Article

Cite This: J. Am. Chem. Soc. 2018, 140, 1812−1823 pubs.acs.org/JACS

Crystal Growth of ZIF-8, ZIF-67, and Their Mixed-Metal Derivatives

Daniel Saliba, Manal Ammar, Moustafa Rammal, Mazen Al-Ghoul,* and Mohamad Hmadeh*
Department of Chemistry, American University of Beirut, P.O. Box 11-0236, Riad El-Solh, 1107 2020 Beirut, Lebanon
*
S Supporting Information

ABSTRACT: A facile method to produce zeolitic imidazolate

frameworks (ZIF-8, ZIF-67, and solid−solution ZIFs (mixed
Co and Zn)) is reported. ZIF crystals are produced via a
reaction−diffusion framework (RDF) by diffusing an outer
solution at a relatively high concentration of the 2-methyl
imidazole linker (HmIm) into an agar gel matrix containing
the metal ions (zinc(II) and/or cobalt(II)) at room temper-
Downloaded by UNIV FED DA PARAIBA at 05:28:38:689 on June 06, 2019

ature. Accordingly, a propagating supersaturation wave,

initiated at the interface between the outer solution and the
gel matrix, leads to a precipitation front with a gradient of
crystal sizes ranging between 100 nm and 55 μm along the
from https://pubs.acs.org/doi/10.1021/jacs.7b11589.

reaction tube. While the precipitation fronts of ZIF-8 and ZIF-67 travel the same distance for the same initial conditions, ZIF-8
crystals therein are consistently smaller than the ZIF-67 crystals due to the disparity of their rate of nucleation and growth. The
effects of the temperature, the concentration of the reagents, and the thickness of the gel matrix on the growth of the ZIF crystals
are investigated. We also show that by using RDF we can envisage the formation mechanism of the ZIF crystals, which consists of
the aggregation of ZIF nanospheres to form the ZIF-8 dodecahedrons. Moreover, using RDF, the formation of a solid−solution
ZIF via the incorporation of Co(II) and Zn(II) cations within the same framework is achieved in a controlled manner. Finally, we
demonstrate that doping ZIF-8 by Co(II) enhances the photodegradation of methylene blue dye under visible light irradiation in
the absence of hydrogen peroxide.

■ INTRODUCTION
Zeolitic imidazolate frameworks (ZIFs) are porous crystalline
materials that consist of tetrahedral clusters of MN4 (e.g., M =
Zn(II), Co(II)) linked by simple imidazolate ligands.1−3 ZIFs
constitute a subfamily of metal−organic frameworks (MOFs)
and are of great interest for their highly desirable physical and
chemical properties such as crystallinity, porosity, rich structural
diversity, and exceptional chemical and thermal stability.4,5
Because of these combined features, ZIFs show significant
potential for gas storage and separation applications.6,7 As in
the case of zeolites, the properties and performance of porous
materials depend on the well-defined size and shape of the
nanocrystals and microcrystals.8 Therefore, it is obvious that
the new synthetic routes have benefited from the profound
understanding of the physicochemical fundamentals of the
crystallization processes.9 ZIFs have been synthesized solvo-
thermally or at room temperature using organic solvents (e.g.,
methanol, DMF) or pure water.10,11 The control of size and
morphology of crystals has been achieved using reverse
microemulsion methods,12,13 microwave,14 ultrasound-assisted
syntheses,15 and coordination modulation methods.16 In
particular, ZIF-8 which possesses a sodalite-type structure is
the most investigated ZIF material for a variety of applications
including catalysis,17,18 gas sorption,19 and sensing.20,21 One of
the interesting features of this structure is that Zn(II) cations
can be substituted by Co(II) or Cu(II) without altering its
topology.22,23 This can be achieved by postsynthetic treatment
or via a one-pot synthesis. The latter approach is more
challenging because of the complexity of adding a second metal
that can lead to the formation of a physical mixture of discrete
structures.24 Very recently, Co-doped ZIF-8 has been used to
tweak the size and the physicochemical properties of the
obtained heterometallic structure.25 More importantly, hetero-
metallic ZIF-8 material exhibits enhanced CO2 and H2 uptakes
and the highest catalytic CO2 conversion as compared to the
homometallic ZIF-8 and ZIF-67.11,26 Moreover, ZIF-8 is used
as a photocatalyst for methylene blue dye degradation.22
However, it is only active under UV−Vis illumination due to its
large optical band gap (Eg = 4.9 eV).27,28 On the other hand,
while ZIF-67 has a band gap of 1.98 eV, which makes it an
excellent visible light absorber, its photocatalytic activity for dye
degradation is shown to be poor, mainly attributed to its weak
chemical stability in aqueous media. As a remedy, Zhang et al.
showed that doping ZIF-67 with Cu(II) enhances chemical
stability and thereby improves the visible light photo-
degradation of methyl orange in the presence of H2O2.22 In a
very recent report, rhombic dodecahedrons of ZIF-67 were
used as photocatalysts for the reduction of Cr(VI) with an
efficiency of around 70% under UV−Vis light irradiation.29
In this paper, we report a new method to synthesize ZIF-8,
ZIF-67, and their heterometallic version via a reaction−
diffusion framework (RDF) where the 2-methyl imidazolate
(HmIm)-based solution (outer electrolyte) is added on top of
Received: October 31, 2017
Published: January 5, 2018

