Lewis Acid Activation of a Hydrogen Bond Donor Metal-Organic Framework for Catalysis

Article abstract of DOI:10.1021/acscatal.6b00424

A new metal-organic framework (MOF) composed of urea-containing tetracarboxylate struts was synthesized, and its hydrogen bonding capabilities were evaluated. The catalytic performance of this heterogeneous framework is enhanced through preactivation with silyl Lewis acids, leading to Friedel-Crafts reaction rates greater than those of common homogeneous hydrogen bond donors.

Full text of DOI:10.1021/acscatal.6b00424

Research Article
pubs.acs.org/acscatalysis
Lewis Acid Activation of a Hydrogen Bond Donor Metal−Organic
Framework for Catalysis
Edward A. Hall, Louis R. Redfern, Michael H. Wang, and Karl A. Scheidt*
Department of Chemistry, Center for Molecular Innovation and Drug Discovery, Northwestern University, 2145 Sheridan Road,
Evanston, Illinois 60208, United States
S
* Supporting Information
ABSTRACT: A new metal−organic framework (MOF)
composed of urea-containing tetracarboxylate struts was
synthesized, and its hydrogen bonding capabilities were
evaluated. The catalytic performance of this heterogeneous
framework is enhanced through preactivation with silyl Lewis
acids, leading to Friedel−Crafts reaction rates greater than
those of common homogeneous hydrogen bond donors.
KEYWORDS: metal−organic framework, catalysis, hydrogen bond donor, urea MOF, Lewis acid
limited innovation toward novel MOF topologies, especially
those with large pore sizes that do not require covalent
postsynthetic modifications.17 In addition, the treatment of
MOFs containing “metalloligand” struts with catalytically active
metals had also led to interesting new heterogeneous
catalysts.18 In this vein, Cohen and Tanabe demonstrated the
incorporation of Cu, Fe, and In into MOF architectures to
catalyze epoxide ring-opening reactions,19 and Ma has reported
the binding of Al(III) using Brønsted acid chelators.20
We imagined that the optimal combination of Lewis acid and
hydrogen bond donor could produce a stable MOF catalyst
with enhanced catalytic activity for new bond-forming reactions
(Figure 1). A novel MOF architecture (NU-GRH-1) using urea
functionalized tetracarboxylate struts (Scheme 1) emerged as a
new scaffold for these studies. The unique bent geometry and
unsymmetrical nature of these struts enabled the generation of
a highly porous framework capable of accepting substrates
larger than those of previous H-bond donor MOF
reports.13,15b,16 We report herein a Lewis acid activation
approach of urea functionalized MOF scaffolds and their
enhanced catalytic activity versus related homogeneous urea
catalysts.
INTRODUCTION
■
Hydrogen bond donor molecules have become selective
organocatalysts for a wide range of applications because of
their tunable reactivity, rigid conformation, and stable
functionality.1 Their emergence as powerful tools for organic
synthesis has led to novel H-bond donors with unique
capabilities enabling new reactions and approaches.2 Since
Schreiner and Wittkopp reported seminal applications of H-
bond-donating thioureas,3 much effort has been directed
toward increasing the potency of these unique organocatalysts.4
The majority of this work has been focused on the
development of new classes of H-bond donor catalysts
including electronically and conformationally activated squar-
amides,5 quinazolines,6 aminobenzothiadiazines,6 2-amino-
benzimidazoles,7 2-aminopyridiniums,8 and silanediols.9 Alter-
natively, dual catalytic activation with Brønsted10 or Lewis
acids11 has also been shown to significantly enhance the
inherent reactivity of simple (thio)urea scaffolds. Urea catalysts
in particular are known for their high thermal and chemical
stability, but their reactivity is often critically attenuated by
undesired catalyst oligomerization (quenching).12 Our recent
efforts to prevent self-quenching by physically separating urea
H-bond donors have shown that modest rate enhancements
can be obtained by incorporating urea catalysts into metal−
organic frameworks (MOFs).13 However, obtaining high levels
of reactivity from a heterogeneous MOF catalyst has proven to
be a challenging endeavor because of restrictive pore sizes and
the inherently low activity of urea catalysts.14
RESULTS
■
The synthesis of urea strut 6 began with commercially available
bromide 1, which was deaminated21 to yield dibromide 2. A
Suzuki coupling with 4-acetylphenylboronic acid followed by
reduction of the nitro functionality with Pd/C provided aniline
4. Two equivalents of the aniline were combined with
carbonyldiimidazole to provide urea 5. A mild methyl ester
An alternative strategy for the installation of catalytic
components into MOFs is postsynthetic modification, which
allows the introduction of new functionality that could not
otherwise be incorporated through de novo solvothermal
synthesis. This approach has been used to incorporate ureas15
and thioureas;16 however, dependence upon this strategy has
Received: February 10, 2016
Revised: April 11, 2016
© XXXX American Chemical Society
3248
DOI: 10.1021/acscatal.6b00424
ACS Catal. 2016, 6, 3248−3252

ACS Catalysis
Research Article
Figure 1. Enhanced hydrogen bonding catalyst strategy.
