Ionic Highways from Covalent Assembly in Highly Conducting and Stable Anion Exchange Membrane Fuel Cells

Article abstract of DOI:10.1021/jacs.9b08749

A major challenge in the development of anion exchange membranes for fuel cells is the design and synthesis of highly stable (chemically and mechanically) conducting membranes. Membranes that can endure highly alkaline environments while rapidly transporting hydroxides are desired. Herein, we present a design using cross-linked polymer membranes containing ionic highways along charge-delocalized pyrazolium cations and homoconjugated triptycenes. These ionic highway membranes show improved performance. Specifically, a conductivity of 111.6 mS cm^-1 at 80 °C was obtained with a low 7.9% water uptake and 0.91 mmol g^-1 ion exchange capacity. In contrast to existing materials, ionic highways produce higher conductivities at reduced hydration and ionic exchange capacities. The membranes retain more than 75% of their initial conductivity after 30 days of an alkaline stability test. The formation of ionic highways for ion transport is confirmed by density functional theory and Monte Carlo studies. A single cell with platinum metal catalysts at 80 °C showed a high peak density of 0.73 W cm^-2 (0.45 W cm^-2 from a silver-based cathode) and stable performance throughout 400 h tests.

Full text of DOI:10.1021/jacs.9b08749

Article
pubs.acs.org/JACS
Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Ionic Highways from Covalent Assembly in Highly Conducting and
Stable Anion Exchange Membrane Fuel Cells
Yoonseob Kim,†,‡ Yanming Wang,§,⊥ Arthur France-Lanord,§,⊥ Yichong Wang,†
You-Chi Mason Wu,† Sibo Lin,† Yifan Li,† Jeffrey C. Grossman,*,§ and Timothy M. Swager*,†
†Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States
‡Department of Chemical and Biological Engineering, Hong Kong University of Science and Technology, Clear Water Bay,
Kowloon, Hong Kong SAR, China
§Department of Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139,
United States
S
* Supporting Information
ABSTRACT: A major challenge in the development of anion
exchange membranes for fuel cells is the design and synthesis of
highly stable (chemically and mechanically) conducting mem-
branes. Membranes that can endure highly alkaline environments
while rapidly transporting hydroxides are desired. Herein, we
present a design using cross-linked polymer membranes containing
ionic highways along charge-delocalized pyrazolium cations and
homoconjugated triptycenes. These ionic highway membranes
show improved performance. Specifically, a conductivity of 111.6
mS cm−1 at 80 °C was obtained with a low 7.9% water uptake and
0.91 mmol g−1 ion exchange capacity. In contrast to existing
materials, ionic highways produce higher conductivities at reduced
hydration and ionic exchange capacities. The membranes retain
more than 75% of their initial conductivity after 30 days of an alkaline stability test. The formation of ionic highways for ion
transport is confirmed by density functional theory and Monte Carlo studies. A single cell with platinum metal catalysts at 80 °C
showed a high peak density of 0.73 W cm−2 (0.45 W cm−2 from a silver-based cathode) and stable performance throughout 400
h tests.
dvances in fuel cell technologies over recent decades have
enabled the production of fuel cell vehicles by major
60−80 °C operating temperatures.9 Avoiding degradation is
necessary for stable, long-term operation of the fuel cells.
Finally, the hydroxide conductivity must be high (>100 mS
cm−1) to meet the desired power output to drive an
automotive motor. The design of stable cationic polymer
networks for fast anion transport is best solved by chemical
principles.10 There has been progress on these fronts with the
incorporation of cross-linked networks to prevent swelling,11,12
ether-free networks to impart improved chemical stability,7,9
and increasing the ion exchange capacity (IEC) to achieve
higher conductivity.13,14 However, in most cases, these
solutions require trade-offs in properties. For example, cross-
linking of hydrocarbon backbones can present challenges and
can give reduced IEC. Similarly, increasing IEC often results in
increased water uptake and polymer swelling. Designs that
address all three AEM criteria simultaneously without
compromises are rare.
A
automobile manufacturers.1 The advantages of fuel cell
vehicles include environmental friendliness, convenient/rapid
refueling, compatibility with current fossil fuel infrastructure,
and the ability to run long range on a single fueling. Current
commercial vehicles use proton exchange membrane fuel cells,
and the high power output is the result of efficient platinum-
catalyzed reactions and fast ion exchange through fluorocarbon
membranes such as Nafion.2 However, these systems are
expensive from both the membrane and catalyst standpoint,
which is hindering broader adoption of this technology.1
Recently, anion exchange membrane based fuel cells
(AEMFCs) that use cationic polymeric hydroxide-conducting
membranes as separators have gained interest because they
function with less expensive metal catalysts.3−7 To realize this
potential, AEMs must address three major issues. First, fuel
cells often operate in a nearly 100% humidity level, which
causes high water uptake and swelling of the charged polymer
membranes. Excessive swelling causes instabilities and loss of
performance,8 and freestanding membranes need to limit
dimensional changes. Second, hydroxide, a strong base and
nucleophile, degrades most cationic organic membranes at
Herein, we report highly conducting AEMs that are
mechanically and chemically stable. Starting with a high free
Received: August 13, 2019
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.9b08749
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 1. Design of ionic highway through a pyrazolium cross-linked triptycene polymer. (a) Schematic of conversion from poly(methoxy-
triptycene ether sulfone) (OMe-PTES) to pyrazolium cross-linked poly(triptycene ether sulfone) (PXm-Tn) via demethylation, in situ
functionalization and film casting, and anion exchange. (b) Schematic for a case of 100% degree of cross-linking. Anion transport is facilitated
through delocalized cationic charges of pyrazoliums over triptycene and homoconjugation in the triptycene. (c) Schematic for a case of 0% degree
of cross-linking. Anions would face much higher activation energy for transport. See Figures S7 and S8 for comprehensive schemes of chemical
structures.
