RESEARCH ARTICLE

www.small-journal.com

CO2 Uptake and Stability Enhancement in

Vinyltrimethoxysilane-Treated SBA-15 Solid Amine-Based
Sorbents
Anthony Vallace, Zachary S. Campbell, Hyun June Moon, William J. Koros,
Christopher W. Jones, and Ryan P. Lively*

and volatile byproduct emissions.[1–6] In

Silica-supported amine absorbents, including materials produced by tethering parallel with these point source tech-
aminosilanes or infusion of poly(ethyleneimine), represent a promising class nologies, numerous strategies have been
of materials for CO2 capture applications, including direct air and point proposed for the direct air capture (DAC)
source capture. Various silica surface treatments and functionalization of CO2 . Among the most promising is
the development of CO2 -selective solid
strategies are explored to enhance stability and CO2 uptake in amine-based
adsorbent materials, which avoid many
solid sorbent systems. Here, the synthesis and characterization of novel of the issues with aqueous-based liquid
vinyltrimethoxysilane-treated Santa Barbara Amorphous-15 (SBA-15) supports systems. In particular, porous silicas,
and the corresponding enhancement in CO2 uptake compared to various including commercially available amor-
SBA-15-based control supports are presented. The relationship between CO2 phous materials (e.g., Syloid C803) and
hierarchically structured materials (e.g.,
diffusion and amine efficiency in these systems is explored using a previously
Santa Barbara Amorphous-15, SBA-15)
reported kinetic model. The synthesized materials are characterized with CO2 have attracted significant attention due
and H2 O isotherms, diffuse reflectance infrared Fourier transform to their abundance, ease of preparation,
spectroscopy, 1 H T1 –T2 relaxation correlation NMR, and rapid thermal cycling and high pore volumes and specific sur-
experiments. The novel support materials are shown to enable high amine face areas. These materials are commonly
efficiencies, approaching a fourfold improvement over standard functionalized with aminosilanes (e.g.,
(3-aminopropyl)triethoxysilane (APTES))
SBA-15-supported amines, while simultaneously exhibiting excellent stability
or infused with oligomeric amines (e.g.,
when cycled rapidly under humid conditions. As the poly(ethyleneimine) poly(ethyleneimine) (PEI)), resulting in
loadings are held constant across the various samples, enhancements in CO2 effective CO2 -selective adsorbents.[7–12]
uptake are attributed to differences in the way the poly(ethyleneimine) Despite the promising nature of these
interacts with the support surface. silica-based adsorbents, there are still sig-
nificant opportunities for improvement,
including increasing thermodynamic
CO2 working capacities, achieving faster
1. Introduction diffusion rates (for shorter cycle times), enhancing stability over
repeated adsorption/desorption cycles in the presence of wa-
Current commercial CO2 capture processes utilize aqueous ter, and mitigating oxidation of the loaded amines.[13] Various
amine scrubbers, which suffer from high regeneration costs, methods have been developed to address these individual issues.
stability issues, large scrubber volumes, equipment corrosion, For example, functionalization of the silica with hexamethyldisi-
lazane (HMDS) creates a hydrophobic surface, thus minimizing
A. Vallace, Z. S. Campbell, H. J. Moon, W. J. Koros, C. W. Jones, R. P. Lively the interaction between the surface silanols and amine groups to
School of Chemical and Biomolecular Engineering improve thermodynamic CO2 uptake and amine efficiency.[14–16]
Georgia Institute of Technology
311 Ferst Dr., Atlanta, GA 30332, USA
Moon et al. grafted different functional groups to SBA-15 and
E-mail: [email protected] characterized the changes in PEI morphology via neutron scatter-
ing and T1 –T2 relaxation NMR as a function of the surface coat-
The ORCID identification number(s) for the author(s) of this article ing moiety. Specifically, when the SBA-15 is made hydrophobic
can be found under https://doi.org/10.1002/smll.202401422
using n-butyl trimethoxysilane, the PEI is mainly occluded in the
© 2024 The Author(s). Small published by Wiley-VCH GmbH. This is an
open access article under the terms of the Creative Commons
pores as aggregates (at low PEI loadings, ≈20 wt% or 40–50% vol-
Attribution-NonCommercial-NoDerivs License, which permits use and umetric pore occupancy) rather than stuck to the pore walls.[17,18]
distribution in any medium, provided the original work is properly cited, Another strategy was reported by Wang et al., where boron was
the use is non-commercial and no modifications or adaptations are incorporated as a heteroatom into SBA-15 and was subsequently
made. etched to increase the overall density of surface silanols. In doing
DOI: 10.1002/smll.202401422 so, it was possible to increase the number of APTES molecules

Small 2024, 20, 2401422 2401422 (1 of 12) © 2024 The Author(s). Small published by Wiley-VCH GmbH
 16136829, 2024, 46, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smll.202401422 by China University Of Petroleum, Wiley Online Library on [04/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.small-journal.com