© 2018 American Chemical Society 1812 DOI: 10.1021/jacs.7b11589

J. Am. Chem. Soc. 2018, 140, 1812−1823
Journal of the American Chemical Society
Figure 1. Theoretical basis of the reaction−diffusion framework (RDF). (A) Diffusion profiles of HmIm (outer) and Zn2+/Co2+ (inner) are depicted
at a given time. Evolution of the reaction zone is also shown and exhibits a decrease of its amplitude and broadening of its width. xf denotes the
location of the peak which also corresponds to the location of the precipitation front. (B) Nucleation of pure ZIF-8 leading to nanosperoids takes
place within the reaction zone. Transition to multisized Co-doped polyhedra with a gradient of composition takes place in its wake.
an agar gel matrix containing the coprecipitating metal cations
(Zn2+ and/or Co2+), serving as so-called inner electrolytes. In
RDF, because the reaction is taking place in a gel matrix which
hinders the kinetics of precipitation, the mechanism of the
formation of the solid ZIF material can be portrayed by the
following consecutive physicochemical processes (Figure 1A):
k
HmIm(aq) + Zn 2 +(aq)/Co2 +(aq) → {ZIF(aq)} → ZIF(s)
where HmIm diffuses and reacts with Zn2+ and/or Co2+ with a
rate constant k to generate the colloidal intermediate {ZIF(aq)}
in a narrow reaction zone defined by the rate function R =
k[HmIm][M2+] (where M = Zn and/or Co) that exhibits
maximum production rate at location xf which can be used to
define the location of the reaction front. This reaction zone that
resembles a Gaussian curve is characterized by a time-varying
amplitude, width, and speed. Because the chemical reaction is
much faster than diffusion, the dynamics of the reaction front is
diffusion controlled and depends mainly on the initial
concentration of the diffusing molecules and their diffusion
coefficients DHmIm, DZn2+, DCo2+, and DZIF.30 The initial
concentrations of the inner and outer electrolytes are chosen
such that their product (initial supersaturation) exceeds by far
their solubility product, which in turn leads to a propagating
supersaturation wave similar in shape and dynamics to the
reaction zone.31 The diffusing species {ZIF(aq)} which is
produced inside the reaction zone will then nucleate if its
concentration exceeds a certain threshold. These nuclei will
subsequently grow into immobile crystals in the wake of the
reaction zone as depicted in Figure 1B. The gel matrix in RDF
provides a solid support for the nucleation and growth of the
ZIF crystals, eliminates any disruptive convective currents, and
slows down the kinetics of the growing crystals. Consequently,
we are able to study the transition from nucleation to
agglomeration leading to the crystal growth of ZIFs in slow
motion (Figure 1B), providing insights into the fundamentals
of ZIF crystallization.32 At high supersaturation, especially at
the early stages of diffusion near the interface, the rate of
nucleation dominates and we obtain small crystals; however, as
the supersaturation weakens far from the interface, growth
dominates and we obtain fewer but larger crystals (Figure 1B).
Hence, the established supersaturation wave leads to a gradient
in crystal size and can range from a few nanometers to a few
micrometers. This size variation is nonlinear and depends on
the coupling of the reaction−diffusion process to nucleation
and growth kinetics.
1813
Article
This synthesis of ZIFs via RDF is also advantageous for many
reasons: (i) it is carried out efficiently at room temperature
without the usage of aging processes or thermal treatments, (ii)
it is scalable where gram scales can be obtained by simply using
larger reactors, (iii) it permits the preparation of solid solution
and generation of a spatial doping gradient along the tubular
reactor, and (iv) it easily provides control over the particle size
and morphology.
Moreover, the photocatalytic properties of the obtained
crystals are explored by examining the extent of methylene blue
(MB) degradation by ZIF-8 and the Co(II)-doped versions
under visible light irradiation without using H2O2.
■ EXPERIMENTAL SECTION
Materials. 2-Methyl imidazole (HmIm), hydrochloric acid (HCl),
dichloromethane (DCM), zinc and cobalt salts, isopropyl alcohol and
N,N-dimethylformamide (DMF) were purchased from Sigma-Aldrich,
while Bacto Agar was obtained from Difco, and methylene blue (MB)
was bought from Acros Organics.
Preparation Method. The inner portion was prepared by adding
agar powder into an aqueous solution containing the desired amount
of zinc nitrate. The mixture was heated and stirred until the agar
dissolved completely; then an equal volume of DMF was added with
continuous stirring. The obtained solution was poured into a test tube
(10 cm × 1 cm or 20 cm × 2.4 cm, depending on the scale of the
experiment) to fill its two-thirds (5 mL for the smaller tube or 50 mL
for the larger one) and then covered with parafilm and left for 2 h at
room temperature to allow complete gelation of agar. After gelation of
the inner portion, the outer solution (HmIm dissolved in a 1:1 mixture
of water/DMF) was poured on top of it (2 mL for the smaller tube
and 10 mL for the larger one). The tube was then covered and left for
2 days to allow the reaction−diffusion process and formation of the
white precipitate front. This process is shown in Scheme 1 and Figure
2A. Interestingly, ZIF crystals can be synthesized using different ratios
of water and DMF. It is noteworthy that in the hydrogel without any
DMF the layered ZIF structure (ZIF-L) was obtained (Figure S1).33
The content of the tubes was divided into equidistant bands of
about 5 mm with a thickness of 1 mm for each one as shown in Figure
2B. Each band was extracted from the tube and then washed in DMF
until complete redissolution of the gel. After centrifugation, the
crystals were collected for analysis. For the ZIF-67, the zinc nitrate was
substituted by cobalt nitrate and a purple precipitate was obtained
(Figure 2B). As for the heterometallic ZIF (zinc/cobalt), the same
preparation method was used where zinc and cobalt nitrate were
mixed in different ratios in the gel. In this case, after reaction−
diffusion, two regions were observed, a relatively narrow leading white
precipitate that gradually transforms into a purple precipitate (Figure
2B and movie in SI). For particle size measurements in this case, the
solid particles from the leading white band are excluded. For atomic
absorbance analysis, consecutive bands were extracted and washed.
DOI: 10.1021/jacs.7b11589
J. Am. Chem. Soc. 2018, 140, 1812−1823
Journal of the American Chemical Society
Scheme 1. Reaction Scheme between the Outer Solution of
HmIm and the Inner Gel solution containing Zn(II) and/or
Co(II) To Give the ZIF Crystals
The dry powder was then dissolved in 1 ml of 1 M HCl. After required
