Modular, active, and robust lewis acid catalysts supported on a metal-organic framework

Article abstract of DOI:10.1021/ic101125m

Metal-organic frameworks (MOFs) have shown promise as heterogeneous catalysts because of their high crystallinity, uniform pores, and ability to be chemically and physically tuned for specific chemical transformations. One of the challenges with MOF-based

Full text of DOI:10.1021/ic101125m

6766 Inorg. Chem. 2010, 49, 6766–6774
DOI: 10.1021/ic101125m
Modular, Active, and Robust Lewis Acid Catalysts Supported on a
Metal-Organic Framework
Kristine K. Tanabe and Seth M. Cohen*
Department of Chemistry and Biochemistry, University of California, San Diego, 9500 Gilman Drive,
La Jolla, California 92093-0358
Received June 3, 2010
Metal-organic frameworks (MOFs) have shown promise as heterogeneous catalysts because of their high
crystallinity, uniform pores, and ability to be chemically and physically tuned for specific chemical transformations.
One of the challenges with MOF-based catalysis is few systems achieve all of the desired features for a heterogeneous
catalyst, including high activity, robustness (recyclability), and excellent selectivity. Herein, postsynthetic modification
(PSM) of a MOF is used to synthesize a series of MOF catalysts that are highly robust and active for epoxide ring-
opening reactions. In the following study, four metalated MOFs (UMCM-1-AMInpz, UMCM-1-AMInsal, UMCM-1-
AMFesal, and UMCM-1-AMCupz) are examined as catalysts for β-azido and β-amino alcohol synthesis with epoxides
of varying sizes and shapes using two different nucleophiles (TMSN3 and aniline). The four MOFs are isostructural,
exhibit good thermal and structural stability, and display different catalytic activities based on the combination of metal
ion and chelating ligand immobilized within the framework. In particular, UMCM-1-AMInpz and UMCM-1-AMInsal act
as robust, single-site catalysts with distinct selectivity for ring-opening reactions with specific nucleophiles. More
importantly, one of these catalysts, UMCM-1-AMInpz, selectively promotes the ring-opening of cis-stilbene oxide in the
presence of trans-stilbene oxide, which cannot be achieved with a comparable molecular Lewis acid catalyst. The
results show that PSM is a promising, modular, and highly tunable approach for the discovery of robust, active, and
selective MOF catalysts that combine the best aspects of homogeneous and heterogeneous systems.
As solid supports and matrixes for catalysis, metal-organic
frameworks (MOFs) have garnered increasing attention
because of their ability to capture advantages of both homo-
geneous and heterogeneous catalysts.1-4 Homogeneous cata-
lysts can be limited by thermal instability, cross reactivity (e.g.,
self-degradation), and the difficulty involved in isolating and
reusing these catalysts.5 Heterogeneous catalysts are gene-
rally more robust than homogeneous catalysts, but they
are not easily functionalized and suffer from mass transport
limitations.6,7 MOF-supported catalysts may alleviate the
limitations of homogeneous and heterogeneous systems while
capturing the best features of each. The high porosity of MOFs
allow for fast mass transport and interaction with substrates.
MOFs also display high thermal stability, and the reactive
active sites can be chemically tuned and modified like a
homogeneous, single-site catalyst. Specifically, MOFs can be
functionalized via their metal nodes or organic ligands through
presynthetic8 or postsynthetic9 approaches. The metal ion and
organic ligand(s) can be carefully chosen to produce a MOF
with favorable attributes such as specific pore apertures, chiral
topologies, and unsaturated metal centers. Likewise, the MOF
can be further tuned after assembly by modifying the organic
substituents on the ligand or by decorating the metal node.10-21
(8) Tranchemontagne, D. J.; Mendoza-Cortes, J. L.; O’Keeffe, M.;
Yaghi, O. M. Chem. Soc. Rev. 2009, 38, 1257–1283.
(9) Wang, Z.; Cohen, S. M. Chem. Soc. Rev. 2009, 38, 1135–1329.
(10) Chui, S. S.-Y.; Lo, S. M.-F.; Charmant, J. P. H.; Orpen, A. G.;
Williams, I. D. Science 1999, 283, 1148–1150.
(11) Seo, J. S.; Whang, D.; Lee, H.; Jun, S. I.; Oh, J.; Jeon, Y. J.; Kim, K.
Nature 2000, 404, 982–986.
(12) Wang, Z.; Cohen, S. M. J. Am. Chem. Soc. 2007, 129, 12368–12369.
(13) Farha, O. K.; Mulfort, K. L.; Hupp, J. T. Inorg. Chem. 2008, 47,
10223–10225.
(14) Ingleson, M. J.; Barrio, J. P.; Guilbaud, J. B.; Khimyak, Y. Z.;
Rosseinsky, M. J. Chem. Commun. 2008, 2680–2682.
(15) Tanabe, K. K.; Wang, Z.; Cohen, S. M. J. Am. Chem. Soc. 2008, 130,
8508–8517.
(16) Banerjee, M.; Das, S.; Yoon, M.; Choi, H. J.; Hyun, M. H.; Park,
S. M.; Seo, G.; Kim, K. J. Am. Chem. Soc. 2009, 131, 7524–7525.
(17) Burrows, A. D.; Frost, C. G.; Mahon, M. F.; Richardson, C. Chem.
Commun. 2009, 4218–4220.
(18) Doonan, C.; Morris, W.; Furukawa, H.; Yaghi, O. M. J. Am. Chem.
Soc. 2009, 131, 9492–9493.
*To whom correspondence should be addressed. E-mail: scohen@
ucsd.edu. Phone: (858) 822-5596. Fax: (858) 822-5598.
(1) Lee, J.; Farha, O. K.; Roberts, J.; Scheidt, K. A.; Nguyen, S. T.; Hupp,
J. T. Chem. Soc. Rev. 2009, 38, 1450–1459.
(2) Ma, L.; Abney, C.; Lin, W. Chem. Soc. Rev. 2009, 38, 1248–1256.
(3) Farrusseng, D.; Aguado, S.; Pinel, C. Angew. Chem., Int. Ed. 2009, 48,
7502–7513.
(4) Corma, A.; Garcia, H.; Llabres i Xamena, F. X. Chem. Rev. 2010,
ASAP Contents.
(5) Li, C. Catal. Rev.-Sci. Eng. 2004, 46, 419–492.
(6) Forster, P. M.; Cheetham, A. K. Top. Catal. 2003, 24, 79–86.
(7) Czaja, A. U.; Trukhan, N.; Muller, U. Chem. Soc. Rev. 2009, 38, 1284–
1293.