that could be grafted to the SBA-15, thus increasing the CO2 up- Table 1. Table of comparable SBA-15-based materials.
take under a variety of conditions.[19] This strategy could be detri-
mental when used with physically impregnated PEI, as the higher Sample Sorbent Uptake Amine
silanol concentration interacts (via H-bonding and acid–base class [mmol g−1 ] efficiency
interaction) with more amines, potentially reducing the CO2 up- 20 wt% PEI V─B─SBA-15-AT (this work) 1 0.59 0.17
take, especially under ultradilute DAC conditions. Amine degra- 20 wt% SBA-1514 1 0.06 0.03
dation is also a significant challenge; chemical degradation can be H─SBA-1514 1 0.19 0.095
induced with O2 , CO2 , and/or H2 O, especially at elevated temper- Amine-grafted SBA-1519 2 0.52 0.19
atures, but this issue is not extensively explored in this work.[20,21]
There tend to be trade-offs between CO2 sorption capacity and
kinetics, which is illustrated by the differences in PEI-loaded ver- Our previously published kinetic adsorption model is used to
sus APTES-functionalized silicas. Branched PEIs are capable of extract reaction rate constants and the diffusivity change as a
high CO2 uptakes due to the high density of primary and sec- function of CO2 uptake to characterize the kinetic performance
ondary amines. When loaded into a porous silica substrate, at of the different supports. Finally, humid cycling tests are used to
low loadings, the PEI tends to coat the walls of the silica, and at probe the cyclic stability under realistic DAC conditions, demon-
higher loadings, it starts to fill the pores, resulting in a “plug” strating that VTMS-treated samples are significantly more sta-
morphology.[14,15] While these high loadings may be desirable ble than HMDS-treated samples, suggesting that the differences
from the standpoint of maximizing uptake, it is deleterious from in hydrophobicity and relative bulk between the vinyl group
the perspective of mass transport and kinetics, as the large plug (VTMS) and methyl groups (HMDS) have significant effects on
domains result in poor CO2 diffusion through the PEI phase and the PEI morphology, stability, and subsequent CO2 capture prop-
occlude the pore, resulting in slow diffusion of CO2 through the erties, which were also supported via characterization with wa-
pore. By comparison, APTES, being a single primary amine teth- ter adsorption isotherms and T1 –T2 relaxation correlation NMR
ered to the pore walls, has the opposite problem. The kinetics and measurements.
mass transport are fast, but the overall uptake is low compared The results reported in this work show that novel surface treat-
to PEI, as the APTES loading is limited by the surface area and ments of SBA-15 with VTMS increase the thermodynamic avail-
silanol concentration of the support, governing how much amine ability of nitrogen sorption sites for CO2 capture under ultra-
can be grafted to the support’s surface. Oligomerizing APTES in dilute conditions. These adsorbents combine the advantages of
the pores to create a higher density of binding sites ultimately high nitrogen activity, which is typically observed at high PEI
results in the same trade-off observed for PEI between kinetics, loadings (in untreated supports), with fast CO2 transport rates
diffusion rates, and capacity.[22,23] With these considerations in that are typically observed at low PEI loadings (in untreated sup-
mind, it is desirable to realize the high CO2 uptakes of PEI-loaded ports). It is critical to note that the samples reported here have sig-
materials while retaining the accelerated mass transport and ki- nificantly lower PEI loadings than typical DAC adsorbents, thus
netics of the APTES-functionalized silicas. allowing for fast CO2 transport rates.
In this work, we synthesized SBA-15-supported amines using
a variety of methods seeking to balance the CO2 uptake (thermo-
dynamics) with CO2 transport (kinetics). We tested two different 2. Results and Discussion
starting substrates, SBA-15 and boron-doped SBA-15 (B─SBA-
15), which were then functionalized with two different hydropho- Each sample and their corresponding treatments are shown
bizing agents, vinyltrimethoxysilane (VTMS) and HMDS. Select in Table 2. A representative flowchart of the functionaliza-
samples were then acid-treated to etch out the B and hydrolyze tion scheme is shown in Figure S1 (Supporting Information).
the remaining methoxy groups present in VTMS. The purpose of Briefly, SBA-15 denotes an unfunctionalized support, whereas
acid treatment on B─SBA-15 was to show the detrimental effect
of high silanol concentration on CO2 uptake, after loading PEI. Table 2. Samples and the corresponding treatments utilized in support
By contrast, the acid treatment of VTMS-treated B-doped SBA- preparation.
15 was used to selectively generate hydroxyl nests in between
vinyl moieties. We hypothesized that PEI would attach to these Sample Conditions
hydroxyl nests due their relative hydrophilicity. PEI was then
B-doped VTMS HMDS Acid
loaded into each support at the same PEI loading of ≈20 wt%.
treatment treatment treatment
The VTMS-treated, B-etched, PEI-loaded SBA-15 demonstrates
increased amine efficiencies with faster uptake kinetics com- SBA-15 – – – –

pared to controls with similar PEI loadings, attributable to a novel B─SBA-15 X – – –

PEI morphology within the pores resulting from differences in B─SBA-15-AT X – – X
PEI–wall interactions. Hydrophobized samples showed stronger V─SBA-15 – X – –
type I isotherms compared to hydrophilic supports. VTMS was V─B─SBA-15 X X – –
shown to be the preferred hydrophobizing agent due to the large V─SBA-15-AT – X – X
amount of thermodynamically available amines (after PEI im- V─B─SBA-15-AT X X – X
pregnation), especially under ultradilute CO2 activities. A table H─SBA-15 – – X –
of comparable class 1 and class 2 SBA-15-based materials may be
H─B─SBA-15-AT X – X X
found below (Table 1).

Small 2024, 20, 2401422 2401422 (2 of 12) © 2024 The Author(s). Small published by Wiley-VCH GmbH
 16136829, 2024, 46, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smll.202401422 by China University Of Petroleum, Wiley Online Library on [04/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.small-journal.com

(“PEI-infused pore volume”), fractional pore fill, and PEI weight

fraction by sample. As expected, the total pore volume in the sam-
ples treated with HMDS or VTMS was lower than the bare SBA-
15, as the hydrophobizing agent is tethered to the pore walls,
resulting in a lower overall pore volume. Support specific sur-
face areas, pore size distributions, N2 isotherms, may be found in
Table S1 and Figure S2 (Supporting Information).
Following the evaluation of the physical properties of the pre-
pared samples, dry isothermal CO2 uptakes at 30 and 60 °C
were measured via TGA under 400 ppm (DAC conditions)
(Figure 1A) and 10% CO2 (point-source conditions) (Figure 1B)
balanced in He. As seen in Figure 1A, at 30 °C, the samples
treated with hydrophobizing agents possessed the highest amine
efficiencies, consistently more than double the untreated sam-
ples. This difference is attributable to the known hydrogen bond-
ing and/or acid–base interaction between CO2 -active primary
and secondary amines in the PEI interacting with silanol groups
on the SiO2 surface.[18,24] The uptake trend for the amine effi-
ciencies of the untreated samples corresponds to the anticipated
number of available surface silanols present. For instance, the
B─SBA-15-AT sample possessed the lowest overall amine effi-
ciency (0.035 mol CO2 :mol N) and is expected to have the high-
est concentration of surface silanols due to acid etching of the
doped boron. Meanwhile, the B─SBA-15 and SBA-15 samples
possessed higher amine efficiencies (0.04 mol CO2 :mol N) as
they likely possessed fewer silanols to interact with the amines
Scheme 1. Depictions of the VTMS and HMDS surface functionalization,
acid etching of the B-doped SBA-15, and PEI infusion into an acid-etched than B─SBA-15-AT. Note that since the PEI loading is equivalent
and B-doped sample. in all cases, a higher amine efficiency corresponds to a higher
CO2 uptake.
a sample denoted B─SBA-15 indicates an originally boron- By comparison, at 400 ppm CO2 , samples treated with HMDS
doped support. Samples having undergone hydrophobization or VTMS exhibited dramatic increases in their amine efficiency.
with HMDS or VTMS are indicated by the presence of H─ and The least effective of the VTMS-treated samples, V─B─SBA-15,
V─, corresponding to HMDS and VTMS, respectively. Finally, all had an amine efficiency nearly twice the uptake of the untreated
acid-treated samples have an -AT appended at the end of the sam- samples (0.067 mol CO2 :mol N, which is an uptake of 0.21 mmol
ple name. For example, the sample V─B─SBA-15-AT indicates CO2 gsorbent −1 ). The V─SBA-15-AT, V─SBA-15, and H─SBA-15
that the SBA-15 was synthesized with the boron dopant, function- possessed even higher amine efficiencies, ranging from 0.09 to
alized with VTMS, and then acid-treated to etch out the boron 0.12 mol CO2 :mol N (0.29–0.38 mmol CO2 gsorbent −1 ). The ob-
sites. The HMDS and VTMS grafting steps, boron etching via served trend for these samples may be attributed to the degree
acid treatment, and PEI infusion are depicted in Scheme 1. of interaction between the pore wall and PEI (i.e., reduced in-
The synthesized supports and PEI-infused samples were char- teraction with pore walls with grafted hydrophobic groups vs
acterized via N2 physisorption and combustion thermogravimet- supports without hydrophobization). Notably, the V─B─SBA-15-
ric analysis (TGA) to evaluate the degree of pore filling and PEI AT sample had the highest amine efficiency, reaching an up-
weight fractions following the PEI infusion. Table 3shows the take of 0.17 mol CO2 :mol N (0.59 mmol CO2 gsorbent −1 ). This
initial pore volume (“support pore volume”), final pore volume suggests a more diffusion-advantaged PEI morphology than the

Table 3. Pore fill fractions and PEI weight loadings of each sample.

