Syntheses, structures and properties of cadmium benzenedicarboxylate metal-organic frameworks

Article abstract of DOI:10.1039/b718947h

The products isolated from the reaction between Cd(NO₃) ₂?4H₂O and 1,4-benzenedicarboxylic acid (H ₂bdc) in DMF are very dependent on the conditions. At 115 °C, the reaction gives [Cd(bdc)(DMF)]_∞1, which has a three-dimensional network structure, whereas at 95 °C, 1 is formed alongside [Cd ₃(bdc)₃(DMF)₄]_∞2, which has a two-dimensional network structure. When the reaction is carried out under pressure, it yields [Cd₃(bdc)₃(DMF)₄] _∞3, which is a supramolecular isomer of 2. The structure of 3 differs from that of 2 regarding the way the Cd₃(O ₂CR)₆ units are interlinked to form layers. When the reaction was carried out in DMF that had undergone partial hydrolysis, the only isolated product was {(NMe₂H₂)₂[Cd(bdc) ₂]?2DMF}_∞4. Compound 4 has a three-dimensional triply-interpenetrated diamondoid structure, with dimethylammonium cations and DMF molecules included within the pores. The reaction between Cd(NO ₃)₂?4H₂O and H₂bdc in DEF gave [Cd(bdc)(DEF)]_∞5, regardless of the solvent quality. Compound 5 has a three-dimensional network structure. The reaction of Cd(NO ₃)₂?4H₂O and 1,3-benzenedicarboxylic acid (H₂mbdc) in DMF gave [Cd(mbdc)(DMF)]_∞6 which has a bilayer structure. The thermal properties of the new materials have been investigated, and the coordinated DEF molecules from 5 can be removed on heating to 400 °C without any change in the powder X-ray diffraction pattern. The H₂ sorption isotherm for the desolvated material shows marked hysteresis between adsorption and desorption, and less adsorption than predicted by simulations. Kinetic data indicate that the hysteresis is not due to mass transfer limitations, and the most likely explanation for this behaviour lies in partial collapse of the framework to an amorphous phase under the conditions of activation. The Royal Society of Chemistry.

Full text of DOI:10.1039/b718947h

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Syntheses, structures and properties of cadmium benzenedicarboxylate
metal–organic frameworks†
Andrew D. Burrows,*a Kevin Cassar,a,b Tina Du¨ren,c Richard M. W. Friend,a Mary F. Mahon,*a
Sean P. Rigby*b and Teresa L. Savaresea
Received 7th December 2007, Accepted 21st February 2008
First published as an Advance Article on the web 19th March 2008
DOI: 10.1039/b718947h
The products isolated from the reaction between Cd(NO3)2·4H2O and 1,4-benzenedicarboxylic acid
(H2bdc) in DMF are very dependent on the conditions. At 115 ◦C, the reaction gives [Cd(bdc)(DMF)]∞
1, which has a three-dimensional network structure, whereas at 95 ◦C, 1 is formed alongside
[Cd3(bdc)3(DMF)4]∞ 2, which has a two-dimensional network structure. When the reaction is carried
out under pressure, it yields [Cd3(bdc)3(DMF)4]∞ 3, which is a supramolecular isomer of 2. The
structure of 3 differs from that of 2 regarding the way the Cd3(O2CR)6 units are interlinked to form
layers. When the reaction was carried out in DMF that had undergone partial hydrolysis, the only
isolated product was {(NMe2H2)2[Cd(bdc)2]·2DMF}∞ 4. Compound 4 has a three-dimensional
triply-interpenetrated diamondoid structure, with dimethylammonium cations and DMF molecules
included within the pores. The reaction between Cd(NO3)2·4H2O and H2bdc in DEF gave
[Cd(bdc)(DEF)]∞ 5, regardless of the solvent quality. Compound 5 has a three-dimensional network
structure. The reaction of Cd(NO3)2·4H2O and 1,3-benzenedicarboxylic acid (H2mbdc) in DMF gave
[Cd(mbdc)(DMF)]∞ 6 which has a bilayer structure. The thermal properties of the new materials have
been investigated, and the coordinated DEF molecules from 5 can be removed on heating to 400 ◦C
without any change in the powder X-ray diffraction pattern. The H2 sorption isotherm for the
desolvated material shows marked hysteresis between adsorption and desorption, and less adsorption
than predicted by simulations. Kinetic data indicate that the hysteresis is not due to mass transfer
limitations, and the most likely explanation for this behaviour lies in partial collapse of the framework
to an amorphous phase under the conditions of activation.
that ‘mistakes’ made in assembling the network can be readily
Introduction
corrected, allowing crystalline compounds to be formed; (ii) the
Coordination networks—otherwise known as metal–organic
ligands are anionic, so combination with metal centres can give
frameworks (MOFs)—have attracted considerable interest over
the past few years.1,2 This is largely a result of their potential
for porosity, and the implications of this in applications such as
gas storage3 and catalysis.4 Perhaps the most successful class of
neutral networks, meaning that counter-ions do not occupy the
pores; (iii) the presence of the SBUs as opposed to single metal
centres means the network nodes are larger, so interpenetration is
less common; and (iv) the bridging coordination modes adopted
ligands for developing coordination networks is that based on di-
within the SBUs appears to lead to robust networks that are able
and tri-carboxylates. These linkers connect together small metal-
to withstand removal of solvent from the pores.
carboxylate aggregates, which are often referred to as secondary
Although the reactions of many transition metals with dicar-
building units (SBUs).1 The structures adopted by the resulting
boxylates have been studied, those of zinc(II) have been particularly
noteworthy.7,8 Yaghi and co-workers have reported a number
of interesting structures, and impressive methane and hydrogen
adsorption properties for compounds such as [Zn4O(bdc)3] (MOF-
5, bdc = 1,4-benzenedicarboxylate, terephthalate) which has a
simple cubic structure, based on Zn4O(O2CR)6 SBUs that are
interconnected by the benzene groups of the benzenedicarboxy-
late ligands.9 Given the interest in zinc(II) dicarboxylates, it is
surprising that the related chemistry of the heavier Group 12
analogue, cadmium, has received rather less attention. Although
there are a number of reports on cadmium(II) mixed ligand
systems containing both dicarboxylates and N-donor ligands,
which can be terminal, such as phenanthroline10 and ammonia,11
or bridging, such as 4,4ꢀ-bipy,12 simple reactions with dicarboxylic
acids which are analogous to the synthesis of MOF-5 have not
been reported. In this paper we report the synthesis of six new
coordination networks are determined largely by the geometry of
the SBUs and the distribution of carboxylate groups in the ligands,
although reports have shown that the synthetic conditions can be
crucial in determining the nature of the isolated product.5,6
Carboxylates have attracted attention in the synthesis of MOFs
for several reasons—(i) the M–O bonds are relatively labile, so
aDepartment of Chemistry, University of Bath, Claverton Down, Bath, UK
BA2 7AY. E-mail: a.d.burrows@bath.ac.uk, m.f.mahon@bath.ac.uk
bDepartment of Chemical Engineering, University of Bath, Claverton Down,
Bath, UK BA2 7AY. E-mail: s.p.rigby@bath.ac.uk
cInstitute for Materials and Processes, School of Engineering and Electronics,
University of Edinburgh, The King’s Buildings, Mayfield Road, Edinburgh,
UK EH9 3JL
† CCDC reference numbers 651815–651819. For crystallographic data in
CIF or other electronic format see DOI: 10.1039/b718947h
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 2465–2474 | 2465
©

View Article Online
◦
a
˚
cadmium(II) networks containing either 1,4-benzenedicarboxylate
or 1,3-benzenedicarboxylate (mbdc) that were all prepared under
relatively mild (i.e. non-solvothermal) conditions. In addition,
we report the thermal properties of the new three-dimensional
network materials, and gas adsorption studies on the one that
retains its framework structure following solvent removal.
