Syntheses, structures and luminescent properties of decorated lanthanide metal-organic frameworks of (E)-4,4′-(ethene-1,2-diyl)dibenzoic acids

Article abstract of DOI:10.1039/c0ce00750a

The functionalized ligand (E)-4,4′-(1,4-bis(methylthio)but-2-ene-2,3- diyl)dibenzoic acid (H₂L_S) and four metal-organic frameworks, [M₂(L_H)₃(DMSO)₄] ·xDMSO (6 and 7 for M = Tb³⁺,

Full text of DOI:10.1039/c0ce00750a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/crystengcomm | CrystEngComm
Syntheses, structures and luminescent properties of decorated lanthanide
metal-organic frameworks of (E)-4,40-(ethene-1,2-diyl)dibenzoic acids†
*
Yu Li and Datong Song
Received 30th August 2010, Accepted 5th November 2010
DOI: 10.1039/c0ce00750a
The functionalized ligand (E)-4,40-(1,4-bis(methylthio)but-2-ene-2,3-diyl)dibenzoic acid (H2LS) and
four metal-organic frameworks, [M2(LH)3(DMSO)4]$xDMSO (6 and 7 for M ¼ Tb3+, Eu3+
respectively; LH ¼ (E)-4,40-(ethene-1,2-diyl)dibenzoate) and [M2(LS)3(DMF)2(H2O)2] (8 and 9 for M ¼
Eu3+, Tb3+) have been synthesized and characterized. All four MOFs adopt the anticipated
connectivity, but the fold of interpenetration in each MOF crystal varies as a result of different solvent
inclusion and ligand functionality. Compounds 6 and 7 contain a large amount of solvent molecules
and tend to lose solvent in air. Compounds 8 and 9 are stable crystals without lattice solvent molecules
in the crystals; the coordinating solvent molecules can be reversibly removed and replenished with the
skeletons intact. All four MOFs show mainly the ligand-based luminescence, while 8 also displays
metal-based luminescence.
dangling functional groups in the pores or channels and
Introduction
exploring their potential applications as sensors. Although the
porosity is not always predictable, there are some general rules to
follow and adjustments to existing examples usually produce
similar structures.11–13 Functional groups inside the pores/chan-
nels of the MOF serve as the binding site for guests, which can be
incorporated either via ligand synthesis prior to MOF assem-
bly4c,14 or post-synthetic modification.15
A large number of metal-organic frameworks (MOFs) with
different functionalities have been reported in the past decade.
Because of their porosity and various inherent properties from
ligands and metal centers involved, these MOFs have found use
in many areas, including gas storage1,2 and separation,2,3
sensing,4–8 and heterogeneous catalysis.9,10 While much effort has
been put into the syntheses of MOFs for gas storage and sepa-
ration, MOF sensors remain relatively underdeveloped. The
known MOF sensors can be classified into two categories with
respect to the guest interacting sites: (1) exposed metal centers
and (2) exposed functional groups on the ligands. Generating
unsaturated metal sites in MOF is more challenging and usually
achieved by serendipity, because of the highly reactive nature of
coordinately unsaturated metal centers. The second strategy can
be utilized to rationally design MOF sensors, because a large
variety of functional groups can be incorporated into the ligand
via relatively simple organic synthesis. However, MOFs sensors
in this category are still rare.4c
We design our functionalized MOFs based on two known
unfunctionalized examples: [Tb2(ADB)3(DMSO)4]$16DMSO
(where ADB ¼ 4,40-azodibenzoate) by Yaghi and co-workers16
and [Eu2L3(DMSO)2(CH3OH)2]$2DMSO (where L ¼ 4,40-
ethyne-1,2-diyldibenzoate) by our group.17 By replacing the
N]N and C^C from the ligands in the known MOFs with
a C]C linkage, different functional groups can be installed in
the central region of the linear ligand. We envision that a similar
skeleton could be anticipated when a functionalized ligand is
used, i.e., each bimetallic node is connected with 6 other nodes to
form 3D porous frameworks with triangular channels along the
body diagonal and functional groups dangling inside the chan-
nels (Scheme 1). It is worth noting that although stilbene dicar-
boxylate has been used in MOF synthesis with transition
metals,18 neither stilbene dicarboxylate nor its derivatives have
not been used in lanthanide MOFs. We chose thioether func-
tional groups for two reasons: (1) the soft sulfur donors do not
interfere with the assembly of lanthanide MOFs; (2) sulfur
Lanthanide MOFs are attractive for sensing applications,
because of not only the characteristic emissions but also the
unique luminescence mechanism, ligand-to-metal energy transfer
(LMET). The interactions between an analyte and functional
groups of the ligands may facilitate or disrupt the LMET
process, causing luminescence signal modulations as the report-
ing mechanism. With such a design in mind, our group is inter-
ested in preparing luminescent porous lanthanide MOFs with
Davenport Chemical Research Laboratories, Department of Chemistry,
University of Toronto, 80 St George Street, Toronto, Ontario, Canada
M5S 3H6. E-mail: dsong@chem.utoronto.ca; Fax: +1-416-978-7013;
Tel: +1-416-978-7014
† Electronic supplementary information (ESI) available: Cif files for
2a-9; TGA results of 6-9 and Na2LH; luminescence spectra of H2LH,
H2LS, and compounds 6-9; powder XRD patterns of 6-9; TGA, FTIR
and powder XRD spectra associated with the desolvation/resolvation
cycle. CCDC reference numbers 791231–791239. For ESI and
crystallographic data in CIF or other electronic format see DOI:
10.1039/c0ce00750a.
