Defect Engineering into Metal-Organic Frameworks for the Rapid and Sequential Installation of Functionalities

Article abstract of DOI:10.1021/acs.inorgchem.7b02391

Postsynthetic treatments are well-known and important functionalization tools of metal-organic frameworks (MOFs). Herein, we have developed a practical and rapid postsynthetic ligand exchange (PSE) strategy using a defect-controlled MOF. An increase in the number of defects amounts to MOFs with enhanced rates of ligand exchange in a shorter time frame. An almost quantitative exchange was achieved by using the most defective MOFs. This PSE strategy is a straightforward method to introduce a functionality into MOFs including bulky or catalytically relevant moieties. Furthermore, some mechanistic insights into PSE were revealed, allowing for a sequential ligand exchange and the development of multifunctional MOFs with controlled ligand ratios.

Full text of DOI:10.1021/acs.inorgchem.7b02391

Article
pubs.acs.org/IC
Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX
Defect Engineering into Metal−Organic Frameworks for the Rapid
and Sequential Installation of Functionalities
Hyojin Park,†,∥ Seongwoo Kim,†,∥ Byunghyuck Jung,‡ Myung Hwan Park,§ Youngjo Kim,†
and Min Kim*,†
†Department of Chemistry and BK21Plus Research Team, Chungbuk National University, Cheongju, 28644, Republic of Korea
‡School of Basic Science, Daegu Gyeongbuk Institute of Science and Technology (DGIST), Daegu, 42988, Republic of Korea
§Department of Chemistry Education, Chungbuk National University, Cheongju, 28644, Republic of Korea
S
* Supporting Information
ABSTRACT: Postsynthetic treatments are well-known and
important functionalization tools of metal−organic frameworks
(MOFs). Herein, we have developed a practical and rapid
postsynthetic ligand exchange (PSE) strategy using a defect-
controlled MOF. An increase in the number of defects amounts
to MOFs with enhanced rates of ligand exchange in a shorter
time frame. An almost quantitative exchange was achieved by
using the most defective MOFs. This PSE strategy is a straightforward method to introduce a functionality into MOFs including
bulky or catalytically relevant moieties. Furthermore, some mechanistic insights into PSE were revealed, allowing for a sequential
ligand exchange and the development of multifunctional MOFs with controlled ligand ratios.
There are two main types of defects that have been
hypothesized in the UiO-66 frameworks: “a vacancy of an
organic linker” and “a vacancy of an entire SBU”.16,17 Both
defects in UiO-66 were discovered and intensively studied by
multiple research groups for their applications in molecular
adsorption and catalysis.18−22 Very recently, Lillerud and co-
workers reported the quantity of defects could be controlled by
judicious choice of modulators.13 Both the amount of and the
acidity (i.e., pKa of acid) of an acid modulator are directly
correlated to the number of defects formed as evidenced by N2
1. INTRODUCTION
Metal−Organic Frameworks (MOFs, also called porous
coordination polymers, PCPs) are three-dimensional highly
porous crystalline materials which consist of metal ions (or
clusters) and multitopic organic ligands.1,2 Although the
repeating coordination bonds between the secondary building
units (SBUs) and the organic ligands produce infinite and
identical pores, and MOFs are the thermodynamic product,
structural defects are often found in the frameworks. In fact,
defects, structural disorder, and heterogeneity of crystalline or
solid-state materials are typical phenomena and have been
widely studied for their physical and chemical properties in
porous materials chemistry.3−5
Among the various MOFs, MOF-5, HKUST-1, and UiO-66
(HKUST = Hong Kong University of Science and Technology,
UiO = University of Oslo) are the most actively studied targets
for defects in the framework.3 The zinc-based MOF-5 and the
copper-based HKUST-1 are legacy MOFs and have been
known since the late 1990s.6,7 The zirconium-based UiO-66
was first reported by Lillerud and co-workers in 2008 and has
superior chemical and physical capabilities to both MOF-5 and
HKUST-1.8 Additionally, the wide range of available function-
alizations makes UiO-66 one of the most popular MOFs.9 Since
UiO-66 is relatively robust and thought of as an “inert”
material, it is hypothesized to be one of the most tolerant
MOFs for both defect generation and engineering. Recently,
Behrens and co-workers reported a practical synthesis of a UiO
series with an acid modulator (i.e., monocarboxylic acid such as
acetic acid or benzoic acid),10 and related studies have revealed
that the acid modulators promote defect generation in the
frameworks without the loss of the bulk crystallinity.11−15
1
adsorption, H NMR, PXRD (powder X-ray diffraction), and
TGA (thermogravimetric analysis). During the defect controls,
the most defective UiO-66 (from 36 equiv of trifluoroacetic
acid as the modulator) showed a 50% higher BET (Brunauer−
Emmett−Teller) surface area than the standard UiO-66 from
traditional synthesis conditions (∼1800 m2/g than ∼1200 m2/
g, respectively).
