Porous Coordination Polymer Based on Bipyridinium Carboxylate Linkers with High and Reversible Ammonia Uptake

Article abstract of DOI:10.1021/acs.inorgchem.6b01119

The zwitterionic bipyridinium carboxylate ligand 1,1′-bis(4-carboxyphenyl)-4,4′-bipyridinium (pc1) in the presence of cadmium chloride affords novel porous coordination polymers (PCPs): [Cd₄(pc1)₃Cl₆]·CdCl₄·gues

Full text of DOI:10.1021/acs.inorgchem.6b01119

Article
pubs.acs.org/IC
Porous Coordination Polymer Based on Bipyridinium Carboxylate
Linkers with High and Reversible Ammonia Uptake
Maxime Leroux,† Nicolas Mercier,*,† Magali Allain,† Marie-Claire Dul,† Jens Dittmer,γ
Abdel Hadi Kassiba,γ Jean-Pierre Bellat,§ Guy Weber,§ and Igor Bezverkhyy*,§
†MOLTECH Anjou, UMR-CNRS 6200, Universite
́
d’Angers, 2 Bd Lavoisier, 49045 Angers, France
du Maine, Avenue O. Messiaen, 72085 Le Mans, France
de Bourgogne-Franche Comte, 9 A. Savary, 21078 Dijon, France
γIMMM, UMR-CNRS 6283, Universite
́
§ICB, UMR-CNRS 6303, Universite
́
́
S
* Supporting Information
ABSTRACT: The zwitterionic bipyridinium carboxylate ligand 1,1′-
bis(4-carboxyphenyl)-4,4′-bipyridinium (pc1) in the presence of
cadmium chloride affords novel porous coordination polymers
(PCPs): [Cd (pc1) Cl ]·CdCl ·guest (1) crystallizing in the P31c
̅
4
3
6
4
space group. In the structure, [Cd4Cl6(CO2)6] building units are linked
together by six pc1 ligands, leading to a 3D high-symmetrical network
exhibiting hexagonal channels along the c axis. The walls of this PCP
consist of cationic electron-acceptor bipyridinium units. The PCP 1
reversibly adsorbs H2O and CH3OH up to about 0.1 g/g at saturation
showing the adsorption isotherms characteristic of a moderately
hydrophilic sorbent. Adsorption of ammonia (NH3) follows a different
pattern, reaching an exceptional uptake of 0.39 g/g (22.3 mmol/g) after
the first adsorption cycle. Although the crystalline structure of 1
collapses after the first adsorption, the solid can be regenerated and
maintains the capacity of 0.29 g/g (17 mmol/g) in the following cycles. We found that the high NH3 uptake is due to a
combination of pore filling taking place below 150 h·Pa and chemisorption occurring at higher pressures. The latter process was
shown to involve two phenomena: (i) coordination of NH3 molecules to Cd2+ cations as follows from 113Cd NMR and (ii)
strong donor−acceptor interactions between NH3 molecules and pc1 ligands.
kinds of ligands often lead to coordination polymer materials
with dense networks as a result of the ability of linkers at the N+
1. INTRODUCTION
Porous coordination polymers (PCPs) or metal−organic
frameworks (MOFs), based on redox-active linkers of the
bypiridinium type (also called viologens), have emerged as an
interesting class of porous compounds.1−8 Such linkers usually
consist of functional units bearing carboxylate groups anchored
to the bipyridine core at the N positions, leading to zwitterionic
or anionic bipyridinium carboxylate ligands (Scheme 1). These
−
site to interact with carboxylate groups through CO2 ···N+
electrostatic interactions.9−14 However, an interesting feature of
these materials, as encountered more generally in those based
on bipyridinium units,15−18 are their photo- or thermochromic
properties due to an electron-transfer process from the
carboxylate electron donor toward the bipyridinium electron
acceptor, giving radical species. The color change in such an
activated redox process is due to the different absorption
domains of viologen V2+ or V+ (UV) and V•+ or V• (visible).19
In such a system, a color change can also originate from the
formation of charge-transfer (CT) complexes of the donor and
bipyridinium acceptor (CT band in the visible region). Thus,
the use of such ligands to build porous compounds appears to
be very interesting not only because the redox viologen site is
able to strongly interact with electron-rich guests but also
because the donor (guest)−acceptor (host) interactions can be
accompanied by a color change, which is useful for chemical
sensor applications. Although PCPs based on such ligands are
Scheme 1. Structural Formulas of Some Zwitterionic (A),
Anionic (B), Symmetrical (a), or Asymmetrical (b)
Bipyridinium Carboxylate Type Ligands
Received: May 7, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.6b01119
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
2H2O) ligand was synthesized according to the literature (see the
Supporting Information).33 Thermogravimetric analyses (TGA) were
performed using a TGA-2050 analyzer (TA Instruments) from room
temperature to 1000 °C with a heating rate of 10 °C/min under a
nitrogen flow. Powder X-ray diffraction (PXRD) analyses were
measured at room temperature on a D8 Bruker diffractometer (Cu
Kα, λ = 1.5418 Å) equipped with a linear Vantec superspeed detector
and have shown that all of the observed reflections could be indexed in
the unit cell parameters obtained from single-crystal X-ray diffraction
experiments (Figure S1).
