carbonation reaction was very fast between 0–5 min. Between 5 and 15 min (the transition period), the adsorption efficiency of adsorbents increased relatively rapidly. However, the adsorption reaction became slow between 15 and 25 min (reaction diffusion stage). In particular, the 0–5 min rapid stage of carbonation ensured good CO2 capture capacity.

#### 3.1.4. Effect of CO2 concentration on the performance of modified adsorbents

This phenomenon was explained as follows. First, the whole carbonation reaction process was divided into two stages: the rapid reaction stage and the reaction diffusion stage. Thereafter, the CO2 diffusion rate in the carbonation stage was determined by temperature and CO2 concentration. In addition, the sintering degree was based on the same calcination temperature and time for all adsorbents. Figures 11 and 12 suggest that the CO2 diffusion rate is mainly depended on CO2 concentration at 650 °C, but determined by temperature and CO2 concentration together at 700 °C. The diffusion rates with 40 and 15 vol% CO2 are similar at 650 °C (both curves essentially overlap). The diffusion rate with 8 vol% CO2 was far slower than that under high CO2 concentration, which led to the sharp decrease in adsorption efficiency.

Figure 12 shows that the diffusion rate of the CO2 molecule was rapid with 40 vol% CO2 concentration at 700 °C, so the adsorption efficiency of adsorbents in rapid reaction stage was also higher. However, the CO2 molecular diffusion rate with 15 vol% CO2 was relatively slow, and the CaCO3 produced in the rapid reaction stage might result in pore blockage, which would ultimately reduce the adsorption efficiency.

Table 1.

Main results of DFT calculation

CaO (100)Al–CaO (110)Zr–CaO (110)Cu–CaO (110)
Ca–O bond angle/° 90 82.74 73.4 83.587
C–O bond length/Å 2.9 2.7962.3291.506
O–C–O bond angle/° 141.1 142.6 119.05 132.76
O–C–O bond length/Å 1.208 1.176 1.3 1.24

### 3.2. Results of DFT calculation

To verify the experimental results for adsorbents under various working conditions, DFT calculations were performed using MS software. The carbonation reaction between CaO-based adsorbents doped with different metals and CO2 molecules was studied in detail. The structure of CO2 molecules and CaO cells has been calculated before [25, 35]. However, in this study, the adsorption reaction between CaO-based adsorbents and CO2 was emphasized. Figure 13 shows the adsorption models of CO2 molecules on the CaO-based adsorbents doped with different metal elements. It could be seen that CaO (110) crystal cells had multiple adsorption sites, including O-top, Ca-top, bridge, and hollow sites. We found that the possibility of forming an anionic carbonate-like $CO32−$ structure was high for CO2 molecules on the O-top site compared with other adsorption sites. Other studies showed that CO2 molecules were also inclined to be adsorbed at the O-top position [26]. Therefore, we only investigated the adsorption conditions at the O-top site. The adsorption results are given in Table 1, including adsorption energy and bond length angle. The adsorption energy of pure CaO was 1.715 eV. The structure of the CaO crystal cell changed slightly upon CO2 adsorption, the Ca–O bond angle remained 90°. The C–O bond length was 2.9 Å, and the bond angle of the CO2 molecule was 141.1°.

The results show that the adsorption energy of CO2 molecules is higher with metal-doped CaO-based adsorbents, especially for Zr. The higher adsorption energy ensured that the C–O bond was more stable, and the adsorbents had much stronger ability to resist sintering. Figures 1417 show the changes in energy band structure of different modified CaO-based adsorbents after adsorption reaction. In summary, the partial density of state of CaO-based adsorbents doped with different metal elements was shifted from right to left. In particular, when the peak height of the waveform in figures is large, then the degree of hybridization between different atoms is higher.

Figure 15 shows the best adsorption performance compared with other results. It was found that the structure in CO2 molecules and CaO crystal cells had also changed and the O–C–O bond angle got smaller, which was caused by the deformation in electrons from the O2− transferred to the C atom. It should be noted that there was no obvious change in the energy band structure of Cu-doped adsorbent, although some researchers have observed benefits for carbon reduction [20, 36, 37, 38]. It might be explained that the method was applied for the CO2 catalytic reduction rather than adsorption reaction. The Cu-doped adsorbent may have melted during the experiment at high calcination temperature. The experimental results are consistent with the DFT calculations for all the carbonation temperatures examined here. It was clear that the Zr-doped adsorbent had the best adsorption efficiency after 20 carbonation–calcination cycles.

## Acknowledgment

This work was supported by The National Key Research and Development Program of China (No. 2019YFE0100100-05).