Scheme 1
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 1821–1830 | 1821

View Article Online
donors may potentially bind with soft metal ions that might be
able to enter the channels of the targeted MOFs, triggering
luminescence signal change. In this paper, we reported the
synthesis of the substituted stilbene dicarboxylic acid ligand with
thioether side chains (H2LS, (E)-4,40-(1,4-bis(methylthio)but-2-
ene-2,3-diyl)dibenzoic acid), the preparation and luminescent
properties of the Eu(III) and Tb(III) MOFs of H2LS and H2LH
((E)-4,40-stilbene dicarboxylic acid).
127.8, 52.2, 21.5. Anal. calcd for C20H20O4: C, 74.06; H, 6.21.
Found: C, 74.13; H, 6.24%. 2a: 1H NMR (CDCl3, 400 MHz, 25
ꢀC) d 8.06 (d, J ¼ 8.0 Hz, 4H), 7.34 (d, J ¼ 8.0 Hz, 4H), 3.93 (s,
6H), 1.88(s, 6H); 13C NMR (CDCl3, 100 MHz, 25 C) d 167.2,
ꢀ
149.2, 133.5, 129.9, 128.5, 52.3, 22.4.
Synthesis of dimethyl 4,40-(1,4-dibromobut-2-ene-2,3-
diyl)dibenzoate (3a trans, 3b cis)
A mixture of 2a and 2b (648 mg, 2 mmol) as obtained above, N-
bromosuccinimide (712 mg, 4 mmol) and benzoyl peroxide (97
mg, 0.4 mmol) were refluxed in cyclohexane (30 mL) for 5 h. The
reaction mixture was cooled to ambient temperature and filtered.
After the removal of solvent from the filtrate, the residual crude
product was purified using silica gel column chromatography
(eluted with 1 : 7 mixture of EtOAc : hexanes), affording color-
less crystals of 3a (460 mg, 48% yield) and 3b (141 mg, 15% yield).
Experimental section
General Information
Elemental analyses were performed in our Chemistry Depart-
ment on a PE 2400 C/H/N/S analyzer. Thermogravimetric
analyses (TGA) were performed on a TA Instruments SDT Q600
instrument under a dinitrogen atmosphere with a heating rate of
10 ꢀC per minute. NMR spectra were recorded on a Varian 400,
1
ꢀ
3a: H NMR (CDCl3, 400 MHz, 25 C) d 8.15 (d, J ¼ 8.4 Hz,
4H), 7.54 (d, J ¼ 8.4 Hz, 4H), 4.00 (s, 4H), 3.96(s, 6H); 13C NMR
(CDCl3, 100 MHz, 25 ꢀC) d 166.8, 142.6, 139.1, 130.2, 128.7,
52.5, 34.6. Anal. calcd for C20H18O4Br2: C, 49.82; H, 3.76.
Found: C, 49.93; H, 3.95%. 3b: 1H NMR (CDCl3, 400 MHz, 25
ꢀC) d 7.92 (d, J ¼ 8.4 Hz, 4H), 7.14 (d, J ¼ 8.4 Hz, 4H), 4.49 (s,
or a Bruker Avance 400 spectrometer. Both H and 13C NMR
1
spectra were referenced and reported relative to the solvent’s
residual signals. Photoluminescence spectra were measured using
a SPEX Fluorolog-3 spectrofluorometer (Jobin Yvon/SPEX,
Edison, New Jersey). The powder XRD experiments were per-
formed on an automated Siemens/Bruker D5000 diffractometer
equipped with a high power line focus Cu-Ka source operating at
50 kV/35 mA. A solid-state Si/Li Kevex detector was used for
removal of Kb lines. The diffraction patterns were collected on
a q/2ꢁq Bragg–Brentano reflection geometry with fixed slits. A
step scan mode was used for data acquisition with a step size of
0.02ꢀ 2ꢁq. Unless otherwise stated, all manipulations were per-
formed in air and all reagents were purchased from commercial
sources and used without further purification. McMurry
coupling reaction was performed under an Ar atmosphere, but
the work-up was carried out under ambient conditions. The THF
solvent for McMurry coupling was purified by John Morris
Scientific IT PureSolv PS-MD-6 system; Zn–Cu couple19 and
methyl 4-acetylbenzoate20 for the coupling reaction was prepared
according to the literature procedures.
4H), 3.87(s, 6H); 13C NMR (CDCl3, 100 MHz, 25 C) d 166.5,
ꢀ
143.8, 140.1, 129.5, 129.2, 52.2, 31.1. Anal. calcd. for
C20H18O4Br2$0.75CH2Cl2: C, 45.66; H, 3.60. Found: C, 45.68;
H, 3.52%.
Synthesis of (E)-dimethyl 4,40-(1,4-bis(methylthio)but-2-ene-2,3-
diyl)dibenzoate (4)
Sodium methanethiolate (108 mg, 1.54 mmol) and 3a (300 mg,
0.62 mmol) were refluxed in methanol (30 mL) under N2 for 3 h.
After the solvent was removed in vacuo, distilled water was added
to dissolve the residual solid. The aqueous solution was acidified
with excess 1 M HCl solution and extracted with EtOAc. The
organic phase was then dried over MgSO4. After filtration the
solution was concentrated to dryness, affording 4 as a light
yellow powder (231 mg, 95% yield). 1H NMR (CDCl3, 400 MHz,
25 ꢀC) d 8.10 (d, J ¼ 8.4 Hz, 4H), 7.44 (d, J ¼ 8.4 Hz, 4H), 3.94 (s,
6H), 3.29 (s, 4H), 1.83(s, 6H); 13C NMR (CDCl3, 100 MHz, 25
ꢀC) d 166.8, 144.8, 137.0, 129.6, 129.3, 129.1, 52.2, 38.8, 16.0.
