Theoretical studies on the degradation of hydrocarbon copolymer ionomers used in fuel cells

Journal of Membrane Science 487 (2015) 229–239

Contents lists available at ScienceDirect

Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci

Theoretical studies on the degradation of hydrocarbon copolymer ionomers used in fuel cells Yuan-yuan Zhao a,n, Eiji Tsuchida b, Yoong-Kee Choe b, Jian Wang c, Tamio Ikeshoji a,b, Akihiro Ohira a,d a Fuel Cell Cutting-Edge Research Center (FC-Cubic), Technology Research Association, AIST Tokyo Waterfront Main Building, 2-3-26 Aomi, Koto-ku, Tokyo 135-0064, Japan b Nanomaterials Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba Central 2, Umezono 1-1-1, Tsukuba 305-8568, Japan c Department of Civil and Environmental Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong, China d Research Institute of Electrochemical Energy, National Institute of Advanced Industrial Science and Technology (AIST), 1-8-31, Midorigaoka, Ikeda, Osaka 563-8577, Japan

art ic l e i nf o

a b s t r a c t

Article history: Received 14 February 2015 Received in revised form 3 April 2015 Accepted 5 April 2015 Available online 13 April 2015

Details on the degradation mechanism of hydrocarbon ionomers for polymer electrolyte membrane fuel cells (PEFCs) have been investigated using density functional theory (DFT). Two model compounds of a hydrocarbon ionomer were selected to investigate the nature of the degradation, where the compounds selected represented the general features of available hydrocarbon polymer electrolyte membranes (PEMs). The results show that two degradation reactions are energetically favorable, indicating two possible weak sites in the hydrocarbon PEMs susceptible to OH or H radical attack. One site has an aryl sulfonated bond and a sulfonate group, and the other site has an aryl ether bond. The alkane chains in the PEMs were calculated to be relatively stable against radical attack. In addition, we found that in contrast to the degradation of perfluorinated PEMs, hydrocarbon PEMs are relatively robust against H radical attack. & 2015 Elsevier B.V. All rights reserved.

Keywords: Hydrocarbon PEMs Chemical stability Degradation reaction mechanism Fuel cell DFT calculations

1. Introduction PEMs play an important role in separating electrodes, by transporting protons in the membrane and in forming membrane electrode assemblies as an ionomer in PEFCs. Good stability and high proton conductivity at low water contents are desired properties for PEMs. According to the generic industry standard, a lifetime durability of 40,000 h and an uninterrupted service life of 8000 h at 480% power are desirable for PEFCs in stationary applications. These figures can be compared with the 20,000 and 6000 operating hours required for buses and cars, respectively [1,2]. Chemical degradation caused by oxidative species produced as a result of chemical reactions in cycling operating fuel cells poses a serious obstacle to the stability of PEMs [3]. Thus, connected with the above stability issues, the details on oxidative species and their formation mechanisms have been extensively studied in PEFCs. For example, hydrogen peroxide, OH radicals, H radicals, and OOH radicals are believed to be prevalent in solution [4–6]. In particular, OH radicals and H radicals are believed to play a major role in membrane

n

Corresponding author. Tel.: þ 81 29 861 2314. E-mail address: [email protected] (Y.-y. Zhao).

http://dx.doi.org/10.1016/j.memsci.2015.04.005 0376-7388/& 2015 Elsevier B.V. All rights reserved.

degradation reactions. Generally, two possible generation mechanisms are discussed in the literature. One is attributed to the crossover of O2 from the cathode to the anode through the membrane and incomplete reaction with H2 on the surface of the anode catalyst, and the other is a result of the reaction of H2 O2 (which is an intermediate of the two-electron reduction of O2 at the cathode) with trace metal. The latter is considered as a main source of OH radical formation in the fuel cell with a hydrocarbon-based membrane [3,7–14]. An understanding of the mechanism of degradation is obviously the first step required for improving the chemical stability of PEMs. Perfluorinated membranes, such as Nafion, are currently widely used in PEFCs, even though they also suffer from chemical degradation caused by oxidative species, and many studies have been reported on their degradation mechanisms [15–19]. Yu et al. proposed that OH radicals can attack the S–C bond in Nafion to form H2SO4 and carbon radicals, with an energy barrier height of 0.96 eV, while H radicals can attack the C–F bond to form HF and carbon radicals based on quantum mechanical calculations [17]. Yu et al. also reported on the degradation mechanisms of Nafion with different end groups (–COOH, –CF ¼CF2, and –CF2H) employed in different fuel cells operating conditions [18]. A recent study by Ghassemzadeh et al. indicated that OH radicals could attack the αO–C bond in the side chain of Nafion, and H radicals could possibly

