Additives enhancing supported amines performance in CO2 capture from air

The utilization of supported amines as adsorbents in direct air capture (DAC) has been demonstrated to be a promising strategy for the reduction of CO2 emissions. To improve the performance of amine‐based adsorbents, the incorporation of additives has been widely adopted. In the present study, we conduct a comprehensive comparison of seven additives on tetraethylenepentamine‐impregnated mesoporous silica as a representative amine‐based adsorbent. The results indicate that minor molecular weight additives with hydroxyl groups show improved adsorption–desorption performance and increase oxidative stability. A proposed mechanism for these improvements is the combined physical and chemical promotion effects of hydroxyl groups. Through a comprehensive review of existing literature, it is found that the effects of additives on amine‐based adsorbents are dependent on factors, such as additive type, pristine adsorbent properties, incorporation method, and testing conditions. Based on these findings, it is recommended that future DAC systems prioritize the use of hydroxyl‐containing additives, whereas higher CO2 concentration and temperature capture may benefit from the incorporation of additives without hydroxyl groups. These conclusions are expected to contribute to the design of efficient adsorbents for CO2 capture.


INTRODUCTION
The excessive emission of greenhouse gases, including anthropogenic CO 2 , has led to the worsening of global warming and raised widespread public concern. 1 Despite a slight decrease in emissions since 2018, when it reached an all-time high of 37.1 billion, it still poses a serious threat to the environment and society. 2 Consequently, it is imperative to focus on developing and deploying efficient CO 2 capture technologies. 3 Direct air capture (DAC), initially proposed by Lackner et al., 4 can mitigate the effects of climate change and reduce CO 2 emissions from distributed sources. 5 The use of adsorbents plays a crucial role in DAC and has sparked a surge of new ideas for developing new materials in the past 5 years. 6,7 Amine-functionalized porous materials, including zeolites, 8,9 activated carbon, 10,11 mesoporous silica, 12,13 metal-organic frameworks, 14,15 and silica sol-gel, 16,17 are widely studied due to their high affinity toward CO 2 , [18][19][20][21] low energy cost, and robust stability even in the presence of moisture. 22 They effectively address the deficiencies of amine solution absorption and interact with CO 2 in a chemisorptive manner, [23][24][25][26][27] resulting in relatively high adsorption heat and steep adsorption isotherms, 28 which have been proved to the benefit of significant CO 2 capacity under air conditions. 20,[29][30][31] Competitive DAC adsorbents should possess (i) high capacity and amine efficiency, (ii) fast adsorption rate and low cost, (iii) good selectivity and recyclability, and (iv) low energy requirement for the regeneration process. An adsorbent with all the merits mentioned above is beneficial to reduce cost and enhance potential for commercialization; however, it is tough to design an adsorbent meeting all requirements. Therefore, seeking novel strategies to improve adsorbent performance is urgent.
Considerable effort has been devoted to promoting the CO 2 adsorption of amine-functionalized materials through various approaches, including the use of different amine sources, 32 adjustments to the structure or porosity of supports, 33 and the exploration of new adsorption routes. 34 These strategies are effective by immobilizing numerous amine sites, reducing steric hindrance, and altering amine speciation. In addition, the incorporation of additives as a novel strategy has been extensively studied to expedite the development of materials for carbon capture. According to the various preparation method of additives' incorporation, the review of research is categorized as co-impregnation, two-step functionalization, and pre-synthesis.
Most of the research focuses on the co-impregnation of additives with amines, 18,27,[35][36][37][38] by which the diffusion limitation of CO 2 can be alleviated due to the inter-actions between the amines and the additives. Jones's group conducted a study wherein various additives were impregnated with polyethyleneimine (PEI) into SBA-15. 27 The results of the study showed that among the additives tested, poly(ethylene glycol), with an average mole weight of 200 (PEG200), had a significantly greater impact on amine efficiency by 60% under dry ambient air. This increase was observed to be higher compared to cetyltrimethylammonium bromide (CTAB), which showed a decrease of 12.7%, and poly(ethylene glycol) (PEG1000), which showed an increase of 12.7%. The findings of this study suggest that PEG200 may be a promising additive for enhancing amine efficiency in SBA-15. Wang et al. incorporated a variety of CO 2 -neutral surfactants into PEI-impregnated hierarchical porous silica monoliths and indicated that all the surfactants enhanced the capacity and amine efficiency in adsorbing pure CO 2 , irrespective of the temperature. Notably, the inclusion of sorbitan monooleate (Span80) resulted in a capacity enhancement of up to 26%. 37 Additionally, they observed similar outcomes when they introduced Span80, CTAB, PEG, and sodium dodecyl sulfate (SDS) into commercial resins and silica-templated mesoporous carbons. 36,[39][40][41] Two-step functionalization, including impregnation followed by grafting (or vice versa), is also used to incorporate additives with mobile amino groups and to increase nitrogen content. [42][43][44] Sanz et al. grafted 3aminopropyltriethoxysilane (APTES) into silica, followed by impregnation of tetraethylenepentamine (TEPA) or PEI and reported that this procedure provided 31% of enhancement on adsorption capacity for TEPA and 38% for PEI under pure CO 2 compared to conventional onestep functionalization procedure. 42 Guo et al. studied a hybrid 3-aminopropyltrimethoxysilane (APTMS) and TEPA-modified MCM-41 sorbent synthesized by a two-step method. The addition of APTMS obtained 62% of increased adsorption capacity from flue gas (CO 2 /N 2 (v/v) = 15/85) and exhibited excellent stability. 44 Pre-synthesis by retaining the templating agent is another feasible way to incorporate additives, which is usually achieved by avoiding calcination at the end of substrate production. 45 and amine, thereby improving the adsorptive capacity of this pure CO 2 adsorption system. 48 As of today, the vast majority of research on additives has been carried out under simulated flue gas or pure CO 2 conditions, with little emphasis on DAC. Given the similar promotion mechanism, it is probable that additives will function effectively in ultra-dilute conditions. 7 Moreover, the use of different amine-based materials, testing conditions, and crucial synthesis parameters has made it difficult to draw conclusions about the impact of additives. Therefore, it is crucial to evaluate a broad range of additives using a set of scientifically rigorous and comprehensive material characterization methods.
Our previous work proposed a comprehensive and standardized test method for evaluating DAC adsorbents. 50 In this study, we examined the performance of seven additives, namely, CTAB, diethanolamine (DEA), P123, phosphatidylcholine (PC), PEG200, SDS, and Span80, which were selected based on previous research. 27,37,48,49,51 The additives were incorporated with TEPA-impregnated SBA-15, and the resulting adsorbents were evaluated under DAC conditions in terms of their intrinsic properties, adsorption capacity and efficiency, kinetics, desorption temperature, thermal stability, and the effect of oxygen and CO 2 concentration. We found that only additives with low molecular weight and abundant hydroxyl exhibited noticeable improvement, which was inconsistent with results from experiments under flue gas or pure CO 2 . Thus, we summarized the performance of additive-modified adsorbents under different CO 2 concentrations and provided recommendations for selecting suitable additives. We also proposed a new promotion mechanism involving the hydroxyl group to explain these experimental results.

