ArticlePDF AvailableLiterature Review

Adsorption and Separation by Flexible MOFs

Advanced Materials

January 2025

DOI:10.1002/adma.202414724

License
CC BY 4.0

Authors:

Irena Senkovska

TU Dresden

Volodymyr Bon

TU Dresden

Show all 5 authorsHide

Flexible metal–organic frameworks (MOFs) offer unique opportunities due to their dynamic structural adaptability. This review explores the impact of flexibility on gas adsorption, highlighting key concepts for gas storage and separation. Specific examples demonstrate the principal effectiveness of flexible frameworks in enhancing gas uptake and working capacity. Additionally, mixed gas adsorption and separation of mixtures are reviewed, showcasing their potential in selective gas separation. The review also discusses the critical role of the single gas isotherms analysis and adsorption conditions in designing separation experiments. Advanced combined characterization techniques are crucial for understanding the behavior of flexible MOFs, including monitoring of phase transitions, framework–guest and guest–guest interactions. Key challenges in the practical application of flexible adsorbents are addressed, such as the kinetics of switching, volume change, and potential crystal damage during phase transitions. Furthermore, the effects of additives and shaping on flexibility and the “slipping off effect” are discussed. Finally, the benefits of phase transitions beyond improved working capacity and selectivity are outlined, with a particular focus on the advantages of intrinsic thermal management. This review highlights the potential and challenges of using flexible MOFs in gas storage and separation technologies, offering insights for future research and application.

Schematic representations of the structural transitions between narrow pore phase (np) and open pore phase (op) induced by a general state variable X: a) free energy F(V; X); b) response V(X), Reproduced with permission.[¹⁶] Copyright 2019, Elsevier Ltd.; as well as schematic representation of the two corresponding framework states (c).

…

Comparison of the typical adsorption isotherms of flexible MOF (X, top) and rigid MOF with comparable pore size (Y, bottom). The working capacity in both cases is indicated by the arrow. Reproduced with permission.[²²] Copyright 2020, The Authors, under CC BY 4.0.

…

a) Pillared layer structure and b) adsorption (filled circles) and desorption (open circles) isotherms of N2, CH4, CO2 and O2 at 298 K for [Cu(dhbc)(4,4′‐bipyridine)]n. Reproduced with permission.[³³] Copyright 2002, Wiley‐VCH GmbH & Co. KGaA.

…

a) Structure and conformation of the DUT‐49 framework and 9,9′‐([1,1′‐biphenyl]−4,4′‐diyl)bis(9H‐carbazole‐3,6‐dicarboxylate) linker in op (left) and cp (right) structures. Guest molecules are omitted for clarity. Reproduced with permission.[³⁵] Copyright 2016, Nature Publishing Group. b) Total volumetric CH4 adsorption isotherm of DUT‐49 at 298 K (adsorption = closed symbols, desorption = open symbols). The working capacity is given between 65 and 5 bar, and 100 and 5 bar. c) CH4 adsorption and desorption isotherms in DUT‐49 at 111 K. Reproduced with permission.[³⁵] Copyright 2016, Springer Nature Limited.

…

+38

a) The structure of H2bdp ligand along with the crystal structures of the closed pore (left) and open pore (right) [Co(bdp)]n phases. b,c) Total CH4 adsorption isotherms for [Co(bdp)]n (b) and [Fe(bdp)]n (c) at 298 K. Filled circles represent adsorption; open circles represent desorption. Reproduced with permission.[²⁶] Copyright 2015, Nature Publishing Group.

…

Figures - available from: Advanced Materials

This content is subject to copyright. Terms and conditions apply.

Access to this full-text is provided by Wiley.

Learn more

Content available from Advanced Materials

This content is subject to copyright. Terms and conditions apply.

REVIEW

www.advmat.de

Adsorption and Separation by Flexible MOFs

Irena Senkovska,* Volodymyr Bon, Antonia Mosberger, Yutong Wang, and Stefan Kaskel*

Dedicated to the 60th birthday of Prof. Omar Yaghi

Flexible metal–organic frameworks (MOFs) oﬀer unique opportunities due to

their dynamic structural adaptability. This review explores the impact of

ﬂexibility on gas adsorption, highlighting key concepts for gas storage and

separation. Speciﬁc examples demonstrate the principal eﬀectiveness of

ﬂexible frameworks in enhancing gas uptake and working capacity.

Additionally, mixed gas adsorption and separation of mixtures are reviewed,

showcasing their potential in selective gas separation. The review also

discusses the critical role of the single gas isotherms analysis and adsorption

conditions in designing separation experiments. Advanced combined

characterization techniques are crucial for understanding the behavior of

ﬂexible MOFs, including monitoring of phase transitions, framework–guest

and guest–guest interactions. Key challenges in the practical application of

ﬂexible adsorbents are addressed, such as the kinetics of switching, volume

change, and potential crystal damage during phase transitions. Furthermore,

the eﬀects of additives and shaping on ﬂexibility and the “slipping oﬀ eﬀect”

are discussed. Finally, the beneﬁts of phase transitions beyond improved

working capacity and selectivity are outlined, with a particular focus on the

advantages of intrinsic thermal management. This review highlights the

potential and challenges of using ﬂexible MOFs in gas storage and separation

technologies, oﬀering insights for future research and application.

1. Introduction

Metal–organic frameworks (MOFs) are crystalline coordination

polymers composed of metal ions or clusters coordinated by

organic ligands, facilitating the formation of inner cavities.[1]

Due to their dual inorganic/organic nature as coordination com-

pounds, the chemistry of MOFs is enormously rich. The vast

number of metal/ligand combinations grants these compounds

exceptional versatility, unlocking immense potential for a wide

I. Senkovska, V. Bon, A. Mosberger, Y. Wang, S. Kaskel

Chair of Inorganic Chemistry I

Technische Universität Dresden

Bergstrasse 66, 01069 Dresden, Germany

E-mail: irena.senkovska@tu-dresden.de;stefan.kaskel@tu-dresden.de

The ORCID identiﬁcation number(s) for the author(s) of this article

can be found under https://doi.org/10.1002/adma.202414724

GmbH. This is an open access article under the terms of the Creative

Commons Attribution License, which permits use, distribution and

reproduction in any medium, provided the original work is properly cited.

DOI: 10.1002/adma.202414724

range of applications, including storage and

separation of molecular species.[2]How-

ever, the vast diversity also turns MOFs into

a “haystack,” making it challenging for re-

searchers to ﬁnd the “needle” when try-

ing to identify the best MOF for a speciﬁc

application.

Recognized early on for their poten-

tial as porous materials, MOFs have un-

dergone signiﬁcant advancements since

their inception. At the early stage of de-

velopment, MOFs faced challenges with

maintaining permanent porosity, and ini-

tial structures often collapsed after remov-

ing guest molecules. However, advances in

desolvation techniques[3]and a better un-

derstanding of building principles[4]led to

the development of robust porous MOFs,

among others, with enhanced chemical sta-

bility. After this breakthrough, MOFs be-

came prominent for their ability to adsorb

and separate gases.

The primary interest in MOFs is,

nonetheless, motivated by their modular

construction and accessibility to deliberate

structural design and reticular chemistry.[5]

Access to isoreticular families of mate-

rials allowed for systematic studies of

structure–properties relationships.[6]The result of these designs

is ultrahigh porosity with surface areas of up to 7850 m2g−1and

pore volume of 5.0 cm3g−1.[7]MOFs still hold records for speciﬁc

surface area and pore volume among microporous adsorbents.

Such new horizons in terms of porosity and storage capacities

catapulted MOFs to high-performance adsorbents.[8]

However, the main feature distinguishing MOFs from other

porous adsorbents, such as activated carbons, is not only the crys-

tallinity of these materials and, as a consequence, the regular

and crystallographically precise pore structure. A unique feature

of some MOFs is the structural ﬂexibility and the ability to un-

dergo stimuli-induced structural transitions.[9]Some MOF repre-

sentatives are able to switch between the crystalline phases with

diﬀerent porosity induced by guest physisorption. A number of

terms have been proposed to manifest the ﬂexibility of MOFs,

including “soft,” “ﬂexible,” “dynamic,” “stimuli-responsive,” or

“switchable.” They all describe solid-state structural phase tran-

sitions initiated by relatively small energetic stimuli and usually

(but not necessarily) include considerable changes in the unit cell

and pore volumes.[9,10]

They are also emerging material classes, including covalent

organic frameworks (COFs)[11]and hydrogen-bonded organic

www.advancedsciencenews.com www.advmat.de

Figure 1. Schematic representations of the structural transitions between narrow pore phase (np) and open pore phase (op) induced by a general state

of the two corresponding framework states (c).

frameworks (HOFs),[12]sharing with MOFs key characteristics,

such as crystallinity, porosity, and ability to undergo structural

transitions. While this review focuses on MOFs, the general ob-

servations and dependencies discussed here are likely applicable

to these materials as well.

Although the variety of phase transitions inducing stimuli

is broad, including pressure, temperature, electric ﬁeld, light

etc.,[13]we are going to focus on the structural changes caused

by the adsorption or desorption of molecules and the bene-

ﬁts/drawbacks of them in ﬁelds of gas storage and separation.

In general, the structural transitions lead to a stepwise increase

(or decrease) of porosity between two extreme cases:[14]i) non-

porous state (closed pore phase, cp), resulting from the dense

packing of the building blocks and ii) maximal possible pore

size and volume for given framework topology (open pore phase,

op). The number of (meta)stable crystal structures (intermedi-

ate phases (ip), or narrow pore phases (np)) between these two

extremes in the given system depends on the free energy land-

scape for the given ﬂuid/framework combination, e.g., the num-

ber of local minima and the height of activation barriers that sep-

arate them (multistability of free energy proﬁle, Figure 1a). The

existence of the energy barrier between two minima in the free

energy landscape is the origin of hysteresis loops in the adsorp-

tion/desorption isotherms (Figure 1b).[15]

Depending on the initial state of the framework and the transi-

tion pathway, the “gate-opening” and “breathing” transitions are

mainly discussed. Sometimes, the authors also use the terms in-

terchangeably. “Gate-opening” behavior is characterized by the

increase in the porosity upon the transition, whereby a less

porous structure (cp or np) expands to a more porous after a

certain threshold gas pressure (pgo) is reached. The physisorp-

tion isotherm, in this case, is characterized by the stepwise in-

crease in the adsorbed amount in a pretty narrow pressure region

(Figure 1c).

“Breathing” usually includes two consecutive transitions dur-

ing gas pressure increase: ﬁrst contraction transition of the more

porous phase (op) to the less porous one (np) at characteristic

contraction pressure (pc), followed by second expansion transi-

tion (gate opening) to a more porous (large pore or op) phase at

characteristic pex. The ﬁrst contraction is usually not easily iden-

tiﬁed in the isotherm (particularly if the phases have a minor

porosity diﬀerence) and have to be often detected by complemen-

tary techniques, such as in situ powder X-ray diﬀraction (PXRD)

(see Section 3).

The isotherms of ﬂexible MOFs can be constructed by super-

imposing the isotherms of two or more structures with varying

adsorption capacities, adsorption enthalpies, etc. However, typi-

cally along the isotherm, the metastable states are observed. The

adsorption or desorption branch (or both) can be kinetically hin-

dered, and the energetic balance strictly depends on the num-

ber of guest molecules in the pores. Thus, the global minimum

state, among all possible conﬁgurations, changes depending on

the external guest pressure.[17]These activation barriers stem 1)

from the solid–solid phase transitions and are particularly high

if the volume change is high, but also 2) the ﬂuid-phase activa-

tion barriers (diﬀusion, nucleation, etc.) may play a role. These

factors lead to hysteresis, which does not vanish even for long

equilibration times because the barriers depend exponentially on

one or more thermodynamic variables and can only be overcome

if the variable reaches a critical magnitude (such as gate open-

ing pressure, see below). Solid–solid phase transitions with high

activation barriers may be characterized as 1st order transitions

leading to steps in the isotherm. If the barriers for solid–solid

transitions are minimal, a quasi-continuous structural change is

observed, particularly when the volume changes are minor be-

cause the pores are already ﬁlled with linker functionalities.[18]

The fascination of ﬂexible MOFs for separations originates

from the vision that an MOF may recognize a molecular stimulus

so speciﬁcally that it opens the pores only for this species, even if

it is only a minor component in a mixture of molecules, similar

to an enzyme recognizing its substrate.[19]However, the bene-

ﬁts of the dynamic nature of MOFs for adsorptive application are

controversially discussed in the literature. Besides the beneﬁts of

structural ﬂexibility discussed below in detail (Sections 2.1–2.3),

there are some critical material-relevant, as well as technologi-

cal aspects related to ﬂexibility. The primary concerns involve the

cyclability and the damage of the crystals in the course of mul-

tiple expansion/shrinkage cycles (Section 4.3) and the manage-

ment of the crystal volume change in the adsorption chamber

or column (Section 4.2). Since the ﬁrst discovery of switchable

MOFs, the number of new MOFs showing pronounced ﬂexibil-

ity and huge changes in pore size is increasing exponentially.

Also, some unique phenomena connected to gas-induced ﬂex-

ibility, such as “negative gas adsorption” were discovered, and,

www.advancedsciencenews.com www.advmat.de

despite open questions, more and more functions resulting from

pore switchability are reported. In the following, we discuss the

concepts of using ﬂexible MOFs as eﬃcient adsorbents, giving

some examples.

2. Impact of Flexibility and Concepts for Gas

Adsorption and Separation Using Flexible MOFs

2.1. Gas Storage Concept

Gas storage is one of the most essential applications of porous

materials, based on gas enrichment in an interfacial layer due to

solid–ﬂuid interactions.

Volumetric and gravimetric capacities are of central impor-

tance for gas storage, although the temperature of operation and

thermal eﬀects upon adsorption are also crucial. The aim is,

therefore, to achieve high volumetric and gravimetric capacities

at a usable temperature.[20]

The main indicator for the comparison of the storage perfor-

mance, however, is the so-called working (also usable) capacity,

reﬂecting the diﬀerence between the amount of gas adsorbed at

the target storage pressure and the amount that still be adsorbed

at the lowest desorption pressure, acceptable for the operation

of the system. In rigid adsorbents, the best performance can be

reached by the adjustment of the pore sizes to the operational

pressure and temperature range. The subatmospheric pressures

require (ultra)microporosity, whereas the mesoporous MOFs or

MOFs with hierarchical pore structure are beneﬁcial for high-

pressure storage.[21]

The stepwise isotherms, characteristic for ﬂexible MOFs,

open a new opportunity for boosting usable capacity, where the

amount of gas adsorbed would be negligible at low pressures (in

the pressure region below the lowest desorption pressure accept-

able) but rise sharply just before the pressure reaches the desired

storage pressure (Figure 2). Flexibility allows the material to close

pores at relatively high pressure and expel the gas molecules from

the pores. As a consequence, more gas can be released from the

pores in comparison to the rigid MOF with a comparable pore

size. In the pressure region above that is needed for the opening

transition, the gas uptake is equal to that of comparable micro-

porous adsorbent. Therefore, the optimal opening and closing

pressures are crucial for achieving high usable capacity. A criti-

cal aspect is hysteresis, which should be as narrow as possible to

achieve a high working capacity and reduce energy consumption.

The second beneﬁt oﬀered by ﬂexible MOFs is thermal man-

agement. There are practical challenges involved in designing

systems with high capacities and managing the temperature

change associated with the adsorption and desorption of gas from

the adsorbent due to the high gas adsorption enthalpies in micro-

porous materials. In practical gas storage application systems, the

heat of adsorption (exothermic) and desorption (endothermic)

can lead to strong temperature variations,[23]which can have a

negative impact on the usable capacity. Rapid charging and dis-

charging are desirable, but they also lead to signiﬁcant tempera-

ture ﬂuctuations. For example, in the case of CH4storage, mea-

surements with commercially available Adsorbed Natural Gas

(ANG) cylinders show a temperature drop of up to 37 °C at high

discharge rates, with performance loss approaching 25% of the

Figure 2. Comparison of the typical adsorption isotherms of ﬂexible MOF

(X, top) and rigid MOF with comparable pore size (Y, bottom). The work-

ing capacity in both cases is indicated by the arrow. Reproduced with

isothermal capacity. The performance loss is expected to be 15–

20% at moderate discharge rates.[24]Analysis indicates that the

thermal capacity of the vessel and external heat transfer condi-

tions signiﬁcantly aﬀect system behavior. MOFs have very low

thermal conductivity,[25]so additional eﬀorts are required to ad-

dress thermal management challenges.

In ﬂexible MOFs, fortunately, the enthalpy associated with the

phase transition (absorption or release of heat upon reversible

phase transitions) can counteract the thermal eﬀect of adsorp-

tion, contributing to positive thermal management (Section 5).[26]

Net heat removed from the ﬂexible system during the desorption

process under adiabatic conditions is the sum of the endothermic

heat associated with guest desorption, the exothermic heat gen-

erated by host shrinkage, and the exothermic heat of the guest

molecules that remain in the host framework due to changes in

host–guest interactions upon shrinkage of the host.[17]

Last but not least, ﬂexible MOFs could potentially minimize

the “ageing” of adsorbent generated by the accumulation of im-

purities. The selectivity in adsorption (discussed in detail in

www.advancedsciencenews.com www.advmat.de

Section 2.2) could lead to the adsorption of only the main, valu-

able components from the gas mixture.

For example, the adsorptive storage of methane often suﬀers

from the adsorption of impurities present in natural gas, such

as higher hydrocarbons (ethane, propane, etc.). These hydrocar-

bons can have a deleterious eﬀect due to accumulation, decreas-

ing methane storage capacity, and aﬀecting the long-term stabil-

ity of the adsorbent.[27]The selective adsorption of the main com-

ponent or forced desorption of the impurities supported by the

ﬂexibility could improve long-term performance.

2.1.1. Examples of Flexible MOFs in Storage Applications

In this section, we explore the unique and critical features of ﬂex-

ible MOFs in terms of usable capacity for gas storage and high-

light the advantages of ﬂexible MOFs through speciﬁc examples.

Many gases have been used to study the adsorption behavior of

ﬂexible porous materials, but most are not considered for stor-

age applications. This is partly because of well-developed stor-

age technologies, e.g., compressed Ar, O2,N

2. The gas storage

in porous solids is mainly motivated by the goal of increasing

the amount of gas within a given volume, e.g., for applications

in the mobility sector. Consequently, only a limited number of

gases, notably H2and CH4, are typically considered for practical

storage purposes.

An additional motivation is the safe storage of hazardous

gases, such as acetylene.

CH4Storage in Flexible MOFs: As the primary constituent of

natural gas, CH4oﬀers a high caloriﬁc value, low carbon dioxide

emissions, and a high research octane number, making it an at-

tractive fuel for vehicular applications. The Advanced Research

Projects Agency-Energy (ARPA-E) under the U.S. Department of

Energy (DOE) has established several ambitious targets for on-

board CH4storage systems.[21,28]The gravimetric storage target

is set at 0.5 g(CH4) per 1 g of adsorbent. The volumetric target (de-

liverable capacity) is 263 cm3(at standard temperature and pres-

sure, STP) per mL of adsorption chamber at 298 K and 65 bar, cor-

responding to compressed methane gas at 250 bar.[28,29]Addition-

ally, accounting for up to 50% volume loss due to the low packing

densities of MOFs,[21]the volumetric storage target for the ma-

terial itself should be even higher. Over the past three decades,

MOFs have been extensively investigated and documented for

their substantial potential in CH4storage applications.[21,30]Com-

putational calculations demonstrated, however, that the usable

(deliverable) capacity of almost all conventional rigid adsorbents

would not be higher than ≈200 cm3methane per cm3of adsor-

bent (deﬁned as STP volume at 298 K and 65–5.8 bar pressure

window),[21,31]which is very low compared to the DOI target. One

of the reasons is that most adsorbents exhibit Type I isotherms,

where a substantial amount of methane remains adsorbed in the

microporous frameworks at pressures around 5 - 6 bar.

Recently, ﬂexible MOFs have gained prominence due to their

ability to markedly increase the usable storage capacity. Methane

storage in ﬂexible MOFs has been recently summarized by For-

rest et al.,[32]therefore we will discuss only a few examples in the

following.

In 2002 and 2003, Seki and Kitagawa reported for the ﬁrst

time a ﬂexible pillared layer MOFs, [Cu2(bdc)2(4,4′-bipyridine)]n

(bdc =terephthalate) and [Cu(dhbc)(4,4′-bipyridine)]n(Hdhbc =

2,5-dihydroxybenzoate), showing the gate opening CH4adsorp-

tion isotherm at high pressures and 298 K (Figure 3).[33]Although

the overall methane uptake is not high, the gate opening and clos-

ing pressures are below 10 bar; thus, this reports opened up new

avenues for methane storage in ﬂexible frameworks.[32]

In 2009, Kaneko and co-workers[34]reported methane ad-

sorption isotherm of ELM-11 ([Cu(BF4)2(4,4′-bipyridine)2]n)

(Figure 37b,g), and discussed the beneﬁts of gating isotherm for

eﬃcient methane storage. The adsorption capacity of the MOF

at65baris55mgg

−1. Taking the packing density of the mate-

rial into account, the storage capacity was calculated to be 155

cm3cm−3(volume of gas/volume of storage vessel).

