Efficient, Scalable Syntheses of Linker Molecules for Metal-Organic Frameworks

Article abstract of DOI:10.1055/s-0034-1381039

Efficient synthesis protocols for five linkers of immediate interest for use in metal-organic frameworks (MOFs) are presented. The importance of scalable, cost-effective, high-yield processes with simple purifications and few steps is emphasized. The prot

Full text of DOI:10.1055/s-0034-1381039

Evaluation Only. Creat3ed7 - 2with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
SYNLETT
0
9
6
-
5
2
1
4
1
4
3
0
9
6
© Georg Thieme Verlag Stuttgart · New York
2015, 26, 1480–1485
letter
1480
K. T. Hylland et al.
Letter
Synlett
Efficient, Scalable Syntheses of Linker Molecules for Metal-Organic
Frameworks
Knut T. Hylland
Sigurd Øien-Ødegaard
Karl Petter Lillerud
Mats Tilset*
Department of Chemistry, University of Oslo, P. O.
Box 1033 Blindern, 0315 Oslo, Norway
mats.tilset@kjemi.uio.no
Dedicated to Professor K. Peter C. Vollhardt on the
occasion of his 69th birthday and 25 years of service
to SYNLETT
Received: 18.03.2015
Accepted after revision: 29.05.2015
Published online: 15.06.2015
DOI: 10.1055/s-0034-1381039; Art ID: st-2015-b0194-l
Abstract Efficient synthesis protocols for five linkers of immediate in-
terest for use in metal-organic frameworks (MOFs) are presented. The
importance of scalable, cost-effective, high-yield processes with simple
purifications and few steps is emphasized. The protocols allow for the
efficient synthesis of biphenyl-, terphenyl-, and quaterphenyl-based
linkers. Four of the linkers have been structurally characterized.
Keywords metal-organic frameworks, ligands, biaryls, coupling, ma-
terials science
Metal-organic frameworks (MOFs) are a class of crystal-
line, porous materials that consist of metal atoms or clus-
ters as cornerstones, often termed connectors, linked by or-
ganic molecules, so-called linkers, via functional groups
with coordination abilities, such as carboxylic acids.1,2
Figure 1 The zirconium-based MOF UiO-67 consists of hexanuclear zir-
MOFs have gained much attention over the last two de-
conium oxo clusters that are interconnected by biphenyl-4,4′-dicarbox-
cades, due to their flexible pore topology and structural di-
versity, which is explored with respect to applications in
gas storage and separation,3–5 catalysis,6,7 medicine,8 and
chemical sensing,9 amongst other. Our group was the first
to report the ubiquitous, thermally robust zirconium-based
MOF UiO-66 and UiO-67 (Figure 1).10–12
Due to the nature of MOFs certain criteria for linkers
and their syntheses need to be fulfilled. Rather than being
final products, the linkers serve as synthetic intermediates
in MOF chemistry. As thorough explorations of a MOF may
consume considerable amounts of material, efficient and
easily scalable syntheses of the linkers are desired, if a MOF
is to find its way to practical applications.13–15
ylate linkers10
inorganic connector.16,17 Added linker functionality is moti-
vated by the opportunity it offers to alter the adsorption
properties,18 pore topology,19 and stability of the MOF.20–23
Functional linkers also provide valuable sites for postsyn-
thetic functionalization, an important feature of MOFs,24–27
as was early demonstrated on UiO-66 derivatives by us.28 In
particular, amino-functionalized MOFs attract attention be-
cause of their potential for CO2 capture, activation, and se-
questration.3,24,29,30 Frequently used linkers may be com-
mercially available, but at high prices and with limited
choices of added functionality. This creates an in-house de-
mand for linker synthesis in MOF research groups.23,31
There is a need to develop linkers which contain func-
tional groups that become available for substrate interac-
tion in the pores of the MOF, that is, not coordinating to the
© Georg Thieme Verlag Stuttgart · New York — Synlett 2015, 26, 1480–1485