Preparation of the Compound. A mixture of H2(pc1)Cl2·2H2O
(50.5 mg, 0.1 mmol), cadmium chloride (CdCl2; 128.4 mg, 0.7 mmol),
dimethylformamide (1.5 mL), dioxane (1.5 mL), and H2O (0.3 mL)
was heated at 100 °C for 48 h in a 25 mL Teflon-lined stainless steel
autoclave. Then the mixture was slowly cooled to room temperature at
a rate of 2 °C/h. Crystals suitable for X-ray analyses were collected,
washed with ethyl acetate, and air-dried: [Cd4(pc1)3Cl6]·CdCl4·8H2O
(1) of 60% yield based on H2(pc1)Cl2·2H2O. Anal. Calcd for
Cd5Cl10C72H64N6O20 (2249.79): C, 38.43; H, 2.86; N, 3.73. Found: C,
37.90; H, 2.62; N, 3.95.
not easy to stabilize, some of them have been reported quite
recently. Most of those with rigid 2D or 3D networks and
clearly defined channels are based on bipyridinium cores
consisting of one pyridinium cycle and one pyridyl chelating
cycle (Scheme 1, column b).1−6 In contrast, those based on
diquaternized bipyridinium, while better electron-acceptor
units, are more scarce.7,8 In one example, a 3D supramolecular
network results from weakly interacting 1D coordination
polymers.7 In another example, a 3D coordination polymer
that is obtained using the 1,1′-bis(4-carboxyphenyl)-4,4′-
bipyridinium (pc1) ligand displays cavities difficult to access.8
Ammonia (NH3) is an important compound widely used in
different industrial processes but also a very toxic gas even at
low concentration. The search for materials for its detection, on
the one hand, or for its capture from air and safe storage and
transportation, on the other hand, is an important subject from
industrial and environmental points of view. PCPs have been
considered as potential materials for such purposes.3,20−28 In
particular, a bipyridinium-linker-based PCP has been proposed
as a visual colorimetric sensor for NH3 detection via a redox
reaction between the NH3 and viologen units, resulting in a
radical-containing colored material.3 However, the detection
limit has not been investigated.3 Adsorption and storage of
NH3 remain great challenges. If some materials, such as zeolites
or activated carbons, are known to exhibit good adsorption
capacities, the progressive desorption of NH3 over time remains
a challenging issue.29,30 To overcome this problem, porous
structures possessing active sites able to retain NH3 molecules
are of high interest. Two main strategies have been followed in
the field of PCPs: the functionalization of ligands by free
2.2. X-ray Crystallography. X-ray diffraction data were collected
on an Agilent Supernova with Cu Kα radiation (λ = 1.5418 Å). Data
were collected at 180 K from a selected single crystal of 1 under a
nitrogen atmosphere. A summary of crystallographic data and
refinement results for 1 is listed in Table 1. The structure was solved
Table 1. Crystallographic Data for 1
fw (g/mol)
2249.79
space group
P31c
̅
a, Å
18.196(5)
18.196(5)
15.127(5)
90
b, Å
c, Å
21,22
+
Brønsted acid groups such as CO2H or NH3
or the
α, deg
introduction of Lewis acid sites.23−26 In most cases, the Lewis
acid sites are metal ions at nodes of networks, especially Cu2+ as
in MOF-7423 or HKUST-1,25,26 which become metal open sites
upon removal of initially coordinated solvent molecules. A main
disadvantage of such materials is their tendency to collapse
upon adsorption of NH3 as a result of metal−NH3 coordination
and consequently a lack of reversibility and recyclability.25−28 In
the neighboring field of covalent organic frameworks, the
introduction of Brønsted acid groups31 or a high density of
Lewis acidic boron atoms into a framework material has also
been successful for NH3 storage with an uptake of 15 mmol/
g.32 Finally, PCPs with bipyridinium linkers represent another
category of materials possessing Lewis acid sites.
In this context, we report on an unprecedented PCP based
on the pc1 bipyridinium carboxylate linker, [Cd4(pc1)3Cl6]·
CdCl4·guest (1), which exhibits a reversible high NH3 uptake.
The trigonal crystal structure of 1 containing channels running
along the c axis of the hexagonal unit cell will first be described.
Then, the adsorption of methanol (CH3OH) and water (H2O)
by 1 will be described, showing the presence of open pores with
moderately hydrophilic walls. Finally, the adsorption properties
of 1 for NH3 will be fully analyzed, showing that in a low-
pressure range a reversible physisorption occurs, while
chemisorption is observed at higher pressure. This process
results in a loss of sample crystallinity; however, a high NH3
capacity is preserved.
β, deg
90
γ, deg
V, Å3
120
4337(2)
2
Z
obsd reflns [I > 2σ(I)] (Rint)
no. of param
2626 (0.029)
175
R1 [I > 2σ(I)]/wR2 (all data)
0.0441/0.1148
by direct methods and refined on F2 by a full-matrix least-squares
method with anisotropic approximation for all non-hydrogen atoms,
using the SHELX97 package. Three H2O molecules were located in
the channels, while five others were deduced from analysis of the
SQUEEZE routine of PLATON, leading to a total of eight H2O
molecules per formula unit, in good accordance with the TGA
experiment. All non-hydrogen H2O atoms were treated with a riding
model. Absorption was corrected by the program SADABS. A
complete list of crystallographic data, along with the atomic
coordinates, anisotropic displacement parameters, and bond distances
and angles for each compound, is given as a CIF file (CCDC
1473082).
2.3. Electron Paramagnetic Resonance (EPR) and NMR
Measurements. Solid-state NMR experiments have been performed
on a Bruker Avance III 300 MHz WB spectrometer equipped with a
2.5 mm H/X probehead. The magic-angle-spinning (MAS) frequency
was 10 kHz. There is only a little heating by MAS friction at this
frequency. 1H−113Cd cross-polarization (CP) NMR spectra have been
accumulated over 24 K (crystalline sample) and 29 K (after NH3
adsorption) repetitions with contact times of 4 and 7 ms, respectively.
For the sample after desorption, the only successful acquisition
strategy for 113Cd was direct excitation with fast recycling (0.5 s) to 64
K repetitions. The DEPTH pulse sequence was applied for
background suppression. The spectra have been referenced to an
aqueous solution of Cd(ClO4)2.
