ARTICLES
Mechanical gas capture and release in a
network solid via multiple single-crystalline
transformations
BRETT D. CHANDLER1 , GARY D. ENRIGHT2 , KONSTANTIN A. UDACHIN2 , SHANE PAWSEY2 ,
JOHN A. RIPMEESTER2 , DAVID T. CRAMB1 * AND GEORGE K. H. SHIMIZU1 *
1
Department of Chemistry, University of Calgary, Calgary, Alberta, T2N 1N4, Canada
Steacie Institute for Molecular Sciences, National Research Council of Canada, Ottawa Ontario, K1A 0R6, Canada
* e-mail: [email protected]; [email protected]
2
Published online: 20 January 2008; doi:10.1038/nmat2101
Metal–organic frameworks have demonstrated functionality stemming from both robustness and pliancy and as such, offer promise
for a broad range of new materials. The flexible aspect of some of these solids is intriguing for so-called ‘smart’ materials in
that they could structurally respond to an external stimulus. Herein, we present an open-channel metal–organic framework that,
on dehydration, shifts structure to form closed pores in the solid. This occurs through multiple single-crystal-to-single-crystal
transformations such that snapshots of the mechanism of solid-state conversion can be obtained. Notably, the gas composing the
atmosphere during dehydration becomes trapped in the closed pores. On rehydration, the pores open to release the trapped gas.
Thus, this new material represents a thermally robust and porous material that is also capable of dynamically capturing and releasing
gas in a controlled manner.
Metal–organic frameworks (MOFs)1–3 represent a class of solids
that, in numerous ways beyond their composition, are truly
hybrid materials. Despite relying on weaker bonds than metal
oxides, coordination networks have demonstrated permanent
microporosity4–11 . Such networks can efficiently physisorb various
gases and are promising materials for storage applications12–16 .
Other coordination solids have shown ‘smart’ behaviour in that
they can convert between different structures via some external
stimulus (for example, heat, guest molecules)17–33 , a property
not typically observed with metal oxides. In this light, we can
consider certain fundamental properties of porous networks and
their respective interdependencies. For example, robustness and
crystallinity are both desirable but, at the same time, are not
necessarily easily reconciled. To compensate for the energetic
penalty of forming a pore, a network solid must be robust
and sustained by sufficiently strong interactions. Often, solids
assembled by such strong interactions precipitate rapidly and
lack long-range order. Thus, stability, and hence porosity, can be
viewed as competing features with crystallinity34 . Further to this
chain of thought, dynamic motion in a network solid should be
fundamentally opposed to each of these three features. Structural
shifting should disfavour stability, and therefore porosity, and
unquestionably, crystallinity should be diminished by movement
of components in the solid35 . Solids that have been observed to
reconcile dynamics and crystallinity typically have components
sustained by non-covalent interactions, at least partially36–42 or
completely43–47 , as the energetic minima in the conversion profile
would be more shallow, although some exceptions do exist48 .
Barium 1,3,5-benzenetrisulphonate, Ba3 L2 , is a network solid
that reconciles all of the above features as, via multiple crystalline
transformations, it converts from open- to closed-pore solid to
nature materials VOL 7 MARCH 2008 www.nature.com/naturematerials
mechanically trap atmospheric gas molecules. This dehydrated
phase is thermally stable to in excess of 550 ◦ C, but it can be
rehydrated to open the pores and allow controlled release of
trapped gaseous guest molecules. The use of the Ba2+ ion and
sulphonate groups in this work is critical as the interaction between
them is of a diffuse ionic nature. Moreover, neither the metal ions
nor the ligands have strong geometrical coordinative preferences
and so the interactions sustaining the network are less directional.
This results in a network material that merges the stability and
porosity of a strongly bonded, rigid solid with the ability of a
weakly bonded solid to flex and optimize its structure. Ultimately, a
thermally robust, crystalline material with controllable gas capture
and release properties is obtained.
Benzene-1,3,5-trisulphonic acid, H3 L, was prepared by
sulphonating sodium benzene-1,3-disulphonate with oleum and
HgSO4 and then metathesized to the tetrabutylammonium
(TBA) salt using TBA(HSO4 ). Crystals of Ba3 (H2 O)3 (L)2 (H2 O)8 ,
compound 1, were grown by diffusion of methanol into an
aqueous solution of (TBA)3 L and BaCl2 ·2H2 O. Owing to their
rapid desolvation, crystals were mounted under a cold stream of
blowoff from liquid N2 . The structure of compound 1, solved
as orthorhombic, Pccn with a = 21.603(2), b = 9.5042(8),
c = 17.932(2) Å, shows hydrated pores down the b and c
axes. Branching out from a single ligand, the structure can
be described as two molecules of L forming π-stacked dimers
(C ··· C = 3.66(1)–3.70(1) Å) that sit in a staggered conformation
offset by 1.38(1) Å. Crosslinking of these dimers with Ba1 results
in the formation of two-dimensional (2D) sheets in the bc plane
(Fig. 1a). Each Ba1 centre is 8-coordinate with sulphonate oxygen
atoms occupying six positions and water the remaining two sites.
