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

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 CO 2 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 CO 2 . 4−6 However, due to the high CO 2 binding energy of amines, amine-based adsorbent materials require higher energy for CO 2 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 CO 2 capture in the presence of moisture due to competitive adsorption between CO 2 and H 2 O, 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 CO 2 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 CO 2 capture performance. A variety of porous support materials have been used for amine impregnation, including silica, 8 47 and mixed metal oxides. 48,49 Despite this, the effects of amine-solid support interactions on CO 2 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 amineimpregnated MIL-101(Cr) reported detailed DAC behavior under sub-ambient temperature conditions (from −20 to 25°C ), including both dry 35 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 highperforming 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 Mg 2 (dobpdc) lead to unique CO 2 adsorption mechanisms 51−57 compared to more ill-defined systems based on MOFs and PEI. 58−61 We hypothesize that the CO 2 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 everchanging environment (outdoor air), 50 an in-depth understanding of the roles of support materials, temperature, humidity on CO 2 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-Al 2 O 3 . TEPA was used for amine impregnation because this chemical mainly consists of primary and secondary amines, which have a higher potential for CO 2 adsorption (e.g., heat of CO 2 adsorption = −80 to −110 kJ/ mol COd 2 ) compared to tertiary amines (−20 kJ/mol COd 2 ), 62,63 as well as shorter amine chains that likely result in enhanced CO 2 kinetics relative to the branched poly(ethylenimine) typically employed, especially under cold temperature conditions. Commercial γ-Al 2 O 3 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' CO 2 adsorption behavior under both dry and humid (70% RH) 400 ppm CO 2 conditions over a range of adsorption temperatures (−20 to 25°C) was probed by CO 2 /H 2 O 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 CO 2 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.
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 CO 2 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 N 2 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 CO 2 /N 2 (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-Al 2 O 3 . To investigate the effect of support material structure and composition on the CO 2 capture behavior of impregnated amines under a similar extent of pore filling (about 25%), MIL-101(Cr) and γ-Al 2 O 3 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 N 2 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-Al 2 O 3 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-Al 2 O 3 (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-Al 2 O 3 at similar loadings. Table 1 shows the extent of pore filling of sorbents determined based on the N 2 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.  Figure S3 shows the dry and humid (70% RH) CO 2 breakthrough curves of the powder sorbents under different temperature conditions. Pseudo-equilibrium 400 ppm CO 2 adsorption capacities of MIL-101(Cr) _TEPA (30) and Al 2 O 3 _TEPA (20) were determined based on the breakthrough curves, and the results are shown in Figure 1a,b, respectively, with amine efficiencies.
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 CO 2 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 Al 2 O 3 _TEPA (20). These different CO 2 capture behaviors of the MIL-101(Cr)_TEPA (30) and Al 2 O 3 _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 CO 2 adsorption mechanism(s) of the impregnated amine (TEPA), CO 2 /H 2 O temperature-programed desorption (TPD) and in situ FT-IR experiments were conducted.
Representative CO 2 /H 2 O TPD experimental results are shown in Figure S4. The desorbed CO 2 and H 2 O during the TPD experiments were quantified as a function of time ( Figure  S5). The total amounts of desorbed H 2 O from the MIL-101(Cr)_TEPA (30) and Al 2 O 3 _TEPA(20) materials after humid (70% RH) 400 ppm CO 2 breakthroughs are plotted in Figure 2a as a function of adsorption temperature. Since the humid CO 2 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 H 2 O uptake by MIL-101(Cr) _TEPA(30) did not consistently change with adsorption temperature conditions. While the H 2 O 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 Al 2 O 3 _TEPA (20), the H 2 O sorption continuously decreased with decreasing adsorption temperature, from 25°C (13.5 mmol/g) to −20°C (8.3 mmol/g).
