Salicylideneanilines-Based Covalent Organic Frameworks as Chemoselective Molecular Sieves

Article abstract of DOI:10.1021/jacs.7b02696

Porous materials such as covalent organic frameworks (COFs) are good candidates for molecular sieves due to the chemical diversity of their building blocks, which allows fine-tuning of their chemical and physical properties by design. Tailored synthesis o

Full text of DOI:10.1021/jacs.7b02696

Subscriber access provided by Grand Valley State | University
Article
Salicylideneanilines-based Covalent Organic
Frameworks as Chemo-Selective Molecular Sieves
Guo-Hong Ning, Zixuan Chen, Qiang Gao, Wei Tang, Zhongxin
Chen, Cuibo Liu, Bingbing Tian, Xing Li, and Kian Ping Loh
J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 15 Jun 2017
Downloaded from http://pubs.acs.org on June 15, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Salicylideneanilines-based Covalent Organic Frameworks as
Chemo-Selective Molecular Sieves
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
GuoꢀHong Ning,1 Zixuan Chen,1 Qiang Gao,1 Wei Tang,2 Zhongxin Chen,1 Cuibo Liu,1 Bingbing Tian,1
Xing Li1 and Kian Ping Loh1*
1Department of Chemistry, National University of Singapore, 3 Science Drive 3, Singapore 117543
2Institute of Materials Research and Engineering, A*STAR, 2 Fusionopolis Way, Innovis, Singapore 138634, Singapore.
ABSTRACT: Porous materials such as covalentꢀorganic frameworks (COFs) are good candidates for molecular sieves due to the
chemical diversity of its building blocks, which allows fineꢀtuning of its chemical and physical properties by design. Tailored synꢀ
thesis of inherently functional building blocks can generate framework materials with chemoꢀresponsivity, leading to controllable
functionalities such as switchable sorption and separation. Herein, we demonstrate chemoꢀselective, salicylideneanilinesꢀbased
COF (SAꢀCOF), which undergo solventꢀtriggered tautomeric switching. This is unique compared to solidꢀstate salicylideneanilines
counterpart, which typically requires high energy input such as photo or thermal activation to trigger the enolꢀketo tautomerisim
and cisꢀtrans isomerization. Accompanying the tautomerization, the ionic properties of the COF can be tuned reversibly, thus formꢀ
ing the basis of sizeꢀexclusion, selective ionic binding or chemoꢀseparation in SAꢀCOF demonstrated in this work.
Introduction
Covalent organic frameworks (COFs) are a class of crystalline
porous polymer that allows atomically precise integration of
functionalities into an extended, porous network.1ꢀ5 Although
COFs have great application potentials in wide ranging fields
including gas storage,6,7 energy storage,8ꢀ10 drug delivery,11
catalysis12ꢀ14 and electronic application,15ꢀ17 molecular separaꢀ
tion on the basis of size, charge and functionality by using
COFs’ ordered pores and channels are not widely reported.
One challenge is the difficulty in tailoring the interactions
between pore surface and sorbents. Wang’s group used preꢀ
synthetic modification to prepare the thioetherꢀfunctionalized
COFꢀLZU8 for sensing and removal of mercury (II) ions.18
Jiang’s group improved CO2 adsorption by postsynthetic modꢀ
ification of COFs’ pore surface with various functional
groups.7 Incorporating inborn functionalities into the pores
which can interact with molecules based on nonꢀcovalent inꢀ
teractions are attractive due to the ability to engender a chemoꢀ
responsive response to the environment. One strategy is to
construct the COF from monomers which can undergo dynamꢀ
ic structural changes via coordination or nonꢀcovalent interacꢀ
tions, i.e. protonationꢀinduced tautomerization or cisꢀtrans
isomerization. Triggered by environmental cues (solvent and
pH), dramatic changes in the acidic character or ionicity of the
functional groups occurs following tautomrization and cis-
trans isomerization, thus imparting inborn selectivity to tarꢀ
geted molecules in the pores. In actuality, the properties of the
porous framework structure are often different from the buildꢀ
ing blocks due to the crossꢀlinking and tight packing of the
molecular layers. The rigidity of the COF framework can also
restrict the rotational or flexural motions of the building
blocks. Thus far there are no reports of reticular synthesis of
COF, which can undergo solventꢀtriggered tautomerization.
Scheme 1. Reversible proton tautomerism of salicylideneaniꢀ
lines under themoꢀ and photoꢀchromism pathway. The transꢀ
keto form bears two accessible hydrogenꢀbonding sites.
Salicylideneanilines (SA), a class of Schiff base comprising
salicylaldehyde and aniline derivatives, have attracted great
attentions due to their unique photoꢀ and thermoꢀchromic
properties in the solidꢀstate, which can find applications in
sensors, optical data storage and displays devices.19ꢀ22 The
mechanism of reversible color change of SA in response to
external simulation (i.e. temperature and light) in the solidꢀ
state is well studied and suggests a tautomeric structural transꢀ
formation promoted by proton transfer.23ꢀ25 Firstly, an equilibꢀ
rium exists between the white colored enol form at lower temꢀ
perature and the yellow colored cisꢀketo form at room temperꢀ
ature (r.t.), and in both forms, a cyclic intermediate stabilized
by intramolecular hydrogen bonding is observed (Scheme 1.).
