Stable, crystalline, porous, covalent organic frameworks as a platform for chiral organocatalysts

Article abstract of DOI:10.1038/nchem.2352

The periodic layers and ordered nanochannels of covalent organic frameworks (COFs) make these materials viable open catalytic nanoreactors, but their low stability has precluded their practical implementation. Here we report the synthesis of a crystalline

Full text of DOI:10.1038/nchem.2352

ARTICLES
PUBLISHED ONLINE: 21 SEPTEMBER 2015 | DOI: 10.1038/NCHEM.2352
Stable, crystalline, porous, covalent organic
frameworks as a platform for chiral organocatalysts
Hong Xu, Jia Gao and Donglin Jiang
*
The periodic layers and ordered nanochannels of covalent organic frameworks (COFs) make these materials viable open
catalytic nanoreactors, but their low stability has precluded their practical implementation. Here we report the synthesis of
a crystalline porous COF that is stable against water, strong acids and strong bases, and we demonstrate its utility as a
material platform for structural design and functional development. We endowed a crystalline and porous imine-based
COF with stability by incorporating methoxy groups into its pore walls to reinforce interlayer interactions. We subsequently
converted the resulting achiral material into two distinct chiral organocatalysts, with the high crystallinity and porosity
retained, by appending chiral centres and catalytically active sites on its channel walls. The COFs thus prepared combine
catalytic activity, enantioselectivity and recyclability, which are attractive in heterogeneous organocatalysis, and were
shown to promote asymmetric C–C bond formation in water under ambient conditions.
ovalent organic frameworks (COFs) are extended structures synthesis of a new COF that combines outstanding stability with
with periodic molecular orderings and inherent porosity1–6
.
a high crystallinity and porosity. The COF is thermally stable up
C
They are structurally predesignable using reticular chemistry to 400 °C and retains its crystallinity and porosity even after treat-
and experimentally synthesized through reversible covalent ments in boiling water, strong acids and strong bases. The COF
bonds1–6. COFs constitute an emerging class of crystalline has a Brunauer–Emmett–Teller (BET) surface area of 2,105 m2 g–1
polymer with a wide range of structures and potential applications. and a Langmuir surface area of 3,336 m2 g–1, which are the highest
Owing to the reversible nature of their synthesis, currently these crys- among 2D COFs reported to date4,7,20,21. This COF combines stab-
talline porous polymers can be prepared with only limited stability; ility, crystallinity and porosity and provides a material platform for
the synthesis of a stable crystalline porous framework that is robust the exploration of new structures and functions. It is compatible
against harsh conditions and environments, such as humidity, with post-synthetic functionalization while retaining the crystallinity
acidity and basicity, remains a major issue that prevents their practi- and porosity of the framework. Modification of the walls converts the
cal implementation. In this study, we report the synthesis of a crystal- achiral COF into a chiral organocatalytic framework in which chiral
line porous COF that is stable in water, strong acids and strong bases, centres and organocatalytic sites are anchored onto the walls of the
and we demonstrate its utility as a polymer platform for the design of open channels. We show that the chiral COFs function as metal-
chiral structures and the development of catalytic functions.
free heterogeneous catalysts for the acceleration of asymmetric
COFs with boroxine and boronate–ester linkages exhibit high Michael C–C formation reactions in water and exhibit high activities,
crystallinities and porosities, but they are unstable in the presence enantioselectivities and recyclabilities.
of water or protic solvents because of the decomposition of their
boronate or boroxine bonds7–12. By contrast, COFs based on Results
imine, hydrazone, triazine, phenazine and azine linkages show Design of stable COFs. We proposed a strategy to soften the
improved stability, but they usually have low crystallinities and polarization influence by delocalizing the lone pairs of electron-
limited porosities13–18. Despite significant efforts in the synthesis donating groups over the positively charged phenyl rings through
of these materials, stability, crystallinity and porosity are rarely resonance effects (Fig. 1a, inset). We developed a novel structure,
found together in a single COF material.
TPB-DMTP-COF (TPB, triphenylbenzene; DMTP, dimethoxy-
The stability of a two-dimensional (2D) COF originates from two terephthaldehyde), that possesses C3-TPB vertices and electron-
factors: the bonding strength between components of the 2D layers donating (lone pairs on the oxygen) methoxy-substituted C2-phenyl
and the interlayer force. The interlayer interactions play a key role in edges (Fig. 1a). We investigated the crystal stacking energy of TPB-
directing the formation of the layered stacking structure and, therefore, DMTP-COF and a control, TPB-TP-COF (TP, terephthaldehyde),
have significant effects on the crystallinity and porosity of the resulting that lacks methoxy groups (Supplementary Fig. 1)19. TPB-DMTP-
COFs. In imine-linked COFs, the C=N bond is polarized to yield par- COF has a crystal stacking energy of 106.862 kcal mol–1, which is
tially positively charged carbon and negatively charged nitrogen much higher than that of TPB-TP-COF (94.084 kcal mol–1)
(Fig. 1a, inset). In a hexagonal 2D COF (Fig. 1), each macrocycle con- (Supplementary Table 1). Therefore, introducing the two electron-
sists of 12 polarized C=N segments; the aggregation of a large number donating methoxy groups to each phenyl edge delocalizes the two
of charged groups causes electrostatic repulsion and destabilizes the lone pairs from the oxygen atoms over the central phenyl ring,
layered structure, as predicted theoretically19. The introduction of which reinforces the interlayer interactions and so stabilizes the
new principles for the design of stable COFs is highly desired.
Here we address this issue by incorporating methoxy groups into
COF and aids in its crystallization22.