volume poly(hydroxy-triptycene ether sulfone) scaffold,
cationic pyrazolium groups are introduced as covalently
bound cross-linkers and pendant groups by a reaction with
dichloro- and monochloro-pyrazolium ions. A series of
pyrazolium cross-linked poly(triptycene ether sulfone)s were
studied (PXm-Tn; 0 ≤ m, n ≤ 100, m denotes the intended
percent of cross-linking, and n denotes the percent of
triptycene repeats in the polymer, Figure 1). Reactions with
the dichloro-pyrazolium cause cross-links, and addition of
monochloro-pyrazolium gives pendant cations. The cross-
linking was optimized for mechanical properties, and the ionic
highways (Figure 1b) form as a combination of delocalized
pyrazolium cations that are organized and homoconjugated by
connection to triptycenes.15 We demonstrate the utility of
these AEMs containing ionic highways in hydroxide ion
conducting fuel cells and show sustained high power output.
substituted ring structures have improved alkaline stability.18
Third, we expected that the covalent assembly of repeating
ionic segmentsionic highwaywould result in higher
conductivity by lowering the activation barriers for ion
transport.19 These collective efforts created repeating segments
of pyrazolium cations and triptycenes, where homoconjugation
to the triptycenes helped to further delocalize the cationic
charge of the ionic highway (Figure 1).
Charge delocalization of cations is an established way to
reduce reactivity with nucleophilic and basic hydroxide anions.
Resonance stabilization can enhance cations’ alkaline stability
as can be seen even in cases of guanidiniums, imidazoliums,
and pyridiniums,16,18,20 although we note that fully substituted
and sterically hindered benzimadoliums and aminophospho-
nium cations go into a top rank for alkaline stability.21 In this
study, we expand the cationic groups in AEMs to include
pyrazolium groups with additional π donors. These are
aromatic resonance-stabilized cations that can be conveniently
incorporated as cross-linkers or pendants.22 Although there are
considerable efforts in creating high-performing AEMs through
microscale phase segregation,23,24 previous investigations have
not pursued covalent assembly of repeating ionic cross-links to
facilitate hydroxide transport.
INTRODUCING IONIC HIGHWAYS IN AN ANION
EXCHANGE MEMBRANE
■
At the onset of this study, we set three design criteria, each
addressing the challenges mentioned above, under a theme of
creating an ionic highway. First, we needed to prevent high
water uptake and swelling by constructing a cross-linked
polymer network.11,12 We decided installing cations as the
cross-linkers would be the preferred strategy, as it balances IEC
and mechanical stability. Second, we targeted fully substituted
cationic rings to avoid reactive C−H groups. Experimental16
and theoretical17 studies have shown that deprotonation of the
cation rings is a major degradation mechanism, and fully
DESIGNING MONOMERS AND POLYMERS AND IN
SITU MEMBRANE PREPARATIONS
■
We have introduced stable cationic rings, generated from 3,5-
dichloropyrazolium, in AEMs. Our efforts were guided by
density functional theory (DFT) calculations of the pKa value
B
DOI: 10.1021/jacs.9b08749
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
of the proton at the alpha methyl of 4-methyl-3,5-diphenoxy-
1,2-diphenyl-pyrazolium tetrafluoroborate (4) and the alpha
methylene of 4-butyl-3,5-diphenoxy-1,2-diphenyl-pyrazolium
tetrafluoroborate (7).17 These calculations reveal that these
pKa’s for 4 and 7 are 39.9 and 47.0, respectively (Figures S9d,
S10d, and Note S2), and both represent particularly high
values for organic cations.17 The five-membered cationic
dichloro-alkyl-diphenyl pyrazolium tetrafluoroborates were
synthesized by adaptation of literature procedures22,25 (Figures
S1 and S4). Two-step procedures from commercial materials
resulted in 4-methyl-3,5-dichloro-1,2-diphenyl-pyrazolium tet-
rafluoroborate (2) and 4-butyl-3,5-dichloro-1,2-diphenyl-pyr-
azolium tetrafluoroborate (6) with high overall yields (>72%).
The cationic nature of these intermediates makes them
excellent substrates for nucleophilic aromatic substitution of
the chlorides with oxygen- or nitrogen-based nucleophiles.
Reactions with phenol result in 4 and 7 which were used for
alkaline stability tests (Figures S3 and S5). The π-electron-
donating phenoxy groups provide additional stabilization of
the cationic pyrazolium groups. For the simplest substituted
pyrazolium, at least four different resonance structures can be
drawn with all atoms in a preferred octet resonance structure
(Figure S11).
on the polymer, wherein 2 adds one cation for every two
reactive hydroxyls. Hence, the ratio of these reactants also
controls the IEC.
As a core polymer structure, we focused on triptycene-
containing poly(ether sulfone), which has been previously
shown to produce molecularly defined porosity (free volume)
and facilitated ion transport while promoting solubility and
enhancing mechanical properties.27,28 In this case, we
introduced modified triptycenes that can present a free
hydroxyl group on each pendant benzene ring. This structure
enables attachment of pyrazolium groups by nucleophilic
substitution as pendants or as a cross-linking group (Figure 1).