dilutions, molar concentrations of cobalt and zinc cations were
Figure 2. Schematic representation of the ZIFs synthesis via the reaction−diffusion framework (A). Formation of the ZIF-8, Co-doped ZIF-8, and
ZIF-67 crystals in the tubular reactor where the extracted bands are indicated as bn (B). log−log plots of the displacement of ZIF-8 (blue) and ZIF-
67 (red) fronts for different inner electrolyte concentrations (C).
1814
Article
measured and the ratio x = Co(II)/(Co(II) + Zn(II)) was calculated.
At the macroscopic scale, the same preparation method was used and a
thin test tube (0.5 cm × 20 cm), in which the reaction is taking place,
was placed under a Canon EOS 450D camera which was controlled to
take a shot every 60 min for 60 h. The smallness of the tube’s diameter
guarantees a pseudo-one-dimentional setup for front propagation.
Using digital photoanalysis, the distances traveled by each of the ZIF
precipitation fronts, d, are measured as a function of time, t, for
different inner electrolyte concentrations and exhibit a scaling law of
the form d ≈ tn as clearly shown in the log(d) vesus log(t) curves
plotted in Figure 2C for ZIF-8 and ZIF-67. The measured distance can
be considered as xf depicted in Figure 1A, as no reaction occurs
beyond the reaction zone. The exponent values, n, are calculated as the
slopes of the fitted lines and found to be ca. 0.5 for both ZIFs,
confirming that the formation of the ZIFs are indeed diffusion
controlled. This is further endorsed by the fact that both ZIFs travel
the same distance at any given time during propagation for the same
inner and outer concentrations. For such a case, it is mathematically
shown that the position of the front depends only on the initial
concentrations of the inner and outer solutions and their diffusion
coefficients, and since DZn2+ are DCo2+ are almost equal,34 the distance
traveled by the two fronts must be equal.35 Moreover, the lower the
concentration of the inner electrolyte, the faster the propagation of the
ZIF front (Figure 2C). Because the system can be considered as 1D as
mentioned above, the effective diffusion coefficient (Deff) for the three
inner concentrations in Figure 2C can be extracted from the equation
d2 = Deff × t and is found to be 0.058 (10 mM), 0.051 (25 mM), and
0.039 cm2/s (50 mM).35 It is noteworthy that although both fronts
DOI: 10.1021/jacs.7b11589
J. Am. Chem. Soc. 2018, 140, 1812−1823
Journal of the American Chemical Society
Figure 3. PXRD patterns of the simulated ZIF-8/67 (A (i)) and the RDF-synthesized ZIF-8 (A (ii)) and ZIF-67 (A (iii)). Thermogravimetric curves
of ZIF-8 (B (i)) and ZIF-67 (B (ii)). Nitrogen adsorption (solid circle)−desorption (empty circle) isotherms at 77 K of ZIF-8 (C) and ZIF-67 (D).
(ZIF-8 and ZIF-67) travel the same distance under the same
conditions, the particle size distributions for both ZIFs are completely
disparate as we are going to explore in the coming sections.
ZIFs Characterization. Powder X-ray diffraction (PXRD) patterns
were recorded on a Bruker D8 discover XRD diffractometer using Cu
Kα radiation (λ = 1.5406 Å) at 40 kV and 40 mA, with 2θ ranging
between 5° and 40° at a scanning rate of 0.5°·min−1. Thermogravi-
metric analysis (TGA) of the precipitates was performed under
nitrogen atmosphere with a heating rate of 5 °C·min−1 and a
temperature ranging from 30 to 900 °C using a TG 209 F1 Iris
(Netzsch, Germany). The N2 adsorption/desorption isotherms,
surface area, and total pore volume were obtained using a
Micromeritics ASAP 2420 analyzer. Prior to the N2 adsorption/
desorption and TGA measurements, the samples were washed for 3
days with DMF followed by washing with DCM for 2 days. The UV−
Vis diffuse reflectance experiments were recorded on a JASCO V-570
UV−Vis−NIR Spectrophotometer. The morphology of the crystals
was examined using a scanning electron microscope (Tescan Mira),
operating at 5 kV with an Oxford detector for energy-dispersive X-ray
(EDX) characterization. The atomic absorption measurements were
carried out by a Thermo Labsystems SOLAAR atomic absorption
spectrophotometer with ASX-510 autosampler and SOLAAR data
acquisition. The photocatalytic activity of the Co-doped ZIF-8 was
investigated by degradation of aqueous methylene blue, acting as a
model pollutant of water, using solar light. The photodegradation
experiments were carried out in a photoreactor equipped with a 100 W
xenon lamp, an AM 1.5 optical filter, a magnetic stirrer, and a jacketed
beaker, placed 20 cm away from the light source, containing 15 mg of
ZIF powder and 10 mg·L−1 of methylene blue (Vtotal = 50 mL). Prior
to irradiation, the ZIF/organic dye mixture was stirred in the dark for
60 min to reach the adsorption−desorption equilibrium and then
exposed to a solar light source. Samples were taken at time intervals,
and the residual methylene blue was detected by a spectrophotometer
at λmax = 665 nm.
1815
Article
■ RESULTS AND DISCUSSION
Characterization of the White (ZIF-8) and Purple (ZIF-
67) Precipitates. The phase purity and structural identity of
the white- and purple-extracted precipitates are determined
using PXRDs. The recorded patterns exhibit sharp peaks that
match the simulated ones from the sodalite structures of ZIF-8
and ZIF-67 (Figure 3A). No extra peaks are observed, which
demonstrates the purity of the ZIF samples synthesized using
RDF. The observed peak broadening in ZIF-8 crystals suggests
the formation of nanosized crystals which are also observed by
high-resolution scanning electron microscopy (HRSEM). The
thermal behavior of the as-synthesized and activated ZIFs is
examined via thermogravimetric analysis (TGA). Similar to the
previously reported ZIFs, the TGA curves in Figure 3B
demonstrate their high thermal stability (up to 500 °C) for
both ZIFs. To assess the porosity and the architectural stability
of the ZIF structures, N2 gas adsorption measurements at 77 K
are performed. The Brunauer−Emmett−Teller (BET) surface
areas for ZIF-8 (Figure 3C) and ZIF-67 (Figure 3D) are
calculated to be 1302 and 1023 m2·g−1, respectively. These
values agree with the reported values in the literature.25
Morphological Features and Particle Size Distribution
of ZIF-8 and ZIF-67. The morphology of the extracted ZIF-8
and ZIF-67 is inspected via HRSEM (Figure 4), which reveals a
rhombic dodecahedral shape with sharp edges and smooth
faces. In order to show the effect of the generated super-
saturation gradient along the tube, one test tube is prepared at
room temperature, where the concentrations of the inner
electrolyte, outer electrolyte, and agar gel are 5 mM, 2 M, and
1% (w/w), respectively. After 48 h of propagation, six
equidistant bands (Figure 2B), having a thickness of 1 mm
DOI: 10.1021/jacs.7b11589
J. Am. Chem. Soc. 2018, 140, 1812−1823
Journal of the American Chemical Society Article

