A comparative study of side-chain-type poly(ether ether ketone) anion exchange membrane functionalized with different hetero-cycloaliphatic quaternary ammonium groups

Article abstract of DOI:10.1039/C8RA09840A

Anion exchange membranes based on side-chain-type quarternized poly(ether ether ketone)s (QPEEKs), containing different hetero-cycloaliphatic quaternary ammonium groups, were prepared via a multi-step procedure, including polycondensation with a monomer containing pendant methylphenyl groups, bromomethylation, and followed by quaternization with 1-methylpyrrolidine (MPY), 1-methylpiperidine (MPRD), 1-methylimidazole (MIDZ) and N-methyl morpholine (MMPH), respectively. The properties of these membranes were then compared with the properties of conventional quarternized poly(ether ether ketone)s containing benzyltrimethylammonium (QPEEK-TMA). Model compounds, QMPY, QMPRD, QMIDZ and QMMPH, were synthesized and used to quantitatively compare the alkaline stability of hetero-cycloaliphatic quaternary ammonium groups using ¹H NMR. The results of this study indicated that the alkaline stability of all these model compounds is in the order of QMPY > QMPRD > QTMA > QMIDZ > QMMPH. These QPEEKs membranes display superior thermal, dimensional and mechanical stability. QPEEK-TMA, QPEEK-MPY and QPEEK-MPRD exhibit higher hydroxide conductivities and lower activation energies than QPEEK-MIDZ and QPEEK-MMPH. Furthermore, after exposure to 1 M NaOH at 60 °C for 24 days, QPEEK-MPY and QPEEK-MPRD demonstrated a loss in hydroxide conductivity of 28.5% and 33.4%, respectively. Both values were lower than that of QPEEK-TMA (37.7%). Therefore, among these side-chain-type QPEEKs membranes, QPEEK-MPY and QPEEK-MPRD, containing benzylmethylpyrrolidinium and benzylmethylpiperidinium as cation head-groups, are promising AEM materials for fuel cell construction.

Full text of DOI:10.1039/C8RA09840A

RSC Advances
View Article Online
View Journal | View Issue
PAPER
A comparative study of side-chain-type poly(ether
ether ketone) anion exchange membrane
functionalized with different hetero-cycloaliphatic
quaternary ammonium groups†
Cite this: RSC Adv., 2019, 9, 7975
a
Dan Yao,a Tonghui Wei,b Lichao Shang,a Hui Naa and Chengji Zhao
*
Anion exchange membranes based on side-chain-type quarternized poly(ether ether ketone)s (QPEEKs),
containing different hetero-cycloaliphatic quaternary ammonium groups, were prepared via a multi-step
procedure, including polycondensation with
a monomer containing pendant methylphenyl groups,
bromomethylation, and followed by quaternization with 1-methylpyrrolidine (MPY), 1-methylpiperidine
(MPRD), 1-methylimidazole (MIDZ) and N-methyl morpholine (MMPH), respectively. The properties of these
membranes were then compared with the properties of conventional quarternized poly(ether ether ketone)
s containing benzyltrimethylammonium (QPEEK-TMA). Model compounds, QMPY, QMPRD, QMIDZ and
QMMPH, were synthesized and used to quantitatively compare the alkaline stability of hetero-cycloaliphatic
quaternary ammonium groups using 1H NMR. The results of this study indicated that the alkaline stability of
all these model compounds is in the order of QMPY > QMPRD > QTMA > QMIDZ > QMMPH. These
QPEEKs membranes display superior thermal, dimensional and mechanical stability. QPEEK-TMA, QPEEK-
MPY and QPEEK-MPRD exhibit higher hydroxide conductivities and lower activation energies than QPEEK-
MIDZ and QPEEK-MMPH. Furthermore, after exposure to 1 M NaOH at 60 ꢀC for 24 days, QPEEK-MPY and
QPEEK-MPRD demonstrated a loss in hydroxide conductivity of 28.5% and 33.4%, respectively. Both values
were lower than that of QPEEK-TMA (37.7%). Therefore, among these side-chain-type QPEEKs membranes,
QPEEK-MPY and QPEEK-MPRD, containing benzylmethylpyrrolidinium and benzylmethylpiperidinium as
cation head-groups, are promising AEM materials for fuel cell construction.
Received 30th November 2018
Accepted 7th February 2019
DOI: 10.1039/c8ra09840a
rsc.li/rsc-advances
In the past decades, hydrogen fuel cells have become the new
and promising energy technology due to their high energy
Introduction
conversion efficiency, superior portability and environmental
friendliness.4,5 As one type of fuel cells, anion exchange
membrane fuel cells (AEMFCs) have drawn tremendous atten-
tion due to their potential advantages over common proton
exchange membrane fuel cells (PEMFCs), including enhanced
oxygen reduction kinetics and faster fuel oxidation kinetics
under alkaline conditions. This technology would allow using
low-cost, non-precious metal catalysts and reducing the overall
cost of AEMFCs greatly.6–10 As the core component in AEMFCs,
anion exchange membrane (AEM) serves as the electrolyte for
hydroxide ion transport between electrodes and separates the
fuel from the oxidant.11 Hence, in order to ensure efficient and
stable operation of AEMFCs, the ideal AEM must possess high
enough hydroxide conductivity (the conductivity must meet the
requirement: >10 mS cmꢁ1 at room temperature and close to
100 mS cmꢁ1 as far as possible at working temperature), satis-
factory thermal and chemical stability, and sufficient lm
forming ability, mechanical strength and dimensional
stability.12
The current pattern of mankind's energy consumption is facing
more and more problems and challenges, including the exces-
sive dependence on traditional fossil fuels, such as coal and oil.