Table 1 Selected bond lengths (A) and bond angles ( ) for 1
Cd(1)–O(1)
2.494(2)
Cd(1)–O(5)
2.274(2)
2.3947(19)
2.349(2)
Cd(1)–O(2)
Cd(1)–O(3)
Cd(1)–O(4)
O(2)–Cd(1)–O(1)
O(4)–Cd(1)–O(3)
O(5)–Cd(1)–O(1)
O(5)–Cd(1)–O(2)
O(5)–Cd(1)–O(3)
2.2788(19) Cd(1)–O(2)ꢀꢀ
2.517(2)
2.256(2)
54.87(7)
54.70(7)
144.16(8)
96.71(7)
86.36(7)
Cd(1)–O(3)ꢀ
O(4)–Cd(1)–O(5)
O(5)–Cd(1)–O(2)ꢀꢀ
O(5)–Cd(1)–O(3)ꢀ
O(4)–Cd(1)–O(1)
125.22(8)
79.34(7)
81.20(7)
90.53(8)
Results and discussion
Synthesis
a Primed atoms generated by the symmetry operation −x + 1, −y, −z + 1.
Double primed atoms generated by the symmetry operation −x, −y, −z
The product from the reaction between Cd(NO3)2·4H2O and
1,4-benzenedicarboxylic acid in DMF is acutely dependent on
the conditions employed. When the reagents were heated to
115 ◦C for 5 h in DMF, the reaction produced [Cd(bdc)(DMF)]∞
1 as the sole product. Carrying out the reaction at a lower
temperature (95 ◦C) led to a mixture of 1 and [Cd3(bdc)3(DMF)4]∞
2, where 2 is the minor product. When the reaction was carried
out in a sealed system at 115 ◦C under a pressure of 1.5–5
atmospheres of dinitrogen or compressed air, the isolated product
was [Cd3(bdc)3(DMF)4]∞ 3, which is a supramolecular isomer
of 2.
The reaction is also sensitive to the quality of the solvent. Use of
DMF in which partial hydrolysis has occurred led to the formation
of {(NMe2H2)2[Cd(bdc)2]·2DMF}∞ 4, with no evidence for the
formation of either 1 or 2. The formation of 4 can also be templated
by addition of [NMe2H2]Cl to the reaction mixture or by addition
of a few drops of acid to reduce the pH to 4.
+ 1.
Structures
[Cd(bdc)(DMF)]∞ 1. The asymmetric unit of 1 consists of one
cadmium centre, which is coordinated to two halves of bdc anions
and one DMF molecule. The remaining portion of each of bdc
anion is generated by inversion symmetry. The carboxylate groups
bridge between two cadmium centres, coordinating j1- to one and
j2- to the other. Each cadmium centre is 7-coordinate and bonded
to two j1-carboxylates, two j2-carboxylates and a DMF ligand.
Selected bond lengths and angles are given in Table 1. The bridging
carboxylate groups link the cadmium centres into infinite chains,
as shown in Fig. 1, which contain non-bonded Cd · · · Cd contacts
˚
of 3.65 and 3.86 A. These chains can be regarded as infinite
{Cd(O2CR)2}∞ SBUs, which are cross-linked by the benzene rings
of the bdc ligands into the three-dimensional network shown in
Fig. 2. The network contains rhomboidal pores, which are partly
filled by the coordinated DMF molecules.
The
reaction
between
Cd(NO3)2·4H2O
and
1,4-
benzenedicarboxylic acid in DEF is less dependent on the synthetic
conditions and the solvent quality, affording [Cd3(bdc)3(DEF)2]∞
5, regardless of the latter. Furthermore, addition of [NEt2H2]Cl
did not lead to a change of product, in contrast to the observations
in DMF or for the related zinc(II) system in either DMF or
DEF. The reaction of Zn(NO3)2·6H2O with H2bdc in DEF gives
[Zn4O(bdc)3] (MOF-5), though on addition of [NEt2H2]Cl the
compound {(NEt2H2)2[Zn3(bdc)4]·2.5DEF}∞ is formed instead.6
In order to investigate the effect of the relative orientation of
the carboxylate groups on the network structure, the reaction of
Cd(NO3)2·4H2O with 1,3-benzenedicarboxylic acid (H2mbdc) was
also investigated. When the reaction was carried out at 90 ◦C, the
compound [Cd(mbdc)(DMF)]∞ 6 was formed as the sole product.
The syntheses of compounds 1–6 are summarised in Scheme 1.
No evidence was observed for any interconversions involving 1–5.
Fig. 1 The cadmium carboxylate chains present in the structure of
[Cd(bdc)(DMF)]∞ 1.
Fig. 2 Part of the three-dimensional network present in the structure of
[Cd(bdc)(DMF)]∞ 1. Hydrogen atoms have been omitted for clarity.
Scheme 1
2466 | Dalton Trans., 2008, 2465–2474
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Fig. 3 The Cd3(l-O2CR)6 SBU present in the structure of [Cd3(bdc)3(DMF)4]∞ 2. Part (a) shows a side-on view and part (b) shows the directions in
which the bdc ligands radiate out from the Cd3 unit.
The structure of 1 is related to that previously reported for
[Cd(bdc)(py)]∞.11 This compound has similar {Cd(O2CR)2}∞
SBUs, with pyridine ligands taking the place of the DMF ligands
in 1. [Cd(bdc)(py)]∞ crystallises in a different space group to 1
(C2/c), however, and only contains one independent carboxylate.
Supramolecular isomers of [Cd3(bdc)3(DMF)4]∞, 2 and 3. The
asymmetric unit of 2 consists of one and a half cadmium centres,
one and a half bdc anions and two DMF molecules. Cd(2) lies on
a twofold axis, and the half bdc ion lies proximate to an inversion
centre. Two-fold rotational symmetry intrinsic to the space group
serves to generate Cd3(O2CR)6 SBUs. Selected bond lengths and
angles are given in Table 2. The DMF molecules are coordinated
Fig. 4 The two-dimensional layer structure present in the structure of
[Cd3(bdc)3(DMF)4]∞ 2. Hydrogen atoms have been omitted for clarity.
to the two outer cadmium centres, as shown in Fig. 3a, and this
prevents further propagation of the Cd3 unit into a chain similar
to that observed for 1.
to two independent Cd3(O2CR)6 SBUs, one containing Cd(1)
and Cd(2), and the other containing Cd(3) and Cd(4). The two
independent SBUs have very similar structures, with the central
metal in each case located at an inversion centre, and one of these
Within each Cd3 unit, the central Cd(2) atom has a distorted
octahedral geometry. It is j1-coordinated to six carboxylate
groups, and the cis angles range from 76.59(4)–100.97(5)◦. The
two symmetry related Cd(1) atoms are 7-coordinate, and are
is shown in Fig. 5a.
j1-coordinated to one carboxylate and j2-coordinated to two
The central Cd(2) and Cd(4) atoms are j1-coordinated to
carboxylates in addition to the two DMF molecules.