Interestingly, the MIL-101(Cr)_TEPA(30) showed significantly higher H 2 O uptake than the Al 2 O 3 _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-Al 2 O 3 (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-Al 2 O 3 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 CO 2 adsorption results shown in Figure 1a,b, one can see that the H 2 O uptake is more significant during the humid DAC process at all temperature conditions. Thus, assuming similar ΔH ads(Hd 2 0) for these materials, more energy will be required for the regeneration of MIL-101(Cr)_TEPA(30) than for Al 2 O 3 _TEPA(20) for both ambient and cold humid DAC processes due to the significant H 2 O uptake. However, it should be noted that the H 2 O 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 CO 2 desorption below 25°C and strong chemisorption is defined as CO 2 desorption above 25°C, as discussed in Supporting Information ( Figure S8). The amounts of weak and strong chemisorption were quantified based on Figure 44 Inside a larger pore (16.1 nm) support with lower pore volume (1.03 cc/g), as in γ-Al 2 O 3 , strong chemisorption was the dominant CO 2 capture mechanism of TEPA. Both weak and strong chemisorption was enhanced under humid conditions. Effect of Support Materials on the CO 2 Adsorption Mechanism(s) of Impregnated Amines (TEPA) under Dry Conditions. To identify the adsorbed CO 2 species associated with weak and strong chemisorption, in situ FT-IR experiments were conducted at 25 and −20°C under dry 400 ppm CO 2 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 CO 2 to observe contributions more clearly from CO 2 adsorption. The relevant peak assignments are summarized in Table S1 with supporting references.
Results shown in Figure 3a,c indicate that the dominant adsorbed CO 2 species on the Al 2 O 3 _TEPA(20) and MIL-101(Cr)_TEPA(30) at 25°C are different. The TEPAimpregnated MIL-101(Cr) appears to dominantly form carbamic acid with a small amount of ammonium carbamate ion pairs, while the TEPA-impregnated γ-Al 2 O 3 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 Figure 3b,d, respectively. It appears that the same adsorbed CO 2 species formed at 25°C were also formed at −20°C on each sorbent, while there are differences in the CO 2 adsorption kinetics and the total amount of adsorbed CO 2 species at equilibrium under the different temperature conditions. Obvious increases in the intensity of adsorbed CO 2 in both adsorbents were observed as a function of time at 25°C, whereas the intensities relatively slowly increased at −20°C, indicating reduced CO 2 adsorption kinetics at cold temperature conditions, as also observed in breakthrough experiments described above. The overall intensity of adsorbed CO 2 on Al 2 O 3 _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 CO 2 adsorption kinetics, indicating the enhanced formation of carbamic acid. These in situ FT-IR results are consistent with the CO 2 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 CO 2 peak at 2335 cm −1 was observed on the MIL-101(Cr)_TEPA (30) at −20°C, while the Al 2 O 3 _TEPA(20) did not show any distinct CO 2 physisorption peak, as depicted in Figure S9. Thus, it can be inferred that the higher pseudo-equilibrium 400 ppm CO 2 uptake of MIL-101(Cr)_TEPA(30) vs Al 2 O 3 _TEPA(20) with lower amine efficiency at −20°C ( Figure 1) is due to a relatively larger amount of CO 2 physisorption over MIL-101(Cr)_TEPA (30). We infer that   (20) shown in Figure 4a clearly show the difference in CO 2 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 Al 2 O 3 _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).
