Front. Chem. Sci. Eng.

2021, 15(2): 437–449

https://doi.org/10.1007/s11705-020-1961-6

RESEARCH ARTICLE

Effects of functional groups for CO2 capture using metal

organic frameworks

Chenkai Gu1, Yang Liu2, Weizhou Wang3, Jing Liu (✉)1, Jianbo Hu1
1 State Key Laboratory of Coal Combustion, School of Energy and Power Engineering, Huazhong University of Science and Technology,
Wuhan 430074, China
2 School of Chemical and Biomolecular Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0100, USA
3 Henan Key Laboratory of Function-Oriented Porous Materials, College of Chemistry and Chemical Engineering, Luoyang Normal University,
Luoyang 471934, China

© Higher Education Press 2020

Abstract Metal organic frameworks (MOFs) are promis- great potentials in short-term CO2 reduction [3]. There are
ing adsorbents for CO2 capture. Functional groups on several approaches to capture CO2 from carbon emitting
organic linkers of MOFs play important roles in improving sources, including amine scrubbing [4,5], membrane
the CO2 capture ability by enhancing the CO2 sorption separation [6,7], and sorbent adsorption [8–10]. Compared
affinity. In this work, the functionalization effects on CO2 to the other methods, capturing CO2 using porous sorbents
adsorption were systematically investigated by rationally is especially attractive owing to the low energy consump-
incorporating various functional groups including –SO3H, tion, high economic benefit and easy tailoring [11–13].
–COOH, –NH2, –OH, –CN, –CH3 and –F into a MOF-177 Comparing with the traditional porous materials such as
template using computational methods. Asymmetries of activated carbons and zeolites, metal organic frameworks
electron density on the functionalized linkers were (MOFs) show more advantages in CO2 capture due to their
intensified, introducing significant enhancements of the highly tunable structures, porosities and chemical proper-
CO2 adsorption ability of the modified MOF-177. In ties [14–16]. In order to improve the CO2 capture
addition, three kinds of molecular interactions between capacities of MOFs, many approaches have been pro-
CO2 and functional groups were analyzed and summarized posed. The pore chemistry adjustment can benefit CO2
in this work. Especially, our results reveal that –SO3H is capture in MOFs, which depends strongly on pore
the best-performing functional group for CO2 capture in topology [17]. Besides, smaller pore size usually leads to
MOFs, better than the widely used –NH2 or –F groups. The higher overlap of potential energy between CO2 molecule
current study provides a novel route for future MOF and MOF, contributing to higher CO2 capture abilities. An
modification toward CO2 capture. and Rosi [18] tuned CO2 adsorption properties of MOFs
through cation exchange and concluded that smaller pores
Keywords metal-organic frameworks, functional groups, in MOFs may be an ideal way to condense CO2 at
CO2 capture, GCMC, DFT temperatures of real-world application. Pal et al. [19]
reported that the selective adsorption of CO2 can be
improved by forming smaller pores via structural inter-
1 Introduction penetrations. Via creating the cage-within-cage architec-
ture, Zheng et al. partitioned the pore space into small
Carbon dioxide (CO2) from fossil fuel combustion is one domains which are among key desirable features for
of the most important greenhouse gases which result in the enhancing CO2 adsorption through a better size match and
global warming [1,2]. Various methods have been stronger charged-induced forces [20].
proposed to prevent the CO2 emission into the atmosphere, For CO2 capture, functional groups can significantly
among which the CO2 capture and storage (CCS) shows affect the CO2 adsorption capacities of MOFs [21,22].
Xiang et al. [23] incorporated polar acidic functionalities
into MOFs and demonstrated that this is an alternative
approach for enhancing CO2 capture. Zheng et al. [24]
Received April 3, 2020; accepted May 19, 2020 found that the acylamide-decorated MOFs show higher
E-mail: [email protected] adsorption heat and improved CO2 adsorption capacity
438 Front. Chem. Sci. Eng. 2021, 15(2): 437–449