230

Y.-y. Zhao et al. / Journal of Membrane Science 487 (2015) 229–239

attack the tertiary carbon C–F bond on both the main and side chains of Nafion [19]. On the other hand, there have been persistent efforts to find viable alternatives for Nafion. In particular, the aromatic polymer family has attracted much attention because of its promising properties, such as low production cost and environmental friendliness. Therefore, various types of hydrocarbon PEMs employing aromatic polymers have been reported so far [20–24], and the durability of these hydrocarbon PEMs has also been investigated extensively [25,7,13,26–28,12]. Hübner et al. employed electron paramagnetic resonance (EPR) techniques to identify degraded membrane fractions caused by OH radicals using eight model sulfonated aromatic compounds [25]. Panchenko proposed various degradation mechanisms for model sulfonated poly(ether ether ketone) (SPEEK) compounds and polyethersulfone (PSU) based on DFT calculations, where the mechanism indicated that the phenyl ring was attacked by OH radicals and oxygen molecules, leading to the loss of the sulfonic acid group [12]. Perrot et al. investigated the degradation of a model compound of sulfonated aryl ether ketones (SPAEK) using nuclear magnetic resonance (NMR), infrared (IR), and mass spectroscopy and demonstrated that the oxidation of the phenolic groups resulted in chain scission after OH radical attack on the aromatic ring [28]. Despite these efforts, a clear and comprehensive understanding of the degradation mechanisms of hydrocarbon PEMs is still lacking because of the complicated nature of the degradation pathway. Recently, Rikukawa's group reported on novel hydrophilic– hydrophobic block copolymer ionomers [29] based on polyphenylenes using a new method, catalyst transfer polycondensation, which was proposed by Yokoyama et al. and Sheina et al. [30,31]. This copolymer exhibits high proton conductivity at low hydration by precisely controlling the diblock lengths and composition of the polymer. As shown in Fig. 1(a), this hydrocarbon copolymer consists of a main chain of polyphenyl rings and two different side chains with hydrophobic and hydrophilic groups and several other functional groups (phenyl rings, alkane chains, ether groups, and sulfonate groups). This type of main chain is often used in hydrocarbon PEMs. In addition, its side chain is similar to that of other hydrocarbon PEMs, such as SPEEK and sulfonated polyether sulfone (SPES) compounds. In this work, we took this new copolymer as an example of hydrocarbon PEMs and investigated its possible degradation mechanisms by applying DFT calculations. The remainder of this paper is organized as follows. The computational details are described in Section 2. Various possible degradation mechanisms initiated by OH or H radicals on the side and main chains are presented in Section 3, and a summary is given in Section 4.

a key sulfonate group for proton donation. It also includes common functional groups used in hydrocarbon PEMs, such as a phenyl ring, an alkane chain, and an ether group. Fig. 1(c) shows the molecular structure of M2, modeling a main chain consisting of phenyl rings. In M2, the methoxy groups were used as a substituent site for the side chain of the copolymer. All the geometries of the stationary points on the potential energy surfaces were optimized at the B3LYP/6-31þG (d) and B3LYP/6-311þ þ G(2d, 2p) levels of theory [32,33]. Since PEMs are hydrated by water in PEFCs, the effect of the solvent was taken into account using the polarizable continuum model (PCM) [34–36]. The resulting structures obtained at the gas phase B3LYP/6311þ þG(2d, 2p) level of theory were then optimized further, including any solvent effects. The vibrational frequencies were calculated to confirm that the transition states had only one imaginary frequency and that the local minima had no imaginary frequency. The reaction enthalpies were obtained in the gas phase and in water at room temperature. The zero point energies (ZPEs) of all the optimized structures were also calculated at the corresponding levels of theory mentioned above and were included in the calculation of the relative energies. All the calculations in this work were carried out using the Gaussian09 software package [37].

3. Results and discussion In this section, we will discuss the degradation reactions of the membranes in terms of the side- and main-chain degradation. For the model compound with a side chain, M1, shown in Fig. 1(b), all of its noted atom sites (C1–C9) could be attacked by oxidative species. For the model compound with a main chain, M2, shown in Fig. 1(c), we note that atoms C1, C2, and C3 were considered to be possible sites that could react with oxidative species. For both models, possible degradation mechanisms occurring at these sites were proposed and are discussed in this section based on our DFT calculations. Concerning the oxidative species involved in the above degradation reactions, we mainly focused on OH and H radicals, but sometimes included O2, based on data from previous studies [17,18,12]. Considering that only trace amounts of oxidative species are present in PEFCs, one OH radical and one H radical were introduced in our computations. These radicals mostly work as an initiator of reactions, and introduce radicals into the membrane. Unless otherwise noted, all the energies mentioned in the following text were obtained at the B3LYP (PCM)/6311 þ þG(2d, 2p) level of theory, including the ZPE corrections, as described in Section 2. The terms R and R 0 are used to represent the inactive fragments in the relevant figures. 3.1. Side chain degradation by OH radicals

2. Computational details Two model compounds of the novel hydrocarbon copolymer ionomer, denoted as M1 and M2, were selected to study the degradation reactions. Fig. 1(b) shows the structure of M1, which is a model compound with a hydrophilic domain on the side chain and

3.1.1. OH radical attack on the sulfonate group It can be seen in Fig. 1(b) that C1 is a carbon atom in a phenyl ring bonded directly to the sulfonate group. The stability of the C– S bond is very important for the performance of PEMs because OH radical attack on this bond can result in the loss of the sulfonate group, causing the membrane to cease being a proton conductor.

Fig. 1. The chemical structures of (a) the copolymer and its model representations, (b) M1 (side chain), and (c) M2 (main chain).

Y.-y. Zhao et al. / Journal of Membrane Science 487 (2015) 229–239

Both the C1 and S atoms in the C–S bond can be attacked by OH radicals. Two forms of M1, a proton-dissociated (M1n) and a proton-undissociated (M1) form were considered in the degradation reaction computations. Four degradation mechanisms involving OH radical attack on the C–S bond, denoted as Reactions (a)–(d), were found. Their details are shown in Fig. 2, along with the values of the barrier heights. Reaction (a) is a OH radical attack on atom C1 in the proton-dissociated form. This mechanism suggests that the attack on C1 leads to the breakage of the C–S bond by a one-step reaction producing a phenol compound and a sulfuric acid radical anion. This reaction occurs through two intermediates (Int1 and Int2) and one transition state (TS). Selected structural parameters are shown in Fig. 2. The optimized distance between the O and C atoms in the TS was calculated to be 2.05 Å, and the H atom in the OH radical was hydrogen bonded to the O atom of the sulfonate group. The barrier height, Ea , was calculated to be 0.23 kJ mol  1, indicating a barrierless reaction. Thermodynamically, Reaction (a) is exothermic by  152.46 kJ mol  1 in terms of the enthalpy. All the obtained energies show that the degradation at atom C1 occurs easily from OH radical attack on M1n. Reaction (b) describes the degradation reaction caused by a OH radical attack on atom C1 in M1. This mechanism is similar to that observed for M1n, i.e., the C–S bond breaking because of the OH radical. The O–C bond distance in the TS was calculated to be 2.04 Å, a value that is almost identical to the corresponding value for Reaction (a). The barrier height for Reaction (b) was calculated to be 6.79 kJ mol  1, which is only slightly higher than that of Reaction (a), and the calculated reaction enthalpy was 123.36 kJ mol  1, which indicates that the reaction was also exothermic. Therefore, atom C1 in the proton-undissociated form is also unstable, just as it is in the dissociated form versus OH radical attack. In addition, the degradation reactions arising from OH radical attack on the S atom in M1n and M1 were defined as Reactions (c) and (d), respectively. These mechanisms show that OH radicals attacking the S atom also give rise to the C–S bond breaking. However, the calculated barrier heights indicate that these two reactions have to overcome barrier energies greater than 140 kJ mol  1 to break the C–S bond. In the case of Nafion degradation, high barrier heights of 92.63 kJ mol  1 and 78.15 kJ mol  1 for the proton-dissociated and -undissociated forms have been

