Anhydrous proton conduction in porous organic networks

Article abstract of DOI:10.1039/C8TA07822J

Solid electrolyte separators within fuel cells enable efficient charge transport and prevent a mass bypass between the two half cells. Hydrated systems, like Nafion, reach unprecedented proton conductivities at ambient temperatures, but the demanding humidity management prevents their use beyond 80 °C, hence limiting the efficiency of current polymer-based systems. As such, water free and chemically inert, solid materials with excellent conductivities between 100 °C and 200 °C, are of high interest. A promising approach is the incorporation of heavier amphoteric molecules into micro- and mesoporous frameworks. Stronger host-guest interactions allow for higher temperatures, while still maintaining sufficient mobility and efficient transport pathways. Here, we present a systematic study investigating the influence of porosity, framework topology and dimensionality as well as framework functionality and charge carrier uptake on the proton conductivity for six porous organic networks (PONs) loaded with imidazole via gas phase adsorption. The resulting materials were thoroughly characterized by multinuclear NMR and IR spectroscopy and physisorption as well as powder X-ray diffraction and DSC experiments, revealing homogeneous distribution of the amphoteric guests within the pore structure. Electrochemical impedance spectroscopy up to 130 °C revealed remarkable conductivities of up to 10^?3 S cm^?1 under anhydrous conditions. We found 3D networks to favour high imidazole loading leading to high proton conductivities based on the Grotthuss mechanism. In contrast, 2D networks showed a lower guest molecule uptake and thus lower proton conductivities, which were governed by vehicle transport. Additional acid/base functionalities within the frameworks seem to have a negative effect on the proton conduction.

Full text of DOI:10.1039/C8TA07822J

Journal of
Materials Chemistry A
View Article Online
View Journal | View Issue
PAPER
Anhydrous proton conduction in porous organic
networks†
Cite this: J. Mater. Chem. A, 2018, 6,
1542
2
a
a
b
and J. Senker *a
C. Klumpen, S. Winterstein, G. Papastavrou
Solid electrolyte separators within fuel cells enable efficient charge transport and prevent a mass bypass
between the two half cells. Hydrated systems, like Nafion, reach unprecedented proton conductivities at
ꢀ
ambient temperatures, but the demanding humidity management prevents their use beyond 80 C, hence
limiting the efficiency of current polymer-based systems. As such, water free and chemically inert, solid
ꢀ
ꢀ
materials with excellent conductivities between 100 C and 200 C, are of high interest. A promising
approach is the incorporation of heavier amphoteric molecules into micro- and mesoporous frameworks.
Stronger host–guest interactions allow for higher temperatures, while still maintaining sufficient mobility
and efficient transport pathways. Here, we present a systematic study investigating the influence of
porosity, framework topology and dimensionality as well as framework functionality and charge carrier
uptake on the proton conductivity for six porous organic networks (PONs) loaded with imidazole via gas
phase adsorption. The resulting materials were thoroughly characterized by multinuclear NMR and IR
spectroscopy and physisorption as well as powder X-ray diffraction and DSC experiments, revealing
homogeneous distribution of the amphoteric guests within the pore structure. Electrochemical impedance
ꢀ
ꢁ3
ꢁ1
spectroscopy up to 130 C revealed remarkable conductivities of up to 10 S cm under anhydrous
conditions. We found 3D networks to favour high imidazole loading leading to high proton conductivities
based on the Grotthuss mechanism. In contrast, 2D networks showed a lower guest molecule uptake and
thus lower proton conductivities, which were governed by vehicle transport. Additional acid/base
functionalities within the frameworks seem to have a negative effect on the proton conduction.
Received 12th August 2018
Accepted 5th October 2018
DOI: 10.1039/c8ta07822j
rsc.li/materials-a
electrochemical stability, are required for the development of
novel membrane materials.
Introduction
3
With respect to the challenges of global warming and an
Despite its nearly 40 year-old history, the peruorosulfonic
increasing world population, fuel cell (FC) technology is among acid Naon® is still state of the art for FC membranes and one of
the most promising solutions for future energy supply as it the few materials that has been utilized in commercial products.
1
provides an efficient and environmentally friendly process. By Naon provides excellent proton conductivity and exhibits a long
converting chemical energy into electricity, fuel cells operating lifetime under operational conditions. However, the high cost of
2
on H are detached from fossil fuels and thus avoid disadvan- Naon (ꢂ700 USD per m ), its demanding humidity management
2
x x x
tageous ue gas (e.g. CO , NO or SO
) as well as sonic pollu- and the problematical waste disposal of uorinated polymers
2,4
2
tion. The heart of a fuel cell setup is the membrane-electrode- instigate the development of new materials. Besides, there are
7
assembly (MEA), which consists of an electro-catalyst and an still few materials with intrinsic proton conductivity and so
electrolyte membrane. While the catalyst is responsible for the a promising approach towards anhydrous proton conduction is
chemical reactions occurring at the anode and cathode of the the substitution of water with other solvents like concentrated
system, the membrane acts as the electrode separator and acids and amphoteric heterocycles. While these provide proper-
2
proton conductor. Due to the harsh electrochemical environ- ties like autoprotolysis and creation of long range hydrogen bond
8
ment within a FC system and a broad range of potential networks, they are prone to early evaporation. For both cases,
ꢀ
applications, special requirements, like physicothermal or operating temperatures beyond 100 C will increase the ion
mobility, decouple the system from complex water management,
3,8,9
and protect the catalysts from CO-poisoning.