than unfunctionalized MOFs. An et al. [25] pointed out

that amine decoration can significantly enhance the CO2
uptake in MOFs through forming favorable interactions.
Ye et al. [26] found that MOFs with rich accessible
nitrogen sites show highly efficient CO2 capture and
separation. Li et al. [17] carried out high-through
computational screening of a large number of computer-
generated MTV-MOFs and concluded that the functiona-
lization and small pores are two beneficial factors for CO2
capture in MOFs. Among various functionalization
methods, incorporations of nitrogen atoms or nitrogen-
containing functional groups are commonly used to
enhance the CO2 interaction ability in previous work.
However, recent studies have shown that incorporating Fig. 1 Structure of MOF-177 (Zn, light blue tetrahedrons; O, red
spheres; C, gray spheres; and H, white spheres).
oxygen onto organic linkers can better enhance the CO2
adsorption in MOFs [27]. Various oxygen-containing
functional groups like –OH and –COOH have been diameter to He (2.6 Å). The structural parameters
investigated for this purpose [28,29]. However, –SO3H mentioned above was calculated with the tool of the
containing more oxygen has not been reported for its Atom Volumes & Surfaces in Materials Studio 2017 [31].
potential in improving CO2 capture capability of MOFs. The pore diameters were obtained by using Poreblazer
In this work, we investigated the functionalization effect v3.0.2 [34].
of –SO3H group on CO2 adsorption capacity of MOFs
using grand canonical Monte Carlo (GCMC) simulations 2.2 DFT calculations
and density functional theory (DFT) calculations. The
incorporation of large functional groups like –SO3H Theoretical calculations were performed to explore the
requires enough pore space in MOFs. The MOF-177 has underlying mechanisms in terms of physical chemistry
suitable pore size and some of the functionalized MOF-177 [35–38]. The entire periodic MOF structure contains
has been successfully synthesized in experiments [30]. hundreds of atoms and will cost too much computational
Therefore, the MOF-177 was treated as the parent material resources to carry out DFT calculations. In order to save
in this work. By comparing with various functional groups computational cost, cluster models were constructed using
including –COOH, –NH2, –OH, –CN, –CH3 and –F in the method as described in our previous studies [28,29].
MOF-177, we verified our hypothesis that –SO3H can The models used in this work were given in Figs. S1–S8
greatly enhance the CO2 adsorption capacity of MOF (cf. Electronic Supplementary Material, ESM). All of the
materials. The current study revealed a novel functional theoretical calculations were performed using DMol3 code
group for benefiting the future adsorbent developments. with exact DFT spherical atomic orbitals [31]. This method
has already minimized the basis set superposition error
effects which strongly affects the binding energy results
2 Model and simulation method [39]. The PW91 exchange-correlation functional was
applied to investigate the interactions between MOFs and
2.1 MOF structure CO2 molecules [40]. The double numeric polarization
basis set was used to describe the atomic orbital [41]. The
The initial structure of MOF-177 was obtained from the DFT semi-core pseudo-pots [42] which was developed
experimental X-ray diffraction [30], as shown in Fig. 1. specially for DMol3 calculations was used to set the core
MOF-177 is a porous crystalline material which consists of treatment type. The convergence tolerance parameters
Zn4O tetrahedrons connected with 1,3,5-benzenetriben- were 2
10–5 (energy), 4
10–3 (force) and 5
10–3
zoate (BTB) ligands. The functional groups were intro- (displacement), respectively.
duced to MOF-177, and then the whole structures were All initial cluster models were first optimized using the
geometry optimized using Forcite module [31]. The calculation methods mentioned above and shown in Fig. 2.
MOF-177 functionalized by –F, –CH3, –NH2 has been The binding sites were classified and signed clearly in
synthesized by Zhang et al. [30]. Besides, the –COOH and Fig. 2(c). For each binding site, three different possible
–SO3H groups have been incorporated to some other orientations of CO2 molecule were investigated, marked as
MOFs in experiments [32,33]. The structural properties of parallel (P), cross (C) and vertical (V), as shown in
functionalized MOF-177 were given in Table 1. The Fig. 2(a). Additional geometry optimizations were per-
accessible surface area was calculated with an equal van formed to the CO2-cluster models. Finally, the single point
der Waals diameter to N2 molecule (3.64 Å). The total pore binding energy of CO2 around functional group was
volume was calculated with an equal van der Waals calculated with the following equation:
 Chenkai Gu et al. Effects of functional groups for CO2 capture using metal organic frameworks 439

Table 1 Structural and limiting CO2 adsorption properties of the MOFs studied in this work
Sacc
MOF dpore/Å rcrsy/(g$cm–3) Vpore/(cm3$g–1) void fraction –1
qst0/(kJ$mol–1)
m $g
2
m2$cm–3
MOF-177 11.99 0.422 1.925 0.811 3977 1677 – 13.41
MOF-177-NH2 11.79 0.455 1.739 0.791 3939 1792 – 16.89
MOF-177-COOH 11.36 0.519 1.475 0.765 3662 1900 – 19.00
MOF-177-CH3 11.76 0.453 1.725 0.781 4076 1845 – 14.17
MOF-177-F 11.98 0.461 1.744 0.804 3731 1722 – 14.67
MOF-177-SO3H 11.54 0.598 1.230 0.736 3260 1950 – 20.52
MOF-177-CN 11.91 0.477 1.642 0.783 3838 1830 – 15.46
MOF-177-OH 11.79 0.457 1.747 0.798 3843 1757 – 16.48

BE ¼ ECO2 cluster – ðECO2 þ Ecluster Þ: (1) 8.8542
10–12 C2$N–1$m–2, and αβ , εαβ are the collision
diameter and well depth, respectively. In this study, the
In the equation above, ECO2 represents the total energy of CO2 was modeled as a rigid linear molecule with two
CO2, Ecluster represents the total energy of cluster, and bonds in the length of 1.16 Å, taken from the transferable
ECO2-cluster represents the total energy of CO2-cluster potentials for phase equilibria force field [44]. The internal
system under equilibrium state. A negative BE corre- quadrupole moment was described through partial charges
sponds to an exothermic binding process and a higher on each charged Lennard-Jones site (qc = + 0.70 e and qo =
negative value means a stronger binding process [43]. –0.35 e). The universal force field [45] has been
successfully applied in the study of CO2 adsorption in
2.3 Force field MOFs, and was used to calculate the interactions between
The interactions considered in this simulation, including adsorbates and adsorbents in this work. All Lennard-Jones
gas-gas part and gas-adsorbent part, were calculated by parameters were mixed with Lorentz-Berthelot rules, as
combining site-site Lennard-Jones (L-J) and Coulombic shown in Table 2. For a given MOF, the determination of
potentials: framework atom charges mainly contains two steps: (1) the
X   12  6   calculation of the electrostatic potential (ESP) in the MOF
αβ αβ qα qβ
α 2 i 4εαβ
uij ðrÞ ¼ – þ , pores [46], (2) the assignment of electronic charges to
β2j gαβ gαβ 4πε0 gαβ MOF atoms to reproduce the ESP using fitting procedures
(2) such as the repeating electrostatic potential extracted
atomic [47], the molecular building block-based [48] and
where ε0 is the permittivity of the vacuum and equals to the density derived electrostatic and chemical [49]

Fig. 2 (a) V, C and P represent the orientations of “vertical”, “cross” and “parallel”; (b) the basic model for binding energy and atom
charge calculations; (c) U, F and R represent the binding site of “top”, “front” and “right”, respectively (Zn, light blue spheres; O, red
spheres; C, gray spheres; H, white spheres and functional group: purple sphere).
440 Front. Chem. Sci. Eng. 2021, 15(2): 437–449