a
Scheme 1. Synthesis of NU-GRH-1
Figure 2. NU-GRH-1 crystal structure: (top panel) view along c axis;
(bottom panel) view along a axis rotated 45°. Urea H atoms added for
clarity.
Dipyridyl ligands within the H-bonding pores of NU-GRH-1
display substantial bending23 (159.7°) potentially caused by the
coordinate bond angles required for Zn metal chelation at both
termini. Cambridge Structure Database (CSD) searches based
on these angular parameters provided previously reported
MOF structures supporting the validity of substantial dipyridyl
bending using Zn,24 Cd,25 and Cu.26 Notably, ligand bending in
NU-GRH-1 is 5° greater than any previously disclosed MOF
structure. Variations in the displacement parameters within the
bent ligands versus the linear BPy ligands in NU-GRH-1
(ORTEP) further suggest the presence of bent pyridyl struts
(see Supporting Information for detailed explanation).
The urea functional groups positioned within NU-GRH-1
were first activated for catalysis by removing the Lewis basic
DMF molecules bound to the urea N−H bonds. The solvent
was exchanged daily with MeNO2 at 50 °C over 1 week. The
crystals were dried under vacuum, digested in 3% D2SO4/
DMSO-d6, and characterized by 1H NMR spectroscopy,
confirming a 1:1 strut-to-BPy ratio. We then employed NU-
GRH-1 in a Friedel−Crafts (FC) reaction with β-nitrostyrene
and indole as a simple benchmark to assess catalytic activity
(Table 1). Our first attempts using dried NU-GRH-1 crystals
surprisingly resulted in no reaction (data not shown). PXRD of
the solvent-free MOF did not give sharp peaks consistent with
the solvent-saturated MOF, suggesting collapse of the pores
a
(a) Pd(OAc)2, 4-acetylphenylboronic acid, Na2CO3, PPh3, THF,
reflux, 18 h; (b) Pd/C, H2, THF, rt, 6 h; (c) CDI, THF, reflux, 18 h;
(d) KOSiMe3, THF, reflux, 18 h; (e) Zn(NO3)2[6(H2O)], 4,4′-
dipyridyl, DMF, 95 °C, 2 d.
cleavage using potassium trimethylsilanolate provided scalable
yields of tetracarboxylate strut 6. This synthetic protocol
provides multigram quantities of the organic strut (>10 g)
without the need for chromatographic purification (see
Supporting Information for details). The solvothermal syn-
thesis of NU-GRH-1 was carried out on gram scale with strut 6,
Zn(NO3)2, and 4,4′-dipyridyl (BPy)22 yielding large, clear-to-
pale yellow crystalline rods having the empirical formula
Zn2(BPy)(6).
These MOF crystals were then subjected to single-crystal X-
ray crystallographic analysis (Figure 2). NU-GRH-1 exists in
the P42/mnm space group and contains large oblong pores
(widest aperture dimensions: 31.6 × 14.0 Å2; distances
measured between atoms) with the N−H bonds of the ureas
pointing inward for catalysis. Powder X-ray diffraction spectra
(PXRD) of bulk samples matched that of the simulated spectra
generated from the cif file (see Supporting Information).
