Operating limits and features of direct air capture on K2CO3/ZrO2 composite sorbent

Vladimir S.Derevschikov*,Janna V.Veselovskaya ,Anton S.Shalygin ,Dmitry A.Yatsenko ,Andrey Z.Sheshkovas2 ,Oleg N.Martyanov

1 Boreskov Institute of Catalysis SB RAS,Akademika Lavrentieva Av.5,630090 Novosibirsk,Russia

2 Novosibirsk State University,Pirogova Str.1,630090 Novosibirsk,Russia

Keywords: Zirconia aerogel Potassium carbonate Carbon dioxide Direct air capture Fourier-transform infrared spectroscopic imaging

ABSTRACT Potassium carbonate-based sorbents are prospective materials for direct air capture(DAC).In the present study,we examined and revealed the influence of the temperature swing adsorption (TSA) cycle conditions on the CO2 sorption properties of a novel aerogel-based K2CO3/ZrO2 sorbent in a DAC process.It was shown that the humidity and temperature drastically affect the sorption dynamic and sorption capacity of the sorbent.When a temperature at the sorption stage was 29°C and a water vapor pressurein the feed air was 5.2 mbar(1 bar=105 Pa),the composite material demonstrated a stable CO2 sorption capacity of 3.4% (mass).An increase in sorption temperature leads to a continuous decrease in the CO2 absorption capacity reaching a value of 0.7% (mass) at T=80 °C.The material showed the retention of a stable CO2 sorption capacity for many cycles at each temperature in the range.Increasingin the inlet air from 5.2 to 6.8 mbar leads to instability of CO2 sorption capacity which decreases in the course of 3 consecutive TSA cycles from 1.7% to 0.8% (mass)at T=29°C.A further increase in air humidity only facilitates the deterioration of the CO2 sorption capacity of the material.A possible explanation for this phenomenon could be the filling of the porous system of the sorbent with solid reaction products and an aqueous solution of potassium salts,which leads to a significant slowdown in the CO2 diffusion in the composite sorbent grain.To investigate the regeneration step of the TSA cycle in situ,the macro ATRFTIR (attenuated total reflection Fourier-transform infrared) spectroscopic imaging was applied for the first time.It was shown that the migration of carbonate-containing species over the surface of sorbent occurs during the thermal regeneration stage of the TSA cycle.The movement of the active component in the porous matrix of the sorbent can affect the sorption characteristics of the composite material.The revealed features make it possible to formulate the requirements and limitations that need to be taken into account for the practical implementation of the DAC process using the K2CO3/ZrO2 composite sorbent.

1.Introduction

Carbon dioxide (CO2) is the major anthropogenic greenhouse gas contributing to global warming.Large emissions of CO2into the atmosphere are caused by the growing consumption of nonrenewable fossil fuels.Technological and natural disasters,which result in uncontrolled combustion of large number of organic substances,also cause atmospheric CO2concentration growth.To reduce the negative effect of these emissions on the environment,the development of CO2capture and storage technologies is required [1].

One of the options for reducing greenhouse gas emissions is to capture CO2from large stationary sources by solid sorbents [2–4].The captured CO2can be permanently stored in deep geological formations [5] or used in the production of fuels [6],chemicals[7],and construction materials [8].CO2capture directly from the atmosphere known as direct air capture (DAC) is currently under consideration as another option for stabilizing global CO2concentrations [9–12].The benefits of DAC as a carbon removal technology include the limited land and water footprint and the possibility of locating capture plants close to suitable storage or utilization sites,eliminating the need for long-distance CO2transport[13–15].On the other hand,DAC is a more challenging process compared to source point capture,because of the very low concentration of CO2in the air (～400 μl?L-1) and the necessity to operate in the presence of moisture excess at ambient temperature and pressure [16].

Regenerable solid sorbents are of interest for DAC due to the favorable kinetics and thermodynamics of the CO2separation from ambient air [17–19].Numerous studies on this topic suggest that potassium carbonate-based sorbents can be used for the effective capture of CO2at ambient temperatures based on the reversibility of the reaction

The solid–gas reaction is characterized by a significant increase in the molar volume of the solid,with the Pilling-Bedworth ratio for the reaction(1)being 1.62.Therefore,as the reaction proceeds,the expanding product layer of around each grain of K2CO3results in a loss of local porosity that reduces the effective diffusivity.The increased diffusional resistance restricts the transport of reactant gas throughout the pellet,thereby lowering the reaction rate.Hence,for the bulk K2CO3,the total rate of the reaction (1) is low.The problem can be solved through the use of composite sorbents containing a porous material that supports K2CO3in a dispersed state and remains stable under conditions of the process.

