Photo-switchable smart metal-organic framework membranes with tunable and enhanced molecular sieving performance

Article abstract of DOI:10.1039/C8TA10541C

Developing a novel MOF membrane material with switchable separation performance is an exciting and challenging research project. In the present work, we report preparation of a new kind of light induced smart MOF membrane, i.e., Cu(AzDC)(4,4′-BPE)_0.5 membrane, which shows (i) enhanced molecular sieving performance, and (ii) is able to respond quickly to external light stimuli. Two photo-switchable moieties are addressed in the Cu(AzDC)(4,4′-BPE)_0.5 membrane: azobenzene and bis(4-pyridyl)ethylene. When the Cu(AzDC)(4,4′-BPE)_0.5 membrane is in situ irradiated with Vis and UV light, the separation factor of a H₂/CO₂ mixture can be switched reversibly between 21.3 and 43.7. This switching effect is mainly caused by reduced CO₂ adsorption in the UV-cis state as proven by independent adsorption studies. For a steric reason, adsorption of CO₂ is limited for the UV-cis state. In full agreement with this model, the adsorption of other gases H₂, CH₄ and N₂ as well as their permeation behaviour is not observably influenced by the trans-cis switching.

Full text of DOI:10.1039/C8TA10541C

Journal of
Materials Chemistry A
View Article Online
View Journal | View Issue
PAPER
Photo-switchable smart metal–organic framework
membranes with tunable and enhanced molecular
sieving performance†
Cite this: J. Mater. Chem. A, 2018, 6,
24949
a
b
c
Chuanyao Liu,ab Yunzhe Jiang, Chen Zhou, Jurgen Caro
*
¨
a
*
and Aisheng Huang
Developing a novel MOF membrane material with switchable separation performance is an exciting and
challenging research project. In the present work, we report preparation of a new kind of light induced
smart MOF membrane, i.e., Cu(AzDC)(4,40-BPE)0.5 membrane, which shows (i) enhanced molecular
sieving performance, and (ii) is able to respond quickly to external light stimuli. Two photo-switchable
moieties are addressed in the Cu(AzDC)(4,40-BPE)0.5 membrane: azobenzene and bis(4-pyridyl)ethylene.
When the Cu(AzDC)(4,40-BPE)0.5 membrane is in situ irradiated with Vis and UV light, the separation
factor of a H2/CO2 mixture can be switched reversibly between 21.3 and 43.7. This switching effect is
mainly caused by reduced CO2 adsorption in the UV-cis state as proven by independent adsorption
studies. For a steric reason, adsorption of CO2 is limited for the UV-cis state. In full agreement with this
model, the adsorption of other gases H2, CH4 and N2 as well as their permeation behaviour is not
observably influenced by the trans–cis switching.
Received 2nd November 2018
Accepted 21st November 2018
DOI: 10.1039/c8ta10541c
rsc.li/materials-a
membranes,5,6 microporous zeolite membranes,7–11 silica
membranes,12 and carbon membranes13 have been developed
Introduction
for the separation of H2 from CO2. Pd-based metal membranes
have shown high H2 selectivity and H2 permeance at high
temperature, but they are easily poisoned by CO or H2S.6
Further, Pd membranes are too expensive to be widely used.
Silica membranes have displayed high H2 selectivity over other
small molecular gases, but suffer from instability even in traces
of steam. Zeolite membranes usually show relatively low H2/CO2
selectivity due to the presence of intercrystalline defects.
Therefore, the development of stable molecular sieving
membranes with high H2 permselectivity is still challenging.
Recently, metal–organic framework (MOF) membranes have
attracted intense attention for the fabrication of superior
molecular sieve membranes due to their highly diversied pore
structures and pore sizes as well as specic adsorption affini-
ties. In the past 10 years, a great deal of research effort has been
focused on the preparation of MOF or MOF-based membranes
with molecular sieving performance.14–25 However, with respect
to the fabrication of MOF membranes, not only is the facile
growth of a dense polycrystalline layer on porous supports ex-
pected, but also the remote control of membrane permeance/
selectivity by external stimuli is highly desired. In this case,
mass transfer and interfacial properties of the membranes can
be reversibly switched by external stimuli, such as temperature,
pH, solution ionic strength, light, electric and magnetic elds,
and chemical cues.26–30 Among these external stimuli, light is
a popular, promising and fascinating candidate because both
its irradiation time and wavelength can be precisely controlled
Due to the change of global climate, crisis of energy security,
and aggravation of local air pollution, there is a growing
demand for a clean and efficient energy in recent years, which is
the concept of “hydrogen economy”.1 Currently, hydrogen is
mainly produced by steam-methane reforming (SMR) followed
by the water–gas shi (WGS) reaction.2 The resulting gas
mixture is mainly composed of H2 (0.29 nm) and CO2 (0.33 nm).