The state-of-the-art approaches to stabilize COFs are based on
the pore walls to reinforce the interlayer interactions for the the introduction of enol–keto tautomerizations or hydrogen-
Department of Materials Molecular Science, Institute for Molecular Science, National Institutes of Natural Sciences, 5–1 Higashiyama, Myodaiji,
Okazaki 444-8787, Japan. e-mail: jiang@ims.ac.jp
*
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
1
© 2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2352
ARTICLES
a
c
H2N
H2N
O
O
O
OMe
OMe
O
NH2
NH2
O
O
MeO
O
MeO
O
H2N
H2N
δ
MeO
DMTA
TAPB
N
1–x DMTA
x BPTA
TAPB
δ
H
C
C
H
δ
N
OMe
δ
Resonance effect of oxygen
lone pairs softens the
interlayer charge repulsion
MeO
O
N
N
N
N
OMe
O
OMe
OMe
N
N
N
N
N
N
MeO
MeO
MeO
OMe
MeO
OMe
N
N
N
3.3 nm
N
N
MeO
OMe
O
O
OMe
N
O
N
N
N
N
MeO
O
MeO
N
MeO
N
N
N
TPB-DMTP-COF
OMe
OMe
[HC C]
x -TPB-DMTP-COFs
x = 0.17, 0.34, 0.50
H
N
N3
NH
b
N
N
N
O
N
N
O
OMe
N
N
N
N
N
N
MeO
MeO
OMe
HN
N
H
N
NH
N
N
N
N
N
N
N
O
N
O
N
N
N
O
N
N
N
N
H
O
MeO
N
N
N
N
HN
OMe
[(S)-Py]x -TPB-DMTP-COFs
x = 0.17, 0.34, 0.50
Figure 1 | Synthesis and structure of stable crystalline porous COFs. a, Synthesis of TPB-DMTP-COF through the condensation of DMTA (blue) and TAPB
(black). Inset: The structure of the edge units of TPB-DMTP-COF and the resonance effect of the oxygen lone pairs that weaken the polarization of the
C=N bonds and soften the interlayer repulsion in the COF. b, Graphic view of TPB-DMTP-COF (red, O; blue, N; grey, C; hydrogen is omitted for clarity).
c, Synthesis of chiral COFs ([(S)-Py]x-TPB-DMTP-COFs, x = 0.17, 0.34 and 0.50; blue, DMTA; black, TAPB; red, BPTA; green, (S)-Py sites) via channel-wall
engineering using a three-component condensation followed by a click reaction.
bonding interactions to the COF skeletons15,23–25. The keto-type show a weight loss of 70 wt% along with the loss of crystallinity
COFs under alkaline conditions (6 M NaOH, three days) lose crys- and porosity24. These examples demonstrate the best stability per-
tallinity because of the backward keto-to-enol conversion25. The formance among the COFs reported to date and indicate the chal-
hydrogen-bonding COFs under acidic conditions (3 M HCl, one lenges faced in the synthesis of a COF material that is robust
week) exhibit a decrease in surface area from 1,305 to 570 m2 g–1. against acids and bases and simultaneously combines a high crystal-
Under alkaline conditions (3 M NaOH, one week), these COFs linity and porosity.
2
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2352
ARTICLES
g
a
b
d
f
1,000
800
600
400
200
0
100
001
DMF
DMSO
Pristine
DMF
24 25 26 27
THF
DMSO
THF
MeOH
Cyclohexanone
HCl (12 M)
Water (100 °C)
0.0 0.2 0.4 0.6 0.8 1.0
MeOH
Relative pressure (P/P0)
c
Cyclohexanone
AA stacking
AB stacking
e
3.0
2.5
2.0
1.5
1.0
0.5
0.0
1.2
1.0
0.8
0.6
0.4
0.2
0.0
HCl (12 M)
Water (100 °C)
Water (25 °C)
NaOH (14 M)
Water (25 °C)
NaOH (14 M)
5
10 15 20 25 30
2θ (°)
0
10 20 30
0
20 40 60 80 100
Residual weight (wt%)
5
10 15 20 25 30
2θ (°)
Pore width (nm)
Figure 2 | Crystallinity, porosity and stability. a, PXRD profiles of TPB-DMTP-COF. Experimentally observed (red), Pawley refined (green) and their
difference (black), simulated using the AA stacking mode (blue) and the staggered AB stacking mode (orange). b, Unit cell of the AA stacking mode (O, red;
N, blue; C, grey; H, white). c, Unit cell of the AB stacking mode (O, red; N, blue; C, grey; H, white; green, a further layer). d, Nitrogen-sorption isotherm
curves measured at 77 K. e, Profiles of the pore size and pore-size distribution. f, Residue weight percentage of TPB-DMTP-COF after treatment for one
week in different solvents. g, PXRD patterns of TPB-DMTP-COF after treatment for one week in different solvents.
Stability of COFs. Thermogravimetric analysis (TGA) revealed that
We further synthesized TPB-DHTP-COF (DHTP, dihydroxy-
TPB-DMTP-COF exhibited no weight loss under N2 to 400 °C, terephthaldehyde) with phenol groups on the edge units that
whereas under O2 it burned out completely above 340 °C allow for the hydrogen-bonding O–H···N interactions between
(Supplementary Fig. 2a,b). To investigate the chemical stability of the OH units and imine groups (Supplementary Fig. 1)24.
TPB-DMTP-COF, we dispersed TPB-DMTP-COF samples for TPB-DHTP-COF exhibited a low crystallinity (intensity = 11 × 103
one week in various solvents, including dimethylformamide counts per second (c.p.s.), FWHM = 0.672° (Supplementary Fig. 6))
(DMF), dimethylsulfoxide (DMSO), THF, MeOH, cyclohexanone, and porosity (BET surface area = 849 m2 g–1 (Supplementary
water (100 and 25 °C), aqueous HCl (12 M) and aqueous NaOH Fig. 7)). On treatment for one week in HCl (12 M), NaOH (14 M)
(14 M) solutions. Figure 2f presents the residue weight percentage or boiling water, TPB-DHTB-COF deteriorated in both crystalli-
of these COF samples. In water and the organic solvents examined, nity and porosity (Supplementary Figs
6 and 7 and
TPB-DMTP-COF exhibited almost no weight loss (<0.1 wt%). Supplementary Table 3). Compared with TPB-DMTP-COF,
Under the ultraharsh conditions of a strong acid (12 M HCl TPB-DHTP-COF exhibited a worse π-stacking between the layers
(concentrated HCl)) and a strong base (14 M NaOH), the residual with a decreased crystal-stacking energy (103 kcal mol–1 per unit
weight percentages were 85 and 92 wt%, respectively. Even on cell (Supplementary Table 1)), which led to low crystallinity
hydrolysis in boiling water for one week, the TPB-DMTP-COF and stability.
retained 72 wt% of its original mass.