Extension in the third dimension, along the a axis, is by the
229
ARTICLES
a
b
c
d
e
f
g
h
Compound 1
a = 21.603 Å, α = 90.0°
b = 9.509 Å, β = 90.0°
c = 17.932 Å, γ = 90.0°
wt% H2O: 17.2%
ρcalc: 2.220 g cm –3
Compound 2a
a = 14.586 Å, α = 90.0°
b = 17.713 Å, β = 96.42°
c = 12.022 Å, γ = 90.0°
wt% H2O: 12.3%
ρcalc: 2.519 g cm –3
Compound 3
a = 9.461 Å, α = 90.0°
b = 17.493 Å, β = 90.0°
c = 15.453 Å, γ = 90.0°
wt% H2O: 3.3%
ρcalc: 2.791 g cm –3
C
Compound 4
a = 9.607 Å, α = 90.0°
b = 9.607 Å, β = 90.0°
c = 16.354 Å, γ = 120.0°
wt% H2O: 0%
ρcalc: 2.689 g cm –3
O
S
Ba
H
Figure 1 Single-crystal X-ray structures of fully hydrated 1, intermediates 2a and 3, and dehydrated 4. a–g, Views of 1 (a), 2a (c), 3 (e) and 4 (g) perpendicular to the
plane of the aryl rings, and 1 (b), 2a (d), 3 (f) and 4 (h) in the plane of the aryl rings showing shifting, contraction and re-expansion of pores (uncoordinated water removed).
coordination of four sulphonate oxygen atoms to the Ba2 centres,
also 8-coordinate with four coordinating water molecules. This
bridging of bc layers along the a axis results in the formation of
pores, visible down the b (Fig. 1b) and c axes, filled by disordered
water molecules. The channel width ranges from 7.60(1)–7.81(1) Å.
The crystal is 17.2% water by mass and has a calculated density
(ρcalc ) of 2.201 g cm−3 .
Facile structural conversion of compound 1 occurs with
dehydration. As water evacuates, the structure responds to two
230
driving factors: (1) the network contracts to fill the void space
created by desolvation; (2) the Ba centres become unsaturated
on dehydration and so the structure flexes to enable further
coordination of sulphonate oxygen atoms. By heating a crystal of
compound 1, it was possible to form the fully dehydrated Ba3 L2
network, 4, in a single-crystal-to-single-crystal manner while also
detecting two intermediate phases, 2 and 3. Unit-cell parameters
are given in Fig. 1. It should be noted that diffraction patterns
yielding intermediate, but statistically poor, unit cells can be
nature materials VOL 7 MARCH 2008 www.nature.com/naturematerials
ARTICLES
a
b
c
d
e
f
Figure 2 X-ray diffraction images of a single crystal of 1 undergoing heating while on the diffractometer. a, 192 K. b, 253 K. c, 263 K. d, 283 K. e, 333 K. f, 393 K. The
image in a represents 1 and that in f represents 4 (see the Supplementary Information). Images in b–d show spots for both 2 and 3. The image in e represents a stage of
rapid heating on the way to conversion to 4. The vertical lines observed in a, b, c and e represent saturation of the detector.
obtained at any point during the dehydration. This is corroborative
of the continuum of intermolecular interactions enabling this
transformation (see below). Thus, the snapshots provided by 2/2a
and 3 are not the only transition points between 1 and 4, but simply
the only ones that could be characterized. Although this presents a
less precise picture of the mechanism, it is, in fact, a more accurate
representation of the dynamic behaviour of this solid.
Figure 2 shows the X-ray diffraction patterns obtained by
heating a crystal of 1 from −100 ◦ C on the CCD (charge-coupled
device) diffractometer from orthorhombic 1 through to trigonal
4. The structural transformations occurring in the dehydration
of 1 reached pseudo-plateaux at compounds 2 and 3. At these
points, it was possible to freeze the sample and collect entire
crystallographic data sets. For 2, the quality of the data was poor as
the crystal did not dehydrate uniformly. This resulted in domains
within the crystal being at different stages of transformation and
considerable structural disorder. That said, the data set of 2
enabled correlation of unit-cell parameters (monoclinic, P 21 /c ,
a = 14.645(5), b = 17.756(5), c = 12.144(5) Å, β = 96.42(1)◦ ) to
the degree of hydration. By diffusing acetone rather than methanol
in the preparation of 1, good single crystals of a Ba3 L2 phase,
with unit-cell parameters and hydration levels equivalent to 2, were
obtained, henceforth referred to as 2a. Calculated spot overlays for
Fig. 2a,f with 1 and 4 are provided as Supplementary Information.
