Lithium silicates synthetized from iron and steel slags as high temperature CO₂ adsorbent materials

Abstract

The use of solid wastes and industrial by-products to prepare CO₂ adsorbents is an alternative to conventional reagent grade raw materials that has recently gained interest. Among waste materials, slag has a high content of silica and calcium and is the largest solid by-product from iron and steel industry, thus its use can reduce the production costs of CO₂ adsorbent materials, such as lithium silicates, which are applied in capture processes at high temperatures. Li₄SiO₄ has potential applications in post-combustion CO₂ capture as well as in H₂ production by sorption enhanced steam reforming process. In this study, Li₄SiO₄ was prepared using solid-state reaction and two iron and steel slags as SiO₂ sources to evaluate their characteristics and CO₂ capture capacities. The slag-derived lithium silicates (S1-Li₄SiO₄ and S2-Li₄SiO₄) were characterized by XRD, adsorption-desorption N₂ and SEM. Different capture tests at CO₂ partial pressures ($P_{{{\text{CO}}_{2} }}$) of 0.05, 0.10, 0.15 and 0.20 were performed using thermogravimetric (TG) and temperature programmed (TPC-TPDC) techniques. The kinetic parameters of the CO₂ capture process were obtained by fitting the experimental results to the Avrami–Erofeev model. Finally, the cyclic behavior of S1-Li₄SiO₄ and S2-Li₄SiO₄ was analyzed in $P_{{{\text{CO}}_{2} }}$ of 0.2 and 0.05. XRD patterns showed that Li₄SiO₄ was the main crystal phase (60 wt%) present in S1-Li₄SiO₄ and S2-Li₄SiO₄ in addition to calcium phases such as Li₂CaSiO₄, Ca₃SiO₅ and CaO. According to the TG and TPC-TPDC tests, the derived lithium silicates showed CO₂ uptake three times greater than the values recorded for Li₄SiO₄ (134 mgCO₂/g sorbent for S1-Li₄SiO₄) produced from pure reagents, at $P_{{{\text{CO}}_{2} }}$ between 0.2 and 0.05 and 650 °C. Furthermore, these materials had kinetic constants at least one order of magnitude higher than those reported for Li₄SiO₄, at the aforementioned operating conditions. Both materials exhibited an excellent stability during 20 cycles of CO₂ adsorption/desorption. These results showed that slags can be used as silica source to produced adsorbents with better performance and stability in the CO₂ capture process at high temperature than the one of Li₄SiO₄ produced from pure reagents, at $P_{{{\text{CO}}_{2} }}$ of 0.2–0.05.

1 Introduction

Industrial by-products are materials produced during the manufacture of a primary product and its use has been encouraged in order to reduce CO₂ emissions, avoid disposing wastes in landfills, increase resource efficiency and generate revenue. Over the past 20 years, the use of the steel industry’s by-products has increased significantly. The main by-products generated during iron and steel production are slags (90% by mass), dust and sludge. According to Euroslag, ferrous slag is considered a by-product in liquid state, directly after its manufacture, with or without processing steps; it is first considered as waste but ceases to be a residue after a number of recovery measures (EUROSLAG 2008). On average, the production of one tonne of crude steel results in around 170 or 400 kg of slags for electric arc furnace (EAF) or blast furnace (BF) routes, respectively. It is estimated that global iron slag output in 2017 was around 300 to 360 million tons, while steel slag was around 170 to 250 million tons (WorldSteel 2016). In 2016, the iron and steel slag production in Europe was of 41 Mt where the mainly uses were cement and concrete additive (46.8%), road construction (29.8%), metallurgical use (6.8%), hydraulic engineering (1%), fertilizer (1.2%), interim storage (4.2%), landfill (5.7%) among others.

At the same time, there is a continuous increase of the worldwide energy consumption and therefore of CO₂ emissions. The average CO₂ concentration in the world on February 2019 was 411.75 ppm, 47% higher than prior to industrial revolution (280 ppm) (U.S. Department of Commerce 2019). In addition to the improvement in energy and process efficiencies in the industry, the reduction of anthropogenic greenhouse gas emissions is one of the main challenges for the coming years. Besides to energy efficiency, other strategies such as emission efficiency (including fuel and feedstock switching and CO₂ capture and storage (CCS), material use efficiency (e.g., less scrap), recycling and re-use of raw materials and products are required (IPCC 2014).