Sample Support pore volume [cc g−1 ] PEI-infused pore volume [cc g−1 ] Pore fill fraction PEI loading [wt%] SiOH density [mmol g−1 ]

SBA-15 1.2 ± 0.04 0.80 ± 0.03 0.35 23 2.38

B─SBA-15 0.98 ± 0.03 0.70 ± 0.02 0.34 20 2.69
B─SBA-15-AT 0.90 ± 0.03 0.50 ± 0.02 0.44 20 6.73
V─SBA-15 0.85 ± 0.03 0.48 ± 0.02 0.43 18 –
V─B─SBA-15 0.82 ± 0.03 0.48 ± 0.02 0.41 17 –
V─SBA-15-AT 0.80 ± 0.03 0.50 ± 0.02 0.41 17 –
V─B─SBA-15-AT 0.69 ± 0.02 0.29 ± 0.01 0.58 19 –
H─SBA-15 0.82 ± 0.03 0.38 ± 0.01 0.53 18 –
H─B─SBA-15-AT 0.84 ± 0.03 0.47 ± 0.02 0.56 19 –

Small 2024, 20, 2401422 2401422 (3 of 12) © 2024 The Author(s). Small published by Wiley-VCH GmbH
 16136829, 2024, 46, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smll.202401422 by China University Of Petroleum, Wiley Online Library on [04/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.small-journal.com

perature in a 3 h TGA experiment. The comparatively low pseu-

doequilibrium amine efficiencies under these conditions were
expected, as the combination of low CO2 concentration and
higher temperature resulted in an overall lower thermodynamic
driving force for CO2 uptake. Despite the reduced uptakes and
reduced variance between sample uptakes, similar trends to the
400 ppm CO2 , 30 °C experiments were apparent. The hydropho-
bized samples once again exhibited the highest uptakes. Interest-
ingly, the uptake for the V─B─SBA-15-AT and H─SBA-15 sam-
ples were similar, indicating that at 60 °C, the mass transport
advantage observed with V─B─SBA-15-AT at low temperature is
mitigated by the reduction in thermodynamic driving force when
compared to H─SBA-15.
Figure 1B shows significant change in the uptakes when the
concentration was increased to 10% CO2 . Under these condi-
tions, higher uptakes were visible for all samples at both tem-
peratures, with uptakes at 30 °C still outperforming the uptakes
at 60 °C. At both temperatures, a trend like the one observed
at 400 ppm is obtained, with the untreated samples exhibiting
lower amine efficiencies than the hydrophobized samples. The
V─B─SBA-15-AT sample possessed the highest amine efficiency
(0.28 mol CO2 :mol N), though the differences were reduced com-
pared to the 400 ppm uptakes due to the significant increase
in the concentration driving force. Meanwhile, the amine effi-
ciency of all samples noticeably decreased when the temperature
was increased to 60 °C, with the H─SBA-15 and V─B─SBA-15-
AT samples exhibiting similar amine efficiencies of ≈0.21 mol
CO2 :mol N. This decrease is expected due to the reduction of
the thermodynamic driving force for adsorption caused by the in-
crease in temperature. Table 4 shows sample uptakes in terms of
amine efficiency, mmol g−1 , and wt% CO2 . A graph of uptake in
mmol g−1 for all samples may be found in Figure S3 (Supporting
Information).
Our kinetic model was employed to determine the diffusion
rate as a function of amine efficiency, where the decay in dif-
fusivity is attributed to PEI cross-linking, which occurs as CO2
forms carbamates during the sorption process.[25] At 400 ppm
and 30 °C (Figure 2A), the diffusivity (D) versus amine efficiency
Figure 1. Pseudoequilibrium CO2 uptake normalized as amine efficiency profiles for samples containing VTMS were higher than samples
at 30 and 60 °C for A) 400 ppm and B) 10% CO2 concentrations. Dry con- without VTMS. V─SBA-15 showed the fastest diffusion profile,
ditions, He balance, 180 min equilibration. Error bars represent 2× stan- which suggests that there is some open pore space between the
dard deviation of duplicate TGA experiments run on different aliquots of likely smaller PEI aggregates for a greater contribution of pore
the same sample.
(Knudsen) diffusion over polymer diffusion. The V─B─SBA-15-
AT sample shows concavity in the D versus amine efficiency pro-
file while a higher initial diffusivity is observed. This could be
morphologies present in the VTMS- and HMDS-treated samples. related to the spacing of PEI domains in the pores as compared
Among the surface-treated samples, the H─B─SBA-15-AT sam- to bare SBA-15, B─SBA-15, and B─SBA-15-AT. In other words,
ple was a notable exception, exhibiting uptakes and amine effi- there may be a more advantageous diffusion pathway in the in-
ciencies similar to the samples without surface-grafted species traparticle pore space in the V─B─SBA-15-AT sample because
(0.05 mol CO2 :mol N, 0.16 mmol CO2 gsorbent −1 ). While the the PEI is preferentially bound to the ─OH nests where boron
HMDS surface coating results in fewer available amines being was etched out, thus creating more ordered spacing of PEI do-
inactivated by silanols, we suspect that the etched B sites serve as mains. At 60 °C (Figure 2B), diffusion was an order of magnitude
nucleation sites where PEI “beads” can form. When combined faster and amine efficiencies were low, consistent with the tem-
with the repulsiveness of the trimethyl silane surface moieties, perature effect on diffusivity and exothermic sorption. The sam-
these PEI beads may result in a low surface area available to CO2 ples showed similar D versus amine efficiency profiles at 60 °C,
reaction/diffusion and longer overall diffusion lengths. suggesting that the uptake is thermodynamically limited under
Using 400 ppm CO2 at 60 °C, the observed differences in these conditions, consistent with the above discussion.
amine efficiency were significantly lower, suggesting that ther- At 10% CO2 (Figure 2C,D) all of the D versus amine efficiency
modynamic equilibrium is achieved at the higher sorption tem- profiles exhibited convex behaviors, consistent with the rapid

Small 2024, 20, 2401422 2401422 (4 of 12) © 2024 The Author(s). Small published by Wiley-VCH GmbH
 16136829, 2024, 46, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smll.202401422 by China University Of Petroleum, Wiley Online Library on [04/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.small-journal.com

Table 4. CO2 uptakes for each sample and experimental condition.