It is relatively easy to introduce not only defects but also a
variety of functional groups affecting the strength and
robustness of the UiO-66 frameworks. Various functional
groups such as NH2, NO2, naphthyl, and Br have been directly
incorporated into UiO-66 using a prefunctionalized ligand (i.e.,
direct synthesis), and amide/cyanide functionalities have been
installed through postsynthetic modification (PSM) strategies.9
Lastly, postsynthetic ligand exchange (PSE) has also been
applied to UiO-66 frameworks, and several interesting
functionalities (e.g., azide - thermally unstable in solvothermal
conditions, dihydroxy - strongly coordinates to the zirconium
Received: September 18, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.7b02391
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
precursor) have been successfully introduced into UiO-66
frameworks.23,24 However, in this last case, the exchange rate
was quite slow, rendering this particular PSE unpractical due to
the lack of efficiency. Typically, only an ∼50% exchange ratio
was obtained for a functionalized BDC (benzene-1,4-dicarbox-
ylic acid) ligand after 5 days.
Another strategy to introduce new functional groups through
defects in MOFs has been recently attempted. Farha and
Hupp’s research team has successfully installed free-carboxylic
acid groups in the copper-based NU-125 MOF (NU =
Northwestern University). The solvent-assisted linker exchange
(SALE) strategy was applied to defects in NU-125 for
introducing BTC (benzene-1,3,5-tricarboxylic acid) ligands.25
Additionally, a monoethanolamine moiety was incorporated
into a UiO-66 framework for CO2 capture using PSE by Zhang
et al.26 The defect sites in UiO-66 were generated by
monodentate acid modulators, and then new functionalities,
such as serine, were installed into those defect sites.27 Inspired
by three independent works, we investigated the correlation
between the amount of defects and postsynthetic ligand
exchange rates to install external functionalities into MOF
pores. Moreover, we hypothesized that the presence of defects
in MOFs could actually enhance the “ligand exchange strategy”
and make it a more practical method for functionalization.
2. RESULTS AND DICUSSION
Defect Controls for Postsynthetic Ligand Exchange
Rates. We chose three different defect-controlled UiO-66
samples from Lillerud’s system. No acid modulator was used for
the standard UiO-66 synthesis (UiO-66(NA), BET surface area
= ∼1200 m2/g), 6 equiv of formic acid was used for UiO-
66(FA) synthesis (BET surface area = ∼1400 m2/g), and, last,
36 equiv of trifluoroacetic acid was used for UiO-66(TFA)
synthesis (BET surface area = ∼1800 m2/g). These three
samples were prepared by following reported procedures
including activating and washing steps.13 As a model reaction,
BDC-NH2 and BDC-Br were used for PSE using these defect-
controlled UiO-66 samples. The exchange was performed in
aqueous conditions at room temperature, and the concen-
tration of the ligand was fixed at 0.1 M for direct comparison
with previous studies.23 Surprisingly, the ligand exchange rates
dramatically increased from the standard UiO-66(NA) and
previous nondefective UiO-66 (i.e., synthesized UiO-66 sample
without using an acid modulator and water).
The amount exchanged (at 3, 6, 12, and 24 h) was
determined by 1H NMR analysis after acid digestion using fully
washed and activated samples (with water and methanol for 3
days; see the Supporting Information for details). While 51
2% of BDC-NH2 was incorporated into UiO-66(NA), a total of
Figure 1. Ligand exchange rate and saturation level of BDC-NH2 and
BDC-Br into defect-controlled UiO-66s.
74
1% of BDC-NH2 and 76
1% of BDC-NH2 were
We believe that there is a chance for both ligand insertion
into defect sites and true ligand exchange. However, the
amount of exchange (76%) is much larger than the amount of
available defects. In the previous study, it was reported that
only 11% of the ligand was changed to the defect site in UiO-
66(NA), 15% for UiO-66(FA), and 33% for UiO-66(TFA),
respectively.13 To confirm the relative contribution of ligand
insertion into the defect site, and overall ligand exchange, first,
we compared the existing amount of both ligands (non-
functionalized BDC and functionalized BDC) in the solid state
(i.e., exchanged UiO-66) and in the aqueous media at the early
stage of exchange before saturation. At 3 h, BDC-NH2 was 60%
incorporated into UiO-66(TFA) frameworks, and the
exchanged into UiO-66(FA) and UiO-66(TFA) in 24 h,
respectively (average numbers from two independent samples,
Figures 1 and S1−S6). Once again, in the previous studies,
pristine UiO-66 showed ∼60% exchange after 5 days.23
Additionally, we have investigated anion effects on PSE. Since
the hydrolyzed dicarboxylate was neutralized with aqueous HCl
solution, the exchanging solution has chloride ions. By
modifying the acidification agent from HCl (aq.) to H2SO4
(aq.), the existing anion is changed to a sulfate ion. In this
sulfate condition, 74% of BDC-NH2 was incorporated into
UiO-66(TFA) in 24 h, which is about the same amount as with
the chloride ions (Figure S7).
B
DOI: 10.1021/acs.inorgchem.7b02391
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 1. Initial Ligand Exchange Rate (at 3 h) for BDC-NH2 and BDC-Br for Defect-Controlled UiO-66s
a
exchanged ligand
UiO-66 (%/h)
UiO-66(NA) (%/h)
UiO-66(FA) (%/h)
UiO-66(TFA) (%/h)
BDC-NH2
BDC-Br
3.3
2.5
10
22
10
22
12
8.3
a
Synthesized without acid modulator and water. Data were obtained from ref 23.