Post-combustion capture provides a short-term approach to mitigate the concentration of the CO₂, generated by power plants, metallurgical and cement industries, among others. In post-combustion, the gas stream contains mainly CO₂ and N₂, with a partial pressure of CO₂ of around 0.05–0.2 bar, where the temperature range depends on the concerned industrial sector. CO₂ capture has also been used in the sorption enhanced methane steaming reforming (SESMR) process, which aims to enrich the hydrogen concentration in the gas stream (95%) (Albo Sánchez 2015; Yancheshmeh et al. 2015). In post-combustion and also in SESMR, high-temperature (450–700 °C) solid sorbents are more cost-effective and efficient than low-temperature amine-based materials as the direct separation of CO₂ from the high-temperature exhaust gases saves large amounts of energy (Dou et al. 2016; Ochoa-Fernández et al. 2005; Yancheshmeh et al. 2015). Recently, Garcia et al. (2017), carried out the integration of lithium looping post-combustion carbon capture technology in a NGCC (natural gas combined cycle) power plant following the EBTF methodology. The results showed that lithium looping have approximately 0.6 percentage points lower energy penalty compared to the best performing chemical absorption capture system. Among high-temperature CO₂ sorbents, lithium-based materials have aroused great interest due to their high CO₂ capture capacity, selectivity, fast kinetics and good regeneration properties. Lithium orthosilicate (Li₄SiO₄), the most studied lithium based ceramic, shows suitable reactivity, thermal stability during several sorption-desorption cycles and the fastest CO₂ sorption rate over a wide range of temperatures and CO₂ concentrations (Albo Sánchez 2015; Amorim et al. 2016; Hu et al. 2019; Kaniwa et al. 2018; Rodríguez-Mosqueda and Pfeiffer 2010; Zhang et al. 2019). Kato et al. (2005) reported that the CO₂ adsorption on Li₄SiO₄ was up to 30 times faster than on Li₂ZrO₃ at 500 °C and 20 vol% CO₂. Although its theoretical adsorption capacity (8.34 mmol CO₂/g) is lesser than the value of calcined limestone (17.8 mmol CO₂/g), this material is a promising adsorbent since shows lower values of energy requirements for regeneration (< 750 °C vs 950 °C) and capture temperature of CO₂ emissions, with respect to CaO (Chen et al. 2016; Kato et al. 2005; Seggiani et al. 2013). The double shell model is widely accepted to explain the CO₂ sorption mechanism on Li₄SiO₄, as well as on other alkaline ceramics, and comprises two stages (e. g. Na₂ZrO₃, Li₅AlO₄, Li₈SiO₆) (Alcérreca-Corte et al. 2008; Avalos-Rendón et al. 2009; Castillo Villa et al. 2015; Durán-Muñoz et al. 2013; Martínez-dlCruz and Pfeiffer 2012; Zhang et al. 2019). In the first stage, the CO₂ reacts with the surface of lithium silicate particles to form an external shell composed of Li₂CO₃ and Li₂SiO₃, according to the following reaction (Amorim et al. 2016; Seggiani et al. 2011; Zhang et al. 2019):

$${\text{Li}}_{{\text{4}}} {\text{SiO}}_{{\text{4}}} \left( {\text{s}} \right)\, + \,{\text{CO}}_{{\text{2}}} \left( {\text{g}} \right){\text{ }} \to {\text{ Li}}_{{\text{2}}} {\text{CO}}_{{\text{3}}} \left( {\text{s}} \right)\, + \,{\text{Li}}_{{\text{2}}} {\text{SiO}}_{{\text{3}}} \left( {\text{s}} \right)$$

In a second stage, the reactants have to diffuse through the external shell to react with each other. The bulk diffusion process begins, and the CO₂ continues reacting with the Li₄SiO₄ particles that remain unreacted. This last stage has been explained from two different perspectives; some works propose that the CO₂ diffusion through the external shell is the dominant process, while others put forward that the intercrystalline diffusion of Li⁺ and O²⁻ ions is the main phenomena (Kato et al. 2005; López Ortiz et al. 2014). To enhance its CO₂ chemisorption Li₄SiO₄ has been modified with different elements such as K, Na, Al, Fe, and V, among others. This promotes the diffusion of lithium and oxygen ions and/or CO₂ due to the formation of vacancies, eutectic phases, or different lithium secondary phases after carbonation process (Albo Sánchez 2015; Gao et al. 2017; Gauer and Heschel 2006; Ortiz-Landeros et al. 2012; Seggiani et al. 2011; Wang et al. 2017a, b). In addition, a Ca-Li₄SiO₄ sorbent was synthetized, considering the high CO₂ capture capacity demonstrated by CaO (0.78 ${\text{g}}_{{{\text{CO}}_{2} }}$ g/_CaO); it was shown that the transformation of the Ca species from Ca₂SiO₄ to Li₂CaSiO₄ during the CO₂ adsorption process promotes the transfer of CO₂ to Li₄SiO₄, and then the inverse process favors the CO₂ desorption (Chen et al. 2016).