Characterization
The molecular formulas of additives are shown in Figure 1. SDS, PEG, PC, Span80, and P123 without nitrogen elements cannot chemisorb CO 2 . A secondary amine in DEA molecule is expected to provide more active reaction sites. Nitrogen at CTAB cannot accept the nucleophilic attack of CO 2 due to its intrinsically positive charge. Although hydroxyl groups in DEA, Span80, PEG, and P123 hardly react with CO 2 , they are widely reported to benefit adsorption. 37,48,52,53 Typically, Yue reported that the hydroxyl group of P123 can facilitate the capture of CO 2 by modifying the interaction mechanism between amine and CO 2 . 48 In another study, they proposed that the presence of the hydroxyl group in DEA enhances the adsorption capacity and amine efficiency by promoting the formation of carbamate-type zwitterions with lower thermal stability through a novel mechanism. 52 Chuang et al. reported that the formation of hydrogen bonds between NH 2 groups of TEPA and OH groups of PEG increases the dispersion of TEPA, thereby improving the contact between CO 2 and amine. 53 Wang et al. found that the improved performance of Span80-promoted sorbents may be influenced by several factors, including the structure of the additive and its selfassembled structure in PEI, the interaction between the additive and PEI, and the CO 2 affinity of the additive. 37 PC and SDS are minor molecular weight additives and are likely to reduce viscosity. The additive-modified adsorbents were denoted as the additive name. For example, PEG200 means the sample synthesized with 10 wt.% of PEG200, 40 wt.% of TEPA, and 50 wt.% of SBA-15. The additive-free adsorbent with 40 wt.% of TEPA and 60 wt.% of SBA-15 was denoted as TEPA. The loading amounts of organic components were verified using TGA, as demonstrated in Figure 2A. The findings indicated that samples without additives and those with additives lost roughly 40 and 50 wt.% of organic components, respectively, which agreed with the expected percentage. Figure 2B displays the differential thermal analysis of the adsorbents utilized to evaluate their thermal stability. The pure TEPA-functionalized adsorbent displayed only one weight loss peak at 231.7 • C. When additives were introduced, the peaks shifted to higher temperatures, and new weight loss peaks appeared, indicating that the thermal stability of the additive-added adsorbents had improved. It is worth noting that the PEG200-modified sample displayed a peak around 570 • C, which was attributed to the hydrogen bond between the amine and hydroxyl groups. Figure 2C is the FT-IR spectrum of pure TEPA-, pure PEG200-, and PEG200-modified adsorbents. The appearance of bands at 1317 and 940/877 cm −1 on PEG200modified adsorbent, which was associated with CH 2 bending and CH 2 rocking motions of the ethylene group, [54][55][56] indicates that PEG200 has been incorporated into the adsorbent. 27 Besides, bands at 1560 and 1470 cm −1 were assigned to the primary and secondary amines, respectively, which proves the successful loading of amines. [57][58][59] The bands for other additives shown in Figure S1 also demonstrate a similar phenomenon.
The pore structure of adsorbents was characterized by N 2 adsorption at 77 K, as shown in Figure S2. All the adsorbents showed a type IV adsorption−desorption isotherm with an H2 hysteresis loop, indicating a typical mesoporous structure of the materials. The BET specific surface area and pore volume of the samples are summarized in Figure 2D. SBA-15 exhibited a surface area of 991.9 m 2 g −1 and a pore volume of 1.076 cm3 g −1 , consistent with value in literature. 60,61 The TEPA-functionalized adsorbent displayed the highest surface area and pore volume due to the lowest organic component loading (40 wt.%). The P123added sample exhibited one of the lowest values, likely due to the large molecular weight of the additive (∼5800). Figure 2E displays the nitrogen content of the samples. The adsorbent modified with DEA had the highest weight percentage of nitrogen (14.54 wt.%) due to the additional secondary amine introduced by the DEA molecule. The promoting nature of hydroxyl-containing substances was evaluated by the OH/amine ratio. 62 Despite PEG200, DEA, and P123 molecules all having two hydroxyl groups, their higher molecular weight resulted in a lower OH/amine ratio. For more specific data, please refer to Table S2 for comparison. F I G U R E 3 CO 2 capacity (black bar) and amine efficiency (green bar) of adsorbents loaded with 40 wt.% tetraethylenepentamine (TEPA) and 40 wt.% TEPA coupled with 10 wt.% of different additives at 25 • C and 400 ppm CO 2 /N 2 mixture.