BDC:BDC-NH2 ratio was determined to be 40:60 in the solid
state (Figure S8). Interestingly, a similar 40:60 ratio for
BDC:BDC-NH2 was observed in the remaining aqueous
solution after exchange. This suggests that the amount of
BDC exchanged out is ∼20% less than the exchanged in BDC-
NH2. We believe that this difference stems from direct ligand
insertion into the defect site without removing ligands. We have
taken this a step further and utilized internal standards for our
NMR studies to determine the definitive amount of defects,
addition of ligands, and exchange. Before PSE, the chemical
formula of UiO-66(TFA) is determined to be
Zr6O4(OH)4(BDC)4.17. Since the chemical formula of defect-
free UiO-66 is Zr6O4(OH)4(BDC)6, approximately 30% of the
BDC ligand thought to be missing (i.e., the amount of defect)
was confirmed, and this number is corroborated with a
previously reported analysis by Lillerud using TGA.13 After
PSE with NH2-BDC for 6 h, a total of 76% of the ligand was
exchanged and the formula of the exchanged sample was
determined to be Zr6O4(OH)4(BDC)1.59(BDC-NH2)3.10 (Fig-
ure S9). This means that, after PSE, the exchanged UiO-
66(TFA)-NH2 sample has approximately 20% defect sites. In
other words, around 1/3 of the defect sites had an inserted
ligand. Although both defect calculations, from NMR-internal
standard and TGA, are not highly accurate,28 this data strongly
indicates that both ligand exchange and insertion processes
exist in the defect-induced PSE process. This phenomenon was
also observed in the case of UiO-66(FA) (Figures S9 and S10,
and Table S1).
incorporated and the acidic phenol derivatives, such as BDC-
OH and BDC-(OH)2, were not selected in this study. EDGs
(NH2, OMe, and (OMe)2) generally showed faster exchange
rates and higher exchange amounts than electron-withdrawing
groups (EWGs; F, Br, and NO2), which are displayed in Figure
2. Among the EDGs, the amino group showed the highest
Figure 2. Ligand exchange rates and saturation level of functionalized
ligands to UiO-66(TFA).
The incorporated ligand ratios were at least 2 times larger
than the amount of defects at the early stage (<3 h), and we
assumed that both UiO-66(FA) and UiO-66(TFA) displayed
the maximum, saturated exchange amount for BDC-NH2
(∼70%/1 d). Figure 1 clearly displays that there are limits on
the BDC-NH2 exchange amount when using 1 equiv of
additional ligand. Both UiO-66(FA) and UiO-66(TFA)
reached the exchange saturation point within 5 h. In the case
of BDC-Br, it generally showed lower exchange rates and
amounts than BDC-NH2, which is the same trend as was first
reported.23 As expected, UiO-66(NA) showed the lowest
exchange amount at 30 1% after 24 h, while UiO-66(FA) and
exchange amount (76 1% exchanged, Figure S5), methoxy
was in the middle (50 1% exchanged, Figures S18 and S19),
and the dimethoxy group demonstrated the lowest amount (43
1% exchanged, Figures S20 and S21) for PSE into UiO-
66(TFA). This indicates that the nucleophilicity of the
carboxylate in the BDC ligand may play a key role in the
PSE process (comparing rates between NH2 > OMe), and the
size of the functional group also contributed to the PSE rates
(comparing methoxy > dimethoxy). In the opposite case, three
EWGs (F, Br, and NO2) showed slower PSE rates and lower
exchange amount than the EDGs, and generally, these three
examples show no significant differences during PSE (F-43
UiO-66(TFA) showed 35
1% and 39
1%, respectively
1%, Br-39
1%, and NO2-33
1% exchanged; average
(average numbers from two independent samples, Figures
S11−S16). Before the exchange saturation (at ∼3 h), the
exchange rates were compared and followed exactly the same
trends with the maximum exchange amount (Table 1). In all
cases, the UiO-66 framework is completely intact after ligand
exchange as evidenced by PXRD patterns (Figure S17).
Functional Group Effects on Saturation Levels of
Postsynthetic Ligand Exchange. To determine which
factors influence the exchange rate and amount, the functional
group effects were also investigated using a variety of ligands.
Electron donating groups (EDG)-containing BDCs, BDC-
OMe, and BDC-(OMe)2 were prepared and utilized to probe
both the electronic effects and the steric issues using UiO-
66(TFA). Since a basic condition (aq. KOH) was employed for
the hydrolysis step during PSE, the alkoxy group was
numbers from two independent samples, Figures 2, S11−S16,
and S22−S25). In both EDG and EWG ligand exchange cases,
the structure of UiO-66 was retained, which is evidenced by
PXRD (Figures S17 and S26).
The first step of PSE is the attack of the dicarboxylate toward
the zirconium clusters. In the case of UiO-66, a ligand
dissociation first (i.e., without the nucleophilic attack of the
dicarboxylate) will not likely occur since UiO-66 is a “relatively
robust” and “chemically stable” MOF. Then, the observation
here is very reasonable; the EDG makes the BDC a better
nucleophile toward the metal cluster for initiating the exchange
process. This suggests that an electronic factor of substituent is
more prominent than the size factor, since size-different BDC-F
and BDC-Br showed similar exchange rates and amounts in the
defect-controlled PSE process (Figure 2).
C
DOI: 10.1021/acs.inorgchem.7b02391
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 2. Exchanged Ratio for Both Functionalized Ligand Exchanged “IN” and “OUT” (at 6 h)
entry
R
ratio of exchanged “IN” (A) (%)
ratio of exchanged “OUT” (B) (%)
1
2
3
4
5
NH2
OMe
(OMe)2
Br
76
50
43
39
33
1
1
1
1
1
33
49
58
60
64
1
1
1
1
1
NO2
Effects of Functional Groups on the Parent MOFs.
