Microporous and Mesoporous Materials 304 (2020) 110377

Contents lists available at ScienceDirect

Microporous and Mesoporous Materials
journal homepage: />
Carbon dioxide capacity retention on elastic layered metal organic
frameworks subjected to hydrothermal cycling
Francisco J. Sotomayor b, Christian M. Lastoskie a, *
a
b

Department of Civil and Environmental Engineering, University of Michigan, 1351 Beal Avenue, Ann Arbor, MI, 48109-2125, USA
Quantachrome Instruments, 1900 Corporate Drive, Boynton Beach, FL, 33426, USA

A R T I C L E I N F O

A B S T R A C T

Keywords:
Flexible adsorbent
Water vapor
Density functional theory
Isotherm
Regeneration

Adsorption of carbon dioxide on elastic layered metal-organic frameworks (ELMs) was investigated during and
after exposure to water. Two ELM variants, ELM-11 and ELM-12, were contacted with water vapor and the
impact of cyclical exposure on the CO2 capacity of the adsorbents was observed. ELM-11 was found to lose CO2
capacity with each successive exposure to water, whereas ELM-12 retained CO2 capacity through four exposure
cycles. Density functional theory calculations were performed to interpret these observations. Changing the

counter-ion from the simple tetrafluoroborate (BFÀ4 ) to the larger and more complex trifluoromethanesulfonate
(CF3SOÀ3 ) anion expands the number of potential binding sites for adsorbate molecules. While CO2 directly
competes with other adsorbates for binding sites in ELM-11, CO2 does not directly compete with other adsorbates
in ELM-12 due to its preference for direct interaction with both fluorine and oxygen atoms in CF3SOÀ3 .

1. Introduction
Metal-organic frameworks (MOFs), also known as porous coordina­
tion polymers (PCPs), are a novel class of hybrid materials assembled
from metal ions with well-defined coordination geometry and organic
bridging ligands [1]. Through careful selection of the metal and organic
building blocks, MOFs can be conceptually designed into networks
possessing finely tuned pore size and crystal structure. Over 20,000
different MOFs have been reported within the past decade [2]. The
structural and chemical diversity of MOFs has fostered extensive
research into their potential applications for gas storage, ion exchange,
molecular separation, and heterogeneous catalysis [3]. Notably, the
exceptional tunability of MOFs has made possible the synthesis of
structures with record-breaking porous material properties such as
surface area and capacity for the storage of hydrogen, methane, or
carbon dioxide by physical adsorption [4]. The large surface areas,
adjustable pore sizes, and controllable surface properties of certain
MOFs make them attractive prospective adsorbents for CO2 capture
from mixed gas streams.
Flexible MOFs, also known as soft porous crystals (SPCs) [5], are a
subset of MOFs that possess both a highly ordered network and struc­
tural transformability. In contrast with rigid MOFs, which retain their

structure and their porosity irrespective of environmental factors, SPCs
undergo structural transformations that are dependent on external
stimuli such as temperature, mechanical pressure, or guest adsorption,

on account of their bi-stable or multi-stable attributes [6]. This facet of
SPCs has led to the observation of previously unanticipated gas
adsorption phenomena. A subset of SPCs are the so-called elastic layered
metal-organic frameworks (ELMs) [7,8]. ELMs are composed of metal
vertex ions, connecting ligands, and charge-balancing counter-ions that
are arranged into two-dimensional sheets that in turn assemble into
three-dimensional stacked structures. These materials show a latent
porosity [9] for the adsorption of gas molecules above a specific pres­
sure, termed the “gate pressure”, that results in an expansion of the
interlayer spacing between the two-dimensional sheets and a corre­
sponding jump in the adsorption isotherm that cannot be classified in
accordance with the conventional IUPAC isotherm designations. The
exotic adsorption characteristics of ELMs are not observed in more
familiar commercial adsorbents such as activated carbons or zeolites.
Nor are they found in the adsorption isotherms of MOFs with rigid pore
structures. These unusual features confer upon ELMs potential advan­
tages for CO2 capture, inasmuch as they combine a high selectivity for
separation of CO2 from gas mixtures with a low energy requirement for
adsorbent regeneration and CO2 recovery [8].

DOI of original article: />* Corresponding author.
E-mail address: (C.M. Lastoskie).
/>Received 19 June 2018; Received in revised form 14 February 2019; Accepted 11 March 2019
Available online 17 June 2020
1387-1811/© 2019 Elsevier Inc. All rights reserved.