On loss of ∼5% of the water from the structure of 1 (noncoordinated water) to give 2/2a, two major structural changes
occur. The Ba2+ ions, which had been aligned into ribbons in
1, begin translation towards a hexagonal pattern about the L
molecules (Fig. 1c). This is accompanied by a redistribution of the
pores in the network. The pores between aryl groups contract,
forming π-stacked tetramers of L (C ··· C = 3.667(3)–3.935(3) Å)
as opposed to dimers in 1. These tetramers are then separated
from each other by 6.69(3)–6.73(3) Å (Fig. 1d). The columns of
tetramers are crosslinked, through sulphonate groups by Ba2+ , into
nature materials VOL 7 MARCH 2008 www.nature.com/naturematerials
a network of 1D pores, as opposed to 2D pores in 1. The net effect
is still a denser phase (ρcalc = 2.519 g cm−3 ). Notably, all Ba centres
retain coordinated water, enabling the coordinated RSO−3 groups to
arrange themselves about the metal in a manner that permits the
aryl groups to stack and obtain an energetic minimum. Ba1 and
Ba3 are both 7-coordinate with two coordinated water molecules
and five sulphonate oxygen atoms. Ba2 is 7-coordinate with four
coordinated water molecules and three sulphonate oxygen atoms.
The loss of the next 9–10% of water from the structure is
accompanied by the gradual progression of the structure towards
a hexagonal pattern as shown by the structure of compound 3
(∼3% water, crystallized as orthorhombic, Pccn with a = 9.461(2),
b = 17.943(4), c = 15.453(4) Å) (Fig. 1e). The most significant
change from 2 to 3, however, occurs perpendicular to the plane
of the L molecules as shown in Fig. 1f. Rather than an alternating
structure of stacked L tetramers and ∼6.7 Å pores observed in
2a, compound 3 shows no prominent pores. Rather, a further
contracted structure is observed (ρcalc = 2.791 g cm−3 ) with two sets
of aryl–aryl distances (3.68(1)–3.74(1) Å and 3.99(1)–4.11(1) Å).
Two two-fold disordered pairs of Ba2+ ions exist in 3: Ba1/Ba2
is 6-coordinate with five ligated sulphonate oxygen atoms and
one water molecule; Ba3/Ba4 is 7-coordinate with homoleptic
sulphonate coordination.
Compound 4, the fully dehydrated phase with
trigonal symmetry, crystallized in P 3c with a = 9.607(2),
b = c = 16.354(3) Å. This is demonstrated by the perfect alignment
of the Ba2+ ions into hexagonal arrays about L molecules. The
Ba centres in the hexagonal pattern are bridged into columns
by the perpendicular crosslinking of SO3 groups of L, (Fig. 1g).
Compound 4 contains two crystallographically unique Ba centres,
each 8-coordinate with homoleptic sulphonate coordination. With
no water coordinated to the Ba ions, the ligated sulphonate
groups distribute as spherically as possible about the sizable Ba
ions (Ba2+ rvdW = 2.14 Å); this requires the aryl groups of L to
231
ARTICLES
a
b
Figure 3 Single crystals of 4. a, On immediate immersion in Paratone oil.
b, Bubbles of trapped gas being released owing to rehydration of 4. 1 H NMR with an
internal standard showed the water content of the Paratone to be <200 p.p.m.
Small bubbles are observed after 5 min, which, owing to the viscosity of the oil and
surface tension, linger on the crystal’s surface until they merge into larger bubbles.
The image in b is after standing for 15 min.
adopt an inefficient packing mode. Two sets of inter-aryl C–C
distances are observed in 4; molecules of L form cofacial dimers
(3.86(1)–3.96(1) Å and 4.09(1)–4.19(1) Å) generating small but
not insignificant voids. Therefore, a key point with respect to the
transformation is that the pores between aryl rings initially contract
on loss of channel water but then, at the lowest degrees of hydration,
they re-expand to enable sulphonate ligation to Ba2+ and achieve
the minimum energy structure. This is corroborated by the fact
that compound 4 (ρcalc = 2.689 g cm−3 ) is actually less dense than
3. This forms the closed pores defined by the aryl rings and the Ba
ions. With the size of Ba2+ , six cations are able to bridge sulphonate
groups and effectively form a continuous belt about two cofacial L
molecules. The L molecules and six Ba ions then frame a cavity in
which gases may be trapped.
At this stage, it is prudent to discuss the nature of
intermolecular interactions in this material and how they contrast
those in other (dynamic) MOF solids. Generally, structural
transformations in MOFs could be classified as arising from
either reorientations of macroscopic components (for example,
shifting of layers) or as due to changes to the metal ion’s
coordination sphere. Most MOFs, to maintain pores, rely on
strong and directional metal–ligand bonding with metal ions of
regular coordinative preferences. Thus, any changes in coordination
resulting from a macroscopic structural shift are clearly defined
and bond cleavage and formation is unambiguous. An excellent
illustration is the recent paper by Rosseinsky and co-workers18 . In
this work, a network incorporates an aquated Co() centre that
H-bonds to an included, non-ligated 4,40 -bipyridine molecule. On
heating, the primary-sphere water molecule is lost and replaced
by the formerly second-sphere bipyridine molecule. Surprisingly,
on rehydration, the bipyridine molecule is replaced by water
and resumes its second-sphere role. Here, we have specifically
chosen the Ba2+ /organosulphonate couple owing to the diffuse
ionic nature of the intermolecular interactions and the adaptable
coordination preferences of both metal and ligand components.
