Support Pore Structure and Composition Strongly Influence the Direct Air Capture of CO₂ on Supported Amines

Abstract

A variety of amine-impregnated porous solid sorbents for direct air capture (DAC) of CO₂ have been developed, yet the effect of amine-solid support interactions on the CO₂ adsorption behavior is still poorly understood. When tetraethylenepentamine (TEPA) is impregnated on two different supports, commercial γ-Al₂O₃ and MIL-101(Cr), they show different trends in CO₂ sorption when the temperature (−20 to 25 °C) and humidity (0–70% RH) of the simulated air stream are varied. In situ IR spectroscopy is used to probe the mechanism of CO₂ sorption on the two supported amine materials, with weak chemisorption (formation of carbamic acid) being the dominant pathway over MIL-101(Cr)-supported TEPA and strong chemisorption (formation of carbamate) occurring over γ-Al₂O₃-supported TEPA. Formation of both carbamic acid and carbamate species is enhanced over the supported TEPA materials under humid conditions, with the most significant enhancement observed at −20 °C. However, while equilibrium H₂O sorption is high at cold temperatures (e.g., −20 °C), the effect of humidity on a practical cyclic DAC process is expected to be minimal due to slow H₂O uptake kinetics. This work suggests that the CO₂ capture mechanisms of impregnated amines can be controlled by adjusting the degree of amine-solid support interaction and that H₂O adsorption behavior is strongly affected by the properties of the support materials. Thus, proper selection of solid support materials for amine impregnation will be important for achieving optimized DAC performance under varied deployment conditions, such as cold (e.g., −20 °C) or ambient temperature (e.g., 25 °C) operations.

Introduction

Fast decarbonization of the energy sector requires the deployment of avoided emission technologies like carbon capture and sequestration from point sources as well as the deployment of negative emission technologies (NETs) such as direct air capture (DAC) coupled with carbon storage.¹ DAC has several advantages over other NETs, such as bioenergy with carbon capture and storage,² due to its lower requirement for land and water utilization.³ However, DAC must overcome challenges associated with high capital costs and moderate energy costs.

DAC has been intensively studied in recent years, and a variety of adsorbent materials have been developed to directly capture CO₂ from the air. Among them, amine-functionalized porous materials are attractive adsorbents for DAC and are widely studied due to their promising performance under humid conditions, lower amine volatilization rates compared to amine solvents, and the strong affinity of amine groups for CO₂.⁴⁻⁶ However, due to the high CO₂ binding energy of amines, amine-based adsorbent materials require higher energy for CO₂ desorption compared to physisorbents such as zeolites and MOFs. While the amine-containing solid adsorbents are tolerant to moisture in the air, it is generally known that physisorbents are not effective adsorbents for CO₂ capture in the presence of moisture due to competitive adsorption between CO₂ and H₂O, which yields high sorbent regeneration costs. Thus, to utilize physisorbents for DAC, a two-stage process for the dehydration of air will likely be required.⁷

The behavior of solid-supported amines can be impacted by the structure and composition of the support materials because the porous structure of the support impacts amine dispersion and thus the access of CO₂ to the amine sites.⁵ Moreover, differences in surface chemistries in support materials can result in different binding motifs between the amine and the support, which also has the potential to influence the CO₂ capture performance. A variety of porous support materials have been used for amine impregnation, including silica,⁸⁻⁴⁰ γ-alumina,⁴¹⁻⁴³ MOFs,⁴⁴⁻⁴⁶ zeolite,⁴⁷ and mixed metal oxides.^48,49 Despite this, the effects of amine-solid support interactions on CO₂ adsorption behavior are still poorly understood. Furthermore, most prior DAC studies are limited to adsorption temperatures above indoor ambient room temperature (>20 °C). Only our previous study on amine-impregnated MIL-101(Cr) reported detailed DAC behavior under sub-ambient temperature conditions (from −20 to 25 °C), including both dry³⁵ and humid gases.⁴⁴ Since there will likely be no optimal sorbent covering all conditions, including day/night, summer/winter, and rainy/dry conditions, designing and operating a DAC process will be complex. A high-performing adsorbent will, therefore, be a sorbent that operates well under all conditions, rather than one that is optimized for a narrow set of conditions. To this end, investigating the DAC performance of sorbent materials under varied conditions is crucial for the practical deployment of DAC technologies.⁵⁰

As noted earlier, the adsorption mechanism (e.g., physisorption, the formation of carbamic acid, the formation of ammonium carbamate ion pairs, and the formation of ammonium bicarbonate) may vary depending on the degree of interaction between the impregnated amines and support. For example, specific molecular interactions between amines and MOFs like Mg₂(dobpdc) lead to unique CO₂ adsorption mechanisms⁵¹⁻⁵⁷ compared to more ill-defined systems based on MOFs and PEI.⁵⁸⁻⁶¹ We hypothesize that the CO₂ adsorption pathways for impregnated amines within different support materials and/or pore structures may differ depending on adsorption temperature and humidities. Thus, to design optimized adsorbents for DAC processes operating in an ever-changing environment (outdoor air),⁵⁰ an in-depth understanding of the roles of support materials, temperature, humidity on CO₂ adsorption kinetics, and thermodynamics is required.

Here, to systematically probe the effect of support materials on the DAC performance of impregnated amines [tetraethylenepentamine (TEPA)], amine-based solid sorbents were synthesized with two different porous solid supports, MIL-101(Cr) and γ-Al₂O₃. TEPA was used for amine impregnation because this chemical mainly consists of primary and secondary amines, which have a higher potential for CO₂ adsorption (e.g., heat of CO₂ adsorption = −80 to −110 kJ/mol_CO2) compared to tertiary amines (−20 kJ/mol_CO2),^62,63 as well as shorter amine chains that likely result in enhanced CO₂ kinetics relative to the branched poly(ethylenimine) typically employed, especially under cold temperature conditions. Commercial γ-Al₂O₃ was selected for comparison with MIL-101(Cr) supports since they are already well-investigated support materials for amine impregnation and currently offer a more accessible pathway for scale up and deployment than MOFs. The materials’ CO₂ adsorption behavior under both dry and humid (70% RH) 400 ppm CO₂ conditions over a range of adsorption temperatures (−20 to 25 °C) was probed by CO₂/H₂O temperature-programed desorption (TPD) experiments and in situ FT-IR spectroscopy. We demonstrate that the properties of the porous solid support materials, such as pore size, pore volume, surface area, and support composition, have a significant effect on the CO₂ adsorption pathways of the impregnated amines. The results presented here suggest that proper porous support materials should be considered for amine impregnation to operate an optimized DAC process, depending on the sub-ambient to ambient temperature conditions and humidities anticipated.

Experimental Section

Material Synthesis

Synthesis of MIL-101(Cr) powders was conducted based on the literature.^46,64 The detailed synthesis process is described in our previous work.⁴⁴ The synthesized MOF powders were dried under a high vacuum (about 10 mTorr) at 150 °C overnight for further analysis and amine impregnation. The MIL-101(Cr) and γ-Al₂O₃ powders were physically impregnated with 30 and 20 wt % TEPA loadings, respectively, to achieve a similar extent of pore filling (about 25%). The detailed procedure of amine impregnation is described in the Supporting Information.