Several different types of porous supports for K2CO3have been proposed including magnesia[20],alumina[21],silica[22],titania[20],yttria [23],and activated carbon [24].CO2sorption–desorption properties of a K2CO3-based sorbent strongly depend on the chemistry and structure of the porous support[25].The disadvantage of many traditional oxide porous matrices is their chemical interaction with the active component (K2CO3).Prior studies on DAC using K2CO3supported on alumina showed that the regeneration of the material must be carried at 250–300°C to decompose KAlCO3(OH)2which is the byproduct of the active component interaction with the support [26].The phase of interaction between K2CO3and Y2O3was found to be stable up to 350 °C[23].An increase in the regeneration temperature allows maintaining a stable CO2absorption capacity of a composite material but simultaneously leads to higher energy costs for the cyclic adsorption–desorption process [19].

Among other supports,zirconia appears to be a good choice for the potassium carbonate-based CO2sorbents.ZrO2exhibits sintering resistance and no reaction with CO2or steam in a broad range of temperatures [27].Leeet al.[27] reported that the K2CO3/ZrO2sorbent demonstrated high and stable values of CO2sorption capacity in the consecutive temperature-swing adsorption (TSA)cycles with the regeneration temperature of 200°C simulating conditions of CO2capture from flue gases.It is worth noting that X-ray diffraction(XRD)analysis showed no evidence of chemical interaction between K2CO3and ZrO2in the material before and after CO2adsorption–desorption experiments [27,28].

In our previous work [29],the composite sorbent obtained by K2CO3deposition on ZrO2aerogel demonstrated a high dynamic CO2sorption capacity in the TSA cycles simulating the DAC process.The K2CO3/ZrO2sorbent showed excellent CO2absorption and regeneration properties and outperformed the composite sorbent K2CO3/γ-Al2O3that was tested under the same conditions [30].Taking into account that K2CO3/ZrO2demonstrates stable CO2sorption capacity values and needs a relatively low temperature for regeneration (200 °C),it should be considered for application in DAC technology [29].

Although the sorbent based on zirconium dioxide was shown to be promising for DAC,it is important to investigate how the variations of operating conditions will affect the sorption properties and durability of the sorbent and find the limitations of the material.Carrying out laboratory experiments in this direction will allow answering the question of what is the optimal configuration of the process and the reactor should be to increase the efficiency of the DAC [31,32].

In this work,we investigated the effect of the adsorption temperature and water vapor pressure in the air on the CO2capture capacity of the sorbent.An attempt was made to assess the advantages and limitations of K2CO3/ZrO2composite materials that are based on a zirconia aerogel since one material is unlikely to be fitting in all potential DAC scenarios.

2.Experimental

2.1.Preparation of the sorbent

ZrO2support was prepared using an epoxide-assisted sol-gel method.During the first step,propylene oxide (PrOX) was vigorously added to the water-ethanol solution of ZrOCl2?8H2O.A magnetic stirrer was used for mixing the reaction mixture to reach uniform gelation.The molar ratio of Zr:PrOX was 1:2.5,while the molar ratio of H2O:Zr was 1:5.Normally,several days were required for the gel aging.To remove the unreacted PrOX and condensate,i.e.,alcohol and ester from the gel formed,it was washed with ethanol for 5 days.Further,the gel was dried in supercritical ethanol using an autoclave (produced by “Autoclave Engineers”,USA).Supercritical drying was carried at a constant pressure of 10.13 MPa while the autoclave was linearly heated to 300 °C.The ZrO2aerogel was grounded and sieved.

Potassium carbonate-based composite sorbent was prepared by impregnating zirconium oxide (ZrO2) with a 3 mol?L-1aqueous solution of K2CO3.Subsequent drying was carried out at 100 °C in the air for 12 h.After that,the sorbent was heated to 200 °C in an argon atmosphere.A detailed description of the synthesis procedure is provided in [29].According to the elemental analysis performed using an ARL Perform’X X-ray fluorescence spectrometer with an X-ray tube with Rh anode (Thermo Scientific,Switzerland),the potassium loading in the composite material is 13.1% (mass),which is equivalent to the K2CO3mass content of 23% .