six carboxylate oxygen atoms. These cadmium centres display
distorted octahedral geometry, with cis angles ranging between
The Cd3(O2CR)6 SBUs are linked together by the benzene rings
of the bdc anions to form a sheet structure (Fig. 4) containing
87.00(5) and 93.00(5)◦ for Cd(2) and between 85.42(5) and
rhombohedral pores that are capped by the coordinated DMF
94.58(5)◦ for Cd(4). The terminal Cd(1) and Cd(3) atoms are
molecules.
coordinated to two j1-carboxylates, one j2-carboxylate and two
DMF molecules. The geometries around these atoms are distorted
octahedral, with the presence of the j2-carboxylate ensuring the
The asymmetric unit of 3 consists of two full cadmium atoms
[Cd(1) and Cd(3)], two half-occupancy cadmium atoms [Cd(2)
and Cd(4)], two full and two halves of bdc anions, the halves lying
distortions are far greater than those for the central cadmium
proximate to inversion centres, and four DMF molecules. Selected
atoms. cis angles range from 55.50(4) to 109.89(6)◦ for Cd(1) and
bond lengths and angles are given in Table 3. In the structure,
from 55.60(4) to 110.10(5)◦ for Cd(3).
the bdc anions bridge between linear Cd3(O2CR)6 SBUs similar
to those observed in 2 to generate a layer structure. The four
independent cadmium atoms in the asymmetric unit are connected
◦
a
˚
Table 3 Selected bond lengths (A) and bond angles ( ) for 3
Cd(1)–O(2)
Cd(1)–O(5)
Cd(1)–O(9)
Cd(1)–O(10)
Cd(1)–O(11)
Cd(1)–O(12)
Cd(2)–O(1)
2.2021(13) Cd(3)–O(3)
2.2296(13) Cd(3)–O(7)ꢀ
2.3641(13) Cd(3)–O(13)
2.3707(13) Cd(3)–O(14)
2.2490(14) Cd(3)–O(15)
2.3977(14) Cd(3)–O(16)
2.2336(13) Cd(4)–O(4)
2.2449(13) Cd(4)–O(8)ꢀ
2.3145(12) Cd(4)–O(14)
2.2158(13)
2.2078(13)
2.3827(13)
2.3565(13)
2.3564(13)
2.2733(14)
2.2161(14)
2.2651(12)
2.3244(12)
◦
a
˚
Table 2 Selected bond lengths (A) and bond angles ( ) for 2
Cd(1)–O(1)
Cd(1)–O(2)
Cd(1)–O(3)
Cd(1)–O(4)
2.3030(13) Cd(1)–O(6)
2.4661(13) Cd(1)–O(7)ꢀ
2.2801(14) Cd(2)–O(8)ꢀ
2.6035(14) Cd(2)–O(2)
2.2886(14) Cd(2)–O(4)
2.2611(13)
2.2970(13)
2.2175(13)
2.4620(13)
2.2863(13)
95.96(5)
Cd(2)–O(6)
Cd(1)–O(5)
Cd(2)–O(10)
O(2)–Cd(1)–O(9)
O(5)–Cd(1)–O(12)
O(11)–Cd(1)–O(10) 150.97(5)
O(9)–Cd(1)–O(10) 55.50(4)
O(1)–Cd(1)–O(2)
O(3)–Cd(1)–O(1)
O(5)–Cd(1)–O(7)ꢀ
55.17(5)
171.85(5)
175.82(5)
O(6)–Cd(1)–O(5)
150.36(5)
173.35(5)
O(3)–Cd(3)–O(15)
167.89(5)
O(2)–Cd(2)–O(2)* 165.66(6)
O(7)ꢀꢀ–Cd(3)–O(13) 145.91(5)
O(16)–Cd(3)–O(14) 165.69(5)
O(8)ꢀꢀ–Cd(2)–O(4)
164.81(5)
a Primed atoms generated by the symmetry operation x, −y, z − 1/2.
Double primed atoms generated by the symmetry operation −x,−y, −z
+ 1. Asterisked atoms generated by the symmetry operation −x, y, −z +
1/2.
O(14)–Cd(3)–O(13)
55.60(4)
a Primed atoms generated by the symmetry operation x, y− 1, z + 1. Double
primed atoms generated by the symmetry operation x, y − 1, z + 1.
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 2465–2474 | 2467
©

View Article Online
Fig. 5 One of the two independent Cd3(l-O2CR)6 SBUs present in the structure of [Cd3(bdc)3(DMF)4]∞ 3. Part (a) shows a side-on view and part (b)
shows the directions in which the bdc ligands radiate out from the Cd3 unit.
The dicarboxylate ligands link the Cd3(O2CR)6 SBUs into layers
containing triangular pores (Fig. 6). Each layer includes both
independent SBUs. Coordination of the DMF molecules to the
terminal Cd(1) and Cd(3) atoms ensures there are only weak
interactions between the layers.
This leads to greater distortions from regular octahedral geometry
around the central atom and, as a result, the carboxylate groups
radiate from the central Cd3 spine in a less regular manner than in
3 (Fig. 3b).
Side views of the sheets in 2 and 3 are shown in Fig. 7,
with the linkers depicted in a simplified form. In both cases, six
dicarboxylates radiate from each Cd3(O2CR)6 SBU, and two of
these connect an ‘upper’ part of the SBU with a ‘lower’ part of
a neighbouring SBU. Looking along an axis generated by the
SBUs, this gives a ‘criss-cross’ of dicarboxylates for 3 but a parallel
arrangement for 2. This difference is related to the symmetry
present in the SBUs, with the central atom of the Cd3(O2CR)6 unit
sitting on an inversion centre in 3 but a 2-fold rotation axis in 2.
Fig. 6 The two-dimensional layer structure present in the structure of
[Cd3(bdc)3(DMF)4]∞ 3. Hydrogen atoms have been omitted for clarity.
Only the oxygen atoms of the coordinated DMF molecules are shown.
Fig. 7 Side on view of the sheets in (a) compound 2 and (b) compound 3.
The bdc ligands are shown as simple lines between the carboxylate groups
for clarity. DMF ligands have also been omitted.
Compounds 2 and 3 have identical formulae and both contain
Cd3(O2CR)6 SBUs. In both cases these units are interlinked to form
sheet structures, but there are significant differences between the
two networks. 2 and 3 can therefore be regarded as supramolecular
isomers.
The Cd3(O2CR)6 SBUs observed in 2 and 3 are similar to
those seen previously in the compound (BMIM)2[Cd3(bdc)3Br2]
(BMIM = 1-butyl-3-methylimidazolium).13 In this case, the bro-
mide ligands terminate the Cd3 units, and the sheets formed
through the bridging bdc ligands have the same triangular pores
as those in 3.
The differences in the sheets are related to differences in the
Cd3(O2CR)6 SBUs. In 3 there are two independent SBUs, but both
are linear, with Cd · · · Cd · · · Cd angles of 180◦. Furthermore, each
of the terminal Cd atoms is connected to the central Cd atom by
two l-j1O,j1Oꢀ carboxylates and one l-j1O,j2O,Oꢀ carboxylate.
Together these factors result in the carboxylate groups radiating
from the central Cd3 spine at approximately 60◦ angles from each
other (Fig. 5b).