Before the produced hydrogen gas can be used, it has to be
separated and puried from the gas mixtures, which are usually
realized by pressure swing adsorption (PSA) or cryogenic
distillation. In comparison with traditional H2 purication
methods, membrane-based separation techniques have attrac-
ted signicant attention due to their lower energy consumption
and investment cost.3 Because organic polymer membranes
usually suffer from the instability problem at high temperature,
inorganic membranes are more attractive under harsh separa-
tion conditions.4 In recent years, dense Pd-based metal
aShanghai Key Laboratory of Green Chemistry and Chemical Processes, Department of
Chemistry, East China Normal University, Dongchuan Road 500, Shanghai 200241,
China. E-mail: huangash@chem.ecnu.edu.cn
bInstitute of New Energy Technology, Ningbo Institute of Materials Technology and
Engineering, Chinese Academy of Sciences, 1219 Zhongguan Road, 315201 Ningbo,
P. R. China. E-mail: zhouchen@nimte.ac.cn
cInstitute of Physical Chemistry and Electrochemistry, Leibniz University Hannover,
Callinstr. 3A, D-30167, Hannover, Germany
† Electronic supplementary information (ESI) available: Experimental details,
FESEM, XRD, FT-IR, TGA, etc. See DOI: 10.1039/c8ta10541c
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A, 2018, 6, 24949–24955 | 24949

View Article Online
Journal of Materials Chemistry A
Paper
without introducing additional chemical species into the
systems.31–34 Therefore, photo-switchable molecules, which can
realize reversible photo-isomerization when irradiated with
light of a certain wavelength, have drawn intense interest. So
far, several photo-responsive molecules have been successfully
incorporated into MOF frameworks to modify and control MOF
properties such as colour,33 release34,35 and adsorption36–42 of
guest molecules. One of the most widely studied photo-
responsive molecules in MOFs is azobenzene, which is well
known for its reversible trans/cis photoisomerization, with
˚
switching its molecular size from 9 to 5.5 A, and the dipole
moment from 0.5 to 3.1 debye.36,37 For the rst time, Wang et al.
reported a photo-switchable MOF [Cu2(AzoBPDC)2(AzoBiPyB)]
membrane which offers dynamic control of selectivity by remote
signals,40 which is realized by assembling linkers containing
photoresponsive azobenzene-side-groups. The azobenzene
moieties can be switched from the trans to the cis conguration
and vice versa by irradiation with ultraviolet or visible light.
Therefore, through precise control of the ratio of cis/trans azo-
benzene by controlled irradiation times or by simultaneous
irradiation with ultraviolet or visible light, it is effective to tune
continuously the membrane permeability and selectivity of the
Fig. 1 The structure of (c) Cu(AzDC)(4,40-BPE)0.5 with the (a) azo-
benzene and (b) bis(4-pyridyl)ethylene photo-switchable groups. The
dynamic photo-switching behavior of Cu(AzDC)(4,40-BPE)0.5 will
result in reversible CO2 adsorption as shown in Fig. S1.†
(Fig. 1a and b), resulting in squeezing out and extrusion of the
adsorbed CO2 in Zn(AzDC)(4,40-BPE)0.5 (Fig. 1c and S1†).
Therefore, it can be expected that the Cu(AzDC)(4,40-BPE)0.5
membrane will display a switchable selectivity in the separation
of H2 from CO2 upon irradiation with UV or Vis light (Fig. S2†).
¨
MOF membrane. Thereaer, Muller et al. prepared a photo-
switchable MOF thin lm functionalized with o-uo-
roazobenzene moieties via layer-by-layer deposition,41 which
can reversibly isomerize between the trans and cis conguration
with visible light. Very recently, Knebel et al. reported Experimental
azobenzene-loaded UiO-67 membranes with photo-switchable
Synthesis of Cu(AzDC)(4,40-BPE)0.5 membranes
separation performance for the separation of H2 and CO2.42
However, the H2/CO2 separation factor of the photo-switchable
MOF membranes is still relatively low due to their large pore
size, i.e., about 8.0 for the Cu2(AzoBPDC)2(AzoBiPyB)
membrane,40 and about 14.7 for the azobenzene-loaded UiO-67
membrane.42 Therefore, the development of photo-switchable
MOF membranes with a small pore size is desired to enhance
the H2/CO2 selectivity.