R=

O

CH2

CH2

CH2

R

+ OH

R

_

SO3

(Int2)

3.1.2. OH radicals and O2 attack on the phenyl group The second type of degradation reaction we studied was an oxidative species attacking the C2, C3, and C4 atoms of the phenyl group in the side chain. The phenyl group is an essential component of the aromatic membrane family, and thus, an understanding of the stability of the phenyl ring and its possible degradation mechanisms in PEFCs is of importance. Several researchers have investigated the mechanistic details of benzene oxidation by OH radicals [38,39]. Recently, Panchenko reported on the possible degradation mechanisms of the phenyl group in SPEEK and PSU based on DFT calculations, but no activation energy was reported [12]. Based on this research, we further investigated the degradation mechanisms of the phenyl ring and calculated the reaction barrier heights. All the mechanisms of benzene or phenyl ring degradation proposed previously [25,27,12,38,39] have suggested that the presence of O2 is necessary for the degradation of the phenyl ring to occur, and O2 is abundant in fuel cell environments, especially at the cathode side. Therefore, two oxidative species, OH radicals and O2 molecules, were considered in our degradation study of the phenyl rings. OH radicals are usually the initiator of a radical reaction, and addition of O2 promotes the degradation of the system. In our study on M1, we calculated the reactions arising from OH radicals attacking the ortho position (C2), the meta position (C3), and the para position (C4) with respect to the sulfonate group in the initial step.

O

2.05

H 1.70 _ (Int1) SO3

OH 1.77 R

suggested for OH radicals attacking the S atom [17], respectively. Our results are consistent with these results, even though the type of membrane studied is different. The calculated reaction enthalpies for Reactions (c) and (d) were  25.65 kJ mol  1 and 2.05 kJ mol  1, which are only slightly exothermic and endothermic, respectively. In these reactions, a radical is introduced into the phenyl ring, which indicates that the aromatic character of the phenyl ring is lost as a result of the introduction of the radical. This fact explains why the reactions exhibit higher barrier heights as well as energetically unstable final products. These energetics demonstrate that the degradation reaction resulting from an attack by OH radicals on the S atom is much more difficult to achieve compared with the result of the C atom in the C–S bond.

CH2 4.49

_ SO3

R

OH + SO3

R

_

SO3H + OH OH R

4.30 SO3H

_

R

SO3

R

SO3H + OH

+ OH

H 1.89 SO3H (Int1)

R

(Int2)

R

O

H 2.10 _ (TS)

SO3

Ea= -0.23 kJ mol-1

SO2 H O 2.04

O R

231

R

OH + SO3H

R

O H 1.96 _ (TS) SO3

R

R

O H 1.84 SO3H (TS)

R

SO3H (TS)

Ea= 6.79 kJ mol-1

SO2

+ HSO4

_

+ H2SO4

Ea= 143.47 kJ mol-1

Ea= 141.04 kJ mol-1

Fig. 2. The proposed degradation mechanisms of OH radical attack on C1 and S in the sulfonate group of M1 and M1n.

232

Y.-y. Zhao et al. / Journal of Membrane Science 487 (2015) 229–239

are shown in Fig. 3. The energy profiles shown in Fig. 3 show that the highest energy structure is TSB3, which connects the OH–M1–OO adduct radical (Int2) and the bridged O–O aromatic compound (Int3). The barrier height of this reaction was calculated to be 65.33 kJ mol  1, being slightly higher than that of Path (e_iiA). The calculated reaction enthalpy of Path (e_iiB) was  86.42 kJ mol  1, indicating an exothermic reaction. It was observed that both paths in Reaction (e) have higher barrier heights than those in Reactions (a) and (b), which indicates that atom C2 is more stable than atom C1 against OH radical attack. A second possible reaction on the phenyl group, denoted as Reaction (f), is where OH radicals and O2 attack atom C3, which is in the meta position with respect to the sulfonate group. The calculated energy profiles shown in Fig. 4 are similar to those of Reaction e) because of their similar reaction sites. Therefore, we will not go into the details of this reaction. Here, we only point out that (1) an initial step for forming the hydroxyl–cyclohexadienyl radical (Int1) has a negative energy barrier, 1.93 kJ mol  1, which is lower than the value obtained in Reaction (e). This is in agreement with a meta-directing effect of –SO3H; (2) O2 still prefers the ortho position located relative to the OH group and on the same side of the ring plane; and (3) the produced intermediate OH–M1–OO adduct radical (Int2) first forms an O–O bridge, followed by O–O and C–C bond-breaking reactions. These reactions finally destroy the aromatic group and cause the ring to be opened, while the C–S bond is retained. The transfer of a H atom from the OH group to its neighboring O atom in the final step leads to a stable compound as the final product with a relative energy of  265.68 kJ mol  1. The energy profile shows that the reaction has a barrier height of 25.41 kJ mol  1, which is lower than that of