In 2012, Stoeck et al.[21b]synthesized a carbazole-based meso-

porous MOF (DUT-49) exhibiting a hierarchical pore system

and showing an exceptionally high speciﬁc surface area of

5476 m3g−1and a large total pore volume of 2.91 cm3g−1,as

well as record gravimetric total methane uptake of 0.56 g g−1

(236 cm3cm−3) at 298 K and 110 bar (Figure 4b). The work-

ing capacity of the MOF in the 5–65 bar range amounts to

177 cm3cm−3.

Later, authors recognized that DUT-49 displays an unusual

breathing type isotherm with a negative step in the adsorption

branch, coined as “negative gas adsorption” (NGA) in methane

isotherm measured at 111 K (Figure 4c).[35]NGA is recognized

as a new, counterintuitive phenomenon based on the release of

already adsorbed gas upon contraction of metastable open pore

phase of DUT-49 to a narrow pore phase, accompanied by the

reduction of the pore volume of the structure by 1.8 cm3g−1.

Krause et al.[36]discovered that by varying the length of the

linker, the number of CH4molecules adsorbed at the intersec-

tion of the open-pore and closed-pore phases increased with the

ligand length. This leads to a more favorable overall change in

adsorption enthalpy for structural contraction, thereby promot-

ing the occurrence of NGA. Then, they investigated the eﬀects of

temperature and adsorbate on the occurrence and extent of NGA

in DUT-49.[37a]Through experiments covering a broad range

of gases and temperatures, the speciﬁc temperature ranges in

which NGA is observable for each guest molecule could be deter-

mined. These ﬁndings were further complemented by molecular

simulations to explain the absence of NGA at higher tempera-

tures and the non-monotonic behavior observed at lower temper-

atures. Although the eﬀect disappears in the methane adsorp-

tion isotherm at room temperature, in future, it can be explored

for the separation of other gases, such as C3–C4 hydrocarbons,

which provoke NGA transitions at room temperature.

In 2015, Long and co-workers reported the [M(bdp)]n(bdp =

1,4-benzenedipyrazolate, Figure 5a)seriesofMOFs(M=Co, Fe)

displaying working capacities for CH4, very close to the high-

est working capacities among all MOF materials reported up to

now,[32]owing to their large pore volume and a high degree of

ﬂexibility.[26]The [Co(bdp)]nshows negligible CH4uptake at low

pressures, with a noticeable step in the adsorption isotherm after

reaching 16 bar (Figure 5b). Desorption exhibits hysteresis, and

the loop closes at 7 bar, indicating a reversible structural phase

transition (Figure 5a), stable over 100 adsorption/desorption cy-

cles. The CH4usable volumetric storage capacity of [Co(bdp)]nis

197 cm3cm−3for 5–65 bar pressure range at 25 °C, calculated

using the crystallographic density of the materials and geometric

www.advancedsciencenews.com www.advmat.de

Figure 3. a) Pillared layer structure and b) adsorption (ﬁlled circles) and desorption (open circles) isotherms of N2,CH

4,CO

2and O2at 298 K for

Figure 4. a) Structure and conformation of the DUT-49 framework and 9,9′-([1,1′-biphenyl]−4,4′-diyl)bis(9H-carbazole-3,6-dicarboxylate) linker in op

b) Total volumetric CH4adsorption isotherm of DUT-49 at 298 K (adsorption =closed symbols, desorption =open symbols). The working capacity

is given between 65 and 5 bar, and 100 and 5 bar. c) CH4adsorption and desorption isotherms in DUT-49 at 111 K. Reproduced with permission.[35 ]

www.advancedsciencenews.com www.advmat.de

Figure 5. a) The structure of H2bdp ligand along with the crystal structures of the closed pore (left) and open pore (right) [Co(bdp)]nphases.

b,c) Total CH4adsorption isotherms for [Co(bdp)]n(b) and [Fe(bdp)]n(c) at 298 K. Filled circles represent adsorption; open circles represent desorption.

pore volume of the expanded phase. High working capacity is ex-

plained by nearly complete desorption of the methane from the

pores (remind adsorbed amount less than 0.2 mmol g−1at 25 °C)

below gate closing pressure.

For isoreticular [Fe(bdp)]n, opening and closing transitions

occur at 24 and 10 bar, respectively (Figure 5c). Nonetheless,

[Fe(bdp)]nmaintains a high volumetric usable capacity of 190

cm3cm−3at 5–65 bar and 298 K. Additionally, [Fe(bdp)]nunder-

goes further expansion at pressures above 40 bar, forming an op

framework with nearly perfect square channels.[26]

The same research group proposed a strategy to adjust the

pressure required for structural transformation by chemical

modiﬁcation of the linker, which led to ﬁve functionalized com-

pounds (Figure 6a), isostructural to [Co(bdp)]n, namely, [Co(F-

bdp)]n, [Co(p-F2-bdp)]n,[Co(o-F

2-bdp)]n,[Co(D

4-bdp)]n, and [Co(p-

Me2-bdp)]n.[38]Experimental results indicate that the functional-

ization of the ligand by introducing ﬂuorine atoms in the phenyl

ring leads to a decrease in phase transition pressure by disrupting

edge-to-face 𝜋–𝜋interactions, while methyl groups increase the

opening pressure by enhancing these interactions (Figure 6b,c).

Yang et al.[39]reported a [NiL2]n(L =4-(4-pyridyl)-biphenyl-

4-carboxylate), ﬂexible MOF with a dia topology and sixfold in-

terpenetration, denoted as X-dia-1-Ni. The framework demon-

strates substantial ﬂexibility, and the initially nonporous X-dia-1-

Ni underwent multiple phase transitions during CO2adsorption

at 195 K (for more crystallographic details, see Section 3.1). High-

pressure CH4adsorption studies at 298 K revealed an S-shaped

adsorption isotherm (Figure 7), with negligible CH4uptake be-

low 20 bar, followed by a sharp increase as the pores open. At 298

K, the working CH4capacity is 162 cm3g−1(147 cm3cm−3)at

1–65bar.

[39,40]

However, low closing pressure observed in the desorption

isotherms signiﬁcantly reduces the working capacities, consid-

ering the working pressure window between 5 and 65 bar. Cobalt

doping in this system (similar to DUT-8(Ni)[41]) enables control

over the gate opening. So, the methane-induced phase transfor-

mations can be ﬁne-tuned by using diﬀerent Ni/Co ratios to en-

hance methane working capacity.[40]The hysteresis was shifted

to a higher pressure, and the amount of CH4retained at 5 bar

during desorption decreased from 60 cm3g−1in X-dia-1-Ni to 19

cm3g−1in X-dia-1-Ni0.89Co0.11 . Therefore, the working capacity

of X-dia-1-Ni0.89Co0.11 reached 202 cm3g−1(5–65 bar) at 298 K

(Figure 7).

Kaskel and co-workers investigated high-pressure CH4adsorp-

tion at 298 K of [Zn2(BPnDC)2(bpy)]n(SNU-9, BPnDC =ben-

zophenone 4,4′-dicarboxylate, bpy =4,4′-bipyridine), the frame-

work reported by Park and Suh already in 2010.[42]The com-

pound demonstrates stepwise adsorption isotherm and hys-

teresis between the adsorption and desorption branches.[43]Al-

though the pore volume of this interpenetrated framework is

moderate (0.37 cm3g−1), the working capacity between 100 and

5 bar at 298 K is 144 cm3cm−3due to the relatively high crystal-

lographic density of the material (1.074 cm3g−1in the open pore

state). In the window between 65 and 5 bar it is, however, much

lower because of the high gate opening pressure (Figure 8).

Two isostructural ﬂexible MOFs of the MIL-53 series, namely,

MIL-53(Al)-OH and MIL-53(Al)-(OH)2, exhibit transitions from a

narrow pore state to an open-pore state during CH4adsorption

www.advancedsciencenews.com www.advmat.de

Figure 6. a) The H2bdp derivatives used as linker. b) CH4adsorption isotherms of [Co(bdp)]n,[Co(Fbdp)]

n, [Co(p-F2-bdp)]n, [Co(o-F2-bdp)]n,and

Society.

at 298 K, and the opening pressures are ≈15 and 46 bar, respec-

tively, with hysteresis in the desorption isotherms leading to clo-

sure pressures of 4.1 and 5.8 bar[46](Figure 9).

The higher transition pressure of MIL-53(Al)-(OH)2is at-

tributed to the additional ─OH groups, which enhance the inter-

molecular interactions and impede the transition of the frame-

work (Figure 9a). At 298 K, the usable capacities for MIL-53(Al)-

OH and MIL-53(Al)-(OH)2in the 5–65 bar range are 71 and 164

cm3cm−3, respectively.

The examples demonstrate the importance of the control over

phase transition pressure at desired thermodynamic conditions.

Such control can be realized by the constitutional changes of the

building blocks (metal in the cluster or substituent on the linker),

by the combination of multiple linkers in the same framework

(solid solution),[47]or by the adjustment of particle size.[48 ]

The mixed ligand strategy was demonstrated by Bolinois et al.

in MIL-53(Al), where the bdc was combined with NH2-bdc in var-

ious ratios to induce ﬂexibility upon high-pressure methane ad-

sorption (Figure 10).[47b]

Very recently, Zhai group was able to show that by intro-

ducing a functional group (in this case, azido-group), a for-

merly rigid CAU-10 ([Al(OH)(1,3-bdc)]n, 1,3-bdc =isophthalate)

framework[49]can be transformed into a ﬂexible one.[50 ]

So far, no MOF structure, rigid or ﬂexible, has been able to

meet the DOE’s current targets for viable methane on-board stor-

age systems in terms of deliverable capacity. Flexible MOFs with

a potential for high deliverable capacity are a promising class of

materials that need further improvement in the pore volume and

pore size optimization.

Hydrogen Storage in Flexible MOFs: Hydrogen is widely ac-

cepted as an environmentally friendly and promising alternative

energy source. However, its extremely low critical temperature

and standard boiling point (33 and 20 K, respectively) pose sig-

niﬁcant challenges for its storage by liquefaction or compression.

Even under high pressures and low temperatures, H2has a rela-

tively low energy density per unit volume, making storage chal-

lenging.

DOE has set the gravimetric and volumetric working capacity

targets of 5.5 wt%; 40 g L−1for 2025, and 6.5 wt%; 50 g L−1as an

ultimate goal.[51]

Although several MOFs have high gravimetric working ca-

pacity reaching the DOE target, simultaneously high volumetric

www.advancedsciencenews.com www.advmat.de

Figure 7. High-pressure CH4physisorption isotherms for X-dia-1-Ni and

Wiley-VCH GmbH.

working capacity is still very challenging, and the current record

holder (NPF-200) has a working capacity of 37.2 g L−1between

100 and 5 bar at 77 K.[52]

To the best of our knowledge, only a small number of MOFs

have been reported where hydrogen adsorption induces gating

due to framework ﬂexibility at 77 K.

In 2008, Long and co-workers reported the proﬁles of hydrogen

high-pressure physisorption isotherms of [Co(bdp)]nbetween 50

and 87 K (Figure 11).[53]At 77 K and 40 bar, the excess adsorption

capacity amounts to 3 wt%, and the corresponding gate opening

and closing pressures are ≈20 and 10 bar. Unfortunately, the au-

thors reported only excess H2uptake values in their work and the

total adsorption and working capacities were not calculated.

Recently, McGuirk and co-workers[54]reported a

mixed linker approach successfully induced ﬂexibility in

[Cd(benzimidazolate)2]n(CdIF-13) framework, which shows

negligible uptake at 77 K in the pressure range up to 100 bar

(Figure 12). The 13% of 2-methyl-5,6-diﬂuorobenzimidazolate

(2M56DFbim) instead of benzimidazolate (b induces gate-

opening for H2. This stepped sorption proﬁle enables a usable

H2capacity of 1.17 wt%, which is, however, far below that of

[Co(bdp)]n.

MIL-53(Al) was demonstrated to show gate opening during hy-

drogen adsorption at 77 and 87 K.[55]Under vacuum and cryo-

genic conditions, the closed pore form is thermodynamically sta-

ble for desolvated MIL-53(Al), resulting in no H2adsorption in

the low-pressure range. According to the isotherm shape, the

opening of the structure and the transition to the intermediate

pore phase occurs at the pressure of ≈1 bar, followed by the open-

ing transition at ≈3 bar at 77 K (Figure 13).

Thus, in the case of H2storage, DOI-recommended values

have not been achieved to date. However, the examples re-

ported so far demonstrate that this area of research holds sig-

niﬁcant promise and is a highly encouraging direction for future

exploration.

Acetylene Storage: Acetylene (C2H2) is a crucial raw mate-

rial in industrial manufacturing and precision electronics. It

is typically stored in acetone, as storing it under pressures ex-

ceeding 1.5 bar can lead to polymerization and potential explo-

sion hazards.[56]However, the presence of acetone vapours poses

a challenge for applications that demand high-purity C2H2.[57]

Therefore, there is a need for a more suitable storage medium

for C2H2. Although extensive research has been conducted on

C2H2adsorption, most studies focus on atmospheric pressure,

and no materials for high-pressure storage have been reported.

For ﬂexible adsorbents, the inﬂexion point of the isotherm

typically appears well below 1.0 bar. Nevertheless, even un-

der low-pressure adsorption conditions, it is possible to iden-

tify highly eﬀective adsorbent materials. For instance, Kitagawa

and co-workers[58]utilized synchrotron X-ray powder diﬀraction

data to determine the structure of a [Cu2(pzdc)2(pyz)]n(pzdc =

pyrazine-2,3-dicarboxylate, pyz =pyrazine) MOF that adsorbs

C2H2molecules, discovering that C2H2adsorption can readily in-

duce signiﬁcant framework distortion. This compound exhibits

a high adsorption aﬃnity for C2H2(adsorption enthalpy 42.5 kJ

mol−1), with an adsorption capacity in saturation of 42 cm3g−1.

In 2009, Zhang et al. reported MAF-2 ([Cu(etz)]n,Hetz=3,5-

diethyl-1,2,4-triazole) MOF, which exhibited a sigmoid adsorp-

tion isotherm for C2H2. This facilitated the release of C2H2in

MAF-2 and increased its usable capacity. The single-crystal struc-

ture of C2H2-loaded MAF-2 revealed the formation of C2H2hex-

amers within the pores. At 298 K and pressures between 1.0 and

1.5 bar, the estimated C2H2uptake in MAF-2 reaches 17 cm3g−1

(equivalent to 20 cm3cm−3), which is over 40 times the usable ca-

pacity of an equal volume gas cylinder (0.5 cm3cm−3)(Figure 14).

Subsequently, the same research group reported that MAF-

123-Cd[59]exhibited even higher working capacity. At 298 K,

the C2H2isotherm was nearly linear, with an uptake of only

2.23 mmol g−1(54.514 cm3g−1or 83.55 cm3cm−3)at1bar.How-

ever, at 273 K, the isotherm displayed a clear S-shape, reaching

an uptake of 6.34 mmol g−1(142.06 cm3g−1or 217.85 cm3cm−3)

at 1 bar, indicating that a substantial usable capacity could be

achieved through temperature swing adsorption. Based on the

extrapolated isotherm (1.0–1.5 bar), the usable storage capacity

of MAF-123-Cd was calculated to be 1.3 mmol g−1(31.79 cm3g−1

or 48.73 cm3cm−3), which is 98 times higher than that of a stan-

dard gas cylinder (0.5 cm3cm−3).

In 2019, Zeng et al.[60]investigated C2H2adsorption on ﬂexi-

ble microporous JNU-1 ([Zn3(OH)2(btca)2]nbtca =benzotriazole-

5-carboxylate) MOF, possessing a high density of open metal

sites (Figure 15a). The isotherms display adsorption steps at

low pressures that gradually disappear with the increase in tem-

perature. Structural contraction of the pores upon the adsorp-

tion and induced-ﬁt eﬀect leads to a signiﬁcant rise in adsorp-

tion enthalpy. The binding between C2H2and the MOF is ex-

ceptionally strong, making complete desorption possible only

under high vacuum and elevated temperatures (Figure 15b).

Therefore, in MOF design, it is crucial to balance binding

aﬃnity with regeneration energy, which involves managing the

trade-oﬀ between thermal management and the ﬁnal usable

capacity.

Moreover, ﬂexible MOFs are highly likely to exhibit remark-

able potential for storing highly toxic gases (such as AsH3,

BF3, and PH3), which are critically important in the semi-

conductor industry. Currently, these gases are stored in cylin-

ders, requiring extremely high levels of sealing. Despite the as-

sociated risks, the necessity of using such highly toxic gases

www.advancedsciencenews.com www.advmat.de

highlights the signiﬁcant application value of ﬂexible adsorbent

materials.

2.2. Mixed Gas Adsorption and Separation Concept

Separation and puriﬁcation technologies account for up to 15%

of global energy consumption and nearly half of industrial energy

usage. Conventional gas separation technologies, such as distil-

lation, involve repeatedly evaporating and condensing mixtures

under harsh conditions. At the same time, liquid absorbents re-

quire heating and cooling large solvent volumes to release ab-

sorbed gases during regeneration. To reduce CO2emissions, for

example, methods like wet scrubbing of exhaust gases have been

applied in industry for many years.[61]The exhaust gases are led

through an absorber ﬁlled with amine solution, which is capable

of reacting with CO2and forming carbamates. These processes

are connected to high energy consumption, as the recovery of

absorbents is only possible through thermal desorption. Hence,

more energy-eﬃcient and sustainable gas separation processes

are needed, and diﬀerent methods and adsorbents have to be

developed.[62]

One of the most important separation problems at present

is the separation of CO2from other gases, making selective

CO2adsorption from diverse gas mixtures an important research

ﬁeld.[63]CO2separation from an exhaust gas stream is one of the

most challenging questions of the carbon capture process. The

separation of these two gases in MOFs is proposed as an alter-

native to conventional materials.[63,64]Thus, it has been shown

that MOFs would be able to signiﬁcantly reduce energy con-

sumption during regeneration, as the desorption of physisorbed

CO2requires less energy input as compared to recovery from

Figure 9. Comparison of gravimetric CH4sorption isotherms at diﬀerent temperatures for a) MIL-53(Al)-OH and b) MIL-53(Al)-(OH)2. Reproduced

www.advancedsciencenews.com www.advmat.de

Figure 10. CH4excess uptakes of MIL-53(Al)–NH2, MIL-53(Al), and MIL-

53(Al) with the BDC/BDC–NH2mixed ligand at 298 K (ﬁlled symbols -

adsorption; empty symbols - desorption). Reproduced with permition.[47]

carbamates. Comparing the so-called parasitic energy, which is

the energy penalty that is used to regenerate an adsorber or scrub-

ber after sorption, it becomes clear that MOF adsorbents pos-

sess a huge advantage in this prospect.[65]State-of-the-art amine

scrubbing technologies (with monoethanol amine; MEA) showed

a parasitic energy higher than 1000 kJ kg−1, whereas rigid MOFs

such as Mg-MOF-74 or Ni-(4PyC)2show much lower values of

727 and 655 kJ kg−1, respectively.[65,66]

The important indicators for eﬃcient adsorptive gas separa-

tion include also the working capacities of each component in a

mixture, enthalpies or heats of adsorption, but also selectivities

and mass transfer (adsorption kinetic).[67]Despite extensive re-

search in the ﬁeld of MOF-based gas separation/puriﬁcation in

more than the last two decades, the trade-oﬀ of adsorption capac-

ity versus selectivity is still a major challenge.[68]

Figure 11. Isotherms for the excess uptake of H2on [Co(bdp)]n, showing

temperature-dependent hysteresis loops at 50, 65, 77, and 87 K. Filled and

open symbols represent adsorption and desorption curves, respectively.

Society.

Figure 12. H2physisorption isotherms at 77 K for CdIF-13 (red) and

[Cd(bim)1.87(2M56DFbim)0.13]n(violet), exhibiting the reduced gate open-

ing pressure threshold for the multivariate MOF whereas CdIF-13 does

not show any gate opening in this operating pressure regime. Closed cir-

cles correspond to adsorption, and open circles correspond to desorp-

Society.

The main advantage of ﬂexible MOFs is the selectivity of the

structural transitions against the triggering ﬂuid. There are many

examples of MOFs and ﬂuids where the structural transition can

be triggered by a speciﬁc ﬂuid only at the speciﬁc conditions

(Figure 16b), or the pressure needed to provoke structural transi-

tion diﬀers for various ﬂuids.[10d,69]In the case of MOFs with “gate

opening” behavior, such phenomena could lead to incredibly

high (near inﬁnite) separation selectivity if such behavior could

be translated to the case of a mixture of ﬂuids (Figure 16a), and

some of them have been introduced as “perfect” adsorbents.[69]

It should also be pointed out that even in a mixture, in many

cases, a relatively high concentration of the “molecular opener”

(i.e., the guest stimulating the phase transition) is required, im-

plying that trace impurity separations based on ﬂexible MOFs are

less likely achievable.

The main concern, however, for ﬂexible MOFs is whether

the structural transition would suppress or enable the coadsorp-

tion of multiple components from the mixture (Figure 16), and

whether the separation factors predicted from the single compo-

nent isotherms can be transferred to the case of multicomponent

adsorption.