Anal. calcd for C22H24O4S2: C, 63.43; H, 5.81. Found: C, 62.89;
H, 5.83%.
Synthesis of dimethyl 4,40-(but-2-ene-2,3-diyl)dibenzoate (2a
trans, 2b cis)
With vigorous stirring TiCl4 (2.7 mL, 0.025 mol) was added
ꢀ
slowly to a cold (0 C) suspension of Zn–Cu couple (3.32 g) in
THF (50 mL). The mixture was then refluxed for 1 h, cooled to
ambient temperature, followed by the addition of a solution of
methyl 4-acetylbenzoate (2.0 g, 0.011 mol) in THF (10 mL). The
reaction mixture was further refluxed for 12 h. After cooling to
ambient temperature, the reaction mixture was quenched with 10
mL of saturated K2CO3 (aq) and extracted with Et2O. The
organic layer was washed with brine, dried over MgSO4. After
MgSO4 was filtered off, the solvents were removed to give the
crude product, which was purified by silica gel column chroma-
tography (eluted with CH2Cl2) to give a trans/cis mixture of
compound 2 (1.0 g, 55% yield) as a light yellow oil. The two
isomers can be separated through chromatography and isolated
as white or light yellow solids, although it is not necessary for the
next step. 2b: 1H NMR (CDCl3, 400 MHz, 25 ꢀC) d 7.74 (d, J ¼
8.4 Hz, 4H), 7.00 (d, J ¼ 8.4 Hz, 4H), 3.85 (s, 6H), 2.18(s, 6H);
13C NMR (CDCl3, 400 MHz, 25 ꢀC) d 167.1, 149.3, 133.8, 129.3,
Synthesis of (E)-4,40-(1,4-bis(methylthio)but-2-ene-2,3-
diyl)dibenzoic acid (H2LS, 5)
Potassium hydroxide (1.1 g, 19 mmol) and 4 (500 mg, 1.2 mmol)
were refluxed in methanol (60 mL) for 2 h. After removal of
solvents in vacuo, distilled water was added to dissolve the
residual solid. The aqueous solution was acidified with excess 1
M HCl to give a white suspension. The suspension was directly
extracted with EtOAc. The organic phase was dried over MgSO4
and concentrated to give 466 mg white powder of 5 in quanti-
tative yield. 1H NMR (CD3OD, 400 MHz, 25 ꢀC) d 8.11 (d, J ¼
8.4 Hz, 4H), 7.58 (d, J ¼ 8.4 Hz, 4H), 3.40 (s, 4H), 1.84(s, 6H);
13C NMR (DMSO-d6, 100 MHz, 25 ꢀC) d 166.9, 144.2, 136.1,
129.6, 129.1, 129.0, 37.6, 14.6. Anal. calcd for C20H20O4S2: C,
61.83; H, 5.19. Found: C, 61.70; H, 5.34%.
1822 | CrystEngComm, 2011, 13, 1821–1830
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Synthesis of [Tb(LH)1.5(DMSO)2]$xDMSO (6)
modeled successfully. The residual diffuse electron density of
unidentified solvent molecules in the lattices of 6 and 7 was
removed with the SQUEEZE function of PLATON program23
and their contributions were not included in the formula,
although elemental analysis results suggest that the removed
electron density may originate at least partially from several
DMSO molecules. All non-hydrogen atoms except for the atoms
involved in the disordered portions were refined anisotropically.
The positions of the hydrogen atoms were calculated using the
riding model. Selected crystallographic data of the MOFs and
synthesis intermediates are summarized in Table 1, while selected
bond lengths and angles of 6–9 are listed in Table 2.
A small vial containing a 2 mL DMSO solution of TbCl3$6H2O
(8 mM) and H2LH (12 mM) was inserted into a large vial con-
taining DMSO (8 mL) and triethylamine (0.024 mL, 0.17 mmol).
The large vial was then sealed. The double-vial system was
allowed to stand for 14 d at room temperature to afford 6 as
colorless X-ray quality crystals. The crystals were collected by
filtration and dried under vacuum for 1 h (7.6 mg, 46% yield).
The synthesis of 6 has poor reproducibility. Anal. Calcd. for
C28H27O8S2Tb$4DMSO: C, 42.10; H, 5.00. Found: C, 42.44;
H, 4.90%.
Synthesis of [Eu2(LH)3(DMSO)4]$xDMSO (7)
Results and discussion
A small vial containing a 2 mL DMSO solution of EuCl3$6H2O
(8 mM) and H2LH (12 mM) was inserted into a large vial con-
taining DMSO (8 mL) and triethylamine (0.024 mL, 0.17 mmol).
The large vial was then sealed. The double-vial system was
allowed to stand for 14 d at room temperature to afford 7 as
colorless X-ray quality crystals. The crystals were collected by
filtration and dried under vacuum for 1 h (8.2 mg, 41% yield).
Anal. calcd. for C56H54O16S4Eu2$14DMSO: C, 40.21; H, 5.54.
Found: C, 39.92; H, 5.17%.