3249
DOI: 10.1021/acscatal.6b00424
ACS Catal. 2016, 6, 3248−3252

ACS Catalysis
Research Article
silicon-based Lewis acids provided the greatest increase of
catalytic activity, with trimethylsilyl trifluoromethanesulfonate
(TMS-OTf) being the most effective. Control experiments with
TMS-OTf without MOF revealed a significant homogeneous
background rate. Furthermore, we observed partial MOF
degradation upon treatment with TMS-OTf leading to
precipitous drops in reactivity when recycling the catalyst.
Consequently, trimethylsilyl chloride (TMS-Cl) was chosen as
the optimal additive moving forward (Figure 3, top panel).
Table 1. Indole Friedel−Crafts Activator Screen
Figure 3. (Top panel) Relative rates of Friedel−Crafts catalysis.
(Bottom panel) TMS-Cl preactivated NU-GRH-1 catalytic recycla-
bility.
Notably, attempts to activate homogeneous urea 8 provided
some rate enhancement (entry 16; Lewis acid was still present
during reaction) but could not replicate the same increase
observed with the MOF/TMS-Cl combination (entry 11;
Lewis acid washed away prior to reaction). Additionally, this
combination is a more active catalyst then strut 5 alone (entry
1) or 5 combined with TMS-Cl (entry 15).
MOF preactivation provides the benefit of washing away the
remaining Lewis acid in solution when compared to standard
homogeneous catalysts, thus preventing substrate degradation
or complexation with strongly Lewis acidic metals. Moreover,
NU-GRH-1/TMS-Cl showed robust recyclability (Figure 3,
bottom). Over five consecutive catalytic cycles, NU-GRH-1/
TMS-Cl maintained high reactivity, without any observable
drop in reaction rate or yield (determined by 1H NMR
spectroscopy). The 1H NMR spectra from unpurified reactions
using recycled TMSCl-activated MOF did not show signals
related to Si-Me (or Si-CH3) to avoid confusion of the possible
presence of TMS-OH or other such compounds.
Control experiments support the heterogeneous catalyst role
of NU-GRH-1 in the FC reaction. The use of solely TMS-Cl
(18 mol %) provides low levels of reactivity (Table 1, entry 14),
which in combination with washing protocols to remove TMS-
Cl from the NU-GRH-1 catalyzed reactions, supports the
conclusion that the Lewis acid additive does not provide rate
enhancement in solution independent of the MOF. Reactions
that included 2,6-ditert-butyl-4-methylpyridine (DTBMP, 20
mol %) during preactivation showed only a minor decrease in
reaction rate (Table 1, entry 13; see Supporting Information for
rate profile), reducing the likelihood of residual Brønsted acid
(presumably HCl) activating NU-GRH-1 or independently
catalyzing the reaction.27 Additionally, NU-GRH-1 in the
presence of TMS−OH (Table 1, entry 17) shows no significant
change in reaction rate compared to the unactivated framework.
a
Used to preactivate MOF then removed by toluene wash prior to
b
reaction. NMR yields using 2-methylnapthalene as internal standard.
c
Average yield between three separate batches of NU-GRH-1.
d
e
Reaction complete in 1.5 h. DTBMP (20 mol %) was used during
MOF preactivation then removed via toluene wash prior to reaction.
f
TMS-Cl (18 mol %) was present during reaction. DTBMP = 2,5-
ditert-butyl-4-methylpyridine.
after removal of solvent. To circumvent this problem, the as-
synthesized crystals were collected with a brief filtration to
remove excess solvent and used without any further drying. The
wet MOF crystals displayed notable catalytic activity, especially
when compared to modern homogeneous urea catalysts (Table
1, entries 1−4).
We then investigated strategies to enhance catalytic activity
of NU-GRH-1 by pretreatment with a range of Lewis acids.