Interestingly, the desorption peak position and the amount of desorbed CO 2 (peak area) of the Al 2 O 3 _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 CO 2 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 CO 2 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 Al 2 O 3 _TEPA (20) show different DAC behavior at ambient and sub-ambient temperature conditions (Figure 1). The CO 2 TPD results shown in Figure 4a indicate that the interaction strength between CO 2 and the impregnated TEPA is different depending on the support materials (MIL-101(Cr) and γ-Al 2 O 3 ) and adsorption temperature conditions. To quantify the binding energy of dry CO 2 onto the adsorbent materials, CO 2 TPD experiments were performed with varied heating rates and the energy of CO 2 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 CO 2 adsorption is an indicator of the strength of the interaction between the CO 2 molecules and amine sites on sorbents, and it is released as heat during the exothermic CO 2 adsorption process. The total energy of CO 2 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 Al 2 O 3 _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 CO 2 was desorbed during the N 2 purging period at 25°C ( Figure  S5e). The heat of dry 400 ppm CO 2 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 COd 2 to 62.4 kJ/mol COd 2 ) 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 CO 2 desorption peak of the Al 2 O 3 _TEPA (20) was not shifted under different adsorption temperature conditions (Figure 4a), resulting in a similar E d (99.9−101.5 kJ/mol COd 2 ), consistent with strong chemisorption (or formation of ammonium carbamate pairs) at all adsorption temperature conditions, assuming ideal mixing (i.e., E H d s , the sorption enthalpy). These results clearly show that the impregnated TEPA has different CO 2 capture mechanisms (carbamic acid vs carbamate) within different support materials, and the ammonium carbamate pair (100.5 kJ/mol COd 2 ) has stronger binding energy than carbamic acid (62.4 kJ/mol COd 2 ). 74,77,78 The regeneration of MIL-101(Cr)_TEPA(30) and Al 2 O 3 _TEPA(20) saturated with dry 400 ppm CO 2 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. CO 2 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 CO 2 by forming both carbamic acid and ammonium carbamate pairs, while the high amine-loaded silica primarily formed ammonium carbamate ion pairs during CO 2 adsorption ( Figure S13a). Another work by  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 CO 2 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 (−NH 2 ) and silanols (−OH) on the surface of SBA-15 but not for interaction with other amines, 82,83 the MONO grafted adsorbents captured CO 2 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 CO 2 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 CO 2 when amine−amine interactions are strong via amine clustering.
In the same manner, one possible cause for the different CO 2 adsorption mechanisms of impregnated TEPA within the two support materials studied here, MIL-101(Cr) and γ-Al 2 O 3 , 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-Al 2 O 3 (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-Al 2 O 3 . This is consistent with the observation that the dominant CO 2 adsorbed species on MIL-101(Cr)_TEPA(30) adsorbents were carbamic acids, while the Al 2 O 3 _TEPA(20) primarily formed ammonium carbamate pairs during dry 400 ppm CO 2 adsorption, as shown in Figure 3.
The The Al 2 O 3 _TEPA(20) may also contain hydroxyl groups on the surface of alumina because it was confirmed that traces of water are generated from γ-Al 2 O 3 during heating, even after activation at 1000°C. 87 Hydroxylation of the alumina surface occurs by the dissociative adsorption of water to Al 3+ , where Lewis acid sites and residing protons on lattice oxygens form surface hydroxyl groups. 88 89 which is comparable to the theoretical maximum of MIL-101(Cr). Although γ-Al 2 O 3 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-Al 2 O 3 (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 CO 2 adsorption by Al 2 O 3 _TEPA(20), probably due to the lower degree of interaction between impregnated TEPA and γ-Al 2 O 3 ( Figure S15c).
The amine-impregnated MIL-101(Cr) powder sorbents also adsorbed 400 ppm CO 2 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 CO 2 TPD 44 and in situ FT-IR studies revealed that both strong (ammonium carbamate) and weak chemisorption (carbamic acid) are involved in the CO 2 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 CO 2 Journal of the American Chemical Society pubs.acs.org/JACS Article 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 CO 2 Adsorption Behavior of Amine (TEPA) within Different Support Materials (MIL-101(Cr) and γ-Al 2 O 3 ).
