Solid solutions of soft porous coordination polymers: Fine-tuning of gas adsorption properties

Article abstract of DOI:10.1002/anie.201000989

(Figure Presented) My flexible friend: Separation of a CO₂/ CH₄ mixture was optimized by varying the composition of solid solutions of interdigitated 2D frameworks comprising two different isophthalate ligands (see picture). The characteristics of the solid solutions were dependent on the ligand ratio, which influenced the inherent flexibility and therefore the adsorption properties of the frameworks.

Full text of DOI:10.1002/anie.201000989

Communications
DOI: 10.1002/anie.201000989
Coordination Polymers
Solid Solutions of Soft Porous Coordination Polymers: Fine-Tuning of
Gas Adsorption Properties
Tomohiro Fukushima, Satoshi Horike,* Yasutaka Inubushi, Keiji Nakagawa, Yoshiki Kubota,
Masaki Takata, and Susumu Kitagawa*
Control of the physical and chemical properties of porous
materials has been an ongoing challenge for the optimization
of functions, such as gas storage, separation, and catalysis. For
example, a high surface area is important for large-volume gas
uptake, and the control of pore shape is also significant for
molecular separation.[1] These requirements are also valid for
porous coordination polymers (PCPs) or metal–organic
frameworks (MOFs), which consist of metal ions and organic
linkers.[2] This class of adsorbent has received attention
because of its structural versatility and physical properties,
such as magnetism and redox activity.[3] Among the PCP
compounds, flexible frameworks have been identified as a
unique type of porous material because of their guest-
responsive transformations.[4] This structural transformation
is often directly related to the functionality of these frame-
works; gas recognition separation or slow drug release are
good examples in this respect.[5]
The flexibility of the network must be modulated to
precisely control these functions and to tailor the network
performance, and much effort has been expended in creating
flexible compounds.[6] However, there have been few reports
on the rational incorporation of flexibility in the known PCP
compounds as their synthesis is difficult.[7] Herein, we
describe the preparation of ligand-based solid solutions of
flexible PCPs and our attempts to overcome difficulties in the
precise flexibility control and resulting gas sorption proper-
ties. A few approaches toward ligand-based solid solutions of
robust metal–organic framework have been reported,[8]
although corresponding structural information and control
of their adsorptive functions has not been observed. We have
synthesized two distinct interdigitated frameworks that con-
tain different organic ligands, and have created a series of
solid solutions based on these frameworks. These compounds
exhibited a range of flexible adsorption properties, and their
bimodal properties enabled them to show an improved
performance compared with the two pure compounds CID-
5 and CID-6 in the separation of a CO2/CH4 mixture.
The two flexible compounds with an interdigitation motif
of 2D layers, [{Zn(5-NO2-ip)(bpy)}(0.5DMF·0.5MeOH)]n
(CID-5ꢀG; 5-NO2-ip = 5-nitroisophthalate, bpy = 4,4’-bipyr-
idyl, and CID = coordination polymer with an interdigitated
structure), and [{Zn(5-MeO-ip)(bpy)}(0.5DMF·0.5MeOH)]n
(CID-6ꢀG; 5-MeO-ip = 5-methoxyisophthalate), were pre-
pared from Zn(NO3)2·6H2O and either of the ligands in a 1:1
v/v DMF/MeOH mixture. The crystal structures are shown in
Figure 1. For both compounds, two carboxylate groups
coordinate to Zn2+ ions to form eight-membered rings, and
the bpy groups coordinate to the axial position of the Zn2+
ions to create the 2D layered structure. The layers are
assembled in an interdigitated fashion with micropores
formed within the structure. Both frameworks can be
[*] T. Fukushima, Dr. S. Horike, K. Nakagawa, Prof. Dr. S. Kitagawa
Department of Synthetic Chemistry and Biological Chemistry
Graduate School of Engineering, Kyoto University, Katsura
Nishikyo-ku, Kyoto 615-8510 (Japan)
Fax : (+81)75-383-2732
E-mail: horike@sbchem.kyoto-u.ac.jp
kitagawa@sbchem.kyoto-u.ac.jp
Prof. Dr. S. Kitagawa
Institute for Integrated Cell-Material Sciences
and
Kitagawa Integrated Pore Project, Exploratory Research for Advanced
Technology (ERATO) Science and Technology Agency (JST) (Japan)
Y. Inubushi
Kurashiki Research Center, Kuraray Co., Ltd. (Japan)
Prof. Dr. Y. Kubota
Osaka Prefecture University (Japan)
Prof. Dr. M. Takata
Japan Synchrotron Radiation Research Institute/SPring-8 Sayo-gun,
Hyogo 679-5198 (Japan)
Figure 1. Partial crystal structures of a) CID-5ꢀG (left) and CID-6ꢀG
(right), and the assembled structures of b) CID-5ꢀG and CID-5 and
c) CID-6ꢀG and CID-6.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201000989.