2.2.Characterization of the sorbent

The phase composition of the composite sorbent after CO2sorption from the air was determined from powder XRD patterns which were collected using the X’TRA diffractometer (Thermo Scientific,Switzerland) with a vertical goniometer of θ/θ geometry and Bragg-Brentano focusing.The generator current was 40 mA,while the voltage was 35 kV.The XRD patterns were collected in a 2θ range from 10° to 70° with 0.05° step.The structural changes of the sorbent upon CO2sorption from the air were studied by XRD analysisin situusing the X-ray diffractometer Bruker D8 Advance(Germany) with a vertical goniometer of θ/θ geometry and Bragg-Brentano focusing.The device is equipped with a Lynx-Eye linear semiconductor energy-dispersive detector.The radiation source for both diffractometers is an X-ray tube with a copper anode.The average radiation wavelength is Cu Kα=0.154184 nm (Cu Kα1=0.15406 nm,Cu Kα2=0.154439 nm).The generator current is 40 mA,while the voltage is 40 kV.XRD patterns were collected in a 2θ range from 23° to 42°,with a 0.05° step size and 4 s accumulation time.X-ray diffractionin situexperiments were carried out in the Anton Paar XRK 900 high-temperature camera(Austria).The samples of the composite sorbent,after the absorption of CO2from humid air,were heated from room temperature to 150°C in a flow of helium (0.025 L?min-1).

Phase analysis was performed using the PDF4+database,the ICSD database was used to obtain structural data.The Rietveld refinement performed with TOPAS v4.2 software was used to quantify the phase composition and determine the crystallite size.For the best fit of the profile with the experiment,the unit cell parameters,sample displacement,atomic coordinates,and thermal parameters were optimized as fitting parameters without physicochemical justification.

Attenuated total reflectance Fourier transform infrared (ATRFTIR) spectra for the samples of composite sorbents were registered by the FTIR spectrometer Vertex 70 V (Bruker,Germany)equipped with the mercury cadmium telluride detector using the diamond ATR accessory (Golden Gate,Specac,UK).The spectra were recorded with a resolution of 2 cm-1.The number of scans was 128.The regeneration of the sample was studiedin situby FTIR spectroscopy using a diamond ATR accessory (Golden Gate Top-Plate,Specac,UK).

The ATR-FTIR spectroscopic imaging experiments were done using a Bruker Vertex 70v spectrometer(Germany)equipped with an imaging macrochamber (IMAC,Bruker Optics) and a 64×64 focal plane array (FPA)detector.The Imaging Golden GateTMaccessory (Specac,UK) with a high-temperature diamond ATR crystal(Golden Gate High-Temperature Top-Plate,Specac,UK)specifically designed for imaging applications was used and placed within an IMAC macrochamber.The heating rate between the spectra registration temperature was 8–10 °C?min-1while recording the spectrum took 3–4 min.ATR-FTIR spectra were recorded in the range from 3900 to 900 cm-1with a spectral resolution of 8 cm-1.The size of the FPA array detector was 0.6 mm×0.55 mm with a lateral spatial resolution of 8–9 μm.The number of co-added scans accumulated for each spectrum was 64 depending on the time scale of the experiment.The data were collected using OPUS 7.5 (Bruker Optics,Germany).The images were created from registered spectra by allocating a color to each pixel depending on the integral absorbance related to the particular spectral band and then plotting its distribution for all pixels to create a 2D map.

For the study,the sample powder was ground in an agate mortar to a very “fine”powder.The experiment was prepared as follows.The sorbent powder was poured into the thinnest possible layer on the ATR surface of the crystal(diamond),and the sapphire anvil was pressed against the crystal surface.Then the crystal was sequentially heated from 100 to 150 °C.The time of heating and spectra recording for the sample was 10 min.

The porous structure of the material before and after sorption experiments was assessed from nitrogen adsorption–desorption isotherms measured at -196 °C on a Nova 1200e Surface Area and Pore Size Analyzer(Quantachrome,USA).Before the measurement,the samples were degassed at 50 °C under vacuum for at least 2 h.Pore size distribution for micro-and mesopores was calculated by a Barrett–Joyner–Halenda (BJH) method.

2.3.DAC cyclic experiments

The DAC process was studiedin temperature TSA mode using an experimental set-up schematically presented in Fig.1.Heating control of a fixed-bed plug-flow reactor with an inner diameter of 1 cm was carried out automatically by a proportional-integralderivative (PID) controller Termodat-13K2 (Control systems Ltd.,Russia).K-type thermocouple located inside the outer electrical heater was used to control the temperature of the reactor.Another K-type thermocouple was used to measurethe temperature inside the sorbent layer throughout the experiment.The weight of the sorbent sample placed in the reactor was 1 g.

During the CO2sorption step of the cycle,the composite sorbent was saturated as indoor air with a CO2concentration of 380–420 μl?L-1was pumped through the reactor with a flow rate of 2.3 L?min-1using a gas pump (YW’Fluid,China).The partial pressureof water vaporin the inlet air was controlled using a saturator filled with distilled water which was thermostated using a cryothermostat LOIP FT-311-25(LOIP,Russia).Therange during the sorption step was from 5.2 to 19.4 mbar (1 bar=105Pa).The temperature of the sorbent during this step was varied from 29 to 80 °C.Thermal regeneration of the composite sorbent was carried out by heating the sample to 200°C in argon flow.The inlet flow rate of 50 ml?min-1was controlled by a gas flow regulator RRG-12-36 with an accuracy of ±0.5 ml?min-1.The duration of the CO2sorption and desorption steps was 6 and 1 h,correspondingly.After the CO2desorption step,the reactor was cooled down to the temperature set for the sorption step.