A visionary goal would be the development of switchable

MOFs in which the pores are accessible for one molecular species

only (able to provoke structural transition and enter the pores)

while the others in the mixture are kept outside (due to the un-

preferable free energy of the system containing multiple compo-

nents) (Figure 16a).

However, predicting the separation performance of ﬂexible

MOFs based solely on single-gas physisorption isotherms is chal-

lenging, as both scenarios—exclusion of a second component or

coadsorption—are theoretically possible, depending on the free

energy proﬁle of the MOF/adsorbate system. As a result, multi-

component adsorption measurements are the primary method

www.advancedsciencenews.com www.advmat.de

Figure 13. High-pressure H2physisorption isotherm of MIL-53(Al) measured at 77 K (black) and 87 K (red). Reproduced with permission.[55a]Copyright

2017, American Chemical Society.

for accurately evaluating the true separation capability of ﬂexible

MOFs.

To evaluate the separation performance of an adsorbent, the

adsorption capacities for the ﬂuids of interest are usually ex-

plored ﬁrst. The most common method for measuring the ad-

sorption capacity of MOFs for various gases is the static adsorp-

tion method. Here, the amount of gas adsorbed is determined

in a state of equilibrium by adding a deﬁned amount of gas to

the sample. Once equilibrium has been reached, the amount ad-

sorbed is determined either by the change in gas pressure (volu-

metric) or by the increase in sample mass (gravimetric).[70]This

method can also be used for gas mixtures, but in this case, the

composition of the gas should be controlled beside the pressure

change, and direct measurement of binary mixture equilibria is

more complicated and time-consuming.[71]

The researchers are still working to improve mathematical

models for predicting mixture adsorption. Indeed, from the ad-

sorption isotherms of individual gases, a prediction can be made

using ideal adsorbed solution theory (IAST) or Grand Canonical

Monte Carlo (GCMC) simulations.[72]However, even for rigid ad-

sorbents, the IAST cannot give information on deviations from

ideal behavior.[73]

The calculations for ﬂexible materials were not feasible for a

long time since the contraction and expansion of the network

could not be taken into account in these calculations, and the ap-

plied models were, therefore, invalid or subjected to signiﬁcant

errors.[72,74]In the past years, however, new methods, like the cou-

pling of adsorbed solution theories and the thermodynamic os-

motic ensemble, resulting in Osmotic Framework Adsorbed So-

lution Theory (OFAST)[75]or hybrid Monte Carlo and Molecular

Dynamics[76]simulations, have been developed to come up with

more accurate predictions.[76,77]The OFAST allows the prediction

of phase transition pressures upon coadsorption, but its major

drawback is that it relies on the IAST to describe adsorption in

each phase of the host material.[78]Thus, true selectivity is not

accessible by this model if the adsorption behavior deviates from

that expected by IAST.

It should also be noted that the OFAST model only deals with

the thermodynamic stability of the phases of the material at equi-

librium and yields no insight into the hystereses that are typically

observed experimentally. Therefore, the theoretical predictions

need to be veriﬁed by experimental data, leaving the most reli-

able method for determining sorption properties measuring the

mixed gas adsorption isotherms.[71]

Breakthrough experiments can be considered as essential

characterization technique for industrially relevant separation

processes. Under dynamic conditions, many new parameters

have to be considered to fully assert the suitability so that dynamic

gas sorption measurements become necessary.[70,79]A break-

through curve is a plot that reﬂects the concentration change of

the adsorptive in the eﬄuent stream at the outlet of a ﬁxed bed

adsorber, giving information not only about the adsorption capac-

ity of the adsorbent but also about the kinetics of the adsorption

processes.

Therefore, dynamic sorption experiments are one of the fre-

quently employed methods for investigating gas separation prob-

lems. The behavior of ﬂexible, phase-changing adsorbents un-

der dynamic breakthrough conditions can undergo several phase

transitions, yielding nonconventional breakthrough proﬁles.[80]

The challenges for applications of ﬂexible compounds in a bed

are, amongst others, discussed below.

There are numerous highlights in the ﬁeld of separation of

diverse molecules on ﬂexible MOFs, but due to limitations and

the need to stay focused, we will limit ourselves to a few examples

(such as some hydrocarbons, CO2/CH4and D2/H2) here.

2.2.1. Examples of Flexible MOFs for Separation

CO2/CH4: The carbon dioxide/methane separation ability of

ﬂexible MOFs is among the characteristics that have been heavily

investigated. Signiﬁcant diﬀerences in the adsorption enthalpies

of the two gases render it likely that the framework will respond

diﬀerently to each during the adsorption.

www.advancedsciencenews.com www.advmat.de

Figure 14. a) Semilogarithmic physisorption isotherms of C2H2(squares)

and CO2(circles) for MAF-2 at 195 K. The ﬁlled and open symbols repre-

sent adsorption and desorption, respectively. b) Volumetric C2H2uptakes

of MAF-2 and other typical materials at room temperature. Two dashed

lines represent the practical working limit of charging and discharging

pressures. The curve marked as MP is shown for a hypothetical micro-

porous adsorbent (following Langmuir isotherm) having the same uptake

as MAF-2 at 1 atm and p0. Reproduced with permission.[57]Copyright

2009, American Chemical Society.

Mixed gas CO2/CH4adsorption was intensively investigated

experimentally and supported by theoretical calculations for MIL-

53(Cr) up to 25 bar at 303 K.[81]Upon adsorption of pure CO2, the

MOF undergoes a breathing-type structural transition.

The adsorption of CH4does not provoke structural transi-

tion in MIL-53(Cr) in the investigated pressure and temperature

range, showing the isotherm typical for microporous rigid ad-

sorbent (Figure 17a). The results of mixed gas adsorption experi-

ments made clear that the composition of the employed gas mix-

ture does have a substantial eﬀect on the ﬂexibility and capacity

of the material. The total amount adsorbed was reduced with an

increasing content of CH4compared to the single gas adsorption

of CO2(Figure 17a). In situ Raman spectroscopic studies of the

narrow pore and open phases revealed that both gases are coad-

sorbed, showing that there is a signiﬁcantly higher aﬃnity toward

CO2, but in the presence of CH4, coadsorption is not completely

hindered.[81]

For MIL-53(Al), Coudert et al.[75a]determined for each com-

position of the CO2/CH4mixture whether breathing occurs and

transition pressures of op to np as well as np to op phase by

solving the OFAST equations numerically. Figure 17b demar-

cates the existence domains of the op (blue) and np forms (col-

orful) as well as selectivity in the (x(CO2), p) phase diagram of

the mixture adsorption. The np phase can be seen as a high-

selectivity island (with values of selectivity in the range of a few

tens), separated from the lower-selectivity background that is the

op phase. The pressure–composition phase diagram resulting

from OFAST predictions is consistent with the experimental re-

sults of Finsy et al.,[82]where two distinct selectivity mechanisms

could be identiﬁed. Up to 5 bar at 303 K and equimolar gas mix-

ture, the adsorption of CO2is dominated by strong speciﬁc inter-

actions with the np framework and average separation factors of

≈7. Above 6 bar, the average separation factor decreases to 4, due

to the framework opening.

Denayer and co-workers proposed a general methodology to

model the behavior of ﬂexible MOFs under dynamic break-

through conditions.[80]Each phase was modeled as a rigid adsor-

bent using a suitable adsorption model in correspondence to the

method proposed by Coudert et al.[83]The phase diagram was

modeled by a sudden transition (SGE method) or a smoother

s-function (PDE method). The s-function varies between 0 and

1 and can be physically correlated to the (mass)fractions of the

occurring phases. The model was used to qualitatively simulate

the breakthrough curves reported by Hamon et al.[75,81]The ex-

perimental and simulated F-proﬁles are shown in Figure 17d.

It was shown that the op to np transformation must be explic-

itly accounted. The feed conditions trigger the MIL-53 to contract

Figure 15. a) Structure of JNU-1. b) C2H2adsorption isotherms at diﬀer-

VCH GmbH & Co. KGaA.

www.advancedsciencenews.com www.advmat.de

Figure 16. a–c) A key scientiﬁc question regarding mixed gas adsorption on ﬂexible MOFs: How do the mixed gas adsorption isotherms for individual

components behave in the pressure range above the phase transition pressure? a) Only one component (ﬂuid triggering the transition) is adsorbed; and

b) both components are coadsorbed, following the adsorption run of a prototypical rigid MOF with the same pore size and composition. c) Schematic

from the op to the np phase in equilibrium with the feed state,

and the breakthrough proﬁle shows an associated step. Simula-

tions could capture a double roll-up, noticed in the experimental

breakthrough proﬁle, and reveal the fractions of both adsorbent

phases.

The selectivity for CO2and CH4mixture was investigated for

[Co(bdp)]nMOF (Figure 5a).[69]This MOF shows the gate open-

ing behavior for both gases in the single-component adsorption

experiments, but the opening transition for CH4occurs at a

much higher pressure than that for CO2(Figure 18) and suggests

that [Co(bdp)]ncould be highly selective for CO2at pressures be-

tween those corresponding phase transition pressures.

The composition of the adsorbate in the bicomponent

CO2/CH4(1:1) equilibrium adsorption experiment was moni-

tored using mass spectrometry. The hypothesis based on the

single-component isotherms could be indeed conﬁrmed, and

[Co(bdp)]nadsorbs approximately no CH4at the examined pres-

sures and the framework is most accurately described as having

near-perfect CO2selectivity under these conditions (Figure 18a).

But, when the framework was exposed to a 6:94 molar ratio

of CO2/CH4at 58.6 bar (the region where the single compo-

nent experiments suggest the opening for both gases), the ma-

terial remained selective for CO2(adsorbing 8.5 mmol g−1), but

a signiﬁcant amount of CH4(2.1 mmol g−1) is also coadsorbed

(Figure 18b).

DUT-8(Ni) ([Ni2(2,6-ndc)2dabco]n, 2,6-ndc =2,6-

naphthalenedicarboxylate, dabco =1,4-diazabicyclo[2.2.2]-

octane), is an interesting and widely investigated model “gate

opening” pillared layer MOF, showing selectivity in the gate

opening for a large variety of molecular species in the gas and

liquid states.[10d,84]It has been shown that the network has a

high aﬃnity toward CO2compared to other gases, including

CH4, only showing the phase transition for CO2at 298 K up

to 60 bar.[84b]To investigate the adsorption selectivity in the

CO2and CH4mixture, the adsorption experiments were per-

formed with the starting gas composition of 75:25 v/v at 215 K.

The compositions of the adsorbed phase and gas phase were

simultaneously monitored in situ by 13C NMR spectroscopy

(Figure 19a,b).[43]To understand the inﬂuence of ﬂexibility on

the separation performance, comparable experiments were also

conducted with the rigid version of DUT-8(Ni). The ﬂexible

and rigid samples diﬀer merely in the particle size. Analyzing

the observed adsorbed and gaseous species of both gases in

the resulting NMR-spectra, it was proven that the ﬂexibility of

DUT-8(Ni) causes nearly perfect selectivity, which is not the case

for rigid MOF, showing signiﬁcant uptake for both gases.

In situ 13C NMR spectroscopy was also applied to investigate

the CO2/CH4selectivity in the adsorption of a 75:25 mixture at

215 K for SNU-9.[43]It is capable of changing its pore structure

between the np and op states during the adsorption of CO2at

195 K through the formation of an intermediate (ip) phase. In

contrast, CH4does not initiate such transitions at 195 K, at least

up to a pressure of 2 bar (Figure 19c).

Interestingly, the phase transition of SNU-9 induces a signiﬁ-

cant change in adsorbed gas composition. Stepwise pressure in-

crease upon adsorption results in a steeply increasing amount of

adsorbed CO2upon the structural np–ip–op transition. The ab-

solute amount of methane coadsorbed with CO2is relatively low

in the np state. It increases in the ip state and decreases after the

transition into the op state, thus passing through a maximum in

the ip state (Figure 19d). At higher pressures, the NMR signal

of adsorbed CH4decreased in intensity while the gaseous CH4

signal increased, indicating the desorption of CH4.Incontrastto

that, no desorption of CO2was observed, resulting in a selectivity

varying strongly with the current phase of the MOF.

Schneemann et al. investigated the coadsorption on a series

of substituted [Zn2(bdc)2(dabco)]n(D-MOF) MOFs belonging to

the pillared layer materials.[18a]D-MOF itself does not show

pronounced ﬂexibility upon adsorption of methane or carbon

dioxide.

By exchanging bdc by DiP-bdc (2,5-diisopropoxy-1,4-

benzenedicarboxylate), however, ﬂexibility could be unlocked.

The single-component isotherms show that the structural tran-

sition can be induced by CO2only, up to 30 bar investigated in

the experiment. In the ﬁrst run of the coadsorption experiment,

the partial CO2pressure was chosen to be slightly below the np

→op transition pressure and during the second experiment, the

partial CO2pressure was above the np →op transition pressure.

From the ﬁrst experiment (Figure 20a), it could be seen that the

adsorbed amount of CO2during the coadsorption experiments

is much higher than for the single component adsorption at

the same pressure (23.47 compared to 6.5 cm3g−1), while the

CH4uptake is only slightly increased (6.5 compared to 4.6 cm3

g−1), suggesting that a part of the sample has been already

www.advancedsciencenews.com www.advmat.de

Figure 17. a) Total adsorbed amount of CO2/CH4mixtures on MIL-53(Cr) at 303 K. Red diamonds: pure CO2, pink squares: 75/25 CO2/CH4, green cir-

Society. b) Calculated CO2/CH4selectivity upon adsorption of a mixture in MIL-53(Al), as a function of the total pressure and mixture composition. The

narrow pore phase corresponds to the central “island,” with high selectivity, while the op phase has lower selectivity. Reproduced with permission.[75 ]

for pure CO2(top panel), pure CH4(bottom panel), and an equimolar mixture of the two (middle panel). Black dashed lines: total adsorbed quantity;

Chemical Society; and breakthrough proﬁles simulated by partial diﬀerential equations (PDEs). Reproduced with permission.[80]Coryright 1016, Royal

SocietyofChemistry.

www.advancedsciencenews.com www.advmat.de

Figure 18. a) Multicomponent adsorption experiments for CO2/CH4mix-

tures in [Co(bdp)]nshow near-perfect CO2selectivity at 6.7, 13.9, and

25.3 bar, under equilibrium CO2/CH4molar ratios of 46:54, 42:58, and

43:57, respectively. b) Multicomponent adsorption experiment performed

under a CH4-rich atmosphere (with an equilibrium CO2/CH4molar ratio

of 6:94) shows that [Co(bdp)]nadsorbs only a small amount of CH4at

this ratio, leading to a selectivity of 61 ±4. (a,b) Purple diamonds rep-

resent the overall amount of gas adsorbed by [Co(bdp)]n(y-axis) from a

CO2/CH4mixture at a given equilibrium pressure (x-axis). Each purple

diamond is paired with a corresponding red and blue star: red stars rep-

resent the CO2adsorbed from the mixture (y-axis) at the equilibrium par-

tial pressure of CO2(x-axis), and blue stars represent the CH4adsorbed

from the mixture (y-axis) at the equilibrium partial pressure of CH4(x-axis).

Single-component isotherms of CO2(red circles) and CH4(blue circles)

American Chemical Society.

transformed to op state upon experiment. In the second exper-

iment (Figure 20b), it can be clearly seen that the CO2uptake

coincides with the CO2uptake from the single-component

experiment and the CH4is coadsorbed.[85]

The adsorption of CO2/CH4mixture on JUK-8

(([Zn(oba)(pip)]n,oba=4,4′-oxybis(benzenedicarboxylate),

pip =4-pyridyl-functionalized benzene-1,3-dicarbohydrazide)

was intensively investigated by Roztocki et al. by a combination

of diﬀerent techniques.[86]JUK-8 possesses a gating type of

ﬂexibility, where the structural transition from cp to op phase

is induced by CO2(Figure 21a). In situ 13C NMR studies at

195 K point on the coadsorption of both gases in the open pore

phase at 5 bar for 1:1 CO2/CH4ratio. At temperatures above

280 K, however, the near-perfect selectivity is achieved, and

only CO2seems to be adsorbed from a CO2/CH4(75:25 v/v)

mixture.

Recently, CO2/CH4[22]separation was extensively investigated

for gate opening ELM-11 by Hiraide et al. for application in pres-

sure vacuum swing adsorption systems. The CO2adsorption

isotherm at 298 K shows two steps. The ﬁrst step (cp to np transi-

tion) occurs at the relative pressure of 10−3(0.75 bar). The second

step (np to op transition) was observed at p/p0≈0.3 (19.2 bar)

(Figure 22c). Methane adsorption isotherm of ELM-11 at 303 K

indicates the opening transition at ≈40 bar and closing transition

upon the desorption is ﬁnished at ≈20 bar).[34]Thus, the open-

ing of the cp phase to the np phase can be provoked by carbon

dioxide but not by methane, and the operation window between

1 and 19 bar should guarantee high selectivity (Figure 22). At

pressures above 20 bar, the compound is expected to adsorb both

components.

To tune the separation performance of ﬂexible MOFs, the

building blocks of the MOF can be adjusted, such as metal in

the cluster or substituent on the linker, as discussed above, but

also the temperature can be adapted since the temperature be-

longs to important thermodynamic parameters and strongly in-

ﬂuences the free energy landscape of the given system.[22,37]In

2020, Dong et al. demonstrated that the separation performance

can be optimized through tuning gate-opening pressure by tem-

perature (Figure 23) on the example of ternary mixture separation

in ﬂexible NTU-65 ([Cu(L1)2SiF6)]n,⋅L1 =1,4-di(1H-imidazol-1-

yl)benzene).[88]

The authors showed that by varying the adsorption tempera-

ture, the uptake was strongly inﬂuenced, allowing for high up-

takes of CO2and C2H2but decreasing the amount adsorbed

of C2H4when optimized. Using the obtained information from

static single gas adsorption, the separation of the ternary mixture

(C2H4/C2H2/CO290/1/9 v/v/v) was possible at the optimized

temperature of 263 K yielding high purity >99.99 % for C2H4

after only one cycle.

Light Hydrocarbon Separation (C2–C3): Light hydrocarbons

are not only important energy resource but also important raw

materials for ﬁne chemicals. The similarity of boiling points,

kinetic diameters, dipole moments, and other physical prop-

erties makes the separation and puriﬁcation of light hydrocar-

bons challenging, making the applications of ﬂexible MOFs very

attractive.[89]

The adsorption of methane, ethane, propane, ethylene, and

propylene was studied using ZIF-7 ([Zn(bim)2]n, bim =benzimi-

dazolate) at 298 K.[90]Except for methane, all the gases show gat-

ing isotherms, with distinct diﬀerences between the gate opening

and closing pressures, leaving a window of selective uptake oper-

ation. It could be shown in breakthrough experiments that ethane

is selectively adsorbed over ethylene in their mixtures, which re-

sults in the direct production of pure ethylene.

It makes ZIF-7 a perfect candidate for the separation of oleﬁns

from paraﬃns since, in contrast to most microporous materials,

the paraﬃn is selectively adsorbed.

www.advancedsciencenews.com www.advmat.de

Figure 19. CO2and CH4single-component physisorption isotherms for a) DUT-8(Ni) and b) SNU-9 measured volumetrically at 195 K. Signal intensities

for adsorbed 13CO2and 13 CH4measured by in situ 13C NMR spectroscopy for c) DUT-8(Ni) and d) SNU-9 pressurized with a 13 CO2/13CH4mixture

(75:25) at 215 K. A full adsorption (ﬁlled symbols and desorption (empty symbols) cycle up to 7.6 bar is shown. No signal of adsorbed CH4is detectable

Figure 20. Depiction of the excess single component and coadsorption measurements for [Zn2(DiP-bdc)2(dabco)]n:CO

2(red), CH4(black), total

www.advancedsciencenews.com www.advmat.de

Figure 21. a) CO2physisorption isotherm of JUK-8 at 195 K and the crystal structures of corresponding cp and op phases. b) 13C NMR spectrum of

the 13CO2/13 CH4gas mixture at 5.5 bar and 195 K. The signal at −10 ppm (grey bar) corresponds to gaseous methane, at −4 ppm to the adsorbed

methane, at 127 ppm to the gaseous carbon dioxide (orange bar), and the broad signal between 90 and 180 ppm to the adsorbed carbon dioxide.

c) Single component methane and carbon dioxide physisorption isotherms at 298 K. d) Isothermal multicomponent adsorption experiments for CO2/CH4

is licensed under CC-BY 4.0.

The potential of [Cu(dhbc)2(4,4′-bipy)]nfor C1–C3 hydrocar-

bon hydrocarbon separation was tested at 298 K.[91]The single-

component isotherms of each C2–C3 hydrocarbon display gating

and diﬀer in gate opening pressures, which inversely correlate

well with the latent heat of vaporization of the corresponding hy-

drocarbon. C3H4has the lowest gate opening pressure, which in-

dicates that the ﬂexible framework can highly selectively adsorb

C3H4at a very low pressure. Several breakthrough experiments

were performed using equimolar C3H4/C3H6/C3H8mixtures at

298 K. C3H6and C3H8break ﬁrst, while C3H4breaks through af-

ter some period of time. C3H4can be eﬀectively separated from

the C3H4/C3H6/C3H8mixtures in a nearly pure form with gas

phase concentrations of more than 99.9%.

