1 Introduction

Carbon dioxide (CO2) is the primary greenhouse gas, which is accountable for about two-thirds of greenhouse effect, and the major source of CO2 emissions is flue gas from coal-fired power plants (Koerner and Klopatek 2002). Therefore, it is urgent to develop the efficient materials and processes for CO2 removal from flue gases resulting from coal burning for the sustainable environment development. Carbon capture, utilization and sequestration (storage) (CCUS) is the most promising alternative for reducing CO2 emissions, and CO2 capture is a critical process (Chu 2009; Hasana et al. 2015).

Currently, proven technologies for the CO2 capture from flue gases mostly rely on absorption, membrane separation, cryogenic distillation and adsorption (Olajire 2010; Li et al. 2013a; Song et al. 2019). Among them, amine scrubbing is currently the most widely used commercial technology for CO2 capture. However, it suffers from some disadvantages, including the high energy consumption of regeneration, the equipment corrosion, and the volatility and degradation of solvent (Rao and Rubin 2002). Adsorption processes is considered as a most promising method for replacing the conventional amine absorption technologies due to its low regeneration energy requirements, no liquid waste, and easy operations (Figueroa et al. 2010). For the adsorption separation technology, the moisture and the higher temperature in flue gases could greatly influenced the CO2 adsorption performance on current adsorbents, and some adsorbents could be easily poisoned and deactivated by gas impurities such as NOx and SOx, making it difficulty in industrial application for the CO2 capture from flue gases. Therefore, the significant research efforts are being undergoing to develop the feasible efficient adsorbents to capture CO2.

Presently, it has been reported that a variety of solid adsorbents (Wang et al. 2011; Billemont et al. 2017; Rocha et al. 2017; Regufe et al. 2018) including zeolites, activated carbons and metal–organic frameworks (MOFs) display a high adsorption capacity of CO2. However, activated carbons always show the low selectivity, and CO2 adsorption capacities on zeolites and MOFs decrease significantly in the presence of moisture that is essentially present in flue gas, so these absorbents alone are difficult to meet Department of Energy (DoE) requirements of 95% CO2 purity, 90% recovery for effective CCUS from flue gases with a low energy consumption (Nikolaidis et al. 2018). To achieve high CO2 capture performance, amine-functionalized adsorbents have been paid great attentions, and mainly prepared by introducing amine functional groups with high affinity for CO2 onto porous solid materials through a physical impregnation or chemical grafting method (Chen et al. 2014). The prepared composite materials combine the advantages of porous support and amine groups, leading to the enhanced mass transfer, high CO2 capture capacity, and desirable CO2 adsorption selectivity. Nonetheless, the mechanical property and thermal stability need to be further improved for amine-based adsorbents.

Recently, adsorbents based on the alkali metals and alkali earth metal oxides have been paid great attentions for the CO2 capture (Yang et al. 2010; Iruretagoyena et al. 2015; Song et al. 2016; Zhao et al. 2018; Kodasma et al. 2019). Among them, MgO is a promising adsorbent to capture CO2 owing to its appropriate basic strength and at a moderate operation temperature than that of CaO, and more cost-effective than Li2O (Jiao et al. 2013; Kim et al. 2014). However, pure MgO exhibits the drawbacks of low surface area and surface formation of a termination layer of carbonate hindering the interaction between MgO and CO2 (Liu et al. 2008). To overcome these drawbacks, extensive efforts have been devoted to the development of various MgO based porous adsorbents by introducing MgO into porous supports such as molecular sieves (Zukal et al. 2013), Al2O3 (Dong et al. 2015), activated carbon (Shahkarami et al. 2016) and TiO2 (Hiremath et al. 2017), allowing more efficient contact between the MgO and CO2. For instance, the supported MgO/γ-Al2O3 adsorbent achieves a CO2 adsorption capacity of 0.8 mmol/g at low-temperature of 298 K and 1 bar (Han et al. 2015). Generally, the MgO-loaded porous adsorbents are mainly prepared by impregnating the porous supports with an aqueous solution of magnesium salt, and then calcining magnesium salt to MgO at high temperatures. The MgO-based porous adsorbents prepared with the impregnation method, apart from the intensive preparation processes, always have lower valid MgO loadings owing to the solvent effect. Solid-state heat dispersion is a simple and effective approach to introduce metal salt or metal oxide into the channels of the porous solid materials, and can achieve more valid metal loadings on the surfaces of the supports compared with the conventional impregnation method (Huang et al. 2015; Gao et al. 2016a). And, after activation at high temperatures, the guest species can highly dispersed on the surfaces of the supports. Herein, zeolite Y supported MgO (MgO@Y) for CO2 capture was prepared by a solid-state heat dispersion method without any solvents and organic templates. The aim of the work is to develop a MgO-based CO2 adsorbent by a facile preparation method. The obtained adsorbents were investigated for CO2 adsorption, and their structural and physical properties were characterized by X-ray powder diffraction (XRD) and N2 adsorption/desorption isotherms. In addition, the CO2 cyclic stability and the isosteric heat of CO2 adsorption on the prepared adsorbent were also studied.