Ligand synthesis
As shown in Scheme 2, ligand H2LS can be synthesized in 5 steps
in a 19% overall yield. Compound 1 was synthesized from
carboxylation of 4-iodo-acetophenone under CO in 80% yield.10
McMurry coupling19 reaction of 1 produces a mixture of trans
(2a) and cis (2b) isomers of 2, in about 1 : 4 ratio according to the
1H NMR spectrum of the crude product. Although the cis isomer
2b is the major product, the subsequent bromination of 2b using
NBS causes the isomerization of the double bond to give 3a as
the major product. Therefore, the mixture of 2a and 2b resulting
from the McMurry coupling step can be used directly for the
bromination reaction and the trans(3a)/cis(3b) isomers of 3 can
be separated after the bromination step using column chroma-
tography. The treatment of the trans compound 3a with NaSCH3
affords 4, which can then be hydrolyzed to produce H2LS, 5. The
geometry at the C]C linkage of 2a, 2b, 3a, 3b and 4 has been
confirmed by X-ray crystallography (see Fig. S1, ESI†).
According to the crystal structures, the two phenyl rings and the
C]C linkage of 2a, 3a and 4 are no longer co-planar due to the
steric effect imparted by the substituents.
Synthesis of [Eu2(LS)3(DMF)2(H2O)2] (8)
Eu(NO3)3$6H2O (17.8 mg, 0.040 mmol), H2LS (23.3 mg, 0.060
mmol), N,N0-dimethylformamide (4 mL) and distilled water
(6 mL) were mixed and sealed in a 23 mL PTFE lined autoclave,
ꢀ
then heated at 120 C for 3 d. After cooling, yellow needle-sha-
ped crystals of 8 were collected by filtration, washed with
ethanol, and dried in air (26.5 mg, 80% yield). Anal. Calcd. for
C66H72Eu2N2O16S6: C, 48.17; H, 4.41; N, 1.70. Found: C, 47.93;
H, 4.17; N, 2.12%.
Synthesis of [Tb2(LS)3(DMF)2(H2O)2] (9)
The synthesis of 9 is similar to that of 8, but with TbCl3$6H2O
(14.9 mg, 0.040 mmol). Yellow needle-shaped crystals of 9 were
collected by filtration, washed with ethanol, and dried in air (22.6
mg, 68% yield). Anal. calcd for C66H72N2O16S6Tb2: C, 47.77; H,
4.37. Found: C, 47.87; H, 4.47%.
MOF preparation and structure description
The ligand (E)-4,40-(ethene-1,2-diyl)dibenzoic acid, H2LH, is
commercially available. The MOFs 6 and 7 can be prepared in
crystalline form in moderate yields by slow diffusion of trie-
thylamine vapor into the mixed solution of H2LH and
MCl3$6H2O (M ¼ Tb in 6 and Eu in 7) at ambient temperature.
Some unidentified amorphous precipitates were also observed
during the preparation of 6 and 7. These amorphous precipitates
have lower density compared to the crystals of 6 and 7 and
therefore, can be flushed away with DMSO. The crystals of 6
and 7 collapse into powders quickly upon contact with acetone
and chloroform; they also lose solvents and become opaque
rapidly upon standing in air. Compounds 6 and 7 are insoluble in
common solvents, suggesting the polymeric nature of their
structures.
X-Ray crystallographic analysis
X-Ray quality single crystals of 6–9 were obtained as described in
the sections above; those of 2a, 2b, 3a, 3b and 4 were obtained
from slow evaporation of the solvent from the corresponding
CH2Cl2 solutions. All crystals were mounted on the tip of
a MiTeGen MicroMount. All single-crystal X-ray diffraction
data were collected on a Bruker Kappa Apex II diffractometer
˚
with graphite-monochromated Mo Ka radiation (l ¼ 0.71073 A)
operating at 50 kV and 30 mA, at 120 K or 150 K controlled by
an Oxford Cryostream 700 series low temperature system. The
data integration and absorption correction were performed with
the Bruker Apex 2 software package.21 All structures were solved
by the direct methods except for compound 9 which was solved
by Patterson method. All structure solution and refinements were
performed using SHELXTL V6.14.22 A disordered phenyl ring of
the ligand in 7 and the thioether side chains in 8 and 9 have been
2ꢁ
Because of the structural similarity between LH and 4,40-
azodibenzoate (ADB) used by Yaghi,16 the corresponding Tb(III)
MOFs, 6 and [Tb2(ADB)3(DMSO)4]$xDMSO, show almost
identical structures, e.g., same space group, similar unit cell
parameters and topology. Compound 6 crystallized in the
monoclinic space group C2/c. As shown in Fig. 1a, each Tb(III)
center adopts distorted tricapped trigonal prismatic coordination
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 1821–1830 | 1823

View Article Online
geometry with seven oxygen donor atoms from the carboxylic
2ꢁ
groups of LH ligands and two oxygen donor atoms from
terminal DMSO ligands occupying the nine coordination sites.
The trigonal prism is defined by O2, O3, O4, O5, O6, and O8; the
three rectangular faces are capped by O1, O4A, and O7,
respectively. The three capping atoms and Tb1 are roughly
coplanar, as indicated by that the sum of the three relevant bond
angles is 355.4(1)ꢀ, close to 360ꢀ. As shown in Fig. 1b, two
adjacent Tb(III) centers are bridged by four carboxylate groups to
˚
form a bimetallic node with the Tb–Tb distance of 4.0866(4) A,
comparable to the literature values.16 There is a crystallographi-
cally imposed center of inversion at the centroid of each di-Tb
node. Because of unsymmetrical bridging, the Tb1–O4A bond
˚
length is 2.851(3) A, while all the remaining Tb–O bond lengths
˚
are within the range of 2.293(3)–2.480(4) A. As shown in Fig. 2a,
each di-Tb node is connected to six neighboring nodes via stil-
bene linkers to form a 3D infinite framework. Similar to Yaghi’s
[Tb2(ADB)3(DMSO)4]$xDMSO, a 2-fold interpenetration has
been observed in the crystal lattice of 6 (Fig. 2b), as a result of the
long linker group employed in the organic ligand.