[Co(vttf)]n(vttf =2,2′-[1,2-bis(4-benzoate)−1,2-

ethanediylidene]bis-1,3-benzodithiole)) reported by Kitagawa

and co-workers, exhibits exclusive gate opening for ethylene,

potentially enabling the discriminatory adsorption of it over

ethane.[92]In the close pore phase, the compound is nonporous

and features crosslinking via the coordination of tetrathiafulva-

lene sulfur atoms with the axial sites of the paddle wheels. The

framework is not responsive to ethane, but the ethylene is able

to coordinate cobalt and, therefore, induces a phase transition,

displacing the tetrathiafulvalene linkers and yielding an open

structure. Once open, however, the framework adsorbs both

ethylene and ethane, resulting in only modest selectivities.

The adsorption ability of triply interpenetrated, ﬂexible SD-

65 ([Zn(NO2ip)(dpe)]n,NO

2ip =5-nitroisophthalate, dpe =1,2-

di(4-pyridyl)ethylene) was studied for seven C4hydrocarbons, in-

cluding 1,3-butadiene, trans-2-butene, cis-2-butene, n-butane, 1-

butene, isobutene, and isobutane at 298 K.[93]Theuptakeofall

butenes and butanes was negligible at 1 bar, but the 1,3-butadiene

shows a gate-opening sorption proﬁle with a gate-opening pres-

sure of 60 kPa, gate-closing pressure of 50 kPa and 40 cm3g−1

uptake at 101 kPa. Therefore, SD-65 can separate 1,3-butadiene

from C4 hydrocarbon mixtures and readily release adsorbed 1,3-

butadiene under the PSA conditions.

H2/D2: Deuterium, constituting 0.016% of total hydrogen

occurring in nature, is a scientiﬁcally and industrially relevant

molecule, but its separation from H2is challenging due to its

similar physical properties.

www.advancedsciencenews.com www.advmat.de

Figure 22. a) Single-component adsorption isotherms of CO2and CH4at 298 K on ELM-11 (ﬁlled symbols: experimental adsorption data, open symbols:

experimental desorption data, lines: adsorption data simulated by GCMC). b) Total and component adsorption isotherms of an equimolar mixture of CO2

and CH4at 298 K for ELM-11 simulated by GCMC. The abscissae correspond to the partial pressures of CO2.Reproduced with permition.[22]Copyright

2020, The Authors, Published by Springer Nature Limited under CC BY 4.0. c) CO2sorption isotherms of ELM-11 measured at diﬀerent temperatures.

Green, 195 K; light blue, 253 K (p0=1.96 MPa); blue, 273 K (p0=3.47 MPa); red, 298 K (p0=6.4 MPa). Open and ﬁlled symbols indicate sorption

Physisorption of H2and D2at diﬀerent temperatures has been

studied in detail for MIL-53(Al) (Figure 24).[55,94]The experiments

show that the selectivity for D2over H2is strongly related to the

state of the pore structure of MIL-53(Al). At temperatures below

≈150 K, the solvent-free MIL-53(Al) forms the closed pore phase

(unit cell volume V=857 Å3), inaccessible for adsorptives.[95]A

two-step transition is observed upon adsorption of D2at the boil-

ing point temperature, where the structure transforms from the

cp to the intermediate pore phase (ip2, V=1349 Å3) and fur-

ther to the open pore phase (V=1531 Å3)(Figure24b). Upon

hydrogen adsorption, however, the framework transforms only

to the ip2 phase at the corresponding boiling point temperature

(Figure 24a). During desorption, additional intermediate phases

are formed, which is evident from the multiple steps in the des-

orption branch. In general, 130 diﬀerent structures of the MIL-53

family are deposited in the Cambridge Structural Database,[96]

showing the complexity of the free energy landscape of this

compound.

The selectivity of H2/D2adsorption was studied by thermal

desorption spectroscopy (TDS), at which the MOF was sub-

jected to the 10 mbar of 1:1 mixture at 25 K. The experiment

shows no outstanding selectivity because both gases transform

the MOF to the thermodynamically stable ip2 phase at these

conditions.

Nevertheless, optimizing the conditions can lead to an in-

creased selectivity of up to 10.5 at 10 mbar and 40 K. At these

conditions, the breathing of the network is in progress, leading to

the optimized environment for quantum sieving (Figure 24d–f).

The ability to separate hydrogen isotopes was also investigated

for the gate-opening DUT-8(Ni) (Figure 25).[97]Low-temperature

adsorption of H2and D2gas, each at their standard boiling

points, reveals pressure-dependent responsivity toward D2.The

adsorbed amount of H2is barely above the detection limit with

a slight increase up to 1 mmol g−1at1bar.Thus,DUT-8(Ni)re-

mains in the cp phase under the H2atmosphere. In stark con-

trast, after little to no uptake at low pressure, there is steep ad-

sorption of D2at 0.24 bar, the gate-opening pressure resulting

in a huge saturation uptake of 41.2 mmol g−1at 1 bar. The des-

orption branch indicates a hysteresis, as expected for a ﬁrst-order

structural transition.

TDS after exposure of the sample to H2/D2(1:1) isotope mix-

ture shows that the D2uptake is approximately ten times higher

than for H2, and the average selectivity is 9.9. The D2uptake in-

creases with increasing total pressure of the mixture to 0.8 bar

www.advancedsciencenews.com www.advmat.de

Figure 23. Illustration of the temperature optimization for ternary mixture separation by ﬂexible MOFs. At low temperatures, all three gases can be

adsorbed, which leads to the coadsorption of target gas and low productivity. At optimized temperature, two impurities are selectively adsorbed, and

pure target gas is obtained in high yield. At high temperatures, only one type of impurity is adsorbed, therefore, the product is not pure. Reproduced

and shows the highest value of 9.44 mmol g−1together with the

best selectivity of 11.6. Nevertheless, the experiments show that

both gases are coadsorbed at the chosen conditions.

2.3. How to Maximize Selectivity in Flexible MOFs

Analyzing the experimental examples of various mixed adsorp-

tion experiments, we propose a hypothesis, answering the ques-

tion posed in Figure 16: whether the structural ﬂexibility would

suppress or enable the coadsorption of multiple components

from the mixture, what is the underlying reason, and what are

the prerequisites for highest selectivity? Analyzing the cases re-

ported, it can be seen that both scenarios are possible: near-

perfect selectivity as well as coadsorption of both components.

Obviously, the reason is hidden in the isotherms runs for

the individual components, or more precisely, in the over-

lay of two isotherms and thermodynamic stability of the

framework/gas_𝛼/gas_𝛽system at experimental condition cho-

sen for coadsorption/separation experiments.

It is important to consider the hysteretic behavior of the

isotherms when designing the separation processes. The hystere-

sis occurs due to the activation barriers separating the phases

with energetic minima and energy penalty associated with the

nucleation and increasing interfacial area upon the phase trans-

formation.

The desorption branch of the gating isotherms is closer

to thermodynamic equilibrium, while the adsorption branch

reﬂects the system accessing metastable states. Thus, the

kinetic barriers determine the opening pressure of the

framework, while the closing pressure determines the

thermodynamic stability of the framework ﬁlled with the

adsorbate.

In the case of breathing, the contraction of the op phase to

the np phase is characterized by a much lower barrier than the

opening (cp to op, or np to op transitions upon the desorption).

Thus, in this case, the adsorption branch upon the ﬁrst breath-

ing transition is closer to thermodynamic equilibrium, while the

desorption branch reﬂects the system accessing metastable state.

Thus, the kinetic barriers determine the expansion pressure

of the framework, while the contraction pressure determines

the thermodynamic stability of the framework ﬁlled with the

adsorbate.

Below some particular cases are considered.

www.advancedsciencenews.com www.advmat.de

Figure 24. Hydrogen (red) and deuterium (blue) physisorption isotherms of MIL-53(Al): a) H2at 20 and 25 K, b) D2at 23 and 25 K, c,d) H2at 40 K, e) D2

2.3.1. Case I: Gate Pressure MOF

The gate opening pressure (pgo) for adsorptive 𝛼is lower than the

gate closing pressure (pgc) for adsorptive 𝛽(Figure 26a). In this

case, three potential pressure ranges exist.

1) Below p(𝛼)go , where the adsorbed amount is neglectable due

to the thermodynamic (p<p(𝛼)gc) or kinetic (p(𝛼)gc <p<

p(𝛼)go stability of the cp phase.

2) Between p(𝛼)gc and p(𝛽)gc is the region of nearly perfect se-

lectivity, where the free energy of the open phase is favor-

able only if the pores contain the adsorptive 𝛼. The partial

pressure of component 𝛼in the experiment should, however,

be larger than p(𝛼)go to initiate the structural transition. (For

example, see the behavior of DUT-8(Ni)[43](Figure 19a,b)or

[Co(bdp)]n[69](Figure 18a)inCO

2/CH4adsorption.)

3) Above p(𝛽)gc, the selectivity is expected to drop to the value

characteristic for the virtually rigid framework with the same

Figure 25. a) H2and D2isotherms for DUT-8(Ni) at 20.3 and 23.3 K. b) TDS spectra for DUT-8(Ni) and D2/H2selectivities (SD2/H2) according to the

Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. Distributed under CC BY-NC 4.0.

www.advancedsciencenews.com www.advmat.de

Figure 26. Hypothetical single component isotherms for a,b) hypothetical gate-opening or c) breathing MOF (op to np transition) (upper row) and

the expected selectivity in the case of an equimolar mixture of the components (middle row). The adsorption is shown as a solid, colored line, and the

desorption is shown as a dashed line of the corresponding color. The grey dashed lines represent the isotherms of the hypothetical “rigid” MOFs with

comparable porosity. The star marks one of the optimal operation pressures. The bottom row shows the thermodynamically stable state of the MOF in

the presence of each gas in the corresponding chemical potential range.

pore size and chemistry. (For example, see the behavior of

[Co(bdp)]n[69](Figure 18b)inCO

2/CH4adsorption at low CO2

concentration.)

2.3.2. Case II: Gate Pressure MOF

The gate closing pressure (pgc) for adsorptive 𝛽is lower than the

gate opening pressure (pgo) for adsorptive 𝛼(Figure 26b).

The pressure regions deﬁning the adsorption selectivity are the

same as in Case I, but the window of the highest selectivity is

much narrower. Moreover, the pressure needed for the frame-

work opening (exemplarily marked by the star in Figure 26b)is

necessarily outside the selectivity window, making the separation

not feasible. Thus, the adsorption selectivity is expected to be sim-

ilar to that of a hypothetical rigid MOF, because the open pore

phase at pressures above p(𝛽)gc is thermodynamically stable in

the presence of both individual components. Case II is expected

for high activation barriers (broad hysteresis) as well as for guests

with similar adsorption enthalpies.

It should be mentioned that if the physisorption isotherms of

the individual components are measured in the limited pressure

range only (below p(𝛽)go), the picture can pretend Case I, leading

to a wrong assessment of the expected selectivity.

2.3.3. Case III: Breathing MOF

The breathing transition usually involves two transformations:

the contraction of the more porous phase in the presence of

a certain amount of adsorbate and the expansion (gate open-

ing) to the more porous phase at higher pressures upon adsorp-

tion. Upon desorption, the framework contracts to the np phase

ﬁrst and, at lower pressure, expands to the initial, guest-free

framework.

For simplicity, we only consider the case where the pressure

of the ﬁrst transition (contraction) at characteristic contraction

pressure of 𝛼component (p(𝛼)c), is much lower than the contrac-

tion pressure for the 𝛽component (p(𝛽)c) and the pressures of

the gate opening (p(𝛼)go and p(𝛽)go and closing (p(𝛼)gc and p(𝛽)gc)

in the second transitions (gating).

The activation barrier of the contraction transition is usually

very low. The expansion upon the desorption, however, is kinet-

ically hindered, resulting in the hysteresis between adsorption

and desorption branches and pex <pc.

If the contraction pressure for adsorptive 𝛼(p(𝛼)c)islower

than the expansion pressure for adsorptive 𝛽(p(𝛽)ex), then

(Figure 26c):

1) At pressures below (p(𝛼)ex) the selectivity is approaching that

of the op phase. (For example, see CO2/CH4separation on

MIL-53(Al)[75]at low pressure, Figure 17.)

2) In the pressure range between p(𝛼)ex and p(𝛽)ex , where the

narrow pore phase is thermodynamically stable only in the

presence of adsorbate 𝛼, the preferable adsorption of compo-

nent 𝛼is expected. (For example, see CO2/CH4separation on

MIL-53(Al)[75a,81]at moderate pressure, Figure 17.)

3) At pressures above (p(𝛽)ex) and below p(𝛼)gc, np phase can be

stabilized by each component, (𝛼or 𝛽), therefore, the selectiv-

ity is approaching that of the virtual, rigid np phase.

www.advancedsciencenews.com www.advmat.de

If we would take into consideration the second gating phase

transition of the breathing (the np to op transition), situations

similar to that described in Cases I and II are feasible, as well as

more complicated cases, depending on the values of pc(𝛼), pc(𝛽),

pex(𝛼), pex (𝛽),p

go(𝛼), pgo (𝛽),p

gc(𝛼), and pgc (𝛽).

But considering the fact, that in the breathing system, the

closed pore phase (unporous state) does not exist, the maximum

selectivity of breathing MOF is expected always to be lower than

that of gate-opening MOF.

Thus, to achieve optimal separation performance, the transi-

tion pressures and hysteresis widths of the individual adsorp-

tion/desorption isotherms must be considered when specifying

the separation conditions. The use of the adsorption branch only

(particularly if the gate opening conditions for the second compo-

nent are not fulﬁlled) will lead to an overestimation of the sepa-

ration performance. Breathing MOFs are expected to have lower

selectivity in comparison to gate-opening MOFs due to the in-

termediate narrow pore phase and the absence of the highly se-

lective closed pore phase.[37b]The highest selectivity for the par-

ticular MOF can be achieved if the adsorption–desorption hys-

tereses of single components are clearly separated on the pres-

sure/concentration axis.

Hence, in order to predict selectivity, it is essential to report

isotherm adsorption and desorption data, ideally in a digital for-

mat based on IUPAC recommendations.[98]Such an approach

will give a wider community the opportunity to screen and eval-

uate potential applications if ﬂexible MOFs in a variety of ﬁelds

with AI tools and machine learning and validate the hypothesis

outlined here. Moreover, it would reduce ambiguity in adsorption

data using IUPACs established and well-deﬁned units for the ad-

sorbed amount (mmol g−1) and improve reproducibility.

3. Advanced Characterization Techniques

3.1. Monitoring the Framework Transitions

In this section, we emphasize the importance of a mechanis-

tic understanding of the guest-induced transitions of the ﬂexi-

ble host structure at the atomic level, which is important to un-

derstand the thermodynamics of the transition and follow the

changes in porosity and pore accessibility.

In the case of “gate-opening” MOFs, the crystal structures of

the guest-free and guest-ﬁlled phases can be determined from

ex situ measurements of solvated and degassed structures. How-

ever, such results should be interpreted with caution, as guest

molecules may inﬂuence the opening degree and do not con-

tain information about the transition pathway.[99]It is, therefore,

recommended to carry out in situ X-ray diﬀraction measure-

ments under conditions close to those used in the gas separa-

tion studies. In the case of minor volume changes (e.g., subnet-

work displacement or linker rotation[9]), in situ experiments can

be performed on the single crystals. Single crystal X-ray diﬀrac-

tion (SCD) allows to follow the changes in the framework and, in

many cases, to determine the position and occupancy of the guest

molecules in the pores. Such experiments can be carried out us-

ing laboratory single-crystal X-ray diﬀractometers equipped with

customized cells, adapted to speciﬁc gas loading conditions.[100]

In all other cases, in situ PXRD can be used.

In some cases, unexpected mechanisms behind the selectivity

can be revealed, for example, self-accelerated and selective CO

sorption in [Cu(aip)(H2O)]n(aip – 5-azidoisophthalate) reported

by Kitagawa and co-workers.[101]In situ PXRD studied in parallel

to N2and CO adsorption at 120 K (Figure 27) demonstrates the

mechanism of stepwise CO sorption, involving the formation of

aCu

2+─CO bond, and inducing a global structural transforma-

tion with the expansion of the squeezed paths. This expansion

promotes additional CO adsorption in the center of the channel,

which is the so-called self-accelerating gas adsorption.

Barbour and co-workers developed and optimized a cell for

in situ single crystal X-ray diﬀraction under static gas loading

conditions.[100]SCD was used to locate the preferable adsorption

sites in [NiNbOF5(pyrazine)2]n(NbOFFIVE-1-Ni) — one of the

state-of-the-art frameworks for direct CO2capture from air.[102]

Interestingly, the CO2molecule occupies an energetically favor-

able position, where the electropositive carbon of the CO2is sur-

rounded by four electronegative ﬂuorine centers from four dis-

tinct (NbOF5)2−pillars and the electronegative oxygen atoms of

the CO2are encaged by pyrazine hydrogens.

The cascade of the phase transitions was observed during

the desolvation of the adamantoid framework with a compo-

sition [NiL2]n(L =4-(4-pyridyl)-biphenyl-4-carboxylate).[39]All

structures were solved and reﬁned from the single crystal X-

ray diﬀraction data collected on the single crystal with CH2Cl2

in the pores at diﬀerent stages of degassing. As a result, three

diﬀerent structures, a1,a2,anda3, all crystallizing in the same

I41cd space group, were visualized. Desolvation of the structure

leads to reductions in cell volume (9528, 8637, and 7441 Å3for

a1–a3, respectively) and solvent-accessible void volume (49%,

43%, and 33% for a1–a3, respectively) induced by the changes

in N–Ni–N/C–Ni–C bond angles: 93.0°/101.5°, 90.7°/101.4°,and

88.0°/102.4°for a1–a3, respectively (Figure 28). The heating of

a1–a3 phases in vacuum yields the completely desolvated and

dense X-dia-1-Ni-c1 structure, showing only 2% of the solvent-

accessible void. In the same work, the authors followed the

CO2physisorption by PXRD, indicating the reversibility of the

transitions. Moreover, multistep isotherm was observed in high-

pressure methane physisorption, which served as a valid precon-

dition for the gas separation studies.

In the follow-up manuscripts, authors developed the strategy

for controlling “gate opening” pressure by utilizing a multivariate

approach and mixing Ni and Co in the structure.[40]Here, in situ

PXRD were collected upon high-pressure methane physisorp-

tion, indicating the phase transition of the framework. Finally, the

mixed-metal MOF with a composition X-dia-1-Ni0.89Co0.11 was

tested in C2H6/C2H4separation and indicated high inverse se-

lectivity toward C2H6with 9.1 times higher uptake compared to

C2H4. In situ PXRD experiments were conducted on X-dia-1-Ni

at 263 K, indicating the phase transition at 40 kPa upon C2H6ad-

sorption, whereas no changes were observed upon the physisorp-

tion of C2H4in the same conditions (Figure 29).

As discussed above, [Co(bdp)]ncompound is a promising can-

didate in terms of gas storage and separation. The structural

ﬂexibility was investigated by in situ PXRD in parallel to high-

pressure physisorption of pure CO2and CO2/CH4mixture at 298

K(Figure 30).

Tanaka and co-workers studied the rearrangement of ELM-11

framework upon CO2adsorption by time-resolved in situ PXRD

www.advancedsciencenews.com www.advmat.de

Figure 27. In situ PXRD/N2and CO sorption measurements on [Cu(aip)]n.a)N

2sorption isotherms at 120 K (insert: crystal structure of NbOFFIVE-

1-Ni). b) PXRD patterns measured at a–d points shown in the N2sorption isotherms in Figure 27a. The calculated pattern for the dried [Cu(aip)]nis

shown at the bottom. c) CO sorption isotherms at 120 K. Adsorption and desorption branches are shown in solid and open circles, respectively. d) PXRD

patterns measured at a–j points of the CO sorption isotherm shown in Figure 27c. The c patterns for the dried [Cu(aip)]nand CO-adsorbed phases are

at diﬀerent temperatures and threshold pressures to evaluate

the switching kinetics in the powdered sample (Figure 31).[22]

The authors conducted experiments at diﬀerent incrementally

increasing CO2pressures all above pgo, showing dependence be-

tween the rate constant and the p–pgo pressure diﬀerence, where

pis the CO2pressure introduced.

To study hydrogen isotope separation, neutron powder diﬀrac-

tion became a useful technique because of the high elastic scat-

tering length of deuterium. In such a case, both the information

about the host and guests can be extracted. Oh and co-authors

demonstrated that the physisorption of H2and D2in MIL-53(Al)

at cryogenic temperatures proceeds in diﬀerent ways.[94]While

hydrogen shows a one-step isotherm, two steps are observed

in deuterium isotherm, measured at the boiling point. Neu-

tron powder diﬀraction (NPD), conducted at deﬁned gas load-

ings, indicates contraction in the case of hydrogen. The same ex-

periment with deuterium indicates the contraction and reopen-

ing of the structure and shines a light on the mechanism of

selective isotope adsorption. Due to nuclear quantum eﬀects,

deuterium shows higher enthalpy of adsorption, which is suf-

ﬁcient to open the structure and keep it in an open pore state.

Besides NPD experiments, proving the structural changes, au-

thors conducted in situ quasi elastic neutron scattering experi-

ments, which showed that the mobility of the hydrogen in this

clamped mode is less than that when the molecule is in the open

structure.