F.J. Sotomayor and C.M. Lastoskie

Microporous and Mesoporous Materials 304 (2020) 110377


However, to be suitable for post-combustion carbon capture (PCC),
prospective adsorbents must selectively adsorb CO2 at low concentra­
tion (4–15 vol%) in the presence of other flue gas constituents. Coal
combustion flue gas, for example, typically contains 5 to 7 vol% water
vapor and tens to hundreds of parts per million of SOx, NOx, and CO
[10], any of which may significantly impact the CO2 capture perfor­
mance and stability of MOFs [11]. In addition, because PCC systems
often assume regeneration of the adsorbent and recovery of the captured
CO2, prospective CO2 capture materials need to perform without ca­
pacity fade through many adsorption and regeneration cycles.
Regeneration of a solid adsorbent is typically accomplished by
temperature swing adsorption (TSA), pressure swing adsorption (PSA),
vacuum swing adsorption (VSA), or some combination of these pro­
cesses [10]. Because of the availability of low-grade waste heat from a
power plant as an energy source for regeneration, TSA is considered
particularly promising for many carbon capture operations. For their use
in TSA recovery, ELMs must demonstrate consistent and reproducible
CO2 capture performance after repeated exposure to unwanted gas
components, thermal stresses, moisture, and trace components of flue
gas.
Many rigid MOFs have relatively high thermal stability. For example,
room temperature CO2 adsorption on Zn4O(bdc)3 (bdc ¼ 1,4-benzene­
dicarboxylate), otherwise known as MOF-5, remains near 3.6 wt% when
the adsorbent is cycled between 30 and 300 � C at atmospheric pressure
[4]. Only above 400 � C does MOF-5 undergo thermal decomposition and
lose its capability to retain CO2. Conversely, flexible MOFs, by defini­
tion, have crystalline structures that are more susceptible to flexing and
distortion upon exposure to external stimuli, leading to the question of
whether the weaker inter-framework interactions characteristic of

flexible MOFs such as ELMs reduces their thermal stability.
Adsorption-desorption cycling of CH4 on ELM-11 at 303 K showed no
degradation of the gated adsorption capacity even after 50 cycles [8].
Thermogravimetric analysis revealed that ELM-11 structure begins to
lose bipyridine and BF4 molecules at around 500 K [12]. The findings
from these two experiments suggest that ELM-11 can be degassed at
temperatures up to approximately 200 � C with no structural degrada­
tion. ELM flexible framework adsorbents are thus expected to have
thermal cycling stability comparable to that of rigid MOFs.
Certain MOFs are well known to be structurally unstable in contact
with water [13]. For example, Cu3(btc)2 (btc ¼ 1,3,5-benzene tri­
carboxylate), also known as HKUST-1, is stable in dry air at room tem­
perature, but its crystallinity progressively declines upon cyclic
exposure to water vapor from air at 30% relative humidity, plateauing at
75% of its original crystallinity after repeated water cycling. The MOF
Ni/dobdc loses CO2 capacity after repeated H2O/CO2 mixture isotherm
measurements. Zn2(bdc)2(dabco) and Ni2(bdc)2(dabco) (dabco ¼ 1,
4-diazabicyclo[2.2.2]octane) are stable after O2 adsorption at 25 � C
from air at 30% relative humidity, but collapse upon exposure to air at
60% relative humidity at the same temperature. Kizzie and coworkers
[14] investigated the effect of humidity on CO2 capture in the M/dobdc
series (where M ¼ Zn, Ni, Co, or Mg; dobdc ¼ 2,5-dioxidobenzene-1,
4-dicarboxylate). They found that although Mg/dobdc had the highest
initial CO2 capacity at the conditions used in their study, exposure to air
at a relative humidity of 70% followed by thermal regeneration resulted
in the retention of only 16% of the initial CO2 capacity. In contrast, 85%
of the CO2 capacity in Co/dobdc was retained under the same condi­
tions. It is evident then that water vapor can both irreversibly after MOF
structures and hinder their adsorption of CO2.
For ELM class materials, Cheng et al. [12] studied the evolution of

the structure of ELM-11 upon a single cycle of dehydration and rehy­
dration, and reported that the CO2 adsorption capacity was largely
preserved after exposure to water vapor. The slight differences observed
in the CO2 adsorption isotherms before and after water vapor exposure
were attributed to stacking faults within the ELM-11 framework. This
study however considered only a single cycle of exposure to moisture,
and did not investigate the stability of other ELM analogues to water.

Considering that other MOFs show significant performance loss
under cyclic exposure to water, and that the substitution of different
framework components can have a significant effect on hydrothermal
stability, the work presented herein seeks to better understand the CO2
capacity retention of ELM variants after water vapor exposure through a
combination of isotherm measurements after water vapor cycling and
density functional theory calculations.
2. Experimental
Material preparation. Two isostructural ELM variants were experi­
mentally tested: Cu(bpy)2(BF4)2, termed ELM-11, and Cu(bpy)2(OTf)2,
termed ELM-12, (where bpy ¼ 4,40 -bipyridine and OTf ¼ CF3SOÀ3 ). Two
methods were used to obtain samples of ELM-11. The first method was
the purchase of the un-activated precursor of ELM-11, [Cu(bpy)
(BF4)2(H2O)2]bpy, termed pre-ELM-11, sold commercially by Tokyo
Chemical Industry Co., Ltd. (CAS Number: 854623-98-6, Product
Number: C2409) at >98% purity. The second method was the synthesis
of pre-ELM-11 following the procedure reported by Tran [15]. 4,
40 -bipyridine (0.312 g; 2 mmol) in 2 mL of ethanol was slowly added to
an 8-mL aqueous solution of Cu(BF4)2⋅H2O (0.309 g; 1 mmol) at room
temperature. A blue precipitate gradually formed. The mixture was
stirred for 4 h at room temperature, after which the solid was allowed to
settle for two days and then filtered off, washed with water and ethanol,