Table 2 Lennard-Jones parameters used in this work

3 Results and discussion
Atom types s/Å ε/(kb$K–1)
CO2–C 2.800 27 3.1 Electrostatic fields and orbitals of the linkers
CO2–O 3.050 79
MOF–Zn 2.462 62.397 The cluster model was geometry optimized followed by
MOF–N 3.261 34.721 calculating the ESP charges as listed in Tables S1–S8 (cf.
MOF–S 3.595 137.877 ESM). For all the MOFs, although the Zn and surrounding
MOF–F 2.997 25.16 O in metal clusters have larger charges than the other atoms
of frameworks, the corner space formed by metal clusters
MOF–C 3.431 52.836
is rather limited for strong CO2 binding. As shown in the
MOF–O 3.118 30.192 tables, introducing functional groups can significantly
MOF–H 2.571 22.141 affect the electrostatic field as evidenced by the changed
atomic charges on the organic linkers. Taking MOF-177-
SO3H as an example, the absolute values of atom charges
methods. In this work, the ESP charges of the MOF in –SO3H group range from 0.338 e to 0.484 e (as shown in
frameworks were calculated via typical MOF-representa- Table S8). These charges are much larger than those of
tive models which have been successfully used in previous other nonmetal atoms with value ranging from –0.229 e to
work [50–52]. + 0.117 e, leading to the enhancement of electrostatic
interactions between CO2 molecules and organic linkers.
2.4 GCMC simulation
Besides, the incorporation of –SO3H group into MOFs will
result in the decrease of pore sizes. This will lead to the
The GCMC simulation was performed to investigate the
stronger overlap of the potential energy fields for CO2
adsorption of CO2 in MOF-177. The simulation box of
molecules. Consequently, the CO2 capture performance
MOF-177 containing 8 (2
2
2) unit cells with periodic
will be highly improved especially at low pressure.
boundary conditions was applied in this work. The
To study the change transfer between CO2 and organic
frameworks studied in this work were treated as rigid
linkers, the lowest unoccupied molecular orbitals
with atoms frozen at their crystallographic positions during
(LUMOs) of functionalized organic linkers are shown in
calculations. A cutoff radius was set to 12.8 Å for the van
Fig. 3. Compared to nonpolar groups like –H and –CH3,
de Waals interactions, and the long-range electrostatic
polar groups like –CN and –SO3H can significantly
interactions were handled using the Ewald summation
intensify the asymmetry of electron density and change
method with the tinfoil boundary condition. The GCMC
the delocalization of electron density. Notably, the –SO3H
calculation in each point consists of 2.0
107 steps. The
group shows more significance on enhancing the deloca-
former half part was used to guarantee the equilibration
lization of LUMO in organic linkers, which further
and the latter half part was used to sample the desired
strengthens the interactions with the orbitals of CO2
thermodynamics properties. The input fugacity was
molecules.
calculated with the state equation of Peng-Robinson. The
isosteric heat qst was determined by the statistical average
3.2 Configurations of CO2 interacting with functionalized
of CO2 binding energies at various sites which contain
linkers
both the sites around functional groups and some other
sites in MOFs, as the following equation:
To explore the binding mechanisms, the configurations of
hUff N i – hUff ihN i hUsf N i – hUsf ihN i CO2 binding around functional groups and metal clusters
qst ¼ RT – – , (3) were analyzed after DFT optimizations. All the stable
hN 2 i – hN ihN i hN 2 i – hN ihN i
configurations and the corresponding binding energies
where N represents the number of adsorbed molecules, R (BEs) are listed in the Supplementary Material (as shown
represent the gas constant, and h i means the ensemble in Figs. S10–S18, cf. ESM). The name of each configura-
average. The first term is the contribution from molecular tion contains three terms. The first term is the functional
thermal energy, the second term is the contribution from group. The second term represents the initial binding site
the interaction energy among adsorbates Uff, and the third before optimizations (U: on the top of functional group; F:
term is the contribution from the adsorbent-adsorbate in front of functional group; R: on the right of functional
interaction energy Usf. The qst0 represents the adsorption group). The third term is the initial orientation before
heat in infinite dilution calculated by NVT simulations optimizations (V: vertical; C: cross; P: parallel). Figure 4
with a single CO2 molecule. In general, the adsorption heat shows several representative configurations of unfunctio-
shows a positive correlation with DFT binding energy, as nalized linkers. In configuration (–H)–F–V, the CO2
shown in Fig. S9 (cf. ESM). All the GCMC simulations molecule was located aside the benzene ring vertically
were performed with the Sorption code [31]. with a distance of 2.812 Å between H atom and C(CO2)
 Chenkai Gu et al. Effects of functional groups for CO2 capture using metal organic frameworks 441

atom. In configuration (–H)–U–C, the CO2 molecule was F–P, and (–NH2)–R–C from Fig. 5. Here, the configuration
lying over the plane of benzene, and the distance between (–SO3H)–F–P was taken as an example. In this configura-
H atom and C(CO2) atom is 2.971 Å. The highest binding tion, the CO2 molecule locates in the middle of the two
energies are –21.14 and –21.83 kJ/mol for (–H)–F–V and S = O bonds and pointing directly to the H atom, which
(–H)–U–C, respectively. ensured that the CO2 molecule can obtain interactions from
The following section is mainly about the configurations H atom and O atoms of –SO3H group simultaneously. The
of functionalized MOF-177 (given in Figs. 5 and 6). distance between O(CO2) and H(–SO3H) is 1.881 Å, which
Except the dispersion forces, we analyzed three kinds of denotes a relatively strong interaction. The H  O = C
interactions between CO2 and frameworks: (i) the hydro- angle is 157.9° within the range of hydrogen bond from
gen bond interactions between carbon atoms in CO2 120° to 180° [53]. The existence of hydrogen bond could
molecules and the hydrogen atoms in acidic groups or significantly improve the affinity of organic linker, which
amino groups; (ii) the electrostatic interactions between is also supported by the large binding energy of
lone pair electrons and the carbon atoms in CO2 molecules; –33.29 kJ/mol between CO2 molecule and organic linker
(iii) the induced dipolar interactions between halogen in configuration (–SO3H)–F–P.
atoms and the carbon atoms in CO2 molecules. Apart from hydrogen bond interaction, the electrostatic
Hydrogen bond interaction is a kind of strong interaction interactions between the lone pair electrons and the carbon
during the adsorption process and can be observed in atoms in quadrupole CO2 molecules also play an important
configurations of (–SO3H)–F–P, (–COOH)–U–C, (–OH)– role. The quadrupole-lonepair interactions can be observed