In summary, the in situ scattering techniques provide unique

structural information on the dynamics of the host framework.

The obtained structures can be used not only for the evaluation

of geometrical porosity but also for more complex calculations,

such as GCMC, osmotic energy, and molecular dynamics simu-

lations. In some particular cases, information about the frame-

work dynamics and guests can be gained. However, much more

information on the guest’s behavior upon adsorption can be

www.advancedsciencenews.com www.advmat.de

Figure 28. Crystal structures of NiL2. a) Structure of the ligand. b) Adamantanoid cage in NiL2. c) Sixfold interpenetrated dia nets in NiL2. d) Rectangular

channels viewed along the c-axis. e) Single crystal structures of the porous (a1–a3) and nonporous (c1) phases of X-dia-1-Ni. f) Crystal morphology and

Wiley-VCH GmbH & Co. KGaA.

derived from spectroscopic techniques, in particular, one that can

diﬀerentiate between adsorptive and adsorbate.

3.2. Monitoring the Guest Behavior

The monitoring of the guest molecules u the adsorption process

in situ is crucial for understanding the role of the guests for the

ﬂexibility and adsorption process.

The main drawback of volumetric coadsorption techniques,

usually used to measure mixed gas adsorption isotherms, is the

limited temperature and pressure range, which is indeed quite

close to the real-world application for CO2/N2,CO

2/CH4,and

alkane/alkene separation. However, it technically does not cover

the cryogenic temperatures and low-pressure range typical for hy-

drogen isotope separation and has certain limitations in terms of

the gas mixture composition. In order to analyze the selectivity

at cryogenic temperatures, TDS can be used, where the gas mix-

ture is adsorbed ﬁrst, followed by applying an ultrahigh vacuum

to remove unadsorbed gas. The desorbed gases are then moni-

tored by mass-spectroscopy upon heating, which allows for the

estimation of the interaction strength and quantiﬁcation of the

species desorbed.[94,97,104]

In situ NMR is an advanced technique that diﬀerentiates and

quantiﬁes the ﬂuid in the gas and adsorbed phases, assuming

that all gases contain an NMR active nucleus. NMR oﬀers the

advantage of not only enabling quantiﬁcation but also providing

insights into the mobility of molecules.

Brunner and co-workers established a high-pressure in situ

cell connected to the homemade gas mixing unit.[97]The instru-

mentation was used to evaluate the coadsorption of CO2/CH4

mixture in DUT-8(Ni), SNU-9, and JUK-8 (as discussed in Sec-

tion 2.2.1).[43,86]The same setup was used for the analysis of

krypton/xenon coadsorption on DUT-8(Ni) at 280 K. 129Xe NMR

experiments on rigid and ﬂexible DUT-8(Ni) at 283 and 237

K using a mixture of xenon and krypton indicates a clear de-

crease in the chemical shift pointing on the partial replacement

of xenon by krypton, i.e., krypton coadsorption. Analysis of the

spectra suggests a slightly higher selectivity for xenon adsorbed

in the ﬂexible version of DUT-8(Ni). This can be explained by the

www.advancedsciencenews.com www.advmat.de

Figure 29. a) C2H6physisorption isotherm at 263 K. b) Corresponding PXRD patterns collected in situ. X-dia-1-Ni undergoes a reversible transformation

from an np phase to an lp phase upon ethane dosing. c) C2H4physisorption isotherm at 263 K. d) Corresponding in situ PXRD patterns. In situ PXRD

patterns indicate that the desolvated phase adsorbs C2H4,butC

2H6is not adsorbed until phase transformation occurs. Reproduced with permission.[103]

overlapping the hysteresis in Xe in Kr isotherms in the analyzed

pressure range; namely, once the “gate opening” pressure for

xenon is reached and the framework is open, coadsorption of

krypton will not change the thermodynamics, op phase remains

energetically favorable phase.

3.3. Combination of Diﬀerent Techniques

The most eﬃcient way to study the gas separation in ﬂexible MOF

is to apply multiple in situ techniques, as shown by Roztocki et al.

for JUK-8 in the mixed gas CO2/CH4adsorption (Figure 32).[86]

In order to explain the observed performance, in situ PXRD was

conducted at 195 K, indicating the phase transition from the JUK-

8_cp to the JUK-8_op phase. To characterize the selectivity in the

low-temperature regime, in situ 13C NMR coadsorption experi-

ment using a 13CO2/13 CH4mixture (molar ratio 1:1) at 195 K

was conducted. The analysis of the spectra indicates that only

minor amount of methane coadsorbs on JUK-8op@CO2, even

at low temperatures and high pressure. In situ IR spectroscopy

conducted under CO2loading suggests the changes in the host

structure, namely, changes in the O–C–O asymmetric stretching

region of the oba2−carboxylate linkers. In addition, two signals

from the adsorbed CO2(at 2340 and 2376 cm−1) were observed

in contrast to one signal observed for gaseous CO2(2345 cm−1).

In summary, the application of in situ diﬀraction techniques

is crucial for understanding the behavior of the host structure

and guest molecules upon the physisorption of the gas mix-

tures. The results provide important information about the host

structure under deﬁned conditions, which is essential for under-

standing the material performance and may be used as input

for in silico studies. Spectroscopic techniques, applied at desired

gas pressure/temperature conditions, can clearly diﬀerentiate be-

tween adsorbed and uadsorbed guest molecules and provide di-

rect proof of the separation performance.[55,97,105]

4. Challenges to Solve for the Application of

Flexible Adsorbents

4.1. Kinetics of Switching

When considering ﬂexible MOFs as adsorbents, not only the

working capacity and selectivity play a crucial role. The kinetics

of switching between the phases have to be taken into account

as well. The adsorption kinetics, in this case, is not only depen-

dent on the crystal surface barriers and gas diﬀusion kinetics but

arises from the framework bistability and is coupled to the phys-

ical properties of the guest molecules with characteristic barriers

of ﬂuid nucleation, diﬀusion, and adsorption processes.[106]Al-

though the importance of kinetics has been known for over 15

years,[107]the kinetics of the adsorption process in ﬂexible MOFs

is still underexplored.[108]

The most frequently applied method to analyze the kinetics of

gas-induced phase transition in an ensemble of crystals (powder

www.advancedsciencenews.com www.advmat.de

Figure 30. Powder X-ray diﬀraction data for [Co(bdp)]ndosedwitha)pureCO

2and b) a 50:50 mixture of CO2/CH4over a range of pressures. In both data

sets, the abrupt appearance or disappearance of peaks indicates discrete phase changes, whereas gradually shifting peaks indicate gradual framework

Figure 31. In situ time-resolved PXRD/CO2physisorption on ELM-11: a) fractions of the open phase and d) for the closed phase at 40.8 kPa and 241 K,

at 41.0 kPa and 264 K, and at 40.8 kPa and 273K. The numbers in (a) and (d) denote the temperature. b) Fractions of the open phase, and e) for the closed

phase at 227 K as a function of the CO2pressure. The numbers in (b) and (e) denote the CO2gas pressure in kPa. The curves after 4.15s in (a–e) were

obtained by ﬁtting the KJMA equation. c) Relationship between the rate coeﬃcients and the pressure diﬀerence between the CO2gas pressure, p,and

the gate-opening pressure, pgo. The error bar represents the standard deviation of the value obtained using the least-square ﬁtting of the KJMA equation

CC BY 4.0.

www.advancedsciencenews.com www.advmat.de

Figure 32. Evaluation of JUK-8 behavior in CO2/CH4adsorption: a) PXRD patterns collected upon CO2physisorption. b) Evolution of unit cell volume

upon CO2physisorption, and c) in situ 13C NMR in parallel to CO2/CH4(50:50) physisorption at 195 K. Reproduced with permission.[86 ]Copyright

2021, The Authors. Published by American Chemical Society. This publication is licensed under CC-BY 4.0.

or pellets) is to monitor the pressure drop in the adsorption cell.

This method is also applied in the industry to evaluate adsor-

bents, particularly for pressure swing adsorption, with high cy-

cling rates. However, this method does not analyze the intrinsic

switching but the macroscopic gas uptake of the sample, which

also involves mass transport and is therefore highly dependent

on the reactor volume and sample amount.

In such a way, Li and co-workers investigated the rate

of adsorption coupled with the structural transition in

a ﬂexible RPM3_Zn ([Zn2(bpdc)2(bpee)]n, bpdc =4,4′-

biphenyldicarboxylate; bpee =1,2-bipyridylethene).[109]The

compound exhibits a signiﬁcant induction period for opening by

N2and Ar at low temperatures (Figure 33). Such a long induction

period is not observed for H2or O2at comparable pressures

and temperatures, suggesting the rate of opening is strongly

inﬂuenced by the gas–surface interaction rather than external

stress.

The induction period leads to severe mass transfer limitations

for adsorption and overestimation of the gate-opening pressure.

The authors reviewed several adsorption rate models and found

that none adequately describes the experimental rate. Statisti-

cally, the rate data are best described by a compressed exponen-

tial function. The resulting ﬁtted parameters exceed the expec-

tations for adsorption but fall within those expected for phase

transition.[109]

By treating adsorption as a phase transition, the generalized

Avrami theory of phase transition kinetics was used to describe

adsorption in ﬂexible hosts.

Gläser and co-workers[110]investigated the kinetics of the n-

butane and iso-butane gas mixture adsorption on gate pressure

Cu-IHMe-pw ([Cu2(H-Me-trz-Ia)2])n[111]MOF. The uptake curves

reveal complex interactions on the outer surface of MOF parti-

cles. The overall rate of adsorption-induced structural transition

depends on several factors, including degree of pressure rise,

temperature, particles size, and the subsequent diﬀusion rate

into newly opened pores. With the aid of a kinetic model based

on the linear driving force (LDF) approach, both rates of diﬀu-

sion and structural transition were studied independently of each

other. The authors claim that the overall velocity of gas uptakes

in ﬂexible MOFs is predominantly determined by the rate of the

structural transition.

It could also be shown that the overall gas uptake is slower

for the gas mixture as compared to the single-gas uptake of n-

butane with the same partial pressure step, although much faster

than compared to the bare iso-butane adsorption (half coverage

is reached after 40 min for the mixture, 1 min for n-butane and

≈1000 min for iso-butane individually) (Figure 34). From the evo-

lution of the gas phase composition, it could be seen that n-

butane is predominantly adsorbed at the beginning, reducing

the total gas-phase fraction to ≈33%, resulting in a total sepa-

ration factor of a maximum of 10 (at time 200 min). Beyond

that, iso-butane continuously enters the opened framework, ex-

Figure 33. Nitrogen adsorption on RPM3_Zn at 77 K shows a pronounced

eﬀect of (ﬁctives) exposure time (dt): (A) dt=5s,(B)dt=20 s, and (D, E)

dt=360 s collected by ASAP 2020 from Micromeritics. (C) Collected using

a fresh sample and Autosorb-1 MP from Quantachrome Instruments. The

allowed time varied from 6 min to 3.7 h for adsorption. A history eﬀect

is also observed in which the number of data points aﬀects the uptake

(e.g., screening experiment D with 4 points versus full isotherm E with 45

points). Times (from the previous data point) for select points are Di =

67 h, Dii =8.5 h, Ei =62 h, Eii =20 h, and Eiii =14 h. The total times for

adsorption (in hours) are A ≈B(∼1) <C (16) <D (80) <E (135). Prior

to the sharp rise, curves C–E follow the A and B curves. Reproduced with

www.advancedsciencenews.com www.advmat.de

Figure 34. Kinetic of gas uptake (top), gas phase composition (bottom, left axis), and adsorption selectivity (bottom, right axis) in dependence of time

for an equimolar mixture of n-butane and iso-butane on Cu-IHMe-pw for a pressure jump of 0–40 kPa. Additionally, the overall gas phase compositions

and sorption kinetics are modeled by the “GO” model and a mass balance. Within the bottom graph, a selectivity larger than 1 indicates a preference

for n-butane, as indicated by the additional colored ribbons. It should be noted that a 50:50 molar mixture was aimed for, but actual results show slight

terms and conditions of the CC BY 4.0.

changes the adsorbed n-butane and incorporates itself within

the framework. This leads to a subsequent increase in the gas

phase fraction of n-butane and a ﬁnal separation factor of 0.9 af-

ter 10 000 min (≈7 days) for n-butane.

In recent years, new methods have been proposed to inves-

tigate the kinetic of phase transitions, involving in situ (single

crystal) XRD, optical, scanning electron microscopy (SEM), and

atomic force microscopy (AFM).

The phase transformation of a [Zn2(1,4-ndc)2(dabco)]n(1,4-

bdc =1,4-benzenedicarboxylate) crystal was investigated in situ

by AFM (with a time resolution of 40 s in DMF solutions with

various concentrations of the biphenyl as a guest molecule) by

the group of Kitagawa.[112]It was found that the lattice structure

of the liquid–solid interface quickly changed (within ≈10 min)

in response to the biphenyl concentration change. It should be

pointed out that in many cases, it is not only the analytical instru-

ment limiting the temporal resolution but also the realization of a

stepwise stimulus increase that poses challenges because of mass

transfer limitations and diﬀusional broadening of a pressure or

concentration pulse peak.

Phase changes were monitored on ≈1μm single crystals of

MIL-53(Al) by in situ environmental SEM.[113]

www.advancedsciencenews.com www.advmat.de

Figure 35. SEM images of MIL-53(Al) obtained at 0, 6, 12, 16, and 20 s of

beam irradiation; bottom right, particle width and length as a function of

vier Inc.

The deformation of single crystals was observed over several

seconds upon electron beam irradiation, which caused partial

desorption of the toluene from the pores. The phase transitions

were assigned through the detected variations in crystal shape

(Figure 35), which was further supported by TEM measurements.

Bon and co-workers demonstrated that the phase transition

kinetic in the MIL-53(Al) system is signiﬁcantly controlled by

the particle size (Figure 36).[114]Samples with the average par-

ticle size of 1.2 and 28 μm were compared upon adsorption of

n-butane at 298 K. It was also found that the contraction transi-

tion (op–np) is the fastest transition for both particle size regimes

and the reopening of the structure proceeds much slower. Hence,

the activation barrier for opening the narrow, butane-ﬁlled pore

is larger, probably due to the diﬀusion limitations of the guest.

The CO2-induced gate opening of ELM-11 was investigated by

time-resolved in situ PXRD (Figures 31 and 37), and a theoreti-

cal kinetic model of this process was developed to gain atomistic

insight into the transition dynamics.[115]

The developed model consists of the diﬀerential pressure from

the gate opening (indicating the ease of structural transition)

and reaction model terms (indicating the transition propagation

within the crystal).

The reaction model of ELM-11 is an autocatalytic reaction with

two pathways for CO2penetration into the framework. How-

ever, gas adsorption analyses of MIL-53(Al), possessing diﬀerent

mechanisms of phase transitions (breathing), indicate that the

kinetics of the adsorption-induced structural transition is highly

dependent on framework structure (Figure 38). The authors con-

ﬁrm the diﬀerences in the kinetics of op to np and np to op tran-

sitions, consistent with the work of Bon et al.[114]

The rate of structural transition of gating in ELM-11 investi-

gated by time-resolved in situ synchrotron Pin orderXRD mea-

surements shows that the structural transition started immedi-

ately after the introduction of CO2at 40.8 kPa and 273 K and

was accomplished in ≈10 s.[22]The MOF accommodating CO2

responded quickly to the decrease in gas pressure at 273 K: the

structural transition was completed in 5 s when the CO2pres-

sure was decreased at the rate of 2.4 kPa s−1. Moreover, the rate

of phase transition increases as the temperature decreases under

the same CO2pressure. At 227K the rate of structural transition

increased with increasing CO2pressure, and the phase transition

was completed within a few seconds at the highest gas pressure.

Furthermore, these data were found to obey the Kolmogorov–

Johnson–Mehl–Avrami (KJMA) equation.[22 ]

Kaskel and co-workers investigated the phase transition kinet-

ics of DUT-8(Ni) using three diﬀerent techniques.[106]The newly

designed microﬂuidic breakthrough apparatus allowed observa-

tion of transformation times for individual crystals varying with

respect to Ni/Co ratios. Comparison with the in situ PXRD data

led to the conclusion that the macroscopically observed switch-

ing rate is mainly governed by variations in the induction period.

The transformation of individual crystals is much faster than that

of an ensemble of crystals, which is, therefore, not easily observ-

able. Crystals with a higher activation barrier (corresponding to

a higher gate opening pressure) show a longer induction period,

leading to an overall slower adsorption kinetic of the ensemble.

However, the individual crystal transformation rate is below 1 s,

which was the temporal limit of resolution in this study. The ex-

act and intrinsic individual crystal transformation rate remains

uncovered, and novel methods approaching a temporal resolu-

tion in the ms or s regime are still to be developed in order to

shed light on this aspect.

4.2. Volume Change of the Crystals Associated with Phase

Transition

Although network ﬂexibility gives rise to signiﬁcant advantages

for gas storage and separation applications, it also poses many

challenges. The structural phase transitions are typically associ-

ated with macroscopic changes in crystal size.[116]This aspect has

been beneﬁcially used for actuators and sensors.[117]

ELM−11 shows two steps upon the CO2adsorption, with a

28% expansion in the interlayer distance in the ﬁrst transition

and a 56% expanded layer structure compared to the initial struc-

ture in the second one.[118]The macroscopic changes in the pow-

der bed upon CO2adsorption have been reported already in

2006.[119]

Kaskel and co-workers studied volume expansion and force ex-

erted by ﬂexible MOFs through expansion for MIL-53(Al).[117b]

The eﬀect of the packing density on the mechanical response was

www.advancedsciencenews.com www.advmat.de

Figure 36. In situ time-resolved PXRD measured during the adsorption of n-butane at 298 K on MIL-53(Al): a) reopening of the crystals with average

crystal size of 28 ±18 μm, and b) breathing transition in the crystals with average crystal size of 1.2 ±0.5 μm. Reproduced with permission.[114]Copyright

2022, Royal Society of Chemistry.

also evaluated. Three diﬀerent regimes were identiﬁed according

to the packing density. The pressure gained from the opening

step was found to be higher than that needed to compress the

empty framework and is speciﬁc for the stimulating ﬂuid.

The macroscopic expansion, however, creates challenges in ad-

sorptive storage and separation applications, especially in the ad-

sorption bed. In most applications, the adsorption column is sev-

eral meters high, and thus, the adsorbent at the bottom of the bed

experiences signiﬁcant gravitational force due to the high weight

of the packed bed above. This lower part of the bed has to exert a

signiﬁcant force against the upper part to open its pores and ac-

complish the volume change. In extreme cases, pore expansion

may be completely suppressed in such a real macroscopic appli-

cation. Another aspect is that the adsorber column has to pro-

vide enough dead volume for expansion, which may limit volu-

metric eﬃciency, which is often overlooked in academic research.

Moreover, the expanding adsorbent will most likely induce a pres-

sure drop in the column.[46a]Hence, MOFs showing minimal vol-

ume changes or ﬂexibility based on the internal movement of the

building blocks are more favorable.

4.3. Crystal Damage Associated with Phase Transition

The huge volume change associated with the mechanical

and thermal stress often leads to the damage of the crystals

(adsorption-induced milling).[120]The magnitude of the eﬀect de-

pends on the structure and chemical composition of the MOF

itself, as well as on the particle size, adsorptive, and adsorption

temperature.

MIL-53(Al) and ELM-11 were proven to be stable in

cyclic experiments with butane at 298 K, withstanding 100

adsorption/desorption cycles without signiﬁcant changes in

performance.[120]Also repeated pressure swings between 0.5 and

www.advancedsciencenews.com www.advmat.de

Figure 37. a) XRD pattern at 42 kPa CO2and b) crystal structure of ELM-11 in the open pore phase. c) Colormap of time resolved XRD patterns of

ELM-11 at 248 K. CO2was introduced 5 s after the start of the measurement with the constant ﬂow rate of 0.8 kPa s−1, until 42 kPa pressure was

reached. e) Time evolution of fraction transformed, calculated from the peak intensity ratio from the XRD patterns in (a) and (f) at regular intervals of

under CC BY 4.0.

2.0 MPa at 298 and 323 K upon six CO2adsorption/desorption

cycles for MIL-53(Al) did not cause the deterioration in the ad-

sorption capacity.[121]For ELM-11, the reproducibility of the gat-

ing was demonstrated in 50 cycles of CH4physisorption at

303 K.[34]

In contrast, for DUT-8(Ni) and SNU-9, the multiple adsorp-

tion/desorption stress upon butane adsorption at room tempera-

ture leads to the reduction of crystallite size, causing changes in

the switching behavior in the initial 10 physisorption runs, and a

characteristic shift of the “gate-opening” pressure to higher val-

uesisobserved.