2 Experimental

2.1 Materials

Mg(NO3)2·6H2O (AR) and Zeolite Y (HY, SiO2/Al2O3 = 7, Na2O = 0.8%) were purchased from Qingdao Zhengye Fine Chemical Research Institute and the Catalyst Plant of Nankai University, respectively. CO2, N2 and helium gases were purchased from Qingdao Dehai Gas Co., Ltd. All the gases had purities of or above 99.99%.

2.2 Sample preparation

MgO@Y adsorbents were prepared by a solid-state dispersion method. In a typical run, 3.846 g Mg(NO3)2·6H2O (15 mmol) and 5.000 g zeolite Y powder was mixed and grinded for 30 min, and then the mixture was activated at 300 °C for 4 h and 550 °C for 4 h in air. Thereby, the sample with magnesium loadings of 3.0 mmol/g zeolite Y was obtained. According to these steps, the sample with different magnesium loadings of 2.0, 3.0, 4.0 and 5.0 mmol/g zeolite Y were prepared and denoted as Mg(NO3)2(X)@Y and MgO(X)@Y before and after the activation, where X represents the magnesium loadings in mmol per gram zeolite Y.

2.3 Characterizations

Powder X-ray diffraction patterns were obtained on a Rigaku D/max-2500 diffractometer employing the graphite filtered Cu Kα radiation (λ = 0.154 nm) in the 2θ ranges from 5° to 80°. Nitrogen adsorption–desorption isotherms of the samples were measured on a Micromeritics ASAP 2020 system at liquid N2 temperature (77 K). Prior to the measurements, all the samples were evacuated at 573 K for 4 h. The Brunauer–Emmett–Teller (BET) method was utilized to calculate the specific surface areas of the samples from the isotherms. The total pore volumes were estimated from the adsorption equilibrium isotherms at a relative pressure (p/p0) of 0.99.

2.4 Adsorption measurements

The adsorption isotherms were measured using a static volumetric method (Gao et al. 2016b, c). The apparatus mainly consisted of an adsorption unit, a loading cell, two pressure transducers, a temperature-controlled electric oven, a mechanical vacuum pump. The volumes of the adsorption unit and the loading cell were measured by water filling and the expansion of helium gas, respectively. Before each measurement, 5 g adsorbent inside the adsorption unit was regenerated in situ at 473 K under the vacuum realized by a mechanical pump to remove the gases impurities retained inside. Usually, most of the adsorbent needs to be regenerated for several hours under vacuum or inert gas purge at high temperatures to achieve a complete sample regeneration. In this work, the pressure value of the adsorption unit reached zero in about 5–10 min under the above regeneration condition and remained unchanged for a long time, but 30 min of regeneration time was used to guarantee that the regeneration fully completed. After regeneration, the adsorption unit was cooled down to the set adsorption temperature. The void volume of the adsorption unit was measured by the expansion of helium gas. After that, CO2 was introduced into the loading cell, and its pressure and temperature were recorded when the values were stabilized. Then the valve connected the adsorption unit and the loading cell was opened, allowing CO2 to contact with the adsorbents. The pressure value of the adsorption unit and loading cell usually reach a constant value in about 10 min, but 30 min of adsorption time was used for each point to guarantee that the adsorption fully reached to equilibrium. After the pressure value was recorded, the valve was closed and an additional amount of CO2 was fed into the loading cell, and the same operation was repeated to obtain another adsorption equilibrium point at a higher pressure. The adsorption measurements of CO2 were carried out at the temperature range of 303 to 333 K and pressures up to 500 kPa. The adsorbed amounts were calculated from the experimental temperatures and pressures using the mass balance equation derived from the equation of state before and after adsorption equilibrium.