However, to
a
comply with this special condition, materials of high electro-
chemical and thermal stability are needed.
Among all the possible candidates, porous organic networks
(PONs) are a promising, yet hardly considered class of mate-
rials. Recently, PONs have been of major interest for gas
University of Bayreuth, Inorganic Chemistry III, Universitaetsstraße 30, 95440
Bayreuth, Germany. E-mail: juergen.senker@uni-bayreuth.de
b
University of Bayreuth, Physical Chemistry II, Universitaetsstraße 30, 95440
Bayreuth, Germany
†
Electronic supplementary information (ESI) available. See DOI:
0.1039/c8ta07822j
1
21542 | J. Mater. Chem. A, 2018, 6, 21542–21549
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
sorption and separation technology as well as for photo- and functionality of the networks. The networks were thoroughly
1
3
15
electrocatalysis due to their high physicochemical stability, characterized by C and N nuclear magnetic resonance
10–14
diverse structures and permanent porosity.
The potential to (NMR) spectroscopy, infrared (IR) spectroscopy, powder X-ray
entrap conductive species in the pores of such frameworks, diffraction (PXRD), and thermogravimetric analysis (TGA) as
using physisorptive interactions with the pore walls should well as argon physisorption. Aer incorporation of imidazole,
prevent the leakage of the incorporated guests, ensuring a long the proton conductivity was investigated by electrochemical
term persistence and thus a permanent conductivity.
impedance spectroscopy in a temperature range from 90 to
ꢀ
In recent publications the incorporation of amphoteric 130 C under a nitrogen atmosphere.
15,16
15
molecules like imidazole
and triazole into the pores of
a framework proved to be a successful approach towards
anhydrous proton conduction in PONs. In this respect, Jiang
Experimental
General information
ꢁ
3
and coworkers presented proton conductivities of 4.37 ꢃ 10
ꢁ
1
ꢁ3
ꢁ1
ꢀ
S cm and 1.1 ꢃ 10 S cm at 130 C under nitrogen by
Argon sorption measurements were carried out on a Quantach-
rome Autosorb-1 pore analyzer at 87.3 K. The data were analyzed
using the ASIQ v 3.0 soware package. Surface areas were esti-
mated by the BET-method, while for pore sizes the Gurvich
method was used. Pore size distributions were calculated with
methods of quasistationary density functional theory (QSDFT).
With respect to the network morphology, we used either
a QSDFT kernel for cylindrical/spherical pores (3D polymers) or
a QSDFT kernel for cylindrical pores (2D polymers) to estimate
the pore size distribution of the respective materials. All poly-
incorporating imidazole and triazole into the covalent organic
15
framework TPB-TMTP. In another study, Xiang et al. incor-
porated imidazole into the mesoporous polyimides Td-PNDI 1
and Td-PPI, and found reasonable conductivities in a tempera-
ꢀ
ꢁ5
ture range from ꢁ40 to 90 C with top values of 9.04 ꢃ 10
S
ꢁ
1
ꢁ4
ꢁ1
ꢀ
cm and 3.49 ꢃ 10 S cm at 90 C and under anhydrous
1
6
conditions, respectively.
Here, we report the synthesis and proton conductivity of six
different porous organic networks (Fig. 1) and their post-
synthetic modication by incorporation of imidazole. Although
the structures have been described previously, to our knowl-
edge, they have never been investigated systematically with
respect to their proton conductivity. We probed the conductivity
as a function of composition and topology for purely carbon–
ꢀ
mers were degassed at 150 C for 12 h before starting the
sorption experiments.
1
13
H and C liquid-NMR spectra of the linker molecules were
6
collected on a Bruker 500 MHz spectrometer using DMSO-d as
the solvent. For infrared spectra (IR) a JASCO FT/IR-6100 Fourier
transform infrared spectrometer with an attenuated total
reectance (ATR) unit was used. Thermogravimetric analysis
9,17
hydrogen networks and benzimidazole based polymers and
thus were able to vary the morphology, porosity and
13
15
Fig. 1 Schematic presentation of the utilized porous organic networks. The assignments correspond to the ones in the C and N NMR spectra.
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A, 2018, 6, 21542–21549 | 21543

View Article Online
Journal of Materials Chemistry A
Paper
e
(
TGA) was carried out on a Mettler Toledo TGA/SDTA851 under were, if not mentioned otherwise, used without further puri-
air. Powder X-ray diffraction (PXRD) measurements were per- cation. The organic solvents were dried by distillation in
formed on a PANalytical X'Pert Pro diffractometer. Here, accordance with the standard procedure. All synthetic proce-
ꢀ
a region from 2 to 30 2q was measured with a 1/4 antiscatter slit dures as well as the apparatus for imidazole loading via the gas
and Cu K
a
radiation (nickel ltered).
phase (Fig. S25) are given in the ESI.†
All solid-state NMR spectra were acquired on a Bruker
1
3
Avance III HD spectrometer operating at a B
¼ 100.6 MHz) MAS spectra were obtained with ramped cross-
polarization (CP) experiments where the nutation frequency nnut
0
eld of 9.4 T.