From another perspective, to confirm the nucleophilicity of
ligands, and the effects of coordination bond strength on the
levels of PSE, we have examined the functional group effect on
the parent MOFs with regards to PSE (Table 2). The
functional group effect on the strut molecule was monitored
when nonfunctionalized BDC was incorporated into the
prefunctionalized UiO-66-R framework via PSE. The function-
alized UiO-66-NH2(as), UiO-66-OMe(as), UiO-66-
(OMe)2(as), UiO-66-Br(as), and UiO-66-NO2(as) (as = as
synthesized) were synthesized by solvothermal conditions with
HCl as a modulator since TFA as a modulator is not ideal for
functionalized UiO-66 synthesis. All functionalized UiO-66-R
displayed identical frameworks evidenced by PXRD (Figure
S27).
“exchanged IN” is much higher for BDC-NH2 versus the
“exchanged OUT” which is much lower in comparison to more
electron deficient ligands, BDC-Br or BDC-NO2. The structure
of the UiO-66 framework was completely retained in both the
exchanged “IN” and “OUT” for all cases, which was evidenced
by PXRD (Figure S27).
Full Exchange and Various Functionalization through
Postsynthetic Ligand Exchanges. Knowing that the
existence of defects accelerates the PSE process, full ligand
exchange was further examined by altering the number of
exchanges and the exchange time (Figure 3). Both soaking with
3 equiv of BDC-NH2 (for 24 h) or two times soaking 1 equiv of
BDC-NH2 (for 3 h + 9 h) for UiO-66(TFA) was attempted at
room temperature, and both protocols produced ∼95%
exchanged UiO-66 frameworks from the nonfunctionalized
From the 6 h incubation at room temperature, the exchanged
amounts generally showed a reverse trend to the previous
exchange “IN” study (Table 2). The most nucleophilic BDC-
NH2 showed lower levels of exchange out of the framework by
additional nonfunctionalized BDC (entry 1(B), Table 2, Figure
S28). At the same time, the less nucleophilic BDC-Br and
BDC-NO2, which had shown the lowest exchanged amount in
the previous functionalization study, showed the highest
amount of exchange, 60% and 64%, respectively (entries 4(B)
and 5(B) in Table 2, Figures S29 and S30). Both methoxy and
dimethoxy substituted ligands showed similar ratio for both
exchanged in and exchanged out experiments (Figures S31 and
S32). Most interestingly, similar ligand ratios were obtained
from either exchanged “IN” or exchanged “OUT” experiments.
For example, installation of BDC-Br into a nonfunctionalized
UiO-66 by an exchanged “IN” experiment provides BDC: 60%
and BDC-Br: 40% incorporated into the UiO-66 framework. In
the opposite case, installation of BDC into the prefunctional-
ized UiO-66-Br by being exchanged “OUT” gives a similar
BDC: 60% and BDC-Br: 40% incorporation. This suggests that
the exchange ratio is determined by the nature of ligands and an
equilibration process with the MOF framework, and this finding
clearly indicates that the coordination bond strength between the
SBU and ligand plays a key role in the PSE process. These kinds
of substituent effects on the metal-carboxylate complexes are
rarely reported in the literature.29 The electron-rich benzene
ring (i.e., EDG-substituted ligand such as BDC-NH2, BDC-
OMe, and BDC-(OMe)2) provides more electron density into
the carboxylate, and these ligands are strongly coordinated to
the zirconium ion on the SBU. Thus, the exchanged amount of
Figure 3. Full ligand exchange and biologically relevant ligand
exchanges on defective UiO-66.
D
DOI: 10.1021/acs.inorgchem.7b02391
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 4. A sequential two-step PSE using BDC-OMe and BDC-Br from nonfunctionalized UiO-66(TFA).
UiO-66(TFA) (Figures S33 and S34). The ∼95% exchanged
respectively, Figures S41 and S42), and once again, the
structure of UiO-66 was completely retained after the BDC-
AM-Pyrenyl and amino acid-BDC exchange process (Figure
S43).
UiO-66(TFA)-NH2 generally displayed similar characteristics
1
to the directly synthesized UiO-66(TFA)-NH2 except for H
NMR after acid digestion. The PXRD and TGA data are
identical for both exchanged and synthesized samples along
with previously reported patterns (Figures S35 and S36).30 The
BET surface area for the exchanged UiO-66(TFA)-NH2 was
1452 m2/g versus the synthesized UiO-66(TFA)-NH2 which
was 1217 m2/g, and the N2 and CO2 isotherms followed typical
UiO-66-NH2 patterns (Figures S37 and S38). However, several
Sequential Ligand Exchange and Core/Shell Type
MOFs. The early saturation of the BDC-AM-Pyrenyl exchange
in conjunction with the lower exchange amount of BDC-AM-
Pro over BDC-AM-Ala strongly suggests that the PSE process
is still controlled by the size of the substituent when it is larger
than the available diffusion channel in MOF. It also implies that
the PSE process starts at the surface of MOF and does not
proceed through the whole particle at the same time.34 To
verify this PSE mechanism, we have designed a sequential two-
step PSE (Figure 4). Nonfunctionalized UiO-66(TFA) was first
exchanged with BDC-OMe and a 56% exchanged sample was
obtained. If the PSE process occurred from the surface inward,
the core of the particle would have 44% nonfunctionalized
BDC remaining. Starting from this exchanged MOF, a second
PSE using BDC-Br was performed. As a result, 28% of BDC-Br
was incorporated into UiO-66 by the second PSE. Surprisingly,
only 28% of the previously installed BDC-OMe was exchanged
out and none of the BDC had exchanged with BDC-Br
(Figures 4, S44 and S45). We believe that this is a strong
evidence for the directionality of PSE, beginning from the
accessible surface and working toward the core of the MOF.