NU-GRH-1 was soaked with a Lewis acid in toluene overnight
at 23 °C, at which point the Lewis acid solution was removed and
the crystals were thoroughly washed with toluene (5×). Boron,
titanium, and aluminum-based Lewis acids all provided
moderate rate enhancements (Table 1, entries 5−8). However,
3250
DOI: 10.1021/acscatal.6b00424
ACS Catal. 2016, 6, 3248−3252

ACS Catalysis
Research Article
Filtration of the MOF after 1 h halted further product
formation, indicating that the heterogeneous framework is the
primary catalyst (Figure 4). Furthermore, activated NU-GRH-1
demonstrated superior reactivity in comparison to other
common lewis acidic MOFs (MIL-100(Fe) and HKUST-1;
see Supporting Information).
between the silicon and urea carbonyl oxygen. The stability of
this interaction may be increased via encapsulation within a
heterogeneous framework. Direct urea silylation is unlikely
given that little effect of in situ generated acid is observed in
experiments employing exogenous base during preactivation.
Furthermore, N,O-bis(trimethylsilyl)acetamide (a strong silylat-
ing reagent) provided no enhancement of catalytic activity
when used to pretreat NU-GRH-1. Urea H-bond donors also
have a propensity to bind anions (such as chloride)28 which
may provide a scenario in which NU-GRH-1 activates the
Lewis acid. In this situation, catalytic activation would then
occur at the electron-deficient silicon center. Attempts at single-
crystal X-ray analysis of MOF/TMS-Cl combinations have thus
far been unable to determine the mode of Lewis acid binding.
Single-crystal data sets in combination with PXRD analysis
indicate the overall topology of NU-GRH-1 remains intact and
unaffected by TMS-Cl treatment. However, both ICP-OES
analysis and 1H NMR spectroscopy of the TMS-Cl activated MOF
indicate silicon is present after digestion of the material (see
Supporting Information for details). While the impact on
catalysis is clear, the precise molecular interactions in the MOF·
TMS-Cl combination responsible for the high levels of activity
are still under investigation.
Figure 4. MOF filtration experiment.
a
Table 2. Substrate Scope
CONCLUSION
■
Lewis acid activation of ureas is feasible through specific design
and judicious incorporation of hydrogen bond donor organo-
catalysts into a heterogeneous metal−organic structure. These
activated MOF catalysts are recyclable and can display superior
reactivity compared to their homogeneous urea counterparts.
The novel MOF topology includes an unusual BPy
conformation and contains large pores amenable to general
catalysis. Continued investigations with these activated
heterogeneous catalysts as well as new MOF strategies for
catalysis with hydrogen bond donors are ongoing.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.6b00424.
■
S
a
Isolated yields. Identical equivalencies to Table 1. Reactions heated at
General experimental details, synthesis of NU-GRH-1,
substrate characterization, ICP-OES and 1H NMR for
detection of Si-based activators, PXRD, bent dipyridyl
analysis, and single-crystal X-ray data (PDF)
Crystallographic data (CIF)
b
60 °C for 24 h. Reactions at 23 °C for 24 h and diluted to 0.5 M.
A substrate scope analysis was performed to probe the
tolerance of NU-GRH-1 to larger substrates (Table 2). In all
examples, NU-GRH-1 accepted expanded substrates and still
provided faster rates than homogeneous urea 8. Crushed MOF
did not provide any change in rate, suggesting that reactions
with large substrates are not exclusively catalyzed on the surface
of the MOF. Less reactive substrates including 4-bromoindole
(10), 3-methylindole (13), and N-tosyl imine (16) highlight
the expanded reactivity profile of NU-GRH-1 relative to
standard urea catalysts. These studies also expand upon our
scope of nitrostyrene electrophiles to include imines with
similarly satisfying results. A common notion is that the
inherent entropic penalties resulting from substrate transport
and active site accessibility in heterogeneous MOF catalysts
render them less reactive than their homogeneous counterparts.
However, these results suggest that new and enhanced
reactivity is indeed possible with proper catalytic design and
novel activation strategies.
AUTHOR INFORMATION
Corresponding Author
*E-mail: scheidt@northwestern.edu.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank AbbVie for support of this work. Single-crystal and
powder X-ray crystallographic experiments were carried out
with support from Amy Sarjeant and Charlotte Stern (NU). We
thank Prof. SonBinh Nguyen (NU) for helpful discussions.
REFERENCES
■
(1) (a) Schreiner, P. R. Chem. Soc. Rev. 2003, 32, 289−296.