C
Results and discussion
Synthesis and characterization
(n
0
on the proton channel was varied linearly from 70 to 100%. The All polymers and linker molecules were synthesized by the
1
3
15
samples were spun at 12.5 kHz ( C) and 10.0 kHz ( N) in literature known procedures, which are provided in the ESI.†
a 4 mm MAS double resonance probe (Bruker). The corre- Both synthesis and characterisation were done repeatedly to
1
3
15
sponding nnut of the C/ N channels and the contact times ensure the reproducibility of the results. We observed only
were adjusted to 70/35 kHz and 3.0/5.0 ms, respectively. In all minor variations in the properties recorded for the materials.
1
9
cases proton broadband decoupling with spinal-64 and nnut
7
¼
PON-1–PON-3, originally synthesized by the groups of Ben,
18
13
20
21
5 kHz was applied during acquisition. The C NMR spectra Kaskel and Han, were derived using metal–organic coupling
are referenced with respect to TMS (tetramethylsilane) using the procedures of the Yamamoto and Suzuki type, respectively.
1
5
22,23
secondary standard adamantane. The N NMR spectra are PON-4–PON-6, rst reported by the groups of El-Kaderi
and
referenced with respect to CH NO using the secondary stan- Senker, were synthesized using a condensation reaction of
24
3
2
dard glycine. Differential scanning microscopy was performed diamines and aldehydes to the respective benzimidazole linked
25
on a Mettler Toledo DSC/SDTA 821e DSC with an auto sampler. compounds. Structural characterisation of the networks was
The measurements were performed over a temperature range done by solid-state NMR and IR spectroscopy, as well as PXRD,
ꢀ
ꢀ
ꢀ
ꢁ1
from 25 C to 200 C at a heating/cooling rate of 5 C min
under a nitrogen atmosphere. The measuring routine was resonances are given in Fig. 3. In the C solid-state CP MAS
cycled three times to ensure the sample persistence.
NMR spectra, for PON-1, PON-2 and PON-4 the tetrahedral
Impedance spectroscopy was carried out on a Zahner (Z) carbon signal at 65 ppm (C-1) is the most characteristic (Fig. 3a).
CHN and TG analysis. The assignments for the respective
1
3
1
3
Ennium potentiostat with an applied perturbation voltage of 50 For PON-4–PON-6, the C signals around 100 ppm as well as
1
5
mV with respect to the preserved response signal. Impedance the N signals at ꢁ147 ppm and ꢁ238 ppm indicate the
data sets were acquired by measuring the frequency range from formation of the benzimidazole bridging unit. For PON-5 an
ꢀ
ꢀ
15
1
MHz to 1 Hz at a temperature range of 90 C to 130 C. To additional signal at ꢁ278 ppm was found in the N NMR
ensure that the shape and size of the pellet remain constant spectrum, which was assigned to the central amino group of the
24
throughout the whole measuring process, the material was aldehyde linker. The absence of C –Br (PON-1, PON-2, and
ar
placed into a cylindrical glass encasement with an inner PON-3) signals expected at 119 ppm and R–CHO (PON-4, PON-5,
1
3
diameter of 8 mm between two copper electrodes. The powder and PON-6) signals expected at 194 ppm in the C spectrum as
+
3
was then compressed by screwing the upper electrode tightly well as the one for NH (PON-4, PON-5, and PON-6) at
1
5
against the lower electrode with a force of 0.7 kN corresponding ꢁ330 ppm in the N spectrum indicate a full conversion of the
to a pressure of about 130 bar during the experiments. The linker molecules and, therefore, a high crosslinking degree for
thickness of the resulting pellets was determined aer the all polymers (Fig. 3). The results are supported by infrared
measurements and is given in Tables S2–S7 in the ESI.† The spectroscopy, which revealed for PON-1–PON-3 vibrational
ꢁ
1
ꢁ1
loaded cell was handled under a dry nitrogen atmosphere. To bands at 3030 cm (C–Hstr.vib.), 1066 cm (–C]Cvib.–) and
ꢁ
1
ensure equilibration before the measurements, an equilibration 800 cm (1,4 subst. benzenes) (Fig. S27†). For PON-4, PON-5
time of 30 minutes for each temperature step was included into and PON-6, the formation of the benzimidazole bridging units
ꢁ
1
the data acquisition protocol. Aer performing the measure- results in stretching vibrations around 3380 cm (N–H) and
2
ꢁ1
ments, the length (l) and radius (r) (A ¼ r p) of the pellet were 1595 cm (C]N), respectively (Fig. S27†). The values from the
used to calculate the conductivity using the following equation: CHN analysis further determine the structural integrity, as they

t well to the calculated ones (ESI†). However, for PON-2 and
l
k ¼
PON-3, some deviations occurred, which we attribute to
remaining impurities in the catalyst.
ðAR Þ
1
1
where the resistance (R ) was taken from the high frequency
semicircle in the appropriate Nyquist plot (Cole–Cole plot),
which has been tted with an equivalence circuit (Fig. 2) using
the Zview soware package (Scribner).