And this is a very interesting three-layered core−shell structure
of MOFs with three different ligands. Previously, the growth of
the second MOF on the first seed MOF for constructing core−
shell or hetero-compositional MOFs has been studied with
several specific examples, and our finding supports that the
sequential PSE of MOFs is a practical way for preparation of a
core−shell structure on MOFs.33−40
Additionally, other mechanistic details were revealed during
time-controlled sequential PSE experiments. By increasing the
2nd PSE exposure time to 24 h, we observed a decrease of the
core part and a leaching out of nonfunctionalized BDC totaling
9%. In other words, the exchanged “free” BDC-OMe by BDC-
Br on the pore near the surface had penetrated into the core
during the elongated reaction time and some exchange between
BDC-OMe and the nonfunctionalized BDC had occurred
(Figures 5, S46, and S47). The PXRD pattern was totally intact
even after the 24 h/24 h sequential PSE (Figure S48). To
summarize, these two results strongly support that the PSE
process is not a dissolution/recrystallization mechanism and
follows a diffusion/replacement mechanism throughout the
MOF particle since the PSE ratio was affected by the PSE
sequence.31,32
1
impurity peaks were observed during the H NMR analysis
after acid digestion. Since the exchanged UiO-66(TFA)-NH2
showed only BDC, BDC-NH2 peaks and remaining solvent
peaks after the same activation treatment, the impurities in the
synthesized samples must be caused from the solvothermal
conditions (Figure S34). In the case of sensitive functional
group-containing ligands, the strong acid modulator (e.g., TFA)
condition is not a good method for preparing functionalized
MOFs and may have degraded the ligands as we observed.
Additionally, the increased surface area of UiO-66(TFA)-NH2
corroborates the existence of remaining defects in the
exchanged UiO-66 framework. The PSE process is not a
dissolution/recrystallization mechanism which is evidenced by
PXRD, TGA, ex situ 1H NMR, solid-state NMR, and
computational studies by several groups.24,31,32 This fully
exchanged result lends weight to the importance of MOF
dynamics and the speed of the ongoing exchange equilibrium.
With this fast and convenient ligand installation strategy in
our hands, as a proof-of-concept study, we thought to install a
bulky ligand that is typically unable to be incorporated into
UiO-66 frameworks through a direct synthesis. This pyrene-
based ligand was prepared through an amide coupling reaction
as a model substrate and applied to the PSE. Interestingly, 16%
ligand exchange was monitored with a retained UiO-66
framework, and the exchange ratio was not increased by either
additional exposure or exchange time (16% for 72 h, Figures
S39 and S40). This result suggests that only surface exchange
occurred; the sterically bulky pyrenyl ligand could only fit on
the surface of the UiO-66 crystal and did not penetrate to the
inside of the pores for further PSE. As the exchanged pyrenyl
ligands potentially blocked the diffusion channel for other
ligands, or the side chain made the BDC ligand simply too large
to fit through the pores, PSE was limited to only the exterior
part of the UiO-66 crystal. Lastly, the more biologically relevant
amino acid-containing BDCs installation was performed
utilizing a defective UiO-66. BDC-AM-Ala and BDC-AM-Pro
were prepared and applied to PSE functionalization. Both
amino acid derivatives displayed moderate exchange amounts
(49 1% and 34 1% for BDC-AM-Ala and BDC-AM-Pro
This sequential PSE has been developed as a practical
method for preparing multifunctional MOFs. From a previous
E
DOI: 10.1021/acs.inorgchem.7b02391
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
4. CONCLUSIONS
In conclusion, we have developed a practical and rapid
postsynthetic ligand exchange strategy using a modulator-
induced, defect-controlled UiO-66 system. Increase of the
defect amounts in the MOF pore enhances the ligand exchange
rate and ratio in a shorter time frame. In particular, using either
3 equiv of functionalized ligands, or 2 times a single equivalent,
for the exposure, solution produces an almost fully exchanged
ratio for the most defective UiO-66(TFA) in 12 h at room
temperature (previous report was ∼60% exchange for 5 days).
This strategy is also a straightforward method to introduce
bulky or catalytically useful functionalities into MOFs. The
pyrenyl-containing BDC was successfully incorporated into the
UiO-66 framework through PSE only on the surface, while the
amino acid groups were successfully installed into the UiO-66
pores. By examining a variety of functionalities for PSE, we have
identified two important mechanistic parameters: the electronic
density of the benzene ring is highly correlated with the
exchange rate, and the size of the pendant substituent is also
essential. Leveraging the surface first exchange, a sequential
PSE was developed to prepare multifunctional MOFs with a
subtler gradient of functionalities in comparison to a core−shell
MOF. This synthetic strategy offers an exciting opportunity to
the preparation of diverse water and chemically stable MOFs
with a wide range of functionalities which are desirable for
various applications.