Characterization

Nitrogen Physisorption

A surface area and porosity (SAP) system (autosorb iQ/Quantachrome) was used for N₂ physisorption experiments at 77 K. About 100 mg of MIL-101(Cr) and γ-Al₂O₃ were activated at 150 and 110 °C, respectively, under a vacuum for 15 h before the measurement. For the TEPA-impregnated MIL-101(Cr) and γ-Al₂O₃ powder sorbents, the activation was conducted at 60 °C under a vacuum for 3 h. The BET surface area was estimated using the N₂ physisorption data in the P/P₀ range of 0.05–0.2. Pore volumes of the materials were determined based on the N₂ physisorption at a partial pressure of 0.995 and normalized per gram of support (MIL-101(Cr) or γ-Al₂O₃) in the powder sorbents.

CO₂ Breakthrough Experiments under Dry and Humid Conditions with 400 ppm CO₂

A custom-built fixed bed reactor described in the literature⁴⁴ was used to perform dry and humid 400 ppm CO₂ breakthrough experiments with TEPA-impregnated MIL-101(Cr) and γ-Al₂O₃ under a wide range of temperatures (−20, −5, and 25 °C) and humidities (0–70% RH). About 50 mg (wet basis) of powder sorbents were loaded inside the 5 cm fixed bed (1/4″ stainless steel tube with 4 mm ID) using glass wool on both ends of the column. A more detailed procedure is described in Supporting Information. In this study, small amounts of powder sorbents (<50 mg) were tested inside a 1/4″ stainless steel fixed bed column under an extremely diluted CO₂ gas stream (400 ppm CO₂/N₂). The bed was always immersed in liquid (50% water/50% ethylene glycol) continuously circulating inside the bath of chiller (Julabo CD-600F) for efficient removal of the heat generated during the 400 ppm CO₂ breakthrough. To confirm an isothermal condition of the fixed bed system, the convective heat transfer coefficient, h, needed to keep the temperature of the powder sorbents less than 1 °C above the flowing simulated air was estimated. For example, the required h was estimated to be 2.36 × 10^–3 W/m²/K for 0.01 W of the heat generation rate during 400 ppm CO₂ adsorption with the TEPA-impregnated γ-Al₂O₃ in the fixed bed system (45 mg of the sorbent, 94 m²/g_sorbent, −ΔH = 100 kJ/mol_CO2, and 2.64 × 10^–4 mmol/g/s adsorption rate). The estimated value was much lower than the h of natural convection in the air (2.5–25 W/m²/K, produced by density differences of air),⁶⁵ indicating that convection cooling by flowing adsorption gas is enough to maintain an isothermal condition due to the slow heat generation rate of CO₂ adsorption under an extremely diluted CO₂ gas stream (400 ppm CO₂/N₂). Thus, it was assumed that the temperature change of the fixed bed column is negligible during the 400 ppm CO₂ adsorption.

Measurement of Energy for Dry/Wet CO₂ and H₂O Desorption

The interaction strength between CO₂ and the amine-impregnated MIL-101(Cr) and γ-Al₂O₃ powder sorbents are expected to be different under varied dry and humid adsorption temperature conditions. Thus, it is crucial to quantify the binding energy of dry and humid CO₂ (also H₂O) to estimate an energy requirement for sorbent regeneration. In this study, TPD experiments were performed with varied heating rates to directly evaluate the energy required for desorption of CO₂ (dry and humid) and H₂O from amine-impregnated MIL-101(Cr) and γ-Al₂O₃ powder sorbents.⁶⁶ Detailed procedures are described in Supporting Information, including a definition of the desorption energy used in this work.

In Situ Diffuse Reflectance Infrared Fourier Transform Spectroscopy

In situ FT-IR spectroscopy was performed on a Nicolet iS10 IR spectrometer with a low-temperature diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) cell (CHC-CHA-4, Harrick Scientific Products Inc.). For the dry and wet 400 ppm CO₂ adsorption studies, a sample holder inside the DRIFTS cell chamber was filled with about 50 mg of sorbent powders. The sample was then activated at 60 °C under 40 sccm N₂ flow for 3 h. After the sample temperature cooled to the adsorption temperature (−20 °C or 25 °C), the inlet gas flow was switched to dry or wet (70% RH) 400 ppm CO₂/N₂ (40 sccm). During the adsorption, spectra were collected with 64 scans at a resolution of 4 cm^–1 every 5 min. The humidity for the wet gas was precisely controlled by a wet gas generator (WETSYS/SETARAM).

Results and Discussion

Characterization of Amine (TEPA)-Impregnated MIL-101(Cr) and γ-Al₂O₃

To investigate the effect of support material structure and composition on the CO₂ capture behavior of impregnated amines under a similar extent of pore filling (about 25%), MIL-101(Cr) and γ-Al₂O₃ porous supports were physically impregnated with TEPA to 30 and 20 wt % loadings, respectively. Key physical properties, including pore volumes, pore sizes, and BET surface areas of the bare support and amine-loaded sorbent materials were determined from N₂ physisorption isotherms (Figure S1), and the results are summarized in Table 1. The particle size of two support materials is also included. The mean particle size of γ-Al₂O₃ was estimated based on SEM images (Figure S2). The properties of the MIL-101(Cr) materials are comparable to the prior literature.^44,46 MIL-101(Cr) has a smaller average pore size (2.4 nm) than γ-Al₂O₃ (16.1 nm), with a much higher pore volume and BET surface area, suggesting that impregnated amines may interact with MIL-101(Cr) more significantly than γ-Al₂O₃ at similar loadings. Table 1 shows the extent of pore filling of sorbents determined based on the N₂ physisorption isotherms of the bare support materials and amine-impregnated powder sorbents (Figure S1). The theoretical extent of pore filling was also calculated with an assumption that all TEPA was impregnated inside the pores of the support materials. The 30 wt % TEPA-impregnated MIL-101(Cr) [MIL-101(Cr)_TEPA(30)] and 20 wt % TEPA-impregnated γ-Al₂O₃ [Al₂O₃_TEPA(20)] show almost identical extents of pore filling (25.4–27.6%). Close values between the measured and theoretical pore filling indicate that TEPA was effectively impregnated inside the pores of both MIL-101(Cr) and γ-Al₂O₃.

Table 1. Physical Properties of Support Materials and Adsorbents.

support materials	MIL-101(Cr)	γ-Al2O3
particle size, [μm]	∼0.544	6.9a
pore volume, [cc/g]	1.84	0.95
pore size (mode), [nm]	2.4	16.1
BET surface area, [m2/g]	3270	153
amine (TEPA) loadings, [wt %]	30	20
powder sorbents	MIL-101(Cr)_TEPA(30)	Al2O3_TEPA(20)
pore filling, [ %]	25.4 (by N2 physisorption)	27.6 (by N2 physisorption)
	23.3 (by calculation)	26.5 (by calculation)

^aMean particle size was determined based on SEM images (Figure S2).

Humid and Dry CO₂ Adsorption Behavior of Amine (TEPA)-Impregnated MIL-101(Cr) and γ-Al₂O₃ at Ambient and Cold Temperatures

To study the effects of temperature and humidity on the CO₂ adsorption behavior of impregnated amines (TEPA) within different porous support materials, dry and humid (70% RH) 400 ppm CO₂ breakthrough experiments were conducted with MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20) powder sorbents at −20, −5, and 25 °C. Figure S3 shows the dry and humid (70% RH) CO₂ breakthrough curves of the powder sorbents under different temperature conditions. Pseudo-equilibrium 400 ppm CO₂ adsorption capacities of MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20) were determined based on the breakthrough curves, and the results are shown in Figure 1a,b, respectively, with amine efficiencies.

Figure 1.