Fig. 3 LUMOs of linkers decorated with –CH3, –CN and –SO3H (C, gray; H, white; O, red; N, dark blue; S, yellow).

Fig. 4 DFT optimized structures of the CO2-linkers in MOF-177 (O, red spheres; C, gray spheres and H, white spheres).
442 Front. Chem. Sci. Eng. 2021, 15(2): 437–449

Fig. 5 DFT optimized structures of the CO2-linkers modified with –SO3H, –COOH, –OH, and –NH2 (S, yellow spheres; O, red spheres;
C, gray spheres; N, dark blue spheres; H, white spheres).

in configurations of (–COOH)–F–V, (–CN)–F–V, (–OH)– configuration (–F)–F–V is mainly different from config-
U–P, (–SO3H)–U–P, and (–NH2)–U–P as listed in Fig. 6. uration (–F)–F–C in terms of the initial orientations of CO2
Taking configuration (–COOH)–F–V as an example, the molecules before optimizations. The CO2 molecules tend
CO2 molecule is tilted in front of the –COOH group with a to stabilize with a final “vertical” orientation after
distance of 3.062 Å between O(–OH) and C(CO2). The optimization at the site in front of the fluorine atom.
O(–OH) is protruding outward and forming a quadrupole- Considering calculation errors, these two configurations
lonepair interaction with C(CO2). The binding energy here is can be regarded as the same terminal configurations, as
–26.43 kJ/mol. There are no hydrogen bond interactions in shown in Fig. S16 (cf. ESM).
this configuration, because the H(–COOH) is pointing away Based on the above analysis, the CO2 binding affinity of
from the CO2 molecule. organic linkers can be significantly enhanced owing to
In addition, the C–F  C = O interaction also contributes enlarged intermolecular interactions after introducing
to the CO2 binding affinity in a different way. In this functional groups. In some configurations, two different
configuration (–F)–F–C of Fig. 6, the fluorine atom is types of interactions can exist simultaneously. For
pointing to the C(CO2) atom with an approximately example, the (–CN)–F–V and (–F)–F–C configurations
orthogonal orientation. The value of characteristic angle contain both electrostatic and induce dipolar interactions.
is (C–F  C) = 173°. This suggests an induced dipolar Besides, hydrogen bond interactions and quadrupole-
interaction between fluorine atoms and the carbon atoms in lonepair interactions occurred simultaneously between
CO2 molecules [54]. The total binding energy between CO2 molecules and the organic linkers of MOF–177–
CO2 molecule and organic linker is –25.39 kJ/mol. The SO3H. Moreover, –SO3H groups contain more O atoms
 Chenkai Gu et al. Effects of functional groups for CO2 capture using metal organic frameworks 443

Fig. 6 DFT optimized structures of the CO2-linkers modified with –COOH, –CN, –OH, –SO3H, –NH2 and –F (S, yellow spheres; O, red
spheres; C, gray spheres; N, dark blue spheres; H, white spheres and F, cyan spheres).

than other functional groups. These O atoms contained tion of –SO3H can significantly improve the total sorbent–
lone pair electrons which can form stronger quadrupole- CO2 binding energy (BE) up to 4 times. The largest
lonepair interactions with C(CO2) atom. This is consistent contributions come from the electrostatic interactions
with its larger binding energy than the others. The energy between –SO3H group and CO2 molecules ( – 35.46
decomposition analyses were also performed to provide kJ/mol) which is much higher than those of the other
some insights of the interactions between CO2 and configurations. It suggests that MOF-177–SO3H is a
functionalized organic linker. The symmetry-adapted promising material in CO2 capture.
perturbation theory (SAPT) [55] were carried out with
MP2/6-31 ++ G(d,p) using Q-Chem computational suites 3.3 CO2 adsorption performance in functionalized MOFs
[56]. The functionalized benzene ring is chosen as the
basic sorbent model for energy decomposition analysis, To quantify the enhancement on the binding energies of
which has been treated as a useful method in previous CO2 with the functionalized MOFs, GCMC simulations
works [57–59]. The energy components including the were performed to calculate the CO2 adsorption properties
electrostatic energy (Ees), the exchange repulsion energy in all the functionalized MOF-177 at 298 K and pressure
(Eexch), the effective dispersion energy (Edisp) and the up to 100 kPa. Specifically, the CO2 capture capacity in
effective induction energy (Eind) are summarized in MOF-177–SO3H was calculated with the pressure up to
Table 3. Among all the functional groups, the incorpora- 4000 kPa. To confirm the reliability of the force field and
444 Front. Chem. Sci. Eng. 2021, 15(2): 437–449

Table 3 SAPT interaction energy decompositions (kJ/mol) between CO2 and functionalized linkers
Functioal
Configurations Total BE Ees Eexch Eind* Edisp*
groups