and dried in air at room temperature. Once pre-ELM-11 is obtained, it is
easily converted to ELM-11 by degassing under vacuum (<10 μmHg) at
403 K for 2 h.
ELM-12 was synthesized using a modified version of the procedure
reported by Kondo et al. [16]. 0.2 g of Cu(OTf)2 was dissolved in 10 mL
deionized water. 0.2 g of 4,40 -bipyridine (bpy) was dissolved in 10 mL
ethanol. To the copper solution, 2 mL methanol was added drop-wise
close to the surface to slow the diffusion-precipitation process. Then
the bpy solution was added drop-wise in the same manner. The solution
was sealed from the atmosphere and allowed to sit undisturbed for two
weeks. The resulting blue microcrystals were rinsed with ethanol,
drained, and air-dried. The powder was then degassed under vacuum
(<10 μmHg) at 403 K for 2 h.
The ELM adsorbents were characterized using X-ray powdered
diffraction (XRD), infrared (IR) spectra, and adsorption isotherm mea­
surements. A Rigaku MiniFlex 600 and a PerkinElmer Spectrum BX FTIR spectrometer were used respectively to collect X-ray powdered dif­
fractions and IR spectra. Adsorption isotherms were measured using a
Micromeritics ASAP 2050 extended pressure volumetric adsorption
analyzer. Ultra-high purity (99.99% or higher) N2, He, and CO2 gases
were used as adsorbates. The temperature in the sample tube was
controlled by an external bath.
Adsorption measurements. To assess the impact of water exposure on
the CO2 capture performance of the ELM adsorbents, cyclic exposure
experiments were performed. The adsorbent, either ELM-11 or ELM-12,
was activated by degassing under vacuum (<10 μmHg) at 403 K for at
least 2 h. A CO2 adsorption isotherm was then measured on the activated
adsorbent to determine its baseline CO2 adsorption capacity. For ELM11, CO2 isotherms were obtained at 273 K to a final pressure
exceeding 1 bar, in keeping with the procedure following by Cheng et al.
[12]. For ELM-12, CO2 isotherms were measured at 298 K to a final
pressure exceeding 3 bar. The ELM adsorbent was then transferred to a

Quantachrome Autosorb-1, where it was regenerated by degassing
under vacuum (<10 μmHg) at 403 K for at least 2 h. An adsorption
isotherm was then measured for pure water vapor at 298 K up to a
relative pressure (P/P0) of 0.8. After exposure to water vapor, the
adsorbent was transferred back to the Micromeritics ASAP 2050, where
it was again regenerated. A second CO2 adsorption isotherm was then
measured. This procedure was repeated three times for each ELM
adsorbent, with four cycles of alternating uptake measurements for CO2
and water.
Density functional theory calculations. To interpret the effect of water
2


F.J. Sotomayor and C.M. Lastoskie

Microporous and Mesoporous Materials 304 (2020) 110377

vapor on the adsorption of CO2 in ELMs, density functional theory (DFT)
calculations were performed. DFT is a quantum mechanical modeling
method used to investigate the electronic structure of many-body sys­
tems. Recent studies have used DFT to investigate the impact of ligands
on CO2 adsorption [17,18], to identify the favorable adsorption sites of
various gas molecules on different MOFs [19–21], as well as adsorption
induced transformations of a flexible MOF [22]. All DFT calculations
were carried out using the Gaussian 09 software package [23]. The
6-31G(d) basis set was used for all atoms except the transition metal Cu
(II), for which the commonly used “LANL2DZ” basis set was chosen [24].
The B3LYP hybrid functional was chosen for use after comparison with
the M06 and M06–2X functionals and with experimental data collected
here and reported elsewhere by various authors [7,25–28].

Model clusters. The ELM-11 unit cluster was constructed from X-ray
powder diffraction data as reported by Kondo et al. [28] and obtained
from the Cambridge crystallographic data center. The representative
cluster for ELM-11 is made up of a single copper atom and its associated
ligands, with the 4,40 -bipyridine (bpy) linkers replaced by more compact
pyridine rings to limit end effects. Atomic positions and bond lengths
were optimized using the geometric optimization option with default
convergence criteria. The ELM-12 unit cluster was similarly constructed
from X-ray powder diffraction data as reported by Kondo et al. [29] and
obtained from the Cambridge crystallographic data center. The
DFT-optimized geometries for ELM-11 and ELM-12 are shown in Fig. 1.
Calculation of binding energies. To determine the binding energies of
adsorbate species on the ELM clusters, atomic coordinates before and
during gas adsorption were first determined using the geometric opti­
mization feature in Gaussian 09 with default convergence criteria. The
coordinates of the ELM cluster and adsorbing gas molecules were not
held fixed during geometry optimizations. For ELM-11, at least two
different starting positions were used for CO2 and H2O to assure the final
optimized positions represent global, not local, lowest energy geome­
tries. For ELM-12, at least three different starting positions were used for
each adsorbate molecule.