Secondly, the isomerization between the cisꢀketo and the red
colored transꢀketo form can be induced by light.23,24 The reꢀ
sulting transꢀketo form relaxes back to the cisꢀketo form via
thermoꢀ or photoꢀisomerization (Scheme 1).25 Accompanying
the reversible color change of SA, the hydrogen and ionic
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. a) Conceptual representation for synthesis of solvatochromic SA-COF and reversible proton tautomerism triggered by water
adsorption and desorption; b) synthesis of reference compound SA-TFP.
bonding properties of SA changes dramatically during the
proton tautomerism (slight acidic in enol form and basic in
cisꢀ and transꢀketo form) (Scheme 1).
mation between cisꢀ and transꢀketo form in solidꢀstate can be
initiated by low energy input i.e. solvent, instead of high
energy input i.e. photoꢀirradiation. Importantly, the tautomꢀ
erization allows a dynamic change in the ionic properties of
the pores in response to chemical cues, thus forming the
basis of sizeꢀexclusion or selective ionic binding. The tunaꢀ
ble functionality of pore surface allows SA-COF to bind to
positively charged molecules in basic condition, while
The reversible switching of functionality in SA motivated
us to consider integrating SA moieties into COFs to make
SA-COF. Interestingly, SA-COF exhibits reversible proton
tautomerism triggered by adsorption and desorption of water
molecules (Fig. 1). This is the first report that the transforꢀ
ACS Paragon Plus Environment

Page 3 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Figure 2. Crystal structure analysis, isotherm profiles of N2ꢀgas sorption and electronic microscope image of SA-COF. a, Experimental
(black line) and refined (red dots) PXRD pattern of SA-COF. b, Difference curve between experimental and refined PXRD patterns. c and
d, Simulated PXRD pattern of SA-COF with AA (red line) and AB (blue line) stacking structure. e and f, Top view of the optimized SA-
COF in the eclipsed AA stacking and staggered AB stacking mode shown as space filing model. g, N2 adsorption (filled) and desorption
(open) isotherm profiles of SA-COF (STP, standard temperature and pressure). h, Poreꢀsize distribution of SA-COF calculated by nonꢀ
local DFT modeling based on N2 adsorption data. i and j showing high resolution TEM image of SA-COF and arrangement of the microꢀ
spores in crystalline domain.
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
excluding it in acidic condition. More interestingly, SA-
COF selectively binds molecules with aromatic hydroxyl
groups over aromatic amino groups following conversion to
the transꢀketo stage (Scheme 1).
ance (665, 608 and 664 nm for MB in neutral, basic and
acidic condition, respectively; 413 nm for AA in basic condiꢀ
tion; 419 nm for CA; 550 nm for RB; 590 nm for DAQ and
546 nm for DHQ). The removal efficiency of dye was calcuꢀ
lated as following equation (1):
Experimental Section
Removal efficiency (%) = (C0 − Ct) / C0 × 100
(1)
Synthesis of SA-COF. TFP (14.3 mg, 0.80 mmol), TAB
(27.5 mg, 0.80 mmol), 0.5 mL of nꢀbutanol, 0.5 mL of 1,2ꢀ
dichorobenzene and 0.1 mL of 6 M aqueous acetic acid was
added into a 10 mL Schlenk storage tube (SynthwareTM,
OD28 x L80mm, high vaccum valve size 0ꢀ8mm, with PTFE
oꢀring, wiper).26 This mixture was sonicated for 15 mins. in
order to get a homogenous dispersion. The tube was then
flash frozen at 77 K (liquid N2 bath) and degassed with three
freezeꢀpumpꢀthaw cycles, and then the tube was sealed and
heated at 120 °C for 7 days. A dark red colored precipitate
produced was collected by decantation and washed and solꢀ
vent exchanged with anhydrous tetrahydrofuran (10 mL)
several times. The resultants were dried at 120 °C under
vacuum overnight to give a light yellow colored powder in
81% (30 mg) isolated yield. FTꢀIR (KBr, cm−1): 2955 (w),
2925 (w), 2863 (w), 1631 (s, C=O), 1582 (s, C=N), 1501 (s,
C=C in enamine), 1451 (s, aromatic C=C), 1141 (s), 829 (s).
Anal. Calcld. For C99O3H63N9: C, 83.35; H, 4.45; N, 8.84%;
found: C, 80.55; H, 5.05; N, 8.04%.
Where C0 and Ct are the concentration of dyes at initial conꢀ
dition and in the filtrate, respectively.
For the molecular separation study, 5 mg of SA-COF was
suspended in 5 mL water solution (pH =12) for 5 mins, and
then passed through a Whatman (0.45 ꢁm) membrane filter
to give SA-COF equipped filter. The resulting SA-COF
equipped filter was washed with 5 mL water solution (pH
=12) once before using it for the separation experiments
without any further process.