2. EXPERIMENTAL SECTION
2.1. Synthesis and Characterization. Materials and General
Methods. All starting materials were of analytical grade and obtained
from commercial sources without further purification. A 4,4′-
bis(carboxyphenyl)bipyridinium dichloride dihydrate (H2pc1Cl2·
EPR experiments were carried out at room temperature on a Bruker
EMX spectrometer working at X band (9.51 GHz). Instrumental
B
DOI: 10.1021/acs.inorgchem.6b01119
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
parameters were fixed to 100 kHz and 0.5 G respectively for the
frequency and magnetic field modulation, while 2 mW was used for
the microwave power. These parameters allow recording of the EPR
signals with good resolution and prevention of any distortion that may
occur from saturation phenomena induced by a high-spin concen-
tration. Dry 2,2-diphenyl-1-picrylhydrazyl radicals were used to
evaluate the g factors in the investigated samples.
2.4. Adsorption Measurements. Adsorption−desorption iso-
therms of CH3OH, H2O, and NH3 on 1 were measured at 298 K using
a home-built McBain-type thermobalance. Before measurements the
samples (ca. 15 mg) were outgassed at 423 K under vacuum (10−6 h·
Pa) overnight. The detailed description of the experimental procedure
can be found elsewhere.6 The estimated experimental error on the
adsorbed amount was about 1 mg/g. The accuracy of the pressure
measurement was 1%, and the temperature was maintained within 1 K.
The adsorption isotherm of carbon dioxide (CO2) in the range of 0.2−
50 bar at 298 K was measured using a Rubotherm magnetic
suspension balance.
Figure 2. General view of the structure of 1 showing the hexagonal
channels running along the c axis (the H2O molecules in the pores
have been omitted for clarity). Inset: View of one hexagonal channel
incorporating H2O molecules that interact with the walls of the
framework through hydrogen bonding.
3. RESULTS AND DISCUSSION
3.1. Crystal Structure Description. The asymmetric unit
of the structure of 1 contains three Cd2+ cations, one pc1
molecule, three chlorides, and one oxygen atom assigned to a
H2O molecule. The Cd(1) atom, located at the crossing of 3-
fold and 2-fold axes (3.2 special position), and the Cd(2) atom,
located on a 2-fold axis, define a tetrameric building unit, as
shown in Figure 1. While Cd(1) is surrounded by six chloride
weak hydrogen bonds between H2O oxygen atoms and slightly
acidic hydrogen atoms of the phenyl rings of pc1 ligands borne
by the carbon atoms in the β positions of N+ sites (Figure 3).
Figure 3. Details of the structure of 1 showing the environment of a
2−
bipyridinum carboxylate (pc1) ligand with neighboring CdCl4
−
anions, N+···Cl [d = 3.408(8) Å] or N+···O(CO2 ) [d = 3.579(8)
Å] contacts, as well as (C)H···O(H2O) hydrogen bonds [d = 2.65(1)
Å] between pc1 and the guest H2O molecules present in the pores.
The other solvent molecules in the center of the channels may
be disordered, which did not allow us their location. After
removal of all of the solvent molecules, the potential of the void
space has been estimated to 10% of the unit volume by the
SQUEEZE routine of PLATON. Using the distance across the
pore between the hydrogen atoms of the bipyridimium units
(7.5 Å) and the value of van der Waals radius of hydrogen (1.2
Å), the accessible pore diameter was calculated to be close to 5
Å.
Of great interest in the field of viologen-based solid-state
materials is analysis of the intermolecular interactions between
the electron-acceptor bipyridinium unit and electron-donor
groups in order to tentatively explain some optical properties
such as colored samples due to CT bands or photo- or
thermochromism. In the particular case of PCPs, the
accessibility of pyridinium cycles by guest solvent molecules
is especially interesting because desolvation can further afford
pyridinium open sites, which would react very rapidly with
electron-donor guest molecules (chemical sensor application).
Unfortunately, in 1, as in most of bipyridinium-based
structures, the viologen core does quite have strong intra-
framework interactions, on the one hand with a chloride
Figure 1. View of the tetrameric inorganic building unit bearing six
pc1 molecules in the trigonal structure of 1 and a disordered CdCl4
tetrahedral anion (inset).
2−
anions in a pseudooctahedral geometry [d = 2.641(1) Å],34,35
Cd(2) is bound to two chlorides through edge-sharing with
Cd(1)Cl6 octahedra and four oxygen atoms of carboxylate
groups belonging to two pc1 ligands. This results in
[Cd4Cl6(CO2)6] building units that are linked together by six
pc1 ligands to construct the porous 3D network. The positive
charge of the [Cd4Cl6(pc1)3]2+ framework is balanced by the
2−
presence of isolated CdCl4 anions located between two
consecutive tetrameric cores along the c axis. Let us note that
this dianion is disordered over two positions with d[Cd(3)−Cl]
= 2.447(4) Å (Cl2) and 2.427(9) Å (Cl3) (Figure 1, inset).
Figure 2 shows a general view of the structure of 1 along the
c axis where the channels with hexagonal windows are clearly
defined. The solvent molecules present in the pores have not all
been located. Only one oxygen atom, assigned to a H2O
molecule, has been refined in a general position with half-
occupation rate (Figure 2). The 12 symmetry-related oxygen
atoms line the walls of the channels, indicating the existence of
C
DOI: 10.1021/acs.inorgchem.6b01119
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
2−
belonging to a CdCl4 anion [d = 3.408(8) Å], and on the
other hand with an oxygen atom of a carboxylate group also
binded to Cd(2) [d = 3.579(8) Å] (Figure 3). However, the
high concentration of the bipyridinium units in the structure of
1 must be noted. This is a consequence of the presence of the
small sized anions of Cl− in 1, while bulky carboxylate-type
organic ligands counterbalance the positive charge of the metal
cations in other such materials. Thus, a high concentration of
the dipolar moments (N+·Cl− or N+···O− types) are present in
the framework and can interact through electrostatic
interactions with polar guest molecules such as H2O, HCOH,
or NH3. These interactions probably come in addition to
hydrogen bonding involving some hydrogen atoms of the pc1
molecules (Figure 3), explaining, for instance, the presence of
H2O molecules lining the walls of the channels (Figure 2).