TPB-DMTP-COF retained its original crystalline structure, as Crystallinity and porosity. The PXRD pattern of TPB-DMTP-COF
indicated by the unchanged intensities and positions of the peaks exhibited six prominent diffraction peaks, with the most-intensive
in its powder X-ray diffraction (PXRD) profile on dispersion for one at 2.76° (FWHM = 0.39°) and the five other peaks at 4.82,
one week in DMF, DMSO, THF, MeOH, cyclohexanone, water 5.60, 7.42, 9.70 and 25.2°; these peaks were assigned to the (100),
(100 and 25 °C), aqueous HCl (12 M) and aqueous NaOH (14 M) (110), (200), (210), (220) and (001) facets, respectively (Fig. 2a,
solutions (Fig. 2g). We evaluated the full-width at half-maximum red curve). The (100) signal is sharp compared with that of
(FWHM) values, which also suggested that these treated TPB-TP-COF under the same PXRD experimental conditions
TPB-DMTP-COF samples retained their crystallinity (Supplementary (qualitatively, 135 × 103 c.p.s. versus 1.7 × 103 c.p.s. in intensity
Table 2). The BET surface areas were 2,081, 2,074 and 2,020 m2 g–1 (Supplementary Fig. 8)). Another feature is that the PXRD pattern
for the COF samples treated for one week in boiling water, strong contains more peaks than those for other imine-linked COFs. A
acid and strong base, respectively; these values are very close to density-functional tight-binding method, which included
a
that of the as-synthesized COF (2,105 m2 g–1 (Supplementary Lennard–Jones dispersion, was used to simulate the optimum
Fig. 3 and Supplementary Table 3)). Moreover, the infrared structures of TPB-DMTP-COF. Using the optimal monolayer
spectra confirmed that the chemical bonds were well preserved structure, AA and staggered AB stacking modes were generated and
(Supplementary Fig. 4). To the best of our knowledge, this COF is optimized. In the stacked frameworks, TPB-DMTP-COF adopts the
the most stable among COFs reported to date (Supplementary AA stacking mode of a space group of P6 with a = b = 37.2718 Å,
Table 4)15,16,23–27. In contrast to TPB-DMTP-COF, TPB-TP-COF interlayer distance (c) of 3.5215 Å, α = β = 90° and γ = 120° (Fig. 2b
without methoxy groups on the walls exhibited a low chemical stab- and Supplementary Table 5). The simulated PXRD pattern of the
ility, which is indicated by the more-significant weight loss after AA stacking mode (Fig. 2a, blue curve) matched the experimental
treatment in acid, base or boiling water (Supplementary Fig. 5).
peak positions and intensities, whereas the staggered AB stacking
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
3
© 2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2352
ARTICLES
mode (Fig. 2a, orange curve) did not reproduce the data (red curve).
The staggered AB mode gave rise to a space group of P-3 with a = b =
36.6669 Å, c = 6.3984 Å, α = β = 90° and γ = 120° (Supplementary
Table 6), which, however, covers the pores of the neighbouring
sheets (Fig. 2c).
Chiral centre
Large space
between
catalytic sites
Pawley refinement (Fig. 2a, green curve) provided a good match
to the observed PXRD pattern, as indicated by the negligible differ-
ence between the simulated and experimental patterns (Fig. 2a,
black curve). The Pawley refinement led to a space group of P6
with a = b = 37.1541 Å, c = 3.5378 Å, α = β = 90° and γ = 120°, and
RP and RWP values of 2.02 and 4.37%, respectively (Supplementary
Table 7). We further used the Rietveld refinement method reported
for the imine-linked COFs28 to refine our crystal structure of
TPB-DMTP-COF, which resulted in a space group of P6 with
a = b = 36.4594 Å, c = 3.5239 Å, α = β = 90° and γ = 120°, and RP
and RWP values of 3.31 and 6.52%, respectively (Supplementary
Table 8). The Rietveld refinement also reproduced the PXRD
pattern well (Supplementary Fig. 9). These refinements indicated
that the peak assignment of the PXRD pattern was correct.
Nitrogen-sorption isotherms of TPB-DMTP-COF measured at
77 K exhibited a rapid uptake at a low pressure of P/P0 < 0.1, fol-
lowed by a sharp step between P/P0 = 0.15 and 0.25 (Fig. 2d, black
curve). This sorption profile is best described as a type IV isotherm,
which is characteristic of mesoporous materials. The BET and
Langmuir surface areas were 2,105 and 3,336 m2 g–1, respectively.
These surface areas are among the highest BET surface areas of 2D
COFs4,7,20,21. By contrast, TPB-TP-COF exhibited a BET surface
area of only 16 m2 g–1 (Supplementary Fig. 10 and Supplementary
Table 3). The pore size and volume of TPB-DMTP-COF were evalu-
ated to be 3.26 nm and 1.28 cm3 g–1, respectively. The theoretical
surface area calculated using the AA stacking mode was 2,098 m2 g–1;
the experimentally observed surface area is nearly identical to the
theoretical one.
Organocatalytic
site
[(S)-Py]0.17-TPB-DMTP-COF
Limited space
between
catalytic sites
[(S)-Py]0.34-TPB-DMTP-COF
Small space
between
catalytic sites
[(S)-Py]0.50-TPB-DMTP-COF
As shown above, TPB-DMTP-COF exhibits an outstanding stab-
ility and possesses high crystallinity and porosity. The introduction
of electron-donating groups to the positively charged edges
reinforces the interlayer interactions that exert a positive and signifi-
cant effect on the stability, crystallinity and porosity. Such a robust
open framework constitutes a useful material platform for structural
design and functional exploration. High-resolution transmission
electron microscopy of TPB-DMTP-COF enables the direct visual-
ization of the hexagonal polygon textures and the stacking layer
structures (Supplementary Fig. 11).
Figure 3 | Channel-wall structure of the chiral organocatalytic COFs. The
open structures of the chiral centres and the catalytic sites on the channel
walls are shown for different COFs (O, red; N, blue; C, grey). The space
between the catalytic sites is required for the face-on stacking of
nitrostyrene over the catalytic sites and plays a role in controlling the activity
of the catalysts.