Degradation starting at C2, denoted by Reaction (e), is shown in Fig. 3. The initial step is OH radical attacking C2 to form a hydroxyl–cyclohexadienyl radical (Int1). The barrier height for this step was calculated to be 5.88 kJ mol  1. Afterwards, the produced radical, Int1, can initiate two other reactions: (e_iiA) and (e_iiB). In Path (e_iiA), a H radical and a phenol compound are generated. The barrier height of this reaction was calculated to be 54.67 kJ mol  1. The reaction is endothermic by 23.43 kJ mol  1. In the case of Path (e_iiB), Int1 first reacts with O2 followed by a series of reactions leading to ring opening and the final loss of the sulfonate group. The first step is an O2 addition reaction. All the carbon atoms in the phenyl ring, except for C2, can react with O2. Moreover, O2 addition can occur in a cis or trans position relative to the OH group. We optimized all possible structures satisfying these conditions and found that the structure of O2 addition at the ortho position has the lowest energy, and the structure of O2 addition at the para position is the second most stable structure. Furthermore, by comparing the energies of the cis and trans conformations, we found that the O2 molecule prefers to remain on the same side of the ring plane as the OH group. Therefore, we took the most stable OH–M1–OO adduct as the intermediate (Int2) for the ensuing steps of the reaction. Subsequently, the –O2 group can attack the meta C or ortho C site to form a bridged O–O aromatic compound. The optimized structures show that an O–O radical bonded to the meta C position (C1) is energetically the most favorable state. The resulting bridged O–O aromatic compound (Int3) further undergoes O–O and C–C bond-breaking reactions, resulting in a ring-opened compound (Int5). In the last stage of the reaction, the C–S bond is broken and the sulfonate group becomes detached from the structure, indicating the degradation of the polymer. All the transition states in these reactions and selected structural parameters

R=

O

CH2

CH2

CH2

CH2

H

OH

H

OH

2.03

e_i) R

SO3H + OH H

e_iiA) R

SO3H (TS1)

R H 1.76

OH

OH

OH

SO3H (TSA2)

R

SO3H (Int1)

SO3H (Int1) Ea= 5.88 kJ mol-1

R

SO3H + H Ea= 54.67 kJ mol-1

R

e_iiB) Erela (kJ.mol-1) O2 2.12 H H

O

R SO3H

R

TSB2 46.57

0.0

H

H

OH

Int1+O2

OH

SO3H O 2.60

TSB3 65.33

R

H R

OH

1.80 O O

SO3H

TSB4 28.02 H

Int2 -2.44

0.00

H

H

Int3 -3.82

O2 H OH H SO3H

R

H

O

O

2.01 O O

R

H

OH H

OH SO3H

Int4 -75.81 H R

H OO

H

SO3H

TSB5 -56.24

O O

R

OH 2.39

SO3H

TSB6 -73.95 -89.01

OH

Int5 -108.10

SO3H

H R

HC O HC OH R O

H OH O O

+ SO3H

SO3H

Fig. 3. The proposed degradation mechanism of OH radical and O2 attack on C2 in the phenyl ring of M1.

Y.-y. Zhao et al. / Journal of Membrane Science 487 (2015) 229–239

R=

O

CH2

CH2

CH2

CH2

O H 2.05

H

f_i) R

SO3H + OH

H SO3H (TS1)

R

H

O2

OH

2.29

HO H

0.0

O

SO3H

SO3H (Int1)

OH H O O2.02

SO3H

TS4

TS3

TS2 Int1+O2

R 2.05 O

R

SO3H

R

1.83

H

1.99

OH

R H

Erela (kJ.mol-1)

f_ii)

233

25.41

20.87

16.38

0.00 Int2 -21.79 HO H H

Int3 O2

-46.55

SO3H

R

HO H

-265.68

H R

O

O

H O

SO3H R

HO

g)

SO3H + OH

R

2.07 SO3H (TS1)

R

HO R

HO R

2.13

SO3H (TS2)

R + HO

SO3H

O

HO

H SO3H

SO3H (Int1)

Ea= 19.36 kJ mol-1

Fig. 4. The proposed degradation mechanisms of OH radicals attack on C3 and C4 in the phenyl ring of M1.

Reaction (e). The calculated reaction enthalpy of Reaction (f) was 267.81 kJ mol  1, indicating an exothermic reaction. Therefore, it is easier for Reaction (f) than Reaction (e) to occur when OH radicals attack the phenyl ring. As shown in Fig. 4, the reaction on atom C4, denoted as Reaction (g), is different from those on atoms C2 and C3 discussed above. In this reaction, the OH radicals' attack on atom C4 directly results in the breaking of the C4–C5 bond and the scission of the chain without a contribution from O2, producing p-hydroxybenzenesulfonic acid and a butyl radical. The barrier height of Reaction (g) was calculated to be 19.36 kJ mol  1, which is obviously lower than that for the reaction on atom C2, and slightly lower than that for the reaction on atom C3. A negative reaction enthalpy of  55.17 kJ mol  1 indicates that the reaction is exothermic. The obtained energetics indicate that atom C4 is weaker than atoms C2 and C3. It was also found that in the phenyl ring, the carbon atoms directly bonded to other groups, such as atoms C1 and C4, are weaker than other carbon atoms.

3.1.3. OH radical attack on the butyl group All the carbon atoms from atoms C5 to C8 belong to the alkane chain, which is a universal functional group used in hydrocarbon PEMs, and there has been speculation on the degradation mechanisms of alkane chains, either from experimental observations or theoretical calculations. Here, we report on the degradation mechanisms of alkane chains based on DFT calculations. We have investigated reactions on all four C atoms, denoted as Reactions (h), (i), (j), and (k).