4820
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 4820 –4824

Angewandte
Chemie
classified as members of a previously reported series of
flexible PCPs.[5c,9] Because the size of the substituent groups
(NO2 and MeO) in the dicarboxylates is comparable, the unit
cell parameters obtained for CID-5ꢀG and CID-6ꢀG are not
markedly different. The thermogravimetric analysis (TGA)
profiles of these samples indicate that DMF and MeOH can
be released on incorporation of a guest to afford a stable
guest-free framework that remains intact up to 3008C. The
weight loss of each compound was 10.8 and 11.5 wt%,
respectively, which is reasonable when their crystal structures
are taken into account.
We investigated the crystal structures of the guest-free
CID-5 and CID-6 forms to obtain direct information on their
contraction behavior. Although the cell parameters of the as-
synthesized compounds are similar, the guest-free structures
were different after degassing. As the guest molecule exits,
the flatness of the Zn2+ eight-membered rings is disturbed
because of the flexibility of this core module, and conse-
quently, a reorientation of the interdigitation for mutual
packing occurs (Figure 1b). There is no guest-accessible void
volume in CID-5, thus a “porous” to “nonporous” trans-
formation occurs. CID-6ꢀG shows contrasting behavior.
Even after complete degassing of the guest molecules, the
crystal structure of CID-6 shows only a small difference when
compared to the as-synthesized structure (Figure 1c), and the
change in void volume decreases from 15.8% to 14.6%. The
Zn2+ eight-membered ring in CID-6 does not show a large
distortion and therefore a small structural change. The
characteristics of the substituent group, such as electronic
properties or shape are the cause of the different behavior of
CID-6.
Figure 2. XRPD patterns of CID-5ꢀG, CID-6ꢀG, and CID-5/6ꢀG
(x=0.52), and simulated pattern of CID-5/6ꢀG (x=0.52).
Figure 3. a) Partial crystal structure (around the Zn2+ center) and
b) interdigitated structure of the solid solution CID-5/6ꢀG (x=0.52).