(x value) in [HC≡C]x-TPB-DMTP-COFs, we plotted two
standard curves (Supplementary Fig. 12) based on the
characteristic stretching bands of a C≡C bond at 2,120 cm–1 and
Chiral COFs engineered with organocatalytic sites on the walls. C≡C–H bond at 3,300 cm–1 using [HC≡C]1.0-TPB-DMTP-COF
Intrigued by its high crystallinity and large mesoporous channels, (equivalent to TPB-BPTP-COF (BPTP, bis(2-propynyloxy)
we developed TPB-DMTP-COF into a catalytic chiral open terephthaldehyde; for the structure see Supplementary Fig. 1)) as
framework by anchoring chiral centres and organocatalytic sites the standard. From the standard curves, the x values were evaluated
onto the channel walls via post-synthetic functionalization to be 0.17, 0.342 and 0.496, based on the bands at 2,120 cm–1, and
(Fig. 1c). The robust open-framework structure provides a 0.173, 0.334 and 0.512, based on the bands at 3,300 cm–1 for
benchmark sample to demonstrate the potential of COFs in [HC≡C]0.17-TPB-DMTP-COF, [HC≡C]0.34-TPB-DMTP-COF and
asymmetric catalysis. We employed
a
three-component [HC≡C]0.50-TPB-DMTP-COF, respectively (Supplementary Table 9).
condensation system with 2,5-bis(2-propynyloxy)terephthalaldehyde On the quantitative click reactions, the characteristic bands of the
(BPTA) and 2,5-dimethoxyterephthalaldehyde (DMTA) as edge C≡C–H units at 2,120 and 3,300 cm–1 disappeared. Moreover, the
units to synthesize the intermediate [HC≡C]x-TPB-DMTP-COFs clicked (S)-Py-triazole units consisted of four nitrogen atoms, which
(Fig. 1c), where x is the percentage of functional groups (either greatly enhanced the nitrogen content of the resulting [(S)-Py]x-
[HC≡C] or, subsequently, (S)-Py moieties) as a fraction of the TPB-DMTP-COFs. The x values based on the nitrogen-content
groups lining the pore walls (that is, including the non- increment were evaluated to be 0.18, 0.33 and 0.51 for [(S)-Py]0.17
functionalized OMe groups of DMTA and the functionalized TPB-DMTP-COF, [(S)-Py]0.34-TPB-DMTP-COF and [(S)-Py]0.50
-
-
[HC≡C] or (S)-Py moieties), so that x = [BPTA]/([BPTA]+[DMTA]); TPB-DMTP-COF, respectively (Supplementary Table 10). In
x = 0, 0.17, 0.34 and 0.50. After a quantitative azide–ethynyl these COFs, each hexagonal macrocycle contains one, two or three
click reaction20,29, the [HC≡C]x-TPB-DMTP-COFs were converted (S)-Py units on average, respectively (Fig. 1c). Figure 3 shows
into chiral organocatalytic [(S)-Py]x-TPB-DMTP-COFs ((S)-Py, the structures of the channel walls with chiral organocatalytic
(S)-pyrrolidine) (Fig. 1c) in which chiral and catalytic (S)-Py sites. The organocatalytic sites are integrated into the walls
sites were anchored onto the channel walls through triazole and form catalytic nanospaces of different sizes (Fig. 3). The
rings in a random manner. To estimate the real [HC≡C] contents [HC≡C]x-TPB-DMTP-COFs and [(S)-Py]x-TPB-DMTP-COFs
4
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2352
ARTICLES
a
b
800
600
400
200
0
≡
[HC C]0.17-TPB-DMTP-COF
[(S)-Py]0.17-TPB-DMTP-COF
5
10
15
20
25
30
0.0
0.2
0.4
0.6
0.8
1.0
2θ (°)
Relative pressure (P/P0)
c
d
e
1.2
1.0
0.8
0.6
0.4
0.2
0.0
2.5
2.0
1.5
1.0
0.5
0.0
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.8
0.6
0.4
0.2
0.0
2.5
2.0
1.5
1.0
0.5
0.0
1.0
0.8
0.6
0.4
0.2
0.0
0
10
20
30
0
10
20
30
0
10
20
30
Pore width (nm)
Pore width (nm)
Pore width (nm)
Figure 4 | Crystallinity and porosity of chiral COFs. a, PXRD patterns of [(S)-Py]0.17-TPB-DMTP-COF (red curve) and of [HC≡C]0.17-TPB-DMTP-COF (black
curve). The chiral [(S)-Py]0.17-TPB-DMTP-COF retains a high crystallinity. b, Nitrogen-sorption isotherm curves of [(S)-Py]0.17-TPB-DMTP-COFs (red curve,
x = 0.17; blue curve, x = 0.34; purple curve, x = 0.50) measured at 77 K. c–e, Profiles of the pore size (spheres) and pore-size distribution (solid lines) of
[(S)-Py]0.17-TPB-DMTP-COF (c), [(S)-Py]0.34-TPB-DMTP-COF (d) and [(S)-Py]0.50-TPB-DMTP-COF (e). The chiral COFs exhibit one type of mesopore.
were unambiguously characterized using various analytical methods [(S)-Py]0.17-TPB-DMTP-COF retained its crystalline structure, as
(Supplementary Figs 13–17 and Supplementary Tables 3 and 10). evidenced by the unchanged intensities, FWHM values and positions
To investigate the possibility of Cu residues in [(S)-Py]x-TPB- of the peaks in the PXRD profiles (Supplementary Fig. 19 and
DMTP-COFs, we conducted inductively coupled plasma atomic Supplementary Table 2). The BET surface areas were 1,924, 1,956
emission spectroscopy (ICP-AES) and observed that the content of and 1,971 m2 g–1 on treatment for one week in boiling water, HCl
Cu was as low as 0.004–0.009 wt% (Supplementary Table 11). (12 M) and NaOH (14 M), respectively; these values are close to that
TGA measurements under O2 also revealed that these [(S)-Py]x- of the as-synthesized [(S)-Py]0.17-TPB-DMTP-COF (1,970 m2 g–1
TPB-DMTP-COFs burned out without residues (Supplementary (Supplementary Fig. 20 and Supplementary Table 3)).
Fig. 2).
The PXRD profiles of the [(S)-Py]x-TPB-DMTP-COFs (Fig. 4a, Heterogeneous asymmetric organocatalysis. Organocatalysis in
red curve) showed diffraction patterns that were identical to that homogeneous systems is problematic because of the difficulty of
of [HC≡C]0.17-TPB-DMTP-COF (black curve), which indicates that separating expensive catalysts for repeated use30–34
.