All four proposed reaction mechanisms are shown in Fig. 5. In all these reactions, the first step is an abstraction reaction of a H atom by a OH radical producing H2O and a relatively stable, secondary (21) alkyl radical. Subsequently, these alkyl radicals undergo C–C or C–O bond-breaking reactions, resulting in the scission of the alkane chain. Obviously, the products of the four reactions are different because of the different carbon radicals formed in the first step. In Reaction (h), the OH radical attacks atom C5, resulting in the breaking of the C6–C7 bond, while in the other three reactions, OH radical attack can result in two bond-breaking reactions. For example, in Reaction (i), either the C4–C5 or C7–C8 bond is broken when OH radicals attack atom C6, while the C5–C6 or C8–O bond is broken when OH radicals attack atom C7 in Reaction (j). Similarly, the C6–C7 or C9–O bond can be broken in Reaction (k) when OH radicals attack atom C8. The calculated barrier heights show that the first step for all four reactions is barrierless, which indicates that the abstraction of a H atom by a OH radical occurs with ease. However, the second reactions for all four carbons have higher barrier heights, greater than 70 kJ mol  1 for breaking the C–C bond, 44.11 kJ mol  1 for breaking the C8–O bond, and 116.01 kJ mol  1 for breaking the C9–O bond. When comparing the four C atoms in the butyl group, Reaction (j) at atom C7 has a lower energy barrier than the other carbon atoms. The obtained reaction enthalpies of Reactions (h)–(k) show that they are endothermic, except for the C8–O bond-breaking reaction, which has a small negative value of  34.51 kJ mol  1. Therefore, although many types of degradation reaction are possible for the alkane chain, such reactions occur with

234

Y.-y. Zhao et al. / Journal of Membrane Science 487 (2015) 229–239

R=

R'= H2

h_i) R O (CH2)3 C H

h_ii) R O (CH2)3 C

SO3H H R O (CH2)3 C

R' + OH

R O (CH2)2

R' (Int1)

R O (CH2)2 + H2C

i_ii) R O (CH2)2

R' (TS)

H2 C R' (Int1) + H2O

R O (CH2)2 C CH2 + R' H H2 2.46 H R O (CH2)2 C C R' (TSA) -1 2.27 H H2 R O C CH2 C C R'(TSB) H2 H2 H R O C C

H2 H2

j_i) R O C C (CH2)2 R' + OH

H2 H

H CH2 C

H R O (CH2)2 C

A H2 C R' (Int1) B

j_ii) R O C C (CH2)2 R' (Int1)

R' (Int1) + H2O

Ea= 139.32 kJ mol-1

CH R'

i_i) R O (CH2)2 CH2 CH2 R' + OH

H C

2.45

(CH2)2

Ea= 146.72 kJ mol

R O CH + H2C

H2 C C

R'

Ea= 107.94 kJ mol-1

R' (Int1) + H2O

R O + H2C C (CH2)2 R' 1.91 H H C C (CH2)2 R' (TSA) Ea= 44.11 kJ mol-1 H2 H2 H 2.15 H2 R O C C C C R' (TSB) R O C C CH2 + H2C R' H2 H2 H R O

A B

Ea= 74.59 kJ mol-1

k_i) R O CH2 (CH2)3 R' + OH

H

k_ii) R O C (CH2)3 R' (Int1)

A B

R O H R O C

H C

(CH2)3

2.34

C H2 H 1.94 O C R

C H2

(CH2)3

R' (Int1) + H2O

H2 C R' (TSA) R' (TSB)

R O C H

H2 CH2 + H2C C R'

Ea= 111.88 kJ mol-1

R + O

C H

(CH2)3

R'

Ea= 116.01 kJ mol-1 Fig. 5. The proposed degradation mechanisms of OH radicals attack on C5–C8 in the alkane chain of M1.

more difficultly than those involving the sulfonate and phenyl groups, such as Reactions (a), (b), (f), and (g). Both the butyl group and the alkane chain are stable against oxidative attack in fuel cell environments.

3.1.4. OH radical attack on the ether group The ether group is a common functional group used in the PEMs of fuel cells, and is found in Nafion, SPEEK, and SPES. In the copolymer compound we were interested in, the ether group connected the main chain and side chain for both the hydrophilic and hydrophobic domains. Recent research has pointed out a weakness in the ether group in Nafion [19]. Here, we propose a possible degradation mechanism caused by OH radical attacking atom C9 in the ether group based on DFT calculations. The mechanism is denoted as Reaction (l), and is shown in Fig. 6 together with its energy profile. In this simple reaction, only two steps are required for degradation to occur. First, OH radical attacks atom C9 to form a hydroxy cyclohexadienyl radical (Int2). Subsequently, the C9–O bond is broken, leading to a phenol molecule and an ether radical, which is in agreement with experimental analysis [28]. The ether radical produced can be degraded further to form smaller fragments, such as formic acid. Optimized structural parameters show that in TS1, the distance between the O atom of the OH radical and atom C9 is 2.03 Å, while in TS2, the distance between atom C9 and the ether–O bond is 1.91 Å. From the energy profile, we found that this reaction had a barrier height of  5.27 kJ mol  1, indicating its barrierless character, which is significantly lower than that of Reaction (k) whose barrier height was calculated to be 116.01 kJ mol  1 for the same

C9–O bond-breaking reaction. In addition, the barrier height of the C8–O bond-breaking reaction (44.11 kJ mol  1) is also higher than that of the C9–O bond-breaking reaction in Reaction (l). Therefore, atom C8 is a more stable site than atom C9 against OH radical attack. The calculated reaction enthalpy was  59.65 kJ mol  1, indicating an exothermic reaction. The low barrier height and exothermicity of this reaction show the weakness of the ether group against the OH radical attack when it is bonded to the phenyl ring. Atom C9 in Reaction (l) is as weak as atom C1 in Reaction (a). The conclusion that the ether group is weak against the OH radical attack when it is bonded to the phenyl ring is also applicable to hydrocarbon based PEMs with ether group in the main chain. For instance, the barrier heights for breaking either of the two C(phenyl) –O(ether) bonds in SPEEK by OH radical are only 3.22 kJ mol  1 and 7.98 kJ mol  1, while those barrier heights for SPES are 8.38 kJ mol  1 and 8.39 kJ mol  1, all calculated at the B3LYP (PCM)/6-311 þ þ G(2d,2p) level of theory (see Supporting information for further details). These barrier heights, although higher than the copolymer case, are relatively low, indicating the weakness of the ether group in the main chain. 3.2. Side chain degradation by H radicals 3.2.1. H radical attack on the phenyl group So far, we have focused on the degradation reactions initiated by OH radicals, but another common aggressive oxidative species in PEFCs is H radicals. Yu et al. reported that H radicals can be produced by OH radicals [17]. Here, we only report on the reactions arising from H radical attacks on atom C1 (denoted as Reaction (m)), atom C9 (denoted as Reaction (n)), abstracting the H