HPLC was performed with an Agilent 1200 HPLC
equipped with a DiodeꢀArray Detector ultraviolet detector
and a Phenomenex® Luna C18 (2) column (2.0 x 150 mm, 5
ꢁm particle size). The sample injection volume was 20 ꢁL,
and the flow rate was 0.7 mL min−1. The mobile phase for
HPLC analysis is 100% methanol. The intensity of the effluꢀ
ent ultraviolet absorbance was monitored at 590 and 468 nm
for DAQ and DHQ, respectively. LC calibration curve was
created for quantitative analysis using four standard soluꢀ
tions, that is, 1, 10, 50, and 100 ꢁM (Supplementary Fig.
S17).
Selective dye separation experiment. All the batch adsorpꢀ
tion experiments were conducted at r.t. in a dark condition.
Initial dye concentrations were fixed to be 50 ꢁM. Typically
5 mg of SA-COF was added into 5 mL of dye solution, and
the mixture was sonicated by 10 second to get a wellꢀ
dispersed solution and then stirred at r.t. on hot plate with
700 r.p.m. stirring rate. At appropriate time interval, the
mixture was filtered in syringes equipped with Whatman
0.45 ꢁm membrane filters. The concentration of dye in the
filtrate was detected using an ultravioletꢀvisible spectrometer
(Shimadzu UVꢀ2600) at a wavelength of maximum absorbꢀ
Results and Discussion
Preparation of salicylideneanilines-based covalent organ-
ic frameworks. When triformylphloroglucinol ligand with
three –OH group is used for synthesis of COFs, it produces a
mechanically rigid βꢀketoenamineꢀlinked COFs and no reꢀ
versible proton tautomerism is observed. This is because the
strong intramolecular hydrogen bonding interactions preꢀ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 9
vents the reversible proton tautomerism and locks the COFs
in the ketoenamine stage.27 To circumvent this, we proposed
that the incorporation of the 2,4,6ꢀtriformayl phenol (TFP)
moiety with only one –OH group would provide a less rigid
βꢀketoenamine intermediate and imparts structural flexibility
around the pores of the COF needed for undergoing reversiꢀ
ble tautomerism.
1
2
3
4
5
6
7
8
9
Phenol, a lowꢀcost and commercially available chemical,
was used to synthesize the TFP building block in gram scale
with 56% yield via a Duff reaction.28 SA-COF was prepared
by the solvothermolysis of a suspension of TFP and TAB
(1,3,5ꢀTris(4ꢀaminophenyl)benzene) in a 5:5:1 (v/v) mixture
of 1ꢀbuthanol, dichlorobenzene and 6 M aqueous acetic acid
(Fig. 1). The Fourier transform infrared (FTꢀIR) spectra of
the SA-COF confirm the formation of imine linkage in the
polymers, supporting by disappearance of the N−H stretches
signals located at 3450 to 3204 cm−1 and exhibition of the
C=N stretches bands located at 1582 cm−1 (Fig. 3a and S2).
The characteristic resonance peaks of imine carbons at 158.9
and 156.1 ppm were observed in solidꢀstate 13C crossꢀ
polarization/magic angle spinning nuclear magnetic resoꢀ
nance (CP/MAS NMR) spectra of SA-COF, evidencing the
existence of imine linkages (Fig. 3e). Thermal gravimetric
analyses (TGA) under N2 atmosphere show that SA-COF is
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
o
highly stable up to 500 C with only 3.8% weight loss (Fig.
S3). Furthermore, scanning electron microscopy (SEM) and
(TEM) reveal that it displays plaitꢀlike morphology, and
consists of highly crystalline nanoꢀflakes (Fig. 2i, j, S5 and
S6).
To analyze the crystal structure of obtained polymer SA-
COF, theoretical simulations and powder Xꢀray diffraction
(PXRD) experiments were performed. The calculations were
conducted using Material Studio version 2016 and the
eclipsed stacking (AA) and staggered stacking (AB) strucꢀ
tures were modeled (See Supporting Information for details).
In the experimental PXRD pattern of SA-COF (Figure 2a
black curve), an intense peak at 5.68° accompanied by four
small peaks at 9.76°, 11.20°, 14.89° and ca. 25.16° were
observed, which can be assigned to (100), (110), (200), (120)
and (001) diffractions. The obtained PXRD pattern matches
well with the calculated PXRD pattern of AA stacking strucꢀ
ture (Fig. 2a, c), suggesting that SA-COF features a highly
uniform pore structure with AA stacking model. Specificalꢀ
ly, Pawley refinement produces a hexagonal space group
with unit cell parameters of a = b = 18.09 Å, c= 3.61 Å, α =
β = 90°, and γ = 120°, with refinement results of Rp = 2.91%
and Rwp = 4.41%. The refined PXRD pattern agrees well
with the experimental PXRD data, as confirmed by the vanꢀ
ishing intensity of the difference plot in Fig. 2a,b.
