Jiangnan Li1, Zhengyang Zhou2, Xue Han1, Xinran Zhang1, Yong Yan1, Weiyao Li1, Gemma L Smith1, Yongqiang Cheng3, Laura J McCormick McPherson4, Simon J Teat4, Mark D Frogley5, Svemir Rudić6, Anibal J Ramirez-Cuesta3, Alexander J Blake7, Junliang Sun2, Martin Schröder1, Sihai Yang1. 1. Department of Chemistry, University of Manchester, Manchester M13 9PL, U.K. 2. BNLMS, College of Chemistry & Molecular Engineering, Peking University, Beijing 100871, China. 3. Neutron Scattering Division, Neutron Sciences Directorate, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, United States. 4. Advanced Light Source, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States. 5. Diamond Light Source, Harwell Science Campus, Oxfordshire OX11 0DE, U.K. 6. ISIS Facility, STFC Rutherford Appleton Laboratory, Chilton, Oxfordshire OX11 0QX, U.K. 7. School of Chemistry, University of Nottingham, Nottingham NG7 2RD, U.K.
Abstract
Structural transitions of host systems in response to guest binding dominate many chemical processes. We report an unprecedented type of structural flexibility within a meta-rigid material, MFM-520, which exhibits a reversible periodic-to-aperiodic structural transition resulting from a drastic distortion of a [ZnO4N] node controlled by the specific host-guest interactions. The aperiodic crystal structure of MFM-520 has no three-dimensional (3D) lattice periodicity but shows translational symmetry in higher-dimensional (3 + 2)D space. We have directly visualized the aperiodic state which is induced by incommensurate modulation of the periodic framework of MFM-520·H2O upon dehydration to give MFM-520. Filling MFM-520 with CO2 and SO2 reveals that, while CO2 has a minimal structural influence, SO2 can further modulate the structure incommensurately. MFM-520 shows exceptional selectivity for SO2 under flue-gas desulfurization conditions, and the facile release of captured SO2 from MFM-520 enabled the conversion to valuable sulfonamide products. MFM-520 can thus be used as a highly efficient capture and delivery system for SO2.
Structural transitions of host systems in response to guest binding dominate many chemical processes. We report an unprecedented type of structural flexibility within a meta-rigid material, MFM-520, which exhibits a reversible periodic-to-aperiodic structural transition resulting from a drastic distortion of a [ZnO4N] node controlled by the specific host-guest interactions. The aperiodic crystal structure of MFM-520 has no three-dimensional (3D) lattice periodicity but shows translational symmetry in higher-dimensional (3 + 2)D space. We have directly visualized the aperiodic state which is induced by incommensurate modulation of the periodic framework of MFM-520·H2O upon dehydration to give MFM-520. Filling MFM-520 with CO2 and SO2 reveals that, while CO2 has a minimal structural influence, SO2 can further modulate the structure incommensurately. MFM-520 shows exceptional selectivity for SO2 under flue-gas desulfurization conditions, and the facile release of captured SO2 from MFM-520 enabled the conversion to valuable sulfonamide products. MFM-520 can thus be used as a highly efficient capture and delivery system for SO2.
Metal–organic
framework (MOF) materials are synthetic porous
materials constructed from bridging metal centers and organic ligands.[1] MOFs can be either rigid[2] or flexible[3] depending upon the metal–ligand
coordination combination, framework topology, and organic functionality.
Rigid MOFs retain a well-defined framework structure and pore interior
against external variables (e.g., temperature, pressure, and guest
inclusion), making them useful for the separation and purification
of substrates.[2,4] In contrast, flexible MOFs can
undergo reversible breathing transitions between typically two well-defined
periodic crystal structures. One form is often described as narrow-pored
and the other form is described as large-pored, and these two forms
therefore show different porosities and properties.[5−10]Figure illustrates
various structural responses to guest binding within a MOF lattice.
Breathing and metastable MOFs show two prevailing types of phase transitions,
namely, periodic-to-periodic and periodic-to-amorphous. Beyond that,
there is also an intriguing class of mesoporous MOFs, allowing the
formation of adsorbate superlattices based upon long-range collective
adsorbate–adsorbate interactions.[11] The periodic-to-aperiodic transition observed here for MFM-520 represents
a new type of framework flexibility in crystalline porous solids.
Superspace group theory[12] has been used
to describe the commensurate structure of crystalline zeolite SSZ-57[13] in higher-dimensional space, and this has been
applied to interpret the incommensurate modulated structures described
here.
Figure 1
Representation of structural transitions in various types of MOFs
upon guest inclusion and removal. MOFs and guest molecules are shown
by solid lines and shaded spheres, respectively. (a) Rigid MOF showing
full retention of the framework structure on gas adsorption and desorption.
(b) Flexible MOF showing reversible framework breathing between a
narrow-pore phase and a large-pore phase. (c) Metastable MOF showing
irreversible framework amorphization. (d) Mesoporous MOF showing the
formation of superlattices of adsorbates. (e) Meta-rigid MOF, MFM-520,
showing reversible incommensurate modulation of the framework structure
on going from (top to bottom) MFM-520·H2O (top) to
MFM-520 to MFM-520·CO2 (CO2 represented
by blue spheres) to MFM-520·SO2 (SO2 represented
by green spheres) (bottom).