Also, it causes environmental pollution and greenhouse gas
emission due to combustion of fossil fuels.1 Therefore, the
development of clean and sustainable energy sources for
replacing fossil fuels has been initiated. At present, there are
many possible alternative energy sources to the combustion of
fossil fuels under investigation, for example, solar energy, wind
energy, hydrogen energy, nuclear energy, tidal energy and bio-
energy. However, utilization of new energy requires energy
generation, storage and release.2,3
aAlan G. MacDiarmid Institute, College of Chemistry, Jilin University, Changchun
130012, P. R. China. E-mail: zhaochengji@jlu.edu.cn; Fax: +86-0431-85168870; Tel:
+86-0431-85168870
bSchool of Mechanical and Aerospace Engineering, Jilin University, Changchun
130025, P. R. China
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c8ra09840a
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 7975–7983 | 7975

View Article Online
RSC Advances
Paper
Quaternary ammonium (QA) functionalized polymer elec- Menshutkin reaction using 1-methylpyrrolidine (MPY), 1-meth-
trolyte membranes, especially those using trimethylamine ylpiperidine (MPRD), 1-methylimidazole (MIDZ) and N-methyl
(TMA) as the quaternary reagent, have been widely investigated morpholine (MMPH) as quaternary reagents, respectively. Quar-
as potential AEM materials. Most of them were prepared by ternized poly(ether ether ketone)s (QPEEKs) containing benzyl-
tethering QA groups to the backbone of chloromethylated or trimethylammonium head-groups were also prepared using
bromomethylated aromatic polymers by immersing them in trimethylamine (TMA) for comparison. The purpose of this study
TMA solution.13,14 However, conventional QA-functionalized was mainly to compare the characteristics of AEMs containing
AEMs exhibit very low ionic conductivity and poor alkaline different aliphatic-heterocyclic QA groups, including ionic
stability at high pH and high temperature conditions. There- exchange capacity (IEC), thermal stability, dimensional stability,
fore, many researchers have concentrated on molecular design mechanical properties, ionic conductivity, especially the effect of
to improve the performance of AEMs. Some achievements have hydroxide conductivity on alkaline stability, mechanical proper-
been made in this eld by nding alternatives to QA functional ties and 1H NMR spectra aer aging in alkaline solution. In
groups. Marino et al. studied the alkaline stability of different addition,
a series of model compounds with aliphatic-
cationic model compounds in high-pH solutions by deter- heterocyclic QA cations were synthesized and used to charac-
mining the half-times of the compounds using NMR analysis. terize degradation of these cations using 1H NMR spectroscopy. It
The results suggested that aliphatic-heterocyclic QA head- would help to clarify the alkaline stability of QPEEKs with rele-
groups exhibited excellent alkaline stability compared to that vant cationic groups on side chain when exposed to alkaline
of the conventional benzyltrimethylammonium head-groups, solutions. In this paper, the AEMs with side-chain-type congu-
especially under extreme accelerated experiment conditions.15 ration and different aliphatic-heterocyclic QA head-groups are
For example, piperidinium demonstrated high resistance to discussed in detail, comparing related properties, especially ionic
both nucleophilic substitution and elimination in 6 M NaOH at conductivity and alkaline stability.
elevated temperatures. Recently, many researchers also re-
ported that AEMs, based on polymers containing N-alkylpyrro-
lidinium and benzyl-N-methylmorpholinium head-groups,
Experimental section
Materials
exhibited higher alkaline stability than the commercial
Tokuyama A201 AEMs.16–18 Ponce-gonzalez et al. reported that
´
4,4-Diuorobenzophenone (DFBP) was purchased from Yanbian
Longjing chemical company. N-Bromosuccinimide (NBS, 99%),
benzoyl peroxide (BPO, 99%), MPY (98%), MPRD (97%), MIDZ
(99%), MMPH (99%), p-benzoquinone (99%) and p-toluidine (AR,
99%) were obtained from Aladdin Reagent Co. Ltd. Trimethyl-
amine in methanol (TMA, 3.2 mol Lꢁ1) was purchased from TCI.
All other reagents and solvents were purchased from Tianjin
Tiantai chemical company and used without further purication.
radiation-graed AEMs, containing saturated-heterocyclic
head-groups, displayed a relatively good alkaline stability,
conductivity and in situ fuel cell performance.19 They believe
that the benzyl-N-methylpyrrolidinium head-group is the most
promising alternative to benzyltrimethylammonium head-
group.
As recently reported, locating the sulfonic acid groups on the
side chains of the polymer backbones could improve the
mobility of ions, thus, promoting phase separation and
increasing proton conductivity. Nowadays, in some studies of
AEMs, a similar method has been utilized to tether cationic
head-groups to the side chain of the polymer for improving the
hydroxide conductivity of AEMs without sacricing their
favourable mechanical properties.20,21
Aromatic polymers, such as poly(ether ether ketone)s
(PEEK), are the most extensively used starting materials for the
preparation of AEMs by graing QA groups onto the main
polymer backbone.22–24 However, it was reported that the back-
bone undergoes cleavage of C–O bonds under the attack of
hydroxide anions when electron-withdrawing cationic head-
groups are in close proximity.25 It was postulated that attach-
ing cationic head-groups to the backbone via a bulky spacer in
the side chain could isolate the electron-withdrawing effect of
the cation from the aromatic C–O bonds, thereby, enhancing
the alkaline stability of AEMs.
Synthesis of model compounds
The synthesis of benzyltrimethylammonium model compound
(QTMA) was described by us earlier.26 One mL (8.69 mmol)
benzyl chloride was dissolved in 40 mL ethyl alcohol in
a 100 mL three-neck round bottom ask, equipped with
a mechanical stirrer and a condenser, and then 5.43 mL (17.38
mmol) TMA was added to the ask. The reaction mixture was
ꢀ
stirred at 78 C for 6 h. Finally, the mixture was washed with
diethyl ether and ltered. The product was washed with diethyl
ether eight times and dried in a vacuum oven at 80 ꢀC for 48 h.
As shown in Scheme 1, benzylmethylpyrrolidine, benzylme-
thylpiperidine,
benzylmethylimidazole
and
In this study, we synthesized a series of PEEKs with pendant
methylphenyl groups, which gave us a chance to prepare side-
chain-type AEMs bearing cationic head-groups, linked to back-
bones via bulky spacers in the side chains. Aer bromomethy-
lation of methylphenyl-containing PEEKs, AEMs, based on
Scheme 1 Chemical structures of model compounds: QTMA (a),
aliphatic-heterocyclic QA head-groups, were prepared via QMPY (b), QMPRD (c), QMIDZ (d) and QMMPH (e).