3.2. Adsorption of CH3OH and H2O. The pores of 1 are
not accessible for N2 at 77 K. This could be either due to the
so-called “gate effect”36−39 taking place in flexible structures or
due to kinetic limitations for the diffusion of nitrogen molecules
in small pores at cryogenic temperature.40,41 Given the
nonflexible nature of 1 (see below), we consider that the
second effect is in action. In contrast, the solid readily adsorbs
CH3OH, H2O, and CO2 at 298 K (Figure 4). The fully
reversible character of the adsorption−desorption process for
all molecules suggests that only physisorption occurs in 1 under
these conditions. Such behavior is different from that of
CH3OH chemisorption observed for other bipyridinium-based
phases described in our previous work.6 While the density of
adsorbed CO2 at 298 K is poorly defined,42 the density of
adsorbed H2O and CH3OH phases can be assumed to be equal
to the density of the bulk liquid at saturation, and their
isotherms can thus be used to estimate the accessible pore
volume of 1. The value determined from the maximum amount
of H2O and CH3OH adsorbed (around 0.1 g/g) is close to 0.10
or 0.13 cm3/g, respectively. Slightly higher pore volume probed
by CH3OH despite the smaller size of the H2O molecule (2.64
vs 3.64 Å) can be attributed to a higher affinity of the material
toward CH3OH (see below).
The shape of the adsorption isotherms of H2O, CH3OH, and
CO2 is different from that of the type I isotherms, showing a
sharp knee at very low pressure.43 In contrast, 1 exhibits S-
shaped isotherms with negligible adsorption in the low-pressure
range (clearly seen on a logarithmic relative pressure scale;
Figure 4). Such behavior in PCPs can be due to either a gate-
opening phenomenon or a moderate strength of the interaction
between the molecules and pore walls.
The first explanation implies that during degassing the
framework structure contracts, provoking shrinking of the
pores, which become thus inaccessible to the adsorbate
molecules. At a certain threshold pressure, the pores open
enough to allow access of the molecules, giving rise to
adsorption. To check this hypothesis, we recorded the PXRD
pattern of the degassed sample 1 sealed in a capillary. A full
profile fitting of the pattern was done using the Le Bail
method44 (Figure S2). It appears that the pattern can be
successfully fitted in the initial P31c space group, and the cell
̅
parameters of the degassed sample (a = b = 18.175 Å and c =
15.012 Å − RT) are close to those of the initial sample
measured in air (a = b = 18.196 Å and c = 15.127 Å −180 K).
This similarity allows one to rule out any pore shrinking at 298
K in degassed 1 and therefore to conclude that a gate opening
is not a plausible explanation of the observed isotherm shape. It
can thus be due to a moderate affinity between the adsorbate
molecules and pore walls.
To estimate the strength of the interaction between the
studied adsorbates and the surface of 1, we fitted the adsorption
isotherms with the Dubinin−Astakhov (DA) equation, which is
well adapted for the description of the adsorption in
microporous sorbents:45
W = W exp −(A/E)n
[
]
0
where A = RT ln(P/P) (R, gas constant; T, temperature; P,
s
pressure; Ps, saturation pressure at 298 K), E is the
characteristic interaction energy (kJ/mol), W is the adsorbed
amount, W0 is the maximum amount adsorbed in the
micropores, and n is the heterogeneity parameter (1 < n <
5). Figure 4 shows that the DA equation gives a fair description
of the S-type shape of all isotherms. The characteristic energies
are equal to 6 kJ/mol for H2O and 7.6 kJ/mol for CH3OH. A
higher value obtained for CH3OH is in line with a larger pore
volume probed by this molecule. An even larger value of E
obtained for CO2 (10.5 kJ/mol) can be due to a high quadruple
moment of this molecule.
The value of the characteristic energy for H2O allows one to
assess the hydrophilicity of a microporous sorbent.46 From this
criterion, our material is closer to a hydrophilic NaY zeolite (10
kJ/mol) than to a hydrophobic activated carbon (1.3 kJ/mol).46
More specifically, the hydrophilicity of 1 can be compared to
that of other PCPs. This can be done using the value of P/Ps at
which the solid adsorbs 50% of its maximum loading.47 The
corresponding value for 1 (P/Ps = 0.1; see Figure 4) is similar
to the values observed for PCPs containing −NH2 groups and
Figure 4. Adsorption isotherms of CH3OH (blue), H2O (red), and
CO2 (green) on 1 at 298 K on linear (top) and logarithmic (bottom)
scales: full symbols, adsorption; empty symbols, desorption. The
connecting lines correspond to the fit with the DA equation (see the
text for details).
D
DOI: 10.1021/acs.inorgchem.6b01119
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
considered to be hydrophilic. PCPs have been considered as
promising materials for H2O adsorption for different
applications such as heat exchangers or the delivery of H2O
in normal conditions,48−53 because they can fulfill the expected
criteria for such a purpose: an easy release of H2O, a high H2O
uptake capacity, a good cycling performance, etc. The PCP
materials based on bipyridinium carboxylate linkers can be of
interest in this field.