CO2concentration in the air was measured at the inlet and outlet of the reactor in the range from 0 to 1000 μl?L-1by an NDIR CO2sensor (Dynament,UK).Another NDIR CO2sensor (Dynament,UK)was used for measuring outlet CO2volume concentrations in the range from 0 to 40% during the CO2desorption step with an accuracy of ±0.2% .The relative humidity and temperature of the inlet air was measured throughout the experiment by a hygrometer IVA-6B (Microfor,Russia).

2.4.H2O vapor sorption isotherms measurement

Equilibrium H2O uptakes depending on the partial water vapor pressurewere measured for the composite sorbent and unmodified ZrO2atT=30 °C.A dry material was placed inside a U-shaped quartz-tube reactor,which was immersed in a water thermostat.The procedure for settingwas similar to the DAC experiment described above.Argon with the 400 ml?min-1flow rate was bubbled through the thermostated water-filled saturator and then was directed to the reactor containing the tested material.The relative humidity (RH) and temperature of the gas at the inlet and outlet of the reactor were measured throughout the experiment by the hygrometer IVA-6B with an RH measurement accuracy of ±1% .The amount of H2O sorbed by the material was determined by weighting the sample before and after H2O sorption at givenusing an electronic balance with an accuracy of ±0.001 g.

3.Results and Discussion

3.1.Influence of air humidity and temperature on the CO2 absorption capacity of the composite material

We investigated the effect of the sorption temperature on the CO2sorption properties of the composite material in the temperature range from 29 to 80°C.Theduring the sorption step was 5.2 mbar in the inlet air flow to the reactor.Fig.2 shows several CO2breakthrough curves measured at different temperatures.One can see that initially,the material demonstrates fast CO2sorption at all the temperatures.The key difference between the curves is observed after the first 5 min of sorption.The outlet CO2concentration reaches the level of the inlet CO2concentration afterca.22 and 33 min at 80 and 43 °C,correspondingly.The sorption curve measured at 29°C differs in shape from those that were measured at the higher temperatures.It shows a plateau as CO2outlet concentration reaches the level ofca.330 μl?L-1and the outlet CO2concentration reaches the level of the inlet CO2concentration only after more than 2.5 h of sorption.Usually,such behavior indicates that a material sorbs CO2as a result of 2 consequent or parallel processes which have substantially different CO2absorption rate.The slower process occurring at 29 °C is responsible for additional contribution to the CO2absorption capacity of the material.

Fig.1.The experimental setup for the cyclic DAC experiments.

Fig.2.CO2 breakthrough curves measured during the sorption step at different temperatures.

Fig.3 shows the experimental temporal dependencies of the outlet CO2concentration measured during the thermal regeneration step with a final regeneration temperature of 200°C following CO2sorption at different temperatures.The lower is the temperature of sorption the higher is maximal CO2concentration measured during the following CO2desorption step.The outlet CO2concentrations measured during the desorption step were used for calculation of the dynamic CO2absorption capacitya(% (mass)) using the following equation:

wherec(t) is the outlet CO2volume concentration (%),tis the corresponding time (min),mis mass of the composite sorbent (g),U0is the inlet flow rate (0.05 L?min-1),Mis the molar mass of CO2(44 g?mol-1),andVmis the molar volume of the ideal gas (24.4 L?mol-1for the standard conditions:T=25 °C andP=1 bar).

Fig.3.Outlet volume concentrations of CO2 measured during the desorption step of the TSA cycle for the composite sorbents saturated with CO2 at different temperatures.

The dependence of the CO2absorption capacity of the composite material (a) on the sorption temperature is presented in Fig.4.One can see thatagradually lowers with the temperature increase from 3.4% (mass)at 29°C to 0.7% (mass)at 80°C.For K2CO3-based composite materials,the decrease in CO2absorption capacity with increasing temperature is expected in thermodynamically controlled sorption processes [33].It is noteworthy that CO2absorption capacity drops by more than 2 times upon temperature increase from 29 to 43 °C,while a subsequent increase in temperature to 80 °C has a less profound effect ona.The CO2sorption–desorption experiment for each sorption temperature was repeated 4 times,and the material showed good stability of CO2absorption capacity regardless of the sorption temperature.

Fig.4.Dependence of CO2 absorption capacity of the composite material(a)on the adsorber temperature during the CO2 sorption step(partial water pressure for all the experiments is 5.2 mbar).The bars in the figure represent the standard error of the measurement.