Moreover, there is enough evidence that the different sources of silica used in the synthesis of Li₄SiO₄ produce changes in the particle size and microstructure, thus generating characteristic behavior of CO₂ sorption (Hu et al. 2019). Bearing in mind that the development of low-cost CO₂ adsorbent materials will undoubtedly enhance the competitiveness of CO₂ capture technologies and other applications such as sorption enhanced steam fuel reforming for hydrogen production and thermochemical energy storage. Some scientific papers have been published on the use of residues and by-products, generated from industrial and agricultural operations, as raw materials for CO₂ adsorbents. These materials are low-cost and abundant, and therefore their use may contribute to reduce the total costs of CO₂ capture technologies and at the same time, show promising CO₂ capture capacities. For this reason different silica sources such as fly ash, rice husk ash and, recently, blast furnace slag, have been used to prepare lithium orthosilicate; the results indicate that, as expected, the adsorption capacity is influenced by the silica source (Olivares-Marín and Maroto-Valer 2012; Sanna et al. 2015; Wang et al. 2011, 2018). Olivares-Marin et al. (2010) investigated lithium-based sorbents made from fly ashes for CO₂ capture at high temperatures. The obtained Li₄SiO₄-based sorbents did not show CO₂ adsorption in 100 vol% CO₂, however, the addition of K₂CO₃ enhanced the sorption capacity to 107 mg CO₂/g sorbent (at 600 °C and 40 mol% K₂CO₃). In addition, Izquierdo et al. (2018), studied the effect of the silica source (pure reagent or fly ash) and the preparation method (solid state reaction and precipitation method) on the CO₂ uptake of the derived Li₄SiO₄. The material prepared from fly ash and Li₂CO₃ by solid state reaction showed just a 5.9 wt% of CO₂ uptake (in 92% CO₂), as the calcium silicates formed limit the CO₂ capture on the CaO. In another work performed by Wang et al. (2018), silica extracted with an acid leaching method from a blast furnace slag was also used to prepare Li₄SiO₄ through the solid state reaction method with Li₂CO₃ at 873 °C. The adsorption capacity (in pure CO₂) obtained at 600–650 °C was 100.8 mg CO₂/g sorbent (28% conversion) and increased to 98% conversion at 700 °C. This was associated to the small particle size and metal impurities present in the slag, such as potassium and calcium.

Based on the above information, so the use of waste or by-product materials as source of silica to prepare Li₄SiO₄ has aroused growing interest due to its potential applications in high temperature post-combustion CO₂ capture as well as in H₂ production by SESR process. In addition to reducing the cost of the sorbent, the use of iron and steel slags, as SiO₂ source, also introduces other elements in to Li₄SiO₄ crystal lattice which could modify the CO₂ capture capacity and the kinetic behavior at low CO₂ partial pressure. Therefore, in this work Li₄SiO₄ was prepared from two different iron and steel slags by solid-state reaction method. The slag-derived lithium silicates were characterized and tested as CO₂ adsorbents at $P_{{{\text{CO}}_{2} }}$ of 0.05, 0.10, 0.15 and 0.20. Furthermore, the kinetic parameters of the CO₂ process and the cyclic performance of the adsorbents were also determined and compared with the Li₄SiO₄ prepared with analytical grade reagents.

2 Experimental

2.1 Sorbent preparation and characterization

Two slag samples, named S1 and S2, were obtained from different iron and steel Mexican industries. S1 is a blast furnace slag while S2 is an electric arc furnace slag. The chemical composition was previously determined by X-ray fluorescence and is shown in Table 1 (Mercado-Borrayo et al. 2013).

Table 1 Chemical composition of slags by X-ray fluorescence [wt%] (Mercado-Borrayo et al. 2013)

The slag-derived lithium silicates were prepared using the solid-state reaction method by mixing Li₂CO₃ reagent grade (99.8%, Meyer) with the iron and steel slags (S1 and S2), as SiO₂ sources, in a Li₂CO₃:SiO₂ molar ratio of 2.1:1 (an excess of 10 mole percent was added due to the lithium sublimation). Both mixtures were calcined at 850 °C for 8 h, and they were named S1-Li₄SiO₄ and S2-Li₄SiO₄. For comparison purposes, pure lithium orthosilicate (Li₄SiO₄) was also prepared from Li₂CO₃ and SiO₂ reagent grade (325 mesh, 99.5%, Aldrich) using the same calcination conditions. All calcined materials were homogenized in an agate mortar to identify their mineralogical phases by the X-ray diffraction (XRD) technic in an Empyrean diffractometer with CuKα radiation and PIXcel3D detector. The XRD measurements were carried out over a 2-θ angle of 5°–70° in steps of 0.003° and 40 s integration time. The identification of phases was performed using the Inorganic Crystal Structure Database (ICSD). The normalized Relative Intensity Ratio (RIR) method was used to conduct a semi-quantitative determination of these phases. In the RIR method, the variation of peak intensities with concentration is considered nonlinear and the former is derived by standards Chung (1974). Recognized that if all phases in a mixture are known and if RIR is known for all of those phases, then the addition of the fractions of all the phases must be equal to 1. N₂ adsorption–desorption isotherms of the materials were obtained on a Minisorp II instrument (Bel Japan) at 77 K using the multipoint technique (N₂ from Praxair, grade 4.8) and the specific surface area was calculated by the Brunauer–Emmett–Teller (BET) method. Prior to analysis, the samples were degassed at room temperature in N₂ flow for 12 h. The microstructural characterization was completed with scanning electron microscopy (SEM); the backscattered electron images were obtained from a Philips XL 20 instrument.