7976 | RSC Adv., 2019, 9, 7975–7983
This journal is © The Royal Society of Chemistry 2019

View Article Online
Paper
RSC Advances
benzylmethylmorpholine model compounds were synthesized
The synthesis of bromomethylated PEEK is shown in
using the similar procedure, employing MPY, MPRD, MIDZ and Scheme 2. Bromomethylation of PEEK, a free-radical substitution
MMPH as the quaternary reagents. These model compounds reaction, is mild and controllable, which distinguishes it from
with aliphatic-heterocyclic QA are denoted as QMPY, QMPRD, the usual chloromethylation reaction. NBS and BPO were used as
QMIDZ and QMMPH, respectively. Taking QMPY as an the brominating reagent and the initiator, respectively. The
example, the synthesis was carried out as follows. One mL (8.69 product was denoted as BrPEEK-X, where X is the ratio of NBS to
mmol) benzyl chloride was dissolved in 40 mL ethyl alcohol in PEEK. Taking BrPEEK-1.2 as an example, the detailed procedure
a 100 mL three-neck round bottom ask equipped with is as follows. The powder of PEEK (3.00 g, 7.94 mmol) was dis-
a mechanical stirrer and a condenser, and then 1.85 mL (17.38 solved in dry 1,1,2,2-tetrachloroethane in a 100 mL three-neck
mmol) MPY was added into the ask. The reaction mixture was round bottom ask, equipped with a mechanical stirrer,
stirred at 78 ꢀC for 6 h. Finally, the product was precipitated in a nitrogen inlet and a condenser. When the color of the mixture
diethyl ether, ltered off, washed with diethyl ether eight times became orange, NBS (1.70 g, 9.52 mmol) and BPO (0.23 g, 0.95
ꢀ
and dried in a vacuum oven at 80 C for 48 h.
mmol) were added into the ask. The mixture was heated to
180 ꢀC for 10 h and turned claret-red gradually. Finally, the light-
yellow polymer was precipitated in ethanol, washed with ethanol
ten times and dried in a vacuum oven at 70 ꢀC for 10 h.
Synthesis of bromomethylated poly(ether ether ketone)
The 4-methylphenyl hydroquinone (He-MQ) monomer was
synthesized according to procedure, used in our previous study.27
Poly(ether ether ketone) containing methylphenyl groups was
synthesized as follows. He-MQ (8.00 g, 40.00 mmol), DFBP (8.72 g,
40.00 mmol), anhydrous potassium carbonate (5.52 g, 40.00
mmol), dimethyl sulfoxide (DMSO, 40 mL) and toluene (18 mL)
were added to a 250 mL three-neck round bottom ask equipped
with a mechanical stirrer, a nitrogen inlet and a condenser.
Toluene was added and the reaction mixture was heated to 140 ꢀC,
reuxed for 4 h and then heated to 180 ꢀC to remove toluene. The
product was precipitated in cold distilled water when the mixture
became very viscous aer 5 h. Finally, the product was ltered off,
smashed, washed with boiling deionized water more than ten
times and dried under vacuum at 80 ꢀC for 48 h.
Synthesis of QPEEKs using TMA, MPY, MPRD, MIDZ, and
MMPH
The powder of BrPEEK-1.2 (3.00 g, 5.01 mmol) was dissolved in N-
methyl-2-pyrrolidone (NMP) in a 100 mL three-neck round
bottom ask equipped with
a mechanical stirrer and
a condenser. Then MPY (1.06 mL, 10.02 mmol) was added into
ꢀ
the ask and the reaction was kept at 60 C for 12 h. Later the
solution was cast onto a at glass plate and thermally treated
ꢀ
under vacuum at 60 C for 48 h to get rid of the residual NMP
solvent. Finally, the membrane was removed from the glass, cut
into pieces, washed with acetone ten times and dried in a vacuum
oven at 80 ꢀC for 48 h. The obtained product was named QPEEK-
MPY. Using a similar procedure, QPEEK-TMA, QPEEK-MPRD,
QPEEK-MIDZ and QPEEK-MMPH were also synthesized.
Preparation of AEMs
Membranes, based on these QPEEKs, were fabricated using the
solution casting method as shown in Scheme 3. For example,
QPEEK-MPY (1.0 g) was dissolved in 10 mL NMP, and then the
mixture was cast onto a at glass plate and thermally treated
under vacuum at 60 ꢀC for 48 h to remove the residual solvent.
The membrane was removed from the glass plate, immersed in
1 M NaOH solution for 24 h at room temperature and washed
with deionized water ten times to remove extra ions. Finally, the
hydroxide (OHꢁ) counterion form membranes were obtained.
All membranes were kept in a nitrogen lled ziplock bag before
testing to avoid access to CO2.
Scheme 2 Synthesis of PEEK, BrPEEK-X and QPEEK.
Scheme 3 Preparation of AEMs.
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 7975–7983 | 7977

View Article Online
RSC Advances
Paper
Characterization and measurement
Dwet ꢁ Ddry
Swelling ratioð100%Þ ¼
ꢂ 100%;
(4)
Ddry
Chemical structures of the model compounds, the monomer,
PEEK, BrPEEK-X and QPEEKs were analyzed using 1H NMR
spectroscopy (in CDCl3 and/or DMSO), performed on a Bruker
510 spectrometer (500 MHz) with tetramethylsilane (TMS) as
a reference. To determine the thermal properties of AEMs,
thermogravimetric analysis (TGA) was performed on a Perkꢁin1
Elmer Pyris 1 TGA instrument with a heating rate of 10 ꢀC min
in the range of 100 to 800 ꢀC under nitrogen atmosphere. A
SHIMADZU AG-1 1KN was used to explore the mechanical
properties of QPEEK membranes at a speed of 2 mm minꢁ1. At
least ten samples (15 mm ꢂ 4 mm) were tested, and the average
values of mechanical properties were calculated and recorded.
All measurements were taken on membranes in OHꢁ form.