Experimental section
All chemicals were purchased from either Sigma-Aldrich
Fig. 2 Equivalent circuit used to fit the obtained results pictured in the
Chemistry GmbH, VWR Chemicals, TCI, or Gr u¨ ssing GmbH and respective Nyquist plots.
21544 | J. Mater. Chem. A, 2018, 6, 21542–21549
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
13
15
Fig. 3
C CP MAS spectra of PON-1–PON-6 with respective assignments (left) and N CP MAS spectra of the corresponding benzimidazole
derived polymers. The labelling is according to Fig. 1.
Morphology and thermal stability were characterized by spherical and cylindrical-pores. In accordance with the calcu-
PXRD and TGA measurements under air, respectively. The lated surface areas the three-dimensional networks
3
ꢁ1
3
ꢁ1
former indicated that all polymers were amorphous (Fig. S31†). showed large pore volumes of 4.73 cm g , 0.34 cm g and
3
ꢁ1
The latter proved the polymers to be temperature stable up to 0.54 cm
g
for PON-1, PON-2 and PON-4, respectively
temperatures between 430 C (PON-3) and 525 C (PON-2) (Table 1). The pore size distribution revealed a primarily
Fig. S29†). The consistently high thermal stabilities are attrib- mesoscopic pore environment for PON-1 (0.20), while PON-2
ꢀ
ꢀ
(
uted to the carbon and carbon–nitrogen based polymer back- (0.6) and PON-4 (0.7) both showed a major microporous
bones, which form strong and persistent networks. Notably, for network morphology (Fig. 4).
the benzimidazole containing polymers a small weight loss
ꢀ
around 100 C was observed. Given that the networks are
hydrophilic due to their inherent functionality, we attribute
this to the incorporation of water molecules upon sample
24
preparation.
The surface area and porosity of the networks were deter-
mined from argon physisorption isotherms measured at 87 K.
The surface area was calculated using the optimized BET
26
equation for microporous compounds. The pore volumes were
27
estimated by the Gurvich method. Further information about
pore size and distribution was calculated based on QSDFT
adsorption branch kernels for carbon materials, which for three
dimensional networks accounts for mixed pores of cylindrical
and spherical shapes (Fig. 4). For two dimensional networks,
however, we used a kernel which accounts solely for cylindrical
pores. Since all networks are amorphous, the given kernels are
not suitable to reect a perfect image of the real porous envi-
ronment and were, therefore, used to derive an estimate only. In
28
accordance with the literature, the highest BET surface area
2
ꢁ1
was found for PON-1 and was about 5000 m g (Table 1). This
network showed a type IV isotherm, typical for primarily mes-
oporous compounds (Fig. 4). For PON-2 and PON-4, which also
reveal a three dimensional morphology, BET surface areas of
2
ꢁ1
2
ꢁ1
700 m
g
and 1270 m
g
were found, respectively. Both
networks exhibit type I isotherms, correlated to primary
29
microporous materials (Fig. 4). The high surface areas of PON-
and PON-4 are the result of their linker morphology, leading
1
to an inherent porosity upon polymer formation. With respect
to their similar buildup, we estimated their pore volume and
pore size distribution using the Gurvich method and the QSDFT
Fig. 4 Argon adsorption isotherms of PONs measured at 87 K (top).
Pore size distribution (PSD) estimated by QSDFT kernels for cylindrical
(PON-3, PON-5, and PON-6) and cylindrical/spherical pores (PON-1,
adsorption branch kernel for carbon materials with mixed PON-2, and PON-4) (bottom). For full PSD see Fig. S26.†
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A, 2018, 6, 21542–21549 | 21545

View Article Online
Journal of Materials Chemistry A
Paper
Table 1 Surface areas and pore volumes taken from Ar adsorption Bragg reections within the PXRD patterns (Fig. S32†) corrob-
isotherms measured at 87 K estimated by BET and DFT kernels for
carbon materials
orates that the frameworks are still amorphous and that imid-
azole is not crystallizing in the inter-particle voids but lls up
a
2
ꢁ1
2
ꢁ1
3
ꢁ1
the micropores. This nding is supported by DSC data, which
indicate the absence of crystallisation and melting peaks
(Fig. S33†).
Polymer
S
A_BET/m g
S
A_DFT/m g
P
V_tot/cm g
PV_mic./PV_tot
b
PON-1b 5043
4165
647
487
1376
649
855
4.73 (4.49)
0.34 (0.33)
0.37 (0.35)
0.54 (0.52)
0.29 (0.28)
0.34 (0.33)
0.2
0.6
0.3
0.7
0.6
0.7
PON-2
PON-3
PON-4
PON-5
PON-6
700
490
1271
593
The amount of incorporated imidazole was determined by
thermogravimetric analysis performed in air, where the weight
c
b
c
c
ꢀ
ꢀ
loss between 150 C and 200 C signals the removal of imidazole
753
(Fig. S30†). We found the highest uptake for imidazole in PON-1
ꢁ1
a
b
(1.08 g g ), which is due to its extended mesoporosity and high
pore volume. Interestingly, by comparing the volume of incor-
porated imidazole (for which we used its liquid density as an
estimate) with the theoretically available pore volume, there still
remains a major part of free space (Fig. 5). We attribute this to
P
V_tot is estimated by the Gurvich method. Estimated by a cylindrical/
spherical pore QSDFT kernel. Estimated by a cylindrical pore QSDFT
c
kernel.