2.2 CO₂ sorption tests

The CO₂ capture capacity of the resultant slag-derived lithium silicates was evaluated by thermogravimetric and temperature programmed techniques. Before the CO₂ sorption tests, the samples were pretreated in N₂ flow (60 mL/min) at 700 °C to eliminate any previous carbonation. Then, temperature programed carbonation-decarbonation (TPC-TPDC) experiments were carried out using a Belcat B (Bel Japan) equipped with a thermal conductivity detector (TCD), which records the changes in CO₂ concentration in the outlet gas stream. In these experiments, 50 mg of sorbents were kept in contact with a gas stream containing 60 mL/min of 5%mol CO₂ (He balance, Praxair, certificated standard), and were heated up to 800 °C using a temperature ramp of 5 °C/min. In addition, thermogravimetric experiments were performed on a Labsys Evo TG analyzer from Setaram, with 20 mg of sample. The dynamic performance of CO₂ sorption was obtained by heating the samples from room temperature to 850 °C at 5 °C/min using $P_{{{\text{CO}}_{2} }}$ of 0.2 and 0.05 (balance with N₂). The CO₂ sorption over the time was measured between 580 and 700 °C. In each test, the temperature was increased in N₂ flow (60 mL/min); afterwards, the flow was switched to a $P_{{{\text{CO}}_{2} }}$ of 0.05 for 180 min. Besides, the influence of different $P_{{{\text{CO}}_{2} }}$ values in the gas mixture ($P_{{{\text{CO}}_{2} }}$ = 0.2, 015, 0.1 and 0.05) was evaluated at the best CO₂ sorption temperature for each material.

2.3 Cyclic performance tests

The cyclic performance of each material was analyzed by consecutive stages of sorption-desorption. Sorption was carried out in $P_{{{\text{CO}}_{2} }}$ of 0.2 and 0.05 at 600 and 650 °C for S2-Li₄SiO₄ and S1-Li₄SiO₄, respectively, and the desorption process was performed at 750 °C in N₂ flow.

3 Results and discussion

3.1 Slag characterization

Figure 1 shows the XRD patterns of iron and steel slags. The XRD profiles of both slags presented characteristic peaks of SiO₂, in addition to aluminite (Al₂SO₄(OH)₄·7H₂O) and tricalcium silicate (Ca₃SiO₅) in S1, and dicalcium silicate (Ca₂SiO₄) and brownmillerite (Ca₂(AlFe)₂O₅) in S2. The textural characterization of both slags is shown in Fig. 2. The N₂ adsorption−desorption isotherms correspond to type II, with a narrow H3-type hysteresis loop, according to the IUPAC classification (Lowell et al. 2004); this behavior corresponds to nonporous materials. The specific surface area was calculated using the BET method with values for S1 and S2 of 4.4 and 1.2 m²/g, respectively.

Fig. 1

XRD patterns of metallurgical iron and steel slags

Fig. 2

N₂ adsorption–desorption isotherms and SEM images of metallurgical slags

In addition, the morphology of the particles was observed by SEM. The backscattering electron micrographs corroborated that both slags were nonporous solids, with leaves and needles shapes for S1, and polyhedral particles for S2.

3.2 Sorbents characterization

Figure 3 shows the presence of Li₄SiO₄ on the XRD patterns of slags-derived lithium silicates. Also, other lithium and calcium compounds (CaO, Li₂CaSiO₄, Ca₃SiO₅ and β-LiAlSiO₄) were identified. According to the semiquantitative analysis of the crystalline phases using a normalized RIR method, Li₄SiO₄ was the main phase in both slag-derived products with 60% w/w, while CaO was the secondary phase with 25 and 14%, in S1-Li₄SiO₄ and S2-Li₄SiO₄, respectively. The Li₂CaSiO₄ can be produced during the synthesis process through a chemical reaction between CaO/CaCO₃, Li₂CO₃ and SiO₂ (Chen et al. 2016). Moreover, the calcium phases such as CaO, Ca₃SiO₅ and Li₂CaSiO₄ might enhance the CO₂ uptake in the capture tests at high temperature. Bejarano Peña (2018), obtained a similar percentage of Li₄SiO₄ crystalline phase (59 w/w%) after the synthesis of another batch using the same steel slag.

Fig. 3

XRD patterns of pure Li₄SiO₄ and slag-derived lithium silicates

The N₂ adsorption–desorption isotherms of lithium silicates presented in Fig. 4 correspond to type II isotherms with a very narrow H3-type hysteresis loop, according to the IUPAC classification (Lowell et al. 2004); this behavior is in accordance with the synthesis method which produces nonporous materials. The specific surface area calculated with the BET method was 0.6 m²/g for the pure Li₄SiO₄ and 1.0 and 1.2 m²/g for S1-Li₄SiO₄ and S2-Li₄SiO₄, respectively. These surface area values are similar to those obtained in other alkaline ceramics prepared in the same way and used as CO₂ sorbents.