Figure 4. Particle size distributions of ZIF-8 (A) and ZIF-67 (B) extracted from six consecutive equidistant (5 mm) bands along the tube. Each band
has a thickness of 1 mm.

Figure 5. Variation of the average ZIF-8 (A) and ZIF-67 (B) particle size with respect to the band distance from the gel/outer electrolyte interface.
Different band numbers are indicated in Figure 2B.

and a separation of 5 mm, are extracted and examined by
HRSEM. The size of the dodecahedrons gradually increases
from 500 nm near the interface (band 1 (b1)) until it reaches 4
μm at the white leading front (band 6 (b6)) for ZIF-8 (Figures
4A and 5A) and from 1 to 10 μm for ZIF-67 (Figures 4B and
5B). Upon addition of the 2-methylimidazole solution (outer
1816
electrolyte), a supersaturation wave is generated by the
interdiffusing species along the tube starting at the gel/outer
interface downward. Due to diffusion, this supersaturation wave
is characterized by a slowly decreasing amplitude and widening
breadth.36−38 Because the rates of nucleation and growth of the
ZIF particles strongly and nonlinearly depend on the
DOI: 10.1021/jacs.7b11589
J. Am. Chem. Soc. 2018, 140, 1812−1823
Journal of the American Chemical Society
Figure 6. Effect of temperature (A), concentration of the inner electrolyte (B), and concentration of the agar gel (C) on the average particle size
distribution of the ZIF-8 crystals. Each band is approximately equal to 1 mm, and the consecutive bands are equidistant (5 mm), as indicated in
Figure 2B. Average ZIFs size is reported as a mean of a sample size of 20 particles, and the error bars represent the standard deviation.
magnitude of the supersaturation, we obtain a gradient of sizes
along the tubular reactor, although the extracted bands are
equidistant. Near the interface, nucleation dominates and
smaller particles are formed; conversely, as we go farther from
the interface, growth starts to dominate and bigger particles are
obtained. Interestingly, in the first bands near the interface we
are able to observe the structural preorganization of the
nanocrystals before formation of the rhombic dodecahedral-
shaped crystals of ZIF-8 and ZIF-67 (Figure 4). The faceted
particles appear far from the gel/outer interface. Such features
of crystal growth cannot be observed in conventional methods
of ZIF synthesis, which makes RDF a very useful and unique
method to capture the different growth stages of the crystals.
Consequently, RDF provides a synthetic tool that easily allows
manipulation of size distributions through alteration of
supersaturation by simply varying different experimental
parameters, such as the temperature of the reaction medium,
the inner electrolyte types and concentrations, and the
thickness of the gel. The effect of these parameters is
investigated for ZIF-8 crystal synthesis.
Effect of Temperature. Four different experiments are
carried out for 48 h at different temperatures, 7, 15, 20, and 25
°C, where the concentrations of the inner electrolyte (zinc
nitrate), outer electrolyte, and gel are kept fixed at 50 mM, 1 M,
and 1% (w/w), respectively. Near to the interface (band 1), the
particle size increases with the temperature: (1) 0.3 μm, 7 °C;
(2) 0.6 μm, 15 °C; (3) 0.7 μm, 20 °C; (4) 0.8 μm, 25 °C. The
same trend is obtained as we go farther from the interface
(Band 2: (1) 0.8 μm, 7 °C; (2) 1.9 μm, 15 °C; (3) 2.9 μm, 20
°C; (4) 3.6 μm, 25 °C. Band 3: (1) 1.1 μm, 7 °C; (2) 2.6 μm,
15 °C; (3) 3.8 μm, 20 °C; (4) 4.2 μm, 25 °C), where bigger
particles are obtained as the temperature increases (Figures 6A
and S2). This trend in particle size over all of the extracted
bands is a direct consequence of the increase of the rate of
growth with temperature as we move away along the
precipitation front.
Effect of Inner Electrolyte. The effect of the inner
electrolyte is also investigated. First, the type of metal salt is
varied, and four different metal sources for zinc and cobalt are
used (acetate, chloride, sulfate, and nitrate). SEM (Figure S3)
measurements confirm the phase purity of ZIF-8 and ZIF-67.
The size distributions obtained from the different reactions are
shown in Figure S4. While the usage of different zinc salts does
not reveal any significant effect on the particle size distribution
of ZIF-8, the effect on the size of ZIF-67 crystals is noticeable
with those derived from the acetate having relatively the biggest
1817
Article
size. In a second set of reactions, three test tubes are prepared
in which the zinc nitrate concentration is varied (10, 50, and
100 mM), while the concentrations of the outer electrolyte and
gel are kept fixed at 1 M and 1% (w/w), respectively. Near to
the interface (band 1), the particle size increases as the inner
concentration increases: (1) 0.12 μm, 10 mM; (2) 0.95 μm, 50
mM; (3) 3.1 μm, 100 mM. The same trend is obtained as we
go farther from the interface (Band 2: (1) 1.7 μm, 10 mM; (2)
2.3 μm, 50 mM; (3) 4.1 μm, 100 mM. Band 3: (1) 2.6 μm, 10
mM; (2) 4.1 μm, 50 mM. Band 4:(1) 3.2 μm, 10 mM; (2) 5.0
μm, 50 mM), where smaller particles are obtained for lower
inner concentrations, as shown in Figure 6B. As the metal ion
(inner) concentration increases, at a fixed imidazole (outer)
concentration, the generated initial supersaturation gradient
will be smaller and the obtained particles are bigger when
compared at the same distance from the gel/outer interface.
Effect of the Gel. The effect of the agar gel concentration
on the ZIF-8 crystals size distribution is also investigated. Three
different experiments are carried out at room temperature and
at a gel concentration of 0.5%, 1%, and 1.5%, where the
concentrations of the inner and outer electrolytes are kept fixed
at 100 mM and 1 M, respectively. From the three bands
extracted from each tube we observe that near the interface
(band 1) the particle size decreases as the gel thickness
increases: (1) 1.6 μm, 0.5%; (2) 1.0 μm, 1%; (3) 0.51 μm,
1.5%. The same trend is obtained as we go farther from the
interface (Band 2: (1) 7.5 μm, 0.5%; (2) 5.9 μm, 1%; (3) 4.4
μm, 1.5%. Band 3: (1) 10.1 μm, 0.5%; (2) 7.7 μm, 1%; (3) 6.3
μm, 1.5%), where smaller particles are obtained for thicker gel,
as shown in Figure 6C. The inversely proportional relation
between the gel concentration and the particle size is due to the
generation of smaller pores as the gel thickness increases.37 By
decreasing the pore size, the growth of the crystals is hindered
and smaller particles are obtained.
Nucleation and Growth Kinetics of ZIFs Dodecahe-
drons. The nucleation and growth kinetics of the ZIF crystals
at their inception is investigated. For this study, ZIF-8 and ZIF-
67 are prepared in 2 identical reactors with 20 mM inner and 1
M outer concentrations. The crystals are then extracted from
the gel/outer interface, where the highest supersaturation is
encountered in each reactor at different times (30, 60, and 120
s) by discarding the outer electrolyte. PXRDs of the solids at
the first minute correspond to the ZIF-8 and ZIF-67 structures
and indicate that the formation of the ZIF crystals is very fast
(Figure S5). The SEM investigation of the gel/outer electrolyte
interface region reveals that the formation of ZIF-8 and ZIF-67
DOI: 10.1021/jacs.7b11589
J. Am. Chem. Soc. 2018, 140, 1812−1823
Journal of the American Chemical Society Article

Figure 7. Nucleation and growth kinetics of ZIFs dodecahedrons. SEM images of the ZIF-8 crystals after 30 (A), 60 (B), and 120 s (C) and of ZIF-
67 crystals after 30 (D), 60 (E), and 120 s (F) of reaction at the interface region.