2.5 Adsorption theories

2.5.1 Adsorption equilibrium isotherm equations

Many adsorption isotherm models including Langmuir (Langmuir 1918), Freundlich (Freundlich 1906), Sips (Langmuir–Freundlich) (Sips 1948), Toth (Toth 1971), Dual-site Langmuir (DSL) (Kapoor et al. 1990), etc., have been established to investigate the adsorbate-adsorbent interactions. The DSL model, which is an extension of the Langmuir model, is a thermodynamically consistent model, and has been widely used to model the adsorptive processes, especially when the adsorbent surface exhibits a heterogeneous nature. In this study, the DSL model was chosen for correlating the isotherm data as the model is easy to implement in the IAST algorithm with a good accuracy, and is given by

$$q = q_{A,sat} \frac{{b_{A} p}}{{1 + b_{A} p}} + q_{B,sat} \frac{{b_{B} p}}{{1 + b_{B} p}},$$
(1)

where q is the adsorption amounts (mmol/g), p is the equilibrium pressure (kPa), and qA,sat and qB,sat are the saturated adsorption amounts of dual sites A and B (mmol/g), respectively. bA and bB are the affinity coefficients of sites A and B (kPa−1), respectively, and their temperature dependence are described by

$$b_{A} = b_{A0} \exp ( - \Delta H_{A} /RT),$$
(2)
$$b_{B} = b_{B0} \exp ( - \Delta H_{B} /RT),$$
(3)

where bA0 and bB0 are the fitting constants, R is the universal gas constant (J mol−1 K−1), T is temperature (K), and ΔH is the heat of adsorption (kJ/mol).

2.5.2 Adsorption selectivity

Multi-component adsorption equilibrium data is very important for designing industrial gas adsorption separation processes. However, current measurement techniques have some difficulty in measuring the adsorption equilibrium data for the mixed gases. Therefore, in the past few decades, a series of theoretical models have been developed to predict multi-component adsorption equilibrium data by single-component adsorption data. Although the correlative multi-component prediction models has some limitations (Rao and Sirca 1999; Li and Tezel 2008), they are widely employed to predict multi-component adsorption equilibrium data by pure component adsorption data. Among them, the Ideal Adsorbed Solution Theory (IAST) presented by Myers and Prausnitz (1965) is the most common model, and many studies affirm the accuracy of IAST in providing good mixture adsorption predictions (Wang et al. 2014; Avijegon et al. 2018). In this work, the DSL model was combined with IAST to calculate the adsorption selectivity for the binary mixture. The adsorption selectivity of component 1 over component 2 (S1/2) for a binary mixture is commonly defined as (Bahamon and Vega 2016):

$$S_{ 1 / 2} = \frac{{x_{1} /x_{2} }}{{y_{1} /y_{2} }},$$
(4)

where x and y represent the mole fractions of components in the adsorbed phase and gas phase, respectively. In this work, components 1 and 2 are CO2 and N2, respectively.

2.5.3 Isosteric heat of adsorption

From the adsorption isotherms collected at different temperatures, the isosteric heat of adsorption can be calculated by the Clausius–Clapeyron equation as follows (Choma et al. 2016):

$$Q_{st} = RT^{2} \left( {\frac{\partial \ln p}{\partial T}} \right)_{{q_{a} }} ,$$
(5)

where Qst is the isosteric heat of adsorption (kJ/mol), and qa is a specific adsorbate loading (mmol/g).