The dry/humid 400 ppm CO 2 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-Al 2 O 3 powder sorbents at both ambient (25°C) and subambient (−5 and −20°C) temperatures. Similar results were also reported in our previous work with the amineimpregnated MIL-101(Cr). 44 To gain in-depth understanding of the effects of humidity, the CO 2 /H 2 O TPD profiles after humid (70% RH) CO 2 adsorption are compared with the results of dry CO 2 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 CO 2 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 CO 2 into the DRIFTS cell, a dramatic peak increase was observed at 3750 cm −1 , which is attributed to H 2 O adsorption. 74,90,91 Much faster H 2 O 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 CO 2 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 CO 2 /H 2 O 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 Al 2 O 3 _TEPA(20), respectively, was enhanced upon humid 400 ppm CO 2 adsorption. The enhancement in CO 2 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 waterstabilized 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 CO 2 adsorption with the Al 2 O 3 _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 CO 2 capture mechanism of Al 2 O 3 _TEPA(20), the formation of bicarbonate within Al 2 O 3 _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 CO 2 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 13 C solid-state NMR in the future. 96 Figure 5a,b, respectively, for different humid CO 2 adsorption temperature conditions (−20, −5, and 25°C). The amount of desorbed CO 2 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 CO 2 adsorption became more significant. The binding energy of CO 2 on the powder sorbents may be affected by H 2 O adsorption. Thus, quantifying the binding energy of CO 2 and H 2 O 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 CO 2 under humid conditions (co-adsorption of CO 2 and H 2 O) is challenging. Thus, the energies of CO 2 and H 2 O desorption, E d , for humid CO 2 adsorption, were determined based on the TPD method described in Supporting Information. Figures S21 and S22 show the  (Figure 4c,d) as a function of adsorption temperature conditions. It should be noted that T m s for some conditions could not be identified in the TPD profiles due to significant CO 2 or H 2 O desorption during the N 2 purging step, and thus, we were not able to determine E d for those conditions. The determined desorption energy of water for Al 2 O 3 _TEPA(20) was 46.4 kJ/mol Hd 2 O , which is comparable to the heat of vaporization of water in the range of 0−40°C and 43.4−45.1 kJ/mol Hd 2 O , supporting the condensation-like sorption of water by Al 2 O 3 _TEPA (20). A previous study directly measured the heat of adsorption of γ-Al 2 O 3 via microcalorimetry of water adsorption, and it was −44 kJ/mol Hd 2 O 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 Hd 2 O ) was slightly higher than that of Al 2 O 3 _TEPA (20). The reported isosteric heat of water adsorption on MIL-101(Cr) is from −53 to −44 kJ/mol Hd 2 O . 70 The higher heat of water desorption on the MIL-101(Cr)_TEPA(30) than Al 2 O 3 _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 H 2 O uptake than Al 2 O 3 _TEPA(20) under all adsorption temperature conditions (Figure 2a).
The calculated energy required for desorption of dry and humid CO 2 , shown in Figure 5c,d, indicates that the binding energy of CO 2 increases under humid conditions, and the effect of moisture on CO 2 adsorption is more significant with decreasing adsorption temperature. The difference between wet and dry CO 2 desorption energies increased with decreasing adsorption temperatures from 25 to −20°C. At −20°C, the desorption energy of CO 2 for Al 2 O 3 _TEPA(20) increased from 100.5 kJ/mol COd 2 (dry) to 121.1 kJ/mol COd 2 under humid conditions. The MIL-101(Cr)_TEPA(30) also showed similar behavior. It can be inferred from the CO 2 TPD profiles in Figure S17e,f that the increased CO 2 desorption energy for humid CO 2 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 Mg 2 (dobpdc) MOFs also show similar trends. While the DFTcalculated heat of CO 2 adsorption was −70 kJ/mol for the formation of ammonium carbamate chains under dry conditions, the calculated humid CO 2 adsorption energy was −88 kJ/mol in the case of the co-adsorption of 1 CO 2 and 1 H 2 O per diamine, indicating that H 2 O enhances the CO 2 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 CO 2 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 H 2 O uptake kinetics when compared to CO 2 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 H 2 O adsorption working capacity could be very low, depending on the adsorption time and the effect of adsorbed H 2 O on the 400 ppm CO 2 adsorption may not be significant due to the kinetically limited H 2 O uptake at subambient temperature conditions (e.g., −20°C).