Y.-y. Zhao et al. / Journal of Membrane Science 487 (2015) 229–239

R= l)

CH2

SO3H

CH2 CH2

CH2

235

O R + OH

Erela (kJ.mol-1) 2.03 OH O

OH

R

O R + OH

O 1.91

TS1

0.0

0.00

TS2

-5.27

Int1

R

-14.92

-16.68 2.73 O

OH

OH

Int2

R

-55.80

OH O

+ O R

-56.03

R

Fig. 6. The proposed degradation mechanism of OH radical attack on C9 in the ether group of M1.

R=

m) R

n)

O

CH2

CH2

CH2

SO3H + H

R'=

CH2

O R' (TS2)

R

p) R R

SO3

2.44

_

SO3H (TS2)

SO3H + H 2.39

SO3H2 (TS2)

_

_

H 1.31 SO3H (TS1)

R

Ea= 27.93 kJ mol-1

Ea= 25.37 kJ mol-1

_

SO3H (Int1)

R

(TS1)

+ HSO3

R

R

+ SO3H

1.20

SO3

R

+H

SO3H

H O R' (Int1)

+ O R' H

CH2 CH2

R

H 1.76 O R' (TS1)

1.92

o) R

CH2

H 1.87 SO3H (TS)

R

O R' + H

_

CH2

+ H2SO3

Ea= 44.16 kJ mol-1

SO2

R

SO3H2 (Int1)

SO2

Ea= 78.61 kJ mol-1

Fig. 7. The proposed degradation mechanisms of H radicals attack on C1, C9, and the sulfonate groups of M1n and M1.

atom from the butyl group, and an attack on the sulfonate groups in M1n and M1 (denoted as Reactions (o) and (p), respectively). All these proposed reaction mechanisms and their energetics are shown in Fig. 7. The reaction mechanisms of Reactions (m) and (n) are similar to those of Reaction (b) and Reaction (l), except that the reaction initiator, OH radicals, is replaced by H radicals. The barrier heights were calculated to be 27.93 kJ mol  1 and 25.37 kJ mol  1 for Reactions (m) and (n), respectively. These values are somewhat higher than those of Reactions (b) and (l) (6.79 kJ mol  1 and 5.27 kJ mol  1), which indicates that H radicals are less aggressive than OH radicals in the degradation of this compound. In addition, we calculated the barrier height for the abstraction of a H atom from the

butyl group by H radicals to compare with the same action by OH radicals. We found that the barrier heights were also higher in the H radical reactions than in the OH radical reactions. It is known that in the degradation reaction of Nafion, an abstraction reaction of F atoms caused by H radicals occurs in a barrierless manner, and it is easier for this reaction to occur than an abstraction by OH radicals [18]. This indicates that hydrocarbon PEMs are more stable than Nafion against H radical attack in the initial steps of the degradation reactions. Two other possible degradation mechanisms are H radicals attacking the O atom of the sulfonate group, shown in Reactions (o) and (p). These mechanisms demonstrate that the reactants are degraded in a twostep reaction. The first step is an attack by H radical on the O atom of

236

Y.-y. Zhao et al. / Journal of Membrane Science 487 (2015) 229–239

OCH3

R'=OCH3

R= H3CO

HO 2.08

R'

q_i) R

R + OH

R (TS1)

R

R'

q_ii)

HO

R'

HO H R'

R'

R'

TS3

R

O 1.80 R O

R

H

R'

64.45

2.18 O2

R'

HO H R'

O2.64 R O

R

R

TS4

H

H

HO

48.44

O O R'

Int1+O2 13.73

Int3

Int2

0.00

H

H

HO

R'

O O

R R

R H R'

-29.36 R H

R'

H

R + OH R' HO R R'

R (Int1)

R R'

R'

HO R (Int2)

HO 2.10 R

R

2.24 R'

R'

R'

R'

O O

R' R H

R H

R' R (TS1)

R' R (TS2)

O O

R

-143.61

O2

HO 2.14 R' 3.15

H

HO

Int4

R'

R

H

R'

R

r)

TS5 -0.90

HO

R'

R

3.01

-6.30 HO

2.33 R'

R

TS2 0.0

Ea= 3.95 kJ mol-1

R (Int1) R'

Erela (kJ.mol-1) H

R'

R

R'

HO

H

R' R + HO

R R'

Ea= 51.26 kJ mol-1 Fig. 8. The proposed degradation mechanisms of OH radicals attack on C2 and C3 in M2.

the sulfonate group to form a H-added compound (Int1). The distance between the O and H atoms in the transition state (TS1) was 1.20 Å for M1n and 1.31 Å for M1. In the second step, Int1 undergoes a C–S bond-breaking reaction, resulting in the loss of the sulfonate group for both M1n and M1. The distances between the C1 and S atoms in TS2 were 2.44 Å for M1n and 2.39 Å for M1. The calculated barrier heights for the entire reaction were 44.16 kJ mol  1 for Reaction (o) and 78.61 kJ mol  1 for Reaction (p), which indicates that the proton-dissociated form is easier to degrade by H radicals than is the proton-undissociated compound. Nevertheless, these reactions have higher barrier energies than do Reactions (m) and (n), and much higher barrier energies than Reactions (a) and (b), which also shows that H radicals are less aggressive for hydrocarbon PEMs degradation than are OH radicals. The calculated reaction enthalpies,  34.02 kJ mol  1 for Reaction (o) and  53.27 kJ mol  1 for Reaction (p), indicate that both are exothermic reactions.