Fig.5.The CO2 absorption capacity of the composite material (a) in consecutive TSA cycles using the air of different humidity at the CO2 sorption step.The bars in the figure represent the standard error of the measurement.

The influence of atmospheric air humidity at the sorption step on the CO2absorption capacity of the composite sorbent was investigated (Fig.5).Throughout all the experiments,the temperature inside the reactor was maintained at 29 °C.The best results were obtained using air withof 5.2 mbar.The material demonstrated a CO2absorption capacity of 3.4% (mass) which remained stable in 3 consecutive cycles with a regeneration temperature of 200 °C.It should be noted,that in our previous work[29],we have tested a similar zirconia-based material using the air with the sameand performing a sorption step at room temperature.The results of the cyclic performance test showed a stable CO2absorption–desorption capacity in 14 consecutive cycles proving that K2CO3/ZrO2sorbents are indeed stable when the air of relatively low humidity is used.Whenin the air was 6.8 mbar,the CO2absorption capacity of the material decreased rapidly from 1.7% to 0.8% (mass).The material demonstrated similar behavior when the air withof 8.8 mbar was used.For the experiment with=19.4 mbar,the CO2absorption capacity of the composite was stable but very low (0.4–0.6% (mass)).

Fig.6.H2O adsorption isotherms at T=30 °C for the composite material and unmodified zirconia.The bars in the figure represent the standard error of the measurement.

Fig.6 shows isotherms of H2O sorption were measured for the K2CO3/ZrO2and ZrO2materials at 30°C.For the zirconia,the water uptake changes from 1.4% to 2.1% (mass) with increasing vapor pressure from 5.2 to 19.4 mbar.For the K2CO3/ZrO2composite sorbent water uptakes are much higher than for the unmodified zirconia in all therange studied due to chemical and physical interaction between the supported species and H2O.Water by the salt was assessed by subtracting from the total water uptake a portion adsorbed by the support taking into account its mass content in the material.It was used for calculating H2O:K2CO3molar ratios for each point of the isotherm.Per this assessment,1 mol of K2CO3absorbs 1.6 mol of H2O at=5.2 mbar and 1.8 mol of H2O at=6.8 mbar.The stoichiometric ratio for the massive salt is 1.5 mol of H2O per 1 mol of K2CO3.

The values we observe are higher than 1.5 and not steady upon the water pressure increase from 5.2 to 6.8 mbar indicating that a part of the active component formed species demonstrating a bivariant type of equilibrium with water vapor.Such behavior is typical for inorganic salts dispersed inside mesoporous oxide matrices[34].There is a sharp increase in the water uptake upon the water vapor increase from 6.8 to 8.8 mbar.The calculated H2O:K2CO3molar ratio is 3.4 at=8.8 mbar.A likely explanation for this effect would be adeliquescence of dispersed K2CO3in water forming a saturated solution that fills the pores of the sorbent.Though it should be noted that for massive salt this process takes place at higher,which is around 18 mbar atT=30 °C according to the empirical model proposed by Greenspan [35].However,for composite sorbents,the equilibrium H2O pressure corresponding to the salt dissolution often shifts,which is explained by size effects arising from the salt dispersion in a porous medium [34].This phenomenon which is also known in Geophysics aspore-size controlled solubility(PCS) occurs as a result of the surface tension associated with crystals growing in confined pores [36].This surface tension gives rise to excess pressure within the crystal,analogous to the capillary pressure at a liquid–vapor interface within a pore.The relationship between pore size and solubility has been demonstrated experimentallyin situby measuring solute concentrations using nuclear magnetic resonance method.In that study,the solubility of hydrated Na2CO3in 10 nm pores was found to be more than twice that of the bulk solubility,showing that PCS is a feasible mechanism for achieving high supersaturations in porous media [37].

Upon the further increase in the H2O vapor pressure from 8.8 to 19.4 mbar,the water uptake continues to grow,but less sharply,which can be explained by the further dissolution of larger salt crystals and the absorption of water by a saturated salt solution.The H2O:K2CO3molar ratio is 4.4 at=19.4 mbar.The experimental estimation shows that during water sorption at19.4 mbar,the mass of the sorbent increases by more than 13% (mass) (Fig.6).Thus,taking into account that the sorbents pore volume is 0.11 cm3?g-1(Table 1),we assume that the sorbent porous system after the sorption step is filled with the salt solution.The pore-filling increases diffusional resistance,which restricts the transport of reactant gas inside the sorbent,thereby lowering the reaction rate and decreasing sorbent absorption capacity at high values of humidity (Fig.5).