The
the lattice structure was not affected by the chiral catalytic sites on development of immobilized, easily recoverable and reusable
the walls of the pores. [(S)-Py]x-TPB-DMTP-COFs are highly catalysts appears to be one of the most-promising strategies for
porous materials, as revealed by nitrogen-sorption isotherm measure- overcoming these problems. Most heterogeneous organocatalysts
ments (Fig. 4b). The BET surface areas of [(S)-Py]0.17-TPB-DMTP- are based on linear polymer supports; however, such catalysts
COF, [(S)-Py]0.34-TPB-DMTP-COF and [(S)-Py]0.50-TPB-DMTP- exhibit low activities as a result of inefficient access to the catalytic
COF were 1,970, 1,802 and 1,612 m2 g–1, whereas their Langmuir sites32–34
.
To overcome this issue, crystalline metal–organic
surface areas were 3,059, 2,812 and 2,300 m2 g–1, respectively framework (MOF)-based organocatalysts have been developed35–37
.
(Supplementary Table 3). Their pore sizes were 3.07, 2.95 and However, their stabilities, small pore sizes, enantioselectivities and
2.86 nm, and their pore volumes were 1.13, 1.04 and 0.83 cm3 g–1, diastereoselectivities need to be improved.
respectively (Supplementary Fig. 18). Profiles of the pore-size distri-
A Michael addition is a typical organocatalytic reaction, is one of
bution revealed the presence of only one type of mesopore (Fig. 4c–e). the basic C–C bond formation reactions and provides a powerful
To evaluate the stability of the COFs with catalytic sites on the synthetic tool for the formation of synthons of many important
walls, we treated [(S)-Py]0.17-TPB-DMTP-COF in different solvents, natural and biologically active products31,38,39. Typically, Michael
including DMF, DMSO, THF, MeOH, cyclohexanone, water (100 and reactions are conducted in organic solvents or in mixed
25 °C), aqueous HCl (12 M) and aqueous NaOH (14 M) solutions, organic–aqueous solutions; the use of neat water as a solvent is
under otherwise identical conditions to those for TPB-DMTP-COF. particularly attractive from the environmental and economic
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
5
© 2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2352
ARTICLES
Table 1 | Asymmetric Michael reaction catalysed with chiral COFs.
Entry
Catalyst
Time (h)
Conversion (%)
Yield (%)
e.e. (%)
d.r.
1
[(S)-Py]0.17-TPB-DMTP-COF
[(S)-Py]0.34-TPB-DMTP-COF
[(S)-Py]0.50-TPB-DMTP-COF
Free (S)-Py-containing moiety, ((S)-1-(pyrrolidin-2-yl)methyl)-
4-phenoxymethyl-triazole
12
17
34
22
100
100
100
100
95
93
95
96
92
91
89
92
90/10
90/10
88/12
91/9
2
3
4
5
[(S)-Py]0.25-H2P-COF*
36
–
–
–
–
Reaction scheme: see Figure 5, top, with R = H; *No reaction.
points of view. The high catalytic activity of [(S)-Py]x-TPB-DMTP-
Discussion
COFs allows Michael reactions to be conducted in neat water at We investigated different reactants to illustrate the generality of the
25 °C and 1 bar. The insolubility of the [(S)-Py]x-TPB-DMTP- catalyst (Fig. 5). We used β-nitrostyrene compounds with different
COFs gives rise to heterogeneous systems that have outstanding substituents on their phenyl ring for the Michael reaction. The
catalytic activities.
ortho-chlorosubstituted β-nitrostyrene exhibited the highest activity
The Michael reaction of cyclohexanone and β-nitrostyrene pro- among the series; the conversion reached 100% in six hours, with
ceeded cleanly and smoothly and achieved 100% conversion in e.e. and d.r. values of 94% and 97/3, respectively. High activities
12 hours with enantioselectivity (e.e.) and diastereoselectivity (d.r.) were achieved for activated β-nitrostyrenes, for example, 4-bromo-
values of 92% and 90/10, respectively (Table 1, entry 1). The mol- β-nitrostyrene and 4-chloro-β-nitrostyrene. The reactions were
ecular catalyst (S)-Py with the same catalytic structure (Table 1, completed in 12 and 10 hours, respectively. The reaction of deacti-
entry 4) was used as a control and required a much longer reaction vated 4-methyl-β-nitrostyrene was completed in 16 hours and
time of 22 hours. The reduced reaction time suggests that reached e.e. and d.r. values of 93% and 92/8, respectively. The
[(S)-Py]0.17-TPB-DMTP-COF exhibits an enhanced catalytic best selectivity was achieved for 4-methoxyl-β-nitrostyrene with
activity; the open channels can accumulate the reactants from the 96% e.e. and 94/6 d.r. in 26 hours.
water phase and promote the reactions in the confined nanopore
Kinetic studies were conducted throughout the entire reaction
space. The similarities in the enantioselectivity and diastereoselec- region between 0 and 100% conversion, with each experiment per-
tivity between [(S)-Py]0.17-TPB-DMTP-COF and (S)-Py indicate formed at least in duplicate. The average conversions were used for
that the chiral catalytic sites on the channel walls retained both the kinetic evaluations. The decrease in the concentration of
enantiocontrol and diastereocontrol. To investigate the effect of trans-2-chloro-β-nitrostyrene and increase in the concentration of
Cu species on the reaction, we conducted the following control the product were straightforward, and the catalytic reactions pro-
experiments (Supplementary Table 12). (1) The Cu powder, CuI ceeded smoothly (Supplementary Fig. 21). The reactions exhibited
and CuSO4·5H2O were utilized separately as catalysts for the apparent first-order behaviour with respect to the concentration
above reaction under otherwise identical conditions; these Cu of trans-2-chloro-β-nitrostyrene (Supplementary Fig. 22). The cal-
species did not yield any products. (2) The Cu powder, CuI and culated rate constant and lifetime (t1/2) were 0.78 per hour and
CuSO4·5H2O in the presence of TPB-DMTP-COF were utilized sep-
arately for the above reaction; these systems again did not yield any
R
products. (3) The Cu powder, CuI and CuSO4·5H2O in the presence
of [(S)-Py]0.17-TPB-DMTP-COF were utilized separately for the
O
O
reaction; [(S)-Py]0.17-TPB-DMTP-COF retained its original cataly-
tic activity, irrespective of the addition of the Cu species. These
results revealed that the Cu species did not affect the catalytic reac-
tions. Supplementary Table 13 summarizes the heterogeneous orga-
nocatalysis reported to date for the Michael and aldol reactions. The
majority of the heterogeneous catalysts are active for only aldol reac-
tions that utilize small and highly active reactants. The large pore
sizes and high activities render the COFs able to catalyse Michael
reactions; to the best of our knowledge, this COF is the first
example of a heterogeneous catalyst based on a crystalline porous
material, including mesoporous silica, MOFs and COFs, for the acti-
vation of low-activity ketones in Michael reactions.