There are also several examples of zinc-based net-
works containing analogous SBUs. The compounds {[Zn3
7
6
(bdc)3(OH2)2]·4DMF}∞ and {[NEt2H2][Zn3(bdc)4] ·2.5DEF}∞
both contain Zn3(O2CR)6 SBUs, though in these cases the terminal
zinc centres are 4-coordinate. In both of these compounds, the
SBUs are connected into layers by the bdc ligands in a similar
manner to those in 3, and for {[NEt2H2][Zn3(bdc)4]·2.5DEF}∞
these layers are cross-linked by an additional bdc to form
For 2, the Cd · · · Cd · · · Cd angle is reduced to 157◦, and this
distortion from linearity occurs in tandem with there being two
l-j1O,j2O,Oꢀ carboxylates and only one l-j1O,j1Oꢀ carboxylate.
2468 | Dalton Trans., 2008, 2465–2474
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
◦
a
˚
Table 4 Selected bond lengths (A) and bond angles ( ) for 4
Cd(1)–O(1)
Cd(1)–O(2)
O(2)–Cd(1)–O(1)
2.4530(18)
2.391(2)
53.66(6)
Cd(1)–O(3)ꢀ
2.3143(15)
117.74(8)
O(3)ꢀ–Cd(1)–O(3)ꢀꢀ
a Primed atoms generated by the symmetry operation −x + 1/4, y− 1/4,
z− 1/4. Double primed atoms generated by the symmetry operation x +
3/4, −y + 3/4, z− 1/4.
a three-dimensional coordination network. The compound
{[Zn3(bdc)3(MeOH)4]·2MeOH}∞ also contains Zn3(O2CR)6
SBUs, but in this instance they are directly linked by the bdc
ligands into a three-dimensional network.14
Fig.
9
The
triply
interpenetrated
network
structure
of
{[NMe2H2]2[Cd(bdc)2]·2DMF}∞ 4. The asymmetric unit of
4 consists of half of a cadmium atom, located on a two-fold
rotational axis, one bdc anion, one NMe2H2 cation, in which
{[NMe2H2]2[Cd(bdc)2]·2DMF}∞ 4. Hydrogen atoms, dimethylammonium
cations and included DMF molecules have been omitted for clarity.
+
one of the carbons is equally disordered over two positions, and
one DMF molecule disordered over two positions. Selected bond
lengths and angles are given in Table 4.
The formation of 4 rather than 1 with poorer quality solvent
arises because of hydrolysis within the DMF.15 This hydrolysis
+
generates NMe2H2 cations, which template the formation of the
The bdc anions bridge between the cadmium centres to generate
a three-dimensional network structure. Each cadmium centre
in 4 is 6-coordinate, and bonded to two j1-carboxylates and
network observed in 4. In order to verify that a templating effect
was indeed occurring, [NMe2H2]Cl was added to the reaction
mixture. In these cases, compound 4 was observed as the only
product regardless of the quality of the DMF used. While the
hydrolysis of DMF is well established, an alternative way of
two j2-carboxylates (Fig. 8). There is some asymmetry in the
2
˚
j -coordination, with Cd(1)–O(1) 2.4530(18) A and Cd(1)–O(2)
1
˚
˚
+
2.391(2) A. For the j -coordination, Cd(1)–O(3) is 2.3143(15) A,
whereas the non-coordinated distance Cd(1) · · · Cd(4) is
generating NMe2H2 is via decarbonylation, and this can occur
when DMF is heated with base.16 Although the absence of added
base implies decarbonylation is not the only process at work,
it is possible that NMe2H formed through hydrolysis is able to
facilitate decarbonylation, and thus amplify the concentrations
˚
2.705(2) A.
Coordination to four carboxylate groups means that each
cadmium centre can be regarded as a flattened tetrahedral node.
The benzene rings of the bdc ligands link these nodes into a
diamondoid network which is triply interpenetrated, as shown in
Fig. 9. The network contains cadmium(II) centres and bdc anions
in the ratio 1 : 2, so consequently the network is anionic. This
charge is balanced by the presence of dimethylammonium cations,
which interact through hydrogen bonds with the carboxylate
+
of NMe2H2 . This may explain why mixtures of 1 and 4 have
not been observed. We have previously shown that the purity of
DMF is important in determining the isolated product from the
reaction of Zn(NO3)2·6H2O with H2bdc. These observations with
Cd(NO3)2·4H2O reveal that the effect is more general.
˚
˚
[Cd3(bdc)3(DEF)2] 5. The asymmetric unit of 5 consists of one
and a half cadmium centres, one and a half bdc anions and one
DEF molecule, which is disordered over two sites in a 70 : 30 ratio.
Selected bond lengths and angles are given in Table 5.
In a similar manner to the structures of 2 and 3, the structure of
5 consists of a Cd3(O2CR)6 trimer with three bridging carboxylate
groups between the central Cd(1) atom and each of the terminal
Cd(2) atoms. The central Cd(1) atom adopts a distorted octahedral
oxygen atoms [N(3) · · · O(4) 2.941 A, H(3A) · · · O(4) 2.21 A,
◦
˚
N(3)–H(3A) · · · O(4) 139 ; N(3) · · · O(3) 2.909 A, H(3A) · · · O(3)
◦
˚
2.30 A, N(3)–H(3A) · · · O(3) 125 ]. The DMF molecules are not
coordinated and are included as guests within the pores, where they
act as hydrogen bond acceptors to the dimethylammonium cations
˚
˚
[N(3) · · · O(5) 2.646 A, H(3B) · · · O(5) 1.80 A, N(3)–H(3B) · · · O(5)
158◦].
◦
a
˚
Table 5 Selected bond lengths (A) and bond angles ( ) for 5
Cd(1)–O(1)
2.1971(15) Cd(2)–O(4)ꢀꢀꢀ
2.3876(13) Cd(2)–O(5)
2.3005(15) Cd(2)–O(6)
2.2209(18) Cd(2)–O(7)
2.4453(13) Cd(2)–O(7A)
2.3649(14)
2.3043(13)
2.5560(15)
2.2741(16)
2.279(5)
Cd(1)–O(3)ꢀꢀ
Cd(1)–O(5)
Cd(2)–O(2)
Cd(2)–O(3)ꢀꢀ
2.203(16)
Cd(2)–O(4)ꢀꢀ
O(1)–Cd(1)–O(5)ꢀ
O(3)ꢀꢀ–Cd(1)–O(3)* 170.63(8)
168.17(7)
O(2)–Cd(2)–O(7)
174.33(16)
O(6)–Cd(2)–O(4)ꢀꢀ 176.21(7)
O(4)ꢀꢀ–Cd(2)–O(3)ꢀꢀ
53.92(4)
a Primed atoms generated by the symmetry operation −x, y, −z + 1/2.
Double primed atoms generated by the symmetry operation x, −y + 1, z +
1/2. Asterisked atoms generated by the symmetry operation −x, −y + 1,
−z. Triple primed atoms generated by the symmetry operation −x + 1/2,
y + 1/2, −z + 1/2.
Fig.
8
The cadmium environment in the structure of
{[NMe2H2]2[Cd(bdc)2]·2DMF}∞ 4.
This journal is The Royal Society of Chemistry 2008
Dalton Trans., 2008, 2465–2474 | 2469
©

View Article Online
geometry, with cis angles in the range 81.36(5)–101.11(10)◦. Cd(1)
lies on a 2-fold axis and the half bdc ions lie proximate to inversion
centres.