Figure 3. Reversible Proton tautomerism evidenced by FTꢀIR
and solidꢀstate CP/MAS NMR. FTꢀIR spectra of a, dry SA-
COF b, SA-COF-H2O after exposure to air, and c, SA-COF
after desorption of water from SA-COF-H2O. d, schematic
presentation of enol and transꢀketo isomer (the cisꢀketo form is
omitted for clarity) Solidꢀstate CP/MAS NMR spectra (100
MHz, 300 K) and assignment of e, SA-COF; f, SA-COF-H2O
and g, SA-COF produced by desorption of water from SA-
COF-H2O; h, 13C NMR (500 MHz, 300 K, CDCl3) spectrum of
reference compound SAꢀTFP, exhibiting the coꢀexistence of
enol and cisꢀketo form. (*C=O and *C=CꢀN represent the carꢀ
bonyl and enamine carbons in cisꢀketo form, respectively)
The eclipsed AA stacking structure of SA-COF is further
supported by the pore size distribution analysis. Nitrogen
adsorption−desorption experiment at 77 K (Fig. 2g) exhibits
a type I N2 adsorption isotherm that is characteristic of a
microporous structure. The Brunauer−Emmett−Teller (BET)
surface area of SA-COF is calculated as 1588.56 m2 g−1, and
the total pore volume is 0.92 cm3 g−1 (P/P0 = 0.99). The simꢀ
ulated eclipsedꢀstacked SA-COF structure using nonlocal
density functional theory (NLDFT) presents a narrow pore
size distribution with an average pore width of ~1.43 nm
(Fig. 2h), which is very close to the theoretical value of 1.45
nm based on the eclipsed AA stacking structure (Fig. 2e).
Proton tautomerism triggered by adsorption and desorp-
tion of water. When asꢀsynthesized SA-COF powder is
exposed to air, its color changes quickly from yellow to deep
red (Fig. S2), which hints at the existence of proton tautomꢀ
erism. To investigate the transformation between enol, cisꢀ
keto and transꢀketo form, UVꢀvis spectroscopic analysis,
FTꢀIR measurement and solidꢀstate CP/MAS NMR studies
ACS Paragon Plus Environment

Page 5 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Figure 4. Dye molecules used in molecular separation experiments based on size and charge. Molecular model (top) calculated from Mateꢀ
rial Studio 2016, displayed in space filling style (Gray, carbon; Red, oxygen; White, hydrogen; Blue, nitrogen; Green, chlorine; Yellow,
sulfur.) and chemical structure of dye molecules (bottom); a, Anthraflavic acid (AA) with size of 8.08 × 12.66 Å. b, Methylene blue (MB)
with size of 8.54 × 15.80 Å. c, Rhodamine B (RB) with size of 13.45 × 16.58 Å. d, Chrome azurol S (CA) with diameter of 14.50 Å.
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
were conducted. In addition, to better understand the proton
tautomerizaion at a molecular level, the molecular reference
compound SA-TFP was prepared (Fig. 1b), for which the
coꢀexsitence of enol and cisꢀketo form in solution and solidꢀ
sate was confirmed by FTꢀIR, Xꢀray, Uvꢀvis and NMR analꢀ
ysis (Fig. 1b and See Supporting Information for details).
UVꢀvis spectroscopy is a sensitive tool for probing the tauꢀ
tomeric equilibrium in SA compounds. The band below 425
nm can be attributed to the enol form whereas a contribution
peak of cisꢀketo form appears between 425 and 550 nm, and
the signals above 550 nm can be assigned to transꢀketo
form.29,30 The UVꢀvis spectra of wellꢀdispersed SA-COF in
anhydrate THF was recorded and it display two major
broadened signals at 300 and 450 nm (Fig. S7a), which indiꢀ
cates the coꢀexistence of enol and cisꢀketo form in the prisꢀ
tine SA-COF at room temperature. In addition, this result is
further supported by UVꢀvis spectra of the reference comꢀ
pound SA-TFP, which shows similar absorption spectrum
(Fig. S7b). Without exposure to water molecules, the yellow
color of SA-COF originates from the cisꢀketo form. To furꢀ
ther elucidate the structure of SA-COF, FTꢀIR measureꢀ
ments were conducted. As shown in Fig. 3a, the FTꢀIR specꢀ
tra of pristine sample shows a broad absorption band in the
region from 3030 to 2850 cm−1 similar to that in SA-TFP