None –5.05 –11.48 18.89 –1.71 –10.75

–SO3H –10.43 –16.32 18.43 –1.60 –10.94

–SO3H –23.94 –35.46 29.96 –7.19 –11.26

–COOH –13.77 –18.79 16.11 –2.07 –9.02

–COOH –11.22 –22.20 30.09 –4.03 –15.09

–COOH –16.43 –21.76 16.64 –2.77 –8.53

–F –5.06 –8.42 11.21 –0.83 –7.01

–F –9.51 –12.70 10.96 –1.21 –6.57

–NH2 –12.92 –21.77 23.46 –2.35 –12.25

–OH –9.05 –15.43 19.00 –1.80 –10.81

–CH3 –7.25 –13.00 20.50 –1.79 –12.97

–CN –11.21 –14.85 10.35 –1.82 –4.89

–CN –11.96 –16.84 14.82 –2.09 –7.86

 Chenkai Gu et al. Effects of functional groups for CO2 capture using metal organic frameworks 445

the method for calculating atomic partial charges employed can significantly improve CO2 adsorption capacity, which
in this work, the adsorption isotherms of CO2 in MOF-177 is attributed to their relatively stronger polarity than that of
were simulated and compared with available experimental the others. Among these groups, the –SO3H makes the
data, as shown in Fig. 7. As can be seen from this figure, largest contribution and its CO2 adsorption amount is
the simulated adsorption isotherm showed good agreement 105.11 mg/g at 1 bar. This value is about 3 times higher
with experimental data [60], which proved the validity of than that of the parent MOF-177, larger than the ZIF-78
the force field. (91 mg/g) [61], comparable to the uptake on HKUST-1
(150 mg/g) [62] at the same condition. Because, hydrogen
bond interactions and quadrupole-lonepair interactions
occurred simultaneously in MOF-177–SO3H, which will
enhance the binding energy during the CO2 adsorption
process.
Figure 9 shows the contribution of these functional
groups, including CO2 adsorption amount as well as
isosteric heat increasement rate. The increased percentages
of CO2 adsorption amount at 1 kPa was defined as, (Nlater –
Npre)/Npre where the Nlater and Npre denote the amount of
CO2 molecules adsorbed in functionalized and unfunctio-
nalized MOF-177, respectively. The increased percentages
of isosteric heat at initial dilution was defined as
ðq0st – later – q0st – pre Þ=q0st – pre , where the q0st – later and q0st – pre
represent the isosteric heat in functionalized and unfunc-
Fig. 7 Comparison of simulated CO2 isotherms and experimen- tionalized MOF-177. The CO2 adsorption amount rises
tal data in MOF-177 up to 4000 kPa at 298 K.
with the increasement of isosteric heat. This indicated that
isosteric heat can affect the adsorption amount significantly
Figure 8 shows the gravimetric CO2 uptake amounts in at low pressure, and the adsorption amount can be enlarged
MOF-177 functionalized by different groups at low by improving the isosteric heat. In addition, the contribu-
pressure. We can observe an enhancement of CO2 tions of different functional groups are various owing to
adsorption amount for all functionalized MOF-177 their different internal characteristics. Among all func-
compared to unfunctionalized form, which indicated that tional groups, the –SO3H group shows the largest
the substitution of functional groups is one of the improvement on CO2 adsorption amount. The increased
approaches to increase the CO2 adsorption amount in percentage of CO2 adsorption amount is much higher than
MOFs at low pressure. The order of the CO2 loading that of –CH3 group and reach up to 176% at 1 kPa. This
amount is shown as the following: –SO3H > –COOH > value is 32 times higher than that of –CH3 (only about
–NH2 > –OH > –CN > –F > –CH3. This order is consis- 5.5%), which demonstrates the potential of –SO3H group
tent with the order of adsorption heats in infinite dilution in improving CO2 adsorption capability. Therefore, we
(qst0 in Table 1). Therefore, it was indicated that the further investigated the CO2 adsorption in MOF-177–
adsorption heat plays an important role on CO2 adsorption SO3H at pressure up to 4000 kPa.
in MOFs at low pressure. The –SO3H, –COOH and –NH2

Fig. 9 Contribution of functional groups: yellow columns

represent increased percentages of isosteric heat at initial dilution;
Fig. 8 CO2 adsorption isotherms of MOF-177 and its functio- blue columns represent increased percentages of CO2 adsorption
nalized derivatives at low pressure. amount at 1 kPa.
446 Front. Chem. Sci. Eng. 2021, 15(2): 437–449

Figure 10 shows the gravimetric and volumetric CO2 linkers with the increasing pressure, which indicated metal
uptake amounts in MOF-177 and MOF-177–SO3H at sites were the prior adsorption sites. This is also supported
pressure up to 4000 kPa. At low pressure, both gravimetric by the higher CO2 binding energy on metal cluster than
and volumetric CO2 adsorption amounts in MOF-177– that on BTB linker, as shown in Figs. S17 and S18 (cf.
SO3H are higher than that in parent MOF-177 owing to the ESM). For the MOF-177–SO3H, the density of CO2
larger isosteric heat (Table 1). At high pressure, MOF-177– molecules is higher than that in the non-modified form.
SO3H shows lower gravimetric CO2 adsorption capacity, Moreover, the CO2 molecules prefer to the sites around the
which can be mainly attributed to both lower crystal metal clusters and –SO3H groups, which indicates that the
density and total free volume (as shown in Table 1). binding affinity of –SO3H can significantly improve the
However, the volumetric CO2 adsorption capacity of adsorption performance in CO2 adsorption. With the
MOF-177–SO3H is higher than that of MOF-177 with the increasing pressure, the CO2 molecule turned to occupy
increasing pressure. This is because the MOF-177–SO3H the whole pore space in MOF–SO3H. It is also observed
has relatively higher volumetric surface area. that the incorporation of –SO3H groups can lead to the
Figure 11 shows the probability distributions of COM of decrease of CO2 binding energy on metal clusters, as
CO2 in MOF-177 and MOF-177–SO3H at 298 K with 3 shown in Fig. S18. However, this negative effect was offset
different pressures: 1, 100 and 4000 kPa. In the case of by the improved CO2 affinity of organic linkers modified
MOF-177, the CO2 molecules were firstly adsorbed around by –SO3H groups and an increased CO2 adsorption
the metal clusters and then gradually adsorbed at the BTB amount was observed especially at low pressure. It should

Fig. 10 CO2 adsorption isotherms in MOF-177 and MOF-177–SO3H at 298 K with pressure up to 4000 kPa: (a) gravimetric adsorption
isotherms and (b) volumetric adsorption isotherms.