3.2.Porous structure and textural properties of the materials

The porous structure of the sorbent samples was examined using nitrogen adsorption porosimetry.Fig.7 displays the isotherms of nitrogen ad-/desorption at-196°C for the sorbent powders before CO2sorption,after sorption,and after regeneration.It should be noted the humidity of the inlet air to the reactor was 5.2 mbar at the sorption stage of the TSA cycle.The sorbent samples presented type IV a isotherms typical for mesoporous materials.The surface area (SBET) of the samples was determined using the Brunauer–Emmett–Teller(BET)method.The total pore volume was determined from the N2amount adsorbed atP/P0=0.99.The values ofSBETand pore volume are placed in Table 1.

N2ad-/desorption isotherms for samples of the sorbent demonstrate type H1 loop according to IUPAC classification,therefore cylindrical pores with the average size of 10–15 nm calculated by Barrett–Joyner–Halenda (BJH) method [38],are expected in these samples (Fig.8).For all the samples,the volume of the macropores with sizes from 50 to 300 nm is less than 1% of the total pore volume calculated by the BJH method.Analysis of N2adsorption isotherms by thet-plot method[39] showed that samples of the sorbent contain no micropores.

Nitrogen adsorption data show that a significant reduction in the pore space of the sorbent occurs during the sorption of CO2and water vapor.At the sorption stage,the surface area of the sorbent decrease,and the mesopore volume is almost halved(Table 1).The changing of the pore volume is due to the collecting of product with higher molar volume in the porous system of sorbent,and,probably,due to the filling of mesopores with the potassium salts brine,that is not completely removed from the sorbent during the preliminary vacuum thermal treatment at 50 °C for the nitrogen isotherm measurement.While regeneration,solid products decompose to the initial state of a fresh sorbent,and the porous structure of regenerated sorbent approaches the fresh one(Fig.8).It should be noted that according to the nitrogen adsorption data the pore volume and specific surface area of the regenerated sorbent are slightly larger than that of the fresh one.It could mean that some part of an active component releases from the pore space to the sorbent surface under the action of capillary forces while thermal regeneration proceeds.It is worth noting that the explanation including the migration of an active component requires a deeper experimental verification,although it is confirmed by the FTIR spectroscopic imaging data presented below.

Table 1 Parameters of porous structure for the sorbent samples in the as-prepared state,after sorption,and after regeneration

Table 2 The CSD sizes (in nm) for crystalline phases detected in the composite sorbent samples after CO2 sorption from the air of different humidity

Fig.7.Nitrogen ad-/desorption isotherms at -196 °C for the as-prepared,after sorption,and after regeneration sorbent samples.

Fig.8.BJH pore size distributions for the as-prepared,after sorption,and after regeneration sorbent samples.

3.3.Chemical and phase composition of the materials

Fig.9 shows XRD patterns obtained for samples of the composite sorbent after the CO2sorption from the air of different humidity.Full-profile modeling of powder diffraction patterns was carried out by the Rietveld method using the TOPAS program(Figs.S1–S4 in the Supplementary Material).The following structural data from the ICSD database cards were used for modeling:#89426 for monoclinic ZrO2,#92090 for tetragonal ZrO2,#43015 for KHCO3,and #280789 for K2CO3?1.5H2O.

Fig.9.Experimental XRD patterns for the samples of the composite sorbent after CO2 sorption from the air of different humidity:(1)=5.2 mbar,(2)=6.8 mbar,(3)=8.8 mbar,(4) =19.4 mbar.

Table 2 presents the coherently scattering domain sizes (CSD)calculated for each crystalline phase detected in the composite materials.Both zirconia phases are nanosized with CSD of 5–7 nm.All the samples show the reflection maxima characteristic for potassium bicarbonate.The value of the coherent scattering domain for KHCO3varies from 9 to 67 nm depending on the air humidity.The samples of the composite sorbent after CO2sorption from the air withof 6.8 and 8.8 mbar additionally contain the crystalline phase of potassium carbonate sesquihydrate K2CO3?1.5-H2O,which is likely to be in equilibrium with an aqueous solution distributed in zirconia pores.

The size of the CSD for this crystalline phase increases from 50 to 133 nm as a result of a change in the water vapor pressure at the sorption stage from 6.8 to 8.8 mbar.

XRDin situanalysis revealed that heating the composite samples after the sorption step to 150 °C leads to the formation of the K2CO3phase(Fig.10).However,reflections attributed to potassium carbonate have different widths dependingin the air at the sorption step.The increase in water vapor partial pressure at the sorption step from 5.2 and 8.8 mbar results in rising crystallinity of K2CO3obtained as a result of thermal regeneration of the material.