Based on the information obtained for CID-5ꢀG and
CID-6ꢀG, we prepared ligand-based solid solutions. Careful
mixing of H25-NO2-ip and H25-MeO-ip with Zn(NO3)2·6H2O
in molar ratios of 1:1:2 in a 1:1 v/v DMF/MeOH solution
resulted in a white powder containing 48% of 5-NO2-ip and
52% of 5-MeO-ip, as confirmed by elemental analysis and
1H NMR spectroscopy after degradation of the powder with
H2SO4 in DMSO (the product [{Zn(5-NO2-ip)1Àx(5-MeO-
ip)x(bpy)}(DMF·MeOH)]n is denoted as CID-5/6ꢀG (x =
0.52), where x is the content of 5-MeO-ip ligand in the
sample) To exclude the possibility that the powder was just
the mixture of pure microcrystals of CID-5 and CID-6, X-ray
powder diffraction data was obtained (Figure 2). The result-
ing pattern was different from either of the pure compounds,
in particular, the peaks that occur at around 2q = 7.28 and
2q = 88 represent an original phase when compared with the
peak positions of CID-5ꢀG and CID-6ꢀG. To find out if the
powder pattern was representative of a single crystalline
phase, we prepared a single crystal of CID-5/6ꢀG (x = 0.52),
and succeeded in solving its structure (Figure 3). Two types of
ligand, which are disordered with a dihedral angle of 38.28,
were observed around the Zn2+ centers. The mixed-ligand
Zn2+ eight-membered rings are connected by bpy ligands to
form the 2D layers, thus resulting in the formation of an
interdigitated framework. The unit-cell parameters and
volume obtained lie between the values obtained for CID-
5ꢀG and CID-6ꢀG. This result is reasonable if we consider
that both the 5-NO2-ip and 5-MeO-ip ligands are evenly
dispersed and form a single-phase crystal. The simulated
powder X-ray pattern of CID-5/6ꢀG (x = 0.52) generated
from the single crystal structure and the experimental pattern
(Figure 2) are in good agreement, and we concluded that the
obtained powder sample could be regarded as a ligand-based
solid solution of CID-5ꢀG and CID-6ꢀG.[10]
We subsequently synthesized a series of CID-5/6ꢀG solid
solutions with different ratios of 5-NO2-ip and 5-MeO-ip in
the range 0.06 < x < 0.92. We observed that the as-synthesized
compounds had unique XRPD patterns, not all of which are a
mixture of the two types of crystal, as shown in Figure 4.
Determination of the content of each ligand in the com-
pounds was carried out using 1H NMR spectroscopy and
elemental analysis. The actual 5-MeO-ip content for all the
compounds was slightly higher than the theoretical value
calculated from the mixed ligand ratio. From the XRPD
patterns, the crystal structure gradually shifted to the pattern
of CID-6ꢀG as the value of x increased. For these com-
pounds, we determined the cell parameters using the LeBail
fitting procedure, and these values also gradually shifted from
those of CID-5ꢀG to CID-6ꢀG. The thermal stability of the
solid solutions determined from the TGA data was checked,
and the profiles were similar to other CID frameworks, thus
indicating that the solid solutions can be used as flexible
porous coordination polymers, similar to pure CID-5ꢀG and
CID-6ꢀG. The key for the successful preparation of solid
Angew. Chem. Int. Ed. 2010, 49, 4820 –4824
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4821

Communications
Figure 4. XRPD patterns of CID-5ꢀG, CID-6ꢀG, and solid solutions of
Figure 5. Adsorption isotherms (closed circles) and desorption iso-
therms (open circles) of H2O at 298 K for: a) CID-5 and solid solution
CID-5/6 (x=0.06–0.22) and b) CID-5/6 (x=0.43–0.82) and CID-6.
CID-5/6ꢀG (x=0.13–0.92).
solutions in this system is that the unit cell parameters of CID-
5ꢀG and CID-6ꢀG are not markedly different, and there
would be little stress in the crystal, even if the two types of
ligand are mixed with arbitrary ratios in the coordination
networks.
5 and CID-6 in their solid solutions. In the case of CID-5, the
intrusion of H2O occurred at a pressure of Pgo = 2.48 kPa, and
desorption occurred at a low pressure. We observed a large
hysteresis in this process, which is denoted as a Type V profile
using the IUPAC classification (Figure 5a).[14] The total
uptake at 3 kPa was just five H2O molecules per unit pore;
this value suggests the formation of water clusters in the
cavities.[15] In contrast, the data from CID-6 shown in
Figure 5b shows a gradual uptake from low pressures without
a large hysteresis curve. The total uptake was three H2O
molecules per unit pore, which is lower than that of CID-5.