Alkaline stability test of model compounds. The alkaline
stability of model compounds was determined using H NMR
aer aging the model compounds in 1 M sodium deuteroxide
deuterium oxide solution (NaOD/D2O) at 60 C periodically.
Calculation of bromination. The average bromine content of
each repeating unit of BrPEEK was determined accurately by 1H
NMR using the following formula:
The hydration number (l) represents the average number of
water molecules, absorbed by each QA group, and was calcu-
lated using the following equation:
WU ꢂ 10
l ¼
(5)
IEC ꢂ 18
Hydroxide conductivity. Hydroxide conductivity (s) was
measured with a Princeton Applied Research 2273 potentiostat/
galvanostat/frequency response analyzer using a four-probe AC
impedance method, frequency ranging from 1 Hz to 1 MHz. The
thickness of all membranes (1 cm ꢂ 4.5 cm) was measured
before measuring s. The test cell with the membrane was
immersed in deionized water to keep 100% relative humidity at
the tested temperature. The hydroxide conductivity (s) was
characterized by the following equation:
1
ꢀ
L
s ¼
;
(6)
RS
12AH
9
where L is the distance between the electrodes (cm), R is the
membrane resistance and S is the cross-sectional area of
membrane (cm2).
The slopes in Arrhenius plots of the logarithm of the ion
conductivity versus 1000/T can be used to estimate the apparent
activation energies (Ea) using the following equation:
DBr
¼
;
(1)
3AH þ 2AH
9
8
where DBr represents the number of bromomethyl functional
groups in each repeating unit. AH and AH are the integral areas of
9
8
the benzyl bromide groups and the methyl protons, respectively.
Ion exchange capacity (IEC). The ion exchange capacity (IEC)
of the membranes in the OHꢁ form was measured using the
acid–base titration method inꢀtriplicate. The membranes were
dried in a vacuum oven at 60 C for 48 h, accurately weighted,
and immersed in 40 mL of a 0.05 M hydrochloric acid solution.
Then the samples were equilibrated for at least 48 h under an
argon atmosphere. The resulting solutions were titrated with
a standardized 0.08 M sodium hydroxide solution. The IEC
(mmol gꢁ1) was described as
s ¼ A exp(ꢁEa/RT),
(7)
where s, A, Ea, R and T represent the hydroxide conductivity,
a frequency factor, the apparent activation energy, the gas
constant and the temperature, respectively. The apparent acti-
vation energy is estimated from the slopes of the ln s vs. 1000/T
curves.
Alkaline stability of AEMs. The membranes in hydroxide
(OHꢁ) counterion form were labeled as 0 day's sample. Then,
QPEEK membranes were kept in 1 M NaOH aqueous solution at
M2V2 ꢁ M1V1
IEC ¼
;
(2)
Wd
ꢀ
60 C for 1 day to 24 days, taken out at each interval, washed
where M1 (M) and V1 (mL) are the concentration and the volume
of hydrochloric acid before equilibrium was reached. M2 (M)
and V2 (mL) refer to the concentration and volume of the
standardized sodium hydroxide solution used for titration. Wd
is the mass of the dry samples.
Water uptake and swelling ratio. The weight and the thick-
nesses of dried QPEEK membranes, placed in a vacuum oven at
60 C for 48 h, were measured and then recorded as Wdry and
Ddry, respectively. All membranes were subsequently immersed
into deionized water at different temperatures for 24 h, and the
weight and the thickness of wet membranes were immediately
measured aer quickly wiping off the extra water on the surface
and recorded as Wwet and Dwet, respectively. The water uptake
(WU) and the swelling ratio (SR) were characterized using the
following equations, respectively:
with deionized water ten times to remove extra ions and tested
for hydroxide conductivity. The alkaline stability of AEMs was
estimated using the remaining hydroxide conductivity aer
immersing them in 1 M NaOH at 60 ꢀC for a desired time
period. In addition, the mechanical properties and the changes
in 1H NMR spectra were also used to evaluate the relative
stabilities of QPEEKs in OHꢁ form aer immersing them in 1 M
ꢀ
ꢀ
NaOH at 60 C for 24 d.
Results and discussion
Evaluating alkaline stability via degradation of model
compounds
1
Fig. 1 shows the H NMR spectra of model compounds QTMA,
QMPY, QMPRD, QMIDZ and QMMPH. For the QMPY model
compound, the chemical shis around 7.58–7.51 ppm (H1,2,3),
4.60 ppm (H4) and 2.90 ppm (H7) were assigned to the protons
Wwet ꢁ Wdry
Water uptakeð100%Þ ¼
ꢂ 100%;
(3)
Wdry
7978 | RSC Adv., 2019, 9, 7975–7983
This journal is © The Royal Society of Chemistry 2019

View Article Online
Paper
RSC Advances
Fig. 1 The 1H NMR spectra of model compounds QTMA, QMPY,
QMPRD, QMIDZ and QMMPH.
on the phenyl, benzyl and methyl groups, respectively. Besides,
the peaks around 3.50 ppm (H5) and 2.2 ppm (H6) were assigned
to the protons on methylene groups of MPY. Similar charac-
teristic peaks were observed in other 1H NMR spectra of model
compounds, which indicated that all the model compounds
were synthesized successfully.
To monitor the process of degradation and thoroughly
investigate the alkaline stability of hetero-cycloaliphatic QA
groups, 1H NMR spectra of model compounds QMPY, QMPRD,
ꢀ
QMIDZ and QMMPH alkali aging in 1 M NaOH/D2O at 60 C
were periodically measured for 18 days, as shown in Fig. 2.
QTMA was also investigated at the same conditions so that its
degradation pattern could be used as a reference for all model
compounds. The water peak at 4.78 ppm was chosen as the
internal reference on account of its integral area being nearly
constant in the process. Then we could obtain the remaining
fraction of the model compounds at a desired period of up to 18
days by calculating the relative integral area of proton reso-
nances on the phenyl groups, which is presented in Fig. 2(f).