This contrast can be illustrated by considering that, in 2, Ba2 is
ligated, including ligands closer than 2.813(3) Å, by three water
molecules and four sulphonate oxygen atoms. However, three
further sulphonate donors lie from 2.884(3)–3.063(3) Å away from
the Ba centre and four others from 4.467(3)–4.701(3) Å. Factoring
the irregular geometry of Ba2+ (ranges from 7–9 coordinate
typically) with the spherical ligation offered by RSO−3 groups (a
180◦ rotation translates an O atom ∼2.9 Å), loss of coordinated
water would enable even slight macroscopic shifts to maintain
oxygen ligation to the metal ion. Thus, rather than an ‘all or
232
300
250
200
150
100
(p.p.m.)
50
0
–50
–100
Figure 4 Hyperpolarized 129 Xe NMR of a sample of 1 on drying. Top trace: After
24 h at 25 ◦C. Middle trace: After 48 h at 25 ◦C. Bottom trace: On further heating at
150 ◦C for 5 h. The peak at 0 p.p.m. is free Xe.
nothing’ ligand-bonding scenario, for which the structural shifts
require a specific orientation of components, the dynamics in this
material are based more on the availability of a continuum of donor
atoms. This illustration is meant to both affirm the exceptional
behaviour of systems such as the Rosseinsky example above (as
well as others cited) but also emphasize that, for networks from
adaptable components, dynamic crystalline behaviour should be
viewed more as the rule and less the exception34 . Thus, 2/2a
and 3 represent two of numerous stages in a continual structural
transformation between 1 and 4. Powder X-ray diffraction of 1 (see
Supplementary Information, Fig. S13) shows conversion to other
crystalline phases on dehydration, still including compounds 2 and
3, and finally to 4 with heating.
Compound 4 is tremendously thermally robust. Differential
scanning calorimetry/thermogravimetric analysis revealed, after
dehydration to 225 ◦ C, a stable plateau to 574 ◦ C (see the
Supplementary Information). Decomposition at this point was
abrupt (∼30% loss in 3–5 ◦ C) and inconsistent with an equilibrium
process. The temperature of the mass loss varied with the particular
gaseous atmosphere. For example, heating 1 in Ar(g) gave
decomposition at 567 ◦ C. These observations infer confinement
of a finite and different number of gas molecules within each
closed pore with different gases. This results in pressure differences
ultimately manifested as variations in decomposition temperature.
During the selection of single crystals of 4, as the crystals were
immersed in Paratone oil, bubbles of gas emerged from the solid
(Fig. 3). Repeating this with various solvent combinations revealed
that water, even in small amounts as present in Paratone, was being
absorbed by crystals of 4. This served to rehydrate the Ba ions
and open the ‘Ba-belted’ closed-pore cavities described previously.
The clear implication of this result was that gas comprising the
atmosphere could be taken up by the network and actually trapped
within the pores of 4. A video that shows the real-time release of gas
from a pellet of 4 is provided as Supplementary Information. Such
regulated gas release has been observed for calixarene-based van der
Waals crystals43,46 but, to our knowledge, never for a coordination
network or any more strongly bonded system.
Several experiments were undertaken to ascertain the nature of
the gas capture. N2 and CO2 sorption measurements on 4 showed
no porosity. However, none would be expected for a closed-pore
system. Sample 1 was then probed for porosity with continuousflow hyperpolarized 129 Xe NMR (Fig. 4). The signal at ∼130 p.p.m.
nature materials VOL 7 MARCH 2008 www.nature.com/naturematerials
ARTICLES
a
c
b
d
Figure 5 Accessibility of the pores in 4 to spheres of different radii representing different gases as modelled in atoms. a, Pores in the crystal structure of 4. b, Surface
accessible to a sphere of 1.4 Å (∼N2 ). c, Surface accessible to a sphere of 2.1 Å (∼Xe). d, Surface accessible to a sphere of 1.2 Å permitting one disordered water molecule.
is consistent with Xe in channels of 0.6–0.8 nm diameter, but,
with drying, the volume of Xe in the material diminished until a
signal for confined Xe is barely detectable. However, the observed
bubbling could be reconciled by modelling the accessibility of the
pores in 4 to gases of different radii (Fig. 5). The pores in 4 can
contain gases with a radius of 1.4 Å (N2 ) and, expectedly, gases with
a radius of 2.1 Å (Xe) to a much smaller extent. Allowing for one
disordered water molecule makes the pores virtually inaccessible
to a sphere of even 1.2 Å radius. Thus, for the larger Xe atom, it
is likely even trace moisture renders it an ineffective probe of this
material. Characterization of atmospheric gases as trapped in the
pores was difficult owing to their thermal motion and the fact that
the same gases surround the crystals. To ultimately show that gases
comprising the external atmosphere could be included in closed
pores of 4, a colourimetric experiment was carried out (see Fig. 6
for details). Crystals of 1 were dehydrated to 4 in an atmosphere of
NH3 (g) then heated under vacuum to liberate externally adsorbed
NH3 . The dried crystals were then placed in an aqueous Co(NO3 )2
solution. The Co2+ ions readily adsorbed to the crystal’s surface
to give a pink colour. This coincided with release of gas from the
crystal as water rehydrated the Ba ions and opened the closed pores.