Meanwhile, the H2O isotherms of the solid solutions of CID-
5/6 represent unique profiles. For CID-5/6 (x = 0.06), whose
structure is close to that of CID-5, gate-opening adsorption
takes place at 2.42 kPa, which is lower than CID-5, and the
width of the hysteresis was narrower than that of CID-5. The
uptake was the same as CID-5; this result suggests that CID-
5/6 (x = 0.06) not only has similar sorption properties to CID-
5 but also possesses some characteristics of CID-6. The shift
of Pgo to lower values is attributed to an increase in the
frameworkꢀs affinity for H2O because of the presence of a
small amount of 5-MeO-ip ligands. The narrowing of the
hysteresis width also arises from the decrease in the frame-
work cooperativity and a decrease in the activation energy for
the structural transformation from the degassed state to the
adsorbed state. As the value of x increased, we observed a
gradual shift in Pgo, and the gradient of the profile also
decreased, thus resulting in a narrowing of the hysteresis
width. In the range 0.06 < x < 0.22 (Figure 5a), the total
We also measured the XRPD patterns of these samples
after degassing (CID-5/6). Each pattern showed some
changes, which were attributed to the flexibility, and the
patterns of CID-5/6 showed a gradual shift from CID-5 to
CID-6 as the value x increased, thus suggesting that each solid
solution has unique cell parameters even after a degassing
procedure. However, we were not successful in determining
the unit-cell parameters of these degassed solid solutions.
We investigated the gas sorption properties of these
compounds and, in particular, control of the “gate-opening”
type sorption behavior, in which a sudden adsorption occurs
at a given pressure point (not zero).[11] This behavior is
observed when the closed phase of a flexible PCP changes to
the open phase as a guest molecule is accommodated. The
gate-opening pressure Pgo is often sensitive to the affinity of
the framework to the gas molecule, and is regarded as an
important factor for gas separation, especially for gases that
have similar properties.[12] As CID-5 is nonporous, it exhibits a
gate-opening sorption behavior for H2O and CO2 molecules,
as shown in Figure 5 and 6, respectively. The value of Pgo is
2.48 kPa for H2O at 298 K and 1.32 kPa for CO2 at 195 K.[13]
On the other hand, the porous CID-6 does not show a gate-
opening-type behavior, but shows a linear uptake for H2O and
typical Type I isotherms for CO2.[14] We tried to produce
various gate-opening isotherms with different values of Pgo by
making use of the complementary sorption behaviors of CID-
4822 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 4820 –4824

Angewandte
Chemie
uptake was five molecules per unit pore, and for these solid
solutions, this water cluster would stabilize the overall
framework. At higher values of x (Figure 5b), the sorption
profiles became more gentle with a decrease in the total
uptake, which finally reached the value of CID-6.
(Figure 7). CID-5 has a closed-pore system, and so only CO2
could be captured with negligible uptake (2.5 mLgÀ1). On the
other hand, CID-6, which has an open-pore system, can
adsorb more CO2 (40 mLgÀ1), but it also simultaneously
The control of Pgo is important for gas molecules because
we have to deal with a range of relative pressures of the target
gas from gas mixtures requiring separation. The sorption
isotherms of CO2 at 195 K were measured for these com-
pounds (Figure 6). As observed for H2O, CID-5 showed a
Figure 7. Total uptake of CO2 (gray) and CH4 (black) for CID-5, CID-5/
6 (x=0.13), and CID-6 from a CO2/CH4 (1:1) gas mixture at a total
pressure of 101.3 kPa at 273 K.
adsorbs some CH4 (2 mLgÀ1), thus resulting in an unsatis-
factory separation for each compound. We utilized a solid
solution of CID-5/6 (x = 0.13) to take advantage of these
properties. CID-5/6 (x = 0.13) retained the advantages of
CID-6 and adsorbed 30 mLgÀ1 of CO2. Moreover, CID-5/6
(x = 0.13) also had the characteristics of CID-5, and so did not
adsorb CH4 at all. The overall selectivity towards CO2/CH4
was clearly improved compared with that of pure CID-5 and
CID-6. The low concentration of the 5-MeO-ip ligand in the
solid solution optimized the selective accommodation of CO2
into the pores.