Representation of structural transitions in various types of MOFs
upon guest inclusion and removal. MOFs and guest molecules are shown
by solid lines and shaded spheres, respectively. (a) Rigid MOF showing
full retention of the framework structure on gas adsorption and desorption.
(b) Flexible MOF showing reversible framework breathing between a
narrow-pore phase and a large-pore phase. (c) Metastable MOF showing
irreversible framework amorphization. (d) Mesoporous MOF showing the
formation of superlattices of adsorbates. (e) Meta-rigid MOF, MFM-520,
showing reversible incommensurate modulation of the framework structure
on going from (top to bottom) MFM-520·H2O (top) to
MFM-520 to MFM-520·CO2 (CO2 represented
by blue spheres) to MFM-520·SO2 (SO2 represented
by green spheres) (bottom).
Experimental Section
Synthesis of MFM-520
All reagents were used as received
from commercial suppliers without purification. The synthesis of H4L (4,4′-bipyridine-2,6,2′,6′-tetracarboxylic
acid) and {[Zn2(L)·4H2O]}∞ (MFM-520·H2O) was carried out using previously reported
methods.[14]
In Situ Synchrotron Single-Crystal X-ray Diffraction
and Structural Characterization
In situ synchrotron
X-ray single-crystal diffraction data were collected on beamline 11.3.1
of the Advanced Light Source using monochromated radiation [λ
= 0.7749(1) Å]. After data collection for as-synthesized sample
MFM-520·H2O, the same single crystal (0.10 ×
0.05 × 0.05 mm3) was placed in a capillary gas handling
cell and evacuated in situ under dynamic vacuum and
a stream of N2 at 340 K centered on the capillary for 4
h. The single-crystal diffraction data of dehydrated MFM-520 were
collected at 290 K. Upon loading of CO2 into MFM-520 at
1.0 bar and 290 K, three diffraction data sets were collected at 270
K to examine the possible dynamic behavior of the structural relaxation
(i.e., adsorbate–adsorbent equilibrium and change in the modulation
vector). The single crystal was then reactivated under vacuum and
heating to remove all adsorbed gases. The in situ single-crystal diffraction data for SO2-loaded MFM-520
were collected using the same method as for the CO2-loaded
sample.The influence of temperature on the periodic-to-aperiodic
phase transition was monitored by collecting in situ single-crystal diffraction data for dehydrated MFM-520 at 180–350
K on a Rigaku Oxford XtaLAB diffractometer using Mo-Kα radiation. No change in the modulation vectors of MFM-520
was observed between 180 and 350 K (Table S2), confirming the negligible impact of temperature on the modulation
vector of MFM-520.To examine the reversibility of the periodic-to-aperiodic
phase
transition observed in MFM-520, a single crystal of rehydrated MFM-520·H2O, formed by the treatment of dehydrated MFM-520 with water,
was studied by single-crystal diffraction at 180 K. This experiment
confirms the absence of satellite reflections of the rehydrated sample
(Figure S6).In the (3 + 2)D incommensurately
modulated structure, each parameter
of modulation can be expanded as a Fourier series using eq :where x4 = q1r μ + t, x5 = q2r μ + u, and t and u are initial phases in the range
of 0 to 1. The real structural parameters can be expressed as eq :The average structure parameters paveμ were determined
by direct methods and least-squares refinements, and the modulations
were refined globally using free variables. All non-hydrogen atoms
were refined anisotropically, and hydrogen atoms were placed using
a riding model. Given the complexity of the structure, soft constraints
on the bond length/angle to both host and guest were applied in the
final stage of refinement. All refinements of the crystal structures
were performed on the JANA2006 software platform.[15]
Gas Adsorption Isotherms
Gravimetric
sorption isotherms
for N2, CO2, SO2, CH4,
CO, and O2 were recorded at various temperatures on a Hiden
Isochema IGA-003 system under high vacuum (10−10 bar) produced by a turbo pumping system. Temperatures were maintained
using a programmed water bath. Ultra-pure research grade gases (99.999%)
were purchased from BOC or Air Liquide. In a typical gas adsorption
experiment, 50 mg of acetone-exchanged MFM-520 was loaded into the
IGA system and activated at 393 K under dynamic high vacuum (10–10 bar) for 1 day to give fully desolvated MFM-520.
BET surface areas and porosity data were obtained from N2 isotherm data at 77 K.
Gas Separation by Breakthrough Experiments
In the breakthrough
experiments, the flow rate of the entering gas mixture was maintained
at 50 mL min–1, and the gas concentration, C, of SO2, CO2, and N2 at
the outlet was determined by mass spectrometry and compared with the
corresponding inlet concentration C0,
where C/C0 = 1 indicates
complete breakthrough. Breakthrough separation of SO2/CO2 was conducted using a mixture containing SO2 (2500
ppm) and 15% CO2 (v/v) diluted in He through a fixed bed
packed with MFM-520 at 298 and 318 K and 1 bar at a flow rate of 50
mL min–1. The final results have been converted
to dimensionless plots as shown in Figure .
Figure 4
Gas adsorption,
selectivity, stability, thermodynamic, spectroscopy,
and dynamic separation data. (a) Adsorption isotherms for SO2, CO2, CH4, CO, O2, and N2 in MFM-520 at 298 K (solid symbols, adsorption; open symbols, desorption).