During the test, QTMA, QMPY, QMPRD and QMMPH solutions
were clear and transparent, but QMIDZ solution turned into
a pale yellow emulsion and produced yellow oil droplets aer 4
days. In order to continue the alkaline stability test of QMIDZ,
the light-yellow supernatant was taken out each day. As shown
in Fig. 2(f), these model compounds showed a fast degradation
in the rst 6 days and then reached a plateau aer the 7th day.
As shown in Fig. 2(d) and (f), the relative integral area of proton
resonances of QMIDZ and QMMPH decreased signicantly as
the aging time increased.
Fig. 2 The 1H NMR spectra of QTMA (a), QMPY (b), QMPRD (c), QMIDZ
(d) and QMMPH (e) model compounds aged in 1 M NaOD/D2O at
60 ꢀC, and the remaining fraction of the model compounds in 1 M
NaOD/D2O (f).
caused by the OHꢁ nucleophilic attack. From these results, it
is clear that the alkaline stability of all model compounds
follows the order of QMPY > QMPRD > QTMA > QMIDZ >
QMMPH. This conclusion can be used to select the type of QAs
for the preparation of QPEEKs and analyze the alkaline
stability of the AEM.
Aer 18 days of aging, QMIDZ and QMMPH exhibited
degradation of 52.5% and 64.2%, respectively. However, the
degradation degree of QMPY and QMPRD with saturated
aliphatic-heterocyclic QA head-groups, calculated from their
1H NMR spectra, was just 24.1% and 28.9%, respectively; these
values were much lower than those of QMIDZ and QMMPH.
Importantly, the degradation degree of QMPY and QMPRD
was also much lower than that of conventional QTMA (47.3%),
indicating that pyrrolidium and piperidium head-groups were
found to have higher stability as compared to benzyl-
trimethylammonium head-group. These degradations, except
for that of QTMA, were mostly due to the ring-opening reaction
Structural characterization of QPEEKs
1
The HNMR spectra of PEEK and BrPEEK are shown in Fig. 3.
The chemical shis at around 4.50 ppm and 2.24 ppm were
assigned to the protons of the benzyl bromide groups (–CH2Br)
and the methyl protons (–CH3), respectively. The conversion
ratio of bromomethylation could be controlled by adjusting the
molar proportions of NBS and PEEK. The amounts of –CHBr2,
–CH2Br and –CH3 aer bromomethylation were quantitatively
determined by calculating their integral areas and the results
are listed in Table 1. The amount of –CHBr2 per repeat unit
increased from 0.04 to 0.48, while the amount of –CH3
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 7975–7983 | 7979

View Article Online
RSC Advances
Paper
Fig. 5 Water uptake and swelling ratio of QPEEKs.
QPEEK-MIDZ (Fig. 5). These QPEEKs were synthesized
successfully and all conversion ratios from BrPEEK were
between 80–85%, obtained by calculating the integration ratio
of quaternized ammonium benzyl groups (H9) to brominated
benzyl groups (H8).
Fig. 3 The 1H NMR spectra of PEEK and BrPEEK-1.2.
Table 1 The amounts of –CHBr2, –CH2Br and –CH3 per repeat unit
after bromomethylation
As presented in Table 2, the theoretical IEC, calculated using
the 1H NMR spectra, was in the range of 1.10 to 1.74 mmol gꢁ1
.
The experimental values obtained by the titration method
varied from 0.99 to 1.63 mmol gꢁ1. The experimental IEC of
these QPEEK membranes was line with the theoretical value,
implying the successful Menschutkin reaction. QPEEK-TMA,
QPEEK-MPY and QPEEK-MPRD exhibited similar IEC values
of 1.63 mmol gꢁ1, 1.49 mmol gꢁ1 and 1.52 mmol gꢁ1, respec-
tively. QPEEK-MIDZ and QPEEK-MMPH had relatively lower
IECs than those of the rst three membranes.
NBS/PEEK
–CHBr2
–CH2Br
–CH3
2.0
1.5
1.4
1.3
1.2
1.1
1.0
0.48
0.32
0.20
0.16
0.12
0.07
0.04
0.46
0.62
0.68
0.74
0.78
0.76
0.74
0
0
0
0
0.04
0.12
0.23
decreased from 0.23 to 0, as NBS/PEEK was adjusted from 1.0 to
2.0. Meanwhile, the amount of the target group –CH2Br per
repeat unit (DBr) increased slightly from 0.74 to 0.78 and then
decreased to 0.46 when NBS/PEEK increased from 1.0 to 2.0. In
the following procedure, the intermediate product BrPEEK with
the highest DBr of 0.78 was chosen to be the precursor for
tethering hetero-cycloaliphatic QA groups.
Water uptake and swelling ratio
Water uptake plays an important role in determining perfor-
mance of ion-exchange membrane materials. According to the
vehicle mechanism, ions can not penetrate membranes unless
they are hydrated by water. However, excessive water uptake
results in a high swelling ratio, which damages the dimensional
stability and the mechanical properties of AEMs. Therefore, it is
very important to control the water uptake and the swelling
ratio in a reasonable range. Water uptake and the swelling ratio
of QPEEKs in the OHꢁ form as a function of temperature,
ranging from 30 ꢀC to 80 ꢀC, are shown in Fig. 5. Both the water
uptake and the swelling ratio of these membranes increased
with increasing temperature. The QPEEKs with hetero-
cycloaliphatic QA head-groups showed much lower water
uptake than QPEEK-TMA (18.7% at 30 C and 36.6% at 80 C).
For example, QPEEK-MPY and QPEEK-MPRD exhibited similar
water uptake of 18.1% and 17.5% at 30 ꢀC as well as 28.9% and
27.2% at 80 ꢀC, respectively. Furthermore, QPEEK-MPY and
QPEEK-MPRD exhibited similar swelling ratios. Different from
water uptake, their swelling ratios were higher than that of
QPEEK-TMA (3.2% at 30 ꢀC and 11.4% at 80%) with similar
IECs. Besides, the corresponding l of QPEEK-TMA, QPEEK-MPY
and QPEEK-MPRD was in the range of 6.37 to 6.76, which was
higher than the values for QPEEK-MIDZ and QPEEK-MMPH. On
the whole, the changes in water uptake and l were in line with
the change in IEC.