(Figure 6). Such a study is quite rare in this field.55 The
spectrum of the crystalline pristine sample (red line) shows
3.3. Adsorption of NH3. The adsorption and desorption
isotherms of NH3 on 1 at 298 K reveal a high capacity of the
solid toward NH3 (Figure 5). The maximum adsorbed amount
Figure 6. Solid-state 1H−113Cd CP NMR spectra of the pristine
sample of 1 (red) and the sample containing 8.2 mmol/g of NH3
(green) and a direct excitation 113Cd NMR spectrum of 1 saturated
with NH3 and then regenerated at 473 K (blue). Asterisks denote
spinning side bands.
three peaks corresponding to cadmium atoms occupying
different crystallographic positions. We tentatively assign the
sharp downfield signal at 475 ppm to Cd(2) with Cd(2)Cl2O4
coordination, the broad signal at 302 ppm to the disordered
Cd(3)Cl42−, and the upfield signal at 133−176 ppm to
Cd(1)Cl6 octahedra (for comparison, nonhydrated CdCl2 was
found at 183 ppm56). The latter signal also shows a certain
disorder. Note that CP spectra are not quantitative, and the
signal intensity depends on the vicinity of the hydrogen atom if
the molecular mobility is similar. The solid containing 8.2
mmol/g of NH3, which was obtained after desorption at 298 K,
yields, in contrast to the pristine sample, only one very intense
peak whose position (262 ppm) is different from those of all
three previous signals (green line). The local environment of
the cadmium ions has changed, and despite the transformation
to an amorphous phase, it has become more homogeneous.
Furthermore, hydrogen atoms must be relatively close because
the CP transfer is efficient. The chemical shift is close to those
of Cd(NH3)6Cl2 in 50:50 NH3/D2O, which have been
determined as 274 and 287 ppm depending on the
concentration.57 (Note that in ref 57 the sense of the chemical
shift is inverted.) These findings support the hypothesis that
NH3 molecules are coordinated to cadmium cations. A
complete elimination of NH3 after heating at 473 K results in
a shift back toward a mixed oxygen−halogen cadmium
environment, as evidenced by a strong broadening of the
NMR signal, which reflects an extremely inhomogeneous
environment of the cadmium ion (blue line). For this sample,
CP transfer from hydrogen did not prove successful, potentially
because of the longer distances of cadmium to hydrogen. In
contrast, the 13C NMR spectrum does not change very much
upon desorption (not shown). This difference in the sensitivity
to the presence of NH3 is another indication that the cadmium
ions play a central role in the absorption of NH3.
Figure 5. Adsorption and desorption of NH3 on 1 at 298 K: full
symbols, adsorption; empty symbols, desorption. Inset: low-pressure
range. The connecting lines are guides for the eye.
of 22.8 mmol/g at 900 h·Pa [corresponding to 0.39 g(NH3)/g]
is one of the highest values reported to date. The isotherm has
a rather complex shape and can be divided into two distinct
regions.
In the low-pressure range (<150 h·Pa; see the inset), the
shape of the isotherm is similar to that of H2O (see Figure 4).
We have also found that, if the adsorption is stopped below 150
h·Pa, the adsorbed NH3 can be completely desorbed under
vacuum at 298 K, and the crystalline structure of the sample
after such an adsorption−desorption cycle is preserved. This
observation shows that this part of the isotherm is fully
reversible, and below 150 h·Pa, the pore filling occurs via the
physisorption of NH3 molecules.
At pressure higher than 150 h·Pa, the adsorbed amount
strongly increases up to 22.3 mmol/g at 900 h·Pa due to
chemisorption of NH3 in 1. This is evidenced by a pronounced
hysteresis accompanied by retention of ca. 8.2 mmol(NH3)/g
even after the sample is maintained under a secondary vacuum
for 48 h at 298 K (Figure 5). A complete desorption of NH3 is
only possible after heating for 1 h at 473 K (see below). The
conclusion about the chemisorption is also confirmed by the
loss of crystallinity in the solid after NH3 adsorption (P > 150
h·Pa) and desorption at either 298 or 473 K (no diffraction
lines in the PXRD patterns). Decomposition of the crystalline
structure of 1 allows one to suppose that the interaction with
NH3 molecules at pressure higher than 150 h·Pa occurs mainly
through their coordination to cadmium cations, which are
known to be hexacoordinated in NH3 complexes.54 This ability
of cadmium allows one to explain a high capacity of 1:
coordination of each cadmium by six NH3 molecules would
yield the loading of 14.25 mmol of NH3 per g of 1.
The conclusions based on the results of 113Cd NMR are
confirmed by Fourier transform infrared (FTIR) spectroscopy.
The spectrum of the sample containing 8.2 mmol/g of NH3
shows the presence of its characteristic bands (Figure S5), but
in addition, some bands in the range 1300−1700 cm−1 are
The evolution of the coordination sphere of cadmium upon
absorption and desorption of NH3 was studied by 113Cd NMR
E
DOI: 10.1021/acs.inorgchem.6b01119
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
2+
shifted (Figure S6). Comparison with the spectra of other
MOFs58,59 allowed us to identify them as symmetric (νs) and
asymmetric (νas) stretching modes of carboxylates. The
splitting between these bands increases from 154 cm−1 in
pristine 1 to 181 cm−1 in the sample containing 8.2 mmol/g of
NH3 (Figure S6). This increase is indicative of the trans-
formation of bidentate carboxylates into ionic ones.60 This
finding confirms therefore our conclusion about the appearance
of NH3 in the coordination of cadmium atoms, which
transforms the carboxylates from bidentate (see the structure
of 1) into “free” ionic groups. Upon NH3 elimination, the
carboxylates become again coordinated to cadmium, as follows
from the close similarity between the spectrum of 1 and those
of the regenerated samples (Figure S6).
above). At low NH3 pressure, Cd(NH3)6 can also partially
decompose because the stoichiometric complex is known to be
stable only in the presence of gaseous NH3.61,62 The NH3
molecules retained in the solid after 48 h of evacuation (8.2
mmol/g) are therefore coordinated to cadmium (3−4 per
cadmium) and/or interacting with the bipyridinium units.