The proximate nature of the hydroxyl groups is key to the
creation of the ionic highway. The polymer synthesis begins
with 1,8-dimethoxy-benzenoanthracene-13,16-dione (10, Fig-
ure S6), wherein the hydroxyls are protected as methoxy
groups.29,30 Monomer 10 was synthesized in three steps from
commercially available and low-cost 1,8-dihydroxyanthraqui-
none with a high overall yield of 73%. Copolymers were
synthesized using 10, with bisphenol A and bis(4-fluorophen-
yl) sulfone, via nucleophilic aromatic substitution (SNAr), with
K2CO3 as a base and dimethylacetamide as a solvent (Figure
S7). Bisphenol A was introduced to lower Tg of the resulting
polymer for increased film compliance. The SNAr reaction
produced high molecular weight poly(methoxy-triptycene
ether sulfone), OMe-PTES (n = 50; Mn: 107 kDa, Mw: 250
kDa, and PDI: 2.33) (Figure S12; molecular weight
information on polymer of n = 75 is also included).
Subsequently, the OMe-PTES was quantitively demethylated
(1H NMR, Figure S7) by reaction with BBr3 at −78 °C to yield
poly(hydroxy-triptycene ether sulfone) (OH-PTES).
The alkaline stability of 4 and 7 was tested by in situ NMR
monitoring solutions at 1, 2, and 5 M KOH solutions in
CD3OH at 80 °C using an internal standard.16 Periodically, the
solutions were analyzed by 1H NMR spectroscopy to
determine key cation signals relative to an internal standard.
This set of tests showed high stability of 4 and 7 under alkaline
conditions, and more than 91, 85, and 82% of the 4 and 93, 88,
and 84% of the 7 persisted after heating at 80 °C in 1, 2, and 5
M KOH solutions over 30 days (Figures 2 and S9a−c). It
We fabricated membranes by in situ reactions of OH-PTES
(n = 50, 12) and pyrazolium dichloride (2) and monochloride
(3) during film casting from dimethylformamide/triethylamine
solution (Figure S8). By simply changing the amount of each 2
and 3 added, we produced three pyrazolium cross-linked
poly(triptycene ether sulfone) membranes with different m,
PX75-T50, PX40-T50, and PX10-T50 (Table S1, Figure S13,
and Figure S14 for solid-state NMR spectra of membranes).
The substitution reaction between hydroxy on the polymer
and chlorides of 2 and 3 resulted in quantitative loss of
chlorides as determined by energy-dispersive X-ray spectros-
copy, and the final membranes were subjected to ion exchange
of BF4− to OH− (Figure S15). We found that polymers with n
of equal to or greater than 63 result in too brittle of
freestanding membranes. Also, for n = 50, 100% and 0% cross-
linking did not result in freestanding membranes (Tables S1
and S2). Within these restrictions, we studied three
membranes, PX75-T50, PX40-T50, and PX10-T50, that
showed a good balance of material properties (Figures 1 and
3). Membranes prepared via slow drying in a vacuum oven
have uniform thickness: 53 4, 55 3, and 64 5 μm for
PX75-T50, PX40-T50, and PX10-T50, respectively (Figure
S13). Our in situ preparation method has distinct advantages
and allows simultaneous functionalization and cross-linking to
tune the polymer’s properties. The cross-linked freestanding
membranes showed satisfactory mechanical properties (Figure
S16) with sufficiently high Young’s moduli (172 and 117 MPa
for PX75-T50 and PX10-T50, respectively) and ultimate
tensile strengths (264 and 144 MPa for PX75-T50 and PX10-
T50, respectively). Hence, these materials were robust enough
for fuel cell testing wherein they must be compressed without
fracturing under the pressure of high-torque bolts (5.5 N m).
Figure 2. Long-term alkaline stability of 4 at 80 °C. Percent
remaining of 4 in 1, 2, and 5 M KOH solutions in CD3OH at 80 °C,
determined by 1H NMR spectroscopy relative to an internal standard.
See Figure S9 for the time-tracking NMR data.
should be emphasized that the NMR spectra remained
relatively clean, except the small peaks at 7.51 and 7.83 ppm
from cleavage of C−O bonds. Although 7 showed slightly
higher alkaline stability (Figure S10a−c,e), we decided to use
the methyl pyrazolium, 4, for further studies. Specifically, we
anticipate more thermal mobility with butyl pyrazolium (the
melting temperatures of 4 and 7 are 216−218 and 148−150
°C, respectively) and potentially lower conductivity with
longer butyl chains.26 We synthesized 4-methyl-3-chloro-5-
phenoxy-1,2-diphenyl-pyrazolium tetrafluoroborate (3, Figure
S2), which can be used to graft cations onto polymers without
introducing cross-linking that occurs by reaction with dichloro-
pyrazolium (2). Hence, by simply changing the ratio between
2 and 3, we can change m values in the membrane.
Additionally, 3 adds one cation for every reactive hydroxyl
C
DOI: 10.1021/jacs.9b08749
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 3. Hydroxide conductivity of pyrazolium cross-linked poly(triptycene ether sulfone) membranes. (a) Temperature-dependent conductivity
of samples. (b) OH− conductivity of PX75-T50 and PX10-T50 in 1 M KOH solution at 80 °C. (c) Hydroxide conductivity versus water uptake.
Data points at 80 °C were collected to be compared. (d) Hydroxide conductivity versus IEC. Experimental data points at 20 °C were collected to
be compared. Legends for (c) and (d) are in (d). References for data points in (c) and (d): QPAEN,31 TQAPEK,32 XE-Imd,11 LDPE-AEM,33
PF,34 R4-C12,35 and C6D40 and BTMA20.36 For all experiments, mean values from triplicate experiments are used unless otherwise noted.
that C−N or C−C bonds may offer additional stability relative
to the phenolic C−O bonds, and previous studies have pointed
to ether linkages as potentially having limited alkaline stability
(see Figures S9 and S10 and Note S1 on the use of polyaryl
ether and alkaline stability).8,11,32,33,37 It is important to note
that the ether linkages in the polymer main chain are also a
potential point of reactivity.