We observed that the densities of the chiral catalytic sites signifi-
cantly affected the reaction rate. [(S)-Py]0.34-TPB-DMTP-COF and
[(S)-Py]0.50-TPB-DMTP-COF required long reaction times of 17
and 34 hours, respectively (Table 1, entries 2 and 3). The prolonged
reaction time suggests that the high density of catalytic sites on the
channel walls causes these sites to overlap with each other and
results in an inefficient utilization of them; the catalysis requires a
face-on attack by nitrostyrenes at the catalytic sites (Fig. 3). The
pore size is also important for the Michael reaction; as the pore
size decreased to the microporous range, the Michael reactions
were suppressed severely. For example, the microporous
[(S)-Py]0.25-H2P-COF with catalytic species identical to those of
[(S)-Py]0.17-TPB-DMTP-COF exhibited a low activity for the
Michael reaction with nearly negligible products after 36 hours
(Table 1, entry 5).
NO2
NO2
Catalyst (10 mol%)
H2O, 25 °C
R
Cl
O
O
O
Cl
NO2
NO2
NO2
12 h, d.r. = 90/10,
e.e. = 92%
10 h, d.r. = 90/10,
e.e. = 90%
6 h, d.r. = 97/3,
e.e. = 94%
Br
OMe
O
O
O
NO2
NO2
NO2
12 h, d.r. = 93/7,
e.e. = 95%
16 h, d.r. = 92/8,
e.e. = 93%
26 h, d.r. = 94/6,
e.e. = 96%
Figure 5 | Scope of reactants. Different β-nitrostyrene derivatives
investigated for the Michael reactions catalysed with chiral COFs, their
products, e.e. yields and d.r. values (red, cyclohexanone; green, newly formed
C–C bond; blue, nitrostyrene derivatives). R, substituent H, Cl, Br, Me or OMe.
6
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2352
ARTICLES
Table 2 | Cycling test of [(S)-Py]0.17-TPB-DMTP-COF catalyst for the Michael reaction.
Cycle number
Time (h)
Conversion (%)
Yield (%)
e.e. (%)
d.r.
RY (wt%)*
1
6
7
9
11
13
100
100
100
100
100
95
93
94
92
92
94
94
94
94
93
97/3
97/3
97/3
97/3
97/3
>99
>99
>99
>99
>99
2
3
4
5
Reaction scheme: see Figure 5, top, with R = Cl in ortho position; *Recovery yield (RY) of the catalyst.
[HC≡C]x-TPB-DMTP-COFs. o-DCB/n-BuOH (0.5/0.5 ml) mixtures of TAPB
(0. 080 mmol, 28.1 mg) and DMTA/BPTA (a total of 0.120 mmol) at different molar
ratios of 5/1, 4/2, 3/3 and 0/1 in the presence of an acetic-acid catalyst (6 M, 0.1 ml)
in a Pyrex tube (10 ml) were degassed via three freeze–pump–thaw cycles. The tubes
were sealed off by flame and heated at 120 °C for three days. The precipitates were
collected via centrifugation, washed six times with THF and then subjected to Soxhlet
extraction with THF as the solvent for one day to remove the trapped guest molecules.
The powders were collected and dried at 120 °C under vacuum overnight to produce
the corresponding [HC≡C]0.17-TPB-DMTP-COF, [HC≡C]0.34-TPB-DMTP-COF,
[HC≡C]0.50-TPB-DMTP-COF and [HC≡C]1.0-TPB-DMTP-COF (equivalent to
TPB-BPTP-COF) in the isolated yields of 80, 79, 81 and 81%, respectively.
0.9 hours, respectively. A plausible explanation for the observed
first-order rate may involve the quick addition of cyclohexanone
to the pyrrolidine catalyst to form an enamine intermediate, fol-
lowed by the face-on-face addition of the styrene substrate to the
enamine intermediate as the rate-determining step, with a sub-
sequent rapid hydrolysis to yield the corresponding products. The
observed (2S,1R) absolute configuration of the major syn Michael
adduct, determined from a literature comparison of the HPLC
elution order, is consistent with the transient state of the nitrostyr-
ene to attack the re-face of the enamine40.
[(S)-Py]x-TPB-DMTP-COFs. (S)-2-(azidomethyl)pyrrolidine (toluene solution,
1 M, 60 μl) was added to a THF/water (2.1/0.7 ml) mixture of [HC≡C]0.17-TPB-
DMTP-COF (65 mg) in the presence of CuI (6 mg) and N,N-diisopropylethylamine
(THF solution, 1 M, 108 μl) in a flask (25 ml). The flask was degassed via three
freeze–pump–thaw cycles and the mixture was stirred at room temperature for four
hours. The precipitate was collected via centrifugation, washed five times with THF
and acetonitrile, and dried at room temperature under vacuum to produce
[(S)-Py]0.17-TPB-DMTP-COF as a dark-green solid in quantitative yield. The
ethynyl groups quantitatively reacted with the azide units, as shown by infrared
spectroscopy. A click reaction of [HC≡C]x-TPB-DMTP-COFs (x = 34 and 50) with
(S)-2-(azidomethyl)pyrrolidine was conducted to produce [(S)-Py]x-TPB-DMTP-
COFs (x = 34 and 50) in quantitative yield according to this method under otherwise
identical conditions.
A long catalyst lifetime and the capability of repeated use are
important for useful applications. [(S)-Py]x-TPB-DMTP-COFs
were separated easily from the reaction mixture and recovered; fil-
tration and rinsing with solvents refreshed the catalyst for the
next round of use. [(S)-Py]0.17-TPB-DMTP-COF at 10 mol%
loading in the first round was subsequently subjected to repeated
use in water at 25 °C. [(S)-Py]0.17-TPB-DMTP-COF retained its
activity, enantioselectivity and diastereoselectivity after five cycles
(Table 2). The extended reaction time required with increased
cycle number is probably related to the side reactions of the pyrro-
lidine catalytic sites, as reported in the literature41. Various analyti-
cal methods revealed that [(S)-Py]0.17-TPB-DMTP-COF retained its
crystallinity and porosity (Supplementary Figs 23–25 and
Supplementary Table 3). Typically, organocatalysts based on
MOFs35–37 and porous organic polymers42 have exhibited drastically
decreased activities that demand a much longer reaction time per
cycle (Supplementary Table 13). Combining all of the above cataly-
tic features, the chiral COFs function as high-performance hetero-
geneous catalysts for asymmetric transformations.