(19) Gadzikwa, T.; Farha, O. K.; Mulfort, K. L.; Hupp, J. T.; Nguyen,
S. T. Chem. Commun. 2009, 3720–3722.
r
pubs.acs.org/IC
Published on Web 06/21/2010
2010 American Chemical Society

Article
Inorganic Chemistry, Vol. 49, No. 14, 2010 6767
The high crystallinity and uniform pores of MOFs suggest they
can catalyze reactions with selectivity based on substrate size
and shape.7,22-24
ring-opening at extremely low catalyst (metal) loadings;
(3) substrate selectivity can be achieved because of the size
and shape of the MOF environment. More importantly, the
substrate selectivity observed with the MOF catalysts are
distinct from those found with purely homogeneous catalysts,
illustrating that the pores of the MOF clearly create a specific
environment for catalysis that can be used to generate novel
patterns of selectivity. These results show that PSM can be
used to prepare active, selective, and robust MOF catalysts.
Despite these excellent features, the overall number of
reported MOF catalysts still remains relatively low when
compared to their use in areas such as gas sorption. Catalytic
MOFs are still in their infancy with few systems showing high
activity, selectivity, and recyclability.1,2 Various MOFs have
been explored as catalysts for transesterfications,11 C-C
bond formations,16,22,24,25 epoxidations,23 and many other
reactions.26-35 However, exerting precise control over the
physical and chemical properties of the framework to im-
prove catalytic activity and selectivity is not straightforward.
Although early results have been promising, only a few
studies provide evidence of MOF catalysts that are both
selective and robust.23,24
In the following study, we show that postsynthetic modi-
fication (PSM)9 can be used to carefully control the catalytic
properties of a MOF. Presented herein are a series of MOFs
that display unique activity and selectivity based on the
features of the specific active site created within the frame-
work. Four distinctMOFs were synthesizedby modifying the
same parent MOF (UMCM-1-NH2)36,37 with different com-
binations of chelating groups and metal ions (Scheme 1). All
of the MOFs have similar structural and thermal stabilities,
yet they display different catalytic behaviors under a given set
of reaction conditions. The differences in catalytic activity
provide strong evidence for the generation and tuning of well-
defined, single-site catalysts within the MOF lattice. The
findings presented demonstrate that (1) PSM can be utilized
to produce a series of single-site MOF catalysts with dif-
ferent combinations of metal and organic components;
(2) the MOFs are highly robust, active catalysts for epoxide
Experimental Methods
General Procedures. Starting materials and solvents were
purchased and used without further purification from commercial
suppliers (Sigma-Aldrich, Alfa Aesar, EMD, TCI, Cambridge
Isotope Laboratories, Inc., and others). (R,R)-N,N0-bis(3,5-di-
tert-butylsalicylidene)-1,2-cyclohexanediaminochromium(III)
chloride was purchased from Sigma Aldrich. Samples were sub-
mitted to Robertson Microlit Laboratories for atomic absorp-
tion (AA) analysis. 1H NMR spectra were recorded on a Varian
FT-NMR spectrometer (400 MHz). UMCM-1-NH2 was synthe-
sized and activated as previously described.36 UMCM-1-AMpz,
UMCM-1-AMsal, UMCM-1-AMFesal, and UMCM-1-AM-
Cupz were synthesized as previously described.38 Formation of
the corresponding TMS protected, β-azido alcohol, and β-amino
alcohol products were confirmed by comparison with reported
literature spectra.39-46
Metalation of UMCM-1-NH2 and Derivatives. Unmodified
UMCM-1-NH2 (56 mg, 0.05 mmol) or modified UMCM-
1-NH2 (UMCM-1-AMpz and UMCM-1-AMsal, 56 mg, 0.05
mmol) in 2 mL of CHCl3 was metalated with 1.0 equiv (0.05 mmol)
of In(acac)3. The mixture was left to stand at room temperature
(RT) for 4 h. After decanting the solution, the crystals were washed
with CHCl3 (4 ꢀ 10 mL) and left soaking in CHCl3 overnight. The
rinse and wash cycle was repeated for an additional 2 days, and the
crystals were left soaking in CHCl3 until used.
Powder X-ray Diffraction (PXRD) Analysis. Approximately
15 mg of modified UMCM-1-NH2 (typically soaked in CHCl3)
was air-dried before PXRD analysis. PXRD data were collected
at ambient temperature on a Bruker D8 Advance diffractometer
(20) Garibay, S. J.; Wang, Z.; Tanabe, K. K.; Cohen, S. M. Inorg. Chem.
2009, 48, 7341–7349.
(21) Taylor-Pashow, K. L.; Rocca, J. D.; Zie, Z.; Tran, S.; Lin, W. J. Am.
Chem. Soc. 2009, 131, 14261–14263.
(22) Wu, C.-D.; Hu, A.; Zhang, L.; Lin, W. J. Am. Chem. Soc. 2005, 127,
8940–8941.
(23) Cho, S.-I.; Ma, B.; Nguyen, S. T.; Hupp, J. T.; Albrecht-Schmitt,
T. E. Chem. Commun. 2006, 2563–2565.
˚
at 40 kV, 40 mA for Cu KR (λ = 1.5418 A), with a scan speed of
5 s/step, a step size of 0.02° in 2θ, and a 2θ range of 2-35°.
The experimental backgrounds were corrected using the Jade
5.0 software package.
(24) Horike, S.; Dinca, M.; Tamaki, K.; Long, J. R. J. Am. Chem. Soc.
2008, 130, 5854–5855.
(25) Fujita, M.; Kwon, Y. J.; Washizu, S.; Ogura, K. J. Am. Chem. Soc.
1994, 116, 1151–1152.
Thermal Analysis. Approximately 10-20 mg of MOF sample
was used for thermogravimetric analysis (TGA) measurements.
The MOF sample was either used directly after gas sorption
analysis or dried at 90 °C under vacuum for 4-5 h. Samples were
analyzed under a stream of dinitrogen using a TA Instrument Q600
SDT running from RT to 600 °C with a scan rate of 5 °C/min.
Brunauer-Emmett-Teller (BET) Surface Area Analysis. Ap-
proximately 40-60 mg of modified UMCM-1-NH2 (stored
in CHCl3) was evacuated on a vacuum line for 2 h at RT.