IONIC CROSS-LINKS REDUCE WATER UPTAKE
BUT ENHANCE OH− CONDUCTIVITY
■
Hydroxide conductivity over a temperature range of 30−80
°C, measured every 5 °C, reveals that PX75-T50 achieved 95.7
and 101.2 mS cm−1 at 65 and 70 °C (Figure 3a). A
conductivity of ≈100 mS cm−1 at a working temperature of
fuel cells (generally 60−80 °C) satisfies the standards of U.S.
Department of Energy for reliable operation. PX75-T50
showed a conductivity of 52.1 mS cm−1 at 30 °C, and PX40-
T50 and PX10-T50 showed progressively lower values of 28.8
and 9.0 mS cm−1 at the same temperature. At 80 °C, PX75-
T50, PX40-T50, and PX10-T50 showed conductivities of
111.6, 73.1, and 26.5 mS cm−1, respectively. Activation
energies (Ea) calculated from conductivity data using an
Arrhenius equation (Figure S17) show the highly cross-linked
membrane, PX75-T50, to have lower Ea of 13.7 kJ mol−1 than
those of PX40-T50 and PX10-T50, which are, respectively,
17.6 and 18.9 kJ mol−1. Low Ea value for PX75-T50 is
attributed to facilitated ion transport along cross-linked ionic
highways.
To probe the long-term alkaline stability, membranes were
soaked in 1 M KOH solution at 80 °C, and conductivity values
were tracked for 720 h. PX75-T50 has an initial conductivity of
111.6 mS cm−1, which decreased to 84.3 mS cm−1 after 720 h,
thereby retaining 76% of the initial performance (Figure 3b).
The initial decrease is likely due to hydroxide-induced
degradation of residual pyrazolium chlorides or nucleophilic
cleavage of cross-links activated by stress on the polymer films.
PX10-T50 retained 26.5 mS cm−1, 61% of its initial value, after
the same test conditions. PX75-T50 displayed a slow decrease
of conductivity to be plateaued after 532 h and stabilized at 84
mS cm−1 after 30 days. NMR studies of the model cations (4
and 7) and membrane stability data support two operative
degradation mechanisms: cleavage of the ether linkages and
deprotonation of alpha protons of the pyrazolium 4-
substituents (Figures 2, 3b, S9, and S10). We do recognize
Highly cross-linked membranes are mechanically resistive to
water uptake (swelling). Our cross-linked membranes display
low water uptakes of less than 10% (Figures 3c and S18a) with
values of 7.9, 8.5, and 8.8% at 80 °C observed for PX75-T50,
PX40-T50, and PX10-T50, respectively. As expected, the water
uptake is inversely correlated to m, and in accord, swelling
ratios of 2.6, 4.3, and 5.8% at 80 °C were observed for PX75-
T50, PX40-T50, and PX10-T50, respectively (Figure S18b). As
can be seen from data summarized from the literature (Figure
3c), hydroxide conductivity generally increases with water
uptake because of the hydration of cationic centers. The
literature reports that samples having higher conductivity over
100 mS cm−1 and water uptake of greater than 50% are
common. Excessive water uptake membrane swelling leads to
decreased fuel cell output through compromised contact with
the two electrodes. Membranes that are highly conductive with
minimal water uptake and swelling are desired, and finding
systems that do not suffer from reduced conductance at low
water uptake levels is rare. In this context, PX75-T50, which
has reduced water uptake and the highest conductivity, is a key
advancement. The effectiveness of our design is also revealed
by IEC analysis. All of our cross-linked membranes have low
IECs of 0.91, 1.03, and 1.12 mmol g−1 for PX75-T50, PX40-
T50, and PX10-T50, respectively, but all maintain high
conductivity (Figure 3d and Table S3). PX75-T50 has the
remarkable property of having conductivity higher than 100
mS cm−1 with an extremely low IEC of 0.91 mmol g−1. Also
within our series, it has the highest conductivity and lowest
IEC, which supports the concept that the assembly of the
D
DOI: 10.1021/jacs.9b08749
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 4. Computational studies. (a−d) DFT analysis of charge delocalization in triptycene-connected pyrazolium cations. (a) Electrostatic
potential (ESP) mapped on the 0.01 e·Å−3 isodensity surface of model structures where two pyrazoliums (Pyr) are connected to a triptycene (Trp)
and the corresponding non-cross-linked reference. (b) LUMO of the chemical structure in (a). (c) ESP mapped on the 0.01 e·Å−3 isodensity
surface of model structures where three pyrazoliums and two triptycenes are connected in an alternating way and the corresponding non-cross-
linked reference. For (a) and (c), the arrows and associated numbers point at the maximum value of the ESP map, and ESP was projected on an
isosurface of the electron density (0.01 e·Å−3). (d) LUMO of the chemical structure in (c). The ends of all methyl or methoxy groups are supposed
to have phenyls. However, they were all omitted for the convenience of calculation. See Figure S21 and Files S2−S5 for chemical structures and the
ones with coordinates, used in (a−d). (e−j) 2D-LMC studies of ion hopping. (e) Anion hopping probability distribution map for a polymer
network constructed by orderly placed cross-linkers for m = 75%. (f, g) Calculated OH− diffusivity and conductivity at different m as a function of
temperature, respectively. (h) Anion hopping probability distribution map for a polymer network constructed by randomly placed cross-linkers for
m = 10%. (i, j) Calculated OH− diffusivity and conductivity at different m values as a function of temperature, respectively. Probability maps of the
entire cases are in Figure S23 and see SI for details of the calculation.
cations in chains through a combination of cross-linking and
grafting creates the facilitated transport we refer to as an ionic
highway.
analysis (Figure S20) showed slightly higher char yields with
heating to 1000 °C for PX75-T50 of 51% relative to PX40-T50
and PX10-T50 with 49 and 46%, which is also consistent with
a higher m. Overall, achieving the highest conductivity with the
lowest water uptake and IEC (Figure 3) demonstrates that the
formation of ionic highways is a promising approach for
efficient hydroxide transport (Figures 1, 3, and S17).