The 1H NMR spectra of QPEEKs are shown in Fig. 4. For
QPEEK-MPY, the chemical shi around 4.70 ppm (H9) was
assigned the protons of –CH2–N groups; meanwhile, 2.80 ppm
(H11) and 2.10 ppm (H12) corresponded to the protons of –CH2–
groups in MPY ring. Similar peaks were found in the other
spectra of QPEEK-TMA, QPEEK-MPRD, QPEEK-MMPH and
ꢀ
ꢀ
Fig. 4 The 1H NMR spectra of QPEEK-TMA, QPEEK-MPY, QPEEK-
MPRD, QPEEK-MIDZ and QPEEK-MMPH.
7980 | RSC Adv., 2019, 9, 7975–7983
This journal is © The Royal Society of Chemistry 2019

View Article Online
Paper
RSC Advances
Table 2 Ion exchange capacity, water uptake at 80 ꢀC, and hydroxide conductivity at 80 ꢀC of QPEEKs
Sample
IECa (mmol gꢁ1
)
IECb (mmol gꢁ1
)
Water uptake (%)
s (mS cmꢁ1
)
l (OHꢁ)
QPEEK-TMA
QPEEK-MPY
QPEEK-MPRD
QPEEK-MIDZ
QPEEK-MMPH
1.74
1.68
1.68
1.14
1.10
1.63
1.49
1.52
1.03
0.99
36.6
28.9
27.2
11.6
19.8
40.5
39.9
38.8
37.4
33.1
6.37
6.76
6.38
5.11
5.93
a
Calculated from 1H NMR spectra. b Estimated by titration. l average number of H2O per quaternary ammonium group at 30 ꢀC. Water uptake and
s were measured at 80 ꢀC.
Thermal stability
Table 3 Mechanical Properties of QPEEKs
The TGA curves of QPEEKs are shown in Fig. 6. All these TGA
Tensile
Elongation
Young's
curves exhibited two thermal decomposition stages with the
temperature variation from 100 ꢀC to 800 ꢀC under nitrogen
atmosphere. The temperature of the rst weight loss, assigned
to the decomposition of QAs on side chains, was approximately
200 ꢀC and no further decomposition was observed. The
temperature of the second weight loss, assigned to the decom-
position of the main chains of QPEEK, was approximately
490 ꢀC. Moreover, QPEEK-TMA, QPEEK-MPY and QPEEK-MIDZ
exhibited similar Td5% of 279.1 ꢀC, 272.8 ꢀC and 273.6 ꢀC,
respectively. QPEEK-MPRD exhibited a relatively higher Td5% of
286.0 ꢀC and QPEEK-MMPH displayed the highest Td5% of
299.3 ꢀC. The Td5% of all these QPEEKs was higher than 270 ꢀC.
Besides, the residual weight percentage of all these QPEEKs at
800 ꢀC was in the range of 59.1ꢃ61.5%. Therefore, all
membranes possessed sufficient thermal stability for potential
application in construction of fuel cells.28
Sample
strength (MPa)
at break (%)
modulus (MPa)
QPEEK-TMA
QPEEK-MPY
QPEEK-MPRD
QPEEK-MIDZ
QPEEK-MMPH
43.8 ꢄ 2.2
46.2 ꢄ 1.0
46.0 ꢄ 0.7
68.8 ꢄ 4.8
60.0 ꢄ 6.7
16.5 ꢄ 2.5
18.0 ꢄ 2.1
10.2 ꢄ 1.2
27.9 ꢄ 3.7
70.0 ꢄ 2.9
590.2 ꢄ 127.7
771.6 ꢄ 63.6
820.9 ꢄ 62.0
911.1 ꢄ 131.8
801.5 ꢄ 75.5
MMPH > QPEEK-QMPY z QPEEK-MPRD > QPEEK-TMA. The
AEMs with hetero-cycloaliphatic QA head-groups had higher
tensile strength than QPEEK-TMA, indicating that they were strong
enough to be applied in fuel cell construction.
Hydroxide conductivity
Hydroxide conductivity is the core parameter used to determine
the performance of AEMs. The hydroxide conductivities of
QPEEKs as the function of temperature are shown in Fig. 7(a).
The hydroxide conductivities of all these membranes increased
with increasing temperature. The hydroxide conductivities of
QPEEK-MPY (39.9 mS cmꢁ1, 80 ꢀC) and QPEEK-MPRD (38.8
Mechanical properties
The mechanical properties of QPEEKs, measured in the dry state,
are listed in Table 3. QPEEKs exhibited tensile strength from
43.8 MPa to 68.8 MPa, Young's modulus from 590.2 MPa to
911.1 MPa and the elongation at break from 10.2% to 70.0%.
QPEEK-MIDZ had a maximum tensile strength of 68.8 MPa and
a maximum Young's modulus of 911.1 MPa, while QPEEK-MMPH
had a maximum elongation at break of 70.0%. According to the
tensile strength of QPEEKs presented in Table 2, the mechanical
strength of QPEEKs follows the order: QPEEK-MIDZ > QPEEK-
mS cmꢁ1, 80 C) wꢀere slightly lower than that of QPEEK-TMA
ꢀ
(40.5 mS cmꢁ1, 80 C), but higher than those of QPEEK-MIDZ
(37.4 mS cmꢁ1, 80 C) and QPEEK-MMPH (33.08 mS cmꢁ1, 80
ꢀ
ꢀC). Using the Arrhenius plots of hydroxide conductivity dis-
played in Fig. 7(b), the apparent activation energies of QPEEKs
were calculated from the slopes of the lines: 13.8 kJ molꢁ1 for
QPEEK-TMA, 14.8 kJ molꢁ1 for QPEEK-MPY, 15.2 kJ molꢁ1 for
QPEEK-MPRD, 16.2 kJ molꢁ1 for QPEEK-MIDZ and
19.0 kJ molꢁ1 for QPEEK-MMPH, respectively. The lower
apparent activation energies of QPEEK-MPY and QPEEK-MPRD
indicated that their hydroxide conductivity could be enhanced,
Fig. 7 Hydroxide conductivity of QPEEKs (a) and Arrhenius plots of
Fig. 6 TGA curves of QPEEKs.
hydroxide conductivity of QPEEKs (b).