3 Results and discussion

3.1 Characterization of MgO@Y adsorbents

Figure 1 shows the XRD patterns of pure MgO, zeolite Y and the supported Mg(NO3)2 samples with different magnesium loadings before and after activation at high temperatures. It can be seen that, before activation, the Mg(NO3)2@Y sample show the diffraction peaks of Mg(NO3)2 at 2θ values of 15.3°, 20.2°, 21.5°, 27.2° and 30.7° (Ling et al. 2017). After activation, for the MgO@Y samples, only the typical diffraction peaks of zeolite Y are observed, and neither Mg(NO3)2 nor MgO are observed for the MgO@Y samples, suggesting that MgO is highly dispersed on the surfaces of zeolite Y support. The absence of MgO reflections in XRD patterns might be owing to that the smaller crystallites of highly dispersed MgO are undetectable by XRD (Chen et al. 2010). A high dispersion of the magnesium salts on zeolite Y made the active components of MgO tend to spread rather than agglomerate on the surfaces of support.

Fig. 1
figure 1

XRD patterns of MgO, zeolite Y and Mg(NO3)2 loaded zeolite Y before and after activation

Full size image

Figure 2 shows the N2 adsorption–desorption isotherms of zeolite Y support and MgO@Y samples with different magnesium loadings at 77 K. It can be seen that the N2 adsorption amounts apparently decrease with increasing the magnesium loadings. The BET surface area, pore volume and average pore diameters of zeolite Y support and MgO@Y samples are summarized in Table 1. The specific surface area and pore volume of the samples decrease gradually with the increasing magnesium loadings, resulting from the occupation of the partial surface areas and the pore volumes of zeolite Y by MgO. For the MgO@Y samples, when increasing the magnesium loading from 2.0 to 3.0 mmol/g, the specific surface area and pore volume of the samples decrease slowly, and when further increase the magnesium loadings, these values decrease sharply, indicating that the specific surface area and pore volume of the samples are occupied more seriously by MgO when the magnesium loadings is higher than 3.0 mmol/g. In addition, the average pore diameter of the samples increases with an increase in magnesium loadings. As more micropores are occupied, it results in an increased average pore sizes because the percentage of the available mesopores in zeolite Y increases (Ramli and Amin 2015).

Fig. 2
figure 2

N2 adsorption/desorption isotherms at 77 K on zeolite Y and MgO@Y samples with different magnesium loadings

Full size image
Table 1 Textural properties of zeolite Y and MgO@Y samples
Full size table

3.2 Adsorption performance of CO2 on MgO@Y samples

Figure 3 illustrates the effect of the magnesium loadings on the adsorption performance of CO2. It can be seen that pure MgO has a low CO2 adsorption capacity, resulting from the low surface area-to-volume ratio. The support of zeolite Y shows a CO2 adsorption capacity of 2.18 mmol/g at 500 kPa, caused by the physical interaction between CO2 and zeolite Y. When introducing MgO, the MgO@Y samples show the higher adsorption capacities compared to the zeolite Y support, implying that the introduction of MgO on zeolite Y is crucial to obtain adsorbents with high CO2 adsorption capacity. When the magnesium loadings increase to 3.0 mmol/g zeolite Y, the CO2 adsorption amounts at 500 kPa increase from 2.18 to 2.78 mmol/g. When the magnesium loadings further increase to 5.0 mmol/g zeolite Y, the CO2 adsorption capacity decreases gradually. This may be due to the accumulation of excessive MgO on the surfaces of samples, resulting in the decrease of the effective active sites and the blockage of pore channels. The results indicate that the monolayer dispersion capacity of MgO on zeolite Y is 3 mmol/g. When the magnesium loadings is higher than 3 mmol/g, the excess magnesium species may cumulate on the monolayer, which is in good agreement with N2 adsorption–desorption results. In addition, it can been that the CO2 adsorption capacities of all the MgO@Y samples are higher than that of the zeolite Y support at low pressures, and the CO2 adsorption amounts on MgO(3.0)@Y at 20 kPa increase from 0.15 to 0.35 mmol/g, ascribed to the occupation of the partial pore volumes and surface areas of the zeolite Y support by MgO, resulting in that the physical absorption sites for CO2 on the adsorbents is gradually replaced by the chemical adsorption between MgO on the adsorbents with CO2 molecules with increasing the magnesium loadings.