Sample 400 ppm 10%

Units 30 °C 60 °C 30 °C 60 °C

SBA-15 mol CO2 :mol N 0.04 ± 0.01 0.01 ± 0.04 0.17 ± 0.02 0.13 ± 0.02
mmol gsorbent −1 0.16 ± 0.03 0.05 ± 0.03 0.67 ± 0.08 0.51 ± 0.08
wt% CO2 0.70 0.22 2.86 2.19
B─SBA-15 mol CO2 :mol N 0.04 ± 0.01 0.01 ± 0.00 0.15 ± 0.01 0.10 ± 0.00
mmol gsorbent −1 0.15 ± 0.03 0.05 ± 0.01 0.55 ± 0.06 0.37 ± 0.01
wt% CO2 0.66 0.22 2.36 1.60
B─SBA-15-AT mol CO2 :mol N 0.04 ± 0.00 0.01 ± 0.00 0.16 ± 0.00 0.10 ± 0.03
mmol gsorbent −1 0.13 ± 0.01 0.04 ± 0.01 0.56 ± 0.00 0.36 ± 0.11
wt% CO2 0.57 0.18 2.40 1.56
V─B─SBA-15 mol CO2 :mol N 0.07 ± 0.01 0.02 ± 0.00 0.20 ± 0.03 0.15 ± 0.00
mmol gsorbent −1 0.21 ± 0.04 0.05 ± 0.01 0.62 ± 0.08 0.46 ± 0.01
wt% CO2 0.92 0.22 2.66 1.98
V─SBA-15-AT mol CO2 :mol N 0.10 ± 0.01 0.02 ± 0.00 0.22 ± 0.00 0.18 ± 0.01
mmol gsorbent −1 0.29 ± 0.04 0.06 ± 0.01 0.65 ± 0.01 0.53 ± 0.01
wt% CO2 1.26 0.26 2.78 2.28
V─SBA-15 mol CO2 :mol N 0.09 ± 0.01 0.03 ± 0.04 0.22 ± 0.01 0.16 ± 0.00
mmol gsorbent −1 0.34 ± 0.03 0.11 ± 0.03 0.80 ± 0.03 0.60 ± 0.06
wt% CO2 1.47 0.48 3.40 2.57
H─SBA-15 mol CO2 :mol N 0.12 ± 0.01 0.04 ± 0.00 0.25 ± 0.02 0.21 ± 0.04
mmol gsorbent −1 0.38 ± 0.03 0.13 ± 0.01 0.82 ± 0.07 0.67 ± 0.13
wt% CO2 1.64 0.57 3.48 2.86
mol CO2 :mol N 0.12 ± 0.01 0.04 ± 0.00 0.25 ± 0.02 0.21 ± 0.04
mmol gsorbent −1 0.38 ± 0.03 0.13 ± 0.01 0.82 ± 0.07 0.67 ± 0.13
wt% CO2 1.64 0.57 3.48 2.86
V─B─SBA-15-AT mol CO2 :mol N 0.17 ± 0.03 0.04 ± 0.04 0.28 ± 0.02 0.21 ± 0.00
mmol gsorbent −1 0.59 ± 0.10 0.13 ± 0.04 0.94 ± 0.07 0.72 ± 0.01
wt% CO2 2.53 0.57 3.97 3.07
H─B─SBA-15-AT mol CO2 :mol N 0.05 ± 0.00 0.01 ± 0.00 0.16 ± 0.01 0.12 ± 0.01
mmol gsorbent −1 0.16 ± 0.00 0.05 ± 0.00 0.54 ± 0.03 0.42 ± 0.02
wt% CO2 0.70 0.22 2.32 1.81

decay in diffusivity due to the higher overall reaction rate. Sam- PEI spreads along the wall to minimize its surface energy, re-
ples without VTMS show some linear portions of their D versus sulting in shorter overall diffusion lengths (perpendicular to pore
efficiency profile at 60 °C, compared to the more rapid decay for direction). The downside is that a large fraction of the amine
samples without VTMS. All samples show some enhancement sorption sites have low chemical activity, as they are interact-
in the diffusivity profiles between 30 and 60 °C, with the sole ex- ing with Si─OH via acid–base and H-bonding, and are thus ren-
ception of the H─SBA-15 sample. This is consistent with the hy- dered unreactive with CO2 under the CO2 partial pressures of
pothesis that the plug domains are so large in H─SBA-15 that the interest.[14,15,17,18] The plug-like domains only (or predominantly)
increased diffusivity decay rate at 10% CO2 is more significant experience amine–amine interactions, and they should minimize
than the 30° difference in temperature, causing lower observed their surface energy by minimizing the local surface area, creat-
diffusivities. ing larger diffusion lengths, but rendering a larger number of po-
Previous work has shown that the PEI loading affects the local tentially active amine sorption sites. When the pores are coated
morphology of the PEI domains confined in the SBA-15 pores. with PEI, VTMS, or HMDS, the magnitude of the amine–amine
PEI initially forms a conformal coating on the silanol-laden, hy- interactions (PEI with itself) relative to the interaction of the PEI
drophilic pore walls and once the loading is high enough, the with an uncoated wall determines the size of the plug domain (for
walls are completely coated, and PEI forms plug-like occlusions instance, stronger amine–amine interactions relative to weaker
in the remaining pore space.[14,15] Two forces are exerted on PEI PEI–wall interactions likely result in larger PEI plugs). Model
determining the size and shape of the domains and the chemi- versus experimental uptake curves and images collected using a
cal activity of the nitrogen sorption sites (for a given control vol- scanning electron microscope showing the particle sizes for the
ume). The pore-wall force consists of H-bonding and acid–base B-doped and bare SBA-15 supports may be found in Figures S4
interactions of surface hydroxyls with the amines. Amine–amine and S5 (Supporting Information), respectively.
(H-bonding) interactions of PEI with itself are also present.[17,18] The apparent reactivity of the amines can be assessed by
In the case of wall-coated PEI, the diffusion length is compar- the shape of the CO2 isotherm, quantified with dual site Lang-
atively small relative to PEI in the plug-like occlusions, as the muir model parameters (Table S2, Supporting Information). Our

Small 2024, 20, 2401422 2401422 (5 of 12) © 2024 The Author(s). Small published by Wiley-VCH GmbH
 16136829, 2024, 46, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smll.202401422 by China University Of Petroleum, Wiley Online Library on [04/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.small-journal.com

Figure 2. Average diffusivity (D) versus amine efficiency curves extracted from the TGA curves via kinetic modeling. A) 400 ppm, 30 °C. B) 400 ppm,
60 °C. C) 10%, 30 °C. D) 10%, 60 °C.