Figure 5. Time-controlled sequential PSE for diffusion mechanism
confirmation: (a) short time (3 h) for 2nd PSE and (b) long time (24
h) for the 2nd PSE.
ASSOCIATED CONTENT
■
study, a mixed ligand strategy allows a maximum of eight
different functionalities into a single MOF pore with a random
distribution via direct synthesis.41 In contrast; this sequential
PSE could be also used to prepare multiple functionalities
within the MOF pores but potentially with a gradient of
functional groups from the inside out.
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.inorg-
chem.7b02391.
Detailed procedures for ligand synthesis, figures, and a
table as described in the text and H and 13C NMR and
1
3. EXPERIMENTAL SECTION
Preparation of Defect-Controlled UiO-66s. The UiO-66 series
was prepared and activated using a modified method from what has
been previously described.13
IR spectra (PDF)
AUTHOR INFORMATION
■
UiO-66(TFA). Terephthalic acid (123 mg, 0.74 mmol), ZrCl4 (172
mg, 0.74 mmol), DMF (20 mL), H2O (0.040 mL, 2.22 mmol), and
trifluoroacetic acid (2.04 mL, 26.6 mmol) were placed in a Teflon
lined autoclave and heated at 120 °C for 72 h. The microcrystalline
powders were isolated by centrifugation. Residual DMF and ligand
precursors were removed from the material by washing with 12 mL of
DMF three times. Then they were incubated with fresh DMF (12 mL)
for overnight at 70 °C. Three further such washes were performed.
Then the solids were washed with 3 × 12 mL of MeOH, after which
the solids were left to soak in 12 mL of MeOH for 1 day. The solids
were centrifuged and dried under vacuum.
General Procedures of Postsynthetic Ligand Exchange
(PSE). Functionalized ligand (0.1 mmol, 0.1 M) was dissolved in
THF (1 mL) and aqueous 4% KOH solution (1 mL). The solution
was stirred at room temperature overnight. The THF was removed by
evaporation, and the remaining aqueous phase was neutralized to pH 7
with 1 M HCl solution. This solution was added to UiO-66 (ca. 29 mg,
0.1 mmol of BDC), and the mixture was incubated at room
temperature. After finishing the exchange, the mixture was centrifuged
and the aqueous phase decanted. The solids were suspended in 5 mL
of water, and solid was centrifuged and the water was decanted. This
water washing was repeated for two times. After water washing, the
solids were washed with 5 mL of MeOH (two times). Then the solids
were suspended in 5 mL of MeOH. After 24 h, the solid was
centrifuged and MeOH was decanted. This MeOH soaking step was
repeated two more times (for 24 h each). The solid phase was finally
dried under vacuum for 6 h.
Corresponding Author
*E-mail: minkim@chungbuk.ac.kr.
ORCID
Myung Hwan Park: 0000-0002-8858-5204
Youngjo Kim: 0000-0001-8571-0623
Min Kim: 0000-0002-1710-2682
Author Contributions
∥The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript. These authors contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Dr. Corinne A. Allen (AAAS Science & Technology
Policy Fellow at the Department of Energy, USA) for helpful
discussions and in preparing this manuscript. This research was
s u p p o r t e d b y t h e S c i e n c e R e s e a r c h C e n t e r
(2016R1A5A1009405) and Basic Science Research Program
(2016R1C1B1011076) through the National Research Foun-
dation of Korea (NRF) funded by the Ministry of Science, ICT
& Future Planning.
F
DOI: 10.1021/acs.inorgchem.7b02391
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(20) Huang, L.; He, M.; Chen, B.; Hu, B. A mercapto functionalized
magnetic Zr-MOF by solvent-assisted ligand exchange for Hg2+
removal from water. J. Mater. Chem. A 2016, 4, 5159−5166.
(21) Liang, W.; Coghlan, C. J.; Ragon, F.; Rubio-Martinez, M.;
D’Alessandro, D. M.; Babarao, R. Defect engineering of UiO-66 for
CO2 and H2O uptake − a combined experimental and simulation
study. Dalton Trans. 2016, 45, 4496−4500.
(22) Pereira, C. F.; Howarth, A. J.; Vermeulen, N. A.; Paz, F. A. A.;
Tome, J. P. C.; Hupp, J. T.; Farha, O. K. Towards hydroxamic acid
linked zirconium metal−organic frameworks. Mater. Chem. Front.
2017, 1, 1194−1199.
(23) Kim, M.; Cahill, J. F.; Su, Y.; Prather, K. A.; Cohen, S. M.
Postsynthetic ligand exchange as a route to functionalization of ‘inert’
metal−organic frameworks. Chem. Sci. 2012, 3, 126−130.
(24) Kim, M.; Cahill, J. F.; Fei, H.; Prather, K. A.; Cohen, S. M.
Postsynthetic Ligand and Cation Exchange in Robust Metal−Organic
Frameworks. J. Am. Chem. Soc. 2012, 134, 18082−18088.
(25) Karagiaridi, O.; Vermeulen, N. A.; Klet, R. C.; Wang, T. C.;
Moghadam, P. Z.; Al-Juaid, S. S.; Stoddart, J. F.; Hupp, J. T.; Farha, O.
K. Functionalized Defects through Solvent-Assisted Linker Exchange:
Synthesis, Characterization, and Partial Postsynthesis Elaboration of a
Metal−Organic Framework Containing Free Carboxylic Acid
Moieties. Inorg. Chem. 2015, 54, 1785−1790.