Compound 5 is formed in DEF regardless of the quality of
solvent used or the presence of added [NEt2H2]Cl. The insensitivity
+
of the reaction in DEF to the presence of NEt2H2 may be
Each Cd(2) atom is 7-coordinate, and bonded to two j2-
carboxylates, two j1-carboxylates and a DEF ligand. One of
the j1-carboxylates derives from a neighbouring trimer, and this
interaction links the trimers into infinite chains, as shown in
Fig. 10. The three structurally independent carboxylate groups
each adopt a different binding mode: that containing O(1) and
O(2) exhibits the j1O,j1Oꢀ-coordination mode, that containing
O(5) and O(6) adopts the j1O,j2O,Oꢀ-coordination mode, and that
containing O(3) and O(4) has the j1O,j2O,Oꢀ,j1Oꢀ coordination
mode.
a consequence of the larger size of this cation with respect to
NMe2H2 , rendering it incapable of being accommodated within
+
the pores of the interpenetrated anionic {Cd(bdc)2}n network
observed in 4.
[Cd(mbdc)(DMF)]∞ 6. Though the gross structure is unam-
biguous, there were problems in refinement of the model in the
crystal structure of 6, as noted in the Experimental section, which
prevents meaningful discussion of the bond length and bond angle
data. The asymmetric unit of 6 contains four cadmium centres,
four mbdc ligands and four molecules of DMF. Although there
are four independent cadmium centres, the differences between
them are minor. Each is bonded to four carboxylates and a
DMF ligand. The cadmium centres are connected into pairs by
two l-j1O,j2O,Oꢀ-carboxylate groups, and these pairs are linked
into chains by two further bridging carboxylate groups. Two
orientations of these chains are shown in Fig. 12.
Fig. 10 The cadmium carboxylate chains present in the structure of
[Cd3(bdc)3(DEF)2]∞ 5. Hydrogen atoms and minor components of the
disordered solvent molecules have been omitted for clarity.
The infinite chains are connected into a three-dimensional
network through the bdc ligands, and the gross structure contains
rhombohedral pores, as shown in Fig. 11, into which the coordi-
nated DEF ligands project. A coordination number of 7 for Cd(2)
occurs together with a distortion from linearity of the Cd3 unit,
as witnessed by a Cd · · · Cd · · · Cd angle of 144◦. It is notable
that a similar distortion was observed in 2, which also contained
rhombohedral pores. In both 2 and 5, the central cadmium in each
Cd3 unit is located on a two-fold axis.
Fig. 12 Two views of the cadmium carboxylate chains present in the
structure of [Cd(mbdc)(DMF)]∞ 6. The DMF ligands have been omitted
for clarity.
The mbdc ligands connect the chains into bilayers, as shown in
Fig. 13. The two sheets that comprise each bilayer are separated
˚
by approximately 3.3 A, whilst the distance between bilayers is
˚
greater (approximately 8.0 A) in order to accommodate the DMF
molecules.
Two cadmium–mbdc structures that do not contain N-donor
ligands have been previously reported. [Cd(mbdc)(H2O)]∞ forms
a two-dimensional structure in which mbdc ligands interlink
zigzag cadmium–carboxylate chains into layers.17 The compound
{(NH2Me2)2[Cd3(mbdc)4]}∞ also contains cadmium–carboxylate
chains, but in this case they are formed from the linking together
of Cd3(l-O2CR)6 units in a similar manner to that observed in 5.18
The chains are interconnected into a three-dimensional network
by the mbdc ligands. As with 4, the dimethylammonium cations
are derived from decomposition of the DMF solvent.
Thermal properties
Of 1. Analysis of 1 by TGA revealed that 1 remains unchanged
until 180 ◦C,◦then loses 11% of its mass by 220 ◦C, and a further
10% by 350 C. These two mass losses together account for the
coordinated DMF (calc. 20.3%). The desolvated material starts to
lose mass more rapidly as the temperature approaches 450 ◦C, and
Fig. 11 Part of the three-dimensional network present in the structure
of [Cd3(bdc)3(DEF)2]∞ 5. Hydrogen atoms and minor components of the
disordered solvent molecules have been omitted for clarity.
2470 | Dalton Trans., 2008, 2465–2474
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
precipitate in new structural forms,19 so it is possible that a similar
process is also occurring here.
Of 5. On heating, 5 remains unchanged until 200 ◦C, then loses
10% of its mass between 200 ◦C and 300 ◦C. This material is then
stable until 370 ◦C, when it loses a further 10% of its mass before
◦
400 C. Both of these steps correspond to the loss of one of the
coordinated DEF molecules (calc. 9.8% for each). The desolvated
material undergoes a further mass loss between 450 ◦C and 550 ◦C,
yielding CdO (experimental mass remaining 36%, calc. 37.3%).
Samples of 5 were heated to 300 ◦C and 400 ◦C, to generate
5a and 5b respectively. Microanalyses confirmed the identities
of these materials as Cd3(bdc)3(DEF) (5a) and Cd(bdc) (5b),
and the reduction of the observed percentage of nitrogen from
2.98% in 5 (calc. 2.71%) to 1.73% in 5a (calc. 1.51%) and 0% in
5b is particularly striking. Powder X-ray diffraction experiments
confirmed that the network structure of 5 was retained in 5a and
5b, though broadening of the peaks may indicate some loss of
crystallinity.
Of 6. Analysis of 6 by TGA revealed that it lost 20% of its
mass between 80 ◦C and 230 ◦C, then remained unaltered until
500 ◦C. The 20% mass loss corresponds to the coordinated DMF
molecules (calc. 20.3%). Between 530 ◦C and 570 ◦C, 6 underwent
a further mass loss to form CdO (experimental mass remaining
37%, calc. 36.7%). In contrast to the isomeric network 1, 6 loses
DMF in a single step.
Fig. 13 Top and side views of one of the bilayers present in the structure
of [Cd(mbdc)(DMF)]∞ 6.
by 550 ◦C it is converted into CdO (experimental mass remaining
36%, calc. 36.7%).
Separate samples of 1 were heated at 250 ◦C and 400 ◦C, to
generate 1a and 1b respectively. Microanalyses on 1a and 1b
confirmed their identities as [Cd(bdc)(DMF)0.5] (1a) and [Cd(bdc)]
(1b). Powder X-ray diffraction patterns revealed many of the peaks
for 1a and 1b were identical to those for 1, though additional
peaks of low intensity were present in both cases. This suggests a
new phase is formed as a minor product on heating, though the
majority of 1 loses coordinated DMF without destruction of the
network.
Gas adsorption properties of 5b
The powder X-ray diffraction experiments revealed that 5b
retained the three-dimensional network structure of 5 after loss
of the coordinated DEF molecules. Therefore, 5b should contain
rhombohedral pores, so was deemed a good candidate for gas
adsorption studies.
Experimental single-component adsorption isotherms on 5b
were obtained for nitrogen, krypton and hydrogen at 77 K. Exam-
ples of these data are shown in Fig. 14 and 15. From Fig. 14 it can
be seen that the adsorption isotherm for nitrogen is of Type I form.
Although it was not developed for microporous materials, the BET
model performs surprisingly well if the two consistency criteria are
taken into account:20 firstly, the pressure range is selected so that
the values v(P0 − P) are increasing with increasing P/P0, and
Of 4. Analysis of 4 by TGA revealed that on heating it remains
unchanged until 120 ◦C. A 22% mass loss between 120 ◦C and
◦
200 C corresponds to removal of the included DMF molecules
(calc. loss 21.5%). The desolvated network remains stable until
250 ◦C, then steadily loses mass until 500 ◦C, after which the mass
◦
loss is accelerated, finally yielding CdO at 550 C (experimental
mass loss 20%, calc. 18.9%).