To support the abovementioned argument, H 2 O uptake profiles of the MIL-101(Cr)_TEPA (30) and Al 2 O 3 _TEPA- (20) were obtained by integrating the H 2 O breakthrough curves in Figure S7 and are plotted in Figure S23 as a function of time. The dashed lines in the figures depict H 2 O adsorption working capacity at the time required to reach pseudo-equilibrium states (95% of breakthrough) of humid 400 ppm CO 2 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 H 2 O adsorption working capacity could dramatically decrease with decreasing temperatures due to significant kinetic limitation of H 2 O uptake at colder temperatures. The H 2 O adsorption working capacities of MIL-101(Cr)_TEPA (30) and Al 2 O 3 _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 H 2 O 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 CO 2 /H 2 O uptakes and total energies for CO 2 and H 2 O 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 CO 2 and H 2 O desorption were calculated based on the measured CO 2 and H 2 O uptakes ( Figure 2) and the E d for CO 2 /H 2 O 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 subambient temperature DAC operation (e.g., −20°C) due to their higher CO 2 uptakes and lower regeneration energies than Al 2 O 3 _TEPA (20). For dry ambient temperature operations (e.g., 25°C), the Al 2 O 3 _TEPA(20) appears to be an advantaged sorbent because of its higher CO 2 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 H 2 O 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 H 2 O uptake. However, as discussed above, the sorbent materials are never fully saturated with H 2 O, especially at sub-ambient temperature conditions (e.g., −20°C) due to kinetically limited H 2 O adsorption. When the estimated H 2 O 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.
For DAC operations, Table 2 indicates that the Al 2 O 3 _TEPA(20) will be better sorbents than MIL-101(Cr) _TEPA(30) due to its lower H 2 O 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 CO 2 /H 2 O 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 CO 2 working capacity, 1.55 mmol/g, than Al 2 O 3 _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 CO 2 capture mechanisms of amineimpregnated sorbent materials, varying from the formation of carbamic acid (weak chemisorption) to ammonium carbamate (strong chemisorption). DAC performances, including CO 2 uptake and energy requirement for sorbent regeneration, of these two different mechanisms were strongly affected by adsorption temperatures and humidity. These conditiondependent 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-Al 2 O 3 , to explore the effect of support materials on CO 2 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 amineimpregnated powder sorbents was investigated under a wide ) under 400 ppm CO 2 conditions. Since MIL-101(Cr) has a much higher BET surface area to the pore volume ratio of 1.78 nm −1 vs γ-Al 2 O 3 (0.16 nm −1 ), impregnated amines may more strongly interact with MIL-101(Cr) than γ-Al 2 O 3 . 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-Al 2 O 3 via amine−amine interactions (as opposed to amine-surface interactions, like in the MIL-101(Cr) samples) during the CO 2 adsorption, probably due to the weaker interaction between impregnated amines and γ- 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 Al 2 O 3 _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 CO 2 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 Al 2 O 3 _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 CO 2 desorption energies also increased with decreasing adsorption temperatures from 25 to −20°C.
Additionally, it appears that the two adsorbent materials have different H 2 O adsorption behaviors. The MIL-101(Cr) _TEPA (30) showed much higher H 2 O uptake (28.8−35.2 mmol/g) than the Al 2 O 3 _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 CO 2 adsorption may not be significant, especially at sub-ambient temperature conditions (e.g., −20°C) due to the kinetically limited H 2 O uptake. The estimated H 2 O 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 H 2 O and CO 2 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 CO 2 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 Al 2 O 3 _TEPA(20) upon humid 400 ppm CO 2 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 CO 2 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 aminesurface 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 CO 2 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 H 2 O 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. ■ ASSOCIATED CONTENT