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
1481
K. T. Hylland et al.
Letter
Synlett
1. NaNO2, H2O/HCl
0–5 °C, 30 min
Me
Me
Me
Me
CuCN, DMF
130 °C, 3 h
H2N
NH2
I
I
2. KI, H2O, 80 °C, 2 h
2a
3a
60–65%
Me
Me
CN
Me
Me
1. 12 M NaOH (aq)
EtOH, reflux, 110 h
NC
HO2C
CO2H
2. H2O, concd HCl, r.t.
1a
85–90%
4a
85–90%
Scheme 1 The synthesis of 3,3′-dimethylbiphenyl-4,4′-dicarboxylic acid (1a)
In this work we present the efficient, multigram synthe-
ses of the linkers 3,3′-dimethylbiphenyl-4,4′-dicarboxylic
acid (1a), 3,3′-dinitrobiphenyl-4,4′-dicarboxylic acid (1b),
and 2′,5′-dimethyl-[1,1′:4′,1′′-terphenyl]-4,4′′-dicarboxylic
acid (1d) scaled up to 5–10 g quantities, as well as the syn-
theses of the linkers 3,3′-diaminobiphenyl-4,4′-dicarboxylic
acid (1c) and 2′,3′′-dimethyl-[1,1′:4′,1′′:4′′,1′′′-quaterphe-
nyl]-4,4′′′-dicarboxylic acid (1e) on a somewhat smaller
scale. The reason for including 1c and 1e in this letter is the
fact that they are conveniently derived from key intermedi-
ates in the syntheses of 1a and 1b.
sis was also conducted on up to 10 gram scale with high
yields and purity. The structure of 1a based on the X-ray
diffraction analysis was determined (Figure 2).
The dicarboxylic acid 1a is an interesting linker for bi-
phenyl-based MOF because the added methyl groups en-
hance the solubility of the biphenyl system compared with
the unfunctionalized biphenyl-4,4′-dicarboxylic acid,
which is used as standard linker for the synthesis of UiO-67.
The enhanced solubility leads to reduced solvent consump-
tion in the MOF synthesis.
A three-step procedure (Scheme 1) starting from com-
mercially available 3,3′-dimethylbenzidine (2a) furnished
1a. In this route, 3,3′-dimethylbiphenyl-4,4′-dinitrile (4a)
serves as a key intermediate and immediate precursor to
1a. Attempts to prepare 4a from 3,3′-dimethylbenzidine
through a double Sandmeyer cyanation always resulted in
low yields and difficult purification procedures; low yields
have also been previously reported for the synthesis of 4a
by the Sandmeyer reaction.32
Instead, the Rosenmund–von Braun cyanation of 4,4′-
diiodo-3,3′-dimethylbiphenyl (3a) provided 4a in excellent
yields and high purity in quantities up to 10 grams. The
diiodide 3a was synthesized from 3,3′-dimethylbenzidine
according to Serre et al.33
The final hydrolysis of 4a to 1a was performed over sev-
eral days in a mixture of aqueous 12 M NaOH and ethanol at
reflux. The disodium salt of 1a crystallized during the
course of the reaction, and this allowed the facile isolation
of the free diacid 1a without contamination from the copi-
ous quantities of NaCl which would be formed by direct
acidification of the reaction mixture with HCl. The hydroly-
Figure 2 ORTEP plots of the molecular structures of 1a, 1b (as its di-
methylammonium salt), 1d, and 1e (as dimethylacetamide solvate) as
determined by X-ray crystallography. Ellipsoids are shown at 50% prob-
ability
Our procedure represents a considerable improvement
over existing ones to prepare 1a. The existing methods re-
ported by Yaghi et al.34 and Serre et al.33 were also tested by
us. While the former lacks some experimental details, se-
vere limitations were found in the latter method when at-
tempts to scale up the process were made. For example, the
synthesis of the key intermediate and diacid precursor di-
ethyl 3,3′-dimethylbiphenyl-4,4′-dicarboxylate (5a) neces-
© Georg Thieme Verlag Stuttgart · New York — Synlett 2015, 26, 1480–1485