Powder X-ray ◦diffraction patterns recorded after heating sam-
ples of 4 to 210 C were very different to those of 4, suggesting
that a substantial rearrangement of the network structure occurs
in conjunction with loss of the included solvent. Attempts to
exchange the included DMF molecules for more volatile guests
proved unsuccessful. Immersion of samples of 4 in chloroform or
acetone led to no changes in the TGA or powder X-ray diffraction
patterns after 3 d, suggesting no guest exchange had occurred. In
contrast, immersion of samples of 4 in methanol led to significant
changes in the powder X-ray diffraction patterns after only 10 min.
The peaks associated with 4 had diminished in intensity, whereas a
new set of peaks had appeared. After further immersion, these new
peaks increased in relative intensity. These observations suggest
that methanol converts 4 into another, as yet uncharacterised,
compound. A number of MOFs are known to dissolve and re-
Fig. 14 Adsorption isotherms for nitrogen (ꢀ) and krypton (᭜) on 5b
obtained at 77 K.
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 2465–2474 | 2471
©

View Article Online
secondly, the intercept of the BET plot must be positive. The BET
surface area calculated in this way is 13 m2 g−1.
A single component Langmuir isotherm can be fitted to the
adsorption isotherm for krypton (r2 = 0.9820). The monolayer
capacity and interaction energy parameters obtained from the fit
were not significantly affected by the choice of reference state
(liquid/solid) for krypton at 77 K. The apparent surface area of
5b calculated, using the customary value for the cross-sectional
area of krypton mentioned in the literature (0.202 nm2),21 from
the Langmuir monolayer capacity was 15 m2 g−1. Considering the
simplicity of the models used and uncertainties in the customary
molecular areas, this value is in good agreement with that obtained
for nitrogen.
The accessible surface area calculated in a geometric manner
from the crystallographic data of the perfect crystal provides
a useful tool to assess the quality of a synthesised sample.22
Calculation of the accessible surface area in the structure of 5b,
determined with a N2 sized probe molecule, suggests a much
higher theoretical value—735 m2 g−1—than that observed in the
nitrogen and krypton adsorption measurements. One possible
explanation for this is that partial structural collapse has occurred
on removal of the DEF ligands. Though there is no evidence for
another structure in the powder X-ray diffraction experiments,
these cannot rule out partial formation of an amorphous phase,
which is either non-porous or significantly less porous than 5b is
expected to be from the crystal structure of 5.
From Fig. 15 it can be seen that the hydrogen isotherm for
5b has a Type I form expected for adsorption in a microporous
solid and/or supercritical adsorption. There is marked hysteresis
between adsorption and desorption isotherms. This hysteresis has
been previously observed23 for MOF equilibrium isotherms, and
in the absence of kinetic data, was attributed to mass transfer
limitations. However, in our studies, data on the kinetics of
adsorption and desorption were obtained simultaneously with
the isotherms. These data demonstrated that each point in the
hydrogen adsorption and desorption isotherms in Fig. 15 reached
equilibrium rapidly, and were not associated with significant
temperature excursions. Hence, the observed hysteresis is not due
to mass transfer effects in this case.
Fig. 16 Simulation of hydrogen adsorption (ꢀ) and desorption (ꢁ) in 5b.
maximum amount of hydrogen adsorbed in the simulation is
24 mg g−1, compared to 4 mg g−1 experimentally. As with the
nitrogen adsorption data, the differences between the experimental
data and that predicted for 5b are consistent with partial collapse
of the porous 5b framework under the high temperature activation
conditions. As the simulations show that hydrogen saturation
loading, even in the perfect crystal (2.4 wt% and 36 g dm−3), is
far from the DOE 2010 targets for hydrogen storage (6 wt% and
45 g dm−3) 5b is not a promising material for hydrogen storage.
Consequently, we did not invest further time in trying to improve
the activation procedure of 5.
Conclusions
We have demonstrated that the outcome of the reaction be-
tween Cd(NO3)2·4H2O and H2bdc in DMF is very depen-
dent on the synthetic conditions, with temperature, pres-
sure and solvent quality all factors that affect the nature
of the observed product(s). Four new coordination networks
have been structurally characterised from this reaction, with
[Cd(bdc)(DMF)]∞ 1 and {(NMe2H2)2[Cd(bdc)2]·2DMF}∞
4
adopting three-dimensional networks while the two supramolec-
ular isomers of [Cd3(bdc)3(DMF)4]∞ (2 and 3) adopt two-
dimensional arrays.
Simulation of hydrogen adsorption/desorption in 5b does not
replicate the hysteresis observed in the experiments, and reveals
the saturation loading is anticipated at approximately 10 bar
(Fig. 16), compared with 0.5 bar as observed. In addition, the
The effect of temperature on MOF synthesis is well
documented,5 and solvothermal conditions are widely used to
prepare MOFs. In this work, the increase in pressure that led to
formation of 3 was not undertaken to access a higher temperature.
Why the change in pressure should lead to a change in product is
not yet well understood, but it is noteworthy that the calculated
density of 3 (1.809 g cm−3) is marginally greater than that of its
supramolecular isomer 2 (1.781 g cm−3).
While it is well known in organometallic chemistry that solvent
quality is important in determining the outcome of a reaction,
this principle is less established in MOF chemistry. A degree of
hydrolysis in the DMF solvent can lead to NMe2H2+ cations, which
template the formation of the anionic network observed in 4. We
have previously noted this effect in the analogous zinc chemistry.6
The reaction between Cd(NO3)2·4H2O and H2bdc in DEF is
much more robust than that in DMF, yielding [Cd(bdc)(DEF)]∞
5 as the only product, regardless of the temperature, pressure or
solvent quality. This is in contrast to the analogous zinc chemistry,
Fig. 15 Adsorption (ꢀ) and desorption (᭜) isotherms for hydrogen on
5b obtained at 77 K.
+
+
in which NEt2H2 plays a similar role to NMe2H2 . Compound
2472 | Dalton Trans., 2008, 2465–2474
This journal is The Royal Society of Chemistry 2008
©

View Article Online
5 loses the coordinated DEF without change in the powder X-
ray diffraction pattern on heating to 400 C, and the desolvated
network reversibly adsorbs hydrogen with rapid adsorption and
desorption kinetics. However, the amount of hydrogen uptake is
lower than anticipated on the basis of simulations, suggesting that
partial framework collapse to an amorphous phase has occurred.
and the mixture maintained at 115 ◦C for a further 5 h. Colourless
crystals of 4 were separated by filtration. Yield 0.27 g (24%).
m(CO2)/cm−1 1675s, 1558s. Found C, 45.1; H, 5.40; N, 6.52%.
Calc. for C23H31CdN3O9 (4 − DMF, 605.9): C, 45.6; H, 5.16; N,
6.93%.
◦
Synthesis of [Cd3(bdc)3(DEF)2]∞ (5)
Experimental
Cd(NO3)2·4H2O (0.086 g, 0.279 mmol) and H2bdc (0.045 g, 0.271
mmol) were dissolved in DEF (10 cm3), and the solution was
heated at 115 ◦C for 5 h. After this time, yellow crystals were
separated by filtration. Yield 0.039 g (41%). Found C, 39.2; H,
3.52; N, 2.98%. Calc. for C34H34Cd3N2O14 (1031.9): C, 39.6; H,
3.32; N, 2.71%.