(b) Comparison of the SO2 adsorption capacity of MFM-520
over 75 cycles at 298 K. (c) Variation of isosteric heats of adsorption
(Qst) and adsorption entropy (ΔS) for SO2 and CO2 uptake in
MFM-520. (d) Comparison of experimental FTIR spectra for bare and
CO2- and SO2-loaded MFM-520. (e) Comparison
of DFT-calculated FTIR spectra for bare and CO2- and SO2-loaded MFM-520. (f) Comparison of the difference plots for
experimental and DFT-calculated INS spectra of D2O-, CO2-, and SO2-loaded MFM-520 (black, experiment; red,
calculation). The discrepancies, particularly for the lattice modes
at low energy (<200 cm–1), are due to the limitation
that DFT simulations can be conducted only on average unit cells,
which does not take the incommensurate modulation into consideration.
(g) Dimensionless breakthrough plots for SO2/N2 mixtures under dry (solid lines) and humid (dashed lines) conditions.
Dry conditions: 2500 ppm SO2, 99.75% N2, total
flow rate 50 mL min–1. Wet conditions: 1.5% H2O, 2500 ppm SO2, 98.25% N2, total flow
rate 50 mL min–1. (h) Dimensionless breakthrough
plots for SO2/CO2 mixtures at 298 K (15% CO2, 2500 ppm SO2, 84.75% He, total flow rate 50 mL
min–1). (i) Dimensionless breakthrough plots for
cyclic SO2 adsorption under both dry and humid conditions.
Dry conditions: 2500 ppm SO2, 99.75% He, total flow rate
50 mL min–1. Wet conditions: 1.5% H2O,
2500 ppm SO2, 98.25% He, total flow rate 50 mL min–1. For experiments under humid conditions, the bed
was first exposed to a flow of 1.5% H2O in He until the
breakthrough of water.
The fixed bed of MFM-520 for the
adsorptive removal of low concentrations of SO2 under wet
conditions was first exposed to a flow of 1.5% H2O in He
until the breakthrough of water. A stream of 0.25% SO2 (i.e.,
2500 ppm) diluted in He was then flowed through the packed bed of
MFM-520 (approximately 1 g) at a total flow rate of 50 mL min–1 at 298 K and 1.0 bar. Dimensionless breakthrough
plots were calculated with the following parameters: bed diameter, d, (7 mm); bed length, L, (120 mm); bed
volume (5.0 mL); sample mass (1.0 g); sample framework density (1.55
g/cm3); and flow rate (50 mL/min). The sample occupied
a volume of 0.65 mL (assuming 100% purity and no framework collapse),
and thus the fractional porosity of the fixed bed, ε, was calculated
to be 0.87. The superficial gas velocity, u, at the
entrance of the bed corresponded to 0.22 m/s, and the characteristic
contact time between the gas and the sample of MFM-520 corresponded
to εL/u = 4.82 s. The dimensionless
time, τ, was obtained by dividing the actual time, t, by the contact time between the gas and the MFM-520 sample, εL/u (i.e., τ = tu/εL).
Inelastic Neutron Scattering
(INS) Experiments
INS
spectra were recorded on the TOSCA spectrometer at the ISIS Facility
at the STFC Rutherford Appleton Laboratory. TOSCA is an indirect geometry
instrument that provides a wide spectral range (−25 to 4000
cm–1) with resolution optimized in the 50–2000
cm–1 range. In this region, TOSCA has a resolution
of 1.25% of the energy transfer. The instrument is composed of 130 3He detectors in forward and backscattering geometries located
17 m downstream from a 300 K Gd poisoned water moderator. A temperature
of 10 ± 0.2 K was maintained during data collection by two He
closed-cycle refrigerators with 30 mbar He as an exchange gas.A sample of MFM-520 was activated at 120 °C and 1 × 10–7 mbar for 1 day and then placed into an 11-mm-diameter
vanadium sample can, which was loaded into a helium closed-cycle refrigerator
(CCR) cryostat and cooled to 10 K for data collection. Defined amounts
of CO2 and SO2 were introduced by warming the
sample to 290 K, and the gas was dosed volumetrically from a calibrated
volume. The gas-loaded sample was then cooled to 10 K over a period
of 2 h to ensure good mobility of adsorbed gases within the crystalline
MFM-520. The sample was retained at 10 K for an additional 30 min
before data collection to ensure thermal equilibrium.
Density Functional
Theory (DFT) Calculations
DFT calculations
were performed using CASTEP (version 19.11).[16] The generalized gradient approximation (GGA) as implemented by Perdew–Burke–Ernzerhof
(PBE) was used to describe the exchange-correlation interactions.
On-the-fly norm-conserving pseudopotentials were employed to
account for the effects of core electrons with the default energy
cutoff (∼870 eV) for the plane-wave basis. The unit cell configuration
determined by XRD was used as the initial structure for the simulations.