3. Results and discussion
Water vapor cycling experiments. Fig. 2 shows the results of cyclic
water vapor exposure on the CO2 isotherms of ELM-11 (A-1) and ELM-12
(A-2). As previously noted, ELM-11 retains its CO2 capacity during
adsorption/desorption cycles when moisture is not present. When
exposed to water vapor between CO2 isotherm measurements, however,
ELM-11 progressively loses CO2 adsorption capacity with each succes­
sive cycle, with a complete loss of CO2 capacity and gated adsorption by

the fifth CO2 uptake measurement. In contrast, ELM-12 demonstrates
resiliency in both its CO2 capacity and the features of its CO2 and water
vapor isotherms, with minimal variations in the adsorption branches of
the isotherms after four complete cycles of water vapor exposure.
Capacity loss of ELM-11. The observed degradation of ELM-11
adsorption capacity is unexpected based on a previous report by
Cheng et al. [12] that concluded that the latent porosity of the dewa­
tered ELM-11 structure could be mostly recovered after water vapor
adsorption. However, a full comparison of the experimental results re­
quires examination not only of the CO2 isotherms but also the water
vapor isotherms. In the Cheng et al. study, the authors reported a water
vapor isotherm in which the maximum capacity approached 65 mg
H2O/g at 303 K. This is significant in that 65 mg/g is equivalent to one
water molecule per BF4 counter-ion (or two water molecules per copper
ion), identical to the ratio present in the pre-ELM-11 structure, sug­
gesting that water “adsorption” is actually the re-incorporation of H2O
into the crystal structure as ELM-11 reverts to pre-ELM-11. In situ XRD
studies by Cheng et al. indicated that the ELM-11 structure returned to
pre-ELM-11 during water vapor exposure, in support of this
interpretation.
By contrast, for the ELM-11 water vapor adsorption isotherms shown
in Fig. 2 (panel B-1), the amount of water adsorbed was 40 mg/g for the
first cycle and did not exceed 55 mg/g in cycles thereafter. This suggests
the ELM-11 adsorbent used in the cycling experiments conducted in this
study was only partially hydrated to the pre-ELM-11 structure. There
was only a slight difference in the temperature of the water vapor
isotherm measurement in this work (298 K) versus that of Cheng et al.
(303 K). So the lower ELM-11 water uptake capacity in this study,
relative to previously reported work, can be more plausibly attributed to
either a difference in the particle sizes of the synthesized ELM-11 sam­

ples, or differences in the gas-phase pressures of water vapor sampled in
the respective isotherm measurements (Cheng et al. sampled ten data
points at relative pressures below P/P0 ¼ 0.1, whereas in this study only
two sampling points were below P/P0 ¼ 0.1). Most likely, the differences
in water vapor adsorption obtained in this study and in Cheng et al. can
be attributed to a combination of differences in particle size and the
regimen used for water loading. If ELM-11 is suddenly exposed to water
at a high vapor pressure, the water will be incorporated into the surface
of the ELM-11 particle, transforming its surface into the nonporous
hydrated pre-ELM-11 form. This nonporous surface layer would inhibit
diffusion of water further into the interior of the ELM-11 particle. It is
therefore suggested that the more aggressive condition under which
water was loaded onto the ELM-11 adsorbent in this work is responsible
for the lower measured H2O capacity in comparison to previously re­
ported findings.
Cheng et al. [12] suggested that the slight difference seen in the CO2
adsorption isotherm before and after water exposure were probably due
to stacking faults and/or different interaction phases. Scanning electron
microscope images revealed a change in morphology, with the
as-synthesized pre-ELM-11 having a distinctly anisotropic plate-like
particle shape whereas the rehydrated form after water exposure did
not exhibit this highly preferential orientation. The presence of different
phases would cause additional stresses or faults in the crystal structure,
because the bipyridine linkers in the partially hydrated form (Fig. 3)
have different preferred orientations than bipyridine linkers in the
anhydrous form (Fig. 1). Given a sufficient number of exposure cycles,
the buildup of faults eventually leads to a breakdown of the crystal

2.1. The adsorption binding energy BE is calculated as
BE ¼ Egas=cluster À Egas À Ecluster


(1)

where Egas and Ecluster are the total electronic energies in isolation of the
optimized adsorbate and model ELM unit cluster geometries, respec­
tively, and Egas/cluster is the total electronic energy of the geometricallyoptimized adsorbate-adsorbent system when the gas molecule is adsor­
bed onto the ELM unit cluster.