CO 2 capacity and amine efficiency
Thermogravimetric and volumetric methods measured the CO 2 adsorption capacity and amine efficiency of adsorbents at 25 • C, and the results are summarized in Figure 3 and Table S3. It was found that not all additives improved the CO 2 sorption capacity and amine efficiency, which was different from the results of other studies. 37,48,49 CTAB mildly reduced the adsorption capacity in the test but showed promotion under pure CO 2 at 75 • C. 63 This may be due to the diffusion force being insufficient under low CO 2 concentration and temperature to penetrate the solid-like interfacial layers formed between CTAB and SiO 2 . 64 The same phenomenon occurred with SDS. 65 Although both values for PC were lower compared to TEPA, the reason for the negative effect was still unclear. DEA obtained 39.9% of enhancement of adsorption capacity and 27.4% of amine efficiency. The promotion was driven by a combination of more amine sites and the existence of hydroxyl groups. 51 PEG had the largest amine efficiency of 0.22 CO 2 /N and the second-largest adsorption capacity of 2.05 mmol g −1 . Considering the lower pore volume and N weight percentage than TEPA, the reason for such superiority was accredited to the effect of the hydroxyl group. Span80, despite also containing hydroxyl, showed an adverse effect, possibly due to its high viscosity (1000-2000 mPa) and larger molecule, which resulted in pore blockage. 66 Likewise, the hydroxyl-containing compound Pluronic P123 exhibited a minimal change due to an offset between the positive effect of the hydroxyl group and the negative effect of the pore structure block. A different perspective has been proposed for the promotion mechanism of hydroxyl-containing additives. Goeppert et al. suggested that the addition of hydroxyl seemed to lower the viscosity of amine, making it more "fluid" and possibly allowing easier access for CO 2 to the active absorption sites. 67 Jones et al. reported that hydroxyl could disperse bulk PEI into smaller agglomerates, resulting in increased amine efficiencies without drastically changing dynamics. Hydroxyl group also promoted intrachain CO 2 adsorption events by protecting neighboring PEI chains against one another, reducing the diffusion barrier, and releasing more active amine sites. 27 With in situ ATR, Chuang et al. reported that the existence of hydroxyl broke up the hydrogen bonding network between amines through the interaction between -OH and primary and secondary amines. 68 Additionally, Baker et al. hypothesized that the positive effect was due to the increased Lewis basicity, but this assumption has yet to be supported by experiments. 69 Considering the hydrogen bond forming between the hydroxyl group and carbonyl group, 70 we proposed that physical and chemical effects coexisted in hydroxylcontaining adsorbents, favoring the CO 2 capture process. As illustrated by Figure 4A/B, in terms of physical effect, the introduction of the hydroxyl group would interact with TEPA's NH + 3 and NH + 2 active sites, leading to the disruption of the hydrogen bonding network between adjacent TEPA groups. 68 The hydrogen bonding between molecules and polymer chains plays a critical role in determining viscosity. 71 As a result, the disruption of the hydrogen bonding network reduced the viscosity of TEPA and facilitated the diffusion of CO 2 through smoothing the diffusion path of CO 2 . Moreover, the breakdown of the hydrogen bonding network between adjacent TEPA groups led to the liberation of more -NH 3 and -NH 2 active sites, thereby increasing the probability of CO 2 adsorption reactions. This assumption is supported by the FT-IR results shown in Figure 2C, where the emergence of the OH⋅⋅⋅NH band at 3321 cm −1 indicates the formation of hydrogen bonding between TEPA and hydroxyl groups of additives. 68 Chemical effects indicate the hydroxyl groups have an F I G U R E 4 Adsorbents modified without (a/c) hydroxyl group and with (b/d) hydroxyl group (a/b) before and (c/d) after exposure to CO 2 . Blue, red, black, and white spheres represent N, O, C, and H atoms, respectively. In situ DRIFTS for (e) PEG200 under dry 400 ppm CO 2 /N 2 . effect on the product of CO 2 adsorption as indicated by Figure 4C/D. 35,72,73 Without hydroxyl groups, carbamate formation was favored due to the reaction of two amine groups with one CO 2 molecule. 74 However, in the presence of hydroxyl groups anchored to the surface of silica or randomly dispersed in the pores, 75 the formation of carbamic acid/zwitterions was stabilized, resulting in the reaction of one amine group with one CO 2 molecule. 35 This supposition is supported by the in situ DRIFTS spectra of PEG200 exposed to 400 ppm CO 2 , as illustrated in Figure 4E. The spectra show an increase in IR characteristic peaks of 1644, 1556, 1411, and 1329 cm −1 , corresponding to NH + 3 deformation, COO − stretching, NH + 2 deformation, and NCOO − skeletal vibration, respectively, indicating the formation of ammonium carbamate upon exposure to air for CO 2 adsorption. Notably, the appearance of the C=O stretching at 1701 cm −1 is observed in the shoulder after exposure to CO 2 for 20 min, demonstrating the formation of hydrogen bond carbamic acid. 62,75,76 To further discuss the effect of additives on solid amine adsorbents, we have summarized previous studies conducted under different test conditions in Figure 5 and Table S1. The works were categorized into three groups based on CO 2 concentration. Most of the research was conducted under pure CO 2 followed by flue gas (∼15 vol.