Pseudo-equilibrium 400 ppm CO₂ adsorption capacities and amine efficiencies of (a) 30 wt % TEPA-impregnated MIL-101(Cr) and (b) 20 wt % TEPA-impregnated γ-Al₂O₃ powder adsorbents under dry and humid (70% RH) conditions at −20, −5, and 25 °C. Both materials have similar pore filling with TEPA.

The 400 ppm CO₂ uptake of MIL-101(Cr)_TEPA(30) under dry conditions was significantly enhanced with decreasing temperature, from 0.11 mmol/g (25 °C) to 1.11 mmol/g (−20 °C) because the overall CO₂ capture capacities are dominantly affected by thermodynamics rather than mass transfer due to the low degree of pore filling (25.4%, see Table 1).⁴⁴ On the other hand, TEPA in γ-Al₂O₃ showed a different trend. The 400 ppm CO₂ uptake of Al₂O₃_TEPA(20) (1.10 mmol/g) at 25 °C gradually reduced with decreasing temperature down to −20 °C (0.81 mmol/g) even with a much larger pore size (16.1 nm > 2.4 nm) and similar extent of pore filling compared to MIL-101(Cr)_TEPA(30) (see Table 1). Interestingly, at −20 °C, the MIL-101(Cr)_TEPA(30) showed higher 400 ppm CO₂ uptake (1.11 mmol/g) with lower amine efficiency (0.14) than Al₂O₃_TEPA(20) (0.81 mmol/g with 0.15 amine efficiency), suggesting that a larger amount of CO₂ physisorption probably occurred in the CO₂ capture by the MIL-101(Cr)_TEPA(30) due to the smaller pore size of MIL-101(Cr) than γ-Al₂O₃ (2.4 nm < 16.1 nm).

Under humid conditions, the overall DAC performances of the two powder sorbents were enhanced. Similar observations for amine sorbents in the presence of humidity have been attributed to more carbamate/carbamic acid species forming due to the freeing of amines from support interactions^17,67,68 or the formation of bicarbonate under humid conditions.⁶⁹ The form(s) of captured CO₂ under dry and humid conditions are further discussed below with insight from in situ FT-IR experiments. Moreover, MIL-101(Cr)_TEPA(30) was shown to be more significantly affected by the presence of humidity than Al₂O₃_TEPA(20). These different CO₂ capture behaviors of the MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20) under dry and humid conditions strongly suggest that the local environment of TEPA within each support is different. To further probe the impact of the support materials (or pore structures) on the CO₂ adsorption mechanism(s) of the impregnated amine (TEPA), CO₂/H₂O temperature-programed desorption (TPD) and in situ FT-IR experiments were conducted.

Representative CO₂/H₂O TPD experimental results are shown in Figure S4. The desorbed CO₂ and H₂O during the TPD experiments were quantified as a function of time (Figure S5). The total amounts of desorbed H₂O from the MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20) materials after humid (70% RH) 400 ppm CO₂ breakthroughs are plotted in Figure 2a as a function of adsorption temperature. Since the humid CO₂ capture experiments were conducted at constant relative humidity (RH) conditions, 70% RH, under varied adsorption temperatures from −20 to 25 °C, the absolute humidity range was very wide (870 ppm–2.2% by volume). As a result, we observed that the H₂O uptake by MIL-101(Cr)_TEPA(30) did not consistently change with adsorption temperature conditions. While the H₂O uptake slightly increased from 25 °C (32.7 mmol/g) to −5 °C (35.2 mmol/g), it decreased at −20 °C (28.7 mmol/g), likely due to the extremely low absolute humidity (870 ppm) at this temperature. Overall, these uptake values are consistent with liquid-like conditions in the pores of the MIL-101(Cr). In the case of Al₂O₃_TEPA(20), the H₂O sorption continuously decreased with decreasing adsorption temperature, from 25 °C (13.5 mmol/g) to −20 °C (8.3 mmol/g).

Figure 2.

(a) Amount of desorbed H₂O from TEPA-impregnated MIL-101(Cr) and γ-Al₂O₃ powder adsorbents during the CO₂/H₂O TPD. Amount of desorbed dry/humid CO₂ from (b) 30 wt % TEPA-impregnated MIL-101(Cr) and (c) 20 wt % TEPA-impregnated γ-Al₂O₃ powder adsorbents during the CO₂/H₂O TPD. Strong chemisorption: CO₂ desorption above 25 °C; weak chemisorption: CO₂ desorption under 25 °C.

Interestingly, the MIL-101(Cr)_TEPA(30) showed significantly higher H₂O uptake than the Al₂O₃_TEPA(20) under all temperature conditions. This is likely due to the much smaller pore size of MIL-101(Cr) (2.4 nm) than γ-Al₂O₃ (16.1 nm) (Table 1), yielding higher capillary condensation at lower partial pressures of water, as shown in Figure S6. It was reported that the process of capillary condensation of MIL-101(Cr) begins at 35% RH at 25 °C,⁷⁰ while γ-Al₂O₃ begins at 65% RH at 22–25 °C,^71,72 which is consistent with measured water vapor adsorption isotherm in this study (Figure S6). When compared to the humid 400 ppm CO₂ adsorption results shown in Figure 1a,b, one can see that the H₂O uptake is more significant during the humid DAC process at all temperature conditions. Thus, assuming similar ΔH_ads(H20) for these materials, more energy will be required for the regeneration of MIL-101(Cr)_TEPA(30) than for Al₂O₃_TEPA(20) for both ambient and cold humid DAC processes due to the significant H₂O uptake. However, it should be noted that the H₂O uptake kinetics are extremely slow at cold temperature conditions due to the low absolute humidity and the effect of temperature on the overall kinetics (Figure S7). Thus, if water sorption is kinetically limited, regeneration energies can be significantly reduced.

In this study, weak chemisorption (including physisorption) is defined as CO₂ desorption below 25 °C and strong chemisorption is defined as CO₂ desorption above 25 °C, as discussed in Supporting Information (Figure S8). The amounts of weak and strong chemisorption were quantified based on Figure S5 and the results are plotted in Figure 2b,c for the MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20), respectively. The figures clearly show that the impregnated TEPA within MIL-101(Cr) and γ-Al₂O₃ exhibit different CO₂-amine interactions. Inside smaller pores (2.4 nm) with higher remaining pore volume (1.84 cc/g), as in MIL-101(Cr), TEPA showed weak chemisorption dominant behaviors.⁴⁴ Inside a larger pore (16.1 nm) support with lower pore volume (1.03 cc/g), as in γ-Al₂O₃, strong chemisorption was the dominant CO₂ capture mechanism of TEPA. Both weak and strong chemisorption was enhanced under humid conditions.

Effect of Support Materials on the CO₂ Adsorption Mechanism(s) of Impregnated Amines (TEPA) under Dry Conditions

To identify the adsorbed CO₂ species associated with weak and strong chemisorption, in situ FT-IR experiments were conducted at 25 and −20 °C under dry 400 ppm CO₂ environments (Figure 3). It should be noted that these spectra were obtained after subtracting the activated sorbent spectra from the IR spectra after a given time of exposure to dry 400 ppm CO₂ to observe contributions more clearly from CO₂ adsorption. The relevant peak assignments are summarized in Table S1 with supporting references.

Figure 3.