hypothesis is based on the hydrophilic interactions of the pore is still relatively low, which can be explained by the increased
wall and PEI, where more hydrophilic surfaces interact more base–base interactions in the PEI domains due to the strong
strongly with PEI, thus deactivating many amine sites for CO2 repulsion-like force of the HMDS methyl groups. SBA-15 and
capture. This deactivation likely results in CO2 isotherms that B─SBA-15 had similar Henry’s constants, followed by B─SBA-
more significantly deviate from type I behavior. Because the PEI 15-AT which has the strongest wall–PEI interaction due to the
loadings are essentially equivalent across samples, the appar- increased surface hydroxyl concentration. The increased CO2 up-
ent changes in isotherm shape are attributable to differences take at higher CO2 pressures in bare SBA-15 and B─SBA-15-AT
in PEI surface affinity. In this work, CO2 isotherms, shown in is likely due to amine sites that were previously interacting with
Figure 3A,B, were collected at 60 °C to better approach pseudoe- hydroxyls at lower CO2 pressures. When the PEI–wall interac-
quilibrium in reasonable measurement times. The V─SBA-15 tion is relatively high, as in SBA-15 and B─SBA-15-AT, the pseu-
isotherm exhibited the strongest type I isotherm behavior, fol- doequilibrium CO2 capacity between 400 ppm and 10% differs
lowed by the V─SBA-15-AT, then the V─B─SBA-15 isotherm. significantly due to the inability of low CO2 concentrations to
This difference is ascribed to the hydrophilicity associated with overcome the strong PEI–wall interactions. Once the CO2 pres-
the hydrolysis of the VTMS methoxy groups and the acidity of sure is increased, these previously occupied amines begin to sorb
boron on the pore wall, respectively. Next is V─B─SBA-15-AT, CO2 , as seen in Figure 4A. V─B─SBA-15-AT, containing some
which will have some additional hydrophilicity from the etching (albeit very little) wall-coated PEI, is less type I than V─SBA-15,
of boron. H─SBA-15 is slightly higher than SBA-15 and B─SBA- V─SBA-AT, and V─B SBA-15. This result suggests that VTMS
15, consistent with the hydrophobization of the pore walls with methoxy groups hydrolyze during acid treatment (Figure 5B) and
methyl groups. The Henry’s constant (Figure 3C) of H─SBA-15 the hydroxyls from the etched boron are interacting with PEI.

Small 2024, 20, 2401422 2401422 (6 of 12) © 2024 The Author(s). Small published by Wiley-VCH GmbH
 16136829, 2024, 46, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smll.202401422 by China University Of Petroleum, Wiley Online Library on [04/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.small-journal.com

HMDS functionalization leads to larger plug domains (kinetic

effect, diffusion length) with the strongest amine–amine interac-
tions (thermodynamic effect) of any sample.
These results suggest that an optimum surface–PEI interac-
tion strength is desired. If the wall interaction is too high (SBA-
15, B─SBA-15-AT), amines are inactivated by hydrogen bonding
and/or acid–base interactions with the wall and if the degree of
wall interaction is too low or even repulsive (H─SBA-15), amines
will be inactivated with stronger base–base interactions. Overall,
the V─SBA-15 strikes a desirable balance between the pore-wall
and amine–amine interactions, leading to the most favorable PEI
morphology for maximum amine reactivity. PEI appears to inter-
act more favorably with the VTMS-treated surface compared to
the HMDS-treated surface, thus counteracting the amine–amine
interactions to a greater extent. The effect of surface treatment
on the CO2 sorption properties is also expressed in the Henry’s
constants, shown in Figure 3C, and follows the same trend as the
isotherm shape discussed above.
TGA– derivative thermogravimetry (DTG) curves (Figure S6,
Supporting Information) were utilized to evaluate the strength of
the PEI–support interactions for each of the tested supports. The
H─SBA-15 sample exhibited the characteristic derivative peak
of PEI loss at the lowest temperature, indicating the weakest
PEI–support interaction, while the B─SBA-15-AT exhibited the
strongest PEI–support interactions (derivative peak at the high-
est temperature). The remaining supports exhibited PEI–support
strengths between the H─SBA-15 and B─SBA-15-AT, with the
V─B─SBA-15-AT having a slightly stronger PEI–support inter-
action compared to H─SBA-15.
With the unexpected behavior exhibited by the VTMS-treated
samples, additional characterization was performed to evaluate
the PEI morphology as well as the hydrophobicity of the dif-
ferent supports and the resulting effects on the PEI morphol-
ogy and CO2 uptake behavior. Figure 4A shows T1 –T2 relax-
ation NMR correlation plots for SBA-15, B─SBA-15-AT, V─SBA-
15, and V─B─SBA-15-AT to evaluate the effects of different sur-
face treatments on the various supports at constant PEI loadings
(≈20 wt%). The 1 H T1 and T2 relaxation times can be linked
to packing and mobility of PEI chains.[17,18] Considering the na-
ture of polymers, higher T1 values generally correlate to more
solid-like domains having less free volume, due to slower 1 H
spin diffusion.[26–28] T2 relaxation times get longer with increas-
ing polymer mobility as polymer motions generally elongate spin
decoherence time for 1 H spins.[17,28,29] T1 –T2 NMR spectra for
the SBA-15 and B─SBA-15-AT are very similar, both exhibiting
signals consistent with PEI pore coating without the presence of
Figure 3. CO2 isotherms and isotherm parameters at 60 °C. A) Full signals at higher T2 values that would indicate the existence of
isotherms, where P0 = 1 atm CO2 . B) Isotherms at lower pressure ranges. higher mobility PEI “aggregates.”
C) Henry’s constants extracted from the dual site Langmuir model. Error Based on these results, it is reasonable to conclude that the
bars represent reported instrument uncertainty of the manifold volume SBA-15, B─SBA-15, and B─SBA-15-AT samples exist in the “Low
and from sample mass.
PEI” regime depicted in Figure 4B, resulting in inactivation
Interestingly, H─SBA-15 did not have a strong type I isotherm. caused by hydrogen bonding and/or acid–base interactions be-
This was initially surprising because the now nonexistent pore tween a significant fraction of available primary amines and the
wall force should give a greater number of active amines, surface silanol groups. This is also supported by the lower amine
since no amine sites are experiencing hydrogen bonding and/or efficiencies observed in Figure 1, which is further evidence of
acid–base interactions with wall silanols. However, the extreme a higher degree of hydrogen bonding and/or acid–base inter-
hydrophobicity of H─SBA-15 may make the amine–amine inter- action with pore-wall silanols compared to the samples treated
actions even stronger, deactivating nitrogen sites at low CO2 pres- with hydrophobizing agents. In addition to the signal consistent
sures because they are interacting with other amines. Thus, the with PEI pore coating, the V─SBA-15 sample also presents an

Small 2024, 20, 2401422 2401422 (7 of 12) © 2024 The Author(s). Small published by Wiley-VCH GmbH
 16136829, 2024, 46, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smll.202401422 by China University Of Petroleum, Wiley Online Library on [04/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.small-journal.com

Figure 4. A) 1 H T1 –T2 relaxation correlation plots for different samples. B) Cartoons displaying the speculated PEI morphologies with different PEI
loadings and surface treatments.