(26) Dybstev, D. N.; Nuzhdin, A. L.; Chun, H.; Bryliakov, K. P.; Talsi,
E. P.; Fedin, V. P.; Kim, K. Angew. Chem., Int. Ed. 2006, 45, 916–920.
ꢀ
(27) Gandara, F.; Gomex-Lor, B.; Gutierrez-Puebla, E.; Iglesias, M.;
Monge, M. A.; Proserpio, D. M.; Snejko, N. Chem. Mater. 2008, 20, 72–76.
(28) Hwang, Y. K.; Hong, D. Y.; Chang, J. S.; Jhung, S. H.; Seo, Y. K.;
Kim, J.; Vimont, A.; Daturi, M.; Serre, C.; Ferey, G. Angew. Chem., Int. Ed.
2008, 47, 4144–4148.
(29) Ingleson, M. J.; Barrio, J. P.; Bacsa, J.; Dickinson, C.; Park, H.;
Rosseinsky, M. J. Chem. Commun. 2008, 1287–1289.
(30) Jiang, D.; Mallat, T.; Krumeich, F.; Baiker, A. J. Catal. 2008, 390–
395.
(31) Tanaka, K.; Oda, S.; Shiro, M. Chem. Commun. 2008, 820–822.
(32) Schroder, F.; Esken, D.; Cokoja, M.; van den Berg, M. W. E.;
Lebedev, O. I.; Van Tendeloo, G.; Walaszek, B.; Buntkowsky, G.; Limbach,
H.-H.; Chaudret, B.; Fischer, R. A. J. Am. Chem. Soc. 2008, 130, 6119–6130.
(33) Schultz, A. M.; Farha, O. K.; Hupp, J. T.; Nguyen, S. T. J. Am.
Chem. Soc. 2009, 131, 4204–4205.
(34) Ohara, K.; Kawano, M.; Inokuma, Y.; Fujita, M. J. Am. Chem. Soc.
2009, 132, 30–31.
(35) Uemura, T.; Ono, Y.; Hijikata, Y.; Kitagawa, S. J. Am. Chem. Soc.
2010, 132, 4917-4924.
(36) Wang, Z.; Tanabe, K. K.; Cohen, S. M. Inorg. Chem. 2009, 48, 296–
306.
(38) Tanabe, K. K.; Cohen, S. M. Angew. Chem., Int. Ed. 2009, 48, 7424–
7427.
(39) Schaus, S. E.; Larrow, J. F.; Jacobsen, E. N. J. Org. Chem. 1997, 62,
4197–4199.
(40) Ami, E. i.; Ohrui, H. Biosci. Biotechnol. Biochem. 1999, 63, 2150–
2156.
(41) Negron, G.; Guerra, N.; Lomas, L.; Gavino, R.; Cardena, J.
ARKIVOC 2003, 11, 179–184.
(42) Sahasrabudhe, K.; Gracias, V.; Furness, K.; Smith, B. T.; Katz,
C. E.; Reddy, S.; Aube, J. J. Am. Chem. Soc. 2003, 125, 7914–7922.
(43) Shivarkar, A. B.; Gupte, S. P.; Chaudhari, R. V. Synlett. 2006, 9,
1374–1378.
(44) Yamashita, H. Bull. Chem. Soc. Jpn. 1988, 61, 1213–1220.
(45) Kureshy, R. I.; Prathap, K. J.; Agrawal, S.; Khan, N.-u. H.; Abdi,
S. H. R.; Jasra, R. V. Eur. J. Org. Chem. 2008, 3118–3128.
(46) Fleming, E. M.; Quigley, C.; Rozas, I.; Connon, S. J. J. Org. Chem.
2008, 73, 948–956.
(37) Koh, K.; Wong-Foy, A. G.; Matzger, A. J. Angew. Chem., Int. Ed.
2008, 47, 677–680.

6768 Inorganic Chemistry, Vol. 49, No. 14, 2010
Tanabe and Cohen
a
Scheme 1. Synthesis (Top) and Postsynthetic Modification (Bottom) of UMCM-1-NH2
a UMCM-1-NH2 is modified with cyclic anhydrides and metalated with metal acacs (M=Fe3þ, Cu2þ, In3þ) to give UMCM-1-AMMsal
(where M = Fe3þ or In3þ) and UMCM-1-AMMpz (where M = Cu2þ or In3þ).
The sample was then transferred to a preweighed sample tube
and degassed at 25 °C on an Micromeritics ASAP 2020 Adsorp-
tion Analyzer for a minimum of 12 h or until the outgas rate was
<5 μmHg. The sample tube was re-weighed to obtain a con-
sistent mass for the degassed MOF sample. BET surface area
(m2/g) measurements were collected at 77 K with dinitrogen
on an Micromeritics ASAP 2020 Adsorption Analyzer using
volumetric technique.
vial followed by cis-2,3-epoxybutane (0.1 mmol) and nucleo-
phile (TMSN3 or aniline, 0.1 mmol). The reaction mixture was
left standing at RT for 24 h, and the supernatant was analyzed
by 1H NMR.
Epoxide Ring-Opening Catalysis with Cr(salen)Cl. The ex-
periments described were adapted from a literature procedure.47
cis or trans epoxide (2,3-epoxybutane or stilbene oxide, 0.1 mmol)
and 5 mol % (R,R)-N,N0-bis(3,5-di-tert-butylsalicylidene)-1,2-
cyclohexanediaminochromium(III) chloride (Cr(salen)Cl) were
dissolved in 50 μL of CDCl3 in a 4 mL dram vial. TMSN3 or
aniline (0.1 mmol) was added, and the vials were placed on a
shaker for 24 h at RT. The reaction mixture was directly analyzed
by 1H NMR without further purification.
Epoxide Ring-Opening Catalysis with MOFs. UMCM-
1-NH2, UMCM-1-NH2 treated with In(acac)3, UMCM-1-AM-
Cupz, UMCM-1-AMInpz, and UMCM-1-AMInsal were dried
under vacuum at 90 °C for 4-5 h. UMCM-1-AMpz and
UMCM-1-AMsal were dried at RT for 4-5 h. UMCM-1-
AMFesal was activated as previously reported.38 Dried MOF
samples (15 mg, 0.014 mmol based on -NH2) were placed into
4 mL dram vials. The MOF was immersed in 1 mL of CDCl3,
followed by epoxide (0.1 mmol), and TMSN3 (0.1 mmol) or
aniline (0.1 mmol). The reaction mixture was left standing at RT
for 24 h and the supernatant was analyzed by 1H NMR.