Fig. 11 Probability distribution plots of COM of CO2 in MOF-177 and MOF-177–SO3H (Zn, light blue spheres; C, gray spheres; H,
white spheres; O, red spheres; S, yellow spheres).
 Chenkai Gu et al. Effects of functional groups for CO2 capture using metal organic frameworks 447

be noted that the CO2 capture of MOFs will also be piperazine blended solutions. International Journal of Greenhouse
affected by pore sizes or topologies when these functional Gas Control, 2016, 50: 135–146
groups are introduced to other MOFs [63]. Our findings 5. Rochelle G T. Amine scrubbing for CO2 capture. Science, 2009, 325
may not apply to all of the MOFs since we only chose (5948): 1652–1654
MOF-177 as a parent material to compare the CO2 binding 6. Lin Y. Metal organic framework membranes for separation
energies of different functional groups. applications. Current Opinion in Chemical Engineering, 2015, 8:
21–28
7. Lin J Y. Molecular sieves for gas separation. Science, 2016, 353
4 Conclusions (6295): 121–122
8. Wu Y, Chen X, Fan M, Jiang G, Kong Y, Bland A E. Development
Due to the functionalization effect, the atomic charge of K and N based composite CO2 sorbents (KN) dried with a
distributions on the linker of MOF-177 were significantly supercritical fluid. Chemical Engineering Journal, 2015, 262: 1192–
intensified according to LUMO analysis. Three different 1198
intermolecular interactions were found between CO2 9. Yang Q, Zhong C, Chen J F. Computational study of CO2 storage in
molecules and functional groups, including hydrogen metal-organic frameworks. Journal of Physical Chemistry C, 2008,
bond interaction, quadrupole-lonepair interaction and 112(5): 1562–1569
induced dipolar interaction. It is beneficial to incorporate 10. Yang S, Liu Z, Yan X, Liu C, Zhang Z, Liu H, Chai L. Catalytic
groups containing lone pair electrons, acidic hydrogens or oxidation of elemental mercury in coal-combustion flue gas over the
halogen atoms to induce forming the interactions men- CuAlO2 catalyst. Energy & Fuels, 2019, 33(11): 11380–11388
tioned above to improve the CO2 affinity of linkers. 11. Yang H, Xu Z, Fan M, Gupta R, Slimane R B, Bland A E, Wright I.
Moreover, a trend of adsorption ability at low pressure was Progress in carbon dioxide separation and capture: a review. Journal
formed as following: –SO 3 H > –COOH > –NH 2 > of Environmental Sciences (China), 2008, 20(1): 14–27
–OH > –CN > –F > –CH 3 . Among these functional 12. Figueroa J D, Fout T, Plasynski S, McIlvried H, Srivastava R D.
groups, the –SO3H group brought the strongest promotion Advances in CO2 capture technology—the US department of
to CO2 adsorption, and was further studied. The CO2 energy’s carbon sequestration program. International Journal of
adsorption capacity of MOF-177 at high pressure declines Greenhouse Gas Control, 2008, 2(1): 9–20
as a result of decreasing total free volume after introducing 13. Liu H, Xie X, Chen H, Yang S, Liu C, Liu Z, Yang Z, Li Q, Yan X.
–SO3H group. However, it still shows a remarkable SO2 promoted ultrafine nano-sulfur dispersion for efficient and
improvement on CO2 capture at the pressure less than stable removal of gaseous elemental mercury. Fuel, 2020, 261:
1000 kPa, and thus has the potential to be a promising CO2 116367
capture material in this pressure range. 14. Zhang Z, Yao Z Z, Xiang S, Chen B. Perspective of microporous
metal-organic frameworks for CO2 capture and separation. Energy
Acknowledgements This work was supported by the National Natural & Environmental Science, 2014, 7(9): 2868–2899
Science Foundation of China (Grant Nos. 51676079 and 21773104), 15. Gu C, Liu J, Hu J, Wu D. Highly selective separations of C2H2/C2H4
Fundamental Research Funds for the Central Universities (No. 2019kfyR-
and C2H2/C2H6 in metal-organic frameworks via pore environment
CPY021) and China Scholarship Council (No. 201906160014).
design. Industrial & Engineering Chemistry Research, 2019, 58(43):
Electronic Supplementary Material Supplementary material is available 19946–19957
in the online version of this article at https://doi.org/10.1007/s11705-020- 16. Hu J, Liu Y, Liu J, Gu C. Chelation of transition metals into MOFs
1961-6 and is accessible for authorized users. as a promising method for enhancing CO2 capture: a computational
study. AIChE Journal. American Institute of Chemical Engineers,
2020, 66(2): e16835
References 17. Li S, Chung Y G, Simon C M, Snurr R Q. High-throughput
computational screening of multivariate metal–organic frameworks
1. Zhao C, Chen X, Anthony E J, Jiang X, Duan L, Wu Y, Dong W, (MTV-MOFs) for CO2 capture. Journal of Physical Chemistry
Zhao C. Capturing CO2 in flue gas from fossil fuel-fired power Letters, 2017, 8(24): 6135–6141
plants using dry regenerable alkali metal-based sorbent. Progress in 18. An J, Rosi N L. Tuning MOF CO2 adsorption properties via cation
Energy and Combustion Science, 2013, 39(6): 515–534 exchange. Journal of the American Chemical Society, 2010, 132
2. Wang M, Liu J, Shen F, Cheng H, Dai J, Long Y. Theoretical study (16): 5578–5579
of stability and reaction mechanism of CuO supported on ZrO2 19. Pal A, Chand S, Das M C. A water-stable twofold interpenetrating
during chemical looping combustion. Applied Surface Science, microporous MOF for selective CO2 adsorption and separation.
2016, 367: 485–492 Inorganic Chemistry, 2017, 56(22): 13991–13997
3. Darensbourg D J, Chung W C, Wang K, Zhou H C. Sequestering 20. Zheng S T, Bu J T, Li Y, Wu T, Zuo F, Feng P, Bu X. Pore space
CO2 for short-term storage in MOFs: copolymer synthesis with partition and charge separation in cage-within-cage indium-organic
oxiranes. ACS Catalysis, 2014, 4(5): 1511–1515 frameworks with high CO2 uptake. Journal of the American
4. Yu J, Wang S, Yu H. Experimental studies and rate-based Chemical Society, 2010, 132(48): 17062–17064
simulations of CO2 absorption with aqueous ammonia and 21. Tanabe K K, Cohen S M. Postsynthetic modification of metal-
448 Front. Chem. Sci. Eng. 2021, 15(2): 437–449