Before synthesis of Cu(AzDC)(4,40-BPE)0.5 membranes on
porous a-Al2O3 disks, the top surface of the a-Al2O3 disks was
treated with dopamine (DPA) according to our previous
reports.20,45 For the synthesis of Cu(AzDC)(4,40-BPE)0.5
membranes, 0.113 g Cu(NO3)$3H2O was dissolved in 50 mL
DMF by ultrasonic treatment, and then the PDA-modied
a-Al2O3 disks were placed into the solution and heated to
373 K for 12 h. Aer the solution was cooled to room temper-
ature, the PDA-modied a-Al2O3 disks were taken out, and AzDC
(0.127 g) and BPE (0.045 g) were added to the above solution of
Cu(NO3)$3H2O and DMF. The pre-treated a-Al2O3 disks were
placed horizontally in a Teon-lined stainless-steel autoclave
which was lled with the synthesis solution and heated at 373 K
in an air-circulating oven for 24 h. Aer the solvothermal
synthesis, the Cu(AzDC)(4,40-BPE)0.5 membranes were washed
with DMF and methanol several times, and then dried in air at
333 K overnight.
Indeed, the tuning of micropores within porous materials is
of great importance for their application in the separation of
small molecules. Chen et al. prepared a triply interpenetrated
porous framework Zn(AzDC)(4,40-BPE)0.5 by using two different
photo-isomerizable linkers,43 namely, 4,40-dicarboxylate (AzDC)
and 1,2-bis(4-pyridyl)ethylene (4,40-BPE). With a small pore size
0
˚
of about 3.4 ꢀ 3.4 A, Zn(AzDC)(4,4 -BPE)0.5 exhibited a highly
selective sorption behavior toward H2/N2, H2/CO, and CO2/CH4.
Lately, Lyndon et al. reported that Zn(AzDC)(4,40-BPE)0.5 is
photosensitive to UV light, triggering the adsorption or release
of CO2 under Vis/UV light irradiation, showing promising
application in the separation of H2/CO2.44 In the present work,
Characterization
we report the design and synthesis of a new photo-switchable The FT-IR spectra were obtained by using a Bruker TENSOR27
MOF Cu(AzDC)(4,40-BPE)0.5 membrane. Cu(AzDC)(4,40-BPE)0.5 impact spectrometer. Raman spectra were obtained by using
has the same topological structure as Zn(AzDC)(4,40-BPE)0.5, in Renishaw in Via-reex Raman spectroscopy. All UV/Vis spectra
which the azobenzene and bis(4-pyridyl)ethylene moieties, as of the samples were recorded in a BaSO4 matrix in reection
the “backbone” of the linker, are incorporated directly into the geometry with a PerkinElmer LAMBDA 950. Thermogravimetric
framework, forming an interpenetrating network structure with analyses (TGA) were carried out using a Mettler Toledo TGA/
˚
a small pore size of about 3.4 ꢀ 3.4 A. In particular, under UV STDA 851e. Samples (10 mg) which had been solvent
irradiation, azobenzene and bis(4-pyridyl)ethylene moieties can exchanged with dry methanol and degassed at 423 K were
be reversibly switched from the trans to the cis conguration placed in 70 mL alumina pans and heated under a N2 gas ow
24950 | J. Mater. Chem. A, 2018, 6, 24949–24955
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
ꢀ
(20 mL minꢁ1) from 20 to 1073 K at a heating rate of 5 K minꢁ1
.
yi;Perm yj;Perm
ꢀ
The BET surface area and pore volume of Cu(AzDC)(4,40-BPE)0.5
were investigated using N2 adsorption–desorption isotherms at
77 K using nitrogen on an automatic volumetric adsorption
apparatus (Micrometrics ASAP 2020). Prior to the measure-
ments, the samples were heated at 423 K for 16 h under vacuum.
Aer adsorption, the samples were regenerated by degassing
under vacuum at room temperature for a few hours.
ai=j
¼
(1)
yi;Ret yj;Ret
The details of chemicals used for the synthesis of
Cu(AzDC)(4,40-BPE)0.5 crystals and membranes, the protocol of
the synthesis of organic ligand azobenzene-4,40-dicarboxylic
acid (AzDC) (Fig S4 and 5†), and the synthesis of Cu(AzDC)(4,40-
BPE)0.5 crystals are shown in the ESI.†
The morphology and thickness of the Cu(AzDC)(4,40-BPE)0.5
membranes were characterized by eld emission scanning
electron microscopy (FESEM). FESEM micrographs were taken
on an S-4800 (Hitachi) with a cold eld emission gun operating
Results and discussion
at 4 kV and 10 mA. The phase purity and crystallinity of the Cu(AzDC)(4,40-BPE)0.5 crystals were prepared by the sol-
Cu(AzDC)(4,40-BPE)0.5 crystals and membranes were conrmed vothermal reaction of copper(II) nitrate trihydrate, AzDC, and
by powder X-ray diffraction (PXRD). The PXRD patterns were BPE in DMF according to ref. 43 with minor modication (for
recorded at room temperature under ambient conditions with details see ESI†). As shown in Fig. S6,† well-shaped crystals with
a Bruker D8 ADVANCE X-ray diffractometer with Cu Ka radia- a block morphology of about 5–10 mm are formed by heating at
tion at 40 kV and 40 mA.