Bibliographie

[1] Y. Zhang; S. Li; T. Luo; J. Gao J. Clean. Product., 265 (2020), 121843

[2] H. Li; M. Qu; Y. Yang; Y. Hu; W. Liu Chem. Eng. J., 374 (2019), pp. 619-625

[3] Y. Zhang; X. Gong; X. Chen; L. Yin; J. Zhang; W. Liu Fuel, 232 (2018), pp. 205-214

[4] S. He; Y. Hu; T. Hu; A. Ma; Q. Jia; H. Su; S. Shan J. Alloys Comp., 701 (2017), pp. 828-833

[5] X. Gong; Z. Wang; Z. Wang; J. Cao; S. Zhang Powder Technol., 336 (2018), pp. 92-101

[6] H. Sun; C. Wu; B. Shen; X. Zhang; Y. Zhang; J. Huang Mater. Today Sustain., 1–2 (2018), pp. 1-27

[7] J. M. Valverde; P. E. Sanchez-Jimenez; L. A. Perez-Maqueda Appl. Energy, 126 (2014), pp. 161-171

[8] X. Liu; X. Ma; L. He; S. Xu Chin. J. Chem. Eng., 25 (2017), pp. 1412-1421

[9] Y. Hu; H. Lu; W. Liu; Y. Yang; H. Li Chem. Eng. J., 396 (2020), 125253

[10] S. Scaccia; G. Vanga; D. M. Gattia; S. Stendardo J. Alloys Compounds, 801 (2019), pp. 123-129

[11] Q. Hu; C. H. Wang Bioresour. Technol., 310 (2020), 123384

[12] H. Wang; Z. Li; N. Cai Chem. Eng. J., 394 (2020), 124892

[13] C. H. Lee; S. W. Choi; H. J. Yoon; H. J. Kwon; H. C. Lee; S. G. Jeon; K. B. Lee Chem. Eng. J., 352 (2018), pp. 103-109

[14] P. Pecharaumporn; S. Wongsakulphasatch; T. Glinrun; A. Maneedaeng; Z. Hassan; S. Assabumrungrat Int. J. Hydrog. Energy, 44 (2019), pp. 20663-20677

[15] H. Guo; J. Feng; Y. Zhao; S. Wang; X. Ma J. CO${}_{2}$ Util., 19 (2017), pp. 165-176

[16] H. J. Yoon; K. B. Lee Chem. Eng. J., 355 (2019), pp. 850-857

[17] X. Tong; W. Liu; Y. Yang; J. Sun; Y. Hu; H. Chen; Q. Li Fuel Process. Technol., 193 (2019), pp. 149-158

[18] N. Wang; Y. Feng; L. Liu; X. Guo Fuel Process. Technol., 173 (2018), pp. 276-284

[19] A. N. Antzara; A. Arregi; E. Heracleous; A. A. Lemonidou Chem. Eng. J., 333 (2018), pp. 697-711

[20] C. Qin; J. Yin; C. Luo; H. An; W. Liu; B. Feng Chem. Eng. J., 228 (2013), pp. 75-86

[21] J. Li; M. Hou; Y. Chen; W. Cen; Y. Chu; S. Yin Appl. Surf. Sci., 399 (2017), pp. 420-425

[22] K. R. Vijisha; M. K Int. J. Greenh. Gas Control, 58 (2017), pp. 62-70

[23] M. D. Esrafili J. Mol. Graph Model., 90 (2019), pp. 192-198

[24] M. D. Esrafili; F. Sharifi; L. Dinparast J. Mol. Graph Model., 77 (2017), pp. 143-152

[25] X. Ma; Y. Li; W. Zhang; Z. Wang; J. Zhao Chem. Eng. J., 370 (2019), pp. 10-18

[26] L. Liu; D. Hong; X. Guo J. CO${}_{2}$ Util., 22 (2017), pp. 155-163

[27] Y. Fan; J. G. Yao; Z. Zhang; M. Sceats; Y. Zhuo; L. Li; P. S. Fennell Fuel Process. Technol., 169 (2018), pp. 24-41

[28] H. Guo; Z. Xu; T. Jiang; Y. Zhao; X. Ma; S. Wang J. CO${}_{2}$ Util., 37 (2020), pp. 335-345

[29] B. Jensen; L. Pettersson J. Phys. Chem. B, 109 (2005), pp. 16774-16781

[30] W. Dai; Z.-H. Shui; K. Li Comput. Theoret. Chem., 967 (2011), pp. 185-190

[31] Y. Xu; C. Luo; Y. Zheng; H. Ding; Q. Wang; Q. Shen; L. Zhang RSC Adv., 6 (2016), pp. 79285-79296

[32] X. Ma; Y. Li; L. Duan; E. Anthony; H. Liu Appl. Energy, 225 (2018), pp. 402-412

[33] P. Yang; L. Duan; H. Tang; T. Cai; Z. Sun Greenh. Gases: Sci. Technol., 8 (2018), pp. 1110-1123

[34] F. Hou; J. Jin; H. Yang; Y. Wang; S. Li Appl. Surf. Sci., 475 (2019), pp. 1033-1042

[35] W. Wang; L. Fan; G. Wang; Y. Li Appl. Surf. Sci., 425 (2017), pp. 972-977

[36] J. Chen; T. Shi; L. Duan; Z. Sun; E. J. Anthony Chem. Eng. J., 393 (2020), 124716

[37] F. N. Ridha; D. Lu; A. Macchi; R. W. Hughes Fuel, 153 (2015), pp. 202-209

[38] I. Martínez; J. R. Fernández; J. C. Abanades; M. C. Romano Energy, 163 (2018), pp. 570-584