In(acac)3 Control Reactions. In(acac)3 (23 mg, 0.056 mmol)
was dissolved in 4 mL of CDCl3 followed by the addition of
ligand (3-hydroxyphthalic anhydride, 2,3-pyrazinedicarboxylic
anhydride, or salicylamide, 0.042 mmol). The mixture was
sonicated for 15 min and was left to sit overnight. A portion
(1 mL) of the CDCl3 solution was transferred to a 4 mL dram
Results
We38 and others18 have demonstrated that UMCM-1-NH2,
a highly porous MOF composed of Zn4O clusters coordi-
nated by 4,40,400-benzene-1,3,5-triyl-tribenzoic acid (BTB)
and 2-amino-1,4-benzenecarboxylic acid (NH2-BDC) could
be postsynthetically modified with a chelating group and
(47) Bartoli, G.; Bosco, M.; Carlone, A.; Locatelli, M.; Massaccesi, M.;
Melchiorre, P.; Sambri, L. Org. Lett. 2004, 6, 2173–2176.

Article
Inorganic Chemistry, Vol. 49, No. 14, 2010 6769
Table 1. Atomic Absorption (AA) Analysis of UMCM-1-NH2 Treated with
In(acac)3, UMCM-1-AMInpz, and UMCM-1-AMInsal
MOF (UMCM-1)
metal
theoretical
experimental
Zn2þ
In3þ
Zn2þ
In3þ
Zn2þ
In3þ
25.26
N/A
22.40
4.92
23.09
3.55
24.30 ( 0.55
0.50 ( 0.28
21.48 ( 0.77
3.76 ( 0.56
22.41 ( 0.46
2.96 ( 0.25
a
-NH2 þ In(acac)3
-AMInpzb,c
-AMInsalb,d
a Based on three independent samples. b Based on six independent
samples. c Assuming 50% modification. d Assuming 35% modification.
subsequently metalated. The successful preparation of two
modified derivatives, UMCM-1-AMpz and UMCM-1-AM-
sal, which are appended with pyrazinedicarboxylate and
salicylate substituents (Scheme 1) was previously described.38
The reaction conditions used convert 50% and 35% of the
amine sites into the desired chelating groups for UMCM-1-
AMpz and UMCM-1-AMsal, respectively. UMCM-1-AM-
sal and UMCM-1-AMpz were metalated with Fe(acac)3 and
Cu(acac)2 to give the metalated MOFs UMCM-1-AMFesal
and UMCM-1-AMCupz. To further exploit this system,
metalation of UMCM-1-AMpz and UMCM-1-AMsal with
different metals was investigated, and the resulting MOFs
were examined as Lewis acid catalysts. After several attempts
with different metal sources, it was determined that UMCM-
1-AMpz and UMCM-1-AMsal could be cleanly metalated
with In(acac)3. Similar to previous studies with Cu(acac)2
and Fe(acac)3, In3þ binding was confirmed by atomic absorp-
tion (AA) analysis (Table 1). UMCM-1-AMpz, UMCM-
1-AMsal, and UMCM-1-NH2 exposed to 1 equiv of In(acac)3
were found to contain 3.76, 2.96, and 0.5 wt % of In3þ. These
values, when compared with theoretical loadings (e.g., one
metal ion per chelator site), indicate that more than 70% of
the metal binding sites are occupied for both UMCM-
1-AMpz and UMCM-1-AMsal, while only residual binding
is observed for the unmodified UMCM-1-NH2.
UMCM-1-AMInpz and UMCM-1-AMInsal were further
examined by PXRD, TGA, and dinitrogen gas sorption.
The PXRD patterns of UMCM-1-AMInpz and UMCM-
1-AMInsal were found to be identical to the parent MOF
(Figure 1), thus confirming that the metalation had no effect
on the structural stability of the framework. Single crystal
X-ray diffraction data of UMCM-1-AMInpz revealed the
framework consists of octahedral Zn4O SBUs coordinated by
four BTB and two NH2-BDC ligands. The MOF has large
hexagonal pores framed by the BTB ligands and smaller pores
defined by both BTB and NH2-BDC ligands (Supporting
Information, Figures S1 and S2). The framework topology,
along with the cell parameters (space group = P63/m; a =
Figure 1. PXRD comparison of UMCM-1-NH2 (black), UMCM-
1-AMInpz (red), and UMCM-1-AMInsal (blue).
aldehyde and silyl enol.24,48 Notably, UMCM-1-AMFesal
exhibited exceptional catalytic activity with low Fe3þ load-
ings. In an attempt to identify new catalysts, UMCM-
1-AMCupz, UMCM-1-AMInpz, and UMCM-1-AMInsal
were tested for activity in the Mukaiyama aldol reaction
under the previously reported conditions. We anticipated that
these MOF catalysts might show comparable or better
activity based solely on increased metal loading relative to
UMCM-1-AMFesal; however, little or no reactivity was seen
with these other metalated MOFs despite attempts to increase
the catalyst loading and reaction time (data not shown).
With this observation in hand, other Lewis acid catalyzed
reactions were explored, in particular reactions that had not
been widely studied with other MOF catalysts.1,2 Epoxide
ring-openings, which are important processes for generat-
ing stereocontrolled organic intermediates, are an attractive
target.49 MOF catalysts are appealing platforms for epoxide
ring-opening reactions because these reactions generally
occur in the presence of a Lewis acid catalyst and a nucleo-
phile (e.g., thiols, alcohols, aromatic amines) under fairly
mild conditions.44,50 Many homogeneous catalysts have been
utilized for epoxide ring-openings, including metal salen
complexes (e.g., Cr3þ, Co2þ, Mn2þ 49,51
and metal salts with
)
or without additives.45 One of the biggest challenges in these
ring-opening reactions is regioselective control as the nucleo-
phile can attack either position of the epoxide ring.52
According to recent MOF reviews,1,2 only three MOFs
have been explored as epoxide ring-opening catalysts. In one
report, Cu(asp)(bpe)0.5(H2O)0.5(MeOH)0.5 (asp= aspartic
acid, bpe=1,2-bis(4-pyridyl)ethylene) was protonated in a
˚
˚
postsynthetic manner to produce Cu(D/L-asp)bpe0.5(HCl)-
b = 41.3685 A, c = 17.5097 A; R = β = 90°, γ = 120°; V =
25950 A ) are similar to the parent MOF.38 A nitrogen atom
3
˚
(H2O) as an active Bronsted acid catalyst for methanolysis of
e
cis-2,3-epoxybutane.29 The MOF was found to be hetero-
geneous and gave the expected alcoholysis product with
yields as high as 70%, but the MOF was unable to catalyze
the reaction with larger epoxides and alcohols because of a
was located and assigned on the benzenedicarboxylate ligand;
however, the modified substituents could not be located in the
electron density map because of incomplete modification of
the sites and positional disorder of the modified NH2-BDC
ligand. Both metalated MOFs were stable up to 400 °C
(Supporting Information, Figure S3), and BET surface area
measurements confirmed the modified MOFs remained
highly porous with surface areas of ∼3200 m2/g.