In conclusion, we have synthesized two interdigitated
frameworks and a series of their ligand-based solid solutions
for the control of their adsorption isotherms. The crystallo-
graphic properties of the solid solutions were dependent on
the ligand ratio; the gate-opening pressure Pgo , the gradient of
the sorption profile, and the total gas uptake were dependent
on the inherent framework flexibility. Optimization of the
gas-separation performance from a mixture of CO2 and CH4
using the solid solutions was demonstrated. This approach
provides potential for the design of functional porous
coordination polymers or metal–organic frameworks, even
though these systems have a flexible nature.
Figure 6. Adsorption isotherms of CO2 at 195 K for CID-5, CID-6, and
their solid solution CID-5/6 (x=0.06–0.82) in the range a) P=0–
100 kPa and b) expansion of the range P=0–4 kPa.
clear gate-opening-type uptake, although the value of Pgo
occurred at a lower pressure (1.60 kPa) and CID-6 showed a
common Type I isotherm (Figure 6b). For CID-5/6, there was
also a gradual shift in the gate-opening pressure and the
gradient of the uptake was in the range 0.06 < x < 0.43. The
total amount adsorbed was also dependent on the value of x:
as x increased, the total amount decreased. CID-5/6 with x >
0.61 showed almost the same isotherms as CID-6. The results
of the adsorption experiments suggest that solid solutions of
CID-5/6 exhibit a range of gate-opening profiles based on the
complementary sorption properties of CID-5 and CID-6.
A target for PCP solid solutions is to incorporate the
advantageous properties of both compounds in the overall
adsorption functionality. The present system provides one
example in this respect: the improvement of gas separation
properties. The selective adsorption uptake of CO2 from a
CO2/CH4 mixture (CO2/CH4 = 1:1 by volume) under the
conditions of P = 101.3 kPa and T= 273 K was investigated
Received: February 17, 2010
Published online: May 20, 2010
Keywords: adsorption · carbon dioxide · coordination polymers ·
.
gas separation · water
[1] a) Z. X. Ma, T. Kyotani, A. Tomita, Chem. Commun. 2000,
2365 – 2366; b) A. Corma, F. Rey, S. Valencia, J. L. Jorda, J. Rius,
Nat. Mater. 2003, 2, 493 – 497; c) Z. X. Yang, Y. D. Xia, R.
Mokaya, J. Am. Chem. Soc. 2007, 129, 1673 – 1679.
[2] a) H. K. Chae, D. Y. Siberio-Perez, J. Kim, Y. Go, M. Eddaoudi,
A. J. Matzger, M. OꢀKeeffe, O. M. Yaghi, Nature 2004, 427, 523 –
527; b) S. Kitagawa, R. Kitaura, S. Noro, Angew. Chem. 2004,
Angew. Chem. Int. Ed. 2010, 49, 4820 –4824
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4823

Communications
116, 2388 – 2430; Angew. Chem. Int. Ed. 2004, 43, 2334 – 2375;
c) G. Fꢁrey, Chem. Soc. Rev. 2008, 37, 191 – 214.
[7] C. Serre, S. Surblꢁ, C. Mellot-Draznieks, Y. Filinchuk, G. Fꢁrey,
Dalton Trans. 2008, 5462 – 5464.
[3] a) D. Maspoch, D. Ruiz-Molina, K. Wurst, N. Domingo, M.
Cavallini, F. Biscarini, J. Tejada, C. Rovira, J. Veciana, Nat.
Mater. 2003, 2, 190 – 195; b) M. P. Suh, H. R. Moon, E. Y. Lee,
S. Y. Jang, J. Am. Chem. Soc. 2006, 128, 4710 – 4718; c) P. D.
Southon, L. Liu, E. A. Fellows, D. J. Price, G. J. Halder, K. W.
Chapman, B. Moubaraki, K. S. Murray, J. F. Letard, C. J. Kepert,
J. Am. Chem. Soc. 2009, 131, 10998 – 11009.
[4] a) A. J. Fletcher, K. M. Thomas, M. J. Rosseinsky, J. Solid State
Chem. 2005, 178, 2491 – 2510; b) G. Fꢁrey, C. Serre, Chem. Soc.