additional major signal, notably with lower mobility than the group when compared to the HMDS methyl groups (1 vinyl:≈2
pore-coating signal. This may be attributed to a steric “caging” ef- silanols vs 3 methyl:1 silanol). The lower density and bulk of
fect, where PEI domains are immobilized on the support surface the vinyl moieties, coupled with their possible rotation, may per-
by the vinyl moieties (VTMS-Treated, Figure 4B) due to their com- mit some PEI strands to lie between the surface vinyl groups
parative bulk and the lower relative surface density of the vinyl while interacting with fewer silanol groups, resulting in lower

Small 2024, 20, 2401422 2401422 (8 of 12) © 2024 The Author(s). Small published by Wiley-VCH GmbH
 16136829, 2024, 46, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smll.202401422 by China University Of Petroleum, Wiley Online Library on [04/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.small-journal.com

amine distribution also plays a major role, and having a favor-

able PEI morphology (e.g., brush-like) can lead to better amine
accessibility for the incoming CO2 .
Water isotherms and IR measurements were conducted on the
bare, VTMS-treated, and HMDS-treated supports to evaluate the
hydrophobicity of the pore surfaces to further understand the sur-
face effect on the PEI morphology in the pores (Figure 5). As ex-
pected, the H─SBA-15 sample exhibited the lowest water uptake,
the V─SBA-15 slightly higher, and the bare SBA-15 possessed
the highest maximal water uptake. The higher water uptake, and
thus lower effective hydrophobicity/higher hydrophilicity, could
potentially be attributed to three specific factors. First, the vinyl
group may be more hydrophilic compared to the methyls intro-
duced by HMDS due to VTMS’s 𝜋 electrons. Second, the dif-
ferent grafted groups (i.e., ─Si(CH3 )3 , ─SiCH═CH2 ) may intro-
duce differences in the density of the hydrophobic functional
groups due to the surface stoichiometry (i.e., 3 CH3 :1 Si─OH
vs 1 vinyl:≈2 Si─OH, on average), where the higher density and
relative bulk of CH3 moieties result in H2 O repulsion by the
HMDS-treated surface while lower density and available rotation
in vinyl groups result in reduced repulsion. Finally, potential hy-
drolysis of the expected residual VTMS methoxy groups could
result in the formation of one or two Si─OH groups adjacent to
the vinyl group, thus enabling greater interaction with water and
lower overall repulsion (decreased steric hindrance associated
with the lower density/bulk of the vinyl groups). Surprisingly, the
V─B─SBA-15-AT sample had higher water uptakes than the bare
SBA-15 at all relative humidities (RHs) below 75%, with its up-
take at 75% RH only marginally below the bare SBA-15. This is
likely attributable to the presence of silanol nests that result from
etching the boron dopant, which would be both preferential sites
for water condensation due to the relative confinement compared
to a full SBA-15 pore, as well as nucleation sites for more sig-
nificant condensation once water fills the silanol nests. Another
Figure 5. A) Water isotherms of bare SBA-15, V─SBA-15, H─SBA-15, and potential source for the observed hydrophilicity of V─B─SBA-15-
V─B─SBA-15-AT supports at 30 °C. B) Diffuse-reflectance infrared spectra AT is the additional Si─OH that is generated from the methoxy
of bare B─SBA-15, V─B─SBA-15, and V─B─SBA-15-AT supports. on the VTMS Si after the acid treatment.
As the molecular-scale interactions between PEI and the sup-
port surface may be attributable to a combination of factors,
mobility. V─B─SBA-15-AT exhibits a similar low-T2 signal, with including the final state of any residual methoxy groups on
higher T1 values, indicating PEI domains of similar mobility with the VTMS, diffuse reflectance infrared Fourier transform spec-
lower free volume of V─SBA-15 and V─B─SBA-AT. troscopy (DRIFTS) experiments were conducted on the B─SBA-
The similar mobility is attributed to the proposed caging ef- 15, V─B─SBA-15, and V─B─SBA-15-AT supports to further elu-
fect of VTMS and the decreased free volume is attributed to cidate these effects. As seen in Figure 5B, the B─SBA-15 support
the increased hydrophilicity of V─B─SBA-15-AT associated with does not exhibit any peaks in the 2800–3100 cm−1 range, while
the boron etching and hydrolyzed VTMS (methoxy → silanol). the V─B─SBA-15 and V─B─SBA-15-AT both possess peaks con-
Amine distribution also plays a major role, and having a favor- sistent with alkene and alkane C─H stretching. Notably, the sig-
able PEI morphology (e.g., brush-like) could return better amine nal at ≈2850 cm−1 , corresponding to the Si─OCH3 moieties, al-
accessibility for the CO2 . The proposed difference in amine dis- most completely disappears in the V─B─SBA-15-AT sample, in-
tribution could also contribute to the observed decrease in free dicating that residual methoxy groups from the tethered VTMS
volume of V─B─SBA-15-AT. A cartoon of the proposed PEI mor- groups are hydrolyzed during the acid treatment step. These dif-
phologies can be found in Figure 4B. Here we note that enhanced ferences also correspond to changes in the hydrogen bonded
CO2 performance of V─B─SBA-15-AT, having “caged” (i.e., lower Si─OH bands around from 3200 to 3400 cm−1 and free Si─OH
free volume) PEI domains than other samples may appear coun- peak found at ≈3700 cm−1 . Full spectra, spectra for other sup-
terintuitive. However, in all the materials studied here where PEI ports (i.e., H─SBA-15), and in situ DRIFTS may be found in
mostly formed conformal coating layers around the walls with Figures S7 and S8 (Supporting Information). NMR spectra (i.e.,
relatively low mobility, one may expect that CO2 diffusion plays a 1D T1 –T2 relaxation NMR spectra, 1 H magic-angle spinning
less prominent effect on CO2 sorption properties when compared (MAS) solid-state NMR spectra) may be found in Figure S9 (Sup-
against very high PEI loading materials. Beyond amine mobility, porting Information).