%), and limited studies were focused on DAC (400 ppm). Test temperatures also varied, ranging from 0 to 80 • C. Through our literature review, we have concluded that the effects of additives are dependent on their types, the properties of the original adsorbents, the methods of incorporation, and the test conditions. First, different additives demonstrated varying effects on the same adsorbents, which is supported by the results discussed above. Additionally, the original adsorbents played a significant role in the over-all performance of the adsorbent. For example, Wang et al. reported that under DAC conditions, Span80 increased the adsorption capacity of PEI-functionalized mesoporous carbon by 50%, while having an adverse effect on TEPA-loaded SBA-15 under similar conditions. 39 We believe that this difference can be attributed to the higher pore volume of mesoporous carbon, which allowed for the greater diffusion of CO 2 into the deeper PEI film, breaking it down into a dual interpenetrated composite. 39 However, the smaller pore volume of SBA-15 was hindered by Span80, blocking the diffusion path of CO 2 and inhibiting its contact with TEPA amine sites.
Furthermore, the method used to incorporate additives was also found to affect their performance. Yue et al. added P123 into TEPA-SBA-15 through pre-synthesis and impregnation, respectively, and reported that the pre-synthesis method showed better adsorption performance. 48 The main reason was that the pre-synthesis method allowed the PO blocks of P123 to disperse within the pores of SBA-15-like branches, whereas the EO blocks were rooted within the framework. However, P123 in the impregnation sample occupied the mesopores and occluded the complementary pores, resulting in agglomeration in the pores and blocking channels between the mesopores. 48 Finally, test conditions such as CO 2 concentration and temperature also played essential roles. Wang et al. incorporated three additives (Span80, CTAB, and PEG) into PEI-HP2MGL and tested them under 400 and 5000 ppm. They discovered that additives displayed a slight promotion effect at 400 ppm CO 2 , but a much more apparent enhancement at 5000 ppm. This might be attributed to the low diffusion force deriving from the low-concentration gradient at 400 ppm CO 2 , weakening the positive impact of the promoter. 36 The study notes that large amounts of F I G U R E 5 Capacity of additive-modified adsorbents in different test conditions. Specific information is listed in Table S1. additives benefited adsorption under relatively higher concentrations and temperatures, whereas additives showed no advantages or even detrimental effects under low CO 2 pressure and temperature. However, hydroxyl-containing additives may not follow such phenomena. The isotherm of PEG200 shown in Figure S3 suggested that the adsorption capacity of PEG200-modified adsorbents showed declining enhancement with CO 2 partial pressure increasing from 0.04 to 15 kPa and enhanced temperature. This discrepancy can be attributed to several factors, including insufficient amine sites and weakened thermodynamics. The incorporation of 10 wt.% of additives might cover some amine sites instead of promoting dispersion. Despite higher partial pressure bringing enhanced diffusion force of CO 2 , the existence of additives prevented amine sites from interacting with the adsorbate. Additionally, the presence of PEG200 may have promoted the formation of unstable carbamic acid with lower adsorption heat, resulting in a decrease in the maximum adsorption temperature due to its exothermic nature. These findings suggest that hydroxyl-containing additives may be more suitable for use in DAC rather than for capturing CO 2 from flue gas. Figure S4 reviews the amine efficiency of additivemodified adsorbents tested under various CO 2 conditions. It was found that values of amine efficiency increased with CO 2 concentration. Pure CO 2 showed the highest average, followed by flue gas and DAC condition. Interestingly, almost all the additives enhanced the amine efficiency compared to pristine samples under CO 2 concentration over 7 vol.% while limited promotion for trace CO 2 adsorption. Figure 6A and Table 1 illustrate the CO 2 uptake of the samples, along with their corresponding time profiles and the Avrami model fitting data. Normally, the CO 2 adsorption for amine-based adsorbents is divided into two stages, as indicated in Figure 6A. The first stage is the chemical reaction-controlled process where CO 2 quickly reacts with the affinity sites on the adsorbent surface. The rapid CO 2 adsorption rate peak at this stage shown in Figure 5B lasts for about 60 min and completes about 90% of the pseudo-equilibrium CO 2 adsorption capacity. This supported hypothesis that chemical sorption played a significant role in CO 2 capture capacity at ultra-dilute CO 2 concentration conditions. The second stage is a diffusioncontrolled process, where CO 2 needs to overcome the diffusion barrier resulting from the built-up and cation-anion interaction of the ammonium carbamate layer produced by the CO 2 adsorption reaction on the surface. The Avrami kinetic model was widely used to describe the CO 2 adsorption behaviors for amine-based sorbents. 77,78 The results of the study indicated a good fit between the model and all the samples, with a comparable rate coefficient k ranging from 0.02 to 0.04. It was observed that most of the additive-modified adsorbents exhibited a higher rate coefficient, which could be attributed to the improved amine dispersion facilitated by the additives.