Figure 8. Cobalt doping of ZIF-8 using RDF for different initial (Co(II)/(Co(II) + Zn(II)) ratios exhibiting a decreasing gradient of cobalt content
as we move away from the outer electrolyte/gel interface (A) along the reaction tube. Visible diffuse reflectance spectra of ZIF-8 (black line), ZIF-67
(blue line), and CoxZn1−x(HmIm)2 (red line) (B).

crystals is preceded by an instantaneous homogeneous
nucleation leading to nanospheroids with an average diameter
of less than 100 nm (Figure 7A and 7D). However, the
appearance of faceted crystals of ZIF-8 (Figure 7B) was slower
than that of ZIF-67 (Figure 7E and 7F) as evidenced by the
SEM images. The aggregation of these nanospheres, probably
via oriented attachment, leads to the formation of faceted
crystals within 120 s for ZIF-8 (Figure 7C) and 60 s for ZIF-67
(Figure 7E).27 The size of resulting ZIF-8 crystals (ca. 100 nm)
at the interface is smaller than that of ZIF-67 (ca. 300 nm).
This can be explained by the fact that ZIF-67 is not water stable
as compared to ZIF-8, and consequently, its rate of nucleation
is expectedly smaller because fewer nuclei survive thermal
fluctuations. Hence, nucleation for ZIF-8 dominates the crystal
formation, whereas growth governs in the case of ZIF-67,
leading to noticeably larger crystals throughout the reaction−
1818
diffusion process (Figure 5). This growth mechanism is
consistent with the one reported in the literature.9,27,39
Doping with Cobalt and Growth Mechanism of ZIF-8.
It has been demonstrated that Co(II) cations substitute Zn(II)
without altering the SOD topology of the ZIF structure.22,29 In
order to apply this doping process in RDF, different
percentages of Co(II) are added to the inner Zn(II) electrolyte.
Interestingly, upon addition of the outer electrolyte (HmIm)
two regions are observed: a leading white precipitate that
gradually transforms into a purple one (Figure 8A). Therefore,
the incorporation of Co(II) into the ZIF-8 structure to form
the purple CoxZn1−x(HmIm)2 can be monitored via a visible
color change of the precipitate. The UV−Vis diffuse reflectance
spectrum, recorded for the purple precipitate, exhibits
characteristic bands of tetrahedral-coordinated Co(II) at 585
DOI: 10.1021/jacs.7b11589
J. Am. Chem. Soc. 2018, 140, 1812−1823
Journal of the American Chemical Society
Figure 9. SEM image of a Co-doped ZIF-8 dodecahedron and its corresponding elemental mapping of Co and Zn.
and 540 nm (Figure 8B), confirming the incorporation of
Co(II) in the ZIF-8 structure.
The PXRD pattern for the purple precipitate is identical to
the ZIF-8 pattern, which indicates that the synthesized Co-
doped ZIF-8 exhibits a SOD topology identical to the parent
framework. Since Co(II) and Zn(II) have slightly different ionic
radii (0.74 Å for Zn(II) and 0.72 Å for Co(II)), it makes it
difficult to uncover any shift toward lower angles for the Co-
doped samples (Figure S6). For the surface composition
analysis, an energy-dispersive X-ray (EDX) analysis is
conducted on the purple precipitate and reveals the coexistence
of the elements Zn and Co in the same crystal, confirming the
formation of a solid solution of ZIF-8 and ZIF-67 and not a
physical mixture of both (Figure 9). To gain more insight into
the doping reaction mechanism, 10 different initial cobalt
percentages are used in the inner electrolytes (5%, 10%, 20%,
30%, 40%, 50%, 60%, 70%, 80%, and 90%). The different tubes
are divided into 4 equidistant bands (both the band separation
and the band thickness are equal to 5 mm) to determine the
homogeneity of the substitution throughout the precipitation
region. The crystals in each of the 4 bands are extracted,
washed with DMF, and examined by EDX and atomic
absorption. Figure 10 shows the atomic percentage of cobalt
in the ZIF-8 crystal structure for the four bands in each of the
10 experiments. Remarkably, the percentage of cobalt
incorporated in the structure gradually decreases as the distance
from the gel/outer interface increases in all tubes. For example,
for an initial mixture of 1:1 Co(II):Zn(II), the value of x in
Figure 10. Variation of the Co(II) incorporation percentage in the
ZIF-8 crystal structure as a function of the Co(II) percentage present
in the inner electrolyte and the distance from the gel/outer interface.
Each band has a thickness of 5 mm, and each of the two consecutive
bands are separated by 5 mm.
1819
Article
CoxZn1−x(HmIm)2 ranges between 0.33 (in band 1) and
0.01(in band 4). Moreover, the maximum precentage of doped
cobalt is found in the first bands and represents 50% to 75% of
the initial concentration ratio and decreases gradually to reach
less than 30% in the leading bands. From this set of doping
experiments, we conclude that the incorporation of Zn is
thermodynamically more favored than that of Co in the ZIF
framework.
The measured particle sizes for the heterometallic
CoxZn1−x(HmIm)2 crystals are consistently smaller than both
the ZIF-8 and the ZIF-67 ones in all bands (Figures 11 and S7).
Figure 11. Variation of the particle size of the ZIF-8 dodecahedrons as
a function of the initial Co(II) percentage present in the inner
electrolyte and the distance from the gel/outer electrolyte interface.
Each band has a thickness of 5 mm, and each of the two consecutive
bands are separated by 5 mm.
This is contrary to what Zareba et al.25 observed for their
heterometallic Co−Zn−ZIFs, where they found that the
particle size increases exponentially with the Co percentage.
Their particle size range was between 20 and 500 nm; however,
in our RDF system, the obtained particle sizes fall in a different
range (from 100 nm to 10 μm). This trend in crystal size of the
CoxZn1−x(HmIm)2 crystals is seemingly due to Co(II) ions that
act as “impurities” with respect to ZIF-8 crystal nucleation that
starts first as evidenced by the leading white front. Therefore, it
looks like a plausible idea that the presence of cobalt enhances
the kinetics of nucleation, leading to doped particles of
relatively smaller size. This is further endorsed by a higher
initial concentration of cobalt ions reaching a 95% ratio (Figure
DOI: 10.1021/jacs.7b11589
J. Am. Chem. Soc. 2018, 140, 1812−1823
Journal of the American Chemical Society
S8). Interestingly, however, at 100% cobalt (pure ZIF-67), the
size of the obtained crystal is distinctively bigger.
The effect of the temperature on the doping reaction is
further investigated, and three different temperatures (7, 23,
and 30 °C) are chosen with an initial Co(II) concentration of
10%. The efficiency of Co doping is increased by a factor of 2
for an increase of the reaction temperature from 7 to 30 °C
(Figure 12). This enhancement in the percentage of Co
Figure 12. Effect of temperature on the cobalt-doping percentage of
the ZIF-8 crystals.
incorporation is observed for all bands in the tube. As
mentioned above, the ZIF-8 formation is kinetically more
favored, and hence, the growth of the ZIF-8 dodecahedrons
brings a local depletion in the concentration of Zn(II) at the
expense of Co(II), which will then be incorporated in the
structure. Since smaller particles are obtained at lower
Figure 13. SEM images of the leading white front (A), transition zone from white to purple (B−E), and upper purple precipitate (F) from a Co-
doped ZIF-8 system showing the formation mechanism of the ZIF dodecahedrons.
1820
Article