Fig. 4

N₂ adsorption–desorption isotherms of pure Li₄SiO₄ and slag-derived lithium silicates

3.3 CO₂ sorption tests

The evaluation of the CO₂ capture properties of the synthesized materials started with a dynamic thermogravimetric test in $P_{{{\text{CO}}_{2} }}$ of 0.2 and 0.05, with N₂ balance (Fig. 5). All samples showed a similar behavior in the CO₂ concentrations tested. The weight for the three materials increased slowly in a low temperature range, from around 150 to 300 °C, and then sharply at 350 and 450 °C, in slag-derived materials and pure Li₄SiO₄, respectively. The diffusion process was activated above 450 °C and CO₂ uptake increased in all samples. Li₄SiO₄ reached the maximum adsorption of 56.4 mgCO₂/g material at 555 °C and 29.8 mgCO₂/g material at 631 °C, for $P_{{{\text{CO}}_{2} }}$ of 0.2 and 0.05, respectively. At higher temperatures, the pure Li₄SiO₄ started to release CO₂ rapidly. In addition, this material diminished by 47% its CO₂ capture capacity when the $P_{{{\text{CO}}_{2} }}$ decreases from 0.20 to 0.05. On the contrary, S1-Li₄SiO₄ and S2-Li₄SiO₄ presented outstanding improvements in CO₂ capture capacities, with values at least two times greater than the capacity of pure Li₄SiO₄. For the CO₂ partial pressure of 0.2, the maximum CO₂ uptakes from 600 °C to T ≥ 815 °C for S1-L₄SiO₄ and S2-L₄SiO₄, were 115.8 mgCO₂/g material and 130.8 mgCO₂/g material, respectively. After this temperature, the desorption process began for both materials. A decrease in the CO₂ concentration ($P_{{{\text{CO}}_{2} }}$ = 0.05) diminished the CO₂ adsorption by almost 15%, obtaining values of 99.7 and 115.8 mgCO₂/g material, in S1-L₄SiO₄ and S2-L₄SiO₄, respectively. In addition, when $P_{{{\text{CO}}_{2} }}$ decreased from 0.2 to 0.05, the desorption temperature shifted from ≥ 815 °C to 770 °C in S1-L₄SiO₄ and 720 °C in S2-L₄SiO₄. These thermal shifts of the desorption process can be related to CO₂ adsorption-desorption equilibrium changes generated by the CO₂ concentration in the gas mixture.

Fig. 5

Dynamic thermogravimetric analyses of pure Li₄SiO₄ and slag-derived lithium silicates in CO₂ partial pressures of 0.2 and 0.05

The differences in the CO₂ adsorption capacities, between pure Li₄SiO₄ and slag-derived lithium silicates are related to the presence of calcium phases in the slag-derived materials. Dicalcium and tricalcium silicates can transfer CO₂ to Li₄SiO₄ during the sorption process to generate CaCO₃ and SiO₂, and then CaCO₃ reacts with Li₄SiO₄ to produce Li₂CaSiO₄ and Li₂CO₃ (Chen et al. 2016). It is important to note that the CO₂ adsorption behavior of Li₄SiO₄ is affected by the type of SiO₂ used, the synthesis method and, as expected, by the CO₂ concentration. Some studies showed a very low CO₂ adsorption capacities, such as 40 mgCO₂/g material (Rodríguez-Mosqueda and Pfeiffer 2010; Romero-Ibarra et al. 2013), even in saturated CO₂ atmosphere, while in others works, it is reported that they almost reached the maximum theoretical CO₂ capture capacity, between 300 and 350 mgCO₂/g material (Chen et al. 2016; Gao et al. 2017; Wang et al. 2016, 2017b, 2018).

The sorption behavior was also observed in the TPC and TPDC results, as illustrated in Fig. 6, which is useful to determine the CO₂ adsorption-desorption capacity and the inversion temperature. In the carbonation process for the pure Li₄SiO₄, a double peak could be observed; the first (420 °C) is assigned to the CO₂ adsorption on the surface and the second (500 °C) to the CO₂ adsorption in the bulk of the material controlled by diffusive processes. This produces the formation of an external shell of Li₂CO₃ and Li₂SiO₃ (Chowdhury et al. 2013; Qi et al. 2013). The inversion temperature was identified at 515 °C, which is in agreement to the thermodynamic calculations reported by Chowdhury et al. (Chowdhury et al. 2013) for the same condition (500 °C and a CO₂ partial pressure of 0.05). Duan et al. (Duan et al. 2013; Duan et al. 2012) also reported the turnover temperature for pure Li₄SiO₄ in pre- and post-combustion conditions, as the temperature above at which lithium silicate cannot adsorb CO₂ and starts to release it according to the CO₂ partial pressure.

In addition, some differences in the TPC-TPDC profiles were observed between pure Li₄SiO₄ and lithium silicates derived from slags. A third peak of adsorption was identified after 500 °C, which can be associated to the CO₂ adsorption on the calcium phases present. Also, the CO₂ adsorption peaks for S1-Li₄SiO₄ appeared at lower temperatures, the maximum TPC peak is observed at 381 °C, while the desorption of CO₂ started after 588 °C. For S2-Li₄SiO₄, the inversion temperature increased slightly to 600 °C with the maximum TPC peak at 540 °C. The desorption peak, in both slags-derived lithium silicates, showed the maximum value at 770 °C.