Stability test. The COF samples (10 mg) were kept for seven days in 1 ml of DMF,
DMSO, THF, cyclohexanone, MeOH, HCl (12 M, equivalent to concentrated HCl),
NaOH (14 M), water (25 °C) or boiling water (100 °C). The samples were washed
with THF (for samples treated in organic solvents) or water (for samples treated in
aqueous solutions), dried under vacuum at 120 °C for 12 hours and subjected to
PXRD, infrared spectroscopy and nitrogen-sorption isotherm measurements.
Asymmetric Michael reactions. Cyclohexanone (2 mmol), nitrostyrene
(0.1 mmol), benzoic acid (0.025 mmol) and N,N-diisopropylethylamine (water
suspension, 0.5 M, 32 μl) were added to a water (0.4 ml) suspension of the
[(S)-Py]x-TPB-DMTP-COFs catalyst (0.01 mmol). The mixture was stirred at room
temperature until 100% conversion was reached. After the addition of EtOH (5 ml),
the organic layer was removed via centrifugation. The catalyst was washed with
EtOH (5 ml twice) and ethyl acetate (5 ml three times), and the organic layers were
combined and concentrated under reduced pressure. 1H NMR spectroscopy was
used to determine the d.r. The e.e. was determined using HPLC on a chiral-phase
Chiralpak AD-H column.
Conclusion
In summary, we have developed a mesoporous COF that combines
stability, crystallinity and porosity by reinforcing the interlayer
interactions between the 2D layers. The COF’s skeleton is compati-
ble with post-synthetic functionalization with its crystallinity and
porosity retained; engineering the channel walls with chiral
centres and organocatalytic species resulted in the development of
two distinct chiral COFs in a simple yet controlled manner. The
resulting crystalline metal-free catalysts display activity, enantio-
selectivity, recyclability and environmental benignity—a set of
characteristics that has remained challenging to engineer together
in heterogeneous organocatalysis. Our strategy may thus facilitate
the design of COFs in which the combination of stability, crystalli-
nity and porosity would be of use for a variety of functions
and applications.
Cycle use of the catalyst. [(S)-Py]0.17-TPB-DMTP-COF was recovered via
centrifugation, washed with ethyl acetate and a mixture of triethylamine/ethanol
(5% by volume) to remove any products and dried under vacuum for use in the next
reaction cycle.
Received 13 June 2015; accepted 13 August 2015;
published online 21 September 2015
References
1. Feng, X., Ding, X. & Jiang, D. Covalent organic frameworks. Chem. Soc. Rev.
41, 6010–6022 (2012).
2. Ding, S. Y. & Wang, W. Covalent organic frameworks (COFs): from design to
applications. Chem. Soc. Rev. 42, 548–568 (2013).
Methods
3. Cote, A. P. et al. Porous, crystalline, covalent organic frameworks. Science 310,
1166–1170 (2005).
4. Tilford, R. W., Gemmill, W. R., zur Loye, H. C. & Lavigne, J. J. Facile synthesis of
a highly crystalline, covalently linked porous boronate network. Chem. Mater.
18, 5296–5301 (2006).
5. Belowich, M. E. & Stoddart, J. F. Dynamic imine chemistry. Chem. Soc. Rev.
41, 2003–2024 (2012).
6. Xiang, Z. H., Cao, D. P. & Dai, L. M. Well-defined two dimensional covalent
organic polymers: rational design, controlled syntheses, and potential
applications. Polym. Chem. 6, 1896–1911 (2015).
TPB-DMTP-COF. An o-dichlorobenzene (o-DCB)/n-BuOH (0.5/0.5 ml) mixture
of 1,3,5-tri-(4-aminophenyl)benzene (TAPB) (0.080 mmol, 28.1 mg) and DMTA
(0.120 mmol, 23.3 mg) in the presence of an acetic-acid catalyst (6 M, 0.1 ml) in a
Pyrex tube (10 ml) was degassed via three freeze–pump–thaw cycles. The tube was
flame sealed and heated at 120 °C for three days. The precipitate was collected via
centrifugation, washed six times with THF and then subjected to Soxhlet extraction
with THF as the solvent for one day to remove the trapped guest molecules. The
powder was collected and dried at 120 °C under vacuum overnight to produce
TPB-DMTP-COF in an isolated yield of 81%.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
7
© 2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2352
ARTICLES
7. Cote, A. P., El-Kaderi, H. M., Furukawa, H., Hunt, J. R. & Yaghi, O. M. Reticular
synthesis of microporous and mesoporous 2D covalent organic frameworks.
J. Am. Chem. Soc. 129, 12914–12915 (2007).
27. Ding, S. Y. et al. Construction of covalent organic framework for catalysis:
Pd/COF-LZU1 in Suzuki–Miyaura coupling reaction. J. Am. Chem. Soc. 133,
19816–19822 (2011).
28. Rabbani, M. G. et al. A 2D mesoporous imine-linked covalent organic framework
for high pressure gas storage applications. Chem. Eur. J. 19, 3324–3328 (2013).
29. Xu, H. et al. Catalytic covalent organic frameworks via pore surface engineering.
Chem. Commun. 50, 1292–1294 (2014).
30. MacMillan, D. W. The advent and development of organocatalysis. Nature 455,
304–308 (2008).
8. Wan, S., Guo, J., Kim, J., Ihee, H. & Jiang, D. A belt-shaped, blue luminescent,
and semiconducting covalent organic framework. Angew. Chem. Int. Ed. 47,
8826–8830 (2008).
9. Tilford, R. W., Mugavero, S. J., Pellechia, P. J. & Lavigne, J. J. Tailoring
microporosity in covalent organic frameworks. Adv. Mater. 20,
2741–2746 (2008).