Kinetics
It has been observed that DEA and PEG200 exhibit a lower k value than TEPA, despite having higher capacities. This suggests that the chemical reaction-controlled processes for DEA and PEG200 are slightly slower and longer, as confirmed by Figure 6B. The increase of 10 wt.% F I G U R E 6 Experimental capacity-time profile with the corresponding Avrami model fitting data (a) of additive-modified adsorbents and its adsorption rate (b).

TA B L E 1 Avrami kinetic model fitted parameters of additive-modified adsorbents.
Sample q e a (mmol of organic components has been proposed as the cause of reduced porosity. However, for substrates with wider pores and channels, this deficiency may be eliminated. 79 The adsorption half-time, which represents the time required for the adsorbent to reach 50% of its saturation capacity, was employed to quantitatively analyze kinetics. A longer adsorption time indicated a slower adsorption rate. Although the resistance distribution during the adsorption process was unclear, this index was obtained from the capacity and time profile, providing an intuitive reflection of the adsorption rate. 50 Figure 7 summarizes the literature on adsorption half-time, facilitating visualiza-tion of the range of adsorption properties of additives. The results showed that the adsorption half-time decreased as the CO 2 concentration increased, indicating that the CO 2 diffusion force was a primary factor affecting the adsorption rate. Under identical test conditions, only a limited number of additives promoted the adsorption rate of the pristine sample. This may be attributed to the fact that the addition of additives occupied the diffusion channels of CO 2 , while their existence appropriately dispersed amine sites.