Fig. 6

CO₂ adsorption–desorption cycle with respect to temperature (TPC-TPDC) of slag-derived lithium silicates and pure Li₄SiO₄ in a flow of 5 mol% CO₂

The effect of temperature and of different $P_{{{\text{CO}}_{2} }}$ (0.05, 0.10, 0.15 and 0.20) was evaluated by performing TG experiments in which CO₂ adsorption was measured over time. First, the CO₂ capture performance of lithium silicates using a $P_{{{\text{CO}}_{2} }}$ of 0.05 and temperatures from 580 to 700 °C (according with the dynamic TG results), is shown in Fig. 7. The maximum CO₂ capture, of both derived lithium silicates, was reached at very short times, such as 15 min, despite the low CO₂ concentration used in the tests. Figure 7a shows the S1-Li₄SiO₄ curves where, as expected, the amount of adsorbed CO₂ increased with the rise in temperature as follows: 84.5, 97.3 and 134.4 mgCO₂/g material at 580, 600 and 650 °C, respectively. The highest CO₂ capture value was observed at 650 °C. Afterwards, an increase in the temperature to 700 °C resulted in a decrease of the CO₂ capture by 40 percent (79.0 mgCO₂/g material), associated with the desorption process and the sintering of the material.

Fig. 7

Effect of temperature on the CO₂ uptake of a S1-Li₄SiO₄ and b S2-Li₄SiO₄ using $P_{{{\text{CO}}_{2} }}$ of 0.05 (the first 15 min are shown in the inset)

S2-Li₄SiO₄ (Fig. 7b) shows a similar behavior for the CO₂ uptake to that of S1-Li₄SiO₄. The adsorption capacities values of S2-Li₄SiO₄ were 106.5, 118.6, 103.0 and 97.1 mgCO₂/g material at 580, 600, 650 and 700 °C, respectively. According to these results, the best CO₂ capture was reached at 600 °C, which is a lower temperature than that of S1-Li₄SiO₄, nevertheless, the capture was higher for S1-Li₄SiO₄ than the one of S2-Li₄SiO₄. The adsorption of CO₂ on S2-Li₄SiO₄ decreased above 600 °C, that is, the adsorption-desorption equilibrium is different and the desorption process started at lower temperatures compared to those of S1-Li₄SiO₄. Considering as calculation basis the theoretical maximum CO₂ capture for the pure Li₄SiO₄ is 8.3 molCO₂/kg sorbent (367 mgCO₂/g sorbent) (Zhang et al. 2019), the slag derived-lithium silicates reached an efficiency of 36.57 and 32.37%, for S1-Li₄SiO₄ and S2-Li₄SiO₄, respectively, using a $P_{{{\text{CO}}_{2} }}$ = 0.05. Figure 8 shows the CO₂ uptake as function of time for the pure Li₄SiO₄, which is compared with that of derived silicates, under the same experimental conditions. With a $P_{{{\text{CO}}_{2} }}$ = 0.20 (Fig. 8a) the CO₂ uptake was very slow and low, where the equilibrium was not reached after 3 h (between 500 and 600 °C). The maximum CO₂ adsorption was 60 mgCO₂/g sorbent (550 °C), similar to data previously reported (Hu et al. 2019; Monica et al. 2013). In addition, when the CO₂ concentration decreased from 0.2 to 0.05, the CO₂ uptake became slower and decreased to only 5 mgCO₂/g sorbent in $P_{{{\text{CO}}_{2} }}$ = 0.05.

Fig. 8

CO₂ uptake by Li₄SiO₄ over the time a effect of temperature and b effect of different $P_{{{\text{CO}}_{2} }}$

Fig. 9

CO₂ uptake profiles over time of a S1-Li₄SiO₄ and b S2-Li₄SiO₄ with different CO₂ partial pressures, at 650 and 600 °C, respectively

Figure 9 shows the CO₂ capture as a function of time on the slag-derived lithium silicates, at the best adsorption temperature (650 °C for S1-Li₄SiO₄ and 600 °C for S2-Li₄SiO₄), using $P_{{{\text{CO}}_{2} }}$ of 0.05, 0.10, 015 and 0.20. In both materials, the CO₂ capture increased as the CO₂ concentration augmented from 0.1 to 0.2, but this behavior was not observed for $P_{{{\text{CO}}_{2} }}$ = 0.05. In fact, the highest adsorption capacity was obtained for $P_{{{\text{CO}}_{2} }}$ = 0.05 on S1-Li₄SiO₄. Figure 9a shows the CO₂ adsorption for S1-Li₄SiO₄, and it is evident that when the $P_{{{\text{CO}}_{2} }}$ increased the process became faster during the first 12 min. Thus, for $P_{{{\text{CO}}_{2} }}$ values of 0.10, 0.15 and 0.2, the CO₂ capture increased from 78.6 to 94.7 and 114.2 mgCO₂/g sorbent, respectively. To verify the reproducibility of these results, a second batch of S1-Li₄SiO₄ was prepared and the CO₂ uptake capacities were 79.1, 95.35 and 111.24 mgCO₂/g for $P_{{{\text{CO}}_{2} }}$ = 0.10, 0.15 and 0.20, respectively. A difference lower than 3% was estimated between both batches. A similar behavior was observed for S2-Li₄SiO₄ (Fig. 9b), since the CO₂ sorption process became faster in the first 10 min and the CO₂ capture capacity increased with an increment in the $P_{{{\text{CO}}_{2} }}$ (88.3, 98.6 and 118 mgCO₂/g material for $P_{{{\text{CO}}_{2} }}$ = 0.10, 0.15 and 0.20, respectively). S2-Li₄SiO₄ showed the same CO₂ sorption capacities in the upper and lower partial pressures values (0.05 and 0.20). It is important to highlight that the reaction rates observed in both materials for the different $P_{{{\text{CO}}_{2} }}$ used in this work were faster than that reported for pure Li₄SiO₄ with low CO₂ concentration (Seggiani et al. 2013; Seggiani et al. 2011; Zhang et al. 2014).