temperature, the local depletion rate in the zinc cations
becomes lower; thus, less cobalt is consumed for growth of the
particles.
From the SEM images of the doped samples, we were able to
capture the structural evolution of the solid−solution ZIFs. A
reorganization of nanospheroidal particles in the white leading
front to form the faceted particles in the purple precipitation
zone can be observed in Figure 13. This comportment was
observed for all of the doped samples as shown, for example, in
the SEM images (Figure S9) for the white front region for three
different Co-doping percentages (1%, 5%, and 10%). Here the
white front is further partitioned into three consecutive bands
(Figure S9A−C) to capture the agglomeration of the
nanospheres (Figure 9A) into polyhedral-shaped structures
(Figure 9B) before assuming well-defined single crystals
(Figure 9C). We then selected the 1% doping case for PXRD
analysis (Figure S10) because it relatively provides the widest
white front (ca. 0.5 cm) with enough powder for analysis. The
lowest band in the white front away from the white−pink
interface is composed of nanospheroids (50−100 nm) of
crystalline pure ZIF-8 (from EDX data) and exhibits broad-
ening of the peaks in the PXRD data (Figure S10A). This
indicates that in addition to its higher thermodynamic stability,
ZIF-8 is also kinetically more favored than the Co-doped ZIF-8.
Furthermore, as the agglomeration of the white spheres starts
near the white−purple interface in the second band of the
white front, the doping begins to take place and the PXRD
peaks sharpen (Figure S10B) until they reach the final cobalt
composition, size, and well-defined polyhedral morphology in
band 3 of the white front located at the white−purple interface
and above (Figure S10C). Moreover, the SEM images in the
transition region in Figure 13 show the existence of hollow
assembly. We believe that some of the formed polyhedral might
DOI: 10.1021/jacs.7b11589
J. Am. Chem. Soc. 2018, 140, 1812−1823
Journal of the American Chemical Society
remain hollow and others are fully filled to form single crystals.
Self-assembly in MOFs and ZIFs crystals to form hollow
macrostructures is common and was recently reviewed.40 This
feature deserves further investigation in the future. This relative
ease in monitoring crystal formation constitutes one of the
main advantages of using RDF and confirms the mechanism
obtained using MD simulations.41
Stability of Co-Doped ZIF-8. Because ZIF-67 is not water
stable, the chemical stability of Co-doped ZIF-8 is investigated.
We extracted doped ZIF-8 crystals with various cobalt
concentrations and washed them with water for 24 h.
Remarkably, the doped products for an initial Co(II)
concentration below 50% are chemically stable as evidenced
from their PXRDs (Figure S11). Nevertheless, for high initial
Co(II) concentrations (70% and above), the obtained solid−
solution ZIFs are not chemically stable.
Photocatalytic Properties of ZIF Crystals. The gen-
erated ZIF crystals are tested for the photocatalytic degradation
of methylene blue dye under visible light irradiation. The band
gaps (Eg) of ZIF-8, ZIF-67, and the heterometallic ZIF are
determined from their diffuse reflectance spectra (converted
using Kubelka−Munk function) (Figure 8B) and found to be
4.9, 2.0, and 2.0 eV for ZIF-8, ZIF-67, and Co-doped ZIF-8
(10%), respectively. This indicates that our doping approach is
useful to control the optical properties and to bring the ZIF-8
material to the realm of visible light absorption and
photocatalysis. In order to test for any enhancement in
photocatalytic performance of ZIF-8, ZIF-67, and Co-doped
ZIF-8, we evaluate their activity during the degradation of MB
under visible light illumination without using H2O2.22 The rate
and extent of MB degradation are determined from the time-
dependent decrease of the MB concentration, which is
calculated from the decrease in the visible absorbance band
(665 nm) of MB. It is clear from Figures S12 and S13 that the
performance of all of the doped ZIF is similar. Therefore, we
choose the 10% Co-doped ZIF-8 (surface area of 1235 m2·g−1
(Figure S14)) to compare with the pure ZIF-8. The control
solution, consisting of pure MB with no catalyst, shows a small
percentage of degradation (around 10%) after 90 min of
illumination (Figures 14 and S15). Moreover, in the absence of
light, the MB solution in the presence of ZIF-8 and the 10%
Co-doped-ZIF-8 displayed a saturation period of 45 min for the
Figure 14. Photocatalytic performance of the ZIF-8 and 10% Co-
dopedZIF-8 in the degradation of methylene blue (MB) under visible
light and for two consecutive cycles.
1821
Article
former and 3 min for the latter (Figure S16). Remarkably, the
Co-doped ZIF-8 exhibits higher dye adsorption (ca. 65%) than
that of ZIF-8 (ca. 40%) at saturation under our experimental
conditions. This notable difference in saturation time and
adsorption capacity between the two solids is one of the
important facets of doping.
Kinetically, the UV−vis spectra of the MB solutions exhibit a
diminution as a function of chromophoric absorbance
maximum of MB centered at 665 nm42 as shown in Figure
S17. Interestingly, 70% of the MB is photodegraded in 90 min
when the Co-doped ZIF-8 (10%) is used, whereas only 45% is
degraded when ZIF-8 is used (Figure 14). The kinetic data in
Figure 14 are fitted to a stretched exponential, and the rate
constants in the case of the pure and Co-doped ZIF-8 are
extracted (Table S2) and found to differ, in the first cycle, by 1
order of magnitude (0.0044 ± 0.0011 and 0.021 ± 0.001 min−1
for ZIF-8 and the Co-doped ZIF-8, respectively). On the other
hand, the high chemical stability of the samples after the
photodegradation reactions is confirmed as evidenced by the
PXRD and SEM shown in Figure S18. Moreover, the
reusability of both photocatalysts is tested, and remarkably,
both ZIFs retained their relative activity in the second cycle
with rate constants of 0.0021 ± 0.001 min−1 for ZIF-8 and
0.016 ± 0.001 min−1 for Co-doped ZIF-8 (Figure 14). On the
other hand, it is noteworthy that since ZIF-67 is not chemically
stable in water, it cannot be used in such applications. It is also
important to mention that the reported dye photodegradation
studies using ZIF materials employed H2O2 to produce •OH
radicals that speed up the oxidation process22 or strong UV
light irradiation,43 whereas our experiments are performed in
aqueous solution in the absence of H2O2 and under visible light
illumination.
Under visible light illumination, electrons are excited from
the HOMO to the LUMO of the ZIF material, generating holes
and electrons. The photogenerated electrons react with the
dissolved or adsorbed oxygen on the surface of ZIF crystals to
produce the superoxide radical anion •O2−, while the holes
react with water to form reactive hydroxyl radicals •OH, which
probably cause the degradation of MB as deduced from the
peaks around 210 nm in the case of ZIF-8 and around 272−280
nm for the Co-doped ZIF-8 (Figure S17). Moreover, only in
the case of ZIF-8, the aforementioned reaction competes with
the decolorization of MB by reduction to yield leuco methylene
blue (LMB), which is usually observed at 256 nm43,44 (Figure
S17A and S17B). The mechanism involving •OH is further
confirmed by performing the same photodegradation reaction
in the presence of isopropyl alcohol as hole scavenger. In this