Due to their seemingly two-dimensional structures, the pore blocking effects due to microporous parts of the inter-
porosity of PON-3, PON-5 and PON-6 is attributed to the stack- connecting pore system. PON-2 and PON-5 revealed similar
ꢁ
1
ꢁ1
ing behaviour of the respective polymer sheets, probably driven uptakes of 0.77 g g and 0.75 g g . For both polymers, the
24
by p–p interactions. These networks revealed surface areas of amount of incorporated imidazole exceeds the respective pore
volume, which might result from the breathing behaviour of the
PON-3), respectively, and type I isotherms (Fig. 4). With respect exible networks. With 0.6 g g and 0.43 g g the amount of
2
ꢁ1
2
ꢁ1
2
ꢁ1
750 m
g
(PON-6), 590 m
g
(PON-5) and 490 m
g
24
ꢁ1
ꢁ1
(
to the shape of the polymers, we used the QSDFT adsorption incorporated imidazole exceeds the pore volume of PON-4 and
branch kernel for carbon materials with cylindrical pores, since PON-6, however, only marginally. We attribute the latter nding
the expression of spherical pores in a two dimensional network to an additional lling of microporous interparticle voids. PON-
ꢁ1
is not reasonable. The pore volumes were estimated to be 3 (0.23 g g ) has the lowest uptake value, as well as an
3
ꢁ1
3
ꢁ1
3
ꢁ1
0
.37 cm g (PON-3), 0.29 cm g (PON-5) and 0.34 cm g
incomplete pore lling. In order to estimate the inuence of the
(PON-6) (Table 1). In accordance with the shape of their benzimidazole backbone of PON-4, PON-5 and PON-6 on the
isotherms, the pore size distribution for PON-5 and PON-6 proton conduction performance, we calculated the ratio of
showed a primary microporous structure. PON-3, however, seems incorporated imidazole per functional group. While for PON-5
to consist mostly of small mesopores (Table 1 and Fig. 4 bottom). 1.8 imidazole molecules per functional group were incorpo-
rated, PON-4 and PON-6 revealed ratios of only 1.4 and 1.0,
respectively.
Incorporation of imidazole
The substitution of water with nitrogen containing, amphoteric
Proton conduction
heterocycles such as imidazole already proved to be a promising
16,30
approach for proton conduction in COFs and PONs.
Since The proton conductivity of the different networks has been
ꢀ
the nitrogen bearing functionalities are strong proton media- determined as a function of temperature (90–130 C) under
+
tors, the formation of protonic charge carriers (C H N H ) , a static nitrogen atmosphere. The data were obtained by
3
3
2
2
driven by self-dissociation, leads to extensive hydrogen bond
8,31,32
interactions and thus creates proton conduction pathways.
The high melting point and stability of imidazole allow
ꢀ
temperatures beyond 100 C to be used and are, therefore,
favourable in terms of ion mobility and catalyst reusability.
Various procedures for the incorporation of imidazole into
porous structures have been reported in the literature, e.g.
16
33
through solvent, via condensation or through a gas phase.
We found that among all reported methods, gas phase incor-
poration provides the most reliable data, since only single
molecules are incorporated into the pores of a network, which
supports an effective pore lling and minimizes condensation
as well as crystallisation of imidazole on the polymer particle
surfaces. Thereby, we were able to incorporate imidazole into
the voids of the respective networks. The success of the imid-
azole incorporation was monitored gravimetrically and by
infrared spectroscopy. For the latter characteristic bands at
ꢁ1
ꢁ1
2
650 cm and 1615 cm were assigned to the N–H/N and
Fig. 5 Imidazole uptake in comparison to the available pore volume of
C–N vibrations of imidazole (Fig. S28†). The absence of any the respective polymers.
21546 | J. Mater. Chem. A, 2018, 6, 21542–21549
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
semicircles in the respective Nyquist plots (Fig. 2 and S36–
34
S41†). The circuit describes two semicircles in the impedance
plots, which we assign to charge transfer resistance, for the high
frequency circle, and additional boundary effects due to the
measuring setup for the ones at low frequencies (electrode
34
contacting, two electrode setup etc.) (Fig. 6). The constant
phase elements (CPE-1 and -2) are used to describe capacitive
parts of the system adapted to inuences of surface roughness
35,36
and system heterogeneity.
Interestingly, for the three
dimensional polymers, the two semicircles are well separated,
while for the two-dimensional polymers, the semicircles sepa-
rate to a lesser extent and showed an additional spur at low
frequencies (Fig. 6, top and middle). We attribute this behav-
iour to the sheet-like morphology of the network, which, in
comparison to the more granular formed three-dimensional
polymers, results in a different contact to the electrodes.
The proton conductivity of the polymers was estimated from
the high frequency resistance R , depicted by the rst semicircle
1
in the respective Nyquist plots (Fig. 6, Table 2 and Fig. S34–S39†).