Fig.11 presents the ATR-FTIR spectra for the composite sorbent samples after the step of CO2sorption from the air of different humidity.For all the samples,the KHCO3characterized absorption peaks are observed at 1691,1647,1624,1397,1371,1007,981,832 and 702 cm-1[27,29].For the sample after CO2sorption at=8.8 mbar,the peaks observed at 1447,1061,881 and 846 cm-1indicate the formation of hydrated potassium carbonate K2CO3?1.5H2O[24,40].A broad band observed at 745 cm-1is due to the stretching vibration of Zr-O-Zr bond characteristic of monoclinic zirconia [41,42].Additional bands at 1567,1511,1338,1310,1257,1183,1085 and 1050 cm-1can be attributed to surface carbonate species [43,44].

Fig.10.Experimental XRD in situ patterns for the composite sorbent samples heated to 150 °C after CO2 sorption from the air of different humidity:(1)=5.2 mbar,(2) =8.8 mbar.

Fig.11.ATR-FTIR spectra for the composite sorbent samples after the step of CO2 sorption from the air of different humidity.

A joint analysis of the ATR-FTIR and XRD data shows that an increase inleads to an increase in the content of sesquihydrate K2CO3?1.5H2O in the sample.The crystalline size of K2CO3?1.5H2O exceeds the median pore size of the porous matrix of the sorbent,and it could mean that the K2CO3?1.5H2O formation proceeds either on the surface of the sorbent particles or in some larger pores,presented inside the sorbent porous system.Another effect that is observed is the increase in potassium hydro carbonate crystallite size with increasing humidity (Table 2).An explanation for this phenomenon can be found within the PCS mechanism.According to the PCS mechanism for systems with spatially heterogeneous pore-size distributions,fluid flowing through a region dominated by small pores will achieve a high degree of supersaturation;upon reaching a fracture or zone with large pores,the fluid will no longer be in equilibrium with the porous matrix and salt precipitation will occur.In systems with a range of pore sizes,the system is much more complex as,locally,precipitation may dominate in large pores,while dissolution may occur in smaller ones,or vice versa [36].Apparently,in this way,under conditions of high humidity and formation of the brine inside the pores,potassium salts can be removed from the mesopores and condense into large crystals in larger pores or on the sorbent surface.The formation of the solid products K2CO3?1.5H2O and KHCO3from the solid reagent K2CO3is accompanied by a significant increase in volume,which may result in an additional pore blockage and heightened resistance to CO2diffusion in the sorbent grain.Thus,a likely reason for the decrease in the absorption capacity of the sorbent is the filling of the pores with an aqueous solution and solid products.

3.4.Macro ATR-FTIR imaging of sorbent regeneration

To get a better insight regarding regeneration processes taking place during the thermal step of the TSA cycle on the sorbent surface,we performed the macro ATR-FTIR spectroscopy imagingin situexperiments.FTIR spectroscopic imaging is a nondestructive analytical technique with minimal sample preparation and chemically specific technique,most importantly,can provide both the time and spatially resolved information of specific chemical bonds,functional groups,and components present in a sample[45,46].The images of chemical species distribution can be generated from the array of spectra by plotting the 2D map of the absorbance distribution of a specific spectral band [47,48].

The macro ATR-FTIR spectroscopic imaging has been utilized to investigate the regeneration of the sorbent saturated with CO2and water.Fig.12 shows ATR-FTIR spectroscopic images of sorbent particle thermal regeneration registeredin situ.FTIR spectroscopic imaging reveals the spatial distribution of KHCO3and K2CO3.The 2D image was obtained using integrated absorbance of the spectral bands at 1740–1475,1025–934 cm-1,which correspond to the stretching and deformation vibration modes of the υas(C=O) and δ(C-O) bonds of,and 1470–1365 cm-1,which correspond to the stretching vibration modes of the υas(C-O) bonds ofgroups [44] respectively.The red color represents the highest absorbance,while the blue color– the lowest integral absorbance for the chosen bands.All the images capture the same region of the sorbent grain.

The first and the second rows depicting the integral absorbance of spectral bands at 1740–1475 cm-1and 1025–934 cm-1respectively indicate the spatial distribution of KHCO3with changing the temperature (regardless of its crystalline or amorphous form).There are no significant differences in KHCO3distribution on the sorbent surface before heating up to 135 °C,indicating the presence of the original KHCO3crystals in these areas.At higher temperatures,the intensity of absorption bands in the spectral ranges mentioned above significantly decreases in the appropriate site of the image due to the thermal decomposition of KHCO3.

Fig.12.ATR-FTIR spectroscopic images of sorbent particle thermal regeneration registered in situ.The field of view of each image covers an area of 600 μm×550 μm.The 2D image was obtained using integrated absorbance of the spectral bands at 1740–1475,1025–934 and 1470–1365 cm-1,which correspond to the υas(C=O) and δ(C-O)vibration modes of ,and υas(C-O) mode of group respectively.