Fig. 3
figure 3

Adsorption isotherms of CO2 on MgO, zeolite Y and the MgO@Y adsorbents with different magnesium loadings at 303 K

Full size image

For the adsorption separation of CO2 from gas mixtures, besides adsorption capacity, the adsorption selectivity is also a key factor for the separation performance. Figure 4a gives the experimental and correlated adsorption isotherms of pure CO2 and N2 on the MgO(3.0)@Y adsorbent at 303 K, and the DSL fitting parameters are listed in Table 2. It can be seen that the adsorption capacity of CO2 on MgO(3.0)@Y is significantly higher than that of N2, which is ascribed to the stronger interaction between the highly dispersed MgO on the adsorbent and CO2 and the weaker van der Waals and electrostatic interactions of N2 with the adsorbent. Based on the single gas adsorption isotherms, the CO2/N2 adsorption selectivity were predicted by IAST for the binary mixture with CO2/N2 molar ratio of 15:85, which is the typical component of flue gases in a coal fired power plant, and depicted in Fig. 4b. It can be seen that the CO2/N2 adsorption selectivity decreases gradually from 39 to 32 with an increase in adsorption pressure, and at a given pressure of 100 kPa, the CO2/N2 adsorption selectivity is as high as 37. In addition, the CO2/N2 adsorption selectivity of the prepared MgO(3.0)@Y adsorbent was compared with that of common porous materials, such as MOFs, carbon materials and zeolites, and the comparison results are summarized in Table 3. Overall, in Table 3, MOFs has higher CO2/N2 adsorption selectivity than carbon materials and zeolites, however, MOFs always suffer from the expensive cost, durability and mechanical strength problems, limiting its industrial application (Lee and Park 2015). In comparison, the MgO(3.0)@Y adsorbent in this work exhibits a higher CO2/N2 adsorption selectivity than other adsorbents except for zeolite 13X in Table 3, and the CO2/N2 adsorption selectivity of MgO(3.0)@Y adsorbent is almost 3.7 times of that of zeolite Y, implying that the introducing of MgO onto the zeolite Y support is crucial to obtain adsorbents with a high CO2/N2 adsorption selectivity. For the zeolite 13X adsorbent, the CO2/N2 adsorption selectivity was greatly influenced by the adsorption pressure, and decreased significantly to ~ 5 at 500 kPa (Belmabkhout and Sayari 2009), which was lower than that of 32 of MgO(3.0)@Y in this work. The higher CO2/N2 adsorption selectivity suggests that the MgO(3.0)@Y adsorbent can preferentially adsorb CO2 over N2 in a binary mixture of CO2/N2.

Fig. 4
figure 4

Adsorption isotherms of CO2 and N2 on MgO(3.0)@Y at 303 K (a) and IAST-predicted CO2/N2 adsorption selectivity for the binary mixture with CO2/N2 molar ratio of 15:85 (b)

Full size image
Table 2 Dual-site Langmuir (DSL) fitting parameters for CO2 and N2 adsorption on MgO(3.0)@Y
Full size table
Table 3 Comparison of CO2/N2 selectivity at 100 kPa and heat of adsorption of CO2 adsorption on different adsorbents
Full size table