Fig. 1. DFT optimized geometries for ELM-11 (left) and ELM-12 (right). Color
scheme is: copper (copper), fluorine (light blue), boron (pink), nitrogen (dark
blue), carbon (grey), oxygen (red), hydrogen (white), and sulfur (yellow). (For
interpretation of the references to colour in this figure legend, the reader is
referred to the Web version of this article.)
3


F.J. Sotomayor and C.M. Lastoskie

Microporous and Mesoporous Materials 304 (2020) 110377

Fig. 2. CO2 isotherms on ELM-11 at 273 K (A-1) and ELM-12 at 298 K (A-2) prior to (blue triangles) and after one (orange circles), two (yellow squares), three (grey
diamonds), and four (green crosses) cycles of exposure to water vapor. The corresponding water vapor isotherms on ELM-11 (B-1) and ELM-12 (B-2) at 298 K are
shown for the first (blue triangles), second (orange circles), third (yellow squares), and fourth (grey diamonds) exposures. Desorption branches for all isotherms have
been removed for clarity. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3. Top (left), front (center), and side views (right) of the DFT-optimized geometry for a partially hydrated ELM-11 cluster. Color scheme is: copper (copper),
fluorine (light blue), boron (pink), nitrogen (dark blue), carbon (grey), oxygen (red), and hydrogen (white). (For interpretation of the references to colour in this
figure legend, the reader is referred to the Web version of this article.)

structure. The declining CO2 capacity of ELM-11 can therefore be

attributed to the progressive accumulation of faults over the hydra­
tion/dehydration cycles. Had Cheng et al. continued their experiment
for additional exposure cycles, it is expected that they too would have
observed an eventual loss of CO2 capacity on ELM-11 due to the inac­
tivating effect of water vapor.
Resilience of ELM-12. ELM-11 and ELM-12 are both synthesized in
water-based solutions. Whereas the use of copper/tetrafluoroborate

solution produces the hydrate pre-ELM-11 as a precipitate, which can
later be activated to form ELM-11, the use of copper/OTf solution yields
the direct production of ELM-12 with no precursor hydrate. Activation
of ELM-12 is needed only to remove solvent molecules that remain in the
adsorbent pore spaces after synthesis. If the degradation of ELM-11 is the
result of an undesired reversion to its as-synthesized hydrated structure,
then the comparative stability of ELM-12 in the presence of water can be
rationalized by the absence of hydrated form for this MOF. This is likely
4


F.J. Sotomayor and C.M. Lastoskie

Microporous and Mesoporous Materials 304 (2020) 110377

a consequence of the strong OTf/copper coordination bond. Previous
work by Tran [15] found that the metal-anion force constant between
copper and oxygen in ELM-12 was twice as large as the metal-anion
force constant between copper and fluorine in ELM-11. This, in combi­
nation with the fact that OTf is a bulkier counter-ion than BF4, likely
prevents water from coordinating with ELM-12 copper vertex and
incorporating itself into the framework structure.

XRD patterns of ELM-12 taken before and after exposure to water
vapor show a return to its original crystal structure (Fig. 4), providing
further evidence of the resiliency of ELM-12 in the presence of moisture.
ELM variants that are anhydrous when synthesized are therefore ex­
pected to be more resistant to capacity loss from water vapor exposure
than ELM variants that require activation to remove water from the
framework structure. Additional XRD patterns before and after water
vapor exposure are shown in Fig. S1 in the Supporting Information.
Density functional theory calculations. Fig. 5 shows the preferred
binding sites of CO2 and H2O on ELM-11. Both adsorbates exhibit a
preference to interact with the fluorine atoms in the BF4 counter-ion.
However, H2O has a significantly stronger binding energy for this site
(À 60.4 kJ/mol for H2O, compared with À 25.8 kJ/mol for CO2). The
large difference in binding energy is of concern, as ELM-11 will selec­
tively adsorb the more strongly binding H2O over CO2 when both are
present in combustion flue gas. In addition, both CO2 and H2O cause a
slight extension of the copper-fluorine coordinate bond (from 2.24 Å to
2.27 Å and 2.34 Å respectively for CO2 and H2O). This is interesting
because the presence of fluorine in the counter-ion is expected to play a
role in modulating the interlayer expansion of ELM frameworks [7].
When gated adsorption of CO2 on ELM-11 occurs, the interlayer spacing
increases from 4.6 Å to 6.8 Å, or a net increase of 2.2 Å [28]. Thus, while
the Cu-F bond extension in ELM-11 is small in comparison to the total
framework layer expansion during CO2 adsorption, adsorption-induced
bond distortion within the layer may nonetheless be important in mo­
lecular identification at the ELM-11 surface, where adsorbate molecules
are principally interacting with one or two surface layers.
Adsorption-induced weakening of the Cu-F bond and resulting
intra-layer distortion may allow for increased interaction between the
framework and adsorbate molecules, and may thereby initiate clathrate

formation typical of guest molecule adsorption in the ELM-11 frame­
work. Adsorbate-induced intra-layer expansion of the surface layers may
also be responsible for the slight increase in volume that occurs just
below the gate pressure as observed by Kondo et al. [28] during
adsorption of CO2 onto ELM-11.
Jiang et al. [30] explored adsorption-desorption induced structural
changes of ELM-11 using the probe molecules CH3OH and CH3CN. Their
investigation found that adsorption of a molecule with a strong dipole
moment like CH3CN weakens the Cu-F coordinate bond between the
copper vertex atom and the BF4 counter-ion, thus increasing framework