In situ FT-IR spectra of (a/b) 20 wt % TEPA-impregnated γ-Al₂O₃ and (c/d) 30 wt % TEPA-impregnated MIL-101(Cr) powder adsorbents as a function of adsorption time at (a/c) 25 °C and (b/d) −20 °C with the activated sample as the background. Orange: carbamic acid, red: carbamate ion, purple: ammonium ion. Adsorption conditions: gas, 400 ppm CO₂/N₂; flow rate, 40 sccm; relative humidity, dry; activation, 60 °C under 40 sccm N₂ for 2–3 h.

Results shown in Figure 3a,c indicate that the dominant adsorbed CO₂ species on the Al₂O₃_TEPA(20) and MIL-101(Cr)_TEPA(30) at 25 °C are different. The TEPA-impregnated MIL-101(Cr) appears to dominantly form carbamic acid with a small amount of ammonium carbamate ion pairs, while the TEPA-impregnated γ-Al₂O₃ adsorbents only or primarily form ammonium carbamate ion pairs at the same conditions. It is known that carbamate ion species (strong chemisorption) have higher binding energies than carbamic acid species (weak chemisorption),^73,74 and the form of adsorbed CO₂ here correlates well with the ease of desorption, as determined by the TPD experiments.

The measured FT-IR spectra of Al₂O₃_TEPA(20) and MIL-101(Cr)_TEPA(30) under dry 400 ppm CO₂ flow at −20 °C (sub-ambient) are shown in Figure 3b,d, respectively. It appears that the same adsorbed CO₂ species formed at 25 °C were also formed at −20 °C on each sorbent, while there are differences in the CO₂ adsorption kinetics and the total amount of adsorbed CO₂ species at equilibrium under the different temperature conditions. Obvious increases in the intensity of adsorbed CO₂ in both adsorbents were observed as a function of time at 25 °C, whereas the intensities relatively slowly increased at −20 °C, indicating reduced CO₂ adsorption kinetics at cold temperature conditions, as also observed in breakthrough experiments described above. The overall intensity of adsorbed CO₂ on Al₂O₃_TEPA(20) at equilibrium slightly decreased from 25 to −20 °C, suggesting that the formation of ammonium carbamate ion pairs may be suppressed at sub-ambient temperature conditions. Given that carbamate formation requires cooperative interaction of two amine sites, lower amine chain mobility at low temperatures may impede carbamate formation. In the case of carbamic acid formation on the MIL-101(Cr)_TEPA(30) powder sorbents, sharper and higher equilibrium intensities were observed at −20 °C, even with reduced CO₂ adsorption kinetics, indicating the enhanced formation of carbamic acid. These in situ FT-IR results are consistent with the CO₂ desorption data shown in Figure 2b,c. The effects of temperature on the formation of carbamic acid and ammonium carbamate ion pairs will be further discussed below.

Also noteworthy is that an apparent physisorbed CO₂ peak at 2335 cm^–1 was observed on the MIL-101(Cr)_TEPA(30) at −20 °C, while the Al₂O₃_TEPA(20) did not show any distinct CO₂ physisorption peak, as depicted in Figure S9. Thus, it can be inferred that the higher pseudo-equilibrium 400 ppm CO₂ uptake of MIL-101(Cr)_TEPA(30) vs Al₂O₃_TEPA(20) with lower amine efficiency at −20 °C (Figure 1) is due to a relatively larger amount of CO₂ physisorption over MIL-101(Cr)_TEPA(30). We infer that TEPA impregnation inside the small pores of MIL-101(Cr) (2.4 nm) does two things: (i) it leaves relatively isolated primary amines leading to carbamic acid formation, and (ii) it leaves behind suitable micropore space for physisorption of CO₂.

The dry CO₂ TPD profiles of MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20) shown in Figure 4a clearly show the difference in CO₂ desorption trends (weak and strong chemisorption). A desorption peak of MIL-101(Cr)_TEPA(30) was shown at a relatively low temperature between 30 and 35 °C for the 25 °C adsorption temperature, and it was shifted to 2.2 °C when decreasing the adsorption temperature down to −20 °C, indicating that the weak chemisorption became more significant at colder adsorption temperatures.⁴⁴ The in situ FTIR results shown in Figure 3 strongly indicate that the weak chemisorption is related to the formation of carbamic acid. In contrast, a desorption peak of Al₂O₃_TEPA(20) was confirmed at a much higher temperature (53 °C), consistent with the strong chemisorption mechanism associated with formation of ammonium carbamate ion pairs (Figure 3).

Figure 4.

(a) CO₂ TPD profiles and (b) energy of CO₂ desorption of 30 wt % TEPA-impregnated MIL-101(Cr) and 20 wt % TEPA-impregnated γ-Al₂O₃ powder adsorbents for dry 400 ppm adsorption at −20, −5, and 25 °C. ΔH_ad,CO2: measured heat of dry 400 ppm CO₂ adsorption by TGA/DSC.

Interestingly, the desorption peak position and the amount of desorbed CO₂ (peak area) of the Al₂O₃_TEPA(20) adsorbent were not significantly changed when decreasing the adsorption temperature, indicating that the formation of ammonium carbamate (a strongly exothermic reaction) is not significantly affected by the adsorption temperature, although lower temperatures enthalpically favor the adsorption process. On the other hand, the formation of carbamic acid (a less exothermic reaction) was enhanced at cold temperatures (Figure 4a). These different trends are probably due to the effects of temperature on amine mobility, in which amine mobility dramatically decreases with decreasing temperature conditions.⁷⁵ The formation of ammonium carbamate ion pairs occurs via a reaction between two amine groups (primary or secondary) and CO₂ with a 2:1 molar ratio, while the reaction ratio is 1:1 for the formation of carbamic acid. We hypothesize that the formation of the ammonium carbamate ion pairs may be more strongly impacted by the decrease in amine mobility at cold temperatures than the formation of carbamic acid due to the higher amine to CO₂ reaction ratio (2 > 1). Thus, the expected large thermodynamic enhancement of the more exothermic carbamate ion pair formation is mitigated to some degree by arrested amine chain mobility. In contrast, chain motion is less important for carbamic acid formation, requiring only a single adsorption site, and thus, thermodynamic effects are more prominent. Thus, the MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20) show different DAC behavior at ambient and sub-ambient temperature conditions (Figure 1).

The CO₂ TPD results shown in Figure 4a indicate that the interaction strength between CO₂ and the impregnated TEPA is different depending on the support materials (MIL-101(Cr) and γ-Al₂O₃) and adsorption temperature conditions. To quantify the binding energy of dry CO₂ onto the adsorbent materials, CO₂ TPD experiments were performed with varied heating rates and the energy of CO₂ desorption, E_d, was calculated based on the method described in the Supporting Information. The procedure is described in more detail in Figures S10 and S11. The heat of CO₂ adsorption is an indicator of the strength of the interaction between the CO₂ molecules and amine sites on sorbents, and it is released as heat during the exothermic CO₂ adsorption process. The total energy of CO₂ desorption, as defined in this work, is the energy required to overcome the thermodynamic barrier for desorption (ΔH_ads) + the energy needed to overcome any kinetic barrier to desorption (EA_des). Sensible heats are typically small under these conditions and are therefore often neglected.