Desorption behavior
The desorption temperature has a significant impact on both the energy consumption of the DAC operation system and the lifespan of the adsorbents, especially for a temperature concentration swing adsorption system, which is commonly used in adsorbent development. [80][81][82] Minimizing the desorption temperature can reduce thermal energy consumption by utilizing inexpensive and/or renewable heat sources. In addition, our previous work has shown that keeping the regeneration temperature of TEPA-modified adsorbents under 90 • C would protect the amine from leaching or cracking and minimize the effect of oxygen on it, 83 which would extend the lifespan of adsorbents and reduce the cost of DAC. 84 However, another work reported that TEPA-modified adsorbents could only obtain complete desorption at 90 • C or above. 12 To study the CO 2 desorption behavior of additivemodified adsorbents, they were subjected to regeneration by temperature swing adsorption with N 2 purge F I G U R E 7 Adsorption half-time of additive-modified adsorbents in different test conditions. Specific information is listed in Table S1. after adsorption saturation at 400 ppm CO 2 flow. Figure  S5 and Figure 8 show the desorption profile of each additive-added adsorbent ranging from 60 to 100 • C and normalized CO 2 desorption process degassing at 80 • C, respectively. The results indicated that 10% of additives showed a limited effect on desorption temperature, and all of them regenerated completely at 90 • C or above within 15 min. Especially, PEG200 showed slightly better desorption performance, regenerating 50% of adsorbed CO 2 within 3.55 min and 90% within 12.15 min. This mild enhancement allowed it to have a better working capacity and stronger endurance in the cyclic test shown in Figure S6. Due to incomplete desorption at 70 • C, the pure TEPA-functionalized sample dropped by 22.2% of its initial capacity. In contrast, the DEA-and PEG200-containing samples showed better retentions at 90% and 91%, respectively. However, other additive-containing adsorbents lost as high a capacity as TEPA. Furthermore, less loss was observed when the desorption temperature was increased to 80 • C for all materials, at 7.1% (TEPA), 2.2% (DEA), and 3.7% (PEG200), elucidating that 80 • C was the optimal desorption temperature for additives and TEPA-modified silica under DAC condition. The promotion could be interpreted as follows: With the aid of the hydroxyl group from PEG, the adsorption mechanism changed, resulting in the production of more carbamic acid during the reaction. According to the antecedent research, the adsorption heat of carbamic acid was lower than the carbamate, implying that the binding reaction between the functionalized polyamine sites and CO 2 molecular was weaker and that less energy consumption was required for regeneration. 35,62,68,85,86