Rev. 2009, 38, 1380 – 1399; c) S. Horike, S. Shimomura, S.
Kitagawa, Nat. Chem. 2009, 1, 695 – 704.
[5] a) B. L. Chen, C. D. Liang, J. Yang, D. S. Contreras, Y. L. Clancy,
E. B. Lobkovsky, O. M. Yaghi, S. Dai, Angew. Chem. 2006, 118,
1418 – 1421; Angew. Chem. Int. Ed. 2006, 45, 1390 – 1393; b) P.
Horcajada, C. Serre, G. Maurin, N. A. Ramsahye, F. Balas, M.
Vallet-Regi, M. Sebban, F. Taulelle, G. Fꢁrey, J. Am. Chem. Soc.
2008, 130, 6774 – 6780; c) D. Tanaka, K. Nakagawa, M. Higuchi,
S. Horike, Y. Kubota, L. C. Kobayashi, M. Takata, S. Kitagawa,
Angew. Chem. 2008, 120, 3978 – 3982; Angew. Chem. Int. Ed.
2008, 47, 3914 – 3918.
[8] a) H. Chun, D. N. Dybtsev, H. Kim, K. Kim, Chem. Eur. J. 2005,
11, 3521 – 3529; b) W. Kleist, F. Jutz, M. Maciejewski, A. Baiker,
Eur. J. Inorg. Chem. 2009, 3552 – 3561; c) K. Koh, A. G. Wong-
Foy, A. J. Matzger, Chem. Commun. 2009, 6162 – 6164; d) H.
Deng, C. J. Doonan, H. Furukawa, R. B. Ferreira, J. Towne, C. B.
Knobler, B. Wang, O. M. Yaghi, Science 2010, 327, 846 – 850.
[9] S. Horike, D. Tanaka, K. Nakagawa, S. Kitagawa, Chem.
Commun. 2007, 3395 – 3397.
[10] D. Williams, J. Kouvetakis, M. OꢀKeeffe, Inorg. Chem. 1998, 37,
4617 – 4620.
[11] a) F. X. Coudert, M. Jeffroy, A. H. Fuchs, A. Boutin, C. Mellot-
Draznieks, J. Am. Chem. Soc. 2008, 130, 14294 – 14302; b) H.
Noguchi, A. Kondoh, Y. Hattori, H. Kanoh, H. Kajiro, K.
Kaneko, J. Phys. Chem. B 2005, 109, 13851 – 13853.
[12] L. Hamon, P. L. Llewellyn, T. Devic, A. Ghoufi, G. Clet, V.
Guillerm, G. D. Pirngruber, G. Maurin, C. Serre, G. Driver, W.
van Beek, E. Jolimaitre, A. Vimont, M. Daturi, G. Fꢁrey, J. Am.
Chem. Soc. 2009, 131, 17490 – 17499.
[13] The gate-opening pressure Pgo was defined as a flexion point of
the sigmoidal adsorption curve.
[6] a) B. L. Chen, S. Q. Ma, F. Zapata, F. R. Fronczek, E. B.
Lobkovsky, H. C. Zhou, Inorg. Chem. 2007, 46, 1233 – 1236;
[14] K. S. W. Sing, D. H. Everett, R. A. W. Haul, L. Moscou, R. A.
Pierotti, J. Rouquerol, T. Siemieniewska, Pure Appl. Chem.
1985, 57, 603 – 619.
˘
b) H. J. Choi, M. Dinca, J. R. Long, J. Am. Chem. Soc. 2008, 130,
7848 – 7850; c) Z. Q. Wang, S. M. Cohen, J. Am. Chem. Soc.
2009, 131, 16675 – 16677.
[15] T. Ohba, H. Kanoh, K. Kaneko, J. Am. Chem. Soc. 2004, 126,
1560 – 1562.
4824 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 4820 –4824