The results are plotted in Figure 4b for MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20), as a function of adsorption temperature. The E_d of MIL-101(Cr)_TEPA(30) for 25 °C adsorption was not determined because most of captured CO₂ was desorbed during the N₂ purging period at 25 °C (Figure S5e). The heat of dry 400 ppm CO₂ adsorption, as measured in situ by TGA/DSC, is also plotted. Since the energy required for the desorption process (E_d) is comparable to the heat released during the adsorption process (ΔH_ads), the kinetic barrier for desorption (e.g., activation energy, EA_des) is very small or negligible, and thus, the desorption energy determined by the TPD method approximates the heat of adsorption.⁷⁶ The E_d of MIL-101(Cr)_TEPA(30) was slightly reduced (69.6 kJ/mol_CO2 to 62.4 kJ/mol_CO2) with decreasing the adsorption temperature from −5 to −20 °C because of enhanced weak chemisorption (or formation of carbamic acid) at cold temperature conditions (Figure 4a). In contrast, the CO₂ desorption peak of the Al₂O₃_TEPA(20) was not shifted under different adsorption temperature conditions (Figure 4a), resulting in a similar E_d (99.9–101.5 kJ/mol_CO2), consistent with strong chemisorption (or formation of ammonium carbamate pairs) at all adsorption temperature conditions, assuming ideal mixing (i.e., , the sorption enthalpy). These results clearly show that the impregnated TEPA has different CO₂ capture mechanisms (carbamic acid vs carbamate) within different support materials, and the ammonium carbamate pair (100.5 kJ/mol_CO2) has stronger binding energy than carbamic acid (62.4 kJ/mol_CO2).^74,77,78 The regeneration of MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20) saturated with dry 400 ppm CO₂ at −20 °C was monitored with in situ FTIR, and the results shown in Figure S12 support the assertion that carbamate species are stronger chemisorption products than carbamic acid.

CO₂ Adsorption Via Carbamic Acid Versus Ammonium Carbamate Ion Pair Formation

The abovementioned results are consistent with the previous work of Bacsik and Hedin (2011), who previously elucidated the formation of carbamate species vs carbamic acid species over amine-based sorbents.^79,80 Their in situ FT-IR study revealed that the relatively low amine-loaded silica (AMS-6) adsorbents captured CO₂ by forming both carbamic acid and ammonium carbamate pairs, while the high amine-loaded silica primarily formed ammonium carbamate ion pairs during CO₂ adsorption (Figure S13a). Another work by Didas et al. (2014) showed similar results (Figure S13b).⁸¹ Yoo et al. (2015)⁷³ controlled the formation of carbamic acid and ammonium carbamate pairs when amines react with CO₂ by functionalizing a mesoporous silica SBA-15 with single (MONO) and triaminosilanes (TRI), as shown in Figure S14. Since the length of the MONO molecules was estimated to be sufficient for the interaction between primary amines (−NH₂) and silanols (−OH) on the surface of SBA-15 but not for interaction with other amines,^82,83 the MONO grafted adsorbents captured CO₂ via interaction between primary amines and surface silanols, forming carbamic acid species stabilized by surface silanols with low isosteric heats of adsorption (∼−46 kJ/mol).⁷³ On the other hand, the TRI grafted adsorbents captured CO₂ via intramolecular amine–amine interactions within a single chain, forming ammonium carbamate ion pairs that have higher heats of adsorption of ∼−73 to −78 kJ/mol than the MONO adsorbents. It can be inferred from this prior study that the formation of carbamic acid is probably dominant when the interaction between impregnated amines and support materials is strong. These studies indicate that enough surface silanols are required (strong amine-support material interactions) to form carbamic acid stabilized by surface silanol groups, while ammonium carbamate pairs are the dominant species of adsorbed CO₂ when amine–amine interactions are strong via amine clustering.

In the same manner, one possible cause for the different CO₂ adsorption mechanisms of impregnated TEPA within the two support materials studied here, MIL-101(Cr) and γ-Al₂O₃, could be the result of different extents of interaction between the impregnated amines and the support materials. As shown in Table 1, MIL-101(Cr) has a much higher BET surface area to pore volume ratio of 1.78 nm^–1 than γ-Al₂O₃ (0.16 nm^–1), indicating that an equivalent amount of impregnated amines (filling a similar pore volume percentage) may more strongly interact with MIL-101(Cr) than γ-Al₂O₃. This is consistent with the observation that the dominant CO₂ adsorbed species on MIL-101(Cr)_TEPA(30) adsorbents were carbamic acids, while the Al₂O₃_TEPA(20) primarily formed ammonium carbamate pairs during dry 400 ppm CO₂ adsorption, as shown in Figure 3.

The empirical formula of MIL-101(Cr) is Cr₃(O)OH(bdc)₃(H₂O), where bdc is benzene-1,4-dicarboxylate. The octahedral trinuclear Cr(III)₃O building units of MIL-101(Cr) are saturated with one hydroxyl group and two water molecules. MIL-101(Cr) is generally activated at 150 °C under a high vacuum to remove the water molecules, creating potential Lewis acid sites.⁸⁴⁻⁸⁶ Since the MIL-101(Cr)_TEPA(30) was activated at 60 °C under 1 atm N₂ flow, the open metal sites of MIL-101(Cr) saturated with water during the amine impregnation process with MeOH solution might be mostly occupied with water molecules even after the activation process due to the mild conditions used. The theoretical maximum amount of hydroxyl groups and water molecules on the surface of MIL-101(Cr) is 4.2 mmol OH and H₂O/g MOFs. Thus, as shown in Figure S15a, the formed carbamic acid on the MIL-101(Cr)_TEPA(30) adsorbents might be stabilized by water molecules or hydroxyl groups occupying the open metal sites of MIL-101(Cr).

The Al₂O₃_TEPA(20) may also contain hydroxyl groups on the surface of alumina because it was confirmed that traces of water are generated from γ-Al₂O₃ during heating, even after activation at 1000 °C.⁸⁷ Hydroxylation of the alumina surface occurs by the dissociative adsorption of water to Al³⁺, where Lewis acid sites and residing protons on lattice oxygens form surface hydroxyl groups.⁸⁸ The reported heat of water adsorption on γ-Al₂O₃ at 25 °C under low water coverage (0.4 H₂O/nm²) is −158.8 kJ/mol,⁷¹ indicating that the activated Al₂O₃_TEPA(20) at 60 °C under N₂ flow certainly contains surface hydroxyl groups. Lagauche et al. also reported that the coverage of surface OH groups on the activated γ-Al₂O₃ at 100 °C for 4 h is about 11.7 OH/nm² or 4.4 mmol OH/g alumina,⁸⁹ which is comparable to the theoretical maximum of MIL-101(Cr). Although γ-Al₂O₃ has a similar amount of surface hydroxyl groups to MIL-101(Cr), the degree of interaction between impregnated amine groups and hydroxyl groups on the surface of alumina will be much lower than in the case of MIL-101(Cr) due to the much lower BET surface area to pore volume ratio of γ-Al₂O₃ (0.16 nm^–1) than that of MIL-101(Cr) (1.78 nm^–1), as discussed above. Thus, it can be inferred that ammonium carbamate pairs are dominantly formed via amine–amine interactions during the CO₂ adsorption by Al₂O₃_TEPA(20), probably due to the lower degree of interaction between impregnated TEPA and γ-Al₂O₃ (Figure S15c).

The amine-impregnated MIL-101(Cr) powder sorbents also adsorbed 400 ppm CO₂ via the formation of ammonium carbamate pairs when amine loading was high enough for significant amine clustering to occur (50 wt % TEPA). As shown in Figure S16, the CO₂ TPD⁴⁴ and in situ FT-IR studies revealed that both strong (ammonium carbamate) and weak chemisorption (carbamic acid) are involved in the CO₂ adsorption by high amine loaded MIL-101(Cr) adsorbents (Figure S15b), while the low amine loaded adsorbents only showed weak chemisorption (carbamic acid) dominant behavior at −20 °C (Figure S15a). It appears that the CO₂ capture mechanisms of amine-impregnated porous solid sorbents, including the formation of carbamic acid (weak chemisorption) or ammonium carbamate (strong chemisorption), can be controlled by adjusting the extent of pore filling and interactions between amine and the surface of support materials. Thus, the selection of support materials for amine impregnation could be an important factor in determining the DAC performance of the amine-impregnated solid adsorbent materials.