ꢀ
All samples were preheated to 90 C to ensure a full equilibration
of the system before starting the measurement routine. Inter-
estingly, although all networks rely on imidazole based proton
transport, the conduction performance differs markedly (Fig. 6,
bottom). While for PON-3, -4 and -6 only low conductivities in the
ꢁ
6
ꢁ7
ꢁ1
range from 10 to 10 S cm were observed, PON-5 and PON-2
ꢁ5
ꢁ1
exhibit conductivities around 10 S cm . The best values were
ꢁ
4
ꢁ1
found for PON-1 (5.2 ꢃ 10 S cm ), making this network
comparable to state-of-the-art porous polymers like Im@TPB-
15
ꢁ3
ꢁ1
ꢀ
16
37
DMTP-COF (4.37 ꢃ 10 S cm , 130 C), H
3 4
PO @TP-Azo
ꢁ4
ꢁ1
ꢀ
ꢁ4
ꢁ1
(
9
9.9 ꢃ 10 S cm , 70 C) and Im@Td-PPI (3.49 ꢃ 10 S cm
,
ꢀ
0 C). The comparison of the estimated proton conductivity with
the amount of incorporated imidazole shows that a high amount
of proton mediators is favourable for an excellent performance.
Besides varying conductivities, different activation energies
for the polymers were obtained from the respective Arrhenius
plots, which are indicative for the conduction mechanism
3
8,39
(
Fig. 6).
activation barriers between 0.1 and 0.4 eV (9.6–38.6 kJ mol
are postulated, while for the vehicle type mechanism activation
For the Grotthuss mechanism (structural diffusion)
ꢁ
1
)
Fig. 6 Exemplary Nyquist plots of PON-1 (top) and PON-6 (middle),
ꢁ
1
ꢀ
A
energies (E ) between 0.5 and 0.9 eV (48.2–86.8 kJ mol ) are
circuit pictured in the experimental section. The Arrhenius plot of reported. For PON-1, PON-2 and PON-4 the activation barriers
PON-1–PON-6 obtained from high frequency resistance data were calculated to be 39 kJ mol and 35 kJ mol , respectively,
measured at 130 C under a N
2
atmosphere, fitted with the equivalent
40
ꢁ1
ꢁ1
ꢀ
ꢀ
measured in a temperature range between 90 C and 130 C under
a N atmosphere (bottom). The colour code is similar to the one in
Fig. 1.
which is within the range typical for the Grotthuss mechanism
2
4
1
(
Fig. 6). We attribute this to the high amount of incorporated
34
imidazole molecules within the networks, which leads to
a homogeneous charge carrier distribution, and therefore
electrochemical impedance spectroscopy in a two electrode
setup. They have been tted to an equivalent circuit as shown in
Fig. 2 and optimized to comply with the shape of the obtained
interconnected proton conduction pathways.
networks PON-6, PON-3 and PON-4 are 44, 53 and 58 kJ mol
and are signicantly larger indicating a vehicle type conduction
E
A
for the
ꢁ1
ꢀ
Table 2 Proton conductivities measured at 130 C under a nitrogen atmosphere and activation energies of the polymers, calculated from the
respective Arrhenius plots
Network
/S cm
PON-1
PON-2
PON-3
PON-4
PON-5
PON-6
ꢁ
1
1
ꢁ4
ꢁ5
ꢁ7
ꢁ6
ꢁ5
ꢁ6
k
130
ꢀ
C
5.2 ꢃ 10
7.3 ꢃ 10
2.1 ꢃ 10
9.7 ꢃ 10
2.0 ꢃ 10
4.9 ꢃ 10
ꢁ
A
E /kJ mol
39
39
53
58
35
44
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A, 2018, 6, 21542–21549 | 21547

View Article Online
Journal of Materials Chemistry A
Paper
Conclusions
We synthesized six different porous organic networks (PONs)
based either on carbon–hydrogen (PON-1–PON-3) or benz-
imidazole (PON-4–PON-6) in order to study the inuence of
network morphology on anhydrous proton conduction. The
1
3
polymers exhibit high crosslinking degrees as estimated by
C
1
5
and N MAS NMR spectroscopy. The BET surface areas are
between 490 m g (PON-3) and 5043 m g (PON-1) and they
have a high thermal resistance up to 525 C (PON-2). Aer post-
2
ꢁ1
2
ꢁ1
ꢀ
synthetic incorporation of imidazole via the gas phase, which
ꢁ
1
allowed a maximal lling of 1.1 g g to be reached for PON-1,
we investigated the proton conduction performance of these
networks as a function of the morphology, chemical function-
ality, and guest loading. The networks reveal conductivities
ꢁ7
ꢁ1
ꢁ4
ꢁ1
ꢀ
Fig.
7
Correlation between imidazole incorporation and proton
between 10 S cm and 10 S cm at 130 C under anhy-
ꢀ
conduction mechanism. The conductivity values were taken at 130 C
under a nitrogen atmosphere.