organic frameworks—a progress report. Chemical Society Reviews, Energy and Combustion Science, 2020, 79: 100844
2011, 40(2): 498–519 37. Wang Z, Liu J, Yang Y, Liu F, Ding J. Heterogeneous reaction
22. Hu J, Liu Y, Liu J, Gu C. Effects of water vapor and trace gas mechanism of elemental mercury oxidation by oxygen species over
impurities in flue gas on CO2 capture in zeolitic imidazolate MnO2 catalyst. Proceedings of the Combustion Institute, 2019, 37
frameworks: the significant role of functional groups. Fuel, 2017, (3): 2967–2975
200: 244–251 38. Wang Z, Liu J, Yang Y, Yu Y, Yan X, Zhang Z. AMn2O4 (A = Cu,
23. Xiang Z, Leng S, Cao D. Functional group modification of metal- Ni and Zn) sorbents coupling high adsorption and regeneration
organic frameworks for CO2 capture. Journal of Physical Chemistry performance for elemental mercury removal from syngas. Journal of
C, 2012, 116(19): 10573–10579 Hazardous Materials, 2020, 388: 121738
24. Zheng B, Bai J, Duan J, Wojtas L, Zaworotko M J. Enhanced CO2 39. Ikeda A, Nakao Y, Sato H, Sakaki S. Binding energy of transition-
binding affinity of a high-uptakerht-type metal-organic framework metal complexes with large p-conjugate systems. Density func-
decorated with acylamide groups. Journal of the American Chemical tional theory vs post-Hartree-Fock methods. Journal of Physical
Society, 2010, 133(4): 748–751 Chemistry A, 2007, 111(30): 7124–7132
25. An J, Geib S J, Rosi N L. High and selective CO2 uptake in a cobalt 40. Ramsahye N, Maurin G, Bourrelly S, Llewellyn P, Serre C, Loiseau
adeninate metal-organic framework exhibiting pyrimidine-and T, Devic T, Ferey G. Probing the adsorption sites for CO2 in metal
amino-decorated pores. Journal of the American Chemical Society, organic frameworks materials MIL-53 (Al, Cr) and MIL-47(V) by
2009, 132(1): 38–39 density functional theory. Journal of Physical Chemistry C, 2008,
26. Ye Y, Zhang H, Chen L, Chen S, Lin Q, Wei F, Zhang Z, Xiang S. 112(2): 514–520
Metal-organic framework with rich accessible nitrogen sites for 41. Delley B. From molecules to solids with the DMol3 approach.
highly efficient CO2 capture and separation. Inorganic Chemistry, Journal of Chemical Physics, 2000, 113(18): 7756–7764
2019, 58(12): 7754–7759 42. Lee T B, Kim D, Jung D H, Choi S B, Yoon J H, Kim J, Choi K,
27. Hu J, Liu Y, Liu J, Gu C, Wu D. High CO2 adsorption capacities in Choi S H. Understanding the mechanism of hydrogen adsorption
UiO type MOFs comprising heterocyclic ligand. Microporous and into metal organic frameworks. Catalysis Today, 2007, 120(3-4):
Mesoporous Materials, 2018, 256: 25–31 330–335
28. Liu Y, Liu J, Chang M, Zheng C. Effect of functionalized linker on 43. Wang Z, Liu J, Yang Y, Yu Y, Yan X, Zhang Z. Insights into the
CO2 binding in zeolitic imidazolate frameworks: density functional catalytic behavior of LaMnO3 perovskite for Hg0 oxidation by HCl.
theory study. Journal of Physical Chemistry C, 2012, 116(32): Journal of Hazardous Materials, 2020, 383: 121156
16985–16991 44. Potoff J J, Siepmann J I. Vapor-liquid equilibria of mixtures
29. Liu Y, Liu J, Chang M, Zheng C. Theoretical studies of CO2 containing alkanes, carbon dioxide, and nitrogen. AIChE Journal.
adsorption mechanism on linkers of metal–organic frameworks. American Institute of Chemical Engineers, 2001, 47(7): 1676–1682
Fuel, 2012, 95: 521–527 45. Rappé A K, Casewit C J, Colwell K, Goddard W III, Skiff W. UFF, a
30. Zhang Y B, Furukawa H, Ko N, Nie W, Park H J, Okajima S, full periodic table force field for molecular mechanics and molecular
Cordova K E, Deng H, Kim J, Yaghi O M. Introduction of dynamics simulations. Journal of the American Chemical Society,
functionality, selection of topology, and enhancement of gas 1992, 114(25): 10024–10035
adsorption in multivariate metal-organic framework-177. Journal 46. Momany F A. Determination of partial atomic charges from ab initio
of the American Chemical Society, 2015, 137(7): 2641–2650 molecular electrostatic potentials. Application to formamide,
31. Accelrys Software Inc. Materials Studio Release Notes, release 4.4. methanol, and formic acid. Journal of Physical Chemistry, 1978,
Accelrys Software Inc.: San Diego, CA, 2008 82(5): 592–601
32. Yang Q, Vaesen S, Ragon F, Wiersum A D, Wu D, Lago A, Devic T, 47. Campañá C, Mussard B, Woo T K. Electrostatic potential derived
Martineau C, Taulelle F, Llewellyn P L, et al. A water stable metal- atomic charges for periodic systems using a modified error
organic framework with optimal features for CO2 capture. functional. Journal of Chemical Theory and Computation, 2009, 5
Angewandte Chemie International Edition, 2013, 52(39): 10316– (10): 2866–2878
10320 48. Argueta E, Shaji J, Gopalan A, Liao P, Snurr R Q, Gómez-Gualdrón
33. Zhou Y X, Chen Y Z, Hu Y, Huang G, Yu S H, Jiang H L. MIL-101- D A. Molecular building block-based electronic charges for high-
SO3H: a highly efficient Brønsted acid catalyst for heterogeneous throughput screening of metal-organic frameworks for adsorption
alcoholysis of epoxides under ambient conditions. Chemistry applications. Journal of Chemical Theory and Computation, 2018,
(Weinheim an der Bergstrasse, Germany), 2014, 20(46): 14976– 14(1): 365–376
14980 49. Manz T A, Sholl D S. Chemically meaningful atomic charges that
34. Sarkisov L, Harrison A. Computational structure characterisation reproduce the electrostatic potential in periodic and nonperiodic
tools in application to ordered and disordered porous materials. materials. Journal of Chemical Theory and Computation, 2010, 6
Molecular Simulation, 2011, 37(15): 1248–1257 (8): 2455–2468
35. Gu C, Liu J, Hu J, Wu D. Metal-organic frameworks chelated by 50. Cornell W D, Cieplak P, Bayly C I, Gould I R, Merz K M, Ferguson
zinc fluorides for ultra-high affinity to acetylene during C2/C1 D M, Spellmeyer D C, Fox T, Caldwell J W, Kollman P A. A second
separations. Fuel, 2020, 266: 117037 generation force field for the simulation of proteins, nucleic acids,
36. Yang Y, Liu J, Wang Z. Reaction mechanisms and chemical kinetics and organic molecules. Journal of the American Chemical Society,
of mercury transformation during coal combustion. Progress in 1995, 117(19): 5179–5197
 Chenkai Gu et al. Effects of functional groups for CO2 capture using metal organic frameworks 449