Stability
Although the concentration of O 2 in flue gas is only about one fifth that of carbon dioxide O 2 , the O 2 concentration in air is over 500 times that of CO 2 . 83 Thus, it is imperative to explore the effects of oxygen on adsorbents. In addition to oxygen concentration, the temperature was another factor that affected oxidative degradation. 87 Based on our previous research, a temperature lower than 90 • C would minimize oxidative extent for the TEPA-functionalized adsorbent. 83 To explore the oxidative resistance of additivemodified adsorbents, we conducted accelerated oxidative treatments at 70 and 80 • C. The effect of oxidative treatment on the stabilities of TEPA-loaded SBA-15 added with varying additives is presented in Figure 9 and Table S4. Following a 12-h exposure to O 2 at 70 • C, the vast majority of adsorbents maintained their initial values, and a slight retention rate exceeding 100% resulted from measurement error, indicating negligible oxidative damage at this temperature. After exposure to O 2 at 80 • C, the sample without additives exhibited a 4% capacity loss, whereas adsorbents containing DEA and PEG200 exhibited lower values of 2.5% and 1.9%, respectively. The modest enhancement of oxidative resistance is expected to be magnified under practical conditions when endured over 10 times longer than oxidative treatment time in this work. Given the excellent thermal stability confirmed by Figure 2A, the observed capacity loss under consecutive exposure to oxygen was attributed to oxidative degradation, rather than any leaking or cracking of amine sites caused by thermal effects. The fact that oxidative treatment at 80 • C maintained a higher working capacity than at 70 • C confirms that 80 • C is the optimal highest working temperature.
The increased anti-oxidative competence might be ascribed to the agglomeration caused by the addition of additives that slowed the diffusion rate of O 2 , which was directly related to the oxidative degradation. 88 Besides, the interaction between -OH from PEG200 or DEA and amine sites also contributed to the anti-oxidative competence. Chuang et al. suggested that the hydrogen bonding due to the interaction prohibited oxidative degradation by blocking the amine sites from the approach of O 2 . 53,62 In another study, they reported that the hydrogen bonding between the -OH groups of poly(vinyl alcohol) and secondary amines of branched PEI was capable of protecting PEI from oxidative degradation. 89 To be economically feasible, CO 2 adsorbents should withstand thousands of adsorption-desorption cycles.82 Thus, the effect of additives on the cyclic stability of adsorbent was investigated, and the average loss per cycle from the literature review is summarized in Figure 10 for evaluation. Specifically, Zhang et al. converted primary amines into secondary amines using acrylamide to react with TEPA, resulting in only a 1% loss of initial capacity after five cycles under pure CO 2 , as opposed to a 48% loss for unmodified material. 90 Jones et al. incorporated glycerol into PAI-PEI-silica to enhance thermal stability. The resulting adsorbents remained ignorable loss after five cycles under flue gas condition (fixed bed, 10 vol.% CO 2 ), which was significant progress compared to pristine adsorbents that lost 14% of capacity within five cycles. 91 The same group also reported additive-promoted stability under ambient air condition (25 • C, 400 ppm), in which incorporation of APTES showed 7.1% loss and tetrapropyl orthotitanate remained stable after four cycles, much more potent than additive-free adsorbents with 32.2% loss. 18 However, additives generally worked out at the expense of CO 2 capacity, which should be considered when designing adsorbents in this manner.

F I G U R E 1 0
The cyclic stability of additive-modified adsorbents. Specific values can be found in Table S1.

Additive evaluation
Seven additives were incorporated with TEPA into SBA-15 through impregnation, and their CO 2 adsorption performance under DAC condition was tested according to a comprehensive and systematic measurement method. The selection of CTAB, PC, P123, SDS, and Span80 was based on their successful performance in previous literature. However, in this study, these additives displayed weak promotional effects on the pristine adsorbents, with lower saturation capacity, pore or diffusion path blockage, unaltered desorption temperature, and inferior working capacity under simulated air (consisting of 21% O 2 , 78% N 2 , and 400 ppm CO 2 ). The mediocre performance of these additives might be attributed to weaker CO 2 diffusion force due to low CO 2 concentrations and temperatures, coupled with factors, such as amines, substrates, and test methods. Span80 and PC have been reported in previous literature to break bulk PEI films or CO 2 sorption products into a dual interpenetrated composite, promoting the diffusion of CO 2 into the deeper PEI film under pure CO 2 and 75 • C. However, in this study, TEPA with low molecular mass inherently had better distribution, as confirmed by Table S2, which obscured the positive effect of the additives. Meanwhile, low concentration and temperature might deprive the incentive of CO 2 to access more amine sites, thus resulting in worse performance. CTAB was also proposed to promote the dispersion of amines and further facilitate the adsorption. 49 However, it is important to note that diffusion force should also be considered. SDS was suggested to disrupt the molecular packing of the amine, allowing a more significant number of amine sites to be accessible, 63 but it also worked out on the premise of enough CO 2 for molecule movements. In contrast, P123 is an additive that promotes adsorption through the hydroxyl group, which alters the chemical reaction. 48 It is worth noting that P123 was included in the substrate preparation process in the original study, whereas in this study, it was introduced via impregnation, which could potentially result in agglomeration within the pore channels of SBA-15.
PEG200 and DEA displayed noticeable advancement in the test condition of this work. The physical and chemical impacts of the hydroxyl group were suggested to contribute to this promotion. Additionally, the greater presence of secondary amine in DEA facilitated the provision of active amine sites for chemisorption, which resulted in enhanced adsorption capacity, lowered desorption temperature, and increased stability upon long-term oxygen exposure. In general, low-molecular-mass additives containing hydroxyl groups exhibit good compatibility with TEPA-based adsorbents for DAC, and their efficacy remains unimpaired even in the presence of water, as evidenced by Figure S7