The atomic coordinates of the guest molecules were first relaxed by
molecular dynamics at 300 K for 10 ps to find the optimal binding
sites that are not constrained by symmetry (the MOF structure was
fixed during this relaxation). The final configurations (including
both the MOF and the guest molecules) were further relaxed using the
conjugate gradient method to allow minimization of the potential energy
and the interatomic forces. The energy tolerance for the electronic
structure calculations was 5 × 10–10 eV, and
the energy tolerance for ionic relaxation was 5 × 10–9 eV. The tolerance for the interatomic forces was 1 meV/Å. After
convergence, the dynamic matrix was obtained using the linear response
method, from which the phonon frequencies and vibrational modes were
calculated. The electronic structure calculations were performed on
a 3 × 3 × 1 Monkhorst–Pack mesh, and the dynamic
matrix was calculated directly on the same mesh and then interpolated
on a 7 × 7 × 3 Monkhorst–Pack mesh. The IR intensities
were directly computed with CASTEP.[16] The
OClimax software[17] was used to convert
the DFT-calculated phonon results to the simulated INS spectra.
Conversion of Captured SO2
MFM-520 was activated
at 393 K under vacuum overnight. The sample was cooled, and MFM-520
was dosed with SO2 for 1 h to reach adsorption equilibrium.
Morpholin-4-amine (128.0 mg, 1.25 mmol) and SO2@MFM-520
(357 mg, equivalent to 1.18 mmol of SO2) were suspended
in CH3CN (3 mL) and stirred for 1 h. 4-Methoxy-aryldiazonium
tetrafluoroborate (55.0 mg, 0.25 mmol) in CH3CN (1 mL)
was added dropwise to the above suspension, and the mixture was stirred
at room temperature for 1 h. The mixture was filtered, and the filtrate
was evaporated. NMR spectroscopy and preparative thin layer chromatography
(TLC) were used to quantify the conversion of 4-methoxy-aryldiazonium
tetrafluoroborate and the yield of the sulfonamide. Benzenaldehyde
was used as an internal standard.
Results and Discussion
MFM-520·H2O, [Zn2(L)]·H2O (H4L= 4,4-bipyridyl-3,3′,5,5′-tetracarboxylic
acid), adopts a 3D open-framework structure comprising square-based
pyramidal five-coordinate Zn(II) centers bridged by the tetracarboxylate
linker[14] (Figure S1). Each Zn(II) center binds to the N(pyridyl)-donor of L4– along the c axis, and four bridging carboxylates
propagate the structure along the diagonals of the ab plane, thus defining an unusual and rigid sqp topology.[18] MFM-520·H2O exhibits bow-tie-shaped
pores with dimensions of 6.6 × 4.0 Å2 which are
filled with guest water molecules which assemble into 1D helical chains
along the [110] direction. These chains are stabilized by strong intermolecular
hydrogen bonds [Ow···Ow = 2.934(11),
2.887(11) Å] (Figure S2). The water
molecules are also hydrogen bonded to the framework oxygen center
of the ligand [Ow···O2 = 3.064(9) Å]
(Figure a-IV), inducing
the overall structure to crystallize in tetragonal space group P42212 [a = b = 7.0211(6), c = 19.9558(15) Å].
These complementary host–guest and guest–guest hydrogen
bonds stabilize the periodic structure of MFM-520·H2O and restrict its freedom for structural relaxation.
Figure 2
Column I: views of in situ synchrotron X-ray diffraction
images of the hk0 plane. Column II: views of the
corresponding crystal structures of MFM-520 as a function of guest
inclusion/removal. 8 × 8 × 2 means the unit cell of the
modulated structure is 8 times that of the average structure along
the a and b axes and 2 times along
the c axis; 7 × 7 × 2 means the unit cell
of the modulated structure is 7 times that of the average structure
along the a and b axes and 2 times
along the c axis. Column III: details of the structure
shown in the red box in column II. Column IV: molecular details of
host–guest interactions. The MOF framework, zinc, oxygen, carbon,
and sulfur are shown in blue, green, red, gray, and yellow, respectively.
Bond distances are shown in Å. (a) MFM-520·H2O, (b) MFM-520, (c) MFM-520·CO2, and (d) MFM-520·SO2. A comparison of all in situ synchrotron
X-ray diffraction images of the hk0.5 and hk1 planes is shown in Figure S5. The reflection conditions of MFM-520·H2O are 00l: l = 2 × integer and h00: h = 2 × integer. According to superspace
group theory,[12] there is a supercentered
setting with a = a, b = b, c = 2c, Q = α(a + b), and Q = α(−a + b), which is equivalent to the basic setting a, b, c, q = α(a + b) + 0.5c, and q = α(−a + b) + 0.5c. Because the
tables of (3 + 2)D superspace groups[19] provide
only reflection conditions with supercenterd settings, we transformed
the basic setting to supercentered settings to determine the reflection
conditions. With the supercentered setting, the hk0, hk0.5, and hk1 planes are transformed
to HK0, HK1, and HK2, respectively (Figure S5). The reflection
conditions of MFM-520, MFM-520·CO2, and MFM-520·SO2 are HKLMN: L + M + N = 2 × integer, HK0MN: M = 2 × integer, HHLM0: L = 2 × integer, HH̅L0N: N = 2 × integer, H0LMM̅: 2H + L + 2M = 4 × integer, and 0KLMM: 2K + L + 2M = 4 × integer, corresponding
to (3 + 2)D superspace group 136.2.69.10, adopting the symbol P42/mnm(α, α, 1/2)00qs(−α, α, 1/2)0sq0 with basic settings.