(b) Taylor, M. S.; Jacobsen, E. N. Angew. Chem., Int. Ed. 2006, 45,
1520−1543. (c) Doyle, A. G.; Jacobsen, E. N. Chem. Rev. 2007, 107,
There are multiple pathways that may account for catalyst
activation. Our current understanding suggests an interaction
3251
DOI: 10.1021/acscatal.6b00424
ACS Catal. 2016, 6, 3248−3252

ACS Catalysis
Research Article
5713−5743. (d) Connon, S. J. Chem. Commun. 2008, 2499−2510.
(e) Zhang, Z.; Schreiner, P. R. Chem. Soc. Rev. 2009, 38, 1187−1198.
(2) (a) Yin, X.-P.; Zeng, X.-P.; Liu, Y.-L.; Liao, F.-M.; Yu, J.-S.; Zhou,
F.; Zhou, J. Angew. Chem., Int. Ed. 2014, 53, 13740−13745. (b) Zhang,
H.; Lin, S.; Jacobsen, E. N. J. Am. Chem. Soc. 2014, 136, 16485−16488.
(25) Ma, Y.; Li, Y.; Wang, E.; Lu, Y.; Wang, X.; Xu, X. J. Solid State
Chem. 2006, 179, 2367−2375.
(26) Xu, W.; Lin, J.-L.; Xie, H.-Z. Acta Crystallogr., Sect. E: Struct. Rep.
Online 2008, 64, m1031.
(27) HCl is able to activate the MOF albeit to a lesser extent (86%
over 4 h). However, MOF preactivation with HCl and DTBMP gives a
significant drop in reaction rate (36% over 4 h), which is not seen with
TMSCl/DTBMP activation.
(28) (a) Hay, B. P.; Firman, T. K.; Moyer, B. A. J. Am. Chem. Soc.
2005, 127, 1810−1819. (b) Schazmann, B.; Alhashimy, N.; Diamond,
D. J. Am. Chem. Soc. 2006, 128, 8607−8614. (c) Babu, J. N.; Bhalla, V.;
Kumar, M.; Mahajan, R. K.; Puri, R. K. Tetrahedron Lett. 2008, 49,
2772−2775. (d) Babu, J. N.; Bhalla, V.; Kumar, M.; Puri, R. K.;
Mahajan, R. K. New J. Chem. 2009, 33, 675−681.
(c) García Mancheno, O.; Asmus, S.; Zurro, M.; Fischer, T. Angew.
̃
Chem., Int. Ed. 2015, 54, 8823−8827.
(3) Schreiner, P. R.; Wittkopp, A. Org. Lett. 2002, 4, 217−220.
(4) Auvil, T. J.; Schafer, A. G.; Mattson, A. E. Eur. J. Org. Chem. 2014,
2014, 2633−2646.
(5) Malerich, J. P.; Hagihara, K.; Rawal, V. H. J. Am. Chem. Soc. 2008,
130, 14416−14417.
(6) Inokuma, T.; Furukawa, M.; Uno, T.; Suzuki, Y.; Yoshida, K.;
Yano, Y.; Matsuzaki, K.; Takemoto, Y. Chem. - Eur. J. 2011, 17,
10470−10477.
́ ́
(7) (a) Almasi̧ , D.; Alonso, D. A.; Gomez-Bengoa, E.; Najera, C. J.
Org. Chem. 2009, 74, 6163−6168. (b) Zhang, L.; Lee, M.-M.; Lee, S.-
M.; Lee, J.; Cheng, M.; Jeong, B.-S.; Park, H.-g.; Jew, S.-s. Adv. Synth.
Catal. 2009, 351, 3063−3066.
(8) Takenaka, N.; Sarangthem, R. S.; Seerla, S. K. Org. Lett. 2007, 9,
2819−2822.
(9) (a) Tran, N. T.; Min, T.; Franz, A. K. Chem. - Eur. J. 2011, 17,
9897−9900. (b) Schafer, A. G.; Wieting, J. M.; Mattson, A. E. Org.
Lett. 2011, 13, 5228−5231.
(10) Weil, T.; Kotke, M.; Kleiner, C. M.; Schreiner, P. R. Org. Lett.
2008, 10, 1513−1516.