3.4 Kinetic analysis

In order to investigate the effect of the CO₂ concentration on the kinetic behavior of the prepared materials, the experimental data of S1-Li₄SiO₄ and S2-Li₄SiO₄ showed in Fig. 9 were analyzed according to the Avrami–Erofeev model (Qi et al. 2013; Zhang et al. 2014). The Avrami–Erofeev model is associated with the reaction mechanism of the formation and growth of reaction product crystals and is based on the typical model for gas-solid reactions:

$$d\alpha /dt =KF\left(\alpha \right)$$

(1)

where

$$F\left(\alpha \right)=n\left(1-\alpha \right){\left[-\text{ln}\left(1-\alpha \right)\right]}^{\left(n-1\right)/n}$$

(2)

where α is the degree of conversion (refers to the conversion of sorbent material towards carbonation products, that is, the ratio between the CO₂ adsorption capacity at given time t and the maximum theoretical CO₂ adsorption), t is the time, K is the kinetic constant and n is the kinetic parameter.

thus

$$\alpha =1-exp\left(-k{t}^{n}\right)$$

(3)

where $k={K}^{n}$

$$\text{ln}\left(-\text{ln}\,\left(1-\alpha \right)\right)=\text{ln}\,k+n\text{ln}\,t$$

(4)

The double logarithmic form of Eq. 3 (Eq. 4) can be successfully applied to the experimental results of the CO₂ adsorption by linear plot of $\text{ln}\left(-\text{ln}\left(1-\alpha \right)\right)$ vs $\text{ln}\,t$ with slope n. This equation was used to estimate the specific kinetic parameters, K and n, where n is a fractional number that accounts for possible changes of the adsorption mechanism during the adsorption process. When n > 1, the carbonation reaction is controlled by the formation rate and growth of the product layer, and, with n < 1, the reaction proceeds under diffusion control (Qi et al. 2013; Zhang et al. 2014; Zhao et al. 2018). Plots of $\text{ln}\left[-\text{ln}\left(1-\alpha \right)\right]$ vs $\text{ln}\,t$ for both slag- derived lithium silicates at different CO₂ partial pressures are shown in Fig. 10. The kinetic parameters n and K were determined using the slope and intercept and are summarized in Table 2.

Fig. 10

Fit of the CO₂ kinetic experimental data with the Avrami–Erofeev model, S1-Li₄SiO₄ and S2-Li₄SiO₄

Table 2 Parameters of Avrami–Erofeev model for the CO₂ sorption process on the slag derived lithium silicates under different CO₂ partial pressures

For both slag-derived lithium silicates in the different $P_{{{\text{CO}}_{2} }}$ studied, n values of rapid reaction stage are > 1, and < 1 for the diffusion-control stage. This means that the formation rate and growth of the carbonate external shell controls the rapid reaction stage, the first step of the whole CO₂ capture process, while the diffusion processes control the second stage. The reaction rate K values of the rapid reaction stage are in general higher than those of K values in the diffusion-control stage. This is in agreement with previous studies and means that the limiting step of the total process is the CO₂ sorption controlled by diffusion processes (Qi et al. 2013; Zhang et al. 2014). In S1-Li₄SiO₄, K values of rapid reaction stage are up to eight orders of magnitude greater than the diffusion-control stage with $P_{{{\text{CO}}_{2} }}$ of 0.05 and 0.10. When the CO₂ partial pressure increased to 0.15 and 0.20 the K values of the diffusion-control stage also increased. Whilst for S2-Li₄SiO₄, the K values of rapid reaction stage are also greater than K values for the diffusion-control stage, but only in three or six orders of magnitude. In addition, it is noticeable that the K values of the rapid reaction and diffusion-control stages, obtained for both slag-derived lithium silicates, are at least one order of magnitude higher than those reported for Li₄SiO₄, in 10 vol% of CO₂ (Zhang et al. 2014) and similar to those obtained in pure CO₂ flow (Qi et al. 2013).