Although switching from Tb(III) to Eu(III) usually produces
isostructural analogues, the Eu(III) MOF of LH2ꢁ, 7, crystallized
in a different space group (P212121) compared to 6. Similar to the
Tb(III) centers in 6, each Eu(III) center adopts distorted tricapped
trigonal prismatic coordination geometry with seven oxygen
donor atoms from the carboxylic groups of LH2ꢁ ligands and two
oxygen donor atoms from terminal DMSO ligands occupying
the nine coordination sites (Fig. 3a). The trigonal prism is defined
by O1, O4, O6, O8, O9, and O13; the three rectangular faces are
capped by O7, O10, and O14, respectively. The three capping
atoms and Eu1 are roughly coplanar, as indicated by that the
sum of the three relevant bond angles is 355.6(2)ꢀ, close to 360ꢀ.
As shown in Fig. 3b, two adjacent Eu(III) centers are bridged by
four carboxylate groups to form a bimetallic node with the Eu–
˚
Eu distance of 4.0743(6) A, comparable to the literature
values.17,23 In contrast to the di-Tb nodes in 6, there is no crys-
tallographically imposed center of inversion at the centroid of the
di-Eu nodes in 7, as a result of the chiral space group in which 7
crystallized. Because of unsymmetrical bridging by O7 and O4,
˚
Eu1–O7 and Eu2–O4 bond lengths are 2.802(6) and 2.813(6) A,
respectively, while all the remaining Eu–O bond lengths are
˚
within the range of 2.346(6)–2.511(3) A. As shown in Fig. 4a,
each bimetallic node is connected to six neighboring nodes via
stilbene linkers to form a 3D infinite framework. Because of the
large spacing generated by the long linker, 2-fold interpenetra-
tion has been observed in the crystal lattice of 7 (Fig. 4b).
Similar to the carboxylates in [Eu2L3(DMSO)2(CH3OH)2]$
2DMSO (L ¼ 4,40-ethyne-1,2-diyldibenzoate),17 three coordina-
tion modes have been observed for the carboxylates in the lattices
of 6 and 7 (Scheme 3): chelation mode, bridging mode, and
chelate-bridging mode. In the lattice of [Eu2L3(DMSO)2-
(CH3OH)2]$2DMSO, the two carboxylates within each L2ꢁ
ligand can adopt different coordination modes. In contrast, the
2ꢁ
two carboxylate ligands within each LH ligand always adopt
the same coordination mode in the lattices of 6 and 7.
MOF 8 and 9 can be prepared as light yellow needle-shaped
crystals in good yields from M(NO3)3$6H2O (M ¼ Eu in 8 and
Tb in 9) and H2LS solution under solvothermal conditions in
DMF/H2O. These two isostructural 3D coordination polymers
1824 | CrystEngComm, 2011, 13, 1821–1830
This journal is ª The Royal Society of Chemistry 2011

View Article Online
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 1821–1830 | 1825

View Article Online
Scheme 2 Synthesis of ligand H2LS (5).
Fig. 2 The extended structure of 6: (a) a drawing of one set of 3D
framework, projection along the b axis; (b) a space filling drawing
showing the two interpenetrating frameworks, one in red and the other in
blue. Hydrogen atoms and lattice solvent molecules are omitted for
clarity.
modes, the carboxylates in 8 and 9 only show two types of
2ꢁ
Fig. 1 ORTEP drawings of 6: (a) the coordination sphere of a Tb(III)
center showing the tricapped trigonal prismatic coordination geometry;
(b) the structure of a dinuclear node [Tb2LH3(DMSO)4] with the stilbene
linkers reduced to the ipso carbons. The thermal ellipsoids are plotted at
50% probabilities. All hydrogen atoms are omitted for clarity.
coordination modes: the chelating mode (for 2/3 of the LS
2ꢁ
ligands) and the chelate-bridging mode (for 1/3 of the LS
ligands). The bridging mode is not observed in 8 or 9. Most Eu–O
˚
and Tb–O bond lengths are in the ranges of 2.390(5)–2.480(5) A
˚
and 2.371(5)–2.463(5) A, respectively, except for those of Eu1–
˚
O3 and Tb1–O3 which are 2.648(5) and 2.653(5) A, respectively,
ꢀ
both crystallized in the triclinic space group P1. Compound 8 will
due to the unsymmetrical bridging of O3 in each case. In the
crystal lattice, each bimetallic node is connected with six adjacent
nodes via six organic linkers to form an infinite 3D framework
be used as an example for structure descriptions below. Each
Eu(III) center adopts a highly irregular coordination geometry
2ꢁ
2ꢁ
2ꢁ
with seven oxygen donor atoms from LS ligands, one oxygen
(Fig. 6a). Compared to the LH ligand, the LS ligand has
bulkier thioether substituents on the central CC double bond.
donor atom from a DMF molecule, and one oxygen donor atom
from a H2O molecule occupying the nine coordination sites.