Fig. 5. Highest energy binding sites for CO2 (left) and H2O (right) on ELM-11.
Binding energies are À 25.8 kJ/mol and À 60.4 kJ/mol respectively for CO2
and H2O.

flexibility. Water, like CH3CN, also has a strong dipole moment, and so
the weakening of the Cu-F coordinate bonds between the copper vertices
and the BF4 counter-ions in ELM-11 explains why ELM-11 can be
returned to pre-ELM-11 when it is exposed to water vapor.
Whereas ELM-11 has the small and roughly spherical BF4 as a
counter-ion, ELM-12 possesses the larger, linear, and more complex OTf
as its counter-ion and offers a greater number of potential binding sites
for gas adsorption. Fig. 6 shows the preferred binding sites of CO2 and
H2O on ELM-12. Although H2O still has a higher calculated binding
energy on ELM-12 than CO2 (À 51.4 kJ/mol for H2O, compared to À 18.0
kJ/mol for CO2), this difference is not as significant as it is for ELM-11,
because H2O and CO2 prefer different binding sites on the OTf group of
ELM-12. The dipolar water molecule prefers to directly interact with
both available oxygens of the OTf group, whereas quadrupolar CO2
prefers a binding configuration that bridges an OTf oxygen and a fluo­

rine. This suggests that the two molecules co-adsorb onto ELM-12, rather
than compete directly for binding sites.
Interestingly, as shown in Fig. 6, molecules adsorbing onto the ELM12 framework have the ability to rotate or otherwise affect the orien­
tation of the OTf group. Such orientation differences are less meaningful
for the spherical BF4 counter-ion in ELM-11, but may be important for
determining when gating occurs in ELM-12, since, as previously noted,
interlayer interactions for ELM adsorbents are putatively modulated by
the counter-ions.

Fig. 4. X-ray diffraction patterns of ELM-12 as synthesized (black) and after water vapor exposure (purple). XRD patterns were collected under normal atmospheric
conditions. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
5


F.J. Sotomayor and C.M. Lastoskie

Microporous and Mesoporous Materials 304 (2020) 110377

from the National Science Foundation under grant number CBET1034116. The authors also appreciate the support of Micromeritics In­
strument Corporation, who provided an ASAP-2050 Xtended Pressure
gas sorption analyzer through an equipment award for the measurement
of the carbon dioxide adsorption isotherms.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.micromeso.2019.03.019.
References
[1] S. Kitagawa, R. Kitaura, S. Noro, Functional porous coordination polymers, Angew.
Chem. 43 (2004) 2334–2375, />[2] H. Furukawa, K.E. Cordova, M. O’Keeffe, O.M. Yaghi, The chemistry and
applications of metal-organic frameworks, Science (New York, N.Y.) 341 (2013)
1230444, />[3] D.M. D’Alessandro, B. Smit, J.R. Long, Carbon dioxide capture: prospects for new