Addition of water, as in the Supplementary Information video, does
also cause dissolution. In this case, dissolution is apparently slowed
by adsorption of Co(). However, in ∼2 min, the crystals changed
colour from pink to a deep green characteristic of hexammine
Co(). As any externally sorbed gases would have been removed
in the desorption cycle, the only source for NH3 (g) would be
nature materials VOL 7 MARCH 2008 www.nature.com/naturematerials
the interior of the solid, confirming that the external atmosphere
was trapped during dehydration. Figure 6b shows spots of green
visible indicating where NH3 is escaping the crystal. Given the ‘pore
re-opening’ mechanism of the gas uptake, little selectivity in the
process was expected. All gases (CO2 , O2 , N2 , CH4 , He, Ar, Xe, NH3 )
examined as the external atmosphere were taken up.
With respect to gas release from the crystal, two processes
are in operation. On rehydrating 4 by addition of drops of 25%
water in methanol, powder X-ray diffraction (see Supplementary
Information, Fig. S14) confirms the solid reverts to hydrated
crystalline forms that require re-opening of the pores. Dissolution
of the solid does occur at moderate to high degrees of hydration,
rapidly releasing gas as shown in the Supplementary Information
video. At controlled low degrees of hydration, it is expected that
gas could be released exclusively by pore opening, as observed with
Paratone (Fig. 3). The fact that release is controllable by rehydration
though is independent of mechanism. Regarding visualization of
gas release, there are several pertinent variables. More gas trapping,
and hence more bubbling, is enhanced by: higher water content
in the rehydrating solvent, higher heating rates in the dehydration
cycle, longer heating cycles and smaller particles of 4. The actual
visualization is enhanced by a more viscous medium as the bubbles
merge on the solid surface.
The mechanical properties of a solid are inherently linked to
the strength and directionality of the interactions that sustain its
connectivity. In attempts to generate porous materials, efforts have
focused on strong bonding and regimented interactions. However,
233
ARTICLES
a
b
c
d
Figure 6 Release of NH3 from 4 on rehydration. Single crystals of 1 were taken
and placed into a stainless-steel cylinder connected to a home-made manometric
instrument. Anhydrous NH3 (30 psi) was introduced into the chamber, and then the
sample was heated for 4 h at 275 ◦C. This sample was then further heated under
vacuum (150 ◦C, 10−4 torr) for 2 h to liberate externally adsorbed NH3 on the
crystals. a, This crystal was then placed in a crystallizing dish containing
Co(NO3 ) 2 ·6H2 O (0.5 g,1.71 mmol) in 8 ml of 18 M distilled water. b, The pink
colouration due to Co2+ adsorption is observed immediately. c, Within 10–15 s, a
colour change to green, characteristic of hexammine Co2+ , is apparent. d, Within
2 min, the crystal forms small fractures as dissolution begins, resulting in a
pronounced colour change to a more intense green. The main illustration of this
experiment is not the mechanism of release, but rather to show that the external
atmosphere was trapped in the solid. On immersion of 4 in water, dissolution is
actually more rapid than this. In this experiment, the Co ions seem to play a role in
stabilizing the surface of the crystal.
solution and stirred for 10 min. The TBA salt of L was extracted from the basic
mixture using dichloromethane extractions (6 × 125 ml). The extracts were
combined, dried (MgSO4 ) and filtered, giving a clear solution. The solvent
was removed to give tris-tetrabutylammonium benzene-1,3,5-trisulphonate,
TBA3 L. Yield: 24.05 g, 32%. Found: C, 61.89; H, 10.92; N, 4.09. Calculated for
TBA3 L: C, 62.20; H, 10.73; N, 4.03. 1 H NMR (300 MHz D2 O) 0.916–0.965
(t, 36H, J = 0.0247 Hz), 1.281–1.404 (m, 24H, J = 0.0239), 1.539–1,679 (m,
24H, J = 0.0256), 3.136–3.193 (t, 24H, J = 0.0274), 8.294 (s, 3H). 13 C NMR
(300 MHz dimethylsulphoxide) 147.18 (C–SO3 ), 122.93 (C–H), 57.28 (α-CH2 ),
22.99 (β-CH2 ), 18.99 (γ -CH2 ), 13.27 (CH3 ).
PREPARATION OF 1 AND 2a
TBA3 (L) (250 mg, 0.239 mmol) and BaCl2 ·2H2 O (87.71 mg, 0.359 mmol) were
dissolved in water (2 ml). Single crystals of 1 were grown by vapour diffusion of
methanol into this solution. In 4–5 days, large (0.5 mm × 0.68 mm × 0.5 mm)
transparent crystalline blocks of Ba3 (H2 O)6 (L)2 (H2 O)n , 1, that were suitable
for X-ray analysis had grown. Using acetone as the diffusing solvent yielded
single crystals of 2a.
TRANSFORMATION OF 1 TO 2 AND 3 WITH DEHYDRATION
Crystals of 1 were left at ambient conditions. After a day, the crystals had
transformed to compound 2 (monoclinic, P 21 /c , a = 14.6451, b = 17.7557,
c = 12.1443 Å, β = 95.967◦ ). The crystal was continually dehydrating at
this point and a high-quality X-ray data set was not attainable even at low
temperature. Slight variations in the cell parameters were observed on the
crystals standing for more or less time. On standing for 7–14 days, the
dehydration reached a more defined plateau and a solvable data set was
obtained for compound 3 (orthorhombic, P 21 /c , a = 9.5625, b = 17.6670,
c = 15.6630 Å).