Small 2024, 20, 2401422 2401422 (9 of 12) © 2024 The Author(s). Small published by Wiley-VCH GmbH
 16136829, 2024, 46, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smll.202401422 by China University Of Petroleum, Wiley Online Library on [04/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.small-journal.com

In situ CO2 DRIFTS spectra, shown in Figure S8 (Support-

ing Information), show two predominant bands for all sam-
ples at ≈1685 and ≈1540 cm−1 and are assigned to the C═O
stretch of carbamic acid (NCOOH) and an O═C═O─ asymmetric
stretch of carbamate anions (presumably ammonium carbamate)
(NHCOO− ), respectively. These results suggest a mixed mecha-
nism where some CO2 adsorbs to form a carbamic acid species,
which has a theoretical stoichiometry of 1 CO2 :1 N and some CO2
adsorbs to form carbamates, which has a 1 CO2 :2 N stoichiome-
try. The ammonium carbamate species are expected as there are
many proximate amine moieties to deprotonate the initial zwitte-
rion that forms when CO2 is adsorbed. The carbamic acid species
are reported to be associated with secondary amines[30] which are
present in all samples, and also may form on isolated amines
far from available base sites. Based on the consistent mechanism
across all samples, it is likely that the difference in observed CO2
uptake is due to the different support surface moieties changing
the availability of amines, but not changing the adsorption mech-
anism. A schematic of this mechanism may be found as Scheme
S1 (Supporting Information).
Based on these results, several representative samples, includ-
ing the SBA-15, H─SBA-15, V─SBA-15, and the V─B─SBA-15-
AT, were selected for further testing in humid cycling experi-
ments to evaluate their overall performance and stability in the
presence of humidity. CO2 uptakes were collected for 11 ad-
sorption/desorption cycles in humid 400 ppm CO2 , and the re-
sults are shown in Figure 6A. It was observed that the SBA-15,
V─SBA-15, and H─SBA-15 exhibited similar amine efficiencies,
around 0.12 mol CO2 :mol N, while the V─B─SBA-15-AT sam-
ple presented amine efficiencies nearly 25% higher (0.15 mol
CO2 :mol N). This result is attributable to superior kinetics in the
V─B─SBA-15-AT sample, as the 30 min cycles used in these ex-
periments did not provide enough time for the SBA-15, V─SBA-
15, and H─SBA-15 to functionally reach pseudoequilibrium due
to their comparatively diffusion-disadvantaged PEI morpholo- Figure 6. Humid cycling experiments for representative samples.
gies. It was also observed that the performance of the H─SBA- 400 ppm CO2 , 40% relative humidity. A) Amine efficiency (mol CO2 :mol
15 appeared to degrade after just 11 cycles, while the CO2 up- N) versus cycle. 30 min adsorption at 30 °C, 60 min desorption in 40% RH
humid N2 at 90 °C. B) % PEI loss following 50 rapid temperature swing
takes in the remaining samples appeared to functionally remain
cycles between 30 and 110 °C with 5 min holds in 400 ppm CO2 (40% RH).
constant, which is likely attributable to the high degree of sup-
port hydrophobicity in the H─SBA-15 sample resulting in neg-
ligible PEI–support interaction. This lack of interaction between cant roles in preventing PEI loss. In this case, the H─SBA-15
PEI and the pore wall likely results in increased PEI migration or has the greatest density of hydrophobic functional groups (and
extraction when exposed to humidity over several cycles, extrud- no surface Si─OH) while the VTMS-treated samples possess a
ing PEI from the pores. lower density of mobile hydrophobic functional groups, thus in-
With the observed degradation in the capacity of the H─SBA- teracting to a greater degree with the PEI, while the VTMS- and
15 in 11 full adsorption/desorption cycles, accelerated degrada- acid-treated supports (i.e., V─B─SBA-15-AT) possess numerous
tion experiments were conducted on the SBA-15, V─SBA-15, surface silanols that anchor the PEI in place. The observed dif-
H─SBA-15, and V─B─SBA-15-AT samples to evaluate longer ferences in PEI affinity of VTMS- versus HMDS-treated surfaces
term stability of the sorbent sample. The samples were exposed to arise both from the ability of the vinyl group to “cage” PEI do-
50 rapid temperature swings in the presence of humid (40% RH) mains and the spacing or surface density of VTMS versus HMDS
400 ppm CO2 , after which the sample mass was compared with moieties on the SiO2 surface. The grafting stoichiometry and ro-
the initial dry sample mass. As the SBA-15 supports are stable tational mobility of the vinyl group may lead to surface “patchi-
under the experimental conditions, any mass loss observed was ness,” contributing to the higher PEI affinity.
attributed to the loss of PEI. As seen in Figure 6B, the H─SBA-
15 sample exhibited the highest mass loss, with a 13% decrease 3. Conclusion
in the mass of PEI, while the V─SBA-15 decreased by 6%, the
SBA-15 decreased by 1%, and the V─B─SBA-15-AT sample ex- The CO2 uptake and stability of PEI-loaded SBA-15 adsorbent
perienced almost no PEI loss (0.3%). These results indicate that materials exposed to different surface treatments with sim-
the interaction between the PEI and the support plays signifi- ilar PEI weight loadings were evaluated under a variety of

Small 2024, 20, 2401422 2401422 (10 of 12) © 2024 The Author(s). Small published by Wiley-VCH GmbH
 16136829, 2024, 46, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smll.202401422 by China University Of Petroleum, Wiley Online Library on [04/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.small-journal.com