These strongly bound molecules can be fully eliminated only
upon heating up to 473 K (Figure S7).
Reusability is an important characteristic of an adsorbent.
Therefore, the sample (1) was tested during three successive
cycles comprising NH3 adsorption at 298 K and regeneration
under vacuum at 473 K (Figure 8). Surprisingly, despite the
In addition to coordination by Cd2+ cations, NH3 molecules
can be involved in redox reactions with bipyridinium-based
ligands.3 The existence of such a reaction is confirmed by
electron transfer from NH3 to pc1 resulting in the formation of
bipyridinium radicals. This process is suggested by the color
change of 1 from brown to black (see the UV−vis spectra in
Figure S4). Direct evidence of the formation of the
bipyridinium radicals was provided by EPR spectroscopy
(Figure 7). The spectrum of the as-synthesized sample contains
Figure 8. Three successive adsorption isotherms of NH3 on 1 at 298
K: blue, first; red, second; brown, third. After each adsorption, the
sample was regenerated at 473 K under vacuum. The connecting lines
are guides for the eye.
amorphous nature of the solid obtained after the first cycle, we
found that it is still able to adsorb an important volume of NH3
(Figure 8). Indeed, the maximal adsorbed amount at 900 h·Pa
is still 17.6 mmol/g after the first adsorption−regeneration
cycle. Moreover, this value remains almost constant during the
next cycle. The observed drop of the capacity between the first
and second adsorption cycles of 5.2 mmol/g [from 22.8 to 17.6
mmol(NH3)/g] is close to the amount adsorbed in the pores of
pristine 1 (∼5 mmol/g; inset in Figure 5). This fact suggests
that the decrease of the capacity observed after the first
adsorption is mostly due to the disappearance of pores
provoked by the structure collapse. It is worth noting that
the maximum adsorbed amount in the second cycle exceeds the
number of NH3 molecules that can be coordinated by cadmium
(14.25 mmol/g). These additional molecules probably interact
with the positively charged bipyridinium-based ligands. Also,
one cannot exclude the existence of small irregular pores in the
amorphous solid that are able to accommodate NH3 molecules.
Figure 7. EPR spectra of pristine 1 and of a sample containing 8.2
mmol/g of NH3.
only a low-intensity signal originating from the EPR cavity. In
contrast, after NH3 adsorption [the sample containing 8.2
mmol(NH3)/g], a clear signal is observed, revealing a quite
high unpaired spin concentration (Figure 7). With such a
concentration, the occurrence of exchange interactions explains
the quasi-isotropic signal with g = 2.0035 and an EPR narrow
line width of about 1.5 G. Moreover, the absence of hyperfine
coupling between the unpaired spins and nitrogen nuclei from
the bipyridinium units or with chlorine nuclei (superhyperfine
coupling) is probably caused by the mobility of the unpaired
electrons in the host structure. Electron transfer from NH3 to
bipyridinium moieties could produce some species containing
nitrogen atoms with lower oxidation degree than in NH3.
Given the NMR and FTIR data, we can interpret the
desorption behavior of the sample after full saturation in the
following way. The sample saturated at 900 h·Pa of NH3 (22.3
mmol/g) contains three types of NH3 molecules: (i)
physisorbed (∼5 mmol/g); (ii) coordinated to cadmium
(14.25 mmol/g for six NH3 per cadmium); (iii) interacting
with bipyridinium units (remaining 3.05 mmol/g). Upon
decreasing NH3 pressure, the physisorbed molecules are
eliminated because of the reversibility of this interaction (see
4. CONCLUSIONS
In this work, the synthesis, crystal structure, and adsorption
properties of a novel PCP based on the zwitterionic
bipyridinium carboxylate ligand pc1 are described. In the
structure of 1, hexagonal channels run along the c axis. We have
first shown that H2O or CH3OH can be reversibly adsorbed,
filling the channels with maximum uptake values at 298 K of
5.33 and 3.16 mmol/g, respectively. The observed shape of the
adsorption isotherms is shown to be due to a moderate
interaction strength between the guest molecules and walls of
the channels. In contrast, 1 exhibits high NH3 uptake among
the best in the field of PCPs. A value of 22.3 mmol/g is reached
F
DOI: 10.1021/acs.inorgchem.6b01119
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(15) Wang, M.-S.; Xu, G.; Zhang, Z.-J.; Guo, G.-C. Chem. Commun.
2010, 46, 361−376.
at P = 900 h·Pa after the first adsorption cycle. NH3 molecules
are first physisorbed in channels at low pressure (<150 h·Pa)
and then chemisorbed at higher pressure. The chemisorption
process involves, on the one hand, Cd2+ centers, as highlighted
by solid-state 113Cd NMR, and, on the other hand, the
bipyridinium units. Although the sample becomes amorphous
after the first adsorption cycle, the repeatability of the
adsorption−desorption process has been proven, with an
average NH3 uptake of ca. 17 mmol/g.
(16) Mercier, N. Eur. J. Inorg. Chem. 2013, 2013, 19−31.
(17) Leblanc, N.; Bi, W.; Mercier, N.; Auban-Senzier, P.; Pasquier, C.
Inorg. Chem. 2010, 49, 5824.
(18) Leblanc, N.; Allain, M.; Mercier, N.; Sanguinet, L. Cryst. Growth
Des. 2011, 11, 2064−2069.