VARIABLE-TEMPERATURE SINGLE-CRYSTAL X-RAY DIFFRACTION AND TRANSFORMATION TO 4
Single crystals of 1 were flash frozen (173 K) using a liquid nitrogen stream,
and glued onto a Pyrex glass fibre using a two-component epoxy. The crystals
were indexed, collected and underwent a heating program of 10 ◦C steps using
the Oxford Cryostream on the diffractometer. At each step, the crystals were
re-indexed. This process continued until the crystals were at room temperature,
at which point they were heated to 380 K for a period of 3 h and a full data set
was collected giving 4.
129
the natural world is replete with examples where weak interactions
function cooperatively to form stable assemblies. The culmination
of this approach is not simply the ability to optimize the assembly
process but to enable dynamics and, ultimately, to derive function
from those dynamics. The Ba 1,3,5-benzenetrisulphonate system
presented herein is a network solid that retains crystallinity
through multiple stages of structural change. The transformation
in structure results in a highly stable, closed-pore solid capable
of trapping molecules of the atmosphere and releasing them in a
controllable manner. This material merges four often conflicting
properties of network solids (robustness, porosity, crystallinity and
dynamics) in a single compound. This reconciliation is enabled
by the cooperative use of weaker, less-directional interactions and
results in a new functional material with properties intermediate to
organic and inorganic solids.
METHODS
PREPARATION OF L
Sodium benzene-1,3-disulphonate (20.0 g, 70.94 mmol) was added to 20%
aqueous oleum (50.0 ml) and HgSO4 (800 mg, 2.68 mmol, 4 wt%). The
reaction mixture was heated at 290 ◦C for 24 h, and then cooled to room
temperature. The dark brown mixture was poured into water (600 ml) and
quenched with 30% NaOH(aq) until the solution pH was >10, and then
diluted to 1,200 ml with water. This solution was filtered using a 5 µm sintered
glass frit to remove Hg(OH)2 , leaving a clear yellow solution. An aqueous
solution (20 ml) of (TBA)HSO4 (38.9 g, 114.57 mmol) was added to this
234
Xe NMR MEASUREMENTS
Crystals of 1 were gravity filtered from their water/alcohol solution and
dabbed dry on filter paper. The crystals were transferred to the NMR
probe, and left overnight under a stream of argon at 80 ◦ C before any
experiments were carried out. Hyperpolarized 129 Xe gas was produced under
continuous-flow conditions using an optical pumping set-up similar to that
previously reported49,50 . The polarizer was located within the fringe field of
the spectrometer’s superconducting magnet and a 30 W continuous-wave
diode laser from OptoPower, operating at a wavelength of 785 nm, was used
as the source of optical excitation. The gas mixture of xenon, helium and
nitrogen with a volume composition of 1%–96%–3% respectively, and flow
rates of 110–160 cm3 min−1 (gas flow normalized to standard conditions), was
delivered through 1.5-mm-inner-diameter plastic tubing from the polarizer
directly to the sample contained in a modified Morris Instruments NMR probe
equipped with a 5 mm coil. The continuous-flow hyperpolarized 129 Xe NMR
experiments were carried out on a Bruker AMX-300 spectrometer operating at
83.03 MHz (magnetic field of 7.04 T). For all single-pulse experiments, 1,024
scans were acquired with a pulse recycle time of 1 s. Variable-temperature
experiments were carried out over the range of 173–368 K at 15◦ intervals
regulated by a Bruker BVT3000 temperature controller unit. The chemical shifts
of 129 Xe were referenced to xenon gas extrapolated to zero pressure. Before
experiments were repeated the next day, the sample was left in the probe under a
room-temperature flow of argon gas. A week later, the sample was probed again,
at which point it was heated in a 120 ◦ C oven for 6 h before NMR experiments
were carried out.
Received 3 May 2007; accepted 4 December 2007; published 20 January 2008.
References
1. Yaghi, O. M. et al. Reticular synthesis and the design of new materials. Nature 423, 705–714 (2003).
2. Kitagawa, S., Kitaura, R. & Noro, S.-i. Functional porous coordination polymers. Angew. Chem. Int.
Edn 43, 2334–2375 (2004).
3. Janiak, C. Engineering coordination polymers towards applications. Dalton Trans. 2781–2804 (2003).
nature materials VOL 7 MARCH 2008 www.nature.com/naturematerials
ARTICLES
4. Li, H., Eddaoudi, M., O’Keeffe, M. & Yaghi, O. M. Design and synthesis of an exceptionally stable and
highly porous metal-organic framework. Nature 402, 276–279 (1999).
5. Férey, G. et al. A chromium terephthalate-based solid with unusually large pore volumes and surface
area. Science 309, 2040–2042 (2005).
6. Chae, H. K. et al. A route to high surface area, porosity and inclusion of large molecules in crystals.
Nature 427, 523–527 (2004).