ꢁ4
drous conditions with maximum values for PON-1 (5.2 ꢃ 10
S
ꢁ
1
ꢀ
cm at 130 C). The latter is comparable to state-of-the-art
15
ꢁ3
ꢁ1
materials like Im@TPB-DMTP-COF (4.73 ꢃ 10 S cm at 130
ꢀ
16
ꢁ4
ꢁ1
ꢀ
mechanism (Fig. 6) in these cases. This might be due to the
C) or TD-PPI (3.49 ꢃ 10 S cm at 90 C). Moreover, we
lower amount of guest molecules and, therefore, an inhomo- estimated the activation energy of the polymers to fall in the
ꢁ1
ꢁ1
geneous charge carrier distribution, which forces charge range of ꢁ39 kJ mol to ꢁ58 kJ mol and to correlate with the
transport via diffusion of protonated imidazole molecules.
amount of incorporated guest molecules. In this respect, three-
However, the data seem to indicate a direct correlation dimensional polymers, based on rigid tetrahedral building
between observed proton conductivity and the amount of incor- units seem to be favourable to provide more homogeneous
porated imidazole (Fig. 7). Interestingly the level of pore lling charge carrier distributions. Interestingly, no further inuence
seems to be subordinated, as PON-1 exhibits the highest of pore geometry and degree of pore lling on the conduction
conductivity with only 20% pore lling, while PON-6 shows a low performance was found. Furthermore, the presence of poly-
conductivity, despite its excessive pore lling. In this respect, benzimidazole groups within the polymer backbone (PON-4–
pore geometry also seems to have only minor effects, since there PON-6) seemed to have an impeding effect on conduction,
is just a slight difference between the two-dimensional PON-5 especially at low ratios of imidazole to functional groups. This
and three-dimensional PON-2 in terms of imidazole incorpora- might be due to the stronger binding of imidazole to the
tion and proton conductivity (Fig. 7). However, three dimensional functional groups causing its immobilisation. In this respect,
polymers seem to be more promising as they tend to express further investigations are part of our future work.
a more rigid porous environment which is favourable for the
incorporation of proton conductive species. This statement is
based on PON-1 and PON-2, both based on rigid tetrahedral
Conflicts of interest
building units, which showed the best performance in this study. There are no conicts of interest to declare.
Furthermore, the presence of benzimidazole units in the poly-
mer backbone like in PON-4, PON-5 and PON-6 seems to have
a negative effect, as the imidazole affixation at the respective
Acknowledgements
functional group is increased. While at lower imidazole to The authors thank Dr Renee Siegel and Beate Bojer for per-
functional group ratios this might impede the carrier move- forming the solid-state NMR spectroscopy measurements. We
ment, due to hindered diffusion, at high ratios the interplay gratefully thank Prof. Josef Breu for access to physisorption
between the guest and functional group might be supported. instruments.
Among the benzimidazole based polymers, PON-5 exhibits the
highest ratio of 1.8 imidazole molecules per functional group
Notes and references
(PON-4: 1.4; PON-6: 1.0), which in combination with its small
repeating unit is indicative of a dense charge carrier concen-
tration. In consequence PON-5 exhibits a higher proton
conduction performance than the competitive benzimidazoles,
despite its two-dimensional morphology. Since we expect the
conductive imidazole units in PON-4 and PON-6 to be primarily
located at the networks of functional groups, not only their
1 R. F. Service, Science, 1999, 285, 682–685.
2 S. J. Peighambardoust, S. Rowshanzamir and M. Amjadi, Int.
J. Hydrogen Energy, 2010, 35, 9349–9384.
3 P. Ramaswamy, N. E. Wong and G. K. H. Shimizu, Chem. Soc.
Rev., 2014, 43, 5913–5932.
4 C. Klumpen, S. G ¨o drich, G. Papastavrou and J. Senker, Chem.
Commun., 2017, 53, 7592–7595.
ꢁ
1
conduction performance is lower, but at ꢁ58 kJ mol (PON-4)
ꢁ1
and ꢁ44 kJ mol (PON-6) their energy of activation is also much
higher with respect to the one of PON-5 (ꢁ35 kJ molꢁ1).
5 R. C. Woudenberg, O. Yavuzcetin, M. T. Tuominen and
E. B. Coughlin, Solid State Ionics, 2007, 178, 1135–1141.
21548 | J. Mater. Chem. A, 2018, 6, 21542–21549
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
6
N. E. Wong, P. Ramaswamy, A. S. Lee, B. S. Gelfand, 24 C. Klumpen, F. Radakovitsch, A. Jess and J. Senker,
K. J. Bladek, J. M. Taylor, D. M. Spasyuk and Molecules, 2017, 22, 1343.
G. K. H. Shimizu, J. Am. Chem. Soc., 2017, 139, 14676–14683. 25 S. I. Alaqeel, J. Saudi Chem. Soc., 2017, 21, 229–237.
D. Umeyama, S. Horike, M. Inukai and S. Kitagawa, J. Am. 26 K. S. Walton and R. Q. Snurr, J. Am. Chem. Soc., 2007, 129,
7
8
9
Chem. Soc., 2013, 135, 11345–11350.
8552–8556.
M. F. H. Schuster and W. H. Meyer, Annu. Rev. Mater. Res., 27 S. Lowell, J. E. Shields, M. A. Thomas and M. Thommes,
2
003, 33, 233–261.
Characterization of Porous Solids and Powders, Springer
Netherlands, Dordrecht, 2004.
28 S. Das, P. Heasman, T. Ben and S. Qiu, Chem. Rev., 2017, 117,
1515–1563.
J. A. Asensio, E. M. S ´a nchez and P. G ´o mez-Romero, Chem.
Soc. Rev., 2010, 39, 3210.
0 G. Zhu and H. Ren, Porous Organic Frameworks, Springer
1
1
1
1
Berlin Heidelberg, Berlin, Heidelberg, 2015.