Effect of Moisture on CO₂ Adsorption Behavior of Amine (TEPA) within Different Support Materials (MIL-101(Cr) and γ-Al₂O₃)

The dry/humid 400 ppm CO₂ breakthrough experiment results described in Figure 1 clearly show that moisture has a positive effect on the DAC performance of the TEPA-impregnated MIL-101(Cr) and γ-Al₂O₃ powder sorbents at both ambient (25 °C) and sub-ambient (−5 and −20 °C) temperatures. Similar results were also reported in our previous work with the amine-impregnated MIL-101(Cr).⁴⁴ To gain in-depth understanding of the effects of humidity, the CO₂/H₂O TPD profiles after humid (70% RH) CO₂ adsorption are compared with the results of dry CO₂ adsorption at different adsorption temperatures (25, −5, and −20 °C) in Figure S17. At the same time, the measured in situ FT-IR spectra (1750–1250 cm^–1) during the humid 400 ppm CO₂ adsorption (70% RH) at 25 and −20 °C are shown in Figure S18 to support the discussion of the effects of humidity. Figure S19 shows overall in situ FT-IR spectra (3750–1250 cm^–1) under the same conditions. After introducing humid 400 ppm CO₂ into the DRIFTS cell, a dramatic peak increase was observed at 3750 cm^–1, which is attributed to H₂O adsorption.^74,90,91 Much faster H₂O uptake was observed at 25 °C, while the uptake kinetics at −20 °C were extremely slow, as discussed in Figure S7. In situ FT-IR desorption experiments were also conducted after humid 400 ppm CO₂ adsorption at 25 °C to investigate the thermal decomposition behavior of carbamic acid and ammonium carbamate ion pairs formed under humid environments (Figure S20).

The CO₂/H₂O TPD and in situ FT-IR results show that the formation of carbamic acid and ammonium carbamate ion pairs on MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20), respectively, was enhanced upon humid 400 ppm CO₂ adsorption. The enhancement in CO₂ adsorption under humid conditions became particularly more significant at −20 °C, indicating that the effect of humidity on the DAC performance becomes more significant with decreasing adsorption temperature. The enhanced carbamic acid formation under humid conditions may be because more water-stabilized carbamic acid was formed (via hydrogen bonding) with adsorbed water molecules.⁷⁴ More ammonium carbamate ion pairs also can be formed with water vapor adsorption due to water’s ability to free amine sites for amine–amine or amine-support interactions.^17,80,92,93 More detailed discussion on the effect of moisture is described in Supporting Information.

Interestingly, a very small peak for the symmetric νCOO^– vibration of bicarbonate anions was observed at 1358 cm^–1 during humid 400 ppm CO₂ adsorption with the Al₂O₃_TEPA(20) at 25 and −20 °C, while clear evidence for the formation of bicarbonate under humid conditions was not observed from the MIL-101(Cr)_TEPA(30) (Figure S18). This is probably because ammonium bicarbonate is formed from a paired ammonium carbamate precursor.^74,81,94 As shown in Figures 3 and S18, since the formation of ammonium carbamate ion pairs is the dominant CO₂ capture mechanism of Al₂O₃_TEPA(20), the formation of bicarbonate within Al₂O₃_TEPA(20) may be more favorable under humid conditions when compared to the MIL-101(Cr)_TEPA(30). However, it has been reported that observation of bicarbonate formation on solid-supported amines under humid CO₂ capture conditions is challenging with in situ FT-IR conditions due either to the small amounts of bicarbonate formed or due to extinction coefficients that make resolution difficult.^95,96 Thus, the formation of bicarbonate under humid conditions should be further studied with ¹³C solid-state NMR in the future.⁹⁶⁻¹⁰⁰

The CO₂/H₂O TPD profiles of MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20) are shown in Figure 5a,b, respectively, for different humid CO₂ adsorption temperature conditions (−20, −5, and 25 °C). The amount of desorbed CO₂ from the two powder sorbents dramatically increased with decreasing adsorption temperatures from 25 to −20 °C, indicating that the effect of moisture on 400 ppm CO₂ adsorption became more significant. The binding energy of CO₂ on the powder sorbents may be affected by H₂O adsorption. Thus, quantifying the binding energy of CO₂ and H₂O under humid conditions is necessary to estimate the energy required for sorbent regeneration. Generally, the isosteric heat of adsorption is calculated from multiple isotherms measured at different temperature conditions. However, measuring the adsorption isotherm of CO₂ under humid conditions (co-adsorption of CO₂ and H₂O) is challenging. Thus, the energies of CO₂ and H₂O desorption, E_d, for humid CO₂ adsorption, were determined based on the TPD method described in Supporting Information. Figures S21 and S22 show the CO₂/H₂O TPD profiles of the MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20), respectively, with varied heating rates for humid CO₂ adsorption.

Figure 5.

(a/b) CO₂/H₂O TPD profiles and (c/d) energy of CO₂ and H₂O desorption of (a/c) 30 wt % TEPA-impregnated MIL-101(Cr) and (b/d) 20 wt % TEPA-impregnated γ-Al₂O₃ powder adsorbents for humid (70% RH) 400 ppm adsorption at −20, −5, and 25 °C.

The determined E_ds of CO₂ and H₂O for humid CO₂ adsorption are plotted in Figure 5c for MIL-101(Cr)_TEPA(30) and (d) for Al₂O₃_TEPA(20) along with the results of dry CO₂ desorption (Figure 4c,d) as a function of adsorption temperature conditions. It should be noted that T_ms for some conditions could not be identified in the TPD profiles due to significant CO₂ or H₂O desorption during the N₂ purging step, and thus, we were not able to determine E_d for those conditions. The determined desorption energy of water for Al₂O₃_TEPA(20) was 46.4 kJ/mol_H2O, which is comparable to the heat of vaporization of water in the range of 0–40 °C and 43.4–45.1 kJ/mol_H2O, supporting the condensation-like sorption of water by Al₂O₃_TEPA(20). A previous study directly measured the heat of adsorption of γ-Al₂O₃ via microcalorimetry of water adsorption, and it was −44 kJ/mol_H2O at a relative water saturation pressure of 0.7 at 25 °C.⁷¹ In the case of MIL-101(Cr)_TEPA(30), the measured desorption energy of water (47.6–59.4 kJ/mol_H2O) was slightly higher than that of Al₂O₃_TEPA(20). The reported isosteric heat of water adsorption on MIL-101(Cr) is from −53 to −44 kJ/mol_H2O.⁷⁰ The higher heat of water desorption on the MIL-101(Cr)_TEPA(30) than Al₂O₃_TEPA(20) might be linked to the smaller pore size of the MOF, as shown in Table 1. Thus, the MIL-101(Cr)_TEPA(30) showed much higher H₂O uptake than Al₂O₃_TEPA(20) under all adsorption temperature conditions (Figure 2a).