Each Eu(III) center is linked with an adjacent Eu(III) center by
two bridging oxygen atoms from two carboxylates, forming a di-
Eu node as shown in Fig. 5. A similar di-Tb node found in 9 is
shown in Fig. 5. There is a crystallographically imposed center of
inversion at the centroid of each bimetallic node. Unlike the
carboxylates in 6 and 7 which show three types of coordination
2ꢁ
Because MOFs 6 and 7 assembled with LH ligands display
a 2-fold interpenetration, we expect no more than 2-fold of
interpenetration in MOFs 8 and 9. To our surprise, 3-fold
interpenetrating frameworks without any non-coordinating
guest molecules were observed in both 8 and 9 (Fig. 6b).
Calculation via PLATON program23 shows no solvent accessible
space, indicating that the addition framework has taken up all
1826 | CrystEngComm, 2011, 13, 1821–1830
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Fig. 4 The extended structure of 7: (a) a drawing of one set of 3D
framework, projection along the c axis; (b) a space filling drawing
showing the two interpenetrating frameworks, one in red and the other in
blue. Hydrogen atoms and lattice solvent molecules are omitted for
clarity.
Fig. 3 ORTEP drawings of 7: (a) the coordination sphere of a Eu(III)
center showing the tricapped trigonal prismatic coordination geometry;
(b) the structure of a dinuclear node [Eu2(LH)3(DMSO)4]. The thermal
ellipsoids are plotted at 50% probabilities. All hydrogen atoms are
omitted for clarity.
channels perpendicular to the a direction impossible, leading to
the non-porous structures (see Fig. S2, ESI†).
the space that is occupied by solvent molecules in MOFs 6 and 7.
It appears that the introduction of the thioether side chains
makes the inter-framework interactions stronger than the
framework–solvent interactions.
Although there is no solvent accessible space in MOF 8, we
envision that the removal of the coordinating solvents may create
solvent accessible void space. The solvent removal can be ach-
ieved by heating the as-synthesized sample of 8 at 140 ꢀC for 4 h.
The resulting de-solvated sample displays little weight loss at
130–165 ꢀC (Fig. S12b, ESI†) compared to as-synthesized 8
(Fig. S12a, ESI†), suggesting the removal of a significant amount
Solvent inclusion and porosity
In MOFs 6–9, the organic spacers and the bimetallic nodes
assemble into approximately orthogonal skeletons. Each set of
framework can also be viewed as an infinite assembly of ‘cubes’
defined by twelve organic spacers and eight bimetallic nodes. If
the frameworks are viewed along the body diagonal direction of
the ‘cubes’, triangular shaped channels can be seen (Fig. 7a).
MOFs 6 and 7 have similar triangular shaped channels in which
the lattice solvent molecules reside. Gas adsorption/desorption
experiments on MOFs 6 and 7 show no porosity, presumably
because the removal of solvents causes the structure to collapse.
As shown in Fig. 7b, in the lattice of MOF 8 (9 is similar to 8) the
thioether side chains fills the ‘trigonal prismatic’ space (along the
a direction) defined by the stilbene skeletons completely. In
addition, the 3-fold interpenetration makes the formation of
Scheme 3 Different coordination modes of carboxylate ligands.
CrystEngComm, 2011, 13, 1821–1830 | 1827
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Fig. 5 ORTEP drawings. M ¼ Eu (8) or Tb (9). Thermal ellipsoids are
2ꢁ
plotted at 50% probabilities. All LS ligands are reduced to the
carboxylic groups and an ipso carbon from the phenyl ring and all
hydrogen atoms are omitted for clarity.
of solvents. Accordingly, the IR spectrum of the de-solvated 8
shows a significant decrease in intensities of the signals at 1661
cmꢁ1 (coordinating DMF molecules)24,25 and 3300 cmꢁ1 (broad
peak from H2O), compared to the as-synthesized 8 (Fig. S13,
ESI†). When the de-solvated 8 is soaked in the mixture of DMF/
Fig. 7 Diagrams showing: (a) MOF 6, space filling model, projection
down the a + b direction; lattice solvent molecules are omitted for clarity.
(b) MOF 8, space filling model for the skeleton and stick model for the
thioether functional groups, projection down the a direction.
H2O (v/v ¼1 : 1) overnight and then washed with ethanol and air-
dried, the resulting re-solvated 8 shows a weight loss at 120–165
ꢀC in TGA (Fig. S12c, ESI†), similar to as-synthesized 8.
Correspondingly, the IR spectrum of the re-solvated 8 shows the
recovery of the intensities of the signals at 1661 and 3300 cmꢁ1
(Fig. S13c, ESI†). Powder XRD experiments show that
compound 8 remains crystalline after the solvent removal and re-
entry processes (Fig. S14, ESI†). Note that ꢀthe crystals are
insoluble in a DMF/H2O mixture even at 120 C and therefore
the dissolution/recrystallization possibility during the soaking
process at room temperature can be ruled out.
Thermogravimetric analysis
Thermal gravimetric analysis (TGA) of MOFs 6 and 7 (Fig. S3,
ESI†) shows that the loss of coordinating and lattice DMSO
solvent molecules starts at ambient temperature and reaches
completion at around 250 ꢀC. The plateau between 130 and 200
ꢀC is calculated to be [M2(LH)3(DMSO)2], partially de-solvated
MOFs. After the complete removalꢀof solvent molecules, the
remaining solid are stable up to 550 C without weight loss. At
Fig. 6 Extended structures of 8 (or 9): (a) one set of framework; (b)
space filling drawing of three interpenetrating frameworks coated with
different colors.
ꢀ
550–600 C, the weight loss can be attributed to the decompo-
sition of the ligand, which is supported by the identical
1828 | CrystEngComm, 2011, 13, 1821–1830
This journal is ª The Royal Society of Chemistry 2011

View Article Online
decomposition temperature displayed by Na2LH (Fig. S5, ESI†).