[120]The reason for such behavior is particle size-

dependence t ﬂexibility.[48]

The comprehensive investigations show the existence of three

main particle size regimes in the case of DUT-8(Ni): Several mi-

crometer large particles show pronounced gating behavior, while

≈500 nm–1 μm sized particles reveal a suppressed opening (sta-

bilization of the metastable cp phase). Particles with sizes below

500 nm have suppressed closing ability (stabilization of the op

phase).[48]

Abylgazina et al.[48b]demonstrated that the changes in the gate

opening pressure with particle size are not linear and follow a log-

arithmic dependence on the facet extension. Moreover, the steep-

ness of the adsorption branch in the gating region decreases with

decreasing crystal size, pointing to the signiﬁcantly broader acti-

vation energy distribution in the small grains compared to the

large crystallites. The threshold of the particle size, at which the

compound experiences rigidiﬁcation and loss of crystallinity, de-

pends in the DUT-8(M) (M =metal) system, on the metal type in

the paddle wheel.[48a,122]The eﬀects of crystal size and morphol-

ogy are critically discussed in a recent review.[48]

4.4. Change in the Flexibility Characteristics due to the Additives

and Shaping

When building an adsorber for gas separations, the adsorbing

material is usually pelletized to improve the gas ﬂow, reduce pres-

sure drop, and improve overall performance. For MOF materials,

this poses a signiﬁcant challenge since many of the well-known

frameworks cannot withstand the stress of shaping into pellets

and, as a consequence, important properties, like sorption ca-

pacity and selectivity, are reduced.[82,123]In ﬂexible MOFs, loss or

massive impact on the ﬂexibility characteristics can be expected

in addition to mechanical deterioration.[124]

www.advancedsciencenews.com www.advmat.de

Figure 38. Atomic structure, a schematic of the structural transition, and measurement and calculation results for a–c) the gate opening of ELM-11 at

248 K, d–f ) op →np transition of MIL-53(Al) at 223 K, and g–i) np→lp transition of MIL-53(Al) (at 195 K). Reproduced with permission.[115 ]Copyright

2023, The Authors, under CC BY 4.0.

MIL-53(Al), however, shows stable performance after shap-

ing while maintaining the textural properties and characteristic

framework dynamics.[125]It was demonstrated that pellets with

5 or 10 wt% of binder reached mechanical resistances compara-

ble with some of the carbon molecular sieves pellets. All exper-

imental insights revealed that the reversible breathing eﬀect—

characteristic for MIL-53—was preserved even though the sam-

ples were shaped into cylindrical extrudates. In situ X-ray diﬀrac-

tion studies under humid conditions conﬁrmed the similarity

in phase transition kinetics for extrudates and powders. More-

over, the phase transition of the MIL-53 extrudates upon ad-

sorption of carbon dioxide was conﬁrmed by high-pressure ad-

sorption isotherms of CH4and CO2on MIL-53 powders and

extrudates, which exhibited similar capacities for both gases

(Figure 39).

Mechanical stress can potentially induce phase transition in

ﬂexible MOFs.[126]In a recent study the group of Song shows

the inﬂuence of mechanical pressures on MIL-53(Al). High me-

chanical pressures and temperatures support the interactions be-

tween CO2, and the hydroxyl groups and the Al centers of the net-

work and increase the sorption capacity because of chemisorptive

behavior.[127]

Kundu et al. tested the pelletization of MIL-53(Al)-OH and

MIL-53(Al).[46a]Although no decrease in the uptake or changes

in the isotherms’ shape were observed, both samples could not

withstand the stress of deformation, resulting in the breaking of

www.advancedsciencenews.com www.advmat.de

Figure 39. High-pressure methane (red) and carbon dioxide (black) ph-

ysisorption isotherms at 303 K on MIL-53 with 3 wt% of methyl cellulose as

binder (dashed lines) in comparison to the adsorption isotherms on MIL-

Wiley-VCH GmbH & Co. KGaA.

the pellets during CH4adsorption. Similar results were shown by

the group of Denayer,[82]who also investigated the pelletization

of MIL-53(Al).

The application of binders to create ﬂexible MOFs containing

composites can also potentially lead to changes in the ﬂexible

behavior.

Along these lines, the Watanabe group was able to show the

inﬂuence of polymer matrix (polyvinylpyrrolidone, PVP) on the

switching behavior of ELM-11.[124]Later on, similar experiments

were also conducted with an interpenetrated [Cu2(bdc)2(bpy)]n

framework (referred to as JG-MOF).[128]Interestingly, the ma-

terial shows a very distinct change in the gate-opening behav-

ior, depending not only on the amount of PVP added but also

on the sample preparation approach. Increasing binder amount

causes a “slacking” of the gate opening. The term describes an

external force-induced shift from a stepwise to a sequential tran-

sition, leading to a smearing of the originally steep adsorption

branch originating from the phase transition.[124]This eﬀect was

explained through free energy analysis, showing a nonsynchro-

nized transformation within a single MOF particle. Furthermore,

the slacking could be modulated by the composite preparation

method as it impacts the force inﬂicted on the material. For

higher external forces, more pronounced slacking was observed.

In line with theoretical calculations, similar experiments on the

less-expanding interpenetrated JG-MOF conﬁrmed this theory.

Due to the reduced expansion of the networks, less force was

exerted in the pellet, thus minimizing slacking.[128]Therefore,

developing methods to reduce the stress inﬂicted by binders is

crucial.[124]

4.5. Slipping Oﬀ Eﬀect

To use ﬂexible MOFs for separation, the set partial pressure

of the stimulus must exceed gate opening pressure to induce

the phase transition and enable the separation. Thus, in break-

through applications, no pure gases can be obtained because the

Figure 40. Comparison of typical breakthrough curves (blue) and

isotherms (purple) for rigid microporous adsorbents (top) and “gating”-

type ﬂexible MOFs (bottom). The gate opening and gate closing pressures

are marked by A and B, respectively.The breakthrough curves for gating ad-

sorbents have a “stepped” feature not seen in rigid adsorbents, deﬁned by

a plateau in the eﬄuent fraction (represented by a “step height”) before the

American Chemical Society.

target gas exhausts until the gate opening pressure is reached.

This phenomenon was named the “slipping-oﬀ” eﬀect.[129]A

comparison of a typical breakthrough curve proﬁle of a rigid

adsorbent and a representative “stepped” breakthrough curve

of a “gating”-type ﬂexible MOF with “slipping-oﬀ” is shown in

Figure 40.

Such an eﬀect was reported several times for the breakthrough

curves of ﬂexible materials. For example, the breakthrough

curves collected for CH4/CO2on MIL-53(Cr) exhibit distinct

changes in the slope of eﬄuent CO2concentration with time.[81]

An even more pronounced slipping-oﬀ breakthrough curve was

observed for the separation of xylenes on MIL-53(Al)[33a]and sep-

aration of CH4and CO2on CID-5, where the partial pressure of

CO2dropped below the gate opening pressure of the adsorbents,

while it was passing the column.[130]

Breakthrough measurements performed on ELM-11 also un-

covered such a breakthrough step, which was discussed by So-

www.advancedsciencenews.com www.advmat.de

Figure 41. a) Pictures of the column system used by Tanaka and co-workers in order to show the slipping-oﬀ eﬀect and its solution b) by combining

a column of ELM-11 and HKUST-1. c,d) Breakthrough curves of CO2and CH4for the system depicted in (a) and (b), respectively. Reproduced with

tomayor and Lastoskie in 2017 and has also been later on found

in [Co(bdp)]n.[69,77a]

Since this eﬀect is very relevant for practical applications, more

than one solution has been proposed.

i) One possible solution is to overpressure the feed gas so that

the partial pressure of the opening gas remains above the

gate pressure along the entire length of the column. This ap-

proach will, however, incur an energy penalty in the work

input required to pressurize the feed gas.[77a]

ii) Another way is to directly inﬂuence the adsorption condi-

tions by increasing or decreasing the temperature to shift

the gate opening pressure. Further functionalization of the

linker shows an impact if, afterward, one of the phases will

be further stabilized with this functionality.[69]

iii) An advanced approach was demonstrated by Kitagawa and

co-workers,[130]who used a solid solution of two MOFs, CID-

5 and CID-6, to cleanly separate methane from carbon diox-

ide and ethane. CID-5 has more structural ﬂexibility than

CID-6 and exhibits selective gated adsorption for CO2and

C2H6over CH4, whereas CID-6 has permanent microp-

orosity and nonselectively adsorbs all three gases. In break-

through experiments with a 60:40 vol% CH4/CO2mixture

and a 90:10 vol% CH4/C2H6mixture on CID-5, the CH4

fraction in the eﬄuent was with ≈90% well below the target

due to the “slipping-oﬀ”. Pure methane could also not be ob-

tained for the same gas mixtures and CID-6, because of coad-

sorption. However, a solid solution of CID-5 and CID-6 with

gated adsorption characteristics for both CO2and C2H6,pre-

pared by substituting 10% of the nitroisophthalate ligands

www.advancedsciencenews.com www.advmat.de

of CID-5 with the methoxyisophthalate ligands of CID-6,

cleanly separated methane from both carbon dioxide and

ethane for retention times of 8 and 25 min, respectively, in

breakthrough column experiments. Hence, by tuning the

gate opening pressures via framework composition, it was

possible to optimize the gas separation performance under

dynamic conditions.[77a,130]

iv) In 2020, a novel approach was demonstrated by Tanaka and

co-workers by successively operating columns containing

ﬂexible and rigid MOF as a guard bed (Figure 41).[129]Signif-

icant improvement of selectivity for CO2was achieved when

connecting a column containing ﬂexible ELM-11 to a column

loaded with rigid microporous HKUST-1. In this setup, al-

most the entire CO2could be adsorbed. The slipping-oﬀ CO2

was collected by the HKUST-1 column and thereby allowed

for very high purity of the exhaust gas while maintaining a

high capacity for CO2.

v) The MOFs showing smooth, quasi-second-order phase tran-

sitions can also be beneﬁcial in overcoming the slipping-oﬀ

problem.[18b]

5. Beyond Improving Working Capacity and

Selectivity: Intelligent Materials and Use of

Switchability

5.1. Thermal Management

A very important problem is the change in the temperature of

the adsorption bed, leading to a reduction of adsorption capac-

ity upon adsorption and an increase of adsorption capacity upon

desorption or even preventing the complete removal of guest

molecules.[131]

Engineering solutions to this problem can be found in adsor-

ber design, which allows for better heat transfer. The trade-oﬀ,

however, remains between achieving high selectivity and capac-

ity while managing the inherent challenges associated with these

properties.[129]Developing novel materials capable of satisfying

both requirements is a key focus of ongoing research.

One potential solution was proposed to use phase change ma-

terials (PCMs) as latent heat storing additives. PCMs are able

to absorb and store some of the heat released during adsorp-

tion through a phase transition.[131]The ﬂexible MOF itself is,

in a certain sense, a PCM (Figure 42).[132]This leads to an intrin-

sic thermal management upon physisorption that can be mon-

itored in calorimetric measurements.[26]The intrinsic thermal

management capability of ﬂexible MOFs should be useful for

storage applications as well as for developing highly eﬃcient PSA

systems.[118]Typical gas physisorption enthalpies are between −5

and −60 kJ per mol of gas,[61]depending on adsorptive and sur-

face chemistry, as well as the pore sizes of the adsorbent. The

structural transition enthalpies in MOFs are connected with the

thermal eﬀects ranging between 6 and 31 kJ per mol of MOF.[37]

In [Co(bdp)]n,diﬀerential enthalpies of CH4and CO2adsorp-

tion reveals signiﬁcant reductions in heat released upon adsorp-

tion during the discrete, endothermic structural phase changes

relative to the regions between these phase changes (Figure 43).

For comparison, the isostructural rigid MOFs [Ni(bdp)]nand

[Zn(bdp)]n, display diﬀerential enthalpies of CO2adsorption of

Figure 42. When adiabatic gas adsorption is considered, the temperature

rise of the system for the ﬂexible MOF, ΔTX, is smaller than that for the

conventional adsorbent, ΔTY, and the resulting decrease in the adsorption

amount can be suppressed because of the smaller net heat of adsorption

of the ﬂexible MOF owing to its intrinsic thermal management capability.

BY 4.0.

−20 kJ mol−1at zero coverage.[133]Notably, these values are very

close to those observed for [Co(bdp)]nat the pressures between

the CO2-induced phase transition. During the ﬁrst phase transi-

tion, [Co(bdp)]nshows values in the range of −24 to −26 kJ mol−1,

which are signiﬁcantly lower in magnitude than what might be

expected. During the second and third CO2-induced phase tran-

sitions, the magnitude of adsorption enthalpy plummets dra-

matically, reaching values as small as −11 and −16 kJ mol−1,

respectively (Figure 43). Thus, the structural phase changes of

[Co(bdp)]n, because of their endothermic nature, can substan-

tially mitigate the amount of heat that must be dissipated during

adsorption.

For ELM-11, the molar integral heats of adsorption in the

phases with increasing porosity are estimated by GCMC simu-

lations to be 40.8, 38.3, and 33.6 kJ per mol CO2, respectively.

The transition enthalpies were determined to be 21.9 kJ per mol

www.advancedsciencenews.com www.advmat.de

Figure 43. Diﬀerential enthalpies of CO2adsorption in [Co(bdp)]nare

shown in purple (standard errors are shown as black bars) as a function

of CO2loading. Local minima in adsorption enthalpy correspond to re-

gions in which [Co(bdp)]nundergoes an endothermic structural expan-

sion, which oﬀsets some of the heat released upon CO2adsorption and

provides intrinsic thermal management. The single-component CO2ad-

sorption isotherm (red circles) is provided for comparison. Reproduced

CO2for the gate closing at higher pressures and 25.5 kJ per mol

CO2for the gate closing at lower pressure, which suggests that

both closings are endothermic processes and that more heat per

mole of CO2is removed from the system for the gate closing at

lower pressure than for that at the higher pressures.

Hiraide et al.,[129]in their study on ELM-11, demonstrated that

during the adsorption of an equimolar CO2/CH4gas mixture at

500 kPa and 298 K, the host material undergoes an endothermic

expansion with an associated enthalpy change of 55.7 J g−1.Con-

currently, the exothermic enthalpy change due to the adsorption

of the gas mixture is 135.3 J g−1. The net heat eﬀect is decreased

to 79.6 J g−1, indicating that 41% of the exothermic heat is oﬀset

by the endothermic process.

Thus, the structural transitions in ﬂexible adsorbents oﬀer ob-

vious advantages for thermal management compared to rigid

adsorbents.

6. Conclusion

Flexible MOFs are emerging materials for advanced applications

in energy storage and gas separation thanks to their unique ad-

sorption properties and responsive behavior. The ﬂexibility of

the frameworks essentially contributes to achieving exceptionally

high selectivity in separation processes and improves deliverable

storage capacity. These materials oﬀer great potential for tackling

challenging (isotope) separations that would be diﬃcult or even

impossible utilizing rigid structures. The responsivity of MOFs

can be ﬁne-tuned by adjusting the metal type, linker substituents,

or by employing mixed metal or mixed linker strategies.

When designing separation processes, it is crucial to consider

not only the single gas adsorption isotherms of both components

involved but also the desorption branches (as discussed in Sec-

tion 4), as this is essential for accurately predicting or leveraging

the enhanced separation capabilities of ﬂexible MOFs. Addition-

ally, the temperature-dependent behavior of these ﬂexible frame-

works must be carefully considered, as it plays a key role in their

performance under diﬀerent conditions.

For column-based applications, MOFs with minimal volume

changes are generally preferable to ensure stable operation, but

it has been shown, that also ﬂexible MOFs can be shaped while

maintaining their storage capacity and ﬂexibility, allowing for

greater adaptability in industrial applications. Additionally, their

intrinsic thermal management capabilities further enhance their

suitability for demanding energy storage and gas separation pro-

cesses. With their versatile and tunable properties, ﬂexible MOFs

are poised to revolutionize these two ﬁelds in the future.

Acknowledgements

The DGF is acknowledged for ﬁnancial support in the frame of the research

unit (FOR 2433). Y.W. gratefully acknowledges funding from the Alexan-

der von Humboldt Foundation. V.B. acknowledges BMBF for funding

within project Nos. 05K22OD1 (“TIMESWITCH”) and 05K22OD2 (“TO-

MOPORE”).

Open access funding enabled and organized by Projekt DEAL.

Conﬂict of Interest

The authors declare no conﬂict of interest.

Keywords

adsorption selectivity, breathing, ﬂexible metal–organic frameworks, gate

opening, gas separation, gas storage, soft porous crystals

Received: September 28, 2024

Revised: November 28, 2024

Published online:

[1] S. R. Batten, N. R. Champness, X.-M. Chen, J. Garcia-Martinez, S.

Kitagawa, L. Öhrström, M. O’Keeﬀe, M. P. Suh, J. Reedijk, Pure Appl.

Chem. 2013,85, 1715.

[2] X. Zhang, Z. Chen, X. Liu, S. L. Hanna, X. Wang, R. Taheri-Ledari, A.

Maleki, P. Li, O. K. Farha, Chem.Soc.Rev.2020,49, 7406.

[3] a) K. R. Wright, K. Nath, A. J. Matzger, Angew. Chem., Int. Ed. 2022,

61, 202213190; b) J. Ma, A. P. Kalenak, A. G. Wong-Foy,A. J. Matzger,

Angew. Chem., Int. Ed. 2017,56, 14618; c) A. J. Howarth, A. W. Peters,

N. A. Vermeulen, T. C. Wang, J. T. Hupp, O. K. Farha, Chem. Mater.

2017,29, 26.

[4] a)K.O.Kirlikovali,S.L.Hanna,F.A.Son,O.K.Farha,ACS Nanosci.

Au 2023,3, 37; b) B. Pramanik, R. Sahoo, M. C. Das, Coord. Chem.

Rev. 2023,493, 215301.

[5] a) O. M. Yaghi, J. Am. Chem. Soc. 2016,138, 15507; b) H. Jiang, D.

Alezi, M. Eddaoudi, Nat. Rev. Mater. 2021,6, 466; c) O. M. Yaghi,

M. J. Kalmutzki, C. S. Diercks, Introduction to Reticular Chemistry.

Metal–Organic Frameworks and Covalent Organic Frameworks,Wiley-

VCH, Weinheim, Germany 2019.d)R.Freund,S.Canossa,S.M.

Cohen, W. Yan, H. Deng, V. Guillerm, M. Eddaoudi, D. G. Madden,

D. Fairen-Jimenez, H. Lyu, L. K. Macreadie, Z. Ji, Y. Zhang, B. Wang,

F. Haase, C. Wöll, O. Zaremba, J. Andreo, S. Wuttke, C. S. Diercks,

Angew. Chem., Int. Ed. 2021,60, 23946.

[6] a)M.R.Ryder,B.Civalleri,J.C.Tan,Phys. Chem. Chem. Phys. 2016,

13, 9079. b) E. Mahmoud, Crystal s 2020,10,700; c) W. Fan, X. Zhang,

www.advancedsciencenews.com www.advmat.de

Z. Kang, X. Liu, D. Sun, Coord. Chem. Rev. 2021,443, 213968; d) F.-F.

Lu, X.-W. Gu, E. Wu, B. Li, G. Qian, J. Mater.Chem. A 2023,11, 1246;

e) S. Krause, J. D. Evans, V. Bon, I. Senkovska, S. Ehrling, U. Stoeck,

P. G. Yot, P. Iacomi, P. Llewellyn, G. Maurin, F.-X. Coudert, S. Kaskel,

J. Phys. Chem. C 2018,122, 19171.

[7] I. M. Hönicke, I. Senkovska, V. Bon, I. A. Baburin, N. Bönisch, S.

Raschke, J. D. Evans, S. Kaskel, Angew. Chem., Int. Ed. 2018,57,

13780.

[8] a) W. Xu, O. M. Yaghi, ACS Cent. Sci. 2020,6, 1348; b) W. Song, Z.

Zheng, A. H. Alawadhi, O. M. Yaghi, Nat. Water 2023,1, 626; c) C.

Charalambous, E. Moubarak, J. Schilling, E. Sanchez Fernandez, J.-Y.

Wang, L. Herraiz, F. McIlwaine, S. B. Peh, M. Garvin, K. M. Jablonka,

S. M. Moosavi, J. Van Herck, A. Y. Ozturk, A. Pourghaderi, A.-Y. Song,

G. Mouchaham, C. Serre, J. A. Reimer, A. Bardow, B. Smit, S. Garcia,

Nature 2024,632, 89; d) J.-B. Lin, T. T. T. Nguyen, R. Vaidhyanathan,

J. Burner, J. M. Taylor, H. Durekova, F.Akhtar, R. K. Mah, O. Ghaﬀari-

Nik, S. Marx, N. Fylstra, S. S. Iremonger, K. W. Dawson, P. Sarkar, P.

Hovington, A. Rajendran, T. K. Woo, G. K. H. Shimizu, Science 2021,

374, 1464.

[9] a) S. Horike, S. Shimomura, S. Kitagawa, Nat. Chem. 2009,1, 695;

b) A. Schneemann, V. Bon, I. Schwedler, I. Senkovska, S. Kaskel, R.

A. Fischer, Chem. Soc. Rev. 2014,43, 6062; c) S. Kitagawa, Flexible

Metal–Organic Frameworks: Structural Design, Synthesis and Proper-

ties, Royal Society of Chemistry, Cambridge, UK 2024.

[10] a) S. Krause, N. Hosono, S. Kitagawa, Angew. Chem., Int. Ed. 2020,

59, 15325; b) J. H. Lee, S. Jeoung, Y. G. Chung, H. R. Moon, Coord.