These observations prompted us to investigate the internal
sample morphology and structures of the three samples having
a different m linking. X-ray scattering experiments revealed that
although the lightly cross-linked membrane, PX10-T50, had
structures with d-spacing of 16.8 nm the more highly cross-
linked membranes, PX40-T50 and PX75-T50, displayed
featureless scattering patterns (Figure S19). The X-ray
scattering data suggest that a high m may remove nanoscale
phase segregations in the membrane. Thermogravimetric
COMPUTATIONAL STUDIES ON CHARGE
DELOCALIZATION AND IONIC HIGHWAYS
■
Motivated by these findings, we decided to investigate the
microscopic origin of charge transport in ionic highways using
E
DOI: 10.1021/jacs.9b08749
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 5. H2/O2 AEMFC performance with PX75-T50. (a) Polarization curves with different catalysts and higher flow rates. Data points with filled
circles and dotted lines are for cell voltage and power density, respectively. (b) AEMFC durability test at a constant current density of 0.4 A cm−2 at
80 °C, with Pt−Ru/C anode, Pt/C cathode, and high gas flow. Reference condition means Pt/C for both electrodes, 0.2 L min−1 for both gases, 80
°C, 100% RH, and zero backpressure. Besides the reference, the legend contains variables changed compared to the reference. High flow means 1.0
L min−1 for both gases. Catalyst loading for all electrodes is 0.5 mg of metal (or bimetal) per unit cm2.
atomic-scale simulations. We focused first on molecular models
of pyrazolium cations cross-linked with triptycene units, which
we could compare directly to non-cross-linked structures. An
analysis of the electrostatic potential (ESP) of model
geometries, which was derived from electronic structure
calculations based on DFT, indicates that the connections of
the pyrazolium (Pyr) to triptycene (Trp) lead to a significant
lowering of the maximum of the electrostatic potential,
corresponding to a softening of the charge density (Figure
4a,c, Note S3; Figure S21 for model chemical structures).
Compared to the corresponding non-cross-linked control
systems, the doubly (Pyr-Trp-Pyr) and triply charged (Pyr-
Trp-Pyr-Trp-Pyr) systems, respectively, show a decrease of 0.2
and 0.5 eV of the ESP maximum. Assuming that a hydroxide
can be reasonably well approximated as a point charge, this
results in weaker hydroxide interactions with the cross-linked
cations, which leads to a lowering of the activation energy for
the hopping between adjacent cationic sites in our molecular
models and improved alkaline stability. Furthermore, the
extensively delocalized lowest unoccupied molecular orbital
(LUMO) of both doubly charged and triply charged cross-
linked systems provides additional evidence of charge
delocalization over the basal benzene wings of the triptycene
units (Figure 4b,d).
To further investigate the system, we constructed model
systems of ionic networks using the 2D lattice Monte Carlo
(2D-LMC) model,38,39 wherein we simulated ionic con-
ductivity in differently formed channels with varying m (Figure
4e−j, Note S4). We built two systems where conducting
channels are orderly or randomly aligned. In the two systems,
m values were changed to be 75, 40, and 10% (Table S1 was
used for equivalence of monomers). The probability of
hydroxide transport between cross-linked and non-cross-linked
sites was determined by the Arrhenius equation, where the
activation energies for transport between linked cations and
nonlinked cations were extrapolated from our experimental
hydroxide conductivity data (see SI for details). Diffusivity and
conductivity values were higher in the ordered systems (Figure
4f, g, i, and j). The conductivity from an ordered system with
10% cross-linking was higher than a random system with 75%
cross-linking. This observation implies that construction of
ionic highways is necessary for the significant improvements in
the conductivity of the network. At a lower m, the cross-links
were scattered, making it challenging to form ion transport
pathways (highways) throughout the entire polymer network.
At the same time, a higher m formed longer cross-linked path
lengths, which should significantly boost the global diffusion of
the anions. The conductivities predicted by the 2D-LMC
model are in good agreement with the experimental measure-
ments, capturing both the positive correlations between
conductivity and m and the Arrhenius dependence of
conductivity on temperature (Figure 4g, j, and Figure S22
and S25). The fact that the ordered channel model gives a
more accurate prediction for m = 75% and the random channel
model gives an accurate prediction for m = 10% reveals the
nature of our membranes at different degrees of cross-linking.
FUEL CELL PERFORMANCE
■
Membrane electrode assemblies using PX75-T50 demonstra-
ted high power density and stability, generating typical
polarization curves that include an activation region, an
Ohmic region, and a mass transport region (Figures 5a and
S26). AEMFC performance progressively improved by
changing catalyst for the anode and increasing gas flow rate.