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 7975–7983 | 7981

View Article Online
RSC Advances
Paper
Fig. 9 The remaining hydroxide conductivity of QPEEKs membranes
in the OHꢁ form after soaking them in 1 M NaOH.
remaining hydroxide conductivity at 80 ꢀC of the QPEEKs in the
OHꢁ form is shown Fig. 9. The hydroxide conductivities of all
these QPEEKs decreased as the immersion time increased. Aer
immersing the AEMs in 1 M NaOH at 60 ꢀC for 24 days, QPEEK-
Fig. 8 Possible OHꢁ promoted degradation pathways in QPEEKs
membranes.
presumably, due to better dissociation, thus, facilitating the
mobility of OHꢁ anions.29,30
MIDZ and QPEEK-MMPH showed
a signicant loss in
hydroxide conductivity (>40%), caused by lower alkaline
stability of benzylmethylimidazolium and benzylmethylmor-
pholinium at high temperature. QPEEK-TMA displayed
a slightly lower loss than QPEEK-MPY and QPEEK-MPRD in the
rst 15 days. But aer 24 days, QPEEK-MPY and QPEEK-MPRD
exhibited the hydroxide conductivity losses of 28.5% and 33.4%,
respectively, which were lower than that of QPEEK-TMA
(37.7%). For QPEEK-TMA, QPEEK-MPY and QPEEK-MPRD,
a new signal appeared around 5.85 ppm in 1H NMR spec-
trum, assigned to the proton of the OH group of phenol, and the
peak shapes showed some differences between 7 ppm and
8 ppm aer storage in 1 M NaOH at 60 ꢀC for 24 days (Fig S1, S2
and S3†), possibly induced by QPEEKs polymer backbone scis-
sion of those membranes. The integrated area ratio of Ar–CH2–
N+ (A1) and –CH2–Br or –CH2–OH (A0) is also recorded in Fig S1,
S2 and S3.† The changes of A1 : A0 before and aer the OHꢁ
treatment were due to the degradation of cationic groups,
possibly via pathways a, b and c, shown in Fig. 8. This would
cause the functional group loss and, then, conductivity loss. The
degradation of polymer backbone could cause cationic group
loss by dissolving low Mw and highly functionalized polymer
fragment, which resulted in conductivity loss. It should be
noted that the alkaline stability of QPEEKs membranes was not
completely consistent with that of model compounds. This
occured mainly because the alkaline stability of QPEEKs was
comprehensively affected by many factors: QPEEKs structure,
different experimental conditions, different water uptake, etc.
Evaluating the alkaline stability of AEMs using hydroxide
conductivity
Since AEMs must be stable in a strongly alkaline environment,
alkaline stability is one of the important parameters of AEMs for
their practical application in AEMFCs. As reported earlier
regarding methyl-containing head groups, there are three
possible pathways of cationic groups' degradation, presented in
Fig. 8. Pathway a is the SN2 substitution of the methyl group,
resulting in loss of methanol and yielding a polymer containing
a benzyl-linked secondary amine.31 Pathway b is a nucleophilic
attack or Hofmann elimination reaction that leads to hetero-
cyclic ring opening.32 Pathway c is a nucleophilic attack at the
benzylic F–CH2–N which is especially sensitive.33 Besides, there
2
is backbone degradation due to cleavage of aryl ether (Csp –O)
bonds, as shown in Fig. 8.25
As shown in Table 4, the mechanical properties of QPEEKs,
such as strength, elongation and modulus loss, declined aer
the OHꢁ treatment for 24 d. This occured mainly because that
aryl ether-containing polymers underwent polymer backbone
degradation via cleavage of C–O bonds under alkaline condi-
tions,34,35 as shown in Fig. 8.
The change in the hydroxide conductivity as a function of
time was used to evaluate the relative stability of QPEEKs in the
OHꢁ form by immersing them in 1 M NaOH at 60 ꢀC. The
Table 4 Mechanical properties of QPEEKs membranes after treat- The anion exchange membrane exhibited a relatively high local
ment with OHꢁ for 24 d
pH in alkaline solution and a low water uptake, which led to
more degradation. It is why the alkaline stability of QPEEK-
MMPH and QPEEK-MIDZ was not consistent with that of
QMIDZ and QMMPH.
Tensile
strength (MPa)
Elongation
at break (%)
Young's
modulus (MPa)
Sample
QPEEK-TMA
QPEEK-MPY
QPEEK-MPRD
QPEEK-MIDZ
QPEEK-MMPH
34.5 ꢄ 5.2
30.7 ꢄ 8.3
15.5 ꢄ 2.7
61.9 ꢄ 2.4
53.3 ꢄ 2.4
5.3 ꢄ 0.5
5.7 ꢄ 0.8
1.8 ꢄ 0.3
17.0 ꢄ 5.4
23.6 ꢄ 2.9
472.3 ꢄ 77.0
421.7 ꢄ 38.1
291.2 ꢄ 42.0
492.6 ꢄ 58.8
528.3 ꢄ 31.3
Conclusions
Several model compounds with aliphatic-heterocyclic QA head-
groups were synthesized and their alkaline stability was
7982 | RSC Adv., 2019, 9, 7975–7983
This journal is © The Royal Society of Chemistry 2019

View Article Online
Paper
RSC Advances
1
characterized by H NMR. Our results indicated that the alka- 12 J. Pan, C. Chen, L. Zhuang and J. Lu, Acc. Chem. Res., 2012,
line stability of these model compounds followed the order of 45, 473.
QMPY > QMPRD > QTMA > QMIDZ > QMMPH. Then, a series of 13 F. Bu, C. Zhao, B. Wang, N. Zhang, H. Lu, Z. Cai, Y. Zhang
AEMs with aliphatic-heterocyclic QA head-groups on the side and H. Na, RSC Adv., 2015, 5, 57067–57075.
chains were prepared and compared with QPEEK-TMA. These 14 H. S. Dang and P. Jannasch, J. Mater. Chem. A, 2016, 4,
QPEEKs exhibited sufficient thermal, dimensional and 11924–11938.
mechanical stability. QPEEK-MPY, QPEEK-MPRD and QPEEK- 15 M. G. Marino and K. D. Kreuer, Chemsuschem, 2015, 8, 513–
TMA exhibited similar OHꢁ conductivities of 39.9 mS cmꢁ1
523.