(Fig. S2), which indicates the presence of strong resonanceꢀ
assisted intermolecular hydrogen bonding, reflected by a
large amplitude of O−H stretching vibration or proton transꢀ
fer equilibrium between O−H enol form and N−H cisꢀketo
form.31 In addition, the intense peaks at 1622 and 1503 cm−1
originate from C=O and C=C stretching, confirming the exꢀ
istence of cisꢀketo form. Furthermore, solidꢀstate CP/MAS
NMR spectra of SA-COF evidence the proton tautomerism
between enol and cisꢀketo form and agree well with the UVꢀ
vis and FTꢀIR results. Fig. 3e shows two very weak characꢀ
teristic 13C resonances of carbonyl (*C=O) and enamine
(*C=C−N) carbons at 188.7 and 114.3 ppm, respectively,
which are similar to that in the reference compound (190.0
and 115.1 ppm, see Fig. S8), confirming the existence of cisꢀ
keto form as minor component. Besides, several characterisꢀ
tic 13C resonances of enol form including phenol (C1) and
imine (C7, C8 and C9) carbons at 162.2, 158.1 and 156.1
ppm are observed, which are further supported by compariꢀ
son with 13C NMR spectrum of the reference compound SA-
TFP (164.3 and 158.4 ppm) (Fig. 3 and S8).32,33
When SA-COF is exposed to air, it quick absorbs water to
give red colored SA-COF-H2O. The UVꢀvis spectra of wellꢀ
dispersed SA-COF-H2O in H2O display two major broad
signals at 430 and 587 nm (Fig. S7), indicating the absence
of the enol form and the coꢀexistence of cisꢀketo and transꢀ
keto form in the SA-COF-H2O. The resulting dark red color
of SA-COF-H2O comes from the transꢀketo form. In addiꢀ
tion, the FTꢀIR profile of SA-COF-H2O (Fig. 3b) further
proves the transformation from enol to keto form by the folꢀ
lowing changes: i) the appearance of a large broad stretching
peaks at ~3420 cm−1 assignable to absorbed H2O molecules;
ii) the vanishing of the O−H stretching signals located
around 2900 cm−1; iii) the increment of peak intensity at
1621 cm−1 attributed to C=O stretching; iv) the decrease of
signal intensity at 1582 cm−1 assigned to C=N stretching; v)
appearance of a new peak at 1531 cm−1 corresponded to C=C
stretching in transꢀβꢀketoenamine, and (vi) the decline of
peck intensity at 1450 cm−1 assignable to the aromatic C=C
stretching. More importantly, the solidꢀstate CP/MAS NMR
spectra of SA-COF-H2O (Fig. 3f) also support such transꢀ
formation by the following evidences: i) increment of signal
intensity of carbonyl carbon (C1’) at 190.7 ppm correspondꢀ
ing to the increment of transꢀketo population; ii) low field
shift of the cisꢀketo at 188.7 ppm to 190.7 ppm attributed to
breaking of intramolecular hydrogen bonding; ii) the disapꢀ
pearance of C1 and C7 resonances in enol form; iii) the rise
of peak intensity at 116.3 ppm attributed to C6’ resonances
in transꢀketo form. Interestingly, when SA-COF-H2O is
o
heated at 120 C under vacuum, adsorbed water molecules
are removed and it returns back to original yellow colored
SA-COF with identical UVꢀvis, FTꢀIR and solidꢀstate
CP/MAS NMR spectra (Fig. 3c,g). All these data clearly
demonstrate reversible proton tautomerism between enol,
cisꢀ and transꢀketo in SA-COF, which is readily triggered by
the adsorption and desorption of water molecules (See the
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 9
proposed mechanism of proton tautomersim in the Supportꢀ showing a large increment of uptake rate compared to the
1
2
3
4
5
6
ing Information)
neutral SA-COF (98.3% removal efficiency at 10 mins.),
suggesting that electrostatic attraction is operational here.
However, when the dye AA and MB were treated with
Size-dependent separation. The removal of waterꢀsoluble
organic pollutants by porous material is an important reꢀ
search field due to rising environmental problems.34ꢀ36 The
tunable functionality of pore surface in SA-COF induced by
proton tautomerism can be exploited for the separation of
7
8
9
molecules based on size, charge and functionality.