373 K for 24 h. As deduced from the PXRD pattern (Fig. S7b†),
Gas adsorption measurements were performed using ultra- all peaks of Cu(AzDC)(4,40-BPE)0.5 match well with those of the
high purity N2, CO2 and CH4. Before gas adsorption measure- simulated ones (Fig. S7a†), indicating that a phase-pure
ments at 298 K, the powder samples were activated at 393 K for Cu(AzDC)(4,40-BPE)0.5 with high crystallinity has been formed.
6 h under vacuum. Light-responsive gas adsorption was per- With UV irradiation, slight changes were observed in the XRD
formed under in situ UV/Vis light irradiation. For the switching pattern (Fig. S7c†). We assumed that the crystal structure of
experiments, two different LED light sources from PrizMatix Cu(AzDC)(4,40-BPE)0.5 will change aer irradiation with UV
were used. UV irradiation was performed with a 365 nm LED (365 nm), and this change should be reected in the XRD
(112 mW). Vis irradiation was performed with a 455 nm LED characterization. However, there are no obvious changes in the
(135 mW). The distance between the powder sample and the crystal structure which was conrmed by using TOPAS soware
LED was about 5 cm.
to rene the two XRD diffraction pattern data (Fig. S8 and S9,
and Table S1†), suggesting that the crystalline structure of
Cu(AzDC)(4,40-BPE)0.5 is not affected by irradiation with UV, i.e.,
by photoisomerization of the azobenzene groups, which is in
Single gas permeation and mixed gas separation
The volumetric ow rates of the single gases H2, CO2, N2, and good agreement with previous reports.40–42 The BET surface area
CH4 as well as of their equimolar binary mixtures of H2 with CO2 and pore volume of the Cu(AzDC)(4,40-BPE)0.5 were investigated
through the Cu(AzDC)(4,40-BPE)0.5 membranes were measured by N2 adsorption–desorption isotherms at 77 K. As shown in
at room temperature by using the Wicke–Kallenbach tech- Fig. S10,† the N2 isotherms of the Cu(AzDC)(4,40-BPE)0.5 show
nique.18–20 Before permeation measurements, the Cu(AzDC)(4,40- a typical hysteresis behaviour,43,44 indicating the permanent
BPE)0.5 membrane was on-stream activated to remove DMF porosity of the Cu(AzDC)(4,40-BPE)0.5 crystals. The calculated
solvent at 453 K. For single gas and mixture gas permeation, the BET surface area of the Cu(AzDC)(4,40-BPE)0.5 is about 227.9 m2
supported Cu(AzDC)(4,40-BPE)0.5 membrane was sealed in gꢁ1, which is similar to the previous report.44 TGA indicates that
a permeation module with silicone O-rings. The feed gases were the initial weight loss of Cu(AzDC)(4,40-BPE)0.5 is found between
fed to the top side of the membrane, and sweep gas N2 (CH4 was 298 and 453 K due to the evaporation of adsorbed DMF, and the
used as a sweep gas when measuring N2 permeation) was fed on activated Cu(AzDC)(4,40-BPE)0.5 is thermally stable up to about
the permeate side to keep the concentration of the permeating 573 K (Fig. S11†).
gas low providing a driving force for permeation. The total
The Cu(AzDC)(4,40-BPE)0.5 powder was measured in reec-
pressure on each side of the membrane was atmospheric. In situ tion geometry by UV-Vis and time-resolved FT-IR spectroscopy
irradiation of the Cu(AzDC)(4,40-BPE)0.5 membrane was achieved to study the light induced conformational trans–cis trans-
by a ber-coupled, monochromatic high power Prizmatix FC-3 formation of Cu(AzDC)(4,40-BPE)0.5. From the UV-Vis spectra
LED (Fig. S3†). Gas permeation was performed under in situ (Fig. 2), it can be seen that both azobenzene and bis(4-pyridyl)
irradiation alternating with l ¼ 365 nm and l ¼ 455 nm at room ethylene groups in the MOF can be reversibly switched by irra-
temperature. For both single and mixture gas permeation, the diation with UV light (365 nm) from trans to cis and back with
uxes of feed and sweep gases were determined with mass ow Vis light (455 nm). The time-resolved FT-IR spectra show
controllers, and a calibrated gas chromatograph (Agilent 7820A) signicant changes in peak intensity in the region of 500–
was used to measure the gas concentrations. The separation 800 cmꢁ1 under UV irradiation (Fig. S12†). Under UV irradia-
factor ai,j of binary mixture permeation is dened as the tion, the AzDC ligands in the Cu(AzDC)(4,40-BPE)0.5 frameworks
quotient of the molar ratios of the components (i, j) in show cis–trans isomerization since the C–C–C and C–C–N
the permeate, divided by the quotient of the molar ratio of the vibrations at 550 cmꢁ1 increase. The BPE ligands are less clear
components (i, j) in the retentate, as shown in eqn (1).