Chem. Rev. 2019,389, 161; c) S. Seth, S. Jhulki, Mater. Horiz. 2021,8,

700; d) I. Senkovska, V. Bon, L. Abylgazina, M. Mendt, J. Berger, G.

Kieslich, P. Petkov, J. Luiz Fiorio, J. O. Joswig, T. Heine, L. Schaper, C.

Bachetzky, R. Schmid, R. A. Fischer, A. Pöppl, E. Brunner, S. Kaskel,

Angew. Chem., Int. Ed. 2023,62, 202218076.

[11] a) X. Liu, Z. Wang, Y. Zhang, N. Yang, B. Gui, J. Sun, C. Wang, J. Am.

Chem. Soc. 2024,146, 11411; b) A. Mauri, R. Vismara, M. Moroni,

E. Roldán-Molina, J. A. R. Navarro, S. Galli, Small Sci. 2024,4,

2400277.

[12] a) J. Li, B. Chen, Chem. Sci. 2024,15, 9874; b) Y. He, S. Xiang, B.

Chen, J. Am. Chem. Soc. 2011,133, 14570; c) C. Jiang, J.-X. Wang,

D. Liu, E. Wu, X.-W. Gu, X. Zhang, B. Li, B. Chen, G. Qian, Angew.

Chem., Int. Ed. 2024,63, 202404734.

[13] Z. Liu, L. Zhang, D. Sun, Chem. Commun. 2020,56, 9416.

[14] J. D. Evans, V. Bon, I. Senkovska, H.-C. Lee, S. Kaskel, Nat. Commun.

2020,11, 2690.

[15] a) S. Watanabe, H. Sugiyama, H. Adachi, H. Tanaka, M. T. Miyahara,

J. Chem. Phys. 2009,130; b) H. Sugiyama, S. Watanabe, H. Tanaka,

M. T. Miyahara, Langmuir 2012,28, 5093.

[16] H. Tanaka, M. T. Miyahara, Curr. Opin. Chem. Eng. 2019,24, 19.

[17] R. Goeminne, S. Krause, S. Kaskel, T. Verstraelen, J. D. Evans, J. Am.

Chem. Soc. 2021,143, 4143.

[18] a) A. Schneemann, Y. Takahashi, R. Rudolf, S.-I. Noro, R. A. Fischer,

J. Mater. Chem. A 2016,4, 12963; b) C. Wang, X.-W. Zhang, X.-X.

Chen, W.-X. Zhang, J.-P. Zhang, J. Am. Chem. Soc. 2024,146, 13886.

[19] D.-D. Zhou, J.-P. Zhang, Acc. Chem. Res. 2022,55, 2966.

[20] D. P. Broom, Adsorption 2024,30, 1565.

[21] a) I. Senkovska, S. Kaskel, Microporous Mesoporous Mater. 2008,112,

108; b) U. Stoeck, S. Krause, V. Bon, I. Senkovska, S. Kaskel, Chem.

Commun. 2012,48, 10841; c) Y. He, W. Zhou, G. Qian, B. Chen,

Chem.Soc.Rev.2014,43, 5657; d) C.-C. Liang, Z.-L. Shi, C.-T. He, J.

Tan, H.-D. Zhou, H.-L. Zhou, Y. Lee, Y.-B. Zhang, J. Am. Chem. Soc.

2017,139, 13300; e) C. M. Simon, J. Kim, D. A. Gomez-Gualdron,

J. S. Camp, Y. G. Chung, R. L. Martin, R. Mercado, M. W. Deem,

D. Gunter, M. Haranczyk, D. S. Sholl, R. Q. Snurr, B. Smit, Energy

Environ. Sci. 2015,8, 1190.

[22] S. Hiraide, Y. Sakanaka, H. Kajiro, S. Kawaguchi, M. T. Miyahara, H.

Tana k a , Nat. Commun. 2020,11, 3867.

[23] a) H. Demir, M. Mobedi, S. Ülkü, Int. Commun. Heat Mass Transfer.

2009,36, 372; b) J. Wieme, S. Vandenbrande, A. Lamaire, V. Kapil, L.

Vanduyfhuys, V. Van Speybroeck, ACS Appl. Mater. Interfaces 2019,

11, 38697.

[24] K. J. Chang, O. Talu, Appl. Therm. Eng. 1996,16, 359.

[25] a) M. M. Mohseni, M. Jouyandeh, S. M. Sajadi, A. Hejna, S.

Habibzadeh, A. Mohaddespour, N. Rabiee, H. Daneshgar, O.

Akhavan, M. Asadnia, M. Rabiee, S. Ramakrishna, R. Luque, M. R.

Saeb, Chem. Eng. J. 2022,449, 137700; b) H. Babaei, K. R. Meihaus,

J. R. Long, Chem. Mater. 2023,35, 6220; c) C. Yuan, S. S. Sørensen,

T. Du, Y. Song, M. M. Smedskjaer, Int. J. Heat Mass Transfer 2024,

233, 126015.

[26] J. A. Mason, J. Oktawiec, M. K. Taylor, M. R. Hudson, J. Rodriguez,

J. E. Bachman, M. I. Gonzalez, A. Cervellino, A. Guagliardi, C. M.

Brown, P. L. Llewellyn, N. Masciocchi, J. R. Long, Nature 2015,527,

357.

[27] X. Zhang, X. Wang, F. Gao, Y. Chen, H. Liu, P. Zhou, Z. Kang, Y.Wang,

W. Fa n , Mater. Adv. 2024,5, 3135.

[28] a) MOVE (Mobility Optimized Vehicle Energy) program by

ARPA-E: https://arpa-e.energy.gov/sites/default/ﬁles/documents/

ﬁles/MOVE_ProgramOverview.pdf (accessed: September 2024). b)

ISO 15403-1:2016 standard: https://www.iso.org/obp/ui/#iso:std:

iso:15403:-1:ed- 1:v1 (accessed: September 2024). c) M. Beckner, A.

Dailly, Appl. Energy 2015,149, 69.

[29] A. Granja-DelRío, I. Cabria, Int. J. Hydrogen Energy 2024,54, 665.

[30] a) Y. Peng, V. Krungleviciute, I. Eryazici, J. T. Hupp, O. K. Farha,

T. Yildirim, J. Am. Chem. Soc. 2013,135, 11887; b) J. Jiang, H.

Furukawa, Y.-B. Zhang, O. M. Yaghi, J. Am. Chem. Soc. 2016,138,

10244; c) M. Zhang, W. Zhou, T. Pham, K. A. Forrest, W. Liu, Y. He,

H.Wu,T.Yildirim,B.Chen,B.Space,Y.Pan,M.J.Zaworotko,J.Bai,

Angew. Chem., Int. Ed. 2017,56, 11426; d) A. Dailly, M. Beckner,

in Nanoporous Materials for Gas Storage, Green Energy and Tech-

nology (Eds: K. Kaneko, F. Rodríguez-Reinoso), Springer, Singapore

2019.

[31] D. A. Gómez-Gualdrón, C. E. Wilmer, O. K. Farha, J. T. Hupp, R. Q.

Snurr, J. Phys. Chem. C 2014,118, 6941.

[32] K. A. Forrest, G. Verma, Y. Ye, J. Ren, S. Ma, T. Pham, B. Space, Chem.

Phys. Rev. 2022,3, 021308.

[33] a) R. Kitaura, K. Seki, G. Akiyama, S. Kitagawa, Angew. Chem., Int.

Ed. 2003,42, 428; b) K. Seki, Phys.Chem.Chem.Phys.2002,4,

1968.

[34] 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, J. Colloid Interface Sci. 2009,334,1.

[35] S. Krause, V. Bon, I. Senkovska, U. Stoeck, D. Wallacher, D. M.

Többens, S. Zander, R. S. Pillai, G. Maurin, F.-X. Coudert, S. Kaskel,

Nature 2016,532, 348.

[36] S. Krause, J. D. Evans, V. Bon, I. Senkovska, P. Iacomi, F. Kolbe, S.

Ehrling, E. Troschke, J. Getzschmann, D. M. Többens, A. Franz, D.

Wallacher, P. G. Yot, G. Maurin, E. Brunner, P. L. Llewellyn, F.-X.

Coudert, S. Kaskel, Nat. Commun. 2019,10, 3632.

[37] a) S. Krause, J. D. Evans, V. Bon, I. Senkovska, F.-X. Coudert, D. M.

Többens, D. Wallacher, N. Grimm, S. Kaskel, Faraday Discuss. 2021,

225, 168; b) S. Rahman, A. Arami-Niya, X. Yang, G. Xiao, G. Li, E. F.

May, Commun. Chem. 2020,3, 186.

[38] M. K. Taylor, T. Runˇ

cevski, J. Oktawiec, M. I. Gonzalez, R. L.

Siegelman, J. A. Mason, J. Ye, C. M. Brown, J. R. Long, J. Am. Chem.

Soc. 2016,138, 15019.

[39] Q. Y. Yang, P. Lama, S. Sen, M. Lusi, K. J. Chen, W. Y. Gao, M.

Shivanna, T. Pham, N. Hosono, S. Kusaka, J. J. Perry, S. Ma, B.

Space, L. J. Barbour, S. Kitagawa, M. J. Zaworotko, Angew. Chem.,

Int. Ed. 2018,57, 5684.

[40] S.-M. Wang, M. Shivanna, P. Lama, Q.-Y. Yang, L. J. Barbour, M. J.

Zaworotko, ChemSusChem 2023,16, 202300069.

www.advancedsciencenews.com www.advmat.de

[41] S. Ehrling, M. Mendt, I. Senkovska, J. D. Evans, V. Bon, P. Petkov, C.

Ehrling, F. Walenszus, A. Pöppl, S. Kaskel, Chem. Mater. 2020,32,

5670.

[42] H. J. Park, M. P. Suh, Chem. Commun. 2010,46, 610.

[43] M. Sin, N. Kavoosi, M. Rauche, J. Pallmann, S. Paasch, I. Senkovska,

S. Kaskel, E. Brunner, Langmuir 2019,35, 3162.

[44] V. Bon, I. Senkovska, S. Kaskel, in Nanoporous Materials for Gas Stor-

age, Green Energy and Technology (Eds: K. Kaneko, F. Rodríguez-

Reinoso), Springer, Singapore 2019.

[45] V. Bon, I. Senkovska, D. Wallacher, D. M. Többens, I. Zizak, R.

Feyerherm, U. Mueller, S. Kaskel, Inorg. Chem. 2014,53, 1513.

[46] a) T. Kundu, B. B. Shah, L. Bolinois, D. Zhao, Chem. Mater. 2019,

31, 2842; b) J. Yang, X. Yan, T. Xue, Y. Liu, RSC Adv. 2016,6, 55266;

c) S. Biswas, T. Ahnfeldt, N. Stock, Inorg. Chem. 2011,50, 9518.

[47] a) M. Bonneau, C. Lavenn, J.-J. Zheng, A. Legrand, T. Ogawa, K.

Sugimoto, F.-X. Coudert, R. Reau, S. Sakaki, K.-i. Otake, S. Kitagawa,

Nat. Chem. 2022,14, 816; b) L. Bolinois, T. Kundu, X. Wang, Y. Wang,

Z.Hu,K.Koh,D.Zhao,Chem. Commun. 2017,53, 8118.

[48] a) L. Abylgazina, I. Senkovska, S. Ehrling, V. Bon, P. St Petkov,

J. D. Evans, S. Krylova, A. Krylov, S. Kaskel, CrystEngComm 2021,

23, 538; b) L. Abylgazina, I. Senkovska, R. Engemann, S. Ehrling,

T. E. Gorelik, N. Kavoosi, U. Kaiser, S. Kaskel, Front. Chem. 2021,

9, 674566; c) S. Ehrling, H. Miura, I. Senkovska, S. Kaskel, Trends

Chem. 2021,3, 291.

[49] H. Reinsch, M. A. van der Veen, B. Gil, B. Marszalek, T. Verbiest, D.

de Vos, N. Stock, Chem. Mater. 2013,25, 17.

[50] H.-P. Li, J.-W. Wang, Z. Dou, L.-Z. Wu, Y. Wang, Y. Liang, Q.-G. Zhai,

Chem. Eng. J. 2024,492.

[51] a) U.S. Department of Energy’s (DOE) technical targets for

onboard hydrogen storage systems in light-duty vehicles:

https://www.energy.gov/eere/fuelcells/doe-technical-targets-

onboard-hydrogen-storage-light-duty-vehicles (accessed: Septem-

ber 2024). b) M. D. Allendorf, Z. Hulvey, T. Gennett, A. Ahmed,

T. Autrey, J. Camp, E. Seon Cho, H. Furukawa, M. Haranczyk, M.

Head-Gordon, S. Jeong, A. Karkamkar, D.-J. Liu, J. R. Long, K. R.

Meihaus, I. H. Nayyar, R. Nazarov, D. J. Siegel, V. Stavila, J. J. Urban,

S.P.Veccham,B.C.Wood,Energy Environ. Sci. 2018,11, 2784.

[52] X. Zhang, R.-B. Lin, J. Wang, B. Wang, B. Liang, T. Yildirim, J. Zhang,

W. Zhou, B. Chen, Adv. Mater. 2020,32, 1907995.

[53] H. J. Choi, M. Dinc˘

a, J. R. Long, J. Am. Chem. Soc. 2008,130, 7848.

[54] A. Halder, R. A. Klein, S. Shulda, G. A. McCarver, P. A. Parilla, H.

Furukawa, C. M. Brown, C. M. McGuirk, J. Am. Chem. Soc. 2023,

145, 8033.

[55] a) J. Y. Kim, L. Zhang, R. Balderas-Xicohtencatl, J. Park, M. Hirscher,

H. R. Moon, H. Oh, J. Am. Chem. Soc. 2017,139, 17743; b) R. A.

Pollock, J.-H. Her, C. M. Brown, Y. Liu, A. Dailly, J. Phys. Chem. C

2014,118, 18197.

[56] H. Li, L. Li, R.-B. Lin, W. Zhou, Z. Zhang, S. Xiang, B. Chen, Energy-

Chem 2019,1, 100006.

[57] J.-P. Zhang, X.-M. Chen, J. Am. Chem. Soc. 2009,131, 5516.

[58] R. Matsuda, R. Kitaura, S. Kitagawa, Y. Kubota, R. V. Belosludov, T.

C. Kobayashi, H. Sakamoto, T. Chiba, M. Takata, Y. Kawazoe, Y. Mita,

Nature 2005,436, 238.

[59] C.-T. He, Z.-M. Ye, Y.-T. Xu, D.-D. Zhou, H.-L. Zhou, D. Chen, J.-P.

Zhang, X.-M. Chen, Chem. Sci. 2017,8, 7560.

[60] H. Zeng, M. Xie, Y. L. Huang, Y. Zhao, X. J. Xie, J. P. Bai, M. Y. Wan,

R. Krishna, W. Lu, D. Li, Angew. Chem., Int. Ed. 2019,58, 8515.

[61] D. M. D’Alessandro, B. Smit, J. R. Long, Angew. Chem., Int. Ed. 2010,

49, 6058.

[62] M. R. Abdul Hamid, Y. Qian, R. Wei, Z. Li, Y. Pan, Z. Lai, H.-K. Jeong,

J. Membr. Sci. 2021,640, 119802.

[63] a) L. Wang, Y. Li, Y. Wang, J. Yang, L. Li, J. Li, Sep. Purif. Technol. 2020,

251, 117311; b) O. T. Qazvini, L. K. Macreadie, S. G. Telfer, Chem.

Mater. 2020,32, 6744; c) O. T. Qazvini, V.-J. Scott, L. Bondorf, M.

Ducamp, M. Hirscher, F.-X. Coudert, S. G. Telfer, Chem. Mater. 2021,

33, 8886.

[64] a) Y. Takakura, S. Sugimoto, J. Fujiki, H. Kajiro, T. Yajima, Y. Kawajiri,

ACS Sustainable Chem. Eng. 2022,10, 14935; b) S. Hiraide, H.

Tanaka, N. Ishikawa, M. T. Miyahara, ACS Appl Mater. Interfaces

2017,9, 41066; c) R. Haldar, M. Inukai, S. Horike, K. Uemura, S.

Kitagawa, T. K. Maji, Inorg. Chem. 2016,55, 4166.

[65] J. M. Huck, L.-C. Lin, A. H. Berger, M. N. Shahrak, R. L. Martin, A. S.

Bhown, M. Haranczyk, K. Reuter, B. Smit, Energy Environ. Sci. 2014,

7, 4132.

[66] a) L. C. Lin, A. H. Berger, R. L. Martin, J. Kim, J. A. Swisher, K.

Jariwala, C. H. Rycroft, A. S. Bhown, M. W. Deem, M. Haranczyk,

B. Smit, Nat. Mater. 2012,11, 633; b) Z. Hu, Y. Wang, B. B. Shah, D.

Zhao, Adv. Sustainable Syst. 2019,3, 1970002.

[67] a) P. Pullumbi, F. Brandani, S. Brandani, Curr. Opin. Chem. Eng.

2019,24, 131; b) Y.-S. Wang, X.-J. Zhang, Y.-Q. Ba, T.-Y. Li, G.-P.

Hao, A.-H. Lu, Research 2022,2022,https://doi.org/10.34133/2022/

9780864.

[68] a) R.-B. Lin, S. Xiang, W. Zhou, B. Chen, Chem 2020,6, 337; b) C.

Jiang, X. Wang, Y. Ouyang, K. Lu, W. Jiang, H. Xu, X. Wei, Z. Wang,

F. Dai, D. Sun, Nanoscale Adv. 2022,4, 2077; c) X. Zhao, Y. Wang, D.

S. Li, X. Bu, P. Feng, Adv. Mater. 2018,30, 1705189; d) X. Zhang, Y.

Li, J.-R. Li, Trends Chem. 2024,6, 22.

[69] M. K. Taylor, T. Runcevski, J. Oktawiec, J. E. Bachman, R. L.

Siegelman, H. Jiang, J. A. Mason, J. D. Tarver, J. R. Long, J. Am.

Chem. Soc. 2018,140, 10324.

[70] J. U. Keller, F. Dreisbach, H. Rave, R. Staudt, M. Tomalla, Adsorption

1999,5, 199.

[71] D. Shade, B. W. S. Bout, D. S. Sholl, K. S. Walton, Ind. Eng. Chem.

Res. 2022,61, 2367.

[72] K. Sumida, D. L. Rogow, J. A. Mason, T. M. McDonald, E. D. Bloch,

Z.R.Herm,T.H.Bae,J.R.Long,Chem. Rev. 2012,112,724.

[73] O. Talu, Adv. Colloid Interface Sci. 1998,76–77, 227.

[74] S. Hiraide, Y. Sakanaka, Y. Iida, H. Arima, M. T. Miyahara, S.

Watanabe, Proc. Natl. Acad. Sci. USA 2023,120, 2305573120.

[75] a) F. X. Coudert, C. Mellot-Draznieks, A. H. Fuchs, A. Boutin, J. Am.

Chem. Soc. 2009,131, 11329; b) F.-X. Coudert, Phys. Chem. Chem.

Phys. 2010,12, 10904.

[76] S. M. J. Rogge, R. Goeminne, R. Demuynck, J. J. Gutiérrez-Sevillano,

S. Vandenbrande, L. Vanduyfhuys, M. Waroquier, T. Verstraelen, V.

Van S pe ybroe ck, Adv. Theory Simul. 2019,2, 1800177.

[77] a) F. J. Sotomayor, C. M. Lastoskie, Langmuir 2017,33, 11670; b) H.

Tanaka, S. Hiraide, A. Kondo, M. T. Miyahara, J. Phys. Chem. C 2015,

119, 11533.

[78] G. Fraux, A. Boutin, A. H. Fuchs, F.-X. Coudert, Adsorption 2018,24,

233.

[79] a) S. Sircar, Ind. Eng. Chem. Res. 2006,46, 2917; b) A. Rajendran, V.

Kariwala, S. Farooq, Chem. Eng. Sci. 2008,63, 2696.

[80] T.R.C.VanAssche,G.V.Baron,J.F.M.Denayer,Dalton Trans. 2016,

45, 4416.

[81] L. Hamon, P. L. Llewellyn, T. Devic, A. Ghouﬁ, G. Clet, V. Guillerm,

G. D. Pirngruber, G. Maurin, C. Serre, G. Driver, W. van Beek, E.

Jolimaitre, A. Vimont, M. Daturi, G. Ferey, J. Am. Chem. Soc. 2009,

131, 17490.

[82] V. Finsy, L. Ma, L. Alaerts, D. E. De Vos, G. V. Baron, J. F. M. Denayer,

Microporous Mesoporous Mater. 2009,120, 221.

[83] F.-X. Coudert, M. Jeﬀroy, A. H. Fuchs, A. Boutin, C. Mellot-

Draznieks, J. Am. Chem. Soc. 2008,130, 14294.

[84] a) N. Klein, C. Herzog, M. Sabo, I. Senkovska, J. Getzschmann,

S. Paasch, M. R. Lohe, E. Brunner, S. Kaskel, Phys. Chem. Chem.

Phys. 2010,12, 11778; b) N. Klein, H. C. Hoﬀmann, A. Cadiau, J.

Getzschmann, M. R. Lohe, S. Paasch, T. Heydenreich, K. Adil, I.