7. Park, K. S. et al. Exceptional chemical and thermal stability of zeolitic imidazolate frameworks. Proc.
Natl Acad. Sci. 103, 10186–10191 (2006).
8. Wu, C.-D., Hu, A., Zhang, L. & Lin, W. A homochiral porous metal–organic framework for highly
enantioselective heterogeneous asymmetric catalysis. J. Am. Chem. Soc. 127, 8940–8941 (2005).
9. Pan, L. et al. Porous lanthanide-organic frameworks: Synthesis, characterization, and unprecedented
gas adsorption properties. J. Am. Chem. Soc. 125, 3062–3067 (2003).
10. Chandler, B. D., Cramb, D. T. & Shimizu, G. K. H. Microporous metal-organic frameworks formed in
a stepwise manner from luminescent building blocks. J. Am. Chem. Soc. 128, 10403–10412 (2006).
11. Dinca, M., Yu, A. F. & Long, J. R. Microporous metal-organic frameworks incorporating
1,4-benzeneditetrazolate: Syntheses, structures, and hydrogen storage properties. J. Am. Chem. Soc.
128, 8904–8913 (2006).
12. Rowsell, J. L. C. & Yaghi, O. M. Strategies for hydrogen storage in metal–organic frameworks. Angew.
Chem. Int. Edn 44, 4670–4679 (2005).
13. Zhao, X. et al. Hysteretic adsorption and desorption of hydrogen by nanoporous metal-organic
frameworks. Science 306, 1012–1015 (2004).
14. Mueller, U. et al. Metal-organic frameworks—prospective industrial applications. J. Mater. Chem. 16,
626–636 (2006).
15. Matsuda, R. et al. Highly controlled acetylene accommodation in a metal–organic microporous
material. Nature 436, 238–241 (2005).
16. Eddaoudi, M. et al. Systematic design of pore size and functionality in isoreticular MOFs and their
application in methane storage. Science 295, 469–472 (2002).
17. Uemura, K., Matsuda, R. & Kitagawa, S. Flexible microporous coordination polymers. J. Solid State
Chem. 178, 2420–2429 (2005).
18. Bradshaw, D., Warren, J. E. & Rosseinsky, M. J. Reversible concerted ligand substitution at alternating
metal sites in an extended solid. Science 315, 977–980 (2007).
19. Maji, T. K., Matsuda, R. & Kitagawa, S. A flexible interpenetrating coordination framework with a
bimodal porous functionality. Nature Mater. 6, 142–148 (2007).
20. Halder, G. H., Kepert, C. J., Moubaraki, B., Murray, K. S. & Cashion, J. D. Guest dependent spin
crossover in a nanoporous molecular framework material. Science 298, 1762–1765 (2002).
21. Soldatov, D. V. et al. Alpha- and
beta-bis(1,1,1-trifluoro-5,5-dimethyl-5-methoxyacetylacetonato)copper(II): Transforming the dense
polymorph into a versatile new microporous framework. J. Am. Chem. Soc. 121, 4179–4188 (1999).
22. Serre, C. et al. Role of solvent-host interactions that lead to very large swelling of hybrid frameworks.
Science 315, 1828–1831 (2007).
23. Dybtsev, D. N., Chun, H. & Kim, K. Rigid and flexible: A highly porous metal-organic framework
with unusual guest-dependent dynamic behavior. Angew. Chem. Int. Edn 43, 5033–5036 (2004).
24. Yamada, K. et al. Metal-complex assemblies constructed from the flexible hinge-like ligand H2 bhnq:
Structural versatility and dynamic behavior in the solid state. Chem.-Eur. J. 10, 2648–2660 (2004).
25. Cussen, E. J., Claridge, J. B., Rosseinsky, M. J. & Kepert, C. J. Flexible sorption and transformation
behavior in a microporous metal–organic framework. J. Am. Chem. Soc. 124, 9574–9581 (2002).
26. Zhang, J. P., Lin, Y. Y., Zhang, W. X. & Chen, X. M. Temperature- or guest-induced drastic
single-crystal-to-single-crystal transformations of a nanoporous coordination polymer. J. Am. Chem.
Soc. 127, 14162–14163 (2005).
27. Soldatov, D. V., Moudrakovski, I. L., Ratcliffe, C. I., Dutrisac, R. & Ripmeester, J. A. Sorption of
xenon, methane, and organic solvents by a flexible microporous polymer
catena-bis(Dibenzoylmethanato)-(4,40 -bipyridyl)nickel(II). Chem. Mater. 15, 4810–4818 (2003).
28. Choi, H. J. & Suh, M. P. Dynamic and redox active pillared bilayer open framework:
Single-crystal-to-single-crystal transformations upon guest removal, guest exchange, and framework
oxidation. J. Am. Chem. Soc. 126, 15844–15851 (2004).