The third row in Fig.12 that corresponding to integral absorption of bands in the ranges at 1470–1365 cm-1shows the evolution of K2CO3.The redistribution of K2CO3into areas where initially where was no KHCO3is observed,which may occur due to the capillary drop spreading of potassium salt brine on the ZrO2surface.Thus,the macro ATR-FTIR images indicate that there is a great tendency for potassium-containing species to migrate on the surface of the ZrO2matrix.It could also mean that the active component migrates from the pores of the support to the surface during the heating of the sorbent.The data obtained indicate that composite sorbents based on potassium carbonate are not completely solid and the carbonate/bicarbonate system has spatiotemporal dynamics in the surface of the granule.Additional clarifying studies are needed to study the dynamics of the distribution of potassium carbonate within the sorbent and to understand the ways of stabilizing the active component in the porous matrix of the support.

3.5.Operating limits and features of direct air capture on K2CO3/ZrO2

The revealed features of the TSA cycle on K2CO3/ZrO2sorbent impose requirements on humidity control during the operation of DAC installations.Specialized drying of the incoming air may be required when using these materials in a humid climate.In northern regions,maintaining the specified humidity can be carried out by pre-cooling the air to the dew point.The use of these sorbents may be most economically feasible in countries with arid climates since the moisture content in the air of deserts is small and no effort is required to pre-dry the air;at the same time,the by-product is clean water,which is always in short supply in deserts [32].Thus,DAC installations based on potassium carbonate-containing sorbents can become a whole industry either for CO2separation from the atmosphere or for water harvesting in arid regions of the Earth [19,49].

The advantage of using the K2CO3/ZrO2composite material is an opportunity to obtain unique operational properties such as stable sorption capacity due to the absence of chemical interaction between the active component (potassium carbonate) and porous support(zirconia aerogel).The disadvantage of the zirconia aerogel as the support is its relatively high cost due to the price of the zirconia precursor and special synthesis conditions.It should be noted that in general the price of zirconia is 3–4 times higher than that of alumina.Apparently,the expediency of K2CO3/ZrO2practical application will be justified if this material can operate for hundreds or thousands of sorption/regeneration cycles,and that’s why it is very important to define operating conditions needed for stable performance of the material.At the same time,it is quite possible to create cheaper K2CO3-based materials for direct air capture using porous supports,which surface is modified with zirconia(or,in general,with compounds that don’t interact chemically with the active component– potassium carbonate),thus making the material more stable and durable.In general,the global trend is a decrease in the cost of CO2capture,and we believe that the optimization of the materials for direct air capture and decreasing of its cost will happen in a natural way.

4.Conclusions

Potassium carbonate-based sorbents are considered as prospective materials for direct CO2capture from ambient air.The deposition of potassium carbonate to the porous inert matrix allows obtaining the materials with high and stable sorption performance.Recent studies have highlighted the need to determine the behavior of these systems in a broad range of exploitation conditions.In the present study,we revealed the influence of the temperature swing adsorption DAC cycle conditions (namely,the relative humidity of air and temperature) on the CO2sorption properties of the K2CO3/ZrO2sorbent.

The results obtained show that the sorbent demonstrated a CO2stable absorption capacity of 3.4% (mass) withof 5.2 mbar in the feed air and the temperature of 29°C.The absorption capacity of the sorbent for CO2drops by almost 5 times with an increase in temperature from 29 to 80°C.The CO2sorption–desorption experiment for each sorption temperature was repeated 4 times,and the K2CO3/ZrO2material showed good stability of CO2absorption capacity regardless of the sorption temperature.

It was shown that the relative humidity drastically affects the dynamic absorption capacity of the composite material.The increasing ofin the inlet air from 5.2 to 6.8 mbar drives to decrease the CO2absorption capacity of the material to 0.8% (mass).A possible explanation for this phenomenon would be the filling of the sorbent porous system with the saturated solution of potassium salts due to the condensation of water vapor inside the sorbent and the collecting of solid products,preventing further CO2sorption.The obtained isotherm of water sorption and XRD data confirm this theory.

In this work,the macro ATR-FTIR spectroscopic imaging was used for the first time to studyin situthe spatio-temporal distribution of the active component on the surface of the sorbent.It was shown that the migration of carbonate-containing species over the surface of sorbent occurs during the thermal regeneration stage of the TSA cycle.The movement of the active component in the porous matrix of the sorbent can affect the sorption characteristics of the composite material,but at this point,it is difficult to quantitatively assess this phenomenon,and further improvement of this experimental method is required.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by Russian Science Foundation (19-73-00079).The authors also thank Leonova A.A.for performing N2adsorption measurements.

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.07.005.