To
demonstrate size selective adsorption, we have chosen severꢀ
al waterꢀsoluble dye molecules (Fig. 4), specifically Methꢀ
ylene blue (MB), Rhodamine B (RB) and Chrome azurol S
(CA). As shown in Fig. 4, the van der Waals sizes of these
molecules are calculated using Material Studio 2016. The
size of these dye molecules follow the order MB < RB ≲
pore size of SA-COF < CA. Ultravioletꢀvisible (Uvꢀvis)
spectroscopy was used to monitor the change of concentraꢀ
tion of dye in fixed time interval. SA-COF has the ability to
quickly adsorb and remove organic dye (MB) from water
(Fig. 5a). The size exclusion effect of SA-COF was tested
by selecting three molecules with different sizes, and we
found that the binding affinity of SA-COF for dye molecules
declines with an increase in molecular size. As shown in
Fig. 5b, the smallest size MB was completely removed withꢀ
in 10 minutes, showing a removal efficiency of 98.3%, while
the larger size RB, which is slightly smaller than the pore
size of SA-COF, exhibits a much slower adsorption rate (~
60 mins), and with 94.1% removal efficiency compared to
that of MB. The largest dye, CA, whose size is larger than
1.45 nm, exhibits no noticeable change of its concentration
during the adsorption tests (Fig. 5b). Meanwhile, when the
dye MB and CA were treated with commercial porous actiꢀ
vated carbon, it does not show good separation performance
(See Fig. S16). The sizeꢀdependent adsorption reflects the
highly uniform pore size distribution of SA-COF (~1.43
nm), rendering it a useful molecular sieve material based on
sizeꢀexclusion. To test this, SA-COF powders was suspendꢀ
ed and stirred in a green colored aqueous solution of MB and
CA (1:1 molar ratio) for 10 mins. at r.t. After filtration, the
UVꢀvis spectra before and after SA-COF treatment show
that the MB and CA mixture can be completely separated,
the concentration of MB is decreased by 99.9%, while CA
does not show any notable change (Fig. 5d). It should be
stated here that the size of the dye molecule may not be the
only factor at play, charge and functional groups in the dye
molecules may act in concert with its size to effect sizeꢀ
dependent adsorption in the SA-COF.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Size and chargeꢀselective separation experiments. a,
UVꢀvis absorption spectra of an aqueous solution of MB after
treatment with SA-COF at different intervals. b, Change in
concentration of dye over time after treatment with SA-COF,
determined by change in absorbance relative to initial absorbꢀ
ance (C / C0). The data is averaged over three test trials. c, UV–
vis absorption spectra of sizeꢀselective separation of CA from a
mixture of CA and MB in water (green and orange line repreꢀ
sent before and after treatment with SA-COF, respectively;
inset, photographs of CA and MB mixture and filtrate after
treatment with SA-COF, showing green and deep yellow color).
d, UV–vis absorption spectra of chargeꢀ selective separation of
AA from mixture of AA and MB in water (green and orange
line represent before and after treatment with SA-COF, respecꢀ
tively; inset, photoꢀimage of green colored CA and MB mixture,
deep yellow colored filtrate after passing though the filter disc
charged with SA-COF powders)
commercial porous activated carbon at the same condition, it
does not show any selective separation (See Fig. S16). To
further test the charge separation effect, a green colored AA
and MB mixture in 1:1 ratio was quickly passed through (~
10 seconds) a filter in which the base treated SA-COF powꢀ
ders were preꢀequipped (details see experimental section).
The UVꢀvis spectra before and after filter treatment revealed
that the AA and MB mixture was clearly separated, and there
was a sharp decrease of MB concentration (95.2%) (Fig. 5d),
however, almost no change was observed for concentration
of AA. More interestingly, the absorption ability of SA-
COF toward MB can be reversed. When the SA-COF powꢀ
der was added into an acidic aqueous solution of MB, the
concentration of MB showed a much slower decrease with
22% removal efficiency within 5 mins (Fig. 5b) than that in
basic condition (99.8% removal efficiency at 10 seconds).
Charge selective separation. Since the –OH or –NH moieꢀ
ties in SA-COF can be deprotonated or protonated in basic
or acidic condition, respectively, SA-COF is expected to a
promising platform for separation of charged molecules due
to its pHꢀcontrollable ionicity. Before the separation test, the
stability of SA-COF in base (pH = 12, 0.01 M NaOH) or
acid (pH = 1, 0.1 M HCl) was examined. SA-COF retained
its original crystalline structure in base or acid, as indicated
by the unchanged intensities and positions of the peaks in its
PXRD profile (Fig. S12). SA-COF powder was added to an
aqueous solution of Anthraflavic acid (AA) (Fig. 4a), a negaꢀ
tive charged dye at basic condition (pH = 12), and no reꢀ
markable change in the concentration of AA was found, as
confirmed by UVꢀvis spectroscopy (Fig. 5b, S11). In conꢀ
trast, MB, which has a similar size to AA but positive
charged instead, was rapidly absorbed within 10 seconds
(99.8% removal efficiency) at the same condition (Fig. 5b),
Chemo-selective separation. Owing to the basic N–H moieꢀ
ty in the transꢀketo form, SA-COF can discriminate funcꢀ
tional groups with different acidity and it selectively bind –
OH over –NH2 groups under neutral condition. Two moleꢀ
cules of the same size, but with different functional groups,
1,4ꢀdihydroxyanthraquinone
(DHQ)
and
1,4ꢀ
diaminoanthraquinone (DAQ), are chosen as model comꢀ
pound (Fig. 6a). SA-COF powder was added to a solution of
DHQ in ethanol and water mixed solvent (VEtOH/VH2O = 2/3),
ACS Paragon Plus Environment

Page 7 of 9
Journal of the American Chemical Society
and a significant decrease in concentration of DHQ with an avenue for designing smart, chemoꢀresponsive COFs,
1
2
3
4
almost 90% removal efficiency within 60 mins was observed
by UVꢀvis spectroscopy. In contrast, the concentration of
DAQ was slightly only decreased with ~ 25% removal effiꢀ
ciency within 60 mins, resulting a high selectivity with a
further attesting to the utility of COFs as flexible and tunable
materials for molecular separation application.