materials, Angew. Chem. 49 (2010) 6058–6082, />anie.201000431.
[4] J.-R. Li, Y. Ma, M.C. McCarthy, J. Sculley, J. Yu, H.-K. Jeong, P.B. Balbuena, H.C. Zhou, Carbon dioxide capture-related gas adsorption and separation in metalorganic frameworks, in: Coordination Chemistry Reviews, vol. 255, Elsevier B.V.,
2011, pp. 1791–1823, />[5] S. Horike, Y. Inubushi, T. Hori, T. Fukushima, S. Kitagawa, A solid solution
approach to 2D coordination polymers for CH4/CO2 and CH4/C2H6 gas separation:
equilibrium and kinetic studies, Chem. Sci. 3 (2012) 116, />c1sc00591j.
[6] D. Bousquet, F.-X. Coudert, A.G.J. Fossati, A.V. Neimark, A.H. Fuchs, A. Boutin,
Adsorption induced transitions in soft porous crystals: an osmotic potential
approach to multistability and intermediate structures, J. Chem. Phys. 138 (2013)
174706, />[7] H. Kajiro, A. Kondo, K. Kaneko, H. Kanoh, Flexible two-dimensional square-grid
coordination polymers: structures and functions, Int. J. Mol. Sci. 11 (2010)
3803–3845, />[8] H. Kanoh, A. Kondo, H. Noguchi, H. Kajiro, A. Tohdoh, Y. Hattori, W.-C. Xu,
M. Inoue, T. Sugiura, K. Morita, H. Tanaka, T. Ohba, K. Kaneko, Elastic layerstructured metal organic frameworks (ELMs), J. Colloid Interface Sci. 334 (2009)
1–7, />[9] H. Noguchi, A. Kondo, Y. Hattori, H. Kanoh, H. Kajiro, K. Kaneko, Clathrateformation mediated adsorption of methane on Cu-complex crystals, J. Phys. Chem.
B 109 (2005) 13851–13853, />[10] K. Sumida, D.L. Rogow, J. Mason, T.M. McDonald, E.D. Bloch, Z.R. Herm, T.H. Bae, J.R. Long, Carbon dioxide capture in metal-organic frameworks, Chem.
Rev. 112 (2012) 724–781, />[11] J. Yu, Y. Ma, P.B. Balbuena, Evaluation of the impact of H2O, O2, and SO2 on
postcombustion CO2 capture in metal-organic frameworks, Langmuir 28 (2012)
8064–8071, />[12] Y. Cheng, A. Kondo, H. Noguchi, H. Kajiro, K. Urita, T. Ohba, K. Kaneko, H. Kanoh,
Reversible structural change of Cu-MOF on exposure to water and its CO2
adsorptivity, Langmuir 25 (2009) 4510–4513, />la803818p.
[13] J. Liu, P.K. Thallapally, B.P. McGrail, D.R. Brown, J. Liu, Progress in adsorptionbased CO2 capture by metal-organic frameworks, Chem. Soc. Rev. 41 (2012)
2308–2322, />[14] A.C. Kizzie, A.G. Wong-Foy, A.J. Matzger, Effect of humidity on the performance of
microporous coordination polymers as adsorbents for CO2 capture, Langmuir 27
(2011) 6368–6373, />[15] T.D. Tran, Molecular Simulation of Carbon Dioxide Capture on Elastic Layered
Metal-Organic Framework Adsorbents (Doctoral Dissertation), University of
Michigan, 2012, />inh_1.pdf.
[16] A. Kondo, H. Kajiro, H. Noguchi, L. Carlucci, D.M. Proserpio, G. Ciani, K. Kato,
M. Takata, H. Seki, M. Sakamoto, Y. Hattori, F. Okino, K. Maeda, T. Ohba,
K. Kaneko, H. Kanoh, Super flexibility of a 2D Cu-based porous coordination
framework on gas adsorption in comparison with a 3D framework of identical
composition: framework dimensionality-dependent gas adsorptivities, J. Am.

Chem. Soc. 133 (2011) 10512–10522, />[17] A. Torrisi, C. Mellot-Draznieks, R.G. Bell, Impact of ligands on CO2 adsorption in
metal-organic frameworks: first principles study of the interaction of CO2 with
functionalized benzenes. I. Inductive effects on the aromatic ring, J. Chem. Phys.
130 (2009) 194703, />[18] A. Torrisi, C. Mellot-Draznieks, R.G. Bell, Impact of ligands on CO2 adsorption in
metal-organic frameworks: first principles study of the interaction of CO2 with
functionalized benzenes. II. Effect of polar and acidic substituents, J. Chem. Phys.
132 (2010) 044705, />[19] G.P. Dangi, R.S. Pillai, R.S. Somani, H.C. Bajaj, R.V. Jasra, A density functional
theory study on the interaction of hydrogen molecule with MOF-177, Mol. Simulat.
36 (2010) 373–381, />[20] L. Grajciar, A.D. Wiersum, P.L. Llewellyn, J. Chang, P. Nachtigall, “Understanding
CO2 adsorption in CuBTC MOF: comparing combined DFT–ab initio calculations

Fig. 6. Highest energy binding sites for CO2 (left) and H2O (right) on ELM-12.
Binding energies are À 18.0 kJ/mol and À 51.4 kJ/mol for CO2 and H2O
respectively.

4. Conclusions
Flexible metal-organic frameworks are prospective adsorbents for
carbon capture with low energies of regeneration. In this work, CO2
capacity retention under repeated exposure to water vapor was inves­
tigated for two elastic-layered metal organic frameworks (ELMs) that
differ only in the composition of their charge-balancing counter-ions.
One framework variant, ELM-11, underwent progressive loss of its CO2
adsorption capacity with each cycle of water vapor exposure. ELM-12,
by contrast, retained its CO2 capacity through at least four hydrother­
mal cycles. From density functional theory calculations, it was found
that the dipolar H2O binds more strongly to both ELM-11 and ELM-12
than the quadrupolar CO2. By leveraging the tunability of metal
organic frameworks, the uptake of strongly binding co-adsorbing species
can however be lessened. Replacing the BFÀ4 counter-ion in ELM-11 with
the larger, more complex CF3SOÀ3 anion in ELM-12, for example, reduces