The calculated energy required for desorption of dry and humid CO₂, shown in Figure 5c,d, indicates that the binding energy of CO₂ increases under humid conditions, and the effect of moisture on CO₂ adsorption is more significant with decreasing adsorption temperature. The difference between wet and dry CO₂ desorption energies increased with decreasing adsorption temperatures from 25 to −20 °C. At −20 °C, the desorption energy of CO₂ for Al₂O₃_TEPA(20) increased from 100.5 kJ/mol_CO2 (dry) to 121.1 kJ/mol_CO2 under humid conditions. The MIL-101(Cr)_TEPA(30) also showed similar behavior. It can be inferred from the CO₂ TPD profiles in Figure S17e,f that the increased CO₂ desorption energy for humid CO₂ adsorption could be mostly attributed to enhanced strong chemisorption (desorption above 25 °C), which is related to the formation of carbamate species. The ammonium carbamate ion pairs formed under humid conditions may have higher binding energies than the carbamate species formed under dry conditions.^101,102 Cyclic diamine 2-(aminomethyl)piperidine (2-ampd) functionalized Mg₂(dobpdc) MOFs also show similar trends. While the DFT-calculated heat of CO₂ adsorption was −70 kJ/mol for the formation of ammonium carbamate chains under dry conditions, the calculated humid CO₂ adsorption energy was −88 kJ/mol in the case of the co-adsorption of 1 CO₂ and 1 H₂O per diamine, indicating that H₂O enhances the CO₂ binding energy for the ammonium carbamate chain formation.¹⁰³

The results in Figure 5c,d initially suggest that much higher desorption energies could be required for operation of DAC process under humid sub-ambient conditions (e.g., −20 °C) than ambient conditions (e.g., 25 °C) due to the significant increase in the binding energy of chemisorbed CO₂ under humid conditions at −20 °C. However, it should be noted that all the data reported in this study are pseudo-equilibrium values. As discussed in Figure S7, the two powder sorbents have extremely slow H₂O uptake kinetics when compared to CO₂ uptake kinetics, especially at sub-ambient temperature conditions (e.g., −20 °C), due to an extremely low absolute humidity level and the effect of cold temperatures on kinetics of chemical processes. Thus, in realistic sub-ambient DAC operation, the H₂O adsorption working capacity could be very low, depending on the adsorption time and the effect of adsorbed H₂O on the 400 ppm CO₂ adsorption may not be significant due to the kinetically limited H₂O uptake at sub-ambient temperature conditions (e.g., −20 °C).

To support the abovementioned argument, H₂O uptake profiles of the MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20) were obtained by integrating the H₂O breakthrough curves in Figure S7 and are plotted in Figure S23 as a function of time. The dashed lines in the figures depict H₂O adsorption working capacity at the time required to reach pseudo-equilibrium states (95% of breakthrough) of humid 400 ppm CO₂ adsorption during the breakthrough experiments in Figure S3. Figure S23 indicates that if the DAC process is operated until the pseudo-equilibrium states at each adsorption temperature, the H₂O adsorption working capacity could dramatically decrease with decreasing temperatures due to significant kinetic limitation of H₂O uptake at colder temperatures. The H₂O adsorption working capacities of MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20) were decreased from 27.0 and 13.5 mmol/g at 25 °C to 7.7 and 4.1 mmol/g at −20 °C, respectively. The H₂O uptakes of both sorbent materials were reduced by 3.3–3.5 times when decreasing the adsorption temperature from 25 to −20 °C, indicating that significant energy saving for water desorption may be achieved under sub-ambient DAC operation.

The CO₂/H₂O uptakes and total energies for CO₂ and H₂O desorption of the two sorbent materials are compared in Table 2 to estimate the energy requirements for sorbent regeneration. These energies are broken down into the energy for desorption (ΔH_ads + EA_des) and the sensible heat for heating the sample from the adsorption to the desorption temperature. The energies of CO₂ and H₂O desorption were calculated based on the measured CO₂ and H₂O uptakes (Figure 2) and the E_d for CO₂/H₂O desorption determined by the TPD method (Figures 4 and 5). Under dry conditions, the MIL-101(Cr)_TEPA(30) sorbents are better sorbent materials for sub-ambient temperature DAC operation (e.g., −20 °C) due to their higher CO₂ uptakes and lower regeneration energies than Al₂O₃_TEPA(20). For dry ambient temperature operations (e.g., 25 °C), the Al₂O₃_TEPA(20) appears to be an advantaged sorbent because of its higher CO₂ uptake under such conditions. When the sorbent materials are fully saturated with water, their DAC performance is significantly enhanced. However, the sorbent regeneration energy requirement dramatically increases at the same time due to significant amounts of H₂O adsorption. In particular, DAC operation with MIL-101(Cr)_TEPA(30) is very energy-intensive if water is completely removed in every cycle because of its high H₂O uptake. However, as discussed above, the sorbent materials are never fully saturated with H₂O, especially at sub-ambient temperature conditions (e.g., −20 °C) due to kinetically limited H₂O adsorption. When the estimated H₂O working capacity (Figure S23) is considered under kinetic conditions in the calculation, both sorbent materials show much lower regeneration energy requirements for DAC operation at −20 than at 25 °C, indicating the advantage of sub-ambient DAC operation.

Table 2. DAC Performance and Energies for CO₂/H₂O Desorption of MIL-101(Cr)_TEPA(30) and Al₂O₃_TEPA(20) Sorbent Materials (Desorption at 60 °C Under N₂).

	MIL-101(Cr)_TEPA(30)		Al2O3_TEPA(20)
adsorption temperature, [°C]	–20	25	–20	25
Dry
qCO2, [mol/kg]	1.11	0.11	0.81	1.10
qH2O, [mol/kg]	0.00	0.00	0.00	0.00
, [kJ/mol]	62.4	62.4	100.5	99.8
, [kJ/mol]	53.5		46.4
energy for CO2 desorption, [GJ/ton CO2]	1.42	1.42	2.28	2.27
energy for H2O desorption, [GJ/ton CO2]	0.00	0.00	0.00	0.00
sensible heat for sorbent, [GJ/ton CO2]	1.46	6.45	1.92	0.62
Kinetically Limited H2O Adsorption
qCO2, [mol/kg]	1.87	0.38	1.98	1.33
qH2O, [mol/kg]	7.70	27.00	4.10	13.46
, [kJ/mol]	91.7	62.4	121.1	104
, [kJ/mol]	53.5		46.4
energy for CO2 desorption, [GJ/ton CO2]	2.08	1.42	2.75	2.36
energy for H2O desorption, [GJ/ton CO2]	5.01	86.4	2.18	10.7
sensible heat for sorbent, [GJ/ton CO2]	0.87	1.87	0.78	0.51
Full H2O Adsorption
qCO2, [mol/kg]	1.87	0.38	1.98	1.33
qH2O, [mol/kg]	28.7	32.7	8.3	13.5
, [kJ/mol]	91.7	62.4	121.1	104
, [kJ/mol]	53.5		46.4
energy for CO2 desorption, [GJ/ton CO2]	2.08	1.42	2.75	2.36
energy for H2O desorption, [GJ/ton CO2]	18.7	104.6	4.44	10.7
sensible heat for sorbent, [GJ/ton CO2]	0.87	1.87	0.78	0.51

For DAC operations, Table 2 indicates that the Al₂O₃_TEPA(20) will be better sorbents than MIL-101(Cr)_TEPA(30) due to its lower H₂O uptake. However, it should be noted that the MIL-101(Cr) has a weak chemisorption dominant mechanism, and thus, a small temperature swing can be applied. The DAC performance and energies of CO₂/H₂O desorption for small temperature swing operation (adsorption at −20 °C and regeneration at 25 °C) are summarized in Table S2. Due to the weak chemisorption dominant mechanism, MIL-101(Cr)_TEPA(30) shows a much higher CO₂ working capacity, 1.55 mmol/g, than Al₂O₃_TEPA(20), 0.43 mmol/g, for the small temperature swing operation, resulting in lower sorbent regeneration energy.