CO2 concentrations and temperatures. The modified support the United States Government or any agency thereof. This work was
materials, including standard SBA-15 functionalized with VTMS performed in part at the Materials Characterization Facility (MCF) at
and a boron-doped VTMS-treated SBA-15 exposed to an acid the Georgia Tech. The MCF is jointly supported by the GT Institute
for Materials (IMat) and the Institute for Electronics and Nanotechnol-
etch to incorporate small pockets of acidic silanols were com-
ogy (IEN), which is a member of the National Nanotechnology Coor-
pared with several controls. These controls included untreated dinated Infrastructure supported by the National Science Foundation
SBA-15 and supports hydrophobized with HMDS. Experiments (Grant No. ECCS-2025462).
utilizing constant PEI loadings (≈20 wt%) across all samples
showed that VTMS-treated SBA-15 exhibited CO2 uptakes supe-
rior to untreated SBA-15 and comparable to HMDS-treated SBA- Conflict of Interest
15, while the boron-doped, VTMS-treated, acid-etched sample ex-
The authors declare the following competing financial interest(s): C.W.J.
hibited the highest overall amine efficiencies, reaching 0.17 mol has a financial interest in several companies that seek to commercialize
CO2 :mol N under 400 ppm CO2 at 30 °C and 0.28 mol CO2 :mol CO2 capture from air. This work is not affiliated with any such firms. C.W.J.
N under 10% CO2 at 60 °C. These differences in CO2 uptake are has a conflict-of-interest management plan in place at Georgia Tech.
attributable to differences in the degree of interaction between
PEI and the support surface. The tested samples were modeled
to extract average diffusivity as a function of amine efficiency, Data Availability Statement
and CO2 isotherms were collected to evaluate the adsorbent be- The data that support the findings of this study are available in the sup-
havior across a range of CO2 partial pressures. Dual-site Lang- plementary material of this article.
muir parameters and Henry’s constants broadly show that the
CO2 isotherm can be shifted toward higher CO2 uptakes at lower
partial pressures (stronger type I isotherm behavior) with VTMS Keywords
treatment and provide insight into the chemical activity of PEI as
adsorption, CO2 capture, functionalization, porous materials, T1 –T2 relax-
a function of the support surface treatment. ation correlation NMR
The adsorbent materials were characterized using 1 H T1 –T2
relaxation correlation NMR to evaluate changes in PEI morphol- Received: February 22, 2024
ogy in the different supports. The water isotherms and diffuse- Revised: August 1, 2024
reflectance infrared spectra of the bare supports were utilized Published online: August 9, 2024
to elucidate the molecular-level interactions between the sup-
port surface and PEI (e.g., presence of Si─OH groups, density
and bulk/mobility of surface functional groups), corresponding
to overall support hydrophilicity and its effect on adsorbent per- [1] S. D. Kenarsari, D. Yang, G. Jiang, S. Zhang, J. Wang, A. G. Russell,
formance. Finally, cycling experiments under humid conditions Q. Wei, M. Fan, RSC Adv. 2013, 3, 22739.
were conducted to evaluate the stability of the different supports [2] J. D. Shakun, P. U. Clark, F. He, S. A. Marcott, A. C. Mix, Z. Liu, B.
Otto-Bliesner, A. Schmittner, E. Bard, Nature 2012, 484, 49.
under realistic DAC conditions. VTMS-treated supports were
[3] S. Solomon, G. K. Plattner, R. Knutti, P. Friedlingstein, Proc. Natl.
shown to be significantly more stable than SBA-15 hydropho- Acad. Sci. USA 2009, 106, 1704.
bized with HMDS. The boron-doped, VTMS-treated, acid-etched [4] E. S. Sanz-Pérez, C. R. Murdock, S. A. Didas, C. W. Jones, Chem. Rev.
sample exhibited almost no PEI mass loss during accelerated ag- 2016, 116, 11840.
ing experiments. [5] K. S. Lackner, Eur. Phys. J.: Spec. Top. 2009, 176, 93.
Key limitations of this study include the narrow range of low [6] D. Aaron, C. Tsouris, Sep. Sci. Technol. 2005, 40, 321.
PEI loadings and surface treatments explored, as well as the [7] K. Il Kim, R. Lawler, H. J. Moon, P. Narayanan, M. A. Sakwa-Novak,
use of PEI as the sole polyamine in preparation of the mate- C. W. Jones, S. S. Jang, ACS Omega 2021, 6, 3390.
rials. Furthermore, the materials presented in this work would [8] B. Min, S. Yang, A. Korde, C. W. Jones, S. Nair, Adv. Mater. Interfaces
also benefit from additional characterization with advanced tech- 2020, 7, 2000926.
[9] Y. Kuwahara, D. Y. Kang, J. R. Copeland, N. A. Brunelli, S. A. Didas,
niques, specifically neutron scattering (i.e., quasielastic neutron
P. Bollini, C. Sievers, T. Kamegawa, H. Yamashita, C. W. Jones, J. Am.
scattering, small-angle neutron scattering), to further elucidate Chem. Soc. 2012, 134, 10757.
PEI morphological differences between the different supports. [10] A. Kumar, D. G. Madden, M. Lusi, K. J. Chen, E. A. Daniels, T. Curtin,
J. J. Perry, M. J. Zaworotko, Angew. Chem., Int. Ed. 2015, 54, 14372.
[11] G. Rim, F. Kong, M. Song, C. Rosu, P. Priyadarshini, R. P. Lively, C. W.
Supporting Information Jones, JACS Au 2022, 2, 380.
[12] Y. J. Min, A. Ganesan, M. J. Realff, C. W. Jones, ACS Appl. Mater. In-
Supporting Information is available from the Wiley Online Library or from
terfaces 2022, 14, 40992.
the author.
[13] N. Sun, Z. Tang, W. Wei, C. E. Snape, Y. Sun, Front. Energy Res. 2015,
3.
[14] A. Holewinski, M. A. Sakwa-Novak, C. W. Jones, J. Am. Chem. Soc.
Acknowledgements 2015, 137, 11749.
A.V. and Z.S.C. contributed equally to this work. The information, data, [15] A. Holewinski, M. A. Sakwa-Novak, J.-M. Y. Carrillo, M. E. Potter, N.
or work presented herein was funded in part by the Advanced Research Ellebracht, G. Rother, B. G. Sumpter, C. W. Jones, J. Phys. Chem. B
Projects Agency-Energy (Grant No. ARPA-E), the US Department of En- 2017, 121, 6721.
ergy, under Award Number DE-AR0001309. The views and opinions of [16] L. Wang, M. Al-Aufi, C. N. Pacheco, L. Xie, R. M. Rioux, ACS Sustain-
authors expressed herein did not necessarily state or reflect those of able Chem. Eng. 2019, 7, 14785.

Small 2024, 20, 2401422 2401422 (11 of 12) © 2024 The Author(s). Small published by Wiley-VCH GmbH
 16136829, 2024, 46, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smll.202401422 by China University Of Petroleum, Wiley Online Library on [04/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.small-journal.com

[17] H. J. Moon, R. S. Sekiya, C. W. Jones, J. Phys. Chem. C 2023, 127, [23] P. Bollini, N. A. Brunelli, S. A. Didas, C. W. Jones, Ind. Eng. Chem. Res.
11652. 2012, 51, 15153.
[18] H. J. Moon, J. Carrillo, J. Leisen, B. G. Sumpter, N. C. Osti, M. Tyagi, [24] C. F. Cogswell, Z. Xie, A. Wolek, Y. Wang, A. Stavola, M. Finkenaur,
C. W. Jones, J. Am. Chem. Soc. 2022, 144, 11664. E. Gilmore, M. Lanzillotti, S. Choi, J. Mater. Chem. A 2017, 5,
[19] Y. Wang, J. T. Anyanwu, Z. Hu, R. T. Yang, Sep. Purif. Technol. 2023, 8526.
309, 123030. [25] A. Vallace, Y. Ren, C. W. Jones, R. P. Lively, Chem. Eng. J. 2023, 472,
[20] Y. A. Guta, J. Carneiro, S. Li, G. Innocenti, S. H. Pang, M. A. Sakwa- 144838.
Novak, C. Sievers, C. W. Jones, ACS Appl. Mater. Interfaces 2023, 15, [26] P. Xv, C. No, K. O. Laboratorium, Physica 1949, 15, 386.
46790. [27] D. N. Theodorou, U. W. Suter, Macromolecules 1985, 6, 1206.
[21] J. S. A. Carneiro, G. Innocenti, H. J. Moon, Y. Guta, L. Proaño, C. [28] M. H. Levitt, Spin Dynamics: Basics of Nuclear Magnetic Resonance,
Sievers, M. A. Sakwa-Novak, E. W. Ping, C. W. Jones, Angew. Chem., 2nd ed., Wiley, NY, USA 2013.
Int. Ed. 2023, 62, 202302887. [29] S. Godefroy, B. Ryland, P. T. Callaghan, 2D Laplace Inversion Instruction
[22] P. Bollini, N. A. Brunelli, S. A. Didas, C. W. Jones, Ind. Eng. Chem. Res. Manual, Academic Press, Wellington, New Zealand 2003.
2012, 51, 15145. [30] Y. Zhai, S. S. C. Chuang, Energy Technol. 2017, 5, 510.

Small 2024, 20, 2401422 2401422 (12 of 12) © 2024 The Author(s). Small published by Wiley-VCH GmbH