in the FT-IR spectra due to the overlapping of their bands with
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A, 2018, 6, 24949–24955 | 24951

View Article Online
Journal of Materials Chemistry A
Paper
wringing out a sponge, the framework structure with light-
responsive groups will be distorted when irradiated with UV
light, leading to signicant decreases in CO2 adsorption and
consequently promising potential in low energy CO2 release.42
However, N2 and CH4 adsorption show lower reduction aer UV
irradiation due to their lower interaction with Cu(AzDC)(4,40-
BPE)0.5, suggesting that Cu(AzDC)(4,40-BPE)0.5 is promising for
dynamitic capture and release of greenhouse gas CO2. Consid-
ering the photo-switchable properties of CO2 adsorption as well
0
˚
as a small pore size (3.4 ꢀ 3.4 A), Cu(AzDC)(4,4 -BPE)0.5 is an
ideal candidate to prepare smart molecular sieving membranes
for remote-control of membrane permeation and separation.
Due to the poor heterogeneous nucleation of Cu(AzDC)(4,40-
BPE)0.5 crystals on the surface of the a-Al2O3 support, it is
extremely difficult to form directly a continuous Cu(AzDC)(4,40-
BPE)0.5 layer on the a-Al2O3 support. As shown in our previous
reports,20 pre-modication of the support surface with polydop-
amine (PDA) is very helpful and effective to promote the nucle-
ation and growth of MOF and zeolite membranes on the support
surface through the formation of strong covalent bonds with the
MOF or zeolite crystals. Therefore, we prepare Cu(AzDC)(4,40-
BPE)0.5 membranes on the PDA-modied a-Al2O3 disks (ꢂfor
details see ESI†). Aer solvothermal reaction for 24 h at 100 C,
the surface of the PDA-modied a-Al2O3 supports has been
completely covered with continuous block-shaped Cu(AzDC)(4,40-
BPE)0.5 crystals (Fig. 4a). Additional Cu(AzDC)(4,40-BPE)0.5 crystals
are loosely packed on the top of the membranes due to sedi-
mentation, but from the cross-sectional view shown in Fig. 4b, the
Cu(AzDC)(4,40-BPE)0.5 membrane is continuous with a thickness
of about ꢃ10 mm. The XRD pattern of the Cu(AzDC)(4,40-BPE)0.5
membrane shows a high degree of crystallinity, and all of the
peaks match well with those of Cu(AzDC)(4,40-BPE)0.5 besides a-
Al2O3 signals from the support (Fig. 4c).
Fig. 2 UV/Vis spectra of the Cu(AzDC)(4,40-BPE)0.5 before irradiation
and after irradiation with UV light (365 nm) and Vis light (455 nm)
different times.
azobenzene groups. Further, the existence of abundant photo-
responsive groups in Cu(AzDC)(4,40-BPE)0.5 frameworks was
also tested using the Raman spectrum, which shows a strong
trans –N]N– vibration intensity at 1441 cmꢁ1 (Fig. S13†).
The photo-responsive properties of Cu(AzDC)(4,40-BPE)0.5
were also conrmed with gas adsorption. As shown in Fig. 3,
Cu(AzDC)(4,40-BPE)0.5 displays obviously dynamic switching
CO2 adsorption by irradiation with UV light (365 nm) and Vis
light (455 nm). With irradiation with UV light (365 nm), the
efficiency of the static CO2 adsorption of Cu(AzDC)(4,40-BPE)0.5
decreases by about 7.8% in comparison with irradiation with
Vis light (455 nm), which is less than the previous report with
42% variations of static CO2 adsorption.44 The azobenzene
groups in the pristine Cu(AzDC)(4,40-BPE)0.5 are in a predomi-
nantly trans conformation, a relatively stable form. Aer UV
irradiation, the azobenzene groups obtain energy and trigger
the trans–cis transformation, which will reduce the size of the
actual aperture to accommodate CO2 molecules. Just like
Fig. 5 shows the single gas permeance of H2, CO2, N2 and
CH4 through the Cu(AzDC)(4,40-BPE)0.5 membrane under Vis or
UV irradiation at 298 K and 1 bar as a function of the kinetic
diameter of the gas molecules. As shown in Fig. 5, the H2
Fig. 4 Top view (a) and cross-section (b) FESEM images of the
Cu(ABDC)(BPE)0.5 membranes prepared on the PDA-modified Al2O3
disk. (c) XRD patterns of the simulated Cu(AzDC)(4,40-BPE)0.5
,
Fig. 3 Effect of reversible in situ trans–cis photo-switching on the Cu(AzDC)(4,40-BPE)0.5 powder and Cu(AzDC)(4,40-BPE)0.5 membranes
adsorption isotherms of CO2, N2, and CH4 on the Cu(AzDC)(4,40- prepared on the PDA-modified Al2O3 disk. (A): Al2O3 support, (not
BPE)0.5 powder at 298 K. Irradiation time was 60 min.
marked): Cu(AzDC)(4,40-BPE)0.5
.