Experimental results of this study demonstrate that adjusting interactions between amine and the surface of the support materials can control the CO₂ capture mechanisms of amine-impregnated sorbent materials, varying from the formation of carbamic acid (weak chemisorption) to ammonium carbamate (strong chemisorption). DAC performances, including CO₂ uptake and energy requirement for sorbent regeneration, of these two different mechanisms were strongly affected by adsorption temperatures and humidity. These condition-dependent results demonstrate the complexity of designing and operating a DAC process, as there will likely be no optimal sorbent under all conditions in most locations (day/night; summer/winter; rainy/dry conditions), and a material that operates with acceptable performance under all or most of the conditions likely to be encountered is required.⁵⁰

Conclusions

TEPA (tetraethylenepentamine) was physically impregnated into two mesoporous solid support materials, MIL-101(Cr) and γ-Al₂O₃, to explore the effect of support materials on CO₂ adsorption mechanisms of a model-impregnated amine. These supports have different BET surface area to pore volume ratios and surface chemistry, and the DAC behavior of the amine-impregnated powder sorbents was investigated under a wide range of temperatures (−20 to 25 °C) and humidities (0–70% RH) conditions.

Analysis of 400 ppm CO₂ and H₂O adsorption/desorption data on comparably filled MIL-101(Cr) and γ-Al₂O₃ provided a means to infer if impregnated amines behave differently within different pore structures when capturing CO₂. The CO₂/H₂O TPD and in situ FT-IR results revealed that the impregnated TEPA within MIL-101(Cr) captures CO₂ primarily by forming carbamic acid (via weak chemisorption, 62.4 kJ/mol_CO2), while the TEPA within γ-Al₂O₃ forms ammonium carbamate ion pairs (via strong chemisorption, 100.5 kJ/mol_CO2) under 400 ppm CO₂ conditions. Since MIL-101(Cr) has a much higher BET surface area to the pore volume ratio of 1.78 nm^–1 vs γ-Al₂O₃ (0.16 nm^–1), impregnated amines may more strongly interact with MIL-101(Cr) than γ-Al₂O₃. Thus, the carbamic acid within the TEPA-impregnated MIL-101(Cr) adsorbents might be stabilized by surface water molecules or hydroxyl groups. In contrast, ammonium carbamate pairs were primarily formed in the TEPA-impregnated γ-Al₂O₃ via amine–amine interactions (as opposed to amine-surface interactions, like in the MIL-101(Cr) samples) during the CO₂ adsorption, probably due to the weaker interaction between impregnated amines and γ-Al₂O₃.

Interestingly, we found that the formation of carbamic acid in the MIL-101(Cr)_TEPA(30) is enhanced at cold temperatures, showing better DAC performance at sub-ambient (e.g., −20 and −5 °C) over ambient (e.g., 25 °C) conditions. Conversely, the formation of ammonium carbamate (a strongly exothermic reaction) on the Al₂O₃_TEPA(20) is not significantly affected by the adsorption temperature. We speculate that the different trends are because amine mobility dramatically decreases with decreasing temperature conditions. Although the formation of both carbamate and carbamic acid is affected by decreased amine mobility at cold temperatures, it appears that the formation of the ammonium carbamate ion pairs is more strongly impacted by the decrease in amine mobility at cold temperatures than the formation of carbamic acid, due to the higher amine to CO₂ reaction stoichiometry (2 vs 1). Thus, the thermodynamically favorable reaction at lower temperatures may be suppressed by the decreased amine mobility at cold temperatures (kinetically limited) in the alumina-supported case, where carbamate ion pairs are mainly formed.

Under humid conditions, the formation of ammonium carbamate ion pairs on the Al₂O₃_TEPA(20) and carbamic acid species on the MIL-101(Cr)_TEPA(30) was enhanced when compared to dry conditions. The enhancement in DAC performance under humid conditions was the most distinct at −20 °C, indicating that the effect of humidity on the DAC performance of the two adsorbents becomes more significant with decreasing adsorption temperatures. The difference between wet and dry CO₂ desorption energies also increased with decreasing adsorption temperatures from 25 to −20 °C. Additionally, it appears that the two adsorbent materials have different H₂O adsorption behaviors. The MIL-101(Cr)_TEPA(30) showed much higher H₂O uptake (28.8–35.2 mmol/g) than the Al₂O₃_TEPA(20) (8.3–13.5 mmol/g) under all temperature conditions, likely due to the much smaller pore size. However, it is expected that the effect of humidity on the 400 ppm CO₂ adsorption may not be significant, especially at sub-ambient temperature conditions (e.g., −20 °C) due to the kinetically limited H₂O uptake. The estimated H₂O adsorption working capacities of two sorbent materials were reduced by 3.3–3.5× via decreasing the adsorption temperature from 25 to −20 °C. These results suggest that significant energy savings associated with H₂O and CO₂ desorption energies could be achieved by operating a DAC process at sub-ambient temperatures (e.g., −20 °C).

There are several limitations to our study. First, although the analysis of different amine-impregnated porous support materials clearly showed that the support materials could affect the CO₂ adsorption mechanism of impregnated amines, more experiments with a variety of support materials will be required to develop an in-depth understanding of the full range of amine-surface effects. Second, while we observed a very small stretch for bicarbonate anions in the FT-IR spectra of the Al₂O₃_TEPA(20) upon humid 400 ppm CO₂ adsorption, it is not clear and definitive evidence for the formation of (or absence of) significant bicarbonate under humid conditions. As discussed in this article, observation of bicarbonate formation on solid-supported amines under humid CO₂ capture conditions is challenging with in situ FT-IR. Finally, this study evaluated powder materials, and any scalable, low pressure drop contactor design will need to incorporate such materials into alternate forms, such as fibers, monoliths, or other structures, which may result in deviations in amine-surface interactions from those found in powder studies.^104−107

While only two different porous support materials were investigated in this study, the results clearly demonstrate that the CO₂ capture mechanisms of impregnated amines could be controlled to vary from the formation of carbamic acid (weak chemisorption) to ammonium carbamate (strong chemisorption) by adjusting interactions between the amine and the surface of the support materials. The results also show that H₂O adsorption behavior is strongly affected by the properties of solid supports. Thus, to achieve optimized DAC performance depending on the range of sub-ambient and ambient temperature operations, proper support materials should be considered for amine impregnation.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.2c12707.

• Experimental section, characterization data (N₂ physisorption and pore size distribution), CO₂ adsorption behavior (dry and humid 400 ppm CO₂ breakthrough and CO₂/H₂O-TPD data) of TEPA-impregnated MIL-101(Cr) and γ-alumina powder sorbents, and in situ FT-IR spectra of TEPA-impregnated MIL-101(Cr) and γ-alumina powder sorbents during CO₂ adsorption (PDF)

The authors declare the following competing financial interest(s): C.W.J. has a financial interest in Global Thermostat LLC, which seeks to commercialize CO₂ capture from air. This work is not affiliated with Global Thermostat, LLC. C.W.J. has a conflict-of-interest management plan in place at Georgia Tech.