(26) Li, L.-J.; Liao, P.-Q.; He, C.-T.; Wei, Y.-S.; Zhou, H.-L.; Lin, J.-
M.; Li, X.-Y.; Zhang, J.-P. Grafting alkylamine in UiO-66 by charge-
assisted coordination bonds for carbon dioxide capture from high-
humidity flue gas. J. Mater. Chem. A 2015, 3, 21849−21855.
(27) Shearer, G. C.; Vitillo, J. G.; Bordiga, S.; Svelle, S.; Olsbye, U.;
Lillerud, K. P. Functionalizing the Defects: Postsynthetic Ligand
Exchange in the Metal Organic Framework UiO-66. Chem. Mater.
2016, 28, 7190−7193.
(28) Epley, C. C.; Love, M. D.; Morris, A. J. Characterizing Defects in
a UiO-AZB Metal−Organic Framework. Inorg. Chem. 2017, 56,
13777−13784.
(29) Sun, Y.-J.; Huang, Q.-Q.; Zhang, J.-J. Series of Structural and
Functional Models for the ES (Enzyme−Substrate) Complex of the
Co(II)-Containing Quercetin 2,3-Dioxygenase. Inorg. Chem. 2014, 53,
2932−2942.
(30) Garibay, S. J.; Cohen, S. M. Isoreticular synthesis and
modification of frameworks with the UiO-66 topology. Chem.
Commun. 2010, 46, 7700−7702.
(31) Gross, A. F.; Sherman, E.; Mahoney, S. L.; Vajo, J. J. Reversible
Ligand Exchange in a Metal−Organic Framework (MOF): Toward
MOF-Based Dynamic Combinatorial Chemical Systems. J. Phys. Chem.
A 2013, 117, 3771−3776.
(32) Taddei, M.; Tiana, D.; Casati, N.; van Bokhoven, J. A.; Smit, B.;
Ranocchiari, M. Mixed-linker UiO-66: structure−property relation-
ships revealed by a combination of high-resolution powder X-ray
diffraction and density functional theory calculations. Phys. Chem.
Chem. Phys. 2017, 19, 1551−1559.
(33) Choi, S.; Kim, T.; Ji, H.; Lee, H. J.; Oh, M. Isotropic and
Anisotropic Growth of Metal−Organic Framework (MOF) on MOF:
Logical Inference on MOF Structure Based on Growth Behavior and
Morphological Feature. J. Am. Chem. Soc. 2016, 138, 14434−14440.
(34) During the review process, a similar concept of core−shell type
MOF preparation of ligand PSE has been published: Boissonnault, J.
A.; Wong-Foy, A. G.; Matzger, A. J. Core−Shell Structures Arise
Naturally During Ligand Exchange in Metal−Organic Frameworks. J.
Am. Chem. Soc. 2017, 139, 14841−14844.
(35) Koh, K.; Wong-Foy, A. G.; Matzger, A. J. MOF@MOF:
microporous core−shell architectures. Chem. Commun. 2009, 6162−
6164.
(36) Yoo, Y.; Jeong, H.-K. Heteroepitaxial Growth of Isoreticular
Metal−Organic Frameworks and Their Hybrid Films. Cryst. Growth
Des. 2010, 10, 1283−1288.
(37) Li, T.; Sullivan, J. E.; Rosi, N. L. Design and Preparation of a
Core−Shell Metal−Organic Framework for Selective CO2 Capture. J.
Am. Chem. Soc. 2013, 135, 9984−9987.
REFERENCES
■
(1) Zhou, H.-C.; Long, J. R.; Yaghi, O. M. Introduction to Metal-
Organic Frameworks. Chem. Rev. 2012, 112, 673−674.
(2) Ockwig, N. W.; Delgado-Friedrichs, O.; O’Keeffe, M.; Yaghi, O.
M. Reticular Chemistry: Occurrence and Taxonomy of Nets and
Grammar for the Design of Frameworks. Acc. Chem. Res. 2005, 38,
176−182.
(3) Fang, Z.; Bueken, B.; De Vos, D. E.; Fischer, R. A. Defect-
Engineered Metal−Organic Frameworks. Angew. Chem., Int. Ed. 2015,
54, 7234−7254.
(4) Cheetham, A. K.; Bennett, T. D.; Coudert, F.-X.; Goodwin, A. L.
Defects and disorder in metal organic frameworks. Dalton Trans. 2016,
45, 4113−4126.
(5) Bennett, T. D.; Cheetham, A. K.; Fuchs, A. H.; Coudert, F.-X.
Interplay between defects, disorder and flexibility in metal-organic
frameworks. Nat. Chem. 2017, 9, 11−16.
(6) Li, H.; Eddaoudi, M.; O’Keeffe, M.; Yaghi, O. M. Design and
synthesis of an exceptionally stable and highly porous metal-organic
framework. Nature 1999, 402, 276−279.
(7) Chui, S. S.-Y.; Lo, S. M.-F.; Charmant, J. P. H.; Orpen, A. G.;
Williams, I. D. A Chemically Functionalized Nanoporous Material
[Cu3(TMA)2(H2O)3]n. Science 1999, 283, 1148−1150.