5
6
7
ASSOCIATED CONTENT
Supporting Information
8
9
The Supporting Information is available free of charge on the
ACS Publications website.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Additional experimental procedures and structural modeling, IR,
TGA, BET analysis, SEM and TEM images, Uvꢀvis spectrum,
Solidꢀstate CP/MAS NMR spectra, the proposed mechanism of
proton tautomersim, selectivity experiment, control experiment
and HPLC analysis (PDF)
AUTHOR INFORMATION
Corresponding Author
* chmlohkp@nus.edu.sg
Notes
The authors declare no competing financial interests.
ACKNOWLEDGMENT
K. P. Loh acknowledges NRFꢀCRP grant “Two Dimensional
Covalent Organic Framework: Synthesis and Applications.”
Grant number NRFꢀCRP16ꢀ2015ꢀ02, funded by National Reꢀ
search Foundation, Prime Minister’s Office, Singapore.
Figure 6. Chemoꢀselective separation experiments. a, UV
visible absorbance spectra of DAQ and DHQ at different
time interval after treatment with SA-COF (purple, DAQ;
orange, DHQ) The data are average of triplicate experiments.
b, HPLC spectra of DAQ and DHQ mixture in 1:1 molar
ratio before and after treatment with SA-COF (full line:
before; dotted line: after; purple (DAQ) and orange (DHQ)
peaks with 0.693 and 0.914 mins retention times, respectiveꢀ
ly. The * marked peak represent the solvent peak)
REFERENCES
(1) Adrien P. Côté, A. P., Benin, A. I., Ockwig, N. W., O’Keeffe,
M., Matzger, A. J., Yaghi, O. M. Science 2005, 310, 1166–
1170.
(2) Waller, P. J., Gándara, F., Yaghi, O. M. Acc. Chem. Res.
2015, 48, 3053–3063.
(3) Feng, X., Ding, X., Jiang, D. Chem. Soc. Res. 2012, 41,
6010–6022.
(4) Ding, S. Y., Wang, W. Chem. Soc. Rev. 2013, 42, 548–568.
(5) Slater, A. G., Cooper, A. I. Science 2015, 348, aaa8075.
(6) Doonan, C. J., Tranchemontagne, D. J., Glover, T. G., Hunt,
J. R., Yaghi O. M. Nat. Chem. 2010, 2, 235–238.
(7) Huang, N., Krishna, R., Jiang, D. J. Am. Chem. Soc. 2015, 37,
7079–7082.
(8) DeBlase, C. R., Silberstein, K. E., Truong, T.ꢀT., Abruña, H.
D., Dichtel, W. R. J. Am. Chem. Soc. 2015, 135, 16821–
16824.
ratio of 1:7 for DHQ : DAQ in the filtrate (Fig. 6a). The moꢀ
lecular selectivity is further demonstrated in a separation
experiment: when a 1:1 mixture of DHQ and DAQ was
treated with SA-COF powder, within 10 mins, the ratio for
DHQ:DAQ were found to be 1:6 in the filtrate, as confirmed
by the HPLC analysis (Fig. 6b).
Conclusion
We have successfully synthesized salicylideneanilinesꢀbased
COF, which exhibits reversible solvatochromism triggered
by adsorption and desorption of water molecules. Our studꢀ
ies show that inborn, chemoꢀresponsive pores can be proꢀ
grammed into COF by reticular synthesis. The reversible
proton tautomerism was confirmed by UVꢀvis spectroscopy,
FTꢀIR and solidꢀstate CP/MAS NMR analysis. The tautomꢀ
erization of SA-COF stimulates dynamic changes in the
ionic and chemical properties of the pores and allows molecꢀ
ular separation on the basis of sizeꢀexclusion, charge separaꢀ
tion and functional group discrimination. Specifically, it
binds positively charged molecules in basic condition, while
excluding them in acidic condition. Moreover, SA-COF can
discriminate functional groups with different degree of acidiꢀ
ty i.e. it prefers to bind to molecules with aromatic hydroxyl
groups rather than aromatic amino groups. Our studies open
(9) Xu, F., Jin, S., Zhong, H., Wu, D., Yang, X., Chen, X., Wei,
H., Fu. R., Jiang, D. Sci. Rep. 2015, 5, 8225.
(10) Liu, W., Luo, X., Bao, Y., Liu, Y. P., Ning, G.ꢀH., Abdelꢀ
wahab, I., Li, L., Nai, C. T., Hu, Z. G., Zhao, D., Liu, B.,
Quek, S.Y., Loh, K. P. Nat. Chem. 2017, 9, 563ꢀ570.
(11) (a) Fang, Q., Wang, J., Gu, S., Kaspar, R. B., Zhuang, Z.,
Zheng, J., Guo, H., Qiu, S., Yan, Y. J. Am. Chem. Soc.,
2015, 137, 8352–8355; (b) Fang, Q., Zhuang, Z., Gu, S., Kasꢀ
par, R. B., Zheng, J., Wang, J., Qiu, S., Yan, Y. Nat. Com-
mun. 2014, 5, 4503.