According to TGA, the compositions of sample 6 and 7
were calculated to be [Tb2(LH)3(DMSO)4]$8DMSO and
[Eu2(LH)3(DMSO)4]$14DMSO respectively, which correspond
well with the elementary analysis results.
closely resemble Yaghi’s [Tb2(ADB)3(DMSO)4]$16DMSO
(where ADB ¼ 4,40-azodibenzoate). The skeletons of MOFs 8
and 9 define similar trigonal prismatic space as seen in 6 and 7.
However, the thioether functional groups fill the space
completely and the 3-fold interpenetration prevents the forma-
tion of channels along other directions, making the as-synthe-
sized 8 and 9 non-porous. The coordiating DMF and H2O
solvents in 8 and 9 can be removed by heating. Such solvent
removal can be reversed by soaking the desolvated crystals in
a DMF and H2O mixture. The desolvated crystals of 8 show no
significant luminescent property change upon exposure to heavy
metal ions. The formation of the 3-fold interpenetration in 8
and 9 are likely for thermal dynamic reasons, i.e., the inter-
framework interactions are stronger than framework–solvent
interactions. By changing the size, number, or nature of the
functional groups on the ligands and solvents, we may poten-
tially create some extra space in the resulting MOFs. Finally, the
installation of thioether side chains enables the LMET and thus
enhances Eu-based luminescence. This effect could be attributed
at least partly to the steric bulk provided by the side chains,
which prevents the ligand from being coplanar and in turn, alters
the ligand energy levels.
TGA of MOF 8 and 9 (Fig. S4, ESI†) showed a first weight
loss of ꢂ11% at 130–165 ꢀC, which is attributed to the complete
removal of coordinating H2O and DMF molecules. After that,
the loss of the ligand proceeds in a stepwise fashion as the
temperature rises slowly.
Luminescence properties
The photoluminescence excitation and emission spectra of
MOFs 6–9 are shown in Fig. S15 (ESI†). Both 6 and 7 show
broad ligand-based emission between 420 and 490 nm with the
emission maximum at 447 nm (Fig. S15a and b, ESI†), which are
similar to the luminescence spectrum of the free ligand H2LH
(Fig. S6, ESI†).
The emission of compound 8 (Fig. S15c, ESI†) consists of both
the ligand-based and metal-based luminescence. The sharp lines
at 393, 464 and 535 nm in the excitation spectrum can be assigned
to the metal-centered transitions 7F0,1 / 5L6, 7F0,1 / 5D2, 7F0,1
/ 5D1, respectively. The weak broad band centered at ꢂ370 nm
in the excitation spectrum can be attributed to the ligand-based
excitation. The sharp emission peaks at 579, 591, 615, and 696
nm in the emission spectrum are characteristic Eu-centered
transitions 5D0–7F0, 5D0–7F1, 5D0–7F2, and 5D0–7F4 respectively.
When 393 nm is used as the excitation wavelength, at which an
intense metal-centered excitation and a weak ligand-centered
excitation co-exist, the ligand-based emission can be observed as
a weak broad band centered at around 490 nm, along with
intense Eu-based emissions. To examine whether LMET
contributes to the Eu-based emissions, a second emission spec-
trum (emission (ii) in Fig. S15c, ESI†) was recorded with 350 nm
excitation wavelength, at which only ligand-based absorption
occurs. The resulting emission spectrum shows not only the
ligand-based emission, but also the characteristic Eu emissions
with relatively low intensity compared to the ligand-based
emission, indicating the occurrence of LMET at low efficiency.
Compound 9 (Fig. S15d, ESI†) shows broad ligand-based exci-
tation and emission at 350–450 nm, 475–600 nm respectively,
while the direct f–f transition of Tb3+ happens under 487 nm
excitation (7F6–5D4) and the emission peaks at 542 nm, 584 nm,
Acknowledgements
This research is supported by grants to D.S. from the Natural
Science and Engineering Research Council (NSERC) of Canada,
the Ontario Ministry of Research and Innovation. We thank
Winnik group for sharing instruments and Meng Zhang for the
assistance on photoluminescence measurements.
References
ꢁ
1 L. J. Murray, M. Dinca and J. R. Long, Chem. Soc. Rev., 2009, 38,
1294–1314.
2 R. Zou, A. I. Abdel-Fattah, H. Xu, Y. Zhao and D. D. Hickmott,
CrystEngComm, 2010, 12, 1337–1353.
3 J.-R. Li, R. J. Kuppler and H.-C. Zhou, Chem. Soc. Rev., 2009, 38,
1477–1504.
4 (a) B. Chen, Y. Yang, F. Zapata, G. Lin, G. Qian and
E. B. Lobkovsky, Adv. Mater., 2007, 19, 1693–1696; (b) Y. Xiao,
Y. Cui, Q. Zheng, S. Xiang, G. Qian and B. Chen, Chem.
Commun., 2010, 46, 5503–5505; (c) B. Chen, L. Wang, Y. Xiao,
F. R. Fronczek, M. Xue, Y. Cui and G. Qian, Angew. Chem., Int.
Ed., 2009, 48, 500–503; (d) B. Chen, L. Wang, F. Zapata, G. Gian
and E. B. Lobkovsky, J. Am. Chem. Soc., 2008, 130, 6718–6719; (e)
B. Chen, S. Xiang and G. Qian, Acc. Chem. Res., 2010, 43, 1115–1124.
5
5
620 nm on emission (ii) can be assigned to D4–7F5, D4–7F4,
5D4–7F3, respectively. Preliminary luminescent response tests
were carried out by measuring the luminescence spectrum of
desolvated compound 8 soaked in HgCl2 solution in THF and
CdCl2 aqueous solution. No significant luminescence signal
change has been observed.