A reference condition (Pt/C, 0.5 mgPt cm−2, for both
electrodes, flow rate of 0.2 L min−1 for both gases, 80 °C,
100% RH, and zero backpressure) enabled a peak power
density of 0.26 W cm−2 at 0.59 A cm−2 (Figure 5a), which is
comparable to benchmark systems at the same working
conditions.6,11,40 Changing the anode catalyst to Pt−Ru/C
(0.5 mgPt cm−2 and 0.25 mgRu cm−2) increased the peak power
density to 0.45 W cm−1. Increasing gas flow rate to 1.0 L min−1
for both gases, while keeping Pt−Ru/C for the anode, further
increased the peak power density to 0.73 W cm−2. As can be
seen from previous studies, more active catalysts reduce
activation barriers for the hydrogen oxidation reaction, and
faster gas transport gives higher rates (Figure S27).6,33 We also
demonstrated that the fuel cells with PX75-T50 give high
performance with more economical silver catalysts. Ag/C, 0.5
mgAu cm−2, on cathode resulted in 0.47 W cm−2 peak power
density, with Pt−Ru/C for the anode and the high flow rate.
Although a greater voltage drop was observed from the
polarization curve with the Ag/C cathode, this is a promising
result considering substantial power output and significantly
lower price of silver (the price of Ag is ≈2% of that of Pt).
To test single-cell durability, a constant current discharge of
0.40 A cm−2 was applied at 80 °C, for 400 h under the
conditions used for “Pt−Ru/C Anode & High Flow”. Although
F
DOI: 10.1021/jacs.9b08749
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Author Contributions
there is some initial loss, only a 10% voltage decrease was
observed (Figure 5b) over this entire test. This reliable long-
term cell operation is consistent with the membrane stability
data (Figure 3b), and the mechanical stability of the
membrane. PX75-T50 displays only 8% water uptake and 3%
swelling ratio at 80 °C. This improved mechanical stability, as
compared to the previous reports,7,11,13 results in reliable long-
term operation by maintaining conformal contact between the
membrane and the electrode assembly.
⊥Y.W. and A.F.-L. contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by a seed grant from the MIT Energy
Initiative. Yanming W. and A.F.-L. acknowledge the Toyota
Research Institute (TRI), the Extreme Science and Engineer-
ing Discovery Environment, supported by the National Science
Foundation grant number ACI-1053575, and the National
Energy Research Scientific Computing Center, a DOE Office
of Science User Facility supported by the Office of Science of
the U.S. Department of Energy under Contract No. DE-AC02-
05CH11231, for providing computational support.
CONCLUSIONS AND OUTLOOK
■
We synthesized a new generation of AEMs from pyrazolium
cross-linked triptycene copolymers, wherein charges are
delocalized to construct ionic highways for lower conduction
barriers and increased chemical/mechanical stability. The
membranes achieve enhanced properties with decreased
water uptake and lower IEC, features that diverge from the
present trends among AEMs. These rare features are the result
of ionic highway networks that are formed from the unique
cross-linked structures in the membranes. Computational
studies on ESP and 2D-LMC reveal how extended networks
of pyrazolium and triptycene produce ionic highways by
lowering diffusion barriers. The peak power density of 0.73 W
cm−2 from platinum-group-metal electrodes, 0.47 W cm−2
from the silver-based cathode, and stable long-term cell
operation make the materials further qualify this approach as
promising. The concept of assembling an ionic highway using
charge-delocalized cations as cross-linkers is herewith estab-
lished to develop new generations of high performing AEMs
for cleaner energy generation.
REFERENCES
■
(1) Cano, Z. P.; Banham, D.; Ye, S.; Hintennach, A.; Lu, J.; Fowler,
M.; Chen, Z. Batteries and Fuel Cells for Emerging Electric Vehicle
Markets. Nat. Energy 2018, 3 (4), 279−289.
(2) Shao, Y.; Yin, G.; Wang, Z.; Gao, Y. Proton Exchange Membrane
Fuel Cell from Low Temperature to High Temperature: Material
Challenges. Journal of Power Sources; Elsevier, 2007; pp 235−242.
(3) Hickner, M. A.; Herring, A. M.; Coughlin, E. B. Anion Exchange
Membranes: Current Status and Moving Forward. Journal of Polymer
Science, Part B: Polymer Physics; John Wiley & Sons, Ltd., 2013; pp
1727−1735.
(4) Ran, J.; Wu, L.; He, Y.; Yang, Z.; Wang, Y.; Jiang, C.; Ge, L.;
Bakangura, E.; Xu, T. Ion Exchange Membranes: New Developments
and Applications. J. Membr. Sci. 2017, 522, 267−291.
(5) Gottesfeld, S.; Dekel, D. R.; Page, M.; Bae, C.; Yan, Y.; Zelenay,
P.; Kim, Y. S. Anion Exchange Membrane Fuel Cells: Current Status
and Remaining Challenges. J. Power Sources 2018, 375, 170−184.
(6) Maurya, S.; Fujimoto, C. H.; Hibbs, M. R.; Narvaez Villarrubia,
C.; Kim, Y. S. Toward Improved Alkaline Membrane Fuel Cell
Performance Using Quaternized Aryl-Ether Free Polyaromatics.
Chem. Mater. 2018, 30 (7), 2188−2192.
(7) Wang, J.; Zhao, Y.; Setzler, B. P.; Rojas-Carbonell, S.; Ben
Yehuda, C.; Amel, A.; Page, M.; Wang, L.; Hu, K.; Shi, L.; et al.
Poly(Aryl Piperidinium) Membranes and Ionomers for Hydroxide
Exchange Membrane Fuel Cells. Nat. Energy 2019, 4, 392−398.
(8) Hossain, M. M.; Hou, J.; Wu, L.; Ge, Q.; Liang, X.; Mondal, A.
N.; Xu, T. Anion Exchange Membranes with Clusters of Alkyl
Ammonium Group for Mitigating Water Swelling but Not Ionic
Conductivity. J. Membr. Sci. 2018, 550, 101−109.
(9) Arges, C. G.; Ramani, V. Two-Dimensional NMR Spectroscopy
Reveals Cation-Triggered Backbone Degradation in Polysulfone-
Based Anion Exchange Membranes. Proc. Natl. Acad. Sci. U. S. A.