38.8 mS cmꢁ1 and 40.5 mS cmꢁ1 at 80 ꢀC and 100% relative 16 K. M. Meek, J. R. Nykaza and Y. A. Elabd, Macromolecules,
humidity. Furthermore, aer a 24 days' test, QPEEK-MPY and 2016, 49, 3382–3394.
QPEEK-MPRD possessed better alkaline stability aer their 17 F. Gu, H. Dong, Y. Li, Z. Sun and Y. Feng, Macromolecules,
immersion in 1 M NaOH solution at 60 ꢀC. They exhibited
2014, 47, 6740–6747.
conductivity loss of 28.5% and 33.4%, respectively, which was 18 C. Morandi, R. Peach, H. Krieg and J. Kerres, J. Mater. Chem.
lower than that of QPEEK-TMA (37.7%). All parameters indicate A, 2014, 3, 1110–1120.
that QPEEK-MPY and QPEEK-MPRD can be potential compo- 19 J. Ponce-gonzalez, D. K. Whelligan, L. Wang,
,
´
nents of AEMFCs, especially QPEEK-MPY, which exhibited the
best overall performance.
R. Bancesoualhi, Y. Wang, Y. Peng, H. Peng,
D. C. Apperley, H. N. Sarode and T. P. Pandey, Energy
Environ. Sci., 2016, 9, 3724–3735.
20 J. Wang, S. Li and S. Zhang, Macromolecules, 2010, 43, 3890–
3896.
Conflicts of interest
There are no conicts to declare.
21 J. Wang, C. Zhao, H. Lin, G. Zhang, Y. Zhang, J. Ni, W. Ma
and H. Na, J. Power Sources, 2011, 196, 5432–5437.
22 J. Pang, S. Feng, Y. Yu, H. Zhang and Z. Jiang, Polym. Chem.,
2014, 5, 1477–1486.
23 B. Qiu, B. Lin, L. Qiu and F. Yan, J. Mater. Chem., 2011, 22,
1040–1045.
Acknowledgements
We acknowledge the nancial support from the Natural Science
Foundation of China (No. 21474036).
24 B. Qiu, B. Lin, Z. Si, L. Qiu, F. Chu, J. Zhao and Y. Feng, J.
Power Sources, 2012, 217, 329–335.
25 L. Liu, J. Ahleld, A. Tricker, D. Chu and P. Kohl, J. Mater.
Chem. A, 2016, 4, 16233–16244.
26 B. Wang, W. Sun, F. Bu, X. Li, N. Hui and C. Zhao, Int. J.
Hydrogen Energy, 2016, 41, 3102–3112.
27 X. Shuai, Z. Gang, Z. Yang, C. Zhao, W. Ma, H. Sun, Z. Na,
L. Zhang, J. Hao and N. Hui, J. Power Sources, 2012, 209,
228–235.
28 J. Wang, X. Wang, D. Zu, H. Yang, L. Yan, S. Yang, H. Wei
and Y. Ding, J. Membr. Sci., 2017, 545, 116–125.
29 M. Irfan, E. Bakangura, N. U. Afsar, M. M. Hossain, J. Ran
and T. Xu, J. Power Sources, 2017, 355, 171–180.
30 A. Le Mong and D. Kim, Electrochim. Acta, 2018, 290, 544–
555.
31 L. Hai, K. Kim and B. S. Pivovar, J. Phys. Chem. C, 2012, 116,
9419–9426.
32 S. A. Nunez, C. Capparelli and M. A. Hickner, Chem. Mater.,
2016, 28, 2589–2598.
Notes and references
1 A. B. Rao and E. S. Rubin, Environ. Sci. Technol., 2002, 36,
4467.
2 N. Li and M. D. Guiver, Macromolecules, 2014, 47, 2175–2198.
3 J. R. Varcoe, P. Atanassov, D. R. Dekel, A. M. Herring,
M. A. Hickner, P. A. Kohl, A. R. Kucernak, W. E. Mustain,
K. Nijmeijer and K. Scott, Energy Environ. Sci., 2014, 7,
3135–3191.
4 G. J. K. Acres, J. Power Sources, 2001, 100, 60–66.
5 Y. J. Wang, J. Qiao, R. Baker and J. Zhang, Chem. Soc. Rev.,
2013, 42, 5768–5787.
6 C. Chen, J. Pan, J. Han, Y. Wang, L. Zhu, M. A. Hickner and
L. Zhuang, J. Mater. Chem. A, 2016, 4, 4071–4081.
7 C. Chen, Y. L. S. Tse, G. E. Lindberg, C. Knight and
G. A. Voth, J. Am. Chem. Soc., 2016, 138, 991.
8 S. Chempath, B. R. Einsla, L. R. Pratt, C. S. Macomber and
J. M. Boncella, J. Phys. Chem. C, 2017, 112, 3179–3182.
9 J. R. Varcoe, R. C. T. Slade and T. N. Danks, J. Mater. Chem.,
2003, 13, 712–721.
33 A. D. Mohanty and C. Bae, J. Mater. Chem. A, 2014, 2, 17314–
17320.
34 A. Christopher G and R. Vijay, Proc. Natl. Acad. Sci. U. S. A.,
2013, 110, 2490–2495.
10 C. E. Diesendruck and D. R. Dekel, Curr. Opin. Electrochem.,
2018, 9, 173–178.
11 C. H. Park, S. Y. Lee, D. S. Hwang, D. W. Shin, D. H. Cho,
K. H. Lee, T. W. Kim, T. W. Kim, M. Lee and D. S. Kim,
Nature, 2016, 532, 480.
35 S. Zhe, P. Ji, J. Guo and Y. Feng, Adv. Sci., 2018, 5, 1800065.
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 7975–7983 | 7983