(8) Cavka, J. H.; Jakobsen, S.; Olsbye, U.; Guillou, N.; Lamberti, C.;
Bordiga, S.; Lillerud, K. P. A New Zirconium Inorganic Building Brick
Forming Metal Organic Frameworks with Exceptional Stability. J. Am.
Chem. Soc. 2008, 130, 13850−13851.
(9) Kim, M.; Cohen, S. M. Discovery, development, and
functionalization of Zr(IV)-based metal−organic frameworks. Crys-
tEngComm 2012, 14, 4096−4104.
(10) Schaate, A.; Roy, P.; Godt, A.; Lippke, J.; Waltz, F.; Wiebcke,
M.; Behrens, P. Modulated Synthesis of Zr-Based Metal−Organic
Frameworks: From Nano to Single Crystals. Chem. - Eur. J. 2011, 17,
6643−6651.
(11) Gutov, O. V.; Hevia, M. G.; Escudero-Adan, E. C.; Shafir, A.
Metal−Organic Framework (MOF) Defects under Control: Insights
into the Missing Linker Sites and Their Implication in the Reactivity of
Zirconium-Based Frameworks. Inorg. Chem. 2015, 54, 8396−8400.
(12) Hu, Z.; Castano, I.; Wang, S.; Wang, Y.; Peng, Y.; Qian, Y.; Chi,
C.; Wang, X.; Zhao, D. Modulator Effects on the Water-Based
Synthesis of Zr/Hf Metal−Organic Frameworks: Quantitative
Relationship Studies between Modulator, Synthetic Condition, and
Performance. Cryst. Growth Des. 2016, 16, 2295−2301.
(13) Shearer, G. C.; Chavan, S.; Bordiga, S.; Svelle, S.; Olsbye, U.;
Lillerud, K. P. Defect Engineering: Tuning the Porosity and
Composition of the Metal−Organic Framework UiO-66 via
Modulated Synthesis. Chem. Mater. 2016, 28, 3749−3761.
(14) Atzori, C.; Shearer, G. C.; Maschio, L.; Civalleri, B.; Bonino, F.;
Lamberti, C.; Svelle, S.; Lillerud, K. P.; Bordiga, S. Effect of Benzoic
Acid as a Modulator in the Structure of UiO-66: An Experimental and
Computational Study. J. Phys. Chem. C 2017, 121, 9312−9324.
(15) Cai, G.; Jiang, H.-L. A Modulator-Induced Defect-Formation
Strategy to Hierarchically Porous Metal−Organic Frameworks with
High Stability. Angew. Chem. 2017, 129, 578−582.
(16) Cliffe, M. J.; Wan, W.; Zou, X.; Chater, P. A.; Kleppe, A. K.;
Tucker, M. G.; Wilhelm, H.; Funnell, N. P.; Coudert, F.-X.; Goodwin,
A. L. Correlated defect nanoregions in a metal−organic framework.
Nat. Commun. 2014, 5, 4176.
(17) Trickett, C. A.; Gagnon, K. J.; Lee, S.; Gandara, F.; Burgi, H.-B.;
Yaghi, O. M. Definitive Molecular Level Characterization of Defects in
UiO-66 Crystals. Angew. Chem., Int. Ed. 2015, 54, 11162−11167.
(18) Wu, H.; Chua, Y. S.; Krungleviciute, V.; Tyagi, M.; Chen, P.;
Yildirim, T.; Zhou, W. Unusual and Highly Tunable Missing-Linker
Defects in Zirconium Metal−Organic Framework UiO-66 and Their
Important Effects on Gas Adsorption. J. Am. Chem. Soc. 2013, 135,
10525−10532.
(19) Hu, Z.; Faucher, S.; Zhuo, Y.; Sun, Y.; Wang, S.; Zhao, D.
Combination of Optimization and Metalated-Ligand Exchange: An
Effective Approach to Functionalize UiO-66(Zr) MOFs for CO2
Separation. Chem. - Eur. J. 2015, 21, 17246−17255.
G
DOI: 10.1021/acs.inorgchem.7b02391
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(38) Hirai, K.; Furukawa, S.; Kondo, M.; Uehara, H.; Sakata, O.;
Kitagawa, S. Sequential Functionalization of Porous Coordination
Polymer Crystals. Angew. Chem., Int. Ed. 2011, 50, 8057−8061.
(39) Szilagyi, P. A.; Lutz, M.; Gascon, J.; Juan-Alcaniz, J.; van Esch, J.;
Kapteijn, F.; Geerlings, H.; Dam, B.; van de Krol, R. MOF@MOF
core−shell vs. Janus particles and the effect of strain: potential for
guest sorption, separation and sequestration. CrystEngComm 2013, 15,
6003−6008.
(40) Hirai, K.; Furukawa, S.; Kondo, M.; Meilikhov, M.; Sakata, Y.;
Sakata, O.; Kitagawa, S. Targeted functionalisation of a hierarchically-
structured porous coordination polymer crystal enhances its entire
function. Chem. Commun. 2012, 48, 6472−6474.
(41) Deng, H.; Doonan, C. J.; Furukawa, H.; Ferreira, R. B.; Towne,
J.; Knobler, C. B.; Wang, B.; Yaghi, O. M. Multiple Functional Groups
of Varying Ratios in Metal-Organic Frameworks. Science 2010, 327,
846−850.
H
DOI: 10.1021/acs.inorgchem.7b02391
Inorg. Chem. XXXX, XXX, XXX−XXX