2013, 110 (7), 2490−2495.
(10) Dekel, D. R. Review of Cell Performance in Anion Exchange
Membrane Fuel Cells. J. Power Sources 2018, 375, 158−169.
(11) Lee, K. H.; Cho, D. H.; Kim, Y. M.; Moon, S. J.; Seong, J. G.;
Shin, D. W.; Sohn, J. Y.; Kim, J. F.; Lee, Y. M. Highly Conductive and
Durable Poly(Arylene Ether Sulfone) Anion Exchange Membrane
with End-Group Cross-Linking. Energy Environ. Sci. 2017, 10 (1),
275−285.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.9b08749.
■
S
Details about materials, characterization methods,
synthetic procedures, characterization data, and compu-
tational studies (PDF)
Calculated structure for 4-methyl-3,5-diphenoxy-1,2-
diphenyl-pyrazolium without H+ (XYZ)
Calculated structure for doubly-charged (Pyr-Trp-
Pyr) (XYZ)
Calculated structure for triply-charged (Pyr-Trp-Pyr-
Trp-Pyr) (XYZ)
Calculated structure for non-cross-linked control system
of the doubly-charged (Pyr-Trp-Pyr) (XYZ)
Calculated structure for non-cross-linked control system
of the triply-charged (Pyr-Trp-Pyr-Trp-Pyr) (XYZ)
AUTHOR INFORMATION
■
Corresponding Authors
*E-mail: jcg@mit.edu.
*E-mail: tswager@mit.edu.
(12) Hu, E. N.; Lin, C. X.; Liu, F. H.; Yang, Q.; Li, L.; Zhang, Q. G.;
Zhu, A. M.; Liu, Q. L. Cross-Linked Poly(Vinylbenzyl Chloride)
Anion Exchange Membranes with Long Flexible Multihead for Fuel
Cells. ACS Appl. Energy Mater. 2018, 1 (7), 3479−3487.
(13) Wang, L.; Peng, X.; Mustain, W. E.; Varcoe, J. R. Radiation-
Grafted Anion-Exchange Membranes: The Switch from Low- to
High-Density Polyethylene Leads to Remarkably Enhanced Fuel Cell
Performance. Energy Environ. Sci. 2019, 12, 1575−1579.
ORCID
Yoonseob Kim: 0000-0002-6892-8281
Yanming Wang: 0000-0002-0912-681X
Arthur France-Lanord: 0000-0003-0586-1945
You-Chi Mason Wu: 0000-0002-6585-7908
Sibo Lin: 0000-0001-5922-6694
Jeffrey C. Grossman: 0000-0003-1281-2359
Timothy M. Swager: 0000-0002-3577-0510
(14) Shin, D. W.; Guiver, M. D.; Lee, Y. M. Hydrocarbon-Based
Polymer Electrolyte Membranes: Importance of Morphology on Ion
G
DOI: 10.1021/jacs.9b08749
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
́
Transport and Membrane Stability. Chemical Reviews; American
Chemical Society, 2017; pp 4759−4805.
(15) Kawasumi, K.; Wu, T.; Zhu, T.; Chae, H. S.; Van Voorhis, T.;
Baldo, M. A.; Swager, T. M. Thermally Activated Delayed
Fluorescence Materials Based on Homoconjugation Effect of
Donor-Acceptor Triptycenes. J. Am. Chem. Soc. 2015, 137 (37),
11908−11911.
(16) Hugar, K. M.; Kostalik, H. A.; Coates, G. W. Imidazolium
Cations with Exceptional Alkaline Stability: A Systematic Study of
Structure-Stability Relationships. J. Am. Chem. Soc. 2015, 137 (27),
8730−8737.
(17) de Paul Nzuwah Nziko, V.; Shih, J.-L.; Jansone-Popova, S.;
Bryantsev, V. S. Quantum Chemical Prediction of PK a Values of
Cationic Ion- Exchange Groups in Polymer Electrolyte Membranes. J.
Phys. Chem. C 2018, 122, 2490−2501.
(33) Wang, L.; Brink, J. J.; Liu, Y.; Herring, A. M.; Ponce-Gonzalez,
J.; Whelligan, D. K.; Varcoe, J. R. Non-Fluorinated Pre-Irradiation-
Grafted (Peroxidated) LDPE-Based Anion-Exchange Membranes
with High Performance and Stability. Energy Environ. Sci. 2017, 10
(10), 2154−2167.
(34) Lee, W. H.; Mohanty, A. D.; Bae, C. Fluorene-Based Hydroxide
Ion Conducting Polymers for Chemically Stable Anion Exchange
Membrane Fuel Cells. ACS Macro Lett. 2015, 4 (4), 453−457.
(35) Ran, J.; Wu, L.; Wei, B.; Chen, Y.; Xu, T. Simultaneous
Enhancements of Conductivity and Stability for Anion Exchange
Membranes (AEMs) through Precise Structure Design. Sci. Rep. 2015,
4, 6486.
(36) Li, N.; Leng, Y.; Hickner, M. A.; Wang, C. Y. Highly Stable,
Anion Conductive, Comb-Shaped Copolymers for Alkaline Fuel
Cells. J. Am. Chem. Soc. 2013, 135 (27), 10124−10133.
(37) Liu, L.; Chu, X.; Liao, J.; Huang, Y.; Li, Y.; Ge, Z.; Hickner, M.
A.; Li, N. Tuning the Properties of Poly(2,6-Dimethyl-1,4-Phenylene
Oxide) Anion Exchange Membranes and Their Performance in H 2
/O 2 Fuel Cells. Energy Environ. Sci. 2018, 11 (2), 435−446.