Microanalysis (C, H and N) were carried out by Mr Alan Carver
(University of Bath Microanalytical Service). IR spectra were
recorded as KBr discs on a Nicolet Nexus FTIR spectrometer.
Powder diffraction data were recorded on a Bruker D8 diffrac-
tometer fitted with Goebal mirrors and a 0.2 mm beam slit, and
in each case, unless noted in the text, the experimental pattern
showed a good match with that simulated from the single crystal
analysis. Thermogravimetric analyses were performed on a Perkin
Elmer TGA 7 instrument from 30 ◦C to 600 ◦C at a heating
Synthesis of [Cd(mbdc)(DMF)]∞ (6)
Cd(NO3)2.4H2O
(0.500
g,
1.62
mmol)
and
1,3-
benzenedicarboxylic acid (H2mbdc) (0.539 g, 3.24 mmol)
were dissolved in DMF (60 cm3), and the solution heated at
90 ◦C for 5 h. Colourless crystals formed on cooling to room
temperature, and these were separated by filtration. Yield 0.255 g
(45%). Found C, 37.8; H, 3.26; N, 4.02%. Calc. for C11H11CdNO5
(349.6): C, 37.8; H, 3.17; N, 4.01%.
◦
rate of 10 C min−1 under a flow of N2. Nitrogen adsorption
measurements were carried out using a Micromeritics ASAP
2400 instrument, while hydrogen and krypton adsorption mea-
surements were conducted using a Hiden intelligent gravimetric
analyser (IGA).
Crystallography
Synthesis of [Cd(bdc)(DMF)]∞ (1)
Single crystals of compounds 2, 3, 4 and 6 were analysed using
a Nonius Kappa CCD diffractometer and Mo-Ka radiation (k =
Cd(NO3)2·4H2O (0.520 g, 1.69 mmol) and H2bdc (0.501 g,
3.02 mmol) were added to DMF (60 cm3) and the mixture was
stirred at 115 ◦C until a colourless solution was obtained. Stirring
was then stopped, and the mixture maintained at 115 ◦C for a
further 5 h. Colourless crystals of 1 were separated by filtration.
Yield 0.240 g (41%). m(CO2)/cm−1 1653s. Found C, 37.8; H, 3.15;
N, 4.05%. Calc. for C11H11CdNO5 (349.6): C, 37.8; H, 3.17; N,
4.01%.
˚
0.71073 A). Data for 1 and 5 were collected at Daresbury Station
˚
˚
9.8 (k = 0.68970 A and 0.68680 A, respectively). Details of the
data collections, solutions and refinements for 1–5 are given in
Table 6. The structures were universally solved using SHELXS-
97 and refined using full-matrix least squares in SHELXL-97.24
Multiscan absorption corrections were applied throughout and
convergence was uneventful, with the following exceptions and
points of note.
Synthesis of [Cd(bdc)(DMF)]∞ (1) and [Cd3(bdc)3(DMF)4]∞ (2)
The asymmetric unit of 4 consists of half of a cadmium atom
(located on a special position), one bdc anion, one dimethylam-
monium cation, in which one of the carbons is equally disordered
over two positions, and one DMF molecule disordered over two
positions.
The asymmetric unit of 5 consists of one whole cadmium, half
of a cadmium located on a special position, one whole bdc, half of
a bdc (with the remainder generated via a space group symmetry
operation) and one whole DEF molecule. The solvent molecule is
disordered over two sites in a 70 : 30 ratio.
The procedure described above to prepare 1 was repeated with
the temperature maintained at 95 ◦C for 5 h, rather than
115 ◦C. Colourless crystals of 1 and 2 were separated from the
solvent by filtration, but could not be separated from each other.
Combined yield 0.415 g. X-Ray powder diffraction data revealed
1 to be the major product.
Synthesis of [Cd3(bdc)3(DMF)4]∞ (3)
Cd(NO3)2·4H2O (0.520 g, 1.69 mmol) and H2bdc (0.501 g,
3.02 mmol) were dissolved in DMF (60 cm3) at 115 ◦C in a Bu¨chi
miniclave reactor, which was then sealed and pressurised with 3
bar of nitrogen. The solution was heated for 5 h, then colourless
crystals of 3 were separated by filtration. Yield 0.515 g (82%).
m(CO2)/cm−1 1653s, 1539s. Found C, 38.5; H, 3.34; N, 4.39%.
Calc. for C36H40Cd3N4O16 (1122.0): C, 38.5; H, 3.59; N, 4.99%.
Several datasets were collected for 6, of which the optimum was
at 250 K. Although the sample appeared single under a polarising
microscope, a data collection at 150 K failed to solve/refine
properly using any intelligent model. Initially it was assumed
that the crystal was extremely twinned, until it was noted that
the diffraction pattern intensities altered perceptibly at the higher
temperature. Borderline absence conditions resulted in attempts
to solve and refine in many space groups, although no one
space group could satisfy all absence conditions. Notwithstand-
ing this, the refinement was adequate to provide unambiguous
determination of the gross structure, though not good enough to
determine metric parameters. Selected crystallographic data for 6:
C11H11CdNO5, M = 349.61, monoclinic, P1121/b, a = 13.1670(4),
Synthesis of {(NMe2H2)2[Cd(bdc)2]·2DMF}∞ (4)
1,4-Benzenedicarboxylic acid (H2bdc) (0.501 g, 3.02 mmol) was
dissolved in DMF (50 cm3) at 115 ◦C, and a solution containing
Cd(NO3)2·4H2O (0.520 g, 1.69 mmol) and dimethylammonium
chloride (0.080 g, 0.98 mmol) in DMF (10 cm3) was added. The
reaction mixture was stirred for 5 min, then stirring was stopped
◦
˚
b = 17.5870(5), c = 22.0960(7) A, a = 90, b = 90, c = 90.000(1) .
Dalton Trans., 2008, 2465–2474 | 2473
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Table 6 Crystallographic data for compounds 1–5
1
2
3
4
5
Formula
M
C11H11CdNO5
349.61
150(2)
Triclinic
¯
P1
7.4523(9)
9.7764(12)
9.8469(12)
118.255(2)
102.118(2)
91.358(2)
611.51(13)
2
1.797
8031
4205
0.0428
C36H40Cd3N4O16
1121.92
150(2)
Monoclinic
C2/c
28.1090(4)
9.1930(1)
17.8780(2)
90
115.112(1)
90
4183.12(9)
4
1.585
C36H40Cd3N4O16
1121.92
150(2)
Triclinic
¯
P1
10.1270(1)
12.6120(1)
16.6790(2)
75.459(1)
88.924(1)
87.271(1)
2059.65(4)
2
1.610
46686
12980
0.0433
C26H38CdN4O10
679.00
150(2)
Orthorhombic
Fdd2
9.7830(2)
21.2010(4)
27.8110(5)
90
90
90
5768.26(19)
8
0.819
15349
4207
C34H34Cd3N2O14
510.91
150(2)
Monoclinic
C2/c
25.0960(11)
10.7885(5)
17.5107(8)
90
128.996(3)
90
3684.6(3)
4
1.785
T/K
Crystal system
Space group
˚
a/A
˚
b/A
˚
c/A
a/◦
b/◦
c /◦
U/A
Z
3
˚
l(Mo-Ka)/mm−1
Reflections collected
Independent reflections
R(int)
29343
4776
0.0393
18981
5385
0.0376
0.0393
R1, wR2 [I>2r(I)]
R indices (all data)
Largest diff. peak, hole/e A
0.0372, 0.0965
0.0391, 0.0972
0.893, −0.757
0.0208, 0.0450
0.0244, 0.0463
0.679, −0.538
0.0265, 0.0605
0.0332, 0.0635
1.424, −1.301
0.0225, 0.0494
0.0268, 0.0509
0.222, −0.762
0.0255, 0.0673
0.0282, 0.0691
0.659, −0.576
−3
˚
Simulations
7 M. Edgar, R. Mitchell, A. M. Z. Slawin, P. Lightfoot and P. A. Wright,
Chem.–Eur. J., 2001, 7, 5168.
The multi-purpose simulation code music25 was used to calculate
the hydrogen adsorption isotherms with grand canonical Monte
Carlo (GCMC) simulations. All interatomic interactions were
modelled with a standard Lennard-Jones potential. Hydrogen
was represented by a two-centred model.26 The parameters for
the framework were taken from the UFF force field27 with the
positions of the framework atoms fixed at their crystallographic
positions. Lorentz–Bertholot mixing rules were used to calculate
8 H. Li, M. Eddaoudi, T. L. Groy and O. M. Yaghi, J. Am. Chem. Soc.,
1998, 120, 8571.
9 M. Eddaoudi, J. Kim, N. Rosi, D. Vodak, J. Wachter, M. O’Keeffe
and O. M. Yaghi, Science, 2002, 295, 469; N. L. Rosi, J. Eckert, M.
Eddaoudi, D. T. Vodak, J. Kim, M. O’Keeffe and O. M. Yaghi, Science,
2003, 300, 1127.
10 X. Shi, G. S. Zhu, Q. R. Fang, G. Wu, G. Tian, R. W. Wang, D. L.
Zhang, M. Xue and S. L. Qiu, Eur. J. Inorg. Chem., 2004, 185.
11 B. Paul, B. Zimmermann, K. M. Fromm and C. Janiak, Z. Anorg. Allg.
Chem., 2004, 630, 1650.
12 J.-C. Dai, X.-T. Wu, S.-M. Hu, Z.-Y. Fu, J.-J. Zhang, W.-X. Du, H.-H.
Zhang and R.-Q. Sun, Eur. J. Inorg. Chem., 2004, 2096.
13 J.-H. Liao and W.-C. Huang, Inorg. Chem. Commun., 2006, 9, 1227.
14 H. L. Li, C. E. Davis, T. L. Groy, D. G. Kelley and O. M. Yaghi, J. Am.
Chem. Soc., 1998, 120, 2186.
15 J. Juillard, in Recommended Methods for Purification of Solvents and
Tests for Impurities, ed. J. F. Coetzee, Pergamon, Oxford, UK, 1982, p.
32.
˚
hydrogen/framework parameters. Interactions beyond 12.8 A
were neglected. For each point of the isotherm, two million GCMC
steps were performed with 800 000 taken for equilibration and
1 200 000 to sample the data. The excess amount adsorbed was
calculated as outlined previously.28
16 A. C. Sudik, A. R. Millward, N. W. Ockwig, A. P. Coˆte´, J. Kim and
O. M. Yaghi, J. Am. Chem. Soc., 2005, 127, 7110.
17 A. Thirumurugan and C. N. R. Rao, J. Mater. Chem., 2005, 15, 3852.
18 N. L. Rosi, J. Kim, M. Eddaoudi, B. L. Chen, M. O’Keeffe and O. M.
Yaghi, J. Am. Chem. Soc., 2005, 127, 1504.
19 A. N. Khlobystov, N. R. Champness, C. J. Roberts, S. J. B. Tendler, C.
Thompson and M. Schro¨der, CrystEngComm, 2002, 4, 426.
20 K. S. Walton and R. Q. Snurr, J. Am. Chem. Soc., 2007, 129, 8552;
J. Rouquerol, P. Llewellyn and F. Rouquerol, “Is the BET equation
applicable to microporous adsorbents?”, in Characterization of porous
solids VII, ed. P. L. Llewellyn, F. Rodriquez-Reinoso, J. Rouquerol and
N. Seaton, Elsevier, Aix-en-Provence, 2007, p. 49.
21 F. Rouquerol, J. Rouquerol and K. Sing, Adsorption by Powders and
Porous Solids: Principles, Methodology and Applications, Academic
Press, London, 1999.
Acknowledgements
The authors thank the EPSRC for financial support and for
support of the Chemical Database Service, and Fernando Acosta
for technical support with the gas adsorption measurements.
References
1 M. Eddaoudi, D. B. Moler, H. L. Li, B. L. Chen, T. M. Reineke, M.
O’Keeffe and O. M. Yaghi, Acc. Chem. Res., 2001, 34, 319.
2 S. L. James, Chem. Soc. Rev., 2003, 32, 276; S. Kitagawa, S. Noro and T.
Nakamura, Chem. Commun., 2006, 701; S. Bourrelly, P. L. Llewellyn,
C. Serre, F. Millange, T. Loiseau and G. Fe´rey, J. Am. Chem. Soc., 2005,
129, 13519.
22 T. Du¨ren, F. Millange, G. Fe´rey, K. S. Walton and R. Q. Snurr, J. Phys.
Chem. C, 2007, 111, 15350.
3 S. Kitagawa, R. Kitaura and S. Noro, Angew. Chem., Int. Ed., 2004, 43,
2334; J. L. C. Rowsell and O. M. Yaghi, Angew. Chem., Int. Ed., 2005,
44, 4670; A. J. Fletcher, K. M. Thomas and M. J. Rosseinsky, J. Solid
State Chem., 2005, 178, 2491.
4 U. Mueller, M. Schubert, F. Teich, H. Puetter, K. Schierle-Arndt and J.
Pastre, J. Mater. Chem., 2006, 16, 626; S.-H. Cho, B. Ma, S. T. Nguyen,
J. T. Hupp and T. E. Albrecht-Schmitt, Chem. Commun., 2006, 2563.
5 P. M. Forster, A. R. Burbank, C. Livage, G. Fe´rey and A. K. Cheetham,
Chem. Commun., 2004, 368.
23 J. Lee, J. Li and J. Jagiello, J. Solid State Chem., 2005, 178, 2527.
24 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 1990,
46, 467; G. M. Sheldrick, SHELXL-97, University of Go¨ttingen,
Germany, 1997.
25 A. Gupta, S. Chempath, M. J. Sanborn, L. A. Clark and R. Q. Snurr,
Mol. Simul., 2003, 29, 29.
26 Q. Yang and C. Zhong, J. Phys. Chem. B, 2005, 109, 11862.
27 A. K. Rappe´, C. J. Casewit, K. S. Colwell, W. A. Goddard and W. M.
Skiff, J. Am. Chem. Soc., 1992, 114, 10024.
6 A. D. Burrows, K. Cassar, R. M. W. Friend, M. F. Mahon, S. P. Rigby
and J. E. Warren, CrystEngComm, 2005, 7, 548.
28 T. Du¨ren, L. Sarkisov, O. M. Yaghi and R. Q. Snurr, Langmuir, 2004,
20, 2683.
2474 | Dalton Trans., 2008, 2465–2474
This journal is
The Royal Society of Chemistry 2008
©