24952 | J. Mater. Chem. A, 2018, 6, 24949–24955
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
separation of H2/CO2 mixtures on other MOF membranes
(Table S3†), the Cu(AzDC)(4,40-BPE)0.5 membrane in the present
work also shows high separation selectivity and comparable
permeance. The Cu(AzDC)(4,40-BPE)0.5 membrane has been
tested for longer than 80 h at 298 K, and its separation perfor-
mance is found to remain unchanged (Fig. S14†), indicating
that the Cu(AzDC)(4,40-BPE)0.5 membrane has high stability
against UV irradiation. Further, since the preparation of the
Cu(AzDC)(4,40-BPE)0.5 membrane is controllable with PDA
modication of the support surface, the reproducibility of the
membrane preparation is high. For three independent
membrane preparations and tests, the Cu(AzDC)(4,40-BPE)0.5
membranes show similar separation performances (Table S4†),
with an average H2/CO2 selectivity of 43.53 ꢄ 0.54 (standard
deviation).
A reversible switching of the molecular sieving performance
of the Cu(AzDC)(4,40-BPE)0.5 membrane is tested for the sepa-
ration of the equimolar H2/CO2 mixture. As shown in Fig. 6,
although the CO2 permeance signicantly decreases in the case
of trans-to-cis isomerization, the H2 permeance is kept almost
unchanged, resulting in an observable enhancement of the H2/
CO2 separation factor from about 20.9 to 43.6 when switching
from the trans to the cis state. Further, aer irradiation with Vis
Fig.
5
Single gas permeance of the Cu(AzDC)(4,40-BPE)0.5
membranes irradiated in situ with Vis light (455 nm) and UV light
(365 nm) as a function of the kinetic diameter at 298 K and 1 bar.
Irradiation time was 60 min. The inset shows the influence of the
trans–cis switching on the mixture separation factor of the
Cu(AzDC)(4,40-BPE)0.5 membrane for H2 over other gases as deter-
mined by gas chromatography.
permeance of about 4.3 ꢀ 10ꢁ7 mol mꢁ2 sꢁ1 Paꢁ1 is much light (455 nm), the initial gas permeance and the H2/CO2
higher than those of the other gases due to its smallest pore size separation factor of 20.9 are found again because the azo-
of 0.29 nm. The ideal separation factors of H2 from CO2, N2 and benzene groups are switched back to the original trans state.
CH4 are 23.9, 15.4, and 13.9, which by far exceed the corre- The complete reversible switching of the membrane permeance
sponding Knudsen coefficients (4.7, 3.7 and 2.8), suggesting and selectivity are conrmed by three continuous cycles of
that the Cu(AzDC)(4,40-BPE)0.5 membrane is expected to display irradiation with UV (365 nm) and Vis light (455 nm), showing
H2-selectivity. As reported above, Cu(AzDC)(4,40-BPE)0.5 shows there is no sign of degradation, which is in good agreement with
strong adsorption for CO2 because of the strong quadrupolar previous reports of the photo-switchable MOF membranes.37,38
interactions of carbon with nitrogen atoms of the organic As reported previously, for a photo-switchable MOF membrane,
linkers. Therefore, the diffusion of CO2 molecules will be the separation factors (between membranes with minimum and
restricted much more than that of N2 and CH4, leading to a low maximum cis ratios) can be facilely controlled by ne-tuning the
CO2 permeance. With UV irradiation, the H2 permeance only
shows slight reduction, while the CO2 permeance decreases
remarkably, thus resulting in an outstanding enhancement of
the ideal separation factor of H2/CO2 from 23.9 to 48.2.
The molecular sieving performance of the Cu(AzDC)(4,40-
BPE)0.5 membrane was conrmed by the separation of equi-
molar mixtures at 298 K and 1 bar. As shown in Table S2,†
comparing the single gas permeance with the mixed gas per-
meance, the H2 permeance in the H2/CO2, H2/N2, and H2/CH4
mixtures shows only slight reduction with a H2 permeance of
about 4.2 ꢀ 10ꢁ7 mol mꢁ2 sꢁ1 Paꢁ1, indicating that the larger
molecules (CO2, N2, and CH4) only slightly hinder the perme-
ation of the highly mobile H2. As shown in the inset of Fig. 5, for
the 1 : 1 binary mixtures, the mixture separation factors of H2/
CO2, H2/N2, and H2/CH4 are 21.3, 18.6, and 13.8, respectively,
which also by far exceed the corresponding Knudsen coeffi-
cients. Similarly, for the separation of the 1 : 1 H2/CO2 mixture,
the H2 permeance only shows slight reduction with UV irradi-
ation, while the CO2 permeance decreases remarkably, thus
leading to a notable increase of the mixture separation factor of
H2/CO2 from 21.3 to 43.7 due to different interactions of the H2
Fig. 6 Mixed gas permeances and H2/CO2 mixed gas separation
factor of the Cu(AzDC)(4,40-BPE)0.5 membrane measured at 298 K and
1 bar. The Cu(AzDC)(4,40-BPE)0.5 membrane is irradiated with Vis light
(455 nm) and UV light (365 nm) for 30 min each for each data point in
and CO2 multipole moments with the cis- or trans-
42
Cu(AzDC)(4,40-BPE)0.5
.