(48) Mukaiyama, T.; Banno, K.; Narasaka, K. J. Am. Chem. Soc. 1974,
96, 7503–7509.
(49) Martinez, L. E.; Leighton, J. L.; Carsten, D. H.; Jacobsen, E. N.
J. Am. Chem. Soc. 1995, 117, 5897–5898.
(50) Paterson, I.; Berrisford, D. J. Angew. Chem., Int. Ed. 1992, 31, 1179–
1180.
(51) Jacobsen, E. N. Acc. Chem. Res. 2000, 33, 421–431.
(52) Bergmeier, S. C. Tetrahedron 2000, 56, 2561–2576.
UMCM-1-AMFesal was previously determined to be an
active, robust catalyst for the Mukaiyama aldol reaction,
which involves C-C bond formation between an aromatic

6770 Inorganic Chemistry, Vol. 49, No. 14, 2010
Tanabe and Cohen
Figure 2. 1H NMR supernatant comparison between different MOF materials with epoxide and TMSN3. Epoxide starting material is indicated by black
squares, alcohol product is indicated by red circles, and TMS protected product is indicated by blue circles.
Scheme 2. MOF Catalyzed Epoxide Ring-Opening Reactions with
a doublet at 1.25 ppm, which corresponds to the methyl
protons (Figure 2). Analysis of the reaction supernatants by
TMSN3 as the Nucleophilea
1H NMR showed the presence of two new multiplets at
3.6 and 3.8 ppm as well as two new doublets in the upfield
region, thus confirming generation of the ring opened
product (Figure 2). Surprisingly, the product did not match
the reported spectra of 2-azido-3-(trimethylsilyloxy)butane,
although trace amounts of the TMS product were detected.
A new peak was also present at 2.05 ppm that had equivalent
a The MOF catalysts produce either a mixture of the TMS protected
and β-azido alcohol products or only the β-azido alcohol.
integration with the protons at 3.6 and 3.8 ppm. Examina-
tion of similar reactions in the literature between epoxides
with TMSN3 (vide infra) indicated that the product obtained
restricted pore size. Two other MOFs, [Cu2(5,50BDA)2-
was the corresponding β-azido alcohol.44 Synthesis of the
(H2O)2]MeOH 2H2O (BDA = 2,20-dihydroxy-1,10-binaph-
3
β-azido alcohol generally requires an additional deprotec-
tion step; however, in situ deprotection takes place with the
MOF catalyst.
thalene-5,50-dicarboxylate) and Cu(Bpy(H2O)2(BF4)2(bpy)
(denoted as Cu-MOF), were examined as catalysts for amino
alcohol synthesis.30,31 Both MOFs were found to be size selec-
tive, but they displayed modest conversions (<50%) and no
evidence was provided about robustness of these catalysts.
Given the encouraging results from these previously re-
ported systems, the metalated UMCM MOFs were analyzed
for their performance as epoxide ring-opening catalysts. We
chose to focus on amino alcohols, which are important
intermediates for the development of pharmaceuticals, chiral
auxiliary reagents, and are common backbones in natural
and synthetic products.52 As a starting point, cis-2,3-epoxy-
butane and trimethylsilylazide (TMSN3) were selected as
the epoxide and nucleophile. The reaction between epoxides
and TMSN3 generally leads to the formation of TMS protec-
ted alcohols, which are subsequently deprotected and re-
duced to give the corresponding amino alcohol (Scheme 2).39
Although cis-2,3-epoxybutane and TMSN3 have been re-
ported with other catalysts,44 this particular combination has
not been demonstrated using a MOF system.
1
Monitoring of these reaction mixtures by H NMR (RT,
24 h) showed that all the metalated MOFs could catalyze the
reaction, but with varying degrees of activity (Figure 3). On
average, UMCM-1-AMInpz had the highest conversion un-
der the reported conditions at 78%. UMCM-1-AMInsal was
less active with 56% conversion, while UMCM-1-AMFesal
and UMCM-1-AMCupz both had significantly lower con-
versions of 30% and 11%, respectively. On the basis of these
results alone, the In3þ metalated MOFs appear to be more
active when compared with Fe3þ and Cu2þ. To confirm that
the catalysis was occurring because of the presence of the
Lewis acid centers and not just the unmodified MOF itself,
several control reactions were performed. UMCM-1-NH2,
UMCM-1-NH2 treated with In(acac)3, UMCM-1-AMpz,
and UMCM-1-AMsal were prepared and examined under
similar reaction conditions. UMCM-1-NH2, UMCM-1-NH2
treated with In(acac)3, and UMCM-1-AMsal all showed
<10% conversion, and UMCM-1-AMpz was slightly more
active with ∼16% conversion. None of the control reactions
displayed the high activity observed with UMCM-1-AMInpz.
Reactions with a different nucleophile were examined
as well. Aniline was chosen because it had been previously
used in other MOF systems,30,31 which provided a point of
comparison (Scheme 2). Similar to the TMSN3 experiments,
Reactions with metalated MOFs (UMCM-1-AMInpz,
UMCM-1-AMInsal, UMCM-1-AMCupz, UMCM-1-
AMFesal) were conducted in CDCl3 for 24 h at RT. After
24 h, the crude supernatants were directly analyzed by
1H NMR to determine percent substrate conversion. The
1H NMR of neat cis-2,3-epoxybutane shows a multiplet
at 3.05 ppm, which is indicative of the meso protons, and

Article
Inorganic Chemistry, Vol. 49, No. 14, 2010 6771
Figure 3. 1H NMR supernatant comparison between metalated MOF catalysts with epoxide and TMSN3. Epoxide starting material is indicated by black
squares, alcohol product is indicated by red circles, and TMS protected product is indicated by blue circles.
Figure 4. 1H NMR supernatant comparison between metalated MOF catalysts with epoxide and aniline. Epoxide starting material is indicated by black
squares, and the product is indicated by red circles.