3.3. Main chain degradation The degradation mechanisms of the main chain of our copolymer were studied by investigating reactions of the model compound M2. As shown in Fig. 1(c), M2 is composed of three phenyl rings. This type of structure is easy to synthesize and possesses good stability, and therefore, is often used as the main chain of PEMs. Considering the repeat structure and symmetry of M2, only atoms C1, C2, and C3 were investigated as possible sites that could react with oxidative species, namely OH radicals and O2. Atom C1 is bonded to the ether group, which has the same degradation mechanism as atom C9 in M1 (shown in Reaction (l)). The addition of OH radicals to C1 results in the breaking of the C1–O bond, and the loss of the side chain. This reaction was calculated to be barrierless, which is the same as Reaction (l) in M1. Nevertheless, the main chain essentially retains its structure,

Y.-y. Zhao et al. / Journal of Membrane Science 487 (2015) 229–239

237

Table 1 Summary of all the simulated degradation mechanisms and the obtained energetics (in kJ mol  1) in the gas phase (g) and in water (w). Reac.

Attacking site

Oxidative species

Degradation site

Ea ðgÞa

Ea ðgÞb

Ea ðwÞc

ΔHðgÞd

ΔHðwÞe

(a) (b) (c) (d)

C1 of M1n C1 of M1 S of M1n S of M1

OH  OH  OH  OH 

C–S C–S C–S C–S

 16.94 7.90 116.10 118.36

 14.34 10.90 122.79 126.95

 0.23 6.79 143.47 141.04

 147.96  129.28  20.84  2.4

 152.46  123.36  25.65 2.05

(e_iiA) (e_iiB) (f) (g)

C2 C2 C3 C4

of of of of

M1 M1 M1 M1

OH  OH  , O2 OH  , O2 OH 

C–H bond C–S bond, phenyl ring Phenyl ring C4–C5 bond

37.18 53.41 26.96 18.11

40.06 61.54 34.22 16.94

54.67 65.33 25.41 19.36

8.28  85.32  260.33  56.22

23.43  86.42  267.81  55.17

(h) (i_iiA) (i_iiB) (j_iiA) (j_iiB) (k_iiA) (k_iiB)

C5 C6 C6 C7 C7 C8 C8

of of of of of of of

M1 M1 M1 M1 M1 M1 M1

OH  OH  OH  OH  OH  OH  OH 

C6–C7 bond C4–C5 bond C7–C8 bond C5–C6 bond C8–O bond C6–C7 bond C9–O bond

146.54 152.83 112.57 79.46 47.92 119.78 122.95

141.24 147.64 106.58 75.55 46.61 113.58 121.19

139.32 146.72 107.94 74.59 44.11 111.88 116.01

117.38 136.83 70.23 24.18  22.38 85.58 79.40

117.16 136.16 70.25 21.56  34.51 82.24 63.67

(l)

C9 of M1

OH 

C9–O bond

 6.52

 5.07

 5.27

 54.39

 59.65

(m) (n) (o) (p)

C1 of M1 C9 of M1 S of M1n S of M1

H H H H

C–S bond C9–O bond C–S bond C–S bond

26.69 25.34 34.70 60.96

27.75 26.35 25.13 73.37

27.93 25.37 44.16 78.61

 151.67  75.11  17.38  53.68

 148.63  82.87  34.02  53.27

(q) (r)

C2 of M2 C3 of M2

OH  , O2 OH 

Phenyl ring C–C bond

76.34 37.94

84.46 38.90

64.45 51.26

 27.16  3.84

 33.51 2.39

bond bond bond bond

a

The barrier height of the overall reaction calculated at the B3LYP/6-31þ G(d) level with ZPEs in the gas phase. The barrier height of the overall reaction calculated at the B3LYP/6-311 þ þ G(2d, 2p) level with ZPEs in the gas phase. The barrier height of the overall reaction calculated at the B3LYP/6-311þ þ G(2d, 2p) level with ZPEs in water. d The reaction enthalpies calculated at the B3LYP/6-311þ þ G(2d, 2p) level with ZPEs in the gas phase. e The reaction enthalpies calculated at the B3LYP/6-311þ þG(2d, 2p) level with ZPEs in water. b c

even if the side chain is detached. Atom C2 is a carbon atom that has no other substituents attached to the phenyl group. The degradation involving C2 is denoted as Reaction (q), and is shown in Fig. 8. This reaction is expected to be similar to that at atoms C2 and C3 in M1. The initial step involves OH radical attacking atom C2 to form a hydroxy cyclohexadienyl radical (Int1) with a low barrier height of 3.95 kJ mol  1. Int1 then further reacts with O2 and undergoes a series of reactions, resulting in the breaking of the C–C bond and ring opening. In contrast to Reactions (e) and (f), in this degradation mechanism, O2 prefers to react at the para position and the opposite side of the ring plane with respect to the OH group. This difference arises because of the steric hindrance from other substituents on the phenyl group in the main chain. The energy profile shows that the barrier height of the entire reaction is 64.45 kJ mol  1 , indicating the considerable stability of atom C2 against attack from oxidative species. Reaction (q) is slightly exothermic, with a reaction enthalpy of 33.51 kJ mol  1 . The degradation mechanism on atom C3, the carbon atom connecting the phenyl rings, denoted as Reaction (r), is also shown in Fig. 8. This is also a simple reaction, consisting of only two steps. First, OH radical attacks C3 to form a hydroxy cyclohexadienyl radical compound (Int1), followed by the C–C3 bondbreaking reaction. We found that Reaction (r) can result in the direct scission of the main chain by overcoming the barrier height of 51.26 kJ mol  1 . This value is lower than that of Reaction (q), but is still higher than those of Reactions (a), (b), and (l) occurring on the side chain. Moreover, the calculated reaction enthalpy was 2.39 kJ mol  1 , suggesting that this is not an exothermic reaction. Therefore, our results suggest that this type of main chain is stable in fuel cell operating conditions.