51. Torrisi A, Bell R G, Mellot-Draznieks C. Predicting the impact of 58. Gu C, Liu Y, Liu J, Hu J, Wang W. Ab initio study of gas adsorption
functionalized ligands on CO2 adsorption in MOFs: a combined in metal-organic frameworks modified by lithium: the significant
DFT and Grand Canonical Monte Carlo study. Microporous and role of Li-containing functional groups. Journal of Physical
Mesoporous Materials, 2013, 168: 225–238 Chemistry C, 2018, 122(32): 18395–18404
52. Gu C, Liu J, Hu J, Wu D. Highly efficient separations of C2H2 from 59. Gu C, Liu J, Hu J, Wang W. Metal-organic frameworks grafted by
C2H2/CO and C2H2/H2 in metal-organic frameworks with ZnF2 univariate and multivariate heterocycles for enhancing CO2 capture:
chelation: a molecular simulation study. Fuel, 2020, 271: 117598 a molecular simulation study. Industrial & Engineering Chemistry
53. Steiner T, Desiraju G R. Distinction between the weak hydrogen Research, 2019, 58(6): 2195–2205
bond and the van der Waals interaction. Chemical Communications, 60. Millward A R, Yaghi O M. Metal-organic frameworks with
1998, (8): 891–892 exceptionally high capacity for storage of carbon dioxide at room
54. Paulini R, Müller K, Diederich F. Orthogonal multipolar interac- temperature. Journal of the American Chemical Society, 2005, 127
tions in structural chemistry and biology. Angewandte Chemie (51): 17998–17999
International Edition, 2005, 44(12): 1788–1805 61. Phan A, Doonan C J, Uribe-Romo F J, Knobler C B, O’Keeffe M,
55. Jeziorski B, Moszynski R, Szalewicz K. Perturbation theory Yaghi O M. Synthesis, structure, and carbon dioxide capture
approach to intermolecular potential energy surfaces of van der properties of zeolitic imidazolate frameworks. Accounts of
Waals complexes. Chemical Reviews, 1994, 94(7): 1887–1930 Chemical Research, 2010, 43(1): 58–67
56. Shao Y, Gan Z, Epifanovsky E, Gilbert A T B, Wormit M, 62. Chowdhury P, Bikkina C, Meister D, Dreisbach F, Gumma S.
Kussmann J, Lange A W, Behn A, Deng J, Feng X, et al. Advances Comparison of adsorption isotherms on Cu-BTC metal organic
in molecular quantum chemistry contained in the Q-Chem 4 frameworks synthesized from different routes. Microporous and
program package. Molecular Physics, 2015, 113(2): 184–215 Mesoporous Materials, 2009, 117(1-2): 406–413
57. Fioretos K A, Psofogiannakis G M, Froudakis G E. Ab-initio study 63. Anderson R, Rodgers J, Argueta E, Biong A, Gómez-Gualdrón D A.
of the adsorption and separation of NOx and SOx gases in Role of pore chemistry and topology in the CO2 capture capabilities
functionalized IRMOF ligands. Journal of Physical Chemistry C, of MOFs: from molecular simulation to machine learning.
2011, 115(50): 24906–24914 Chemistry of Materials, 2018, 30(18): 6325–6337