Scheme 3. MOF Catalyzed Epoxide Ring-Opening Reactions with Aniline As the Nucleophile
1
the crude reaction supernatant was analyzed by H NMR
after reaction for 24 h at RT. Product formation was easily
confirmed by the presence of two new multiplets at 3.8 and
4.7 ppm along with a new broad peak at around 2.6 ppm,
which corresponds to the hydroxyl group of the product
(Figure 4). These peak shifts were identical to reported
β-amino alcohol spectra.45 After testing all the metalated
MOFs, the activity of the catalyst was found to be different
with aniline as the nucleophile (Scheme 3). In this case,
UMCM-1-AMInsal had the highest conversion at 86%,
followed by UMCM-1-AMFesal with 34% and UMCM-
1-AMInpz with 30% (Table 2). UMCM-1-AMCupz did not
show any activity despite previous evidence that Cu2þ MOFs
are capable of promoting epoxide aminolysis.30,31 Analogous
control experiments, as described with TMSN3 above, were
also performed with aniline as the nucleophile. All of the
control reactions showed less than ∼13% conversion.
Additional controls were performed to confirm that the
metalated MOFs were indeed the catalytic source for
epoxide ring-opening. Initial control reactions with no
MOF showed that epoxide ring-opening did not occur
at all. Likewise, removal of the metalated MOFs (e.g.,
UMCM-1-AMInpz and UMCM-1-AMInsal) from the
reaction supernatant completely halted the progress of
the reaction (Supporting Information, Table S3, Figures
S6 and S7). Control reactions with In(acac)3 and various

6772 Inorganic Chemistry, Vol. 49, No. 14, 2010
Tanabe and Cohen
a
Table 2. Percent Conversions of MOF Catalyzed Reactions between Different Epoxides and TMSN3
epoxide
UMCM-1-
NH2
UMCM-1-
NH2 þ In3þ
10 ( 4
4 ( 3
27 ( 7
1 ( 1
no rxn
UMCM-1- UMCM-1-
AMpz
UMCM-1-
AMInpz
UMCM-1-
AMInsal
UMCM-1-
AMCupz
UMCM-1-
AMFesal
R
R0
no MOF
AMsal
Me
-(CH2)3-
H
Ph
Ph
Me
no rxn
no rxn
no rxn
no rxn
no rxn
8 ( 6
2 ( 1
19 ( 8
2 ( 2
no rxn
16 ( 7
4 ( 2
7 ( 4
78 ( 9b
53 ( 3b
70 ( 2
48 ( 8
no rxn
56 ( 2 b
43 ( 5b
54 ( 6
17 ( 10
no rxn
11 ( 5
3 ( 4
32 ( 7
4 ( 2
no rxn
35 ( 14b
26 ( 3
53 ( 1
9 ( 2
5 ( 4
Ph
41 ( 9
1 ( 1
no rxn
19 ( 11
3 ( 1
no rxn
Phc
Phd
no rxn
a All values are the result of three independent experiments. b Based on four independent trials. c Cis. d Trans.
Table 3. Percent Conversions of MOF Catalyzed Reactions between Different Epoxides and Anilinea
epoxide
UMCM-1-
NH2
UMCM-1-
NH2 þ In3þ
UMCM-1- UMCM-1-
AMpz
UMCM-1-
AMInpz
UMCM-1-
AMInsal
UMCM-1-
AMCupz
UMCM-1-
AMFesal
R
R0
no MOF
AMsal
Me
-(CH2)3-
H
Ph
Ph
Me
no rxn
no rxn
no rxn
no rxn
no rxn
no rxn
no rxn
5 ( 4
no rxn
no rxn
9 ( 7
8 ( 6
13 ( 2
1 ( 1
12
no rxn
no rxn
30 ( 7
20 ( 7
49 ( 1
no rxn
no rxnb
86 ( 11
58 ( 12
∼99
1 ( 2
34 ( 6
22 ( 8
43 ( 1
no rxn
no rxn
1 ( 1
2 ( 2
no rxn
16 ( 1
no rxn
no rxnb
Ph
14 ( 3
no rxn
no rxnb
20 ( 1
no rxn
no rxnb
Phc
Phd
6 ( 2
no rxn
a All values are the result of three independent experiments. b Based on two independent experiments. c Cis. d Trans.
ligands (salicylamide, 3-hydroxyphthalic anhydride, and
2,3-pyrazinedicarboxylic anhydride) were also set up to
determine if trace amounts of the metal source and ligands
could act as homogeneous catalysts. Despite using 10 mol %
loadings, no significant reactivity was seen from any of the
controls, indicating that In(acac)3 metal source alone is a
poor catalyst for epoxide ring-opening.
oxide was examined under the same reaction conditions
as used for cis-2,3-epoxybutane. For the TMSN3 reac-
tions, UMCM-1-AMInpz had the highest conversion
of 53%, followed by UMCM-1-AMInsal with 43%,
UMCM-1-AMFesal with 26%, and UMCM-1-AMCupz
with <5%. Switching the nucleophile to aniline resulted
in UMCM-1-AMInsal having the highest conversion at
58%, while UMCM-1-AMInpz and UMCM-1-AMFesal
both had similar conversions of ∼20% and UMCM-
1-AMCupz gave no conversion.
The metalated MOFs were more active with styrene
oxide in comparison with cyclopentene oxide. UMCM-
1-AMInpz gave the highest conversion of 70% for the
reaction with TMSN3 while UMCM-1-AMInsal and
UMCM-1-AMFesal both gave conversions of ∼53%.
Furthermore, UMCM-1-AMCupz showed 32% conver-
sion while the other MOF controls had conversions as
high as 40%. Despite the unexpectedly high conversions
obtained with the control MOFs, the results still indicate
that the metalated systems, such as UMCM-1-AMInpz,
show the highest activity.
Similar trends were also seen when cyclopentene oxide
and styrene oxide were coupled with aniline. UMCM-
1-AMCupz and the control MOFs were still active, but
showed lower conversions of ∼20%, which may be due to
the larger size of aniline in comparison with TMSN3.
UMCM-1-AMInpz and UMCM-1-AMFesal had similar
conversions between 40 and 50%. UMCM-1-AMInsal
essentially catalyzed the reaction to completion (∼99%);
however, two products were discovered in the reaction
mixture. The high activity of UMCM-1-AMInsal unex-
pectedly resulted in two products: the major product
(68%) was determined to be 2-phenylamino-2-phenyl-
ethanol by 1H and 13C NMR.43,46,53 The minor product
(32%) was identified as the bis-alkylated product, 2,20-
(phenylazanediyl)bis(2-phenylethanol), as confirmed
by high resolution mass spectrometry and 13C NMR.