31. List, B. Proline-catalyzed asymmetric reactions. Tetrahedron 58, 5573–5590 (2002).
32. Benaglia, M., Puglisi, A. & Cozzi, F. Polymer-supported organic catalysts. Chem.
Rev. 103, 3401–3429 (2003).
33. Lee, J. et al. Metal–organic framework materials as catalysts. Chem. Soc. Rev.
38, 1450–1459 (2009).
34. Yoon, M., Srirambalaji, R. & Kim, K. Homochiral metal–organic frameworks for
asymmetric heterogeneous catalysis. Chem. Rev. 112, 1196–1231 (2012).
35. Banerjee, M. et al. Postsynthetic modification switches an achiral framework to
catalytically active homochiral metal–organic porous materials. J. Am. Chem.
Soc. 131, 7524–7525 (2009).
36. Dang, D., Wu, P., He, C., Xie, Z. & Duan, C. Homochiral metal–organic frameworks
for heterogeneous asymmetric catalysis. J. Am. Chem. Soc. 132, 14321–14323 (2010).
37. Lun, D. J., Waterhouse, G. I. & Telfer, S. G. A general thermolabile protecting
group strategy for organocatalytic metal–organic frameworks. J. Am. Chem. Soc.
133, 5806–5809 (2011).
38. Notz, W., Tanaka, F. & Barbas, C. F. III. Enamine-based organocatalysis with
proline and diamines: the development of direct catalytic asymmetric
aldol, Mannich, Michael, and Diels–Alder reactions. Acc. Chem. Res. 37,
580–591 (2004).
39. Berner, O. M., Tedeschi, L. & Enders, D. Asymmetric Michael additions to
nitroalkenes. Eur. J. Org. Chem. 2002, 1877–1894 (2002).
40. Bock, D. A., Lehmann, C. W. & List, B. Crystal structures of proline-derived
enamines. Proc. Natl Acad. Sci. USA 107, 20636–20641 (2010).
41. Luo, S., Li, J., Zhang, L., Xu, H. & Cheng, J. P. Noncovalently supported
heterogeneous chiral amine catalysts for asymmetric direct aldol and Michael
addition reactions. Chem. Eur. J. 14, 1273–1281 (2008).
42. Wang, C. A. et al. ‘Bottom-up’ embedding of the Jorgensen–Hayashi catalyst
into a chiral porous polymer for highly efficient heterogeneous asymmetric
organocatalysis. Chem. Eur. J. 18, 6718–6723 (2012).
10. Wan, S., Guo, J., Kim, J., Ihee, H. & Jiang, D. A photoconductive covalent
organic framework: self-condensed arene cubes composed of eclipsed 2D
polypyrene sheets for photocurrent generation. Angew. Chem. Int. Ed.
48, 5439–5442 (2009).
11. Campbell, N. L., Clowes, R., Ritchie, L. K. & Cooper, A. I. Rapid microwave
synthesis and purification of porous covalent organic frameworks. Chem. Mater.
21, 204–206 (2009).
12. Spitler, E. L. & Dichtel, W. R. Lewis acid-catalysed formation of two-dimensional
phthalocyanine covalent organic frameworks. Nature Chem. 2, 672–677 (2010).
13. Wan, S. et al. Covalent organic frameworks with high charge carrier mobility.
Chem. Mater. 23, 4094–4097 (2011).
14. Uribe-Romo, F. J., Doonan, C. J., Furukawa, H., Oisaki, K. & Yaghi, O. M.
Crystalline covalent organic frameworks with hydrazone linkages. J. Am. Chem.
Soc. 133, 11478–11481 (2011).
15. Kandambeth, S. et al. Construction of crystalline 2D covalent organic
frameworks with remarkable chemical (acid/base) stability via a combined
reversible and irreversible route. J. Am. Chem. Soc. 134, 19524–19527 (2012).
16. Dalapati, S. et al. An azine-linked covalent organic framework. J. Am. Chem. Soc.
135, 17310–17313 (2013).
17. Kuhn, P., Antonietti, M. & Thomas, A. Porous, covalent triazine-based
frameworks prepared by ionothermal synthesis. Angew. Chem. Int. Ed. 47,
3450–3453 (2008).
18. Guo, J. et al. Conjugated organic framework with three-dimensionally ordered
stable structure and delocalized π clouds. Nature Commun. 4, 2736 (2013).
19. Lukose, B., Kuc, A. & Heine, T. The structure of layered covalent-organic
frameworks. Chem. Eur. J. 17, 2388–2392 (2011).
20. Nagai, A. et al. Pore surface engineering in covalent organic frameworks. Nature
Commun. 2, 536 (2011).
21. Dogru, M., Sonnauer, A., Gavryushin, A., Knochel, P. & Bein, T. A covalent
organic framework with 4 nm open pores. Chem. Commun. 47,
1707–1709 (2011).
22. Chen, X., Addicoat, M., Irle, S., Nagai, A. & Jiang, D. Control of crystallinity and
porosity of covalent organic frameworks by managing interlayer interactions
based on self-complementary π-electronic force. J. Am. Chem. Soc. 135,
546–549 (2013).
23. Biswal, B. P. et al. Mechanochemical synthesis of chemically stable isoreticular
covalent organic frameworks. J. Am. Chem. Soc. 135, 5328–5331 (2013).
24. Kandambeth, S. et al. Enhancement of chemical stability and crystallinity in
porphyrin-containing covalent organic frameworks by intramolecular hydrogen
bonds. Angew. Chem. Int. Ed. 52, 13052–13056 (2013).
25. Chandra, S. et al. Phosphoric acid loaded azo-based covalent organic framework
for proton conduction. J. Am. Chem. Soc. 136, 6570–6573 (2014).
26. Du, Y. et al. Experimental and computational studies of pyridine-assisted post-
synthesis modified air stable covalent-organic frameworks. Chem. Commun.
48, 4606–4608 (2012).
Acknowledgements
D.J. acknowledges the support of a Grant-in-Aid for Scientific Research (A) (24245030)
from the Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT).
Author contributions
D.J. conceived the project, designed the experiments and provided funding. H.X. conducted
the experiments and J.G. performed computational calculations. D.J. and H.X. wrote
the manuscript.
Additional information
Supplementary information is available in the online version of the paper. Reprints and
permissions information is available online at www.nature.com/reprints. Correspondence and
requests for materials should be addressed to D.J.
Competing financial interests
The authors declare no competing financial interests.
8
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2015 Macmillan Publishers Limited. All rights reserved