(38) Avramov, I.; Milchev, A.; Argyrakis, P. Diffusion in a Random
Medium: A Monte Carlo Study. Phys. Rev. E: Stat. Phys., Plasmas,
Fluids, Relat. Interdiscip. Top. 1993, 47 (4), 2303−2307.
(39) Jacoboni, C.; Reggiani, L. The Monte Carlo Method for the
Solution of Charge Transport in Semiconductors with Applications to
Covalent Materials. Rev. Mod. Phys. 1983, 55 (3), 645−705.
(40) Wang, X. Q.; Lin, C. X.; Liu, F. H.; Li, L.; Yang, Q.; Zhang, Q.
G.; Zhu, A. M.; Liu, Q. L. Alkali-Stable Partially Fluorinated
Poly(Arylene Ether) Anion Exchange Membranes with a Claw-Type
Head for Fuel Cells. J. Mater. Chem. A 2018, 6 (26), 12455−12465.
(18) Kim, D. S.; Fujimoto, C. H.; Hibbs, M. R.; Labouriau, A.; Choe,
Y. K.; Kim, Y. S. Resonance Stabilized Perfluorinated Ionomers for
Alkaline Membrane Fuel Cells. Macromolecules 2013, 46 (19), 7826−
7833.
(19) Swager, T. M. The Molecular Wire Approach to Sensory Signal
Amplification. Acc. Chem. Res. 1998, 31 (5), 201−207.
(20) Mohanty, A. D.; Bae, C. Mechanistic Analysis of Ammonium
Cation Stability for Alkaline Exchange Membrane Fuel Cells. J. Mater.
Chem. A 2014, 2 (41), 17314−17320.
(21) Hugar, K. M.; You, W.; Coates, G. W. Protocol for the
Quantitative Assessment of Organic Cation Stability for Polymer
Electrolytes. ACS Energy Lett. 2019, 4 (7), 1681−1686.
́
(22) Fernandez, I.; Dyker, C. A.; DeHope, A.; Donnadieu, B.;
Frenking, G.; Bertrand, G. Exocyclic Delocalization at the Expense of
Aromaticity in 3,5-Bis(π-Donor) Substituted Pyrazolium Ions and
Corresponding Cyclic Bent Allenes. J. Am. Chem. Soc. 2009, 131 (33),
11875−11881.
(23) Pan, J.; Chen, C.; Li, Y.; Wang, L.; Tan, L.; Li, G.; Tang, X.;
Xiao, L.; Lu, J.; Zhuang, L. Constructing Ionic Highway in Alkaline
Polymer Electrolytes. Energy Environ. Sci. 2014, 7 (1), 354−360.
(24) Ge, Q.; Ran, J.; Miao, J.; Yang, Z.; Xu, T. Click Chemistry
Finds Its Way in Constructing an Ionic Highway in Anion-Exchange
Membrane. ACS Appl. Mater. Interfaces 2015, 7 (51), 28545−28553.
(25) Lavallo, V.; Dyker, C. A.; Donnadieu, B.; Bertrand, G. Synthesis
and Ligand Properties of Stable Five-Membered-Ring Allenes
Containing Only Second-Row Elements. Angew. Chem. 2008, 120
(29), 5491−5494.
(26) Fan, J.; Willdorf-Cohen, S.; Schibli, E. M.; Paula, Z.; Li, W.;
Skalski, T. J. G.; Sergeenko, A. T.; Hohenadel, A.; Frisken, B. J.;
Magliocca, E.; Mustain, W. E.; Diesendruck, C. E.; Dekel, D. R.;
Holdcroft, S. Poly(Bis-Arylimidazoliums) Possessing High Hydroxide
Ion Exchange Capacity and High Alkaline Stability. Nat. Commun.
2019, 10, 2306.
(27) Kim, Y.; Moh, L. C. H.; Swager, T. M. Anion Exchange
Membranes: Enhancement by Addition of Unfunctionalized Tripty-
cene Poly(Ether Sulfone)S. ACS Appl. Mater. Interfaces 2017, 9 (49),
42409−42414.
(28) Moh, L. C. H.; Goods, J. B.; Kim, Y.; Swager, T. M. Free
Volume Enhanced Proton Exchange Membranes from Sulfonated
Triptycene Poly(Ether Ketone). J. Membr. Sci. 2018, 549, 236−243.
(29) Li, Y.; Cao, R.; Lippard, S. J. Design and Synthesis of a Novel
Triptycene-Based Ligand for Modeling Carboxylate-Bridged Diiron
Enzyme Active Sites. Org. Lett. 2011, 13 (19), 5052−5055.
(30) Navale, T. S.; Rathore, R. A Practical Synthesis of 1,4,5,8-
Tetramethoxyanthracene from Inexpensive and Readily Available 1,8-
Dihydroxyanthraquinone. Synthesis 2012, 44 (5), 805−809.
(31) Hu, E. N.; Lin, C. X.; Liu, F. H.; Wang, X. Q.; Zhang, Q. G.;
Zhu, A. M.; Liu, Q. L. Poly(Arylene Ether Nitrile) Anion Exchange
Membranes with Dense Flexible Ionic Side Chain for Fuel Cells. J.
Membr. Sci. 2018, 550, 254−265.
(32) Li, L.; Wang, X. Q.; Zhu, A. M.; Lin, C. X.; Yang, Q.; Liu, Q. L.;
Zhang, Q. G. Highly Conductive Anion Exchange Membranes with
Long Flexible Multication Spacer. J. Membr. Sci. 2018, 553, 209−217.
H
DOI: 10.1021/jacs.9b08749
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX