2
J. Xu et al. / Journal of Organometallic Chemistry xxx (2014) 1e7
the single net (Fig. 1d). It is noteworthy that there is still free space
(
Fig. 1c) which was calculated by the PLATON [31] analysis of 49.3%
3
(3992.4 Å per unit) filled by DMF molecules and its decomposed
þ
product (C
2
NH
8
).
Stability and gas sorption property of complex 1
Stability information about the porous structure is required and
the results of the TGA analysis not only indicate the thermal sta-
bility of the MOF, but also turn out to be decisive data to ensure an
optimized temperature of activation to remove the guest mole-
cules. For complex 1, the TGA data revealed a steady weight loss
ꢂ
between about 50 and 250 C, which correspond to the loss of the
þ
solvent DMF molecules (calcd. 20.38%, found 20.45%) and C
2
NH
8
cations (calcd. 5.71%, found 5.85%). No weight loss then occurred
ꢂ
until approximately 330 C, when the entire structure starts to
decompose as shown in Fig. 3. According to the TGA results, the
ꢂ
framework 1 is stable even up to 330 C. Before the gas adsorption
tests, the activation for removing the guest molecules from the
pores by heating in a vacuum is a required pre-treatment and the
as-synthesized crystal samples were heated at the optimized
ꢂ
temperature of 100 C for 20 h. The adsorption isotherm for 1 is
shown in Fig. 3 inset, in which complex 1 shows a comparatively
3
ꢁ1
high adsorption amount of CO
at STP) corresponding to 1 CO
uptake is nearly 4 times higher than that of N
at STP). The Langmuir and BET surface areas estimated from the CO
2
at 195 K and 1 atm. (40.13 cm g
2
molecule per formula unit. The CO
2
3
ꢁ1
2
at 77 K (9.17 cm g
2
2
ꢁ1
adsorption isotherm are 119.73 and 85.64 m g , respectively. The
high affinity for CO of 1 may be due to the unsaturated metal sites
Scheme 1.
2
in 1 associated with the pore structure as revealed in other MOF
materials [32,33]. Thus, complex 1 may have a potential application
in the separation of CO /N mixtures and it is possible to adsorb
2 2
Results and discussion
other molecules in the cavity or on the surface of its framework.
Description of crystal structures
Complex {Zn
X-ray diffraction analysis revealed that 1 crystallizes in the ortho-
rhombic form with the space group of Pna2 and the asymmetric
unit has a Ru(II) atom, two Zn(II) atoms, four 4,4 -biphenyldi-
Ru
4 2
(bpdc)
4
$4C
2
NH
8
$9DMF}
n
(1). The results of the
Preparation and characterization of Ru@MOF and B12eRu@MOF
3 2
We used the MOF as a supporter to immobilize [Ru(II) (bpy) ]Cl
and B12 complexes (Scheme 1) by a cation exchange reaction. This
cation or anion exchange of MOFs is useful method for function-
1
0
2
ꢁ
þ
carboxylate ligands (bpdc ), four C
2
NH
8
(protonated dimethyl
amine) molecules decomposed by the DMF solvent and nine DMF
alizing MOFs [34e42]. First, we tried to absorb [Ru(II) (bpy)
form Ru@MOF. The color of the [Ru(II) (bpy) ]Cl ethanol solution
disappeared by stirring with MOF 1, and the UVevis spectra of the
3 2
]Cl to
molecules. The bpdc2 ligands have three coordination modes
Scheme 1). Every Ru(II) center with an octahedral coordination
geometry is six-coordinated by six oxygen atoms from six different
ꢁ
3
2
(
2
þ
solution proved the adsorption of [Ru(II) (bpy)
3
]
in the cavity of
2
ꢁ
bpdc ligands (four (b) and two (c) coordinated modes). Crystal
view of 1 is shown in Fig. 1. The schematic drawings of 1 based on
the crystal structure are also shown in Fig. 2. And as shown in
Fig. 1a, two Zn(II) atoms are five-coordinated with a distorted
trigonal bipyramid coordination geometry by five oxygen atoms
the MOF (Fig. 4). The IR spectra showed that the surface of the
Ru@MOFs is the same as MOF though the two samples have
different colors (Fig. S1). It was calculated from the ICP-MS data
that molar ratio of Ru/MOF is 3:1. The cation exchange efficiency of
þ
Ru photosensitizer considered as {[Ru(bpy)
as 75% in Ru@MOF.
3
]Cl} form is estimated
2
ꢁ
from four different bpdc ligands (one (a), two (b) and one (c)
coordinated modes). The RueO and ZneO bond lengths are in the
range of 2.049(1)e2.109(9) Å and 1.909(1)ꢁ1.970(1) Å, respectively
as listed in Table 2.
Next, it is very important to adsorb the B12 complex in order to
construct the B12eRu@MOF as a catalysis system. We used a hep-
tamethyl cobyrinate perchlorate for the B12 complex (Scheme 1).
The UVevis spectra illustrated that the heptamethyl cobyrinate was
adsorbed by Ru@MOF (Fig. 5). The evidence for the adsorption of
the B12 complex onto Ru@MOF was obtained by an ESR analysis,
and the well-defined Co(II) low spin signal typical for heptamethyl
cobyrinate was observed as shown in Fig. 6 [43]. The amount of B12
molecules was also calculated from the ICP-MS analysis of which
the molar ratio of B12/MOF is 1:15. Given the size of the B12 and the
cavity size of Ru@MOF, the number of adsorbed B12 molecules were
By carefully observing the structure, the connections of the
2
ꢁ
central ZneRueZn and linear linker of the bpdc ligand make the
complex have a 3D framework with large channels as illustrated in
Fig. 1b. Each ZneRueZn unit acts as an 8-connecting node, and the
topology can be described as hex net with the short Schl €a fli sym-
6
18
3
bols of (3 .4 .5 .6) [29,30]. There is a large void in the single net of
the complex as mentioned above, the open channel sizes along the
[
010] direction are ca.1.4 nm ꢀ 1.4 nm (Fig.1b), and according to the
2
þ
famous Aristotle's observation that “nature abhors a vacuum”,
interpenetration inevitably occurred. We found that there were
two individual 3D hex nets interpenetrating due to the large void in
smaller than that of [Ru(bpy)
3
]
. Actually, when we first adsorbed
B
12 complex in the MOF, the obtained B12-MOF did not adsorb
2þ
[Ru(bpy)
3
]
. Thus, it is likely that B12 complex is existed around
2
þ
the surface of MOF, whereas small [Ru(bpy)
3
]
is existed in cavity