1 C. Klumpen, M. Breunig, T. Homburg, N. Stock and
J. Senker, Chem. Mater., 2016, 28, 5461–5470.
2 N. Popp, T. Homburg, N. Stock and J. Senker, J. Mater. Chem. 30 X. Meng, H.-N. Wang, S.-Y. Song, H.-J. Zhang, S. Kitagawa,
A, 2015, 3, 18492–18504.
29 M. Thommes, K. Kaneko, A. V. Neimark, J. P. Olivier,
F. Rodriguez-Reinoso, J. Rouquerol and K. S. W. Sing, Pure
Appl. Chem., 2015, 87, 1051–1069.
G. K. H. Shimizu, H. Y. Zang, G. S. Zhu, N. Gadegaard,
M. Mirza, D. Paul, J. Poblet, L. Cronin, K. D. Demadis and
A. Cabeza, Chem. Soc. Rev., 2017, 46, 464–480.
3 A. B. Soliman, M. H. Hassan, T. N. Huan, A. A. Abugable,
W. A. Elmehalmey, S. G. Karakalos, M. Tsotsalas,
M. Heinle, M. Elbahri, M. Fontecave and M. H. Alkordi, 31 K. D. D. Kreuer, A. Fuchs, M. Ise, M. Spaeth and J. Maier,
ACS Catal., 2017, 7, 7847–7854. Electrochim. Acta, 1998, 43, 1281–1288.
4 R. S. Sprick, B. Bonillo, M. Sachs, R. Clowes, J. R. Durrant, 32 B. S. Hickman, M. Mascal, J. J. Titman and I. G. Wood, J. Am.
1
D. J. Adams and A. I. Cooper, Chem. Commun., 2016, 52,
0008–10011.
5 H. Xu, S. Tao and D. Jiang, Nat. Mater., 2016, 15, 722–726.
Chem. Soc., 1999, 121, 11486–11490.
33 S. Liu, Z. Yue and Y. Liu, Dalton Trans., 2015, 44, 12976–
12980.
1
1
1
6 Y. Ye, L. Zhang, Q. Peng, G.-E. Wang, Y. Shen, Z. Li, L. Wang, 34 Z. Xie, C. Song, B. Andreaus, T. Navessin, Z. Shi, J. Zhang and
X. Ma, Q.-H. Chen, Z. Zhang and S. Xiang, J. Am. Chem. Soc.,
015, 137, 913–918.
7 J. C. Persson and P. Jannasch, Chem. Mater., 2003, 15, 3044–
045.
8 B. M. Fung, A. K. Khitrin and K. Ermolaev, J. Magn. Reson.,
000, 142, 97–101.
9 T. Ben, H. Ren, S. Ma, D. Cao, J. Lan, X. Jing, W. Wang, J. Xu,
F. Deng, J. M. Simmons, S. Qiu and G. Zhu, Angew. Chem.,
Int. Ed., 2009, 48, 9457–9460.
0 M. Rose, N. Klein, W. B ¨o hlmann, B. B ¨o hringer, S. Fichtner
and S. Kaskel, So Matter, 2010, 6, 3918.
1 Q. Chen, J.-X. Wang, F. Yang, D. Zhou, N. Bian, X.-J. Zhang,
C.-G. Yan and B.-H. Han, J. Mater. Chem., 2011, 21, 13554.
2 M. G. Rabbani and H. M. El-Kaderi, Chem. Mater., 2012, 24,
S. Holdcro, J. Electrochem. Soc., 2006, 153, E173.
35 A. Lasia, Electrochemical Impedance Spectroscopy and its
Applications, Springer New York, New York, NY, 2014.
36 R. Yadav and P. S. Fedkiw, J. Electrochem. Soc., 2012, 159,
340–346.
37 S. Chandra, T. Kundu, S. Kandambeth, R. Babarao,
Y. Marathe, S. M. Kunjir and R. Banerjee, J. Am. Chem.
Soc., 2014, 136, 6570–6573.
38 M. F. H. Schuster, W. H. Meyer, M. Schuster and
K. D. Kreuer, Chem. Mater., 2004, 16, 329–337.
39 W. M u¨ nch, K.-D. Kreuer, W. Silvestri, J. Maier and G. Seifert,
Solid State Ionics, 2001, 145, 437–443.
40 V. Zima, D. S. Patil, D. S. Raja, T.-G. Chang, C.-H. Lin,
K. Shimakawa, T. Wagner, D. Senthil Raja, T.-G. Chang,
C.-H. Lin, K. Shimakawa and T. Wagner, J. Solid State
Chem., 2014, 217, 150–158.
2
1
1
1
3
2
2
2
2
2
1511–1517.
3 M. G. Rabbani, T. E. Reich, R. M. Kassab, K. T. Jackson,
H. M. El-Kaderi, W. Zhang, J. Lipkowski, J. Guenther, 41 W. Munch, K.-D. Kreuer, W. Silvestri, J. Maier and G. Seifert,
J. Bl u¨ mel, R. Krishna, Z. Li and H.-C. Zhou, Chem.
Commun., 2012, 48, 1141–1143.
Solid State Ionics, 2001, 145, 437–443.
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A, 2018, 6, 21542–21549 | 21549