experiment, we performed photodegradation tests with the 10%
Co-doped ZIF-8. In this case the degradation was almost
completely inhibited as shown in Figure S19. A similar
mechanism had been proposed recently for the degradation
of organic dyes in the presence of metal−organic frame-
works.45,46 Under similar experimental conditions, Jelle et al.
recently reported the photodegradation of MB over FeOOH
materials and showed that only 20% of MB is degraded after 90
min, and to reach 70% of degradation more than 300 min is
needed.42 Their turnover frequency, TOF (TOF = 2.2 × 10−4
min−1), is found to be relatively lower than ours for the 10%
Co-doped ZIF-8 (TOF = 3.0 × 10−4 min−1). In addition to the
enhancement of visible light absorption by the doping strategy,
the efficiency of our catalyst can also be attributed to the
remarkably enhanced adsorption of MB on the surface of the
doped material, as mentioned above, which is a favored
DOI: 10.1021/jacs.7b11589
J. Am. Chem. Soc. 2018, 140, 1812−1823
Journal of the American Chemical Society
precursor for heterogeneous catalysis. The obtained results
demonstrate the importance of Co doping on enhancing the
photocatalytic properties of ZIF-8.
■ CONCLUSIONS
We synthesized and fully characterized ZIF-8 and ZIF-67
crystals using a reaction−diffusion framework (RDF) where the
precipitation reaction resulted from the interdiffusion of initially
separated metal ions and organic linkers in agar gel as a matrix.
The supersaturation gradient in RDF provided us a means to
effectively control the size of the ZIF crystals. Crystal size
distributions were also studied as a function of the
concentration of the inner electrolytes, the temperature of
the reaction medium, and the thickness of the gel.
Consequently, we demonstrated that this method is effective
in tuning with ease the crystal size distribution at ambient
temperature. We also studied the doping of ZIF-8 by cobalt
ions by means of RDF. This also revealed a gradient of doping
concentration along the tubular reaction reactor. Furthermore,
we took advantage of the slow rates of nucleation and growth
inside the gel pores to visualize the growth mechanism of the
ZIF-8 crystals and their Co-doped analogues that consisted of
an aggregation of ZIF-8 nanospheres to form well-resolved
polyhedra. Finally, the photocatalytic properties of the Co-
doped ZIF-8 crystals were evaluated by studying the
degradation of MB under visible light irradiation. We
demonstrated the importance of Co doping in tuning the
optical and photocatalytic activity of ZIF-8 for dye degradation
under visible light irradiation.
■
ASSOCIATED CONTENT
* Supporting Information
S V.; Soyer, E.; Quoineaud, A.-A.; Bats, N. J. Am. Chem. Soc. 2010, 132,
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.7b11589.
Characterization of the ZIF-L crystals; SEMs, BET,
particle size distribution of the ZIF crystals for different
counter anions, stability of the Co-doped ZIF-8 in
aqueous solution, and full spectra of the photo-
degradation of MB using different ZIFs (PDF)
Movie for the front propagation and formation of Co-
doped ZIF-8 in the reaction-diffusion framework.
(MOV)
■
AUTHOR INFORMATION
Corresponding Authors
*mazen.ghoul@aub.edu.lb
*mohamad.hmadeh@aub.edu.lb
ORCID
Mohamad Hmadeh: 0000-0003-3027-3192
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
The authors gratefully acknowledge the funding provided by
the American University of Beirut Research Board and by the
Lebanese National Council for Scientific Research (LCNSR).
M.H. acknwoledges funds from The World Academy of Science
(TWAS) grant #103277. We also acknowledge the help of Mr.
Antranik Jonderian in the elemental mapping experiment.
1822
Article
■ REFERENCES
(1) Eddaoudi, M.; Kim, J.; Rosi, N.; Vodak, D.; Wachter, J.; O’keeffe,
M.; Yaghi, O. M. Science 2002, 295, 469−472.
(2) Rosi, N. L.; Eckert, J.; Eddaoudi, M.; Vodak, D. T.; Kim, J.;
O’keeffe, M.; Yaghi, O. M. Science 2003, 300, 1127−1129.
(3) Furukawa, H.; Ko, N.; Go, Y. B.; Aratani, N.; Choi, S. B.; Choi,
E.; Yazaydin, A. Ö .; Snurr, R. Q.; O’Keeffe, M.; Kim, J.; Yaghi, O. M.
Science 2010, 329, 424−428.
(4) Chen, R.; Yao, J.; Gu, Q.; Smeets, S.; Baerlocher, C.; Gu, H.; Zhu,
D.; Morris, W.; Yaghi, O. M.; Wang, H. Chem. Commun. 2013, 49,
9500−9502.
(5) Tang, J.; Salunkhe, R. R.; Liu, J.; Torad, N. L.; Imura, M.;
Furukawa, S.; Yamauchi, Y. J. Am. Chem. Soc. 2015, 137, 1572−1580.
(6) Hu, Y.; Liu, Z.; Xu, J.; Huang, Y.; Song, Y. J. Am. Chem. Soc. 2013,
135, 9287−9290.
(7) Rodenas, T.; Luz, I.; Prieto, G.; Seoane, B.; Miro, H.; Corma, A.;
Kapteijn, F.; Llabres i Xamena, F. X. L.; Gascon, J. Nat. Mater. 2015,
14, 48−55.
(8) Tosheva, L.; Valtchev, V. P. Chem. Mater. 2005, 17, 2494−2513.
(9) Venna, S. R.; Jasinski, J. B.; Carreon, M. A. J. Am. Chem. Soc.
2010, 132, 18030−18033.
(10) Jian, M.; Liu, B.; Liu, R.; Qu, J.; Wang, H.; Zhang, X. RSC Adv.
2015, 5, 48433−48441.
(11) Kaur, G.; Rai, R. K.; Tyagi, D.; Yao, X.; Li, P.-Z.; Yang, X.-C.;
Zhao, Y.; Xu, Q.; Singh, S. K. J. Mater. Chem. A 2016, 4, 14932−
14938.
(12) Sun, W.; Zhai, X.; Zhao, L. Chem. Eng. J. 2016, 289, 59−64.
(13) Shang, W.; Kang, X.; Ning, H.; Zhang, J.; Zhang, X.; Wu, Z.;
Mo, G.; Xing, X.; Han, B. Langmuir 2013, 29, 13168−13174.
(14) Bao, Q.; Lou, Y.; Xing, T.; Chen, J. Inorg. Chem. Commun. 2013,
37, 170−173.
(15) Safarifard, V.; Morsali, A. Coord. Chem. Rev. 2015, 292, 1−14.
(16) Guo, H.; Zhu, Y.; Wang, S.; Su, S.; Zhou, L.; Zhang, H. Chem.
Mater. 2012, 24, 444−450.
(17) Chizallet, C.; Lazare, S.; Bazer-Bachi, D.; Bonnier, F.; Lecocq,
12365−12377.
(18) Kuo, C.-H.; Tang, Y.; Chou, L.-Y.; Sneed, B. T.; Brodsky, C. N.;
Zhao, Z.; Tsung, C.-K. J. Am. Chem. Soc. 2012, 134, 14345−14348.
(19) Banerjee, R.; Furukawa, H.; Britt, D.; Knobler, C.; O’Keeffe, M.;
Yaghi, O. M. J. Am. Chem. Soc. 2009, 131, 3875−3877.
(20) Lu, G.; Hupp, J. T. J. Am. Chem. Soc. 2010, 132, 7832−7833.
(21) Liu, S.; Xiang, Z.; Hu, Z.; Zheng, X.; Cao, D. J. Mater. Chem.
2011, 21, 6649−6653.
(22) Yang, H.; He, X.-W.; Wang, F.; Kang, Y.; Zhang, J. J. Mater.
Chem. 2012, 22, 21849−21851.
(23) Tang, J.; Salunkhe, R. R.; Zhang, H.; Malgras, V.; Ahamad, T.;
Alshehri, S. M.; Kobayashi, N.; Tominaka, S.; Ide, Y.; Kim, J. H. Sci.
Rep. 2016, 6, 30295.
(24) Zhou, K.; Mousavi, B.; Luo, Z.; Phatanasri, S.; Chaemchuen, S.;
Verpoort, F. J. Mater. Chem. A 2017, 5, 952−957.
(25) Zaręba, J. K.; Nyk, M.; Samoć, M. Cryst. Growth Des. 2016, 16,
6419−425.
(26) Kuruppathparambil, R. R.; Babu, R.; Jeong, H. M.; Hwang, G.-
Y.; Jeong, G. S.; Kim, M.-I.; Kim, D.-W.; Park, D.-W. Green Chem.
2016, 18, 6349−6356.
(27) Cravillon, J.; Schröder, C. A.; Nayuk, R.; Gummel, J.; Huber, K.;
Wiebcke, M. Angew. Chem. 2011, 123, 8217−8221.
(28) Wang, C.-C.; Li, J.-R.; Lv, X.-L.; Zhang, Y.-Q.; Guo, G. Energy
Environ. Sci. 2014, 7, 2831−2867.
(29) Park, H.; Amaranatha Reddy, D.; Kim, Y.; Ma, R.; Choi, J.; Kim,
T. K.; Lee, K.-S. Solid State Sci. 2016, 62, 82.
(30) Abi Mansour, A.; Al Ghoul, M. Phys. Rev. E 2011, 84, 026107.
(31) Al-Ghoul, M.; Ammar, M.; Al-Kaysi, R. O. J. Phys. Chem. A
2012, 116, 4427−4437.
(32) Al-Ghoul, M.; Issa, R.; Hmadeh, M. CrystEngComm 2017, 19,
608−612.
(33) Low, Z.-X.; Razmjou, A.; Wang, K.; Gray, S.; Duke, M.; Wang,
H. J. Membr. Sci. 2014, 460, 9−17.
DOI: 10.1021/jacs.7b11589
J. Am. Chem. Soc. 2018, 140, 1812−1823
Journal of the American Chemical Society Article

(34) Lide, D. R.; Kehiaian, H. V.; CRC Handbook of Thermophysical

and Thermochemical Data; CRC Press: Boca Raton, FL, 1994.
(35) Gálfi, L.; Rácz, Z. Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38,
3151.
(36) Saliba, D.; Ezzeddine, A.; Sougrat, R.; Khashab, N. M.; Hmadeh,
M.; Al-Ghoul, M. ChemSusChem 2016, 9, 800−805.
(37) Saliba, D.; Al-Ghoul, M. CrystEngComm 2016, 18, 8445−8453.
(38) Saliba, D.; Ezzeddine, A.; Emwas, A.-H.; Khashab, N. M.; Al-
Ghoul, M. Cryst. Growth Des. 2016, 16, 4327−4335.
(39) Seoane, B.; Castellanos, S.; Dikhtiarenko, A.; Kapteijn, F.;
Gascon, J. Coord. Chem. Rev. 2016, 307, 147−187.
(40) Furukawa, S.; Reboul, J.; Diring, S.; Sumida, K.; Kitagawa, S.
Chem. Soc. Rev. 2014, 43, 5700−5734.
(41) Pan, Y.; Heryadi, D.; Zhou, F.; Zhao, L.; Lestari, G.; Su, H.; Lai,
Z. CrystEngComm 2011, 13, 6937−6940.
(42) Jelle, A. A.; Hmadeh, M.; O’Brien, P. G.; Perovic, D. D.; Ozin,
G. A. ChemNanoMat 2016, 2, 1047−1054.
(43) Pegu, R.; Majumdar, K. J.; Talukdar, D. J.; Pratihar, S. RSC Adv.
2014, 4, 33446−33456.
(44) Impert, O.; Katafias, A.; Kita, P.; Mills, A.; Pietkiewicz-Graczyk,
A.; Wrzeszcz, G. Dalton Trans. 2003, 348−353.
(45) Jing, H.-P.; Wang, C.-C.; Zhang, Y.-W.; Wang, P.; Li, R. RSC
Adv. 2014, 4, 54454−54462.
(46) Wang, C.-c.; Jing, H.-p.; Wang, P. J. Mol. Struct. 2014, 1074, 92−
99.

1823 DOI: 10.1021/jacs.7b11589

J. Am. Chem. Soc. 2018, 140, 1812−1823