Column I: views of in situ synchrotron X-ray diffraction
images of the hk0 plane. Column II: views of the
corresponding crystal structures of MFM-520 as a function of guest
inclusion/removal. 8 × 8 × 2 means the unit cell of the
modulated structure is 8 times that of the average structure along
the a and b axes and 2 times along
the c axis; 7 × 7 × 2 means the unit cell
of the modulated structure is 7 times that of the average structure
along the a and b axes and 2 times
along the c axis. Column III: details of the structure
shown in the red box in column II. Column IV: molecular details of
host–guest interactions. The MOF framework, zinc, oxygen, carbon,
and sulfur are shown in blue, green, red, gray, and yellow, respectively.
Bond distances are shown in Å. (a) MFM-520·H2O, (b) MFM-520, (c) MFM-520·CO2, and (d) MFM-520·SO2. A comparison of all in situ synchrotron
X-ray diffraction images of the hk0.5 and hk1 planes is shown in Figure S5. The reflection conditions of MFM-520·H2O are 00l: l = 2 × integer and h00: h = 2 × integer. According to superspace
group theory,[12] there is a supercentered
setting with a = a, b = b, c = 2c, Q = α(a + b), and Q = α(−a + b), which is equivalent to the basic setting a, b, c, q = α(a + b) + 0.5c, and q = α(−a + b) + 0.5c. Because the
tables of (3 + 2)D superspace groups[19] provide
only reflection conditions with supercenterd settings, we transformed
the basic setting to supercentered settings to determine the reflection
conditions. With the supercentered setting, the hk0, hk0.5, and hk1 planes are transformed
to HK0, HK1, and HK2, respectively (Figure S5). The reflection
conditions of MFM-520, MFM-520·CO2, and MFM-520·SO2 are HKLMN: L + M + N = 2 × integer, HK0MN: M = 2 × integer, HHLM0: L = 2 × integer, HH̅L0N: N = 2 × integer, H0LMM̅: 2H + L + 2M = 4 × integer, and 0KLMM: 2K + L + 2M = 4 × integer, corresponding
to (3 + 2)D superspace group 136.2.69.10, adopting the symbol P42/mnm(α, α, 1/2)00qs(−α, α, 1/2)0sq0 with basic settings.On heating MFM-520·H2O at 340 K
under dynamic vacuum,
additional satellite reflections appear at positions defined by modulation
vectors 0.1240(3)(a + b) + 0.5c and 0.1240(3)(−a + b) + 0.5c on the plane diffraction patterns (Figure b-I), indicating
that an intriguing structural transition has occurred. The loss of
guest water molecules removes the stabilizing network of hydrogen
bonding and thus creates voids and freedom for the [ZnO4N] polyhedra to relax. Bond distances Zn–N1, Zn–O1,
and Zn–O2 in MFM-520·H2O [2.002(6), 2.214(6),
and 1.959(5) Å, respectively] change with the loss of water and
lie in ranges of 2.07–2.12, 2.25–2.58, and 1.79–2.11
Å, respectively, for dehydrated MFM-520. This is accompanied
by the modulation of the electron density at the Zn(II) center with
the formal oxidation state ranging from 1.70(10) to 1.93(10) in MFM-520
as determined by bond valence sum (BVS) calculations[20] for the different Zn(II) centers (Figure a). In contrast, those in MFM-520·H2O have the same charge of 1.974(14).
Figure 3
Maps of the bond valence
sum (BVS) analyses of the oxidation state
of Zn(II) centers as a function of the modulation vector u and t in (a) MFM-520, (b) MFM-520·CO2, and (c) MFM-520·SO2. The coordination geometries
of the [ZnO4N] polyhedra for the lowest and highest (highlighted
in blue and brown, respectively) BVS analysis maps for MFM-520, MFM-520·CO2, and MFM-520·SO2 are illustrated above the
contour maps. Bond distances are in Å. BVS parameters R(Zn–O), b0(Zn–O), R(Zn–N), and b0(Zn–N)
are 1.704, 0.37, 1.72, and 0.37 Å, respectively. MFM-520·H2O has a periodic structure and shows no modulation by BVS
analysis; the formal oxidation state of all Zn(II) centers in MFM-520·H2O is 1.974(14).
Maps of the bond valence
sum (BVS) analyses of the oxidation state
of Zn(II) centers as a function of the modulation vector u and t in (a) MFM-520, (b) MFM-520·CO2, and (c) MFM-520·SO2. The coordination geometries
of the [ZnO4N] polyhedra for the lowest and highest (highlighted
in blue and brown, respectively) BVS analysis maps for MFM-520, MFM-520·CO2, and MFM-520·SO2 are illustrated above the
contour maps. Bond distances are in Å. BVS parameters R(Zn–O), b0(Zn–O), R(Zn–N), and b0(Zn–N)
are 1.704, 0.37, 1.72, and 0.37 Å, respectively. MFM-520·H2O has a periodic structure and shows no modulation by BVS
analysis; the formal oxidation state of all Zn(II) centers in MFM-520·H2O is 1.974(14).The resulting structure
of dehydrated MFM-520 is aperiodic in 3D
space but periodic in (3 + 2)D space, and the two modulated vectors
(q1 and q2) are projections of the additional
two dimensions along the three directions of 3D reciprocal space.