Senkovska, E. Brunner, S. Kaskel, J. Mater. Chem. 2012,22, 10303;

c) N. Kavoosi, T. Savchenko, I. Senkovska, M. Maliuta, V. Bon, A.

www.advancedsciencenews.com www.advmat.de

Eychmüller, S. Kaskel, Microporous Mesoporous Mater. 2018,271,

169.

[85] A. Schneemann, Y. Jing, J. D. Evans, T. Toyao, Y. Hijikata, Y. Kamiya,

K. I. Shimizu, N. C. Burtch, S. I. Noro, Dalton Trans. 2021,50, 10423.

[86] K. Roztocki, M. Rauche, V. Bon, S. Kaskel, E. Brunner, D. Matoga,

ACS Appl. Mater. Interfaces 2021,13, 28503.

[87] M. Ichikawa, A. Kondo, H. Noguchi, N. Kojima, T. Ohba, H. Kajiro,

Y. Hattori, H. Kanoh, Langmuir 2016,32, 9722.

[88] Q. Dong, X. Zhang, S. Liu, R. B. Lin, Y. Guo, Y. Ma, A. Yonezu, R.

Krishna, G. Liu, J. Duan, R. Matsuda, W. Jin, B. Chen, Angew. Chem.,

Int. Ed. 2020,59, 22756.

[89] B. R. Barnett, M. I. Gonzalez, J. R. Long, Trends Chem. 2019,1, 159.

[90] C. Gücüyener, J. van den Bergh, J. Gascon, F. Kapteijn, J. Am. Chem.

Soc. 2010,132, 17704.

[91] L. Li, R. Krishna, Y. Wang, J. Yang, X. Wang, J. Li, J. Mater. Chem. A

2016,4, 751.

[92] S. Sen, N. Hosono, J.-J. Zheng, S. Kusaka, R. Matsuda, S. Sakaki, S.

Kitagawa, J. Am. Chem. Soc. 2017,139, 18313.

[93] K. Kishida, Y. Okumura, Y. Watanabe, M. Mukoyoshi, S. Bracco, A.

Comotti, P. Sozzani, S. Horike, S. Kitagawa, Angew. Chem., Int. Ed.

2016,55, 13784.

[94] J. Y. Kim, J. Park, J. Ha, M. Jung, D. Wallacher, A. Franz, R. Balderas-

Xicohténcatl, M. Hirscher, S. G. Kang, J. T. Park, I. H. Oh, H. R.

Moon, H. Oh, J. Am. Chem. Soc. 2020,142, 13278.

[95] a) Y. Liu, J.-H. Her, A. Dailly, A. J. Ramirez-Cuesta, D. A. Neumann,

C. M. Brown, J. Am. Chem. Soc. 2008,130, 11813; b) M. Mendt, B.

Jee, N. Stock, T. Ahnfeldt, M. Hartmann, D. Himsl, A. Pöppl, J. Phys.

Chem. C 2010,114, 19443.

[96] C.R.Groom,I.J.Bruno,M.P.Lightfoot,S.C.Ward,Acta Cryst. B

2016,72, 171.

[97] L. Bondorf, J. L. Fiorio, V. Bon, L. Zhang, M. Maliuta, S. Ehrling, I.

Senkovska, J. D. Evans, J. O. Joswig, S. Kaskel, T. Heine, M. Hirscher,

Sci. Adv. 2022,8, 7035.

[98] a) J. D. Evans, V. Bon, I. Senkovska, S. Kaskel, Langmuir 2021,

37, 4222; b) IUPAC project, focussing on standardized reporting

of gas adsorption isotherms: https://iupac.org/project/2021-016-

1-024 (accessed: September 2024).

[99] V. Bon, N. Klein, I. Senkovska, A. Heerwig, J. Getzschmann, D.

Wallacher, I. Zizak, M. Brzhezinskaya, U. Mueller, S. Kaskel, Phys.

Chem. Chem. Phys. 2015,17, 17471.

[100] T. Jacobs, G. O. Lloyd, J.-A. Gertenbach, K. K. Müller-Nedebock, C.

Esterhuysen, L. J. Barbour, Angew. Chem., Int. Ed. 2012,51, 4913.

[101] H. Sato, W. Kosaka, R. Matsuda, A. Hori, Y. Hijikata, R. V.

Belosludov, S. Sakaki, M. Takata, S. Kitagawa, Science 2014,343,

167.

[102] P. M. Bhatt, Y. Belmabkhout, A. Cadiau, K. Adil, O. Shekhah, A.

Shkurenko, L. J. Barbour, M. Eddaoudi, J. Am. Chem. Soc. 2016,138,

9301.

[103] S.-M. Wang, M. Shivanna, S.-T. Zheng, T. Pham, K. A. Forrest, Q.-Y.

Yang, Q. Guan, B. Space, S. Kitagawa, M. J. Zaworotko, J. Am. Chem.

Soc. 2024,146, 4153.

[104] a) E. Tal-Gutelmacher, D. Eliezer, E. Abramov, Mater. Sci. Eng., A

2007,445, 625; b) L. Zhang, S. Jee, J. Park, M. Jung, D. Wallacher,

A. Franz, W. Lee, M. Yoon, K. Choi, M. Hirscher, H. Oh, J. Am.

Chem. Soc. 2019,141, 19850; c) J. Y. Kim, L. Zhang, R. Balderas-

Xicohténcatl, J. Park, M. Hirscher, H. R. Moon, H. Oh, J. Am. Chem.

Soc. 2017,139, 17743; d) M. Liu, L. Zhang, M. A. Little, V. Kapil, M.

Ceriotti, S. Yang, L. Ding, D. L. Holden, R. Balderas-Xicohténcatl, D.

He, R. Clowes, S. Y. Chong, G. Schütz, L. Chen, M. Hirscher, A. I.

Cooper, Science 2019,366, 613.

[105] L. Zhang, S. Jee, J. Park, M. Jung, D. Wallacher, A. Franz, W. Lee,

M. Yoon, K. Choi, M. Hirscher, H. Oh, J. Am. Chem. Soc. 2019,141,

19850.

[106] a) H. Miura, V. Bon, I. Senkovska, S. Ehrling, N. Bonisch, G. Mader,

S. Grunzner, A. Khadiev, D. Novikov, K. Maity, A. Richter, S. Kaskel,

Adv. Mater. 2023,35, 2207741; b) J. D. Evans, V. Bon, I. Senkovska,

H. C. Lee, S. Kaskel, Nat. Commun. 2020,11, 2690.

[107] D. Tanaka, K. Nakagawa, M. Higuchi, S. Horike, Y. Kubota, T. C.

Kobayashi, M. Takata, S. Kitagawa, Angew. Chem., Int. Ed. 2008,47,

3914.

[108] L. Schaper, R. Schmid, Commun. Chem. 2023,6, 233.

[109] a) S. Sircar, H. Wu, J. Li, A. D. Lueking, Langmuir 2011,27, 14169;

b) A. D. Lueking, C.-Y. Wang, S. Sircar, C. Malencia, H. Wang, J. Li,

Dalton Trans. 2016,45, 4242.

[110] H. Preißler-Kurzhöfer, M. Lange, J. Möllmer, O. Erhart, M. Kobalz,

H. Krautscheid, R. Gläser, Nanomaterials 2024,14, 241.

[111] M. Kobalz, J. Lincke, K. Kobalz, O. Erhart, J. Bergmann, D. Lässig,

M. Lange, J. Möllmer, R. Gläser, R. Staudt, H. Krautscheid, Inorg.

Chem. 2016,55, 3030.

[112] N. Hosono, A. Terashima, S. Kusaka, R. Matsuda, S. Kitagawa, Nat.

Chem. 2019,11, 109.

[113] B. Seoane, S. Sorribas, Á. Mayoral, C. Téllez, J. Coronas, Microporous

Mesoporous Mater. 2015,203, 17.

[114] V. Bon, N. Busov, I. Senkovska, N. Bönisch, L. Abylgazina,

A. Khadiev, D. Novikov, S. Kaskel, Chem. Commun. 2022,58,

10492.

[115] Y. Sakanaka, S. Hiraide, I. Sugawara, H. Uematsu, S. Kawaguchi, M.

T. Miyahara, S. Watanabe, Nat. Commun. 2023,14, 6862.

[116] A. V. Neimark, F. X. Coudert, C. Triguero, A. Boutin, A. H. Fuchs, I.

Beurroies, R. Denoyel, Langmuir 2011,27, 4734.

[117] a) J. Troyano, D. Maspoch, Chem. Commun. 2023,59, 1744; b) P.

Freund, I. Senkovska, B. Zheng, V. Bon, B. Krause, G. Maurin, S.

Kaskel, Chem. Commun. 2020,56, 7411.

[118] S. Hiraide, H. Tanaka, N. Ishikawa, M. T. Miyahara, ACS Appl. Mater.

Interfaces 2017,9, 41066.

[119] A. Kondo, H. Noguchi, S. Ohnishi, H. Kajiro, A. Tohdoh, Y. Hattori,

W. C. Xu, H. Tanaka, H. Kanoh, K. Kaneko, Nano Lett. 2006,6,

2581.

[120] V. Bon, N. Kavoosi, I. Senkovska, S. Kaskel, ACS Appl. Mater. Inter-

faces 2015,7, 22292.

[121] J. Kim, W. Y. Kim, W.-S. Ahn, Fuel 2012,102, 574.

[122] M. Maliuta, I. Senkovska, V. Romaka, M. Roslova, Z. Huang, P.

Petkov, V. Bon, S. Kaskel, CCS Chem 2023,5, 2225.

[123] D. Bazer-Bachi, L. Assié, V. Lecocq, B. Harbuzaru, V. Falk, Powder

Technol. 2014,255, 52.

[124] S. Hiraide, H. Arima, H. Tanaka, M. T. Miyahara, ACS Appl Mater.

Interfaces 2021,13, 30213.

[125] M. Kriesten, J. Vargas Schmitz, J. Siegel, C. E. Smith, M. Kaspereit,

M. Hartmann, Eur. J. Inorg. Chem. 2019,2019, 4700.

[126] K. Titov, Z. Zeng, M. R. Ryder, A. K. Chaudhari, B. Civalleri, C. S.

Kelley, M. D. Frogley, G. Cinque, J. C. Tan, J. Phys. Chem. Lett. 2017,

8, 5035.

[127] J. Liu, B. Li, V. Martins, Y. Huang, Y. Song, J. Phys. Chem. C 2024,

128, 8007.

[128] H. Arima, S. Hiraide, M. T. Miyahara, S. Watanabe, ACS Appl Mater.

Interfaces 2023,15, 36975.

[129] S. Hiraide, Y. Sakanaka, H. Kajiro, S. Kawaguchi, M. T. Miyahara, H.

Tana k a , Nat. Commun. 2020,11, 3867.

[130] S. Horike, Y. Inubushi, T. Hori, T. Fukushima, S. Kitagawa, Chem. Sci.

2012,3, 116.

[131] J. F. Horstmeier, A. G. Lopez, D. W. Agar, Int. J. Greenhouse Gas Con-

trol 2016,47, 364.

[132] S. Horike, S. Shimomura, S. Kitagawa, Nat. Chem. 2009,1, 695.

[133] V. Colombo, C. Montoro, A. Maspero, G. Palmisano, N. Masciocchi,

S.Galli,E.Barea,J.A.R.Navarro,J. Am. Chem. Soc. 2012,134,

12830.

www.advancedsciencenews.com www.advmat.de

Irena Senkovska obtained her diploma in chemistry from Ivan Franko National University of Lviv,

Ukraine, in 1995, and earned her Ph.D. in natural sciences from Ulm University,Germany, in 2004.

Since 2005, she has been a senior scientist at the Chair of Inorganic Chemistry at Dresden University

of Technology. Her research focuses on the design, synthesis, and application of metal–organic frame-

works, mainly focusing on ﬂexibility phenomena and adsorption processes.

Volodymyr Bon received his Ph.D. from the Institute of General and Inorganic Chemistry, National

Academy of Sciences of Ukraine in 2008 and his Venia Legendi from TU Dresden in 2024. His research

interests include crystal engineering, synthesis and crystallographic characterization of novel crys-

talline porous solids for adsorption-related applications. A particular focus is the development of

advanced in situ X-ray diﬀraction/adsorption characterization techniques in large scale facilities. He

is currently coordinating a subgroup focused on the synthesis of new ﬂexible MOFs and the in-depth

study of stimuli-induced switching phenomena.

Antonia Mosberger obtained her B.Sc. in 2021 from the University of Bayreuth, working under the

supervision of Prof. Jürgen Senker.After relocating to Dresden, she completed her Master’s studies

in the research group of Prof. Stefan Kaskel, focusing on developing new metal–organic frameworks

(MOFs) for gas storage and separation applications. Since early 2024, she has continued her research

as a Ph.D. candidate, broadening her focus to include ﬂexible MOFs for enhanced gas separation.

YutongWang obtained his Ph.D. degree (2022) in chemistry under the supervision of Prof. Daofeng

Sun from the China University of Petroleum (East China). Then he joined the group of Prof. Stefan

Kaskel at TU Dresden as a postdoctoral researcher under the support of the Humboldt Fellowship. His

research interests focus on the structural control of ﬂexible functionalized metal–organic frameworks

for hydrocarbon adsorption and separation.

Plastic Pores for Switchable and Optimized Adsorption Behaviors

Article

Mar 2025

Mobile Constituent-Boosted Dynamic Separation of C2H2/C2H4/CO2 Ternary Mixtures in Metal-Organic Frameworks

Article

Mar 2025
J AM CHEM SOC

Investigating the Dynamics of a Soft Crystalline Covalent Organic Framework during Benzene and Cyclohexane Adsorption by in situ Powder X‐ray Diffraction

Article

Full-text available

Oct 2024

Due to their similar boiling points, separation of benzene and cyclohexane mixtures is among the current challenging processes faced by the petrochemical industry. As recently assessed, the soft imine‐based covalent organic framework [(TAM)(BDA)2] (COF‐300; TAM = tetrakis(4‐aminophenyl)methane, BDA = terephthaldehyde) possesses higher affinity for benzene than cyclohexane in both static conditions at 298 K and dynamic conditions in the range of 298–348 K. As shown in this contribution, in situ powder X‐ray diffraction while dosing benzene and cyclohexane vapors in the range of 0.01–4.74 bar on the narrow‐pore form of COF‐300 confirmed the coherent switchability of its framework, unveiling the progressive formation of different intermediate‐ and large‐pore forms. In addition, a basket of otherwise inaccessible key crystallochemical details—“on/off” structural‐feature changes cooperating to adsorption, primary adsorption sites, and host–guest and guest–guest interactions—was successfully retrieved. Overall, these findings allowed to shed light on the framework dynamics underneath the previously observed selectivity toward benzene over cyclohexane, completing this case of study and providing relevant information for the design of new‐generation adsorbents for this applicative context.

A holistic platform for accelerating sorbent-based carbon capture

Article

Full-text available

Jul 2024
NATURE

Reducing carbon dioxide (CO2) emissions urgently requires the large-scale deployment of carbon-capture technologies. These technologies must separate CO2 from various sources and deliver it to different sinks1,2. The quest for optimal solutions for specific source–sink pairs is a complex, multi-objective challenge involving multiple stakeholders and depends on social, economic and regional contexts. Currently, research follows a sequential approach: chemists focus on materials design³ and engineers on optimizing processes4,5, which are then operated at a scale that impacts the economy and the environment. Assessing these impacts, such as the greenhouse gas emissions over the plant’s lifetime, is typically one of the final steps⁶. Here we introduce the PrISMa (Process-Informed design of tailor-made Sorbent Materials) platform, which integrates materials, process design, techno-economics and life-cycle assessment. We compare more than 60 case studies capturing CO2 from various sources in 5 global regions using different technologies. The platform simultaneously informs various stakeholders about the cost-effectiveness of technologies, process configurations and locations, reveals the molecular characteristics of the top-performing sorbents, and provides insights on environmental impacts, co-benefits and trade-offs. By uniting stakeholders at an early research stage, PrISMa accelerates carbon-capture technology development during this critical period as we aim for a net-zero world.

Flexible hydrogen-bonded organic frameworks (HOFs): opportunities and challenges

Article

Full-text available

May 2024

Flexible behavior is one of the most fascinating features of hydrogen-bonded organic frameworks (HOFs), which represent an emerging class of porous materials that are self-assembled via H-bonding between organic building units. Due to their unique flexibility, HOFs can undergo structural changes or transformations in response to various stimuli (physical or chemical). Taking advantage of this unique structural feature, flexible HOFs show potential in multifunctional applications such as gas storage/separation, molecular recognition, sensing, proton conductivity, biomedicine, etc. While some other flexible porous materials have been extensively studied, the dynamic behavior of HOFs remains relatively less explored. This perspective highlights the inherent flexible properties of HOFs, discusses their different flexible behaviors, including pore size/shape changes, interpenetration/stacking manner, H-bond breaking/reconstruction, and local dynamic behavior, and highlights their potential applications. We believe that this perspective will not only contribute to HOF chemistry and materials science, but will also facilitate the ongoing extensive research on dynamic porous materials.

Supramolecular Entanglement in a Hydrogen‐Bonded Organic Framework Enables Flexible‐Robust Porosity for Highly Efficient Purification of Natural Gas

Article

Full-text available

May 2024
ANGEW CHEM INT EDIT

The development of porous materials with flexible‐robust characteristics shows some unique advantages to target high performance for gas separation, but remains a daunting challenge to achieve so far. Herein, we report a carboxyl‐based hydrogen‐bonded organic framework (ZJU‐HOF‐8a) with flexible‐robust porosity for efficient purification of natural gas. ZJU‐HOF‐8a features a four‐fold interpenetrated structure with dia topology, wherein abundant supramolecular entanglements are formed between the adjacent subnetworks through weak intermolecular hydrogen bonds. This structural configuration could not only stabilize the whole framework to establish the permanent porosity, but also enable the framework to show some flexibility due to its weak intermolecular interactions (so‐called flexible‐robust framework). The flexible‐robust porosity of ZJU‐HOF‐8a was exclusively confirmed by gas sorption isotherms and single‐crystal X‐ray diffraction studies, showing that the flexible pore pockets can be opened by C3H8 and n‐C4H10 molecules rather by C2H6 and CH4. This leads to notably higher C3H8 and n‐C4H10 uptakes with enhanced selectivities than C2H6 over CH4 under ambient conditions, affording one of the highest n‐C4H10/CH4 selectivities. The gas‐loaded single‐crystal structures coupled with theoretical simulations reveal that the loading of n‐C4H10 can induce an obvious framework expansion along with pore pocket opening to improve n‐C4H10 uptake and selectivity, while not for C2H6 adsorption. This work suggests an effective strategy of designing flexible‐robust HOFs for improving gas separation properties.

Metal–organic frameworks for petroleum-based platform compound separations

Article

Full-text available

Mar 2024

The advancement of separation technology is not only conducive to lowering energy consumption, but also opens an avenue to obtain the world's key resources. Metal–organic frameworks (MOFs), a novel type of porous materials, have unique advantages in separation and purification compared with traditional adsorbents and separation technologies, especially in the separation of petroleum-based platform compounds, owing to their advantages of easy structure regulation and functionalization. Significant advancements have been achieved in the separation of binary component petroleum-based platform compounds based on MOFs. However, industrial gas mixtures often exist in multiple components. In contrast to the rapid development of binary mixture separations, the separation of ternary and even multi-component mixtures is much more difficult and rarely achieved with a single material. This review briefly introduces the adsorptive separation technology and its mechanism based on MOFs. Next, recent advances in MOFs for the separation of multiple-component petroleum-based platform compounds (C2H2/C2H4/C2H6/CO2, C3H4/C3H6/C3H8, butane/butene/isobutene/1,3-butadiene, trans-2-pentene/1-pentene/isoprene, n-hexane/2-methylpentane/3-methylpentane/2,2-dimethylbutane/2,3-dimethylbutane, o-xylene/m-xylene/p-xylene/ethylbenzene, etc.) are presented to guide more complex chemical separation processes. Furthermore, inherent obstacles and future development prospects from academia to eventual industrial implementation are presented. Based on adsorptive separation technology, we are committed to exploring an alternative separation route that is energy-efficient and environmentally friendly and strives to achieve the high-efficiency separation of multi-component petroleum-based platform compounds.

Hydrocarbon Sorption in Flexible MOFs—Part III: Modulation of Gas Separation Mechanisms

Article

Full-text available

Jan 2024

Single gas sorption experiments with the C4-hydrocarbons n-butane, iso-butane, 1-butene and iso-butene on the flexible MOFs Cu-IHMe-pw and Cu-IHEt-pw were carried out with both thermodynamic equilibrium and overall sorption kinetics. Subsequent static binary gas mixture experiments of n-butane and iso-butane unveil a complex dependence of the overall selectivity on sorption enthalpy, rate of structural transition as well as steric effects. A thermodynamic separation favoring iso-butane as well as kinetic separation favoring n-butane are possible within Cu-IHMe-pw while complete size exclusion of iso-butane is achieved in Cu-IHEt-pw. This proof-of-concept study shows that the structural flexibility offers additional levers for the precise modulation of the separation mechanisms for complex mixtures with similar chemical and physical properties with real selectivities of >10.

Vibrational dynamics dictate heat transport in metal-organic frameworks with gas adsorbates

Article