(11) (a) Hughes, M. P.; Shang, M.; Smith, B. D. J. Org. Chem. 1996,
61, 4510−4511. (b) Hughes, M. P.; Smith, B. D. J. Org. Chem. 1997,
62, 4492−4499. (c) So, S. S.; Burkett, J. A.; Mattson, A. E. Org. Lett.
2011, 13, 716−719. (d) Nickerson, D. M.; Angeles, V. V.; Auvil, T. J.;
So, S. S.; Mattson, A. E. Chem. Commun. 2013, 49, 4289−4291.
(12) (a) Deshapande, S. V.; Meredith, C. C.; Pasternak, R. A. Acta
Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem. 1968, B24, 1396−
1397. (b) George, S.; Nangia, A.; Lynch, V. M. Acta Crystallogr., Sect.
C: Cryst. Struct. Commun. 2001, C57, 777−778. (c) Slawin, A. M. Z.;
Lawson, J.; Storey, J. M. D.; Harrison, W. T. A. Acta Crystallogr., Sect.
E: Struct. Rep. Online 2007, E63, 2925−2927.
(13) Roberts, J. M.; Fini, B. M.; Sarjeant, A. A.; Farha, O. K.; Hupp, J.
T.; Scheidt, K. A. J. Am. Chem. Soc. 2012, 134, 3334−3337.
(14) McGuirk, C. M.; Katz, M. J.; Stern, C. L.; Sarjeant, A. A.; Hupp,
J. T.; Farha, O. K.; Mirkin, C. A. J. Am. Chem. Soc. 2015, 137, 919−
925.
(15) (a) Dugan, E.; Wang, Z.; Okamura, M.; Medina, A.; Cohen, S.
M. Chem. Commun. 2008, 3366−3368. (b) Dong, X.-W.; Liu, T.; Hu,
Y.-Z.; Liu, X.-Y.; Che, C.-M. Chem. Commun. 2013, 49, 7681−7683.
(16) Luan, Y.; Zheng, N.; Qi, Y.; Tang, J.; Wang, G. Catal. Sci.
Technol. 2014, 4, 925−929.
(17) Nouar, F.; Eubank, J. F.; Bousquet, T.; Wojtas, L.; Zaworotko,
M. J.; Eddaoudi, M. J. Am. Chem. Soc. 2008, 130, 1833−1835.
(18) (a) Lee, J.; Farha, O. K.; Roberts, J.; Scheidt, K. A.; Nguyen, S.
T.; Hupp, J. T. Chem. Soc. Rev. 2009, 38, 1450−1459. (b) Oisaki, K.;
Li, Q.; Furukawa, H.; Czaja, A. U.; Yaghi, O. M. J. Am. Chem. Soc.
2010, 132, 9262−9264. (c) Fei, H.; Cohen, S. M. J. Am. Chem. Soc.
2015, 137, 2191−2194. (d) Manna, K.; Zhang, T.; Greene, F. X.; Lin,
W. J. Am. Chem. Soc. 2015, 137, 2665−2673.
(19) Tanabe, K. K.; Cohen, S. M. Inorg. Chem. 2010, 49, 6766−6774.
(20) Li, B.; Leng, K.; Zhang, Y.; Dynes, J. J.; Wang, J.; Hu, Y.; Ma, D.;
Shi, Z.; Zhu, L.; Zhang, D.; Sun, Y.; Chrzanowski, M.; Ma, S. J. Am.
Chem. Soc. 2015, 137, 4243−4248.
(21) Carlin, R. B.; Forshey, W. O. J. Am. Chem. Soc. 1950, 72, 793−
801.
(22) Subramanian, S.; Zaworotko, M. J. Angew. Chem., Int. Ed. Engl.
1995, 34, 2127−2129.
(23) The degree of bending was calculated between the terminal
nitrogens and a centroid on C−C bond between the two pyridine
rings.
(24) (a) Shukla, A. D.; Dave, P. C.; Suresh, E.; Das, A.; Dastidar, P. J.
Chem. Soc., Dalton Trans. 2000, 4459−4463. (b) Lei, C.; Mao, J.-G.;
Sun, Y.-Q.; Song, J.-L. Inorg. Chem. 2004, 43, 1964−1968.
3252
DOI: 10.1021/acscatal.6b00424
ACS Catal. 2016, 6, 3248−3252