Bis-alkylation between epoxides and aniline has been
Stability of MOF Catalysts. Overall, UMCM-1-
AMInpz and UMCM-1-AMInsal were found to be the
most active catalysts, but for different ring-opening reac-
tions. UMCM-1-AMInpz was highly active for the reac-
tion with TMSN3 while UMCM-1-AMInsal showed a
preference with aniline as the nucleophile. Several tests
were performed with the two MOFs to understand their
chemical stability postcatalysis. Overall, both MOFs
remained stable after three catalytic cycles with complete
retention of reactivity. On average, UMCM-1-AMInpz
maintained 76% conversion over three cycles, while
UMCM-1-AMInsal gave 82% conversion (Supporting
Information, Table S1). It should be noted that UMCM-
1-AMInpz and UMCM-1-AMInsal represent extremely
low loadings of In3þ under the reaction conditions used.
As mentioned earlier, UMCM-1-AMInpz and UMCM-
1-AMInsal have metal loadings of 3.76 and 2.96 wt %,
which corresponds to 0.5 mol % (0.00053 mmol) and
0.4 mol % (0.00041 mmol) of In3þ. PXRD indicated that the
MOFs maintained their structural integrity after catalysis
(Supporting Information, Figures S4 and S5). Samples were
prepared for AA analysis pre- and postcatalysis to determine
if metal was leaching from the MOF. On the basis of
three independent samples, it was determined that <10%
leaching of either Zn2þ or In3þ occurs for UMCM-1-
AMInpz and UMCM-1-AMInsal after three catalytic cycles
(Supporting Information, Table S2).
Size Selectivity. The metalated MOFs were examined
with other epoxides of varying sizes and shapes to deter-
mine if the MOF catalysts demonstrated any substrate
selectivity (Tables 2 and 3). Other epoxides that have been
commonly studied in literature, such as cycloalkane ep-
oxides and terminal epoxides were employed, specifically
cyclopentene oxide and styrene oxide. Cyclopentene
(53) Bedore, M. W.; Zaborenko, N.; Jensen, K. F.; Jamison, T. F. Org.
Process Res. Dev. 2010, 14, 432–440.

Article
Inorganic Chemistry, Vol. 49, No. 14, 2010 6773
previously reported; however, the specific product formed
here, 2,20-(phenylazanediyl)bis(2-phenylethanol), has not
been previously reported, to the best of our knowledge
(Supporting Information, Figure S31).53 As expected,
running the same reaction with increased amounts of
styrene oxide (2 equiv) results in the production of more
2,20-(phenylazanediyl)bis(2-phenylethanol), while using
increased amounts of aniline (4 equiv) results in ∼90%
conversion to 2-phenylamino-2-phenylethanol (Supporting
Information, Figure S32).
To complete the size selectivity study, cis- and trans-
stilbene oxide were examined. cis-Stilbene oxide under-
went ring-opening in the presence of TMSN3 with some
of the metalated MOF catalysts (Table 2). UMCM-
1-AMInpz gave 48% conversion while UMCM-1-AMIn-
sal, UMCM-1-AMFesal, and UMCM-1-AMCupz had
conversions of 17%, 9%, and 4%, respectively. The
overall activity dropped drastically for all MOFs upon
switching the nucleophile to aniline (Table 3). UMCM-
1-AMInsal was the only MOF that showed some catalytic
activity, but the yield was very low (6%). In contrast
to cis-stilbene oxide, trans-stilbene oxide did not undergo
ring-opening with either nucleophile in the presence of
any of the MOF catalysts. To ensure that trans-stilbene
oxide was not poisoning the catalyst, the MOF catalysts
(e.g., UMCM-1-AMInpz and UMCM-1-AMInsal) were
recycled and shown to still be competent for the reac-
tion between cis-2,3-epoxybutane and both nucleophiles
(TMSN3 with UMCM-1-AMInpz and aniline with
UMCM-1-AMInsal). All control reactions showed neg-
ligible activity with cis- or trans-stilbene oxide and either
TMSN3 or aniline.
Size selectivity studies were also performed with differ-
ent aniline derivatives using UMCM-1-AMInsal as the
catalyst. As mentioned previously, UMCM-1-AMInsal
was found to be the most active material for the ring-
opening reaction between cis-2,3-epoxybutane and ani-
line. Therefore, the reaction of cis-2,3-epoxybutane with
aniline, 2-methylaniline, and 2,6-dimethylaniline in the
presence of UMCM-1-AMInsal was examined under
identical reaction conditions. All three aniline derivatives
were found to readily react with the epoxide, with 86%,
93%, and 56% conversion for aniline, 2-methylaniline,
and 2,6-dimethylaniline, respectively (Supporting Infor-
mation, Figure S33). The sterically encumbered 2,6-di-
methylaniline is not as good of a substrate, but it does not
completely shut down the reaction, and hence, the MOF
catalyst appears reasonably tolerant of substituents on
the nucleophile for this reaction.
the MOFs display different catalytic activity based on their
metal ion and supporting ligand combinations.
The difference in catalytic activity is most clearly demon-
strated with UMCM-1-AMInpz and UMCM-1-AMInsal.
Both MOFs have similar In3þ loadings and microporo-
sities, yet their organic substituent, a pyrazinedicarboxylate
versus a salicylate, dictates their catalytic activity and
specificity for different reactions. UMCM-1-AMInpz was
found to be best when using TMSN3 as the nucleophile for
epoxide ring-opening, while UMCM-1-AMInsal proved to
be better when the nucleophile was aniline. As shown in
Tables 2 and 3, the most active MOF catalyst for a given
reaction (i.e., nucleophile) is the same for every epoxide
substrate, highlighting the importance of having a specific
metal-ligand combination active site. In general, both
In3þ loaded MOFs were better catalysts for these epoxide
ring-opening reactions than either UMCM-1-AMFesal or
UMCM-1-AMCupz.
Catalytic activity can also be affected by the orientation
and accessibility of the catalytic sites (e.g., metal-ligand
combinations) within the framework. Previous studies with
UMCM-1-NH2 indicated that the efficiency of PSM is
influenced by a combination of both pore and reagent size/
shape.36 The results obtained here show that catalysis can be
similarly affected. The orientation and accessibility of the
catalytic site is dependent on how the metal-ligand unit fits
within the pore based on the size and shape of the chelating
ligand, the metal ion coordination geometry, and the size and
shape of the pore. Depending on the outcome, the position of
the catalytic site can affect how the substrates (e.g., epoxide
and nucleophiles) interact with the catalytic site and with
each other. For this particular MOF catalytic system, the
exact composition of the catalytic site has not been fully
characterized (i.e., the ancillary ligands on the metal center
have not been determined) and the metal sites have not been
located by single crystal X-ray diffraction (to determine
their precise orientation within the MOF). However, the
reactions with different substrates suggests certain proper-
ties about the catalytic site. For example, the metalated
MOFs are catalytically competent for the ring-opening of
cis-stilbene oxide, but not trans-stilbene oxide. As men-
tioned previously, the UMCM MOFs have two types of
pores: a large hexagonal pore and a smaller pore. On the
basis of the framework topology and the crystal struc-
ture obtained of UMCM-1-AMInpz, the catalytic site can
be directed into either pore. On the basis of the catalysis
results with cis/trans epoxides, we infer that the catalytic
site is more localized within the smaller pore, where access
and nucleophilic attack of the trans-epoxides would be
more constrained. This is consistent with the structure of
the UMCM MOFs (Supporting Information, Figure S2),
where the BDC ligands are tightly arranged around the
smaller pores, while the BTB ligands frame the larger pores.