(19) Monk, P. M. S. The Viologens: Physicochemical Properties,
Synthesis, and Application of the Salt of 4,4′-Bipyridine; Wiley: New
York, 1998.
(20) Britt, D.; Tranchemontagne, D.; Yaghi, O. M. Proc. Natl. Acad.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.inorg-
chem.6b01119.
Sci. U. S. A. 2008, 105, 11623−11627.
■
(21) Morris, W.; Doonan, C. J.; Yaghi, O. M. Inorg. Chem. 2011, 50,
6853−6855.
S
(22) Spanopoulos, I.; Xydias, P.; Malliakas, C. D.; Trikalitis, P. N.
Inorg. Chem. 2013, 52, 855−862.
(23) Katz, M. J.; Howarth, A. J.; Moghadam, P. Z.; DeCoste, J. B.;
Snurr, R. Q.; Hupp, J. T.; Farha, O. K. Dalton Trans. 2016, 45, 4150−
4153.
Synthesis procedures, summary of the X-ray crystal data,
PXRD patterns of 1 and dehydrated 1, TGA, UV−vis
and FTIR spectra, and desorption curve of NH3 under
heating in a vacuum (PDF)
(24) Kajiwara, T.; Higuchi, M.; Yuasa, A.; Higashimura, H.; Kitagawa,
S. Chem. Commun. 2013, 49, 10459−10461.
(25) Petit, C.; Huang, L.; Jagiello, J.; Kenvin, J.; Gubbins, K. E.;
Bandosz, T. J. Langmuir 2011, 27, 13043−13051.
(26) Peterson, G. W.; Wagner, G. W.; Balboa, A.; Mahle, J.; Sewell,
T.; Karwacki, C. J. J. Phys. Chem. C 2009, 113, 13906−13917.
(27) Saha, D.; Deng, S. J. Colloid Interface Sci. 2010, 348, 615−620.
(28) Kajiwara, T.; Higuchi, M.; Watanabe, D.; Higashimura, H.;
Yamada, T.; Kitagawa, H. Chem. - Eur. J. 2014, 20, 15611−15617.
(29) Petit, C.; Karwacki, C.; Peterson, G.; Bandosz, T. J. J. Phys.
Chem. C 2007, 111, 12705−12714.
CIF file (CIF)
AUTHOR INFORMATION
Corresponding Authors
*E-mail: nicolas.mercier@univ-angers.fr.
*E-mail: igor.bezverkhyy@u-bourgogne.fr.
Author Contributions
The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript.
■
(30) Petit, C.; Bandosz, T. J. J. Phys. Chem. C 2007, 111, 16445−
16452.
(31) Van Humbeck, J. F.; McDonald, T. M.; Jing, X.; Wiers, B. M.;
Zhu, G.; Long, J. R. J. Am. Chem. Soc. 2014, 136, 2432−2440.
(32) Doonan, C. J.; Tranchemontagne, D. J.; Glover, T. G.; Hunt, J.
R.; Yaghi, O. M. Nat. Chem. 2010, 2, 235−237.
(33) Das, R. R.; Lee, J.-M.; Noh, C.-H.; Jeon, S.-J.U.S. Patent Appl.
Publ. 20120176658 A1 20120712, 2012.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
M.L. thanks the University of Angers for a Ph.D. grant.
■
(34) Banerjee, S.; Ghosh, A.; Wu, B.; Lassahn, P.-G.; Janiak, C.
Polyhedron 2005, 24, 593−599.
(35) Banerjee, S.; Lassahn, P.-G.; Janiak, C.; Ghosh, A. Polyhedron
2005, 24, 2963−2971.
(36) Mondal, S. S.; Bhunia, A.; Kelling, A.; Schilde, U.; Janiak, C.;
Holdt, H.-J. Chem. Commun. 2014, 50, 5441−5443.
(37) Mondal, S. S.; Bhunia, A.; Kelling, A.; Schilde, U.; Janiak, C.;
Holdt, H.-J. J. Am. Chem. Soc. 2014, 136, 44−47.
(38) Mondal, S. S.; Dey, S.; Baburin, I. A.; Kelling, A.; Schilde, U.;
Seifert, G.; Janiak, C.; Holdt, H.-J. CrystEngComm 2013, 15, 9394−
9399.
REFERENCES
■
(1) Higuchi, M.; Tanaka, D.; Horike, S.; Sakamoto, H.; Nakamura,
K.; Takashima, Y.; Hijikata, Y.; Yanai, N.; Kim, J.; Kato, K.; Kubota, Y.;
Takata, M.; Kitagawa, S. J. Am. Chem. Soc. 2009, 131, 10336−10337.
(2) Kanoo, P.; Matsuda, R.; Sato, H.; Li, L.; Jeon, H. J.; Kitagawa, S.
Inorg. Chem. 2013, 52, 10735−10737.
(3) Tan, B.; Chen, C.; Cai, L.-X.; Zhang, Y.-J.; Huang, X.-Y.; Zhang, J.
Inorg. Chem. 2015, 54, 3456−3461.
(4) Yao, Q.-X.; Pan, L.; Jin, X.-H.; Li, J.; Ju, Z.-F.; Zhang, J. Chem. -
Eur. J. 2009, 15, 11890−11897.
(5) Lin, J.-B; Shimizu, G. K. H. Inorg. Chem. Front. 2014, 1, 302−305.
(6) Toma, O.; Mercier, N.; Allain, M.; Kassiba, A. A.; Bellat, J.-P.;
Weber, G.; Bezverkhyy, I. Inorg. Chem. 2015, 54, 8923−8930.
(7) Aulakh, D.; Varghese, J. R.; Wriedt, M. Inorg. Chem. 2015, 54,
1756−1764.