Compared with literature data of the an in situ irradiation device as shown in Fig. S3.†
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A, 2018, 6, 24949–24955 | 24953

View Article Online
Journal of Materials Chemistry A
Paper
ratios of trans/cis azobenzene. Therefore, the H2/CO2 separation
factor is expected to switch not only between the minimum and
maximum values, but also each value in between realized by
choosing appropriate irradiation times, offering a potential and
promising road for remote control of membrane permeance/
selectivity by external stimuli.
5 F. Gallucci, E. Fernandez, P. Corengia and M. van Sint
Annaland, Chem. Eng. Sci., 2013, 92, 40–66.
6 S. Uemiya, T. Matsuda and E. Kikuchi, J. Membr. Sci., 1991,
56, 315–325.
7 M. Hong, S. Li, J. L. Falconer and R. D. Noble, J. Membr. Sci.,
2008, 307, 277–283.
8 J. L. Falconer, M. A. Carreon, S. Li and R. D. Noble, U.S. Pat.,
no. 8302782, 6 Nov 2012.
9 A. Huang and J. Caro, J. Mater. Chem., 2011, 21, 11424–11429.
10 M. Kanezashi, J. O'Brien-Abraham, Y. S. Lin and K. Suzuki,
AIChE J., 2008, 54, 1478–1486.
11 J. van den Bergh, A. Tihaya and F. Kapteijn, Microporous
Mesoporous Mater., 2010, 132, 137–147.
12 R. M. De Vos and H. Verweij, Science, 1998, 279, 1710–1711.
13 M. B. Shiett and H. C. Foley, Science, 1999, 285, 1902–1905.
14 R. Ranjan and M. Tsapatsis, Chem. Mater., 2009, 21, 4920–
4924.
15 A. Huang, H. Bux, F. Steinbach and J. Caro, Angew. Chem.,
2010, 122, 5078–5081.
16 X. Wu, C. Liu, J. Caro and A. Huang, J. Membr. Sci., 2018, 559,
1–7.
17 A. Huang, N. Wang, C. Kong and J. Caro, Angew. Chem., Int.
Ed., 2012, 51, 10551–10555.
18 A. Huang, Y. Chen, N. Wang, Z. Hu, J. Jiang and J. Caro,
Chem. Commun., 2012, 48, 10981–10983.
Conclusions
In conclusion, we have designed and prepared a novel smart
Cu(AzDC)(4,40-BPE)0.5 membrane. In the Cu(AzDC)(4,40-BPE)0.5
framework, the azobenzene and bis(4-pyridyl)ethylene moieties as
the “backbone” of the linker are incorporated directly into the
framework, forming an interpenetrating network structure with
˚
a small pore size of about 3.4 ꢀ 3.4 A. In particular, under irra-
diation with UV light, azobenzene and bis(4-pyridyl)ethylene
moieties can be switched from the trans to cis conguration,
resulting in squeezing out of the adsorbed CO2. Due to different
interactions of the H2 and CO2 multipole moments with cis- or
trans-Cu(AzDC)(4,40-BPE)0.5, the H2 permeance only shows slight
reduction with UV irradiation, while the CO2 permeance decreases
remarkably, thus leading to a notable enhancement of the mixture
separation factor of H2/CO2 from 21.3 to 43.7. Further, the
molecular sieving performance of the Cu(AzDC)(4,40-BPE)0.5
membrane can be facilely controlled by ne-tuning the ratios of
trans/cis azobenzene, offering a potential and promising road for
remote control of membrane permeance and selectivity by
external stimuli.
19 A. Huang, W. Dou and J. Caro, J. Am. Chem. Soc., 2010, 132,
15562–15564.
20 Q. Liu, N. Wang, J. Caro and A. Huang, J. Am. Chem. Soc.,
2013, 135, 17679–17682.
21 Y. S. Lin, Curr. Opin. Chem. Eng., 2015, 8, 21–28.
22 S. R. Venna and M. A. Carreon, Chem. Eng. Sci., 2015, 124,
3–19.
23 S. Qiu, M. Xue and G. Zhu, Chem. Soc. Rev., 2014, 43, 6116–
6140.
Conflicts of interest
There are no conicts to declare.