3.5 Cyclic performance

Finally, the slag-derived lithium silicates were tested in twenty cycles of CO₂ sorption-desorption in order to evaluate their regeneration properties and thermal stability. The cyclic performance in $P_{{{\text{CO}}_{2} }}$ of 0.2 and 0.05 was analyzed at 600 and 650 °C for S2-Li₄SiO₄ and S1-Li₄SiO₄, respectively, temperatures at which the highest CO₂ capture capacities were obtained. For the desorption step, the flow was switched to N₂ and the temperature was raised to 750 °C and maintained for 20 min. Figure 11a shows the results obtained for S1-Li₄SiO₄, in $P_{{{\text{CO}}_{2} }}$ of 0.2 an uptake of 110.1 mgCO₂/g sorbent was reached for the first cycle, and after 20 cycles, the CO₂ capture decreased to 77.32 mgCO₂/g sorbent (1.76 mmolCO₂/g sorbent). When the $P_{{{\text{CO}}_{2} }}$ diminished to 0.05, the CO₂ adsorption was 121.5 mgCO₂/g sorbent in cycle number one and, in cycle fourteen, it decreased by 20% stabilizing in 96 mgCO₂/g sorbent (2.18 mmolCO₂/g sorbent). S2-Li₄SiO₄ showed less thermal stability and after 20 cycles the CO₂ capture diminished by 32% in $P_{{{\text{CO}}_{2} }}$ = 0.2 and almost 40% in $P_{{{\text{CO}}_{2} }}$ = 0.05.

Fig. 11

Cyclic performance of a S1-Li₄SiO₄ and b S2-Li₄SiO₄, during 20 cycles of sorption-desorption with $P_{{{\text{CO}}_{2} }}$ of 0.2 and 0.05

The morphological changes in both materials after multicycle analysis were analyzed using backscattering electron micrographs, presented in Fig. 12. Before CO₂ capture, the slag derived lithium silicates consisted of dense polyhedral particles with a compact and non-porous surface, with small sizes ranging from 5 to 18 µm. Some agglomerates could be observed with a non-uniform size distribution, of 35–70 µm in S1-Li₄SiO₄ and of around 40 µm in S2-Li₄SiO₄ (Fig. 12a, b). However, during 20 cycles of CO₂ sorption-desorption (sorption in $P_{{{\text{CO}}_{2} }}$ = 0.05) the sintering process occurred in both materials due to the high value of the regeneration temperature (750 °C). Thus, the agglomerates of S1-Li₄SiO₄ and S2-Li₄SiO₄ particles became larger after 20 cycles as it was previously reported for other lithium-based sorbents (Fig. 12c, d). The increase of agglomerates size was greater in S2-Li₄SiO₄ than in S1-Li₄SiO₄ particles, from 40 to 214 µm (5 times their size) and from 71 to 160 µm, respectively. This increase makes more difficult the CO₂ diffusion in the sorption step and also in the desorption process, which has a negative effect in the regeneration of the materials during cyclic tests, similar to that observed in other lithium-based ceramics (Chen et al. 2016; Rodríguez-Mosqueda and Pfeiffer 2010; Wang et al. 2016; Xiang et al. 2015).

Fig. 12

Backscattered electrons micrographs a, b before and c, d after 20 cycles of CO₂ sorption–desorption for S1-Li₄SiO₄ and S2-Li₄SiO₄ samples

4 Conclusions

In this work, lithium orthosilicate adsorbents were successfully prepared using two different iron and steel slags as silica sources. In addition, the prepared materials were characterized and tested as high temperature CO₂ adsorbents in CO₂ partial pressures between 0.20 and 0.05. Both slag-derived lithium silicates presented better CO₂ capture capacities, at least thrice higher than the one of pure Li₄SiO₄. The most promising slag-derived sorbent prepared in this work was S1-Li₄SiO₄, which had the highest CO₂ capture, 134 mgCO₂/g sorbent at 650 °C with a $P_{{{\text{CO}}_{2} }}$ of 0.05, a CO₂ uptake higher than that of pure Li₄SiO₄ and other lithium based materials prepared with fly ash and tested with low CO₂ concentrations ($P_{{{\text{CO}}_{2} }}$ = 0.05–0.20). In both slag-derived silicates, the increase in the CO₂ partial pressure from 0.1 to 0.2 enhanced the CO₂ uptake and the reaction rate of the CO₂ adsorption process. The presence of calcium phases and small amounts of Mg, Fe and Al in the slag-derived lithium silicates improved the CO₂ uptake as well as the kinetic behavior. The kinetic parameters, calculated according to the Avrami–Erofeev model, showed that the formation and growing of the carbonate external shell controls the rapid reaction stage in both materials. The reaction rate K of rapid reaction stage was at least four orders of magnitude higher than the K values of the diffusion control stage, i.e., the diffusion control stage was the limiting step of the total CO₂ capture process. The cyclic tests indicated that the S1-Li₄SiO₄ sorbent had a good thermal stability and high CO₂ capture capacity after 20 cycles (2.18 mmol CO₂ per g of material), despite the low CO₂ concentration. All these results showed that the slag derived silicates are promising materials to be used in CO₂ capture processes operated at high temperatures (T ≥ 600 °C) with low CO₂ partial pressures (0.05 ≤  $P_{{{\text{CO}}_{2} }}$ ≤ 0.20), such as sorption enhanced reforming and in the looping of a post-combustion CO₂ capture into natural gas combined cycle (NGCC) plants. Therefore, future research will be focused on the structural and textural modifications of the slags derived silicates, using other gases in the mixture (such as NO_x, SO_x, CO, O₂ and water vapor), in order to guarantee a good cyclic stability.