competitive adsorption between CO2 and H2O by providing separate
preferred binding sites for the two adsorbates.
In this study, ELM-11 and ELM-12 were cyclically exposed to mois­
ture by adsorbing pure water vapor at room temperature. It bears noting
that the conditions maintained in these static adsorption experiments
differ significantly from the combustion flue gas conditions encountered
in prospective CO2 capture operations at power plants and industrial
facilities, which are likely to incur higher temperatures in flow-through
systems with low absolute humidity. The logical next step of analysis is
to test the capture performance of ELMs by measuring the breakthrough
curves of humidified gases in dynamic flow experiments over a range of
temperatures to understand whether low absolute humidity would limit
some of the worst impacts of water vapor exposure. Similar break­
through measurements were recently reported for ELM-11 in purge
displacement experiments involving dry gas combinations of CO2, ni­
trogen, methane, and helium [31]. Given that ELMs and other flexible
framework adsorbents have crystalline structures that are highly sensi­
tive to atmospheric conditions, analysis of the structural stability of
ELMs would benefit from additional experiments carried out under flow
conditions for varying regimes of temperature and moisture. Likewise,
computational and experimental studies of sulfur dioxide adsorption on
ELM adsorbents would be instructive, as SO2 is known to competitively
adsorb with carbon dioxide on metal-organic framework adsorbents
when produced from the combustion of sulfur-rich coals.
Funding
The authors gratefully acknowledge sponsorship of this research
6


F.J. Sotomayor and C.M. Lastoskie


[21]

[22]

[23]

[24]

[25]

Microporous and Mesoporous Materials 304 (2020) 110377

with microcalorimetry experiments, J. Phys. Chem. C 115 (2011) 17925–17933,
/>N.A. Ramsahye, G. Maurin, S. Bourrelly, P.L. Llewellyn, C. Serre, T. Loiseau,
T. Devic, G. Ferey, Probing the adsorption sites for CO2 in metal organic
frameworks materials MIL-53 (Al, Cr) and MIL-47 (V) by density functional theory,
J. Phys. Chem. C 112 (2008) 514–520, />N. Nijem, P. Thissen, Y. Yao, R.C. Longo, K. Roodenko, H. Wu, Y. Zhao, K. Cho,
J. Li, D.C. Langreth, Y.J. Chabal, Understanding the preferential adsorption of CO2
over N2 in a flexible metal-organic framework, J. Am. Chem. Soc. 133 (2011)
12849–12857, />Gaussian 09. 2010. Revision C.01, M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E.
Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A.
Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F. Izmaylov, J.
Bloino, G. Zheng, J.L. Son. Gaussian, Inc., Wallingford CT.
Y. Yang, M.N. Weaver, K.M. Merz, Assessment of the 6-31ỵG** þ LANL2DZ’
mixed basis set coupled with density functional theory methods and the effective
core potential: prediction of heats of formation and ionization potentials for firstrow-transition-metal complexes, J. Phys. Chem. A 113 (2009) 9843–9851, https://
doi.org/10.1021/jp807643p.
D. Li, K. Kaneko, Hydrogen bond-regulated microporous nature of copper complexassembled microcrystals, Chem. Phys. Lett. 335 (2001) 50–56, />10.1016/S0009-2614(00)01419-6.


[26] H. Noguchi, A. Kondo, Y. Hattori, H. Kajiro, H. Kanoh, K. Kaneko, Evaluation of an
effective gas storage amount of latent nanoporous Cu-based metal-organic
framework, J. Phys. Chem. C 111 (2007) 248–254, />jp064035i.
[27] A. Kondo, H. Noguchi, L. Carlucci, D.M. Proserpio, G. Ciani, H. Kajiro,
Tomonori Ohba, H. Kanoh, K. Kaneko, Double-step gas sorption of a twodimensional metal-organic framework, J. Am. Chem. Soc. 129 (2007)
12362–12363, />[28] A. Kondo, H. Noguchi, S. Ohnishi, H. Kajiro, A. Tohdoh, Y. Hattori, W.-C. Xu,
H. Tanaka, H. Kanoh, K. Kaneko, Novel expansion/shrinkage modulation of 2D
layered MOF triggered by clathrate formation with CO2 molecules, Nano Lett. 6
(2006) 2581–2584, />[29] A. Kondo, A. Chinen, H. Kajiro, T. Nakagawa, K. Kato, M. Takata, Y. Hattori,
F. Okino, T. Ohba, K. Kaneko, H. Kanoh, Metal-ion-dependent gas sorptivity of
elastic layer-structured MOFs, Chem. Eur J. 15 (2009) 7549–7553, />10.1002/chem.200901208.
[30] Y. Jiang, J. Huang, B. Kasumaj, G. Jeschke, M. Hunger, T. Mallat, A. Baiker,
Adsorption-desorption induced structural changes of Cu-MOF evidenced by solid
state NMR and EPR spectroscopy, J. Am. Chem. Soc. 131 (2009) 2058–2059,
/>[31] F.J. Sotomayor, C.M. Lastoskie, “Predicting the breakthrough performance of
‘gating’ adsorbents using osmotic framework-adsorbed solution theory”, Langmuir
33 (2017) 11670–11678, />
7