ꢂ
5 B. V. Harbuzaru, A. Corma, F. Rey, J. L. Jorda, D. Ananias,
L. D. Carlos and J. Rocha, Angew. Chem., Int. Ed., 2009, 48, 6476–
6479.
6 B. Zhao, X.-Y. Chen, P. Cheng, D.-Z. Liao, S.-P. Yan and Z.-
H. Jiang, J. Am. Chem. Soc., 2004, 126, 15394–15395.
7 W.-G. Lu, L. Jiang, X.-L. Feng and T.-B. Lu, Inorg. Chem., 2009, 48,
6997–6999.
8 Z. Xie, L. Ma, K. E. deKrafft, A. Jin and W. Lin, J. Am. Chem. Soc.,
2010, 132, 922–923.
Conclusions
9 J. Lee, O. K. Farha, J. Roberts, K. A. Scheidt, S. T. Nguyen and
J. T. Hupp, Chem. Soc. Rev., 2009, 38, 1450–1459.
10 L. Ma, C. Abney and W. Lin, Chem. Soc. Rev., 2009, 38, 1248–1256.
11 O. M. Yaghi, M. O’Keeffe, N. W. Ockwig, H. K. Chae, M. Eddaoudi
and J. Kim, Nature, 2003, 423, 705–714.
We have designed and synthesized a new ligand with thioether
side chain, and successfully assembled both non-functionalized
LH2ꢁ and functionalized LS2ꢁ ligand into luminescent 3D MOFs
6–9 with lanthanide metal ions. MOFs 6, 7, and 9 display ligand-
based luminescence, while MOF 8 showed both ligand-based and
Eu-centered luminescence. The structures of MOFs 6 and 7
12 J. J. Perry IV, J. A. Perman and M. J. Zaworotko, Chem. Soc. Rev.,
2009, 38, 1400–1417.
ꢂ
13 D. J. Tranchemontagne, J. L. Mendoza-Cortes, M. O’Keeffe and
O. M. Yaghi, Chem. Soc. Rev., 2009, 38, 1257–1283.
This journal is ª The Royal Society of Chemistry 2011
CrystEngComm, 2011, 13, 1821–1830 | 1829

View Article Online
14 (a) H. Deng, C. J. Doonan, H. Furukawa, R. B. Ferreira, J. Towne,
C. B. Knobler, B. Wang and O. M. Yaghi, Science, 2010, 327, 846–
850; (b) J. S. Seo, D. Whang, H. Lee, S. I. Jun, J. Oh, Y. J. Jeon
and K. Kim, Nature, 2000, 404, 982–986; (c) S. Hasegawa,
S. Horike, R. Matsuda, S. Furukawa, K. Mochizuki, Y. Kinoshita
and S. Kitagawa, J. Am. Chem. Soc., 2007, 129, 2607–2614; (d)
S. Horike, S. Bureekaew and S. Kitagawa, Chem. Commun., 2008,
471–473.
15 Z. Wang and S. M. Cohen, Chem. Soc. Rev., 2009, 38, 1315–1329.
16 T. M. Reineke, M. Eddaoudi, D. Moler, M. O’Keeffe and
O. M. Yaghi, J. Am. Chem. Soc., 2000, 122, 4843–4844.
17 B. T. Nguyen Pham, L. M. Lund and D. Song, Inorg. Chem., 2008, 47,
6329–6335.
Mater., 2009, 21, 95–101; (d) J. Yang, J.-F. Ma, Y.-Y. Liu and
S. R. Batten, CrystEngComm, 2009, 11, 151–159; (e) L. Zhang, Y.-
L. Yao, Y.-X. Che and J.-M. Zheng, Cryst. Growth Des., 2010, 10,
528–533.
19 J. E. McMurry, M. P. Fleming, K. L. Kees and L. R. Krepski, J. Org.
Chem., 1978, 43, 3255–3266.
20 J. Liu, J. Chen and C. Xia, J. Catal., 2008, 253, 50–56.
21 Apex 2 Software Package; Bruker AXS inc., 2008.
22 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr., 2008,
64, 112–122.
23 (a) A. L. Spek, Acta Crystallogr., 1990, A46, C34; (b) A. L. Spek,
PLATON,
A
Multipurpose Crystallographic Tool, Utrecht
University, Utrecht, The Netherlands, 2000.
24 S. Viswanathan and A. de Bettencourt-Dias, Inorg. Chem., 2006, 45,
10138–10146.
25 D. C. Wilson, S. Liu, X. Chen, E. A. Meyers, X. Bao, A. V. Prosvirin,
K. R. Dunbar, C. M. Hadad and S. G. Shore, Inorg. Chem., 2009, 48,
5725–5735.
18 (a) C. A. Bauer, T. V. Timofeeva, T. B. Settersten, B. D. Patterson,
V. H. Liu, B. A. Simmons and M. D. Allendorf, J. Am. Chem. Soc.,
2007, 129, 7136–7144; (b) J. Yang, J.-F. Ma, S. R. Batten and Z.-
M. Su, Chem. Commun., 2008, 2233–2235; (c) F. P. Doty,
C. A. Bauer, A. J. Skulan, P. G. Grant and M. D. Allenforf, Adv.
1830 | CrystEngComm, 2011, 13, 1821–1830
This journal is ª The Royal Society of Chemistry 2011