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
1482
K. T. Hylland et al.
Letter
Synlett
B2pin2
O2N
HO2C
O2N
EtO2C
PdCl2(dppf) (0.5 mol%)
Pd(OAc)2 (0.5 mol%)
EtOH, 95% H2SO4
reflux, 24 h
Br
Br
K2CO3, DMF
95 °C, 2 h
2b
3b
85–90%
O2N
EtO2C
NO2
O2N
HO2C
NO2
LiOH
THF–H2O, r.t., 2 h
CO2Et
CO2H
4b
75–85%
1b
90%
Scheme 2 The synthesis of 3,3′-dinitrobiphenyl-4,4′-dicarboxylic acid (1b)
sitated large quantities of n-butyllithium and ethyl chloro-
formate and required purification by column chromatogra-
phy.
was obtained in high purity and high yields by hydrolysis of
4b with aqueous LiOH in THF at room temperature in >6
gram scale. The X-ray crystal structure of 1b, in the form of
its dimethylammonium salt, was also obtained (Figure 2).
A major motivation for choosing 1b as a target for a
linker synthesis is that the nitro groups may be readily re-
duced to form amino groups that offer a host of possibilities
for further manipulation and exploration. Most notably,
amine-functionalized MOF linkers have the potential to in-
teract with CO2, as addressed in the introduction. Accord-
ingly, 1c has now been synthesized starting with the reduc-
tion of 4b with SnCl2·2H2O. The product, diethyl 3,3′-diami-
nobiphenyl-4,4′-dicarboxylate (2c), is then subjected to
hydrolysis to furnish 1c (Scheme 3). At this time, the syn-
thesis of 1c has been explored only on a smaller scale
(< 0.5 g) than the syntheses of 1a and 1b. We expect that an
efficient scalable synthesis will soon be in place.
By using a methodology that is analogous to that de-
scribed for the biphenyl linkers, we find that it is possible to
construct even longer linkers, namely terphenyl- and qua-
terphenyl-based systems. Thus, the two linkers 1d and 1e36
were synthesized according to Scheme 4, involving a double
Suzuki coupling of aromatic dihalides with 4-ethoxycar-
bonylphenylboronic acid (2d), to construct the diethyl ester
precursors of 1d and 1e, namely diethyl 2′,5′-dimethyl-
Linker 1b was synthesized by a different route (Scheme
2). The lack of a suitable, inexpensive biphenyl-based start-
ing material necessitated the construction of the biphenyl
system through catalytic coupling chemistry. A procedure
described by Bräse35 was adapted for the key intermediate
diethyl 3,3′-dinitrobiphenyl-4,4′-dicarboxylate (4b). The
precursor, ethyl 4-bromo-2-nitrobenzoate (3b), was ob-
tained through the esterification of commercially available
4-bromo-2-nitrobenzoic acid (2b).
While the coupling reaction proceeded well for 2b fol-
lowing the procedure established by Bräse, it was found
necessary to modify the procedure in order to facilitate
larger-scale preparations of 4b. The rather expensive cata-
lyst PdCl2(dppf) is originally used with a loading of 4 mol%,
which is rather unacceptable at the desired scale (mini-
mum 10 grams). We find that the catalyst loading can be
decreased to 0.5 mol% with the addition of an extra 0.5
mol% of the relatively inexpensive Pd(OAc)2. The reaction
was performed in DMF at 95 °C and was complete in as lit-
tle as two hours with minimal formation of byproducts.
This reaction has been scaled to >15 grams, and the yields
are consistently high, in the range of 75–85%. Finally, 1b
O2N
EtO2C
NO2
CO2Et
H2N
EtO2C
NH2
CO2Et
SnCl2⋅2H2O (8 equiv)
concd HCl, EtOH
reflux, 1 h
2c
4b
75–80%
H2N
HO2C
NH2
3 M NaOH (aq)
EtOH, reflux, 1 h
CO2H
1c
80%
Scheme 3 The synthesis of 3,3′-diaminobiphenyl-4,4′-dicarboxylic acid (1c)
© Georg Thieme Verlag Stuttgart · New York — Synlett 2015, 26, 1480–1485