2.7
Strategy for selecting proper additives Figure 5 summarizes a large amount of additivefunctionalized cases, and readers can refer to similar pristine materials and CO 2 concentrations and accordingly select proper additives for their materials. Given that the performance of additives is influenced by a range of factors, determining the optimal additive for all situations can be challenging. However, we have compiled a list of the top-performing additives for each CO 2 concentration category, which readers can consult as a helpful reference.
Qi et al. reported that the adsorbent (MMSN-30/TEPA-50) with the appropriate amount of brush-like cetyltrimethylammonium chloride (CTAC) exhibited higher CO 2 adsorption capacities (3.68 mmol g −1 under simulated dry air conditions) than that of the adsorbents with none or fewer additives, 45 which is the most significant enhancement in low CO 2 concentration range. Further, the researchers added bis-(γ-triethoxysilylpropyl)tetrasulfide to the CTAC-modified adsorbents, which led to an even greater increase in adsorption values under flue gas conditions (75 • C, 15 vol.%). 46 In another study, PEG400-modified PEI-fumed silica showed the most substantial enhancement, reaching 71.8% in the pure CO 2 case. This improvement might be attributed to the beneficial effect of the hydroxyl group. 67 In terms of amine efficiency, ionized liquid 1-ethyl-3-methylimidazoliumacetate ([emim][Ac]) increased the amine efficiency of PEI-SBA-15 from 0.1 to 0.37 CO 2 /N (30 • C, pure CO 2 ), which was the most significant enhancement in all cases, as a result of [emim][Ac] with low viscosity enhancing the diffusion of CO 2 in composites. 79 In the low concentration range (400-5000 ppm), the PEG200-modified sample of this work obtained the highest amine efficiency at 0.22 CO 2 /N, which was derived from more formation of carbamic acid caused by hydroxyl. 52 In general, it is recommended that the DAC systems prioritize the use of additives containing hydroxyl groups, which can enhance the efficiency of amines by changing the reaction mechanisms. This approach can help to spare more amine sites to make contact with weakly diffused CO 2 molecules. When it comes to carbon capture from flue gas or pure CO 2 at higher temperatures, additives should be used to break up the sorption products of CO 2 (ammonium carbamate) from a compact entity into separated fractions. This allows for the diffusion of more CO 2 into the deeper amine film. Consequently, additives without hydroxyl groups are preferred because they hardly interact with amines or products.

CONCLUSION
The results of our study indicate that the incorporation of minor molecular weight additives with hydroxyl functionalities can enhance the efficacy of DAC processes.
Among the additives investigated, polyethylene glycol 200 (PEG200) demonstrated exceptional performance, exhibiting a capacity of 2.051 mmol g −1 and an amine efficiency of 0.223 CO 2 /N. Furthermore, PEG200 exhibited full regeneration upon heating at 80 • C and retained 98.1% of its initial capacity after exposure to oxygen for 12 h, making it an ideal additive for the TEPA-loaded SBA-15 adsorbent used in DAC applications.
In evaluating the impact of various additives on adsorbent performance, we found that the effects of these additives are contingent on the type of additive, the pristine adsorbent, the method of incorporation, and the testing conditions. Based on these observations, we offer guidance for the selection of appropriate additives for different pristine adsorbents. For systems prioritizing hydroxyl-containing additives, DAC should be considered, whereas for systems targeting higher CO 2 concentration and temperature capture, additives without hydroxyl functionalities should be taken into consideration.
Finally, our study puts forth a novel hypothesis regarding the promotion of carbon capture by hydroxyl groups. We suggest that the hydroxyl group weakens the hydrogen bonding between adjacent amine groups, thereby increasing the availability of active amine sites, while simultaneously stabilizing the formation of carbamic acid by anchoring to the silica surface. Further investigation is necessary to fully validate this assumption.

EXPERIMENTS
Detailed experimental materials and methods can be found in the Supporting Information section.

A C K N O W L E D G M E N T S
The authors express their gratitude to the Science and Technology Commission of Shanghai Municipality (STCSM) for the financial support (No. 21DZ1206200). Dr. Li and Zhu extend their sincere appreciation to the National Natural Science Foundation of China for funding their research (Nos. 72140008 and 52006135), respectively. Furthermore, the authors would like to acknowledge Ms. Yan Zhu of the Instrumental Analysis Center at Shanghai Jiao Tong University for her invaluable assistance in performing the elemental analysis.

C O N F L I C T O F I N T E R E S T S TAT E M E N T
The authors declare no conflict of interest.