Although their sizes are similar, the shape of cis- and trans-
stilbene oxide are quite different, and this likely affects the
ability of the latter to effectively interact with and be
activated by the catalytic sites. While trans-stilbene may
be able to access the interior of the MOF, it may be (a)
sterically unable to interact with the metal site, or it (b) can
interact with the metal site, but is in an orientation that
makes nucleophilic attack unfavorable.
Discussion
PSM is an advantageous approach for modifying materials
that would otherwise be difficult or impossible to synthesize
through conventionalsolvothermal MOF synthesis. Multiple
functionalities can be incorporated into the MOF with
better control over substituent type and degree of modifica-
tion.20 For example, UMCM-1-NH2 was transformed into
UMCM-1-AMpz and UMCM-1-AMsal, which were furth-
ered metalated with Fe3þ, Cu2þ, or In3þ. As a result, four
MOFs catalysts with different metals and supporting ligands
were readily produced by simply modifying a well-characteri-
zed MOF. It is important to note that all these MOFs are
isostructural and have similar thermal stabilities; however,
Literature reports for the catalytic ring-opening of stil-
bene oxide are scarce. Unfortunately, reactions involving the

6774 Inorganic Chemistry, Vol. 49, No. 14, 2010
Tanabe and Cohen
ring-opening with TMSN3 were not found;54 however,
several reports with aniline are available.45,47 According to
one literature report, cis- and trans-stilbene oxide successfully
undergo ring-opening with aniline in the presence of
Cr(salen)Cl with yields of ∼90%.47 Under similar reaction con-
ditions as described for the MOF catalysts, this Cr(salen)Cl
catalyst can couple both cis- and trans-stilbene oxides with
either nucleophile. The Cr(salen)Cl catalyst gave yields of
94% and 68% with cis- and trans-stilbene oxide, respectively,
in the presence of TMSN3. Similarly, with aniline, conver-
sions of 43% and 60% were obtained with cis- and trans-
stilbene oxide. To highlight the differences in substrate
selectivity of the soluble versus MOF-based catalyst,
Cr(salen)Cl and UMCM-1-AMInpz were tested as catalysts
with a 1:1 mixture of cis- and trans-stilbene oxide using
TMSN3 as the nucleophile. Fortunately, the epoxide start-
ing materials have sufficiently distinct 1H NMR spectra
(particularly the resonances at 4.38 ppm and 3.89 ppm for
cis- and trans- epoxides, respectively) to monitor the conver-
sion of each substrate within the mixture. As expected,
Cr(salen)Cl catalyzed the conversion of both cis- and trans-
stilbene oxide, while UMCM-1-AMInpz only catalyzed the
ring-opening of cis-stilbene oxide (Supporting Information,
Figures S34 and S35). To further confirm the selectivity, the
same experiment was performed with cis-2,3-epoxybutane
and trans-2,3-epoxybutane. By itself, trans-2,3-epoxybutane
is essentially ∼99% converted with Cr(salen)Cl, while only
∼20% conversion is observed with UMCM-1-AMInpz
as the catalyst (Supporting Information, Figure S36). A 1:1
mixture of cis- and trans-2,3-epoxybutane with TMSN3 was
examined with both Cr(salen)Cl and UMCM-1-AMInpz. In
the mixed system, both epoxides were ∼99% converted using
Cr(salen)Cl; however, cis-2,3-epoxybutane was the dominant
epoxide turned over with UMCM-1-AMInpz (Supporting
Information, Figure S37) with 1H NMR analysis suggesting
a conversion of ∼90% for cis and ∼10% for trans. This
shows that the MOF-based UMCM-1-AMInpz catalyst
elicits unique, stereochemistry-based substrate selectivity,
not achieved by a soluble catalytic system, while maintaining
comparable activity at a lower catalyst loading.
Conclusions
The MOF UMCM-1-NH2 was functionalized with chelat-
ing groups and metal ions to produce active Lewis acid
catalysts. The systematic investigation of several UMCM-
1-NH2 derivatives have revealed two MOFs (UMCM-
1-AMInpz and UMCM-1-AMInsal) to be excellent catalysts
for epoxide ring-opening reactions to generate β-azido alco-
hols and β-amino alcohols. This is the first description of a
MOF catalyst capable of preparing β-azido alcohols. These
MOF catalysts show excellent thermal and structural stability,
retain good activity over several cycles, and display selectivity
with respect to the epoxide substrate not seen with other Lewis
acid catalysts. By using PSM, a single MOF platform has been
modified with different combinations of metal ions and
chelating ligands to produce MOFs with distinctive catalytic
activities. Future work will continue to employ PSM as a
method toward fine-tuning and enhancing the catalytically
properties of MOFs. For example, functionalization with
multiple substituents that can impart additional interac-
tions (e.g., chirality, hydrogen bonding, etc.) may be used to
improve catalytic activity as well as substrate specificity and
regioselective reactivity. It may be possible, with the optimum
combination of substituents, to discover MOF catalysts with
the best qualities of both homogeneous and heterogeneous
systems and with biomimetic activity and selectivity.
Acknowledgment. We thank Sergio J. Garibay for help-
ful discussions and Dr. Y. Su for performing the mass
spectrometry experiments. This work was supported by
UCSD, the National Science Foundation (CHE-0546531,
new MOF synthesis; CHE-9709183, CHE-0116662 and
CHE-0741968 instrumentation grants), and the Depart-
ment of Energy (DE-FG02-08ER46519, MOF modifica-
tion for gas sorption).
Supporting Information Available: Further details are given in
Figures S1-S33 and Tables S1-S4. This material is available
free of charge via the Internet at http://pubs.acs.org.
(54) Lupattelli, P.; Bonini, C.; Caruso, L.; Gambacorta, A. J. Org. Chem.
2003, 68, 3360–3362.