(12) Ding, S.ꢀY., Gao, J., Wang, Q., Zhang, Y., Song, W.ꢀG., Su,
C.ꢀY., Wang, W. J. Am. Chem. Soc., 2011, 133, 19816–
19822.
(13) Lin, S., Diercks, C. S., Zhang, Y.ꢀB., Kornienko, N., Nichols,
E. M., Zhao, Y., Paris, A. R., Kim, D., Yang, P., Yaghi, O.
M., Chang, C. J. Science 2015, 349, 1208–1213.
(14) Xu, H., Gao, J., Jiang, D. Nat. Chem. 2015, 7, 905–912.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 9
(15) Ding, X., Guo, J., Feng, X., Honsho, Y., Guo, J., Seki, S.,
Bai, L., Zhang, X., Wang, P., Li, P., Zeng, Y., Zou, R., Zhao,
Maitarad, P., Saeki, A., Nagase, S., Jiang, D. Angew. Chem.
Int. Ed. 2011, 50, 1289–1293.
Y. Chin. J. Chem. 2015, 33, 90ꢀ94.
1
2
3
4
5
6
7
8
(27) Kandambeth, S., Mallick, A., Lukose, B., Mane, M. V., Heiꢀ
ne, T., Banerjee, R. J. Am. Chem. Soc., 2012, 134, 19524–
19527.
(28) Anderson, A. A., Goetzen, T., Shackelford, S. A., Tsank, S.
Synth. Commun., 2007, 30, 3227–3232.
(29) Fujiwara, T., Harada, J., Ogawa, K., J. Phys. Chem. B, 2004,
108, 4035–4038.
(30) Jacquemin, P.ꢀL., Robeyns, K., Devillers, M., Garcia, Y.
Chem. Eur. J. 2015, 21, 6832–6845.
(31) Ambroziaka, K., Rozwadowskia, Z., Dziembowskaa, T.,
Biegb, B. J. Mol. Struct. 2002, 615, 109–120.
(32) Schilfa, W., Kamieńskia, B., Kołodziejb, B., Grechb, E.,
Rozwadowskib, Z., Dziembowskab, T. J. Mol. Struct. 2002,
615, 141–146.
(33) Claramunt, R. M., López, C., María, M. D. S., Sanz, D. and
Elguero, J. Prog. NMR Spectrosc. 2006, 49, 169–206.
(34) Schwarzenbach, R. P., Escher, B. I., Fenner, K., Hofstetter, T.
B., Johnson, C. A., Gunten, U., Wehrli, B. Science 2006, 313,
1072–1077.
(35) Alsbaiee1, A., Smith, B. J., Xiao, L., Ling, Y., Helbling, D.
E., Dichtel, W. R. Nature 2016, 529, 190–194.
(36) Byun, J., Patel, H. A., Thirion, D., Yavuz, C. T. Nat. Com-
mun. 2016, 7, 13377.
(16) Guo, J., Xu, Y., Jin, S., Chen, L., Kaji, T., Honsho, Y., Adꢀ
dicoat, M. A., Kim, J., Saeki, A., Ihee, H., Seki, S., Irle, S.,
Hiramoto, M., Gao, J., Jiang, D. Nat. Commun. 2013, 4, 2736.
(17) Xu, H., Tao, S., Jiang, D. Nat. Mater. 2016, 15, 722–726.
(18) Ding, S.ꢀY., Dong, M., Wang, Y.ꢀW., Chen, Y.ꢀT., Wang, H.ꢀ
Z., Su, C.ꢀY., Wang, W. J. Am. Chem. Soc., 2016, 138, 3031–
3037.
(19) Cohen, M. D., Schmidt, G. M. J., Flavian, S. J. Chem. Soc.,
1964, 2041–2051.
(20) Hadjoudis, E., Vittorakis, M., Mavridis, I. M. Tetrahedron
1987, 43, 1345–1360.
(21) Amimoto, K., Kawato, T. J. Photochem. Photobiol., C, 2005,
6, 207–226.
(22) Hadjoudis, E., Mavridis, I. M. Chem. Soc. Res. 2004, 33,
579–588.
(23) Cohen, M. D., Flavian, S., Leiserowitz, L. J. Chem. Soc., B,
1967, 329–334.
(24) Ogawa, K., Kasahara, Y., Ohtani, Y., Harada, J. J. Am. Chem.
Soc., 1998, 120, 7107–7108.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(25) Harada, J., Uekusa, H., Ohashi, Y. J. Am. Chem. Soc., 1999,
121, 5809–5810.
(26) (a) Gao, Q., Bai, L., Zeng, Y., Wang, P., Zhang, X., Zou, R.,
Zhao, Y. Chem. Eur. J. 2015, 21, 16818–16822; (b) Gao, Q.,
ACS Paragon Plus Environment

Page 9 of 9
Journal of the American Chemical Society
TOC
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
9
ACS Paragon Plus Environment