For potential application, an ideal adsorbent should not only display the high adsorption capacity and high adsorption selectivity, but also exhibit a good cyclic stability performance. Thus, the adsorption/desorption cycles with the adsorption at 303 K and desorption at 473 K for 30 min under vacuum were repeated for ten times on MgO(3.0)@Y, and the result is depicted in Fig. 5a. It can be seen that, in the ten adsorption/desorption cycles, the CO2 adsorption capacities are nearly identical. In addition, to mimic a vacuum pressure swing adsorption (VPSA) process, the adsorption/desorption cycles with the adsorption and desorption at the same temperature of 303 K were repeated for five times on MgO(3.0)@Y. The results in Fig. 5b show that the CO2 adsorption capacities are nearly identical in the five cycles at the operation temperature of 303 K, and the reversible CO2 adsorption capacity is 2.46 mmol/g, which is close to that of 2.78 mmol/g with the desorption at 473 K. The result implies that the adsorbed CO2 on MgO(3.0)@Y could be desorbed at ambient temperature under vacuum, which is attributed to the relatively low isosteric heat of CO2 adsorption on the MgO(3.0)@Y sample (see Fig. 6). The results indicate that the adsorbent displays a good cyclic stability for CO2 adsorption, demonstrating that this adsorbent can be repeatedly used for the CO2 adsorption and desorption in CO2 separation processes with the CO2 adsorption capacity being not changed.

Fig. 5
figure 5

The cyclic CO2 adsorption capacity at 500 kPa on MgO(3.0)@Y with the adsorption at 303 K and desorption at 473 K (a), and with the adsorption and desorption at 303 K (b)

Full size image
Fig. 6
figure 6

CO2 adsorption isotherms at different temperatures (a) and isosteric heats of CO2 adsorption (b) on MgO(3.0)@Y sample

Full size image

The above results indicate that the prepared MgO@Y adsorbent with a high CO2 adsorption capability, high adsorption selectivity and good stability is a promising adsorbent for the effective separation of CO2 from gas mixtures.

Isosteric heat of adsorption is an important parameter for investigating the interactions between adsorbate and adsorbent, and providing useful information to design and operate the practical adsorption process. In this work, CO2 adsorption isotherms at different temperatures (303 K, 313 K, 323 K and 333 K) presented in Fig. 6a were used to calculate the isosteric heat of CO2 adsorption on MgO(3.0)@Y sample. The DSL models were used to correlate the experimental data, and the fitting parameters are listed in Table 2. According to Eq. (5), the isosteric heats of adsorption as a function of the CO2 loading were calculated, and depicted in Fig. 6b. From Fig. 6a, the CO2 adsorption capacity decreases with increasing the adsorption temperatures, indicating that the CO2 adsorption on MgO(3.0)@Y is an exothermic process. From Fig. 6b, the isosteric heats of CO2 adsorption on MgO(3.0)@Y decrease gradually from 27.8 to 20.0 kJ/mol with increasing the CO2 adsorption amounts to 1.0 mmol/g. This variation may be associated with the heterogeneity of surface property for the material. The calculated isosteric heat value of adsorption and the variation is similar to that on MgO/C adsorbent obtained by Li et al. (2013b). This relatively low isosteric heat, which is slightly higher than that of zeolite Y (see Table 3), suggests that the interaction strength between CO2 and MgO-based adsorbents is relatively weaker than the strong acid–base interactions, resulting in a moderate operation temperature for CO2 capture and a low regeneration energy consumption using MgO-based adsorbents.

4 Conclusions

A series of zeolite Y supported MgO (MgO@Y) adsorbents for selective adsorption for CO2 has been successfully prepared using Mg(NO3)2 as precursor via a facile solid-state heat dispersion approach without any solvents and organic templates. Mg(NO3)2 supported on zeolite Y can be converted to highly dispersed MgO after activation at high temperatures. The obtained MgO@Y sample with the magnesium loading of 3 mmol/g zeolite Y achieves a high CO2 adsorption capacity of 2.78 mmol/g and a high CO2/N2 adsorption selectivity of 32 at 303 K and 500 kPa. In addition, this adsorbent also shows an excellent cyclic stability in ten adsorption/desorption cycles for CO2. The MgO(3.0)@Y adsorbent with good CO2 adsorption capacity, good adsorption selectivity, and excellent cyclic stability as well as its facile preparation process shows an attractive candidate for CO2 capture from flue gas vents.