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
1483
K. T. Hylland et al.
Letter
Synlett
Me
Me
Pd(OAc)2 (2.5 mol%)
Ph3P (10 mol%)
(HO)2B
KOH
CO2Et
Br
Br
EtO2C
CO2Et
+
K2CO3 (6 equiv)
DMF, 95 °C, 48 h
Me
Me
3d
4d
75–80%
2d
Me
dioxane–H2O, reflux, 18 h
HO2C
CO2H
Me
1d
85–90%
Pd(OAc)2 (2.5 mol%)
Ph3P (10 mol%)
Me
Me
Me
Me
EtO2C
CO2Et
I
I
+
2d
K2CO3 (6 equiv)
DMF, 95 °C, 24 h
2e
3a
77%
Me
Me
KOH
dioxane–H2O, reflux, 18 h
HO2C
CO2H
1e
85%
Scheme 4 The synthesis routes for 1d and 1e
[1,1′:4′,1′′-terphenyl]-4,4′′-dicarboxylate (4d) and diethyl
2′,3′′-dimethyl-[1,1′:4′,1′′:4′′,1′′′-quaterphenyl]-4,4′′′-dicar-
boxylate (2e36), respectively.
In this letter the syntheses of five linkers for metal-
organic frameworks have been described and discussed.
Each synthesis has been optimized to be easy scalable in
terms of the complexity of each synthesis route, the meth-
ods for purification, and the cost of starting materials. The
strategies established may be applied in synthesis of other
functionalized linkers based on the biphenyl, terphenyl, or
quaterphenyl moieties. The successful use of these linkers
in MOF synthesis will be presented elsewhere.
In the case of 1d, the commercially available 1,4-dibro-
mo-2,5-dimethylbenzene (3d) is the substrate for the dou-
ble coupling reaction, whereas in the case of 1e, the diio-
dide 3a (which already serves as an intermediate for the
synthesis of 1a) is the starting material. Both coupling pro-
cedures proceed with good yields for gram-scale reactions,
and column chromatography is again avoided as a purifica-
tion method in both cases. In both coupling reactions inex-
pensive Pd(OAc)2 is used as palladium source, but the draw-
backs are rather high catalyst loadings (2.5 mol%) and long
reaction times (24–48 h). Both 4d and 2e are easily hydro-
lyzed to the corresponding diacids in good yields, and it was
possible to obtain the crystal structure of both these
linkers.
Acknowledgment
The Norwegian Research Council is kindly acknowledged for funding
through projects no. 228157 (stipend to K.T.H.) and 215735 (stipend
to S.Ø.Ø.).
Supporting Information
The molecular structures of 1a, 1b, 1d, and 1e as deter-
mined by single-crystal X-ray crystallography are shown as
ORTEP plots in Figure 2.37 Bond distances are in accordance
with the reported structures.38 Neighboring aromatic
planes are rotated by 12.7° (1a), 10.3° (1b), 35.6° (1d),
43.8°, 24.6°, and 45.8° (1e) with respect to each other. Fur-
ther details concerning molecular structures, packing inter-
actions (involving H-bonding networks and π stacking in-
teractions) may be found in the cif files, available as Sup-
porting Information.
Supporting information for this article is available online at
http://dx.doi.org/10.1055/s-0034-1381039.
S
u
p
p
ortiInfogrmoaitn
S
u
p
p
o
nrtogI
f
rmoaitn
References and Notes
(1) Long, J. R.; Yaghi, O. M. Chem. Soc. Rev. 2009, 38, 1213.
(2) Horike, S.; Kitagawa, S. In Metal-Organic Frameworks: Applica-
tions from Catalysis to Gas Storage; Farrusseng, D., Ed.; Wiley-
VCH: Weinheim, 2011, 1–21.
© Georg Thieme Verlag Stuttgart · New York — Synlett 2015, 26, 1480–1485