(8) Liu, J.-J.; Guan, Y.-F.; Lin, M.-J.; Huang, C.-C.; Dai, W.-X. Cryst.
Growth Des. 2015, 15, 5040−5046.
(9) Yao, Q. X.; Ju, Z. F.; Jin, X. H.; Zhang, J. Inorg. Chem. 2009, 48,
1266−1268.
(10) Sun, J.-K.; Wang, P.; Chen, C.; Zhou, X.-J.; Wu, L.-M.; Zhang,
Y.-F.; Zhang, J. Dalton Trans. 2012, 41, 13441−13446.
(11) Sun, J.-K.; Wang, P.; Yao, Q.-X.; Chen, Y.-J.; Li, Z.-H.; Zhang,
Y.-F.; Wu, L.-M.; Zhang, J. J. Mater. Chem. 2012, 22, 12212−12219.
(12) Sun, J.-K.; Jin, X.-H.; Cai, L.-X.; Zhang, J. J. Mater. Chem. 2011,
21, 17667−17672.
(13) Jin, X.-H.; Sun, J.-K.; Xu, X.-M.; Li, Z.-H.; Zhang, J. Chem.
Commun. 2010, 46, 4695−4697.
(14) Li, H.-Y.; Wei, Y.-L.; Dong, X.-Y.; Zang, S.-Q.; Mak, T. C. Chem.
Mater. 2015, 27, 1327−1331.
(39) Mondal, S. S.; Bhunia, A.; Baburin, I. A.; Jager, C.; Kelling, A.;
̈
Schilde, U.; Janiak, C.; Holdt, H.-J. Chem. Commun. 2013, 49, 7567−
7570.
(40) Bhunia, A.; Boldog, I.; Moller, A.; Janiak, C. J. Mater. Chem. A
̈
2013, 1, 14990−14999.
(41) Zhao, T.; Cuignet, L.; Dîrtu, M. M.; Wolff, M.; Spasojevic, V.;
Boldog, I.; Rotaru, A.; Garcia, Y.; Janiak, C. J. Mater. Chem. C 2015, 3,
7802−7812.
(42) Cazorla-Amoros, D.; Alcaniz-Monge, J.; Linares-Solano, A.
Langmuir 1996, 12, 2820−2824.
(43) Rouquerol, F.; Rouquerol, J.; Sing, K. S. W.; Llewellyn, P.;
Maurin, G. Adsorption by Powders and Porous Solids; Academic Press:
Oxford, U.K., 2014.
(44) Le Bail, A.; Duroy, H.; Fourquet, J. L. Mater. Res. Bull. 1988, 23,
447−452.
(45) Dubinin, M. M. Carbon 1989, 27, 457−467.
(46) Pires, J.; Pinto, M. L.; Carvalho, A.; de Carvalho, M. B.
Adsorption 2003, 9, 303−309.
(47) Canivet, J.; Fateeva, A.; Guo, Y.; Coasne, B.; Farrusseng, D.
Chem. Soc. Rev. 2014, 43, 5594−5617.
G
DOI: 10.1021/acs.inorgchem.6b01119
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(48) Henninger, S. K.; Habib, H. A.; Janiak, C. J. Am. Chem. Soc.
2009, 131, 2776−2777.
(49) Jeremias, F.; Khutia, A.; Henninger, S. K.; Janiak, C. J. Mater.
Chem. 2012, 22, 10148−10151.
(50) Khutia, A.; Rammelberg, H. U.; Schmidt, T.; Henninger, S.;
Janiak, C. Chem. Mater. 2013, 25, 790−798.
(51) Jeremias, F.; Lozan, V.; Henninger, S. K.; Janiak, C. Dalton
Trans. 2013, 42, 15967−15973.
(52) Jeremias, F.; Frohlich, D.; Janiak, C.; Henninger, S. K. RSC Adv.
̈
2014, 4, 24073−24082.
(53) Frohlich, D.; Henninger, S. K.; Janiak, C. Dalton Trans. 2014,
̈
43, 15300−15304.
(54) Nilsson, K. B.; Eriksson, L.; Kessler, V. G.; Persson, I. J. Mol. Liq.
2007, 131−132, 113−120.
(55) Haldar, R.; Inukai, M.; Horike, S.; Uemura, K.; Kitagawa, S.;
Maji, T. K. Inorg. Chem. 2016, 55, 4166−4172.
(56) Sakida, S.; Shojiya, M.; Kawamoto, Y. Solid State Commun. 2000,
115, 553−558.
(57) Cardin, A. D.; Ellis, P. D.; Odom, J. D.; Howard, J. W., Jr. J. Am.
Chem. Soc. 1975, 97, 1672−1679.
(58) Bernini, M. C.; Garro, J. C.; Brusau, E. V.; Narda, G. E.; Varetti,
E. L. J. Mol. Struct. 2008, 888, 113−123.
́
(59) Bernini, M. C.; Gandara, F.; Iglesias, M.; Snejko, N.; Gutierrez-
Puebla, E.; Brusau, E. V.; Narda, G. E.; Monge, M. A. Chem. - Eur. J.
2009, 15, 4896−4905.
(60) Deacon, G. B.; Phillips, R. J. Coord. Chem. Rev. 1980, 33, 227−
250.
(61) Jenkins, T. E.; Ferris, L. T. H.; Bates, A. R.; Gillard, R. D. J. Phys.
C: Solid State Phys. 1977, 10, L521−L526.
(62) Migdal-Mikuli, A.; Mikuli, E.; Zabinska, K. J. Phys. C: Solid State
Phys. 1982, 15, 6565−6571.
H
DOI: 10.1021/acs.inorgchem.6b01119
Inorg. Chem. XXXX, XXX, XXX−XXX