Importantly, all satellite reflections of dehydrated MFM-520 disappear
on rehydration to MFM-520·H2O, demonstrating the reversibility
of this periodic-to-aperiodic phase transition (Figure S6). No change was observed in the modulation vectors
for dehydrated MFM-520 between 180 and 350 K (Table S2). Interestingly, for the periodic-to-aperiodic phase
transitions in inorganic materials, metal alloys, and organic compounds,
which are induced by changes in temperature and electric and magnetic
fields,[21−24] the periodic-to-aperiodic transition observed in MFM-520 is controlled
by host–guest interactions, representing the first example
of such a phenomenon for an inorganic–organic hybrid material
or porous solid.Dehydrated MFM-520 displays a BET surface area
of 313 m2 g–1 and a pore volume of 0.146
cm3 g–1,[14] and single-component
adsorption isotherms for SO2, CO2, CH4, CO, O2, and N2 in MFM-520 have been recorded
at 273–318 K up to 1.0 bar (Figure a, Figures S9–S12). The excess adsorption uptake of SO2 at 298 K and 1.0 bar is 3.38 mmol g–1,
which is higher than those of CO2, CH4, CO,
O2, and N2 (0.77, 0.71, 0.36, 0.19, and 0.18
mmol g–1, respectively). More significantly, at
low pressure (e.g., 0.01 bar, 298 K), where all other gases show negligible
uptakes (0.09 mmol g–1 for CO2; <0.01
mmol g–1 for CH4, CO, O2,
and N2), MFM-520 displays a notably higher isothermal uptake
of SO2 (1.66 mmol g–1), accompanied by
fast adsorption kinetics typically within minutes (Figures S13 and S14). The static SO2 uptake of
MFM-520 is comparable to those of the best-behaving sorbents but is
lower than the dynamic SO2 uptake of SIFSIX-2-Cu-i,[25] which possesses very narrow pores (Table S3). Often, there is a trade-off between
high adsorption at low pressure and difficulties in the complete regeneration
of narrow-pored MOFs. Facile regeneration of the SO2-saturated
MFM-520 can be readily achieved solely via a pressure swing, and no
loss of SO2 adsorption capacity or crystallinity of MFM-520
was observed after 75 cycles of SO2 adsorption–desorption
(Figure b).Gas adsorption,
selectivity, stability, thermodynamic, spectroscopy,
and dynamic separation data. (a) Adsorption isotherms for SO2, CO2, CH4, CO, O2, and N2 in MFM-520 at 298 K (solid symbols, adsorption; open symbols, desorption).
(b) Comparison of the SO2 adsorption capacity of MFM-520
over 75 cycles at 298 K. (c) Variation of isosteric heats of adsorption
(Qst) and adsorption entropy (ΔS) for SO2 and CO2 uptake in
MFM-520. (d) Comparison of experimental FTIR spectra for bare and
CO2- and SO2-loaded MFM-520. (e) Comparison
of DFT-calculated FTIR spectra for bare and CO2- and SO2-loaded MFM-520. (f) Comparison of the difference plots for
experimental and DFT-calculated INS spectra of D2O-, CO2-, and SO2-loaded MFM-520 (black, experiment; red,
calculation). The discrepancies, particularly for the lattice modes
at low energy (<200 cm–1), are due to the limitation
that DFT simulations can be conducted only on average unit cells,
which does not take the incommensurate modulation into consideration.
(g) Dimensionless breakthrough plots for SO2/N2 mixtures under dry (solid lines) and humid (dashed lines) conditions.
Dry conditions: 2500 ppm SO2, 99.75% N2, total
flow rate 50 mL min–1. Wet conditions: 1.5% H2O, 2500 ppm SO2, 98.25% N2, total flow
rate 50 mL min–1. (h) Dimensionless breakthrough
plots for SO2/CO2 mixtures at 298 K (15% CO2, 2500 ppm SO2, 84.75% He, total flow rate 50 mL
min–1). (i) Dimensionless breakthrough plots for
cyclic SO2 adsorption under both dry and humid conditions.