29. Chen, C. L., Goforth, A. M., Smith, M. D., Su, C. Y. & zur Loye, H. C.
[Co2 -(ppca)2 (H2 O)(V4 O12 )0.5 ]: A framework material exhibiting reversible shrinkage and expansion
through a single-crystal-to-single-crystal transformation involving a change in the cobalt
coordination environment. Angew. Chem. Int. Edn 44, 6673–6677 (2005).
nature materials VOL 7 MARCH 2008 www.nature.com/naturematerials
30. Ohmori, O., Kawano, M. & Fujita, M. Crystal-to-crystal guest exchange of large organic molecules
within a 3D coordination network. J. Am. Chem. Soc. 126, 16292–16293 (2004).
31. Wu, C. D. & Lin, W. B. Highly porous, homochiral metal-organic frameworks:
Solvent-exchange-induced single-crystal to single-crystal transformations. Angew. Chem. Int. Edn 44,
1958–1961 (2005).
32. Seki, K. Dynamic channels of a porous coordination polymer responding to external stimuli. Phys.
Chem. Chem. Phys. 4, 968–1971 (2002).
33. Mäkinen, S. K., Melcer, N. J., Parvez, M. & Shimizu, G. K. H. Highly selective guest uptake in a silver
sulfonate network imparted by a tetragonal to triclinic shift in the solid state. Chem.-Eur. J. 7,
5176–5182 (2001).
34. Shimizu, G. K. H. Assembly of metal ions and ligands with adaptable coordinative tendencies as a
route to functional metal-organic solids. J. Solid State Chem. 178, 2519–2526 (2005).
35. Khuong, T. A. V., Nunez, J. E., Godinez, C. E. & Garcia-Garibay, M. A. Crystalline molecular
machines: A quest toward solid-state dynamics and function. Acc. Chem. Res. 39, 413–422 (2006).
36. Kitagawa, S. & Uemura, K. Dynamic porous properties of coordination polymers inspired by
hydrogen bonds. Chem. Soc. Rev. 34, 109–119 (2005).
37. Suh, M. P. & Cheon, Y. E. Recent advances in the dynamics of single crystal to single crystal
transformations in metal-organic open frameworks. Aust. J. Chem. 59, 605–612 (2006).
38. Uemura, K., Kitagawa, S., Fukui, K. & Saito, K. A contrivance for a dynamic porous framework:
Cooperative guest adsorption based on square grids connected by amide–amide hydrogen bonds. J.
Am. Chem. Soc. 126, 3817–3828 (2004).
39. Suh, M. P., Ko, J. W. & Choi, H. J. A metal-organic bilayer open framework with a dynamic
component: Single-crystal-to-single-crystal transformations. J. Am. Chem. Soc. 124,
10976–10977 (2002).
40. Takaoka, K., Kawano, M., Tominaga, M. & Fujita, M. In situ observation of a reversible single
crystal-to-single-crystal apical-ligand exchange reaction in a hydrogen-bonded 2D coordination
network. Angew. Chem. Int. Edn 44, 2151–2154 (2005).
41. Zhang, J. P., Lin, Y. Y., Zhang, W. X. & Chen, X. M. Temperature- or guest-induced drastic
single-crystal-to-single-crystal transformations of a nanoporous coordination polymer. J. Am. Chem.
Soc. 127, 14162–14163 (2005).
42. Takamizawa, S., Nakata, E., Yokoyama, H., Mochizuki, K. & Mori, W. Carbon dioxide inclusion
phases of a transformable 1D coordination polymer host [Rh2 (O2 CPh)4 (pyz)]n . Angew. Chem. Int.
Edn 42, 4331–4334 (2003).
43. Enright, G. D., Udachin, K. A., Moudrakovski, I. L. & Ripmeester, J. A. Thermally programmable gas
storage and release in single crystals of an organic van der Waals host. J. Am. Chem. Soc. 125,
9896–9897 (2003).
44. Atwood, J. L., Barbour, L. J., Jerga, A. & Schottel, B. L. Guest transport in a nonporous organic solid
via dynamic van der Waals cooperativity. Science 298, 1000–1002 (2002).
45. Ananchenko, G. S. et al. Guest exchange in single crystals of van der Waals nanocapsules. Angew.
Chem. Int. Edn 45, 1585–1588 (2006).
46. Atwood, J. L., Barbour, L. J. & Jerga, A. Storage of methane and freon by interstitial van der Waals
confinement. Science 298, 1000–1002 (2002).
47. Leontiev, A. V. & Rudkevich, D. M. Encapsulation of gases in the solid state. Chem. Commun.
1468–1469 (2004).
48. Souza, D. C. S., Pralong, V., Jacobson, A. J. & Nazar, L. F. A reversible solid-state crystalline
transformation in a metal phosphide induced by redox chemistry. Science 296, 2012–2015 (2002).
49. Moudrakovski, I. L. et al. Chemical shift imaging with continuously flowing hyperpolarized xenon
for the characterization of materials. J. Magn. Reson. 144, 372–377 (2000).
50. Driehuys, B. et al. High-volume production of laser-polarized Xe-129. Appl. Phys. Lett. 69,
1668–1670 (1996).
Acknowledgements
D.T.C. and G.K.H.S. thank the Natural Sciences and Engineering Research Council (NSERC)
of Canada.
Correspondence and requests for materials should be addressed to D.T.C. or G.K.H.S.
Supplementary Information accompanies this paper on www.nature.com/naturematerials.
Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/
235