24 Z. Kang, L. Fan and D. Sun, J. Mater. Chem. A, 2017, 5, 10073–
10091.
25 X. Wu, W. Wei, J. Jiang, J. Caro and A. Huang, Angew. Chem.,
Int. Ed., 2018, 57, 15354–15358.
26 F.-X. Coudert, Chem. Mater., 2015, 27, 1905–1916.
27 B. Gui, Y. Meng, Y. Xie, K. Du, A. C. H. Sue and C. Wang,
Macromol. Rapid Commun., 2018, 39, 1700388.
28 A. Knebel, C. Zhou, A. Huang, J. Zhang, L. Kustov and J. Caro,
Chem. Eng. Technol., 2018, 41(2), 224–234.
29 J. Zhang, Q. Zou and H. Tian, Adv. Mater., 2013, 25, 378–399.
30 A. Knebel, B. Geppert, K. Volgmann, D. I. Kolokolov,
A. G. Stepanov, J. Twiefel, P. Heitjans, D. Volkmer and
J. Caro, Science, 2017, 358, 347–351.
Acknowledgements
Financial support by the National Natural Science Foundation
of China (21878100 and 21761132003), the Fundamental
Research Funds for the Central Universities (40500-20101-
222093), the K. C. Wong Education Foundation (rczx0800), and
the Key Research Program of Frontier Sciences of CAS (QYZDB-
SSW-JSC037) is acknowledged. Prof. Armin Feldhoff (Institut fur
Physikalische Chemie und Elektrochemie, Leibniz Universitat
Hannover, Germany) is thanked for TOPAS soware Pawley
analysis of XRD patterns using TOPAS soware and fruitful
discussions.
¨
¨
31 M. M. Russew and S. Hecht, Adv. Mater., 2010, 22, 3348–
3360.
References
1 S. Dunn, Int. J. Hydrogen Energy, 2002, 27, 235–264.
32 R. Klajn, Chem. Soc. Rev., 2014, 43, 148–184.
2 J. R. Rostrup-Nielsen and T. Rostrup-Nielsen, CATTECH, 33 D. Liu, Z. G. Ren, H. X. Li, J. P. Lang, N. Y. Li and
2002, 6, 150–159.
B. F. Abrahams, Angew. Chem., Int. Ed., 2010, 49, 4767–4770.
3 A. J. Brown, N. A. Brunelli, K. Eum, F. Rashidi, J. Johnson, 34 A. Modrow, D. Zargarani, R. Herges and N. Stock, Dalton
W. J. Koros, C. W. Jones and S. Nair, Science, 2014, 345, 72–
Trans., 2011, 40, 4217–4222.
75.
35 A. Modrow, D. Zargarani, R. Herges and N. Stock, Dalton
Trans., 2012, 41, 8690–8696.
4 N. W. Ockwig and T. M. Nenoff, Chem. Rev., 2007, 107, 4078–
4110.
24954 | J. Mater. Chem. A, 2018, 6, 24949–24955
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
36 F. Luo, C. B. Fan, M. B. Luo, X. L. Wu, Y. Zhu, S. Z. Pu, 41 K. Muller, A. Knebel, F. Zhao, D. Bleger, J. Caro and
W. Y. Xu and G. C. Guo, Angew. Chem., Int. Ed., 2014, 53,
9298–9301.
37 I. H. Park, A. Chanthapally, Z. Zhang, S. S. Lee,
L. Heinke, Chem.–Eur. J., 2017, 23, 5434–5438.
42 A. Knebel, L. Sundermann, A. Mohmeyer, I. Strauß, S. Friebe,
P. Behrens and J. Caro, Chem. Mater., 2017, 29, 3111–3117.
M. J. Zaworotko and J. J. Vittal, Angew. Chem., Int. Ed., 43 B. Chen, S. Ma, E. J. Hurtado, E. B. Lobkovsky and
2014, 53, 414–419. H.-C. Zhou, Inorg. Chem., 2007, 46, 8490–8492.
38 G. K. Kole and J. J. Vittal, Chem. Soc. Rev., 2013, 42, 1755– 44 R. Lyndon, K. Konstas, B. P. Ladewig, P. D. Southon,
1775.
P. C. J. Kepert and M. R. Hill, Angew. Chem., Int. Ed., 2013,
52, 3695–3698.
45 A. Huang, Q. Liu, N. Wang and J. Caro, J. Mater. Chem. A,
2014, 2, 8246–8251.
39 B. Garai, A. Mallick and R. Banerjee, Chem. Sci., 2016, 7,
2195–2200.
40 Z. Wang, A. Knebel, S. Grosjean, D. Wagner, S. Brase,
¨
¨
C. Woll, J. Caro and L. Heinke, Nat. Commun., 2016, 7, 13872.
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A, 2018, 6, 24949–24955 | 24955