Dry conditions: 2500 ppm SO2, 99.75% He, total flow rate
50 mL min–1. Wet conditions: 1.5% H2O,
2500 ppm SO2, 98.25% He, total flow rate 50 mL min–1. For experiments under humid conditions, the bed
was first exposed to a flow of 1.5% H2O in He until the
breakthrough of water.Further analysis of the
adsorption thermodynamics of MFM-520 reveals
that the isosteric heat of adsorption (Qst) for CO2 is 32–48 kJ mol–1 and
the entropy of adsorption (ΔS) decreases steadily
with increasing surface coverage, consistent with the expected ordering
of adsorbed CO2 molecules in MFM-520. In comparison, Qst for SO2 adsorption reaches ∼120
kJ mol–1 at low surface coverage and decreases gradually
to 60 kJ mol–1 with increasing loading of SO2. The calculated value of Qst is
consistent with the calorimetric measurements of 43 and 96 kJ mol–1 for CO2 and SO2 adsorption,
respectively (Figures S15 and S16). The
higher value of Qst for SO2 uptake is consistent with the expected stronger interaction between
the framework and SO2 compared with CO2. However,
the ΔS for SO2 adsorption increases
steadily from an initial −260 to −135 J K–1 mol–1 with increasing loading of SO2. This observation is opposite to the prevailing understanding of
gas adsorption in porous materials, where ordering of the gas molecules
results in a decrease in ΔS on gas loading[26−28] and implies increasing disorder of the system on SO2 loading
in MFM-520.We sought to understand the origin of this intriguing
pattern in ΔS and have determined the host–guest
structures
of CO2- and SO2-loaded MFM-520 by in
situ synchrotron X-ray single-crystal diffraction (Table S1). On loading CO2 into MFM-520
at 270 K and 1.0 bar, adsorbed CO2 molecules are clearly
observed in the pores, corresponding to a crystallographic uptake
of 2.18 mmol g–1, in excellent agreement with the
isotherm uptake (2.14 mmol g–1). All guest CO2 molecules in MFM-520·CO2 were found to be
incommensurate (Figure c). Thus, instead of showing a constant distance between adjacent
CO2 molecules, the intermolecular distances range from ca. 6.6 to 7.4 Å (Figure S3), indicating no apparent interaction between adjacent CO2 molecules. The distances between CO2 and phenyl rings
(C1···pyridyl ring) and between CO2 and
aromatic −CH groups (O3···H1C2–C2) are
longer than 3.0 Å (Figure c-IV), suggesting only very weak supramolecular interactions
between the host framework and CO2. Therefore, MFM-520
experiences no change in structure upon CO2 adsorption,
and both the oxidation state of Zn(II) centers [ranging from 1.73(9)
to 2.07(9)] and the geometry of [ZnO4N] nodes in MFM-520·CO2 [Zn–N1, Zn–O1, and Zn–O2 bond distances
ranging between approximately 2.00 and 2.07, 2.20 and 2.50, and 1.72
and 2.25 Å, respectively] are similar to those of bare MFM-520
(Figure b, Figures S7 and S8). Thus, the two modulation
vectors [0.1243(3)(a + b) + 0.5c and 0.1243(3)(−a + b) + 0.5c] in MFM-520 remain upon adsorption of CO2.Upon loading of SO2 into MFM-520 at 270
K and 1.0 bar,
a crystallographic uptake of SO2 of 3.16 mmol g–1 is observed, in excellent agreement with the isotherm uptake (3.38
mmol g–1). Since the occupancy of adsorbed SO2 molecules is 0.359(13) in the refined structural model of
MFM-520·SO2, half of the closely contacted SO2 molecules (ca. 1.7 to 1.9 Å) were removed,
and the remaining SO2 molecules have an intermolecular
distance of ca. 3.5 Å (Figure
S4), which is shorter than that in MFM-520·CO2, suggesting stronger guest–guest interactions between adsorbed
SO2 molecules.On the other hand, strong MOF–SO2 binding is
indicated by the short distances between SO2 and aromatic
−CH groups (O3···H1C2–C2: ca. 2.33 to 2.75 Å) and between SO2 and [ZnO4N] polyhedra (S1···O2: ca. 2.78 to
3.34 Å) (Figure d-IV, Figures S7 and S8). As a result,
the structure of MFM-520 is heavily affected by adsorbed SO2 molecules. The ranges of bond distances Zn–N1, Zn–O1,
and Zn–O2 in MFM-520·SO2 vary from 1.97 to
2.14, 1.92 to 2.73, and 1.75 to 2.19 Å, respectively, much wider
than those observed in MFM-520 and MFM-520·CO2. The
oxidation state of the Zn(II) centers also shows a greater range between
1.79(9) and 2.29(9) (Figure c) compared with that in MFM-520 and MFM-520·CO2. In MFM-520·SO2, the modulation vectors become 0.1407(6)(a + b) + 0.5c and 0.1407(6)(−a + b) + 0.5c, demonstrating
a unique incommensurately ordered dynamic equilibrium for SO2 adsorption. Interestingly, before reaching this equilibrium state,
we also identified a transient structure of MFM-520·SO2 with a corresponding crystallographic SO2 uptake of 0.532
mmol g–1 that has intermediate modulation vectors
of 0.1331(4)(a + b) + 0.5c and 0.1331(4)(−a + b) + 0.5c (Table S1). The change
in the incommensurate modulation of MFM-520 induced by the uptake
of SO2 is a dynamic process as a function of diffusion
and the settling of SO2 molecules within the pores of the
MOF. This result rationalizes the unusual increase in ΔS on SO2 loading and confirms that it emanates from a subtle
increase in the structural flexibility of meta-rigid MFM-520 upon
SO2 adsorption.Overall, the incommensurate modulated
structure reported here will
stimulate a reassessment of a wider range of host–guest systems
with unusual adsorption behavior[29,30] and encourage
the search for new structural flexibility in other MOFs and porous
solids that otherwise may be regarded as entirely rigid.The
binding dynamics and effects on the structural relaxation of
MFM-520 have been analyzed further using in situ synchrotron
infrared microspectroscopy, inelastic neutron scattering (INS), and
DFT modeling (Figure d–f and Figures S19–S30).
Dehydrated MFM-520 has a number of characteristic peaks at 3104, 1080,
1689, 1571, and 1390 cm–1 (denoted as I, II, III,
IV, and V), which are assigned based upon DFT modeling to the stretching
(I) and bending (II) modes of the aromatic C–H groups on the
pyridyl ring and the symmetric stretching (III) and asymmetric stretching
(IV) modes of C–O and the bending (V) mode of C–H groups,
respectively. On dosing CO2 (0–1 bar), peak I shows
a blue shift (Δ = 9 cm–1) to 3113 cm–1, indicating a stiffening of the framework as a result of the binding
of CO2 molecules to −CH groups. No notable changes
occur in peaks II–V on CO2 loading, consistent with
the very weak MOF–CO2 interactions. Upon loading
SO2, a larger blue shift of 12 cm–1 is
observed for peak I, accompanied by near-depletion of this band on
increasing SO2 loading, demonstrating a stronger −CH···O=S=O
supramolecular interaction. Furthermore, distinct red shifts of peaks
III and IV (C–O, carboxyl) to 1682 and 1566 cm–1 (Δ = 6 and 5 cm–1, respectively) and peak
V (C–H, pyridyl ring) to 1385 cm–1 (Δ
= 5 cm–1) are observed, entirely consistent with
the presence of structural distortion as observed in crystallographic
studies and DFT-calculated IR spectra (Figure e). Difference INS spectra recorded (Figure f) between bare and
CO2-, SO2-, and D2O-loaded MFM-520
reveal shifts of peaks associated with deformation and twisting modes
of the pyridyl rings at low frequencies (200–300 cm–1) as well as C–H bending modes at higher frequencies (900–1300
cm–1), as confirmed by DFT modeling. SO2 and D2O both display notably stronger interactions with
MFM-520 than CO2, again consistent with the crystallographic
studies.Unlike conventional once-through FGD (flue gas desulfurization)
methods where SO2 permanently binds to sorbent materials
to form solid inorganic waste, the SO2 captured by MFM-520
remains available to undergo chemical transformation to valuable products.[31,32] A proof-of-concept experiment was thus conducted by coupling 4-methoxyl-aryldiazonium
tetrafluoroborates and hydrazine in a solution containing a suspension
of MFM-520·SO2 to synthesize aryl N-aminosulfonamides (Scheme ).[34] A nearly quantitative
conversion of 4-methoxyl-aryldiazonium tetrafluoroborates and an 80%
yield of the sulfonamide were achieved. Immobilization and then release
of SO2 using a stable MOF can thus be used to drive organic
reactions effectively, removing the high risk of the large-scale use
of caustic and toxic SO2. In line with the observed structural
stability of MFM-520, we have confirmed its reusability for three
cycles to produce sulfonamide (Table S6). Thus, MFM-520 represents a stable and efficient platform for the
capture, storage, and conversion of SO2.
Scheme 1
Conversion of Captured
SO2 in MFM-520
Conclusions
The metal–organic framework material MFM-520 exhibits a
reversible periodic-to-aperiodic structural transition. The aperiodic
crystal structure of MFM-520 has no three-dimensional (3D) lattice
periodicity but shows translational symmetry in higher-dimensional
(3 + 2)D space. Filling of MFM-520 with CO2 and SO2 reveals that while CO2 has a minimal structural
influence, SO2 can further modulate the structure incommensurately.
In addition, MFM-520 shows high selectivity for SO2 under
flue gas desulfurization conditions, and the facile release of captured
SO2 from MFM-520 can be used as an SO2 delivery
system for subsequent reactivity.
Authors: Xiang Lin; Alexander J Blake; Claire Wilson; Xue Zhong Sun; Neil R Champness; Michael W George; Peter Hubberstey; Robert Mokaya; Martin Schröder Journal: J Am Chem Soc Date: 2006-08-23 Impact factor: 15.419
Authors: Sihai Yang; Leifeng Liu; Junliang Sun; K Mark Thomas; Andrew J Davies; Michael W George; Alexander J Blake; Adrian H Hill; Andrew N Fitch; Chiu C Tang; Martin Schröder Journal: J Am Chem Soc Date: 2013-03-25 Impact factor: 15.419
Authors: Andrew L Goodwin; Mark Calleja; Michael J Conterio; Martin T Dove; John S O Evans; David A Keen; Lars Peters; Matthew G Tucker Journal: Science Date: 2008-02-08 Impact factor: 47.728
Authors: Grigorii Skorupskii; Khoa N Le; Dmitri Leo Mesoza Cordova; Luming Yang; Tianyang Chen; Christopher H Hendon; Maxx Q Arguilla; Mircea Dincă Journal: Proc Natl Acad Sci U S A Date: 2022-08-15 Impact factor: 12.779
Authors: Jiangnan Li; Gemma L Smith; Yinlin Chen; Yujie Ma; Meredydd Kippax-Jones; Mengtian Fan; Wanpeng Lu; Mark D Frogley; Gianfelice Cinque; Sarah J Day; Stephen P Thompson; Yongqiang Cheng; Luke L Daemen; Anibal J Ramirez-Cuesta; Martin Schröder; Sihai Yang Journal: Angew Chem Int Ed Engl Date: 2022-07-25 Impact factor: 16.823