3.4. Discussion As discussed above, various types of possible degradation reactions of hydrocarbon PEMs were studied by performing DFT calculations on the model compounds M1 and M2. All these reactions, including their reactants, degradation sites, barrier heights, and reaction enthalpies are summarized in Table 1. We found that the relative energetics obtained show similar trends at the different levels of theory. A comparison of the barrier heights of all the degradation reactions indicates that Reactions (a) and (l) are the most energetically favorable (barrierless) reactions. Reaction (a) involves OH radicals attacking atom C1 of the C1–S bond in the proton-dissociated M1 model, resulting in the loss of the sulfonate group, while Reaction (l) involves OH radicals attacking atom C9 of the C9–O bond, leading to the scission of the side chain. Both reactions involve a single OH radical. These results suggest that the two weakest sites of the hydrocarbon PEMs are the C–S bond in the benzenesulfonate group and the C–O bond in the ether group. Reactions (b), (g), and (f) also have low barrier heights of 6.79 kJ mol  1, 19.36 kJ mol  1, and 1 25.41 kJ mol , respectively. Other degradation reactions possess either moderate or high barrier heights, indicating their relative stability, especially for the main chain. We also found that all the chemical bonds directly bonded to the phenyl ring, for example, the C1–S bond, the C4–C5 bond, and the C9–O bond, are easily broken by radical attack. All the degradation reactions from Reactions (m) to (p) are initiated by H radicals. When compared with reactions initiated by OH radicals, H radicals are less reactive in the initiation step of all the degradation reactions, which is different from the case of Nafion. Proposing effective strategies to improve the stability of membranes is the main purpose of many degradation studies. On the

238

Y.-y. Zhao et al. / Journal of Membrane Science 487 (2015) 229–239

one hand, reducing the number of oxidative species is a good way to prevent several oxidation reactions. Yu et al. have proposed modifying the catalysts to prevent the formation of hydrogen peroxide, which potentially leads to OH radical formation [17,18]. On the other hand, we could increase the antioxidative capability of membranes to make degradation more difficult. To this end, the structure should be optimized to remove, or stabilize, the labile sites determined by theoretical studies. For example, in Reaction (a), atom C1 is weak against OH radical attack. Therefore, it may be useful to replace the benzenesulfonate group by a phenylethanesulfonic acid group, because the C atom in the alkane chain is more stable than that in the phenyl group. Concerning atom C9 in Reaction (l), it may be a good idea to avoid using ether groups in synthesis. Similarly, atom C3 in M1 may be stabilized if the H atom in the phenyl ring is replaced by a F atom or a CH3 group. Another possibility is to find compounds that can protect the membrane from oxidative damage, just like vitamins C and E protect our body from the destructive effects of free radicals, as reported in recent research [40]. To find such compounds, we could borrow ideas from the field of biochemistry.

4. Conclusions Herein, we have presented a theoretical study on the chemical stability of hydrocarbon PEMs. In total, 18 degradation reaction mechanisms for the model compounds of the studied polymer have been proposed. These reactions were all induced by OH radicals, H radicals, and O2, which are considered to be the abundant species in PEMFCs. The obtained energetics show that for Reactions (a) and (l), the barrier heights were  0.23 kJ mol  1 and 5.27 kJ mol  1, respectively, and that both were barrierless reactions against OH radical attack. Reactions (m) and (n) also had relatively low barrier heights against the H radical attack on the same C atoms as for Reactions (a) and (l). The benzenesulfonate group (C1) and the ether group bonded to the phenyl ring (C9) have been calculated to be the two weakest sites against both OH and H radical attack in our model compounds. Two other low barrier heights of 19.36 kJ mol  1 and 25.41 kJ mol  1 appear in Reactions (g) and (f), namely OH radical attack on atoms C4 and C3 in M1. On the other hand, the alkane chain shows good stability against radical attack. The proposed degradation mechanisms for the main chain have clarified that the reactants have to overcome a barrier height of 64.45 kJ mol  1 on C2 and 51.26 kJ mol  1 on C3 for degradation to occur, which indicates a moderate stability of a main chain composed of phenyl groups. Suggestions for strengthening the weak sites in the hydrocarbon PEMs have also been made according to the proposed degradation mechanisms. These include a reduction in the number of oxidative species and modification of the polymer structure, such as replacing the benzenesulfonate group by a phenylethanesulfonic group. Alternatively, the chemical stability of the membranes may be improved by introducing an antioxidant to protect the membrane from radical attack.

Acknowledgments We gratefully acknowledge financial support from the New Energy and Industrial Technology Development Organization (NEDO). We thank Professor Masahiro Rikukawa for helpful discussions.

Appendix A. Supplementary data Supplementary data associated with this paper can be found in the online version at http://dx.doi.org/10.1016/j.memsci.2015.04. 005.

References [1] J. Healy, C. Hayden, T. Xie, K. Olson, R. Waldo, M. Brundage, H. Gasteiger, J. Abbott, Aspects of the chemical degradation of PFSA ionomers used in PEM fuel cells, J. Power Sources 127 (2004) 127–134. [2] S.D. Knights, K.M. Colbow, J. St-Pierre, D.P. Wilkinson, Aging mechanisms and lifetime of PEFC and DMFC, Fuel Cells 5 (2005) 302–308. [3] A. Collier, H. Wang, X.Z. Yuan, J. Zhang, D.P. Wikinson, Degradation of polymer electrolyte membranes, Int. J. Hydrog. Energy 31 (2006) 1838–1854. [4] A. Panchenko, H. Dilger, J. Kerres, M. Hein, A. Ullrich, T. Kaz, E. Roduner, In situ spin trap electron paramagnetic resonance study of fuel cell processes, Phys. Chem. Chem. Phys. 6 (2004) 2891–2894. [5] M. Kitazawa, A.Y. Nosaka, Y. Nosaka, Radical formation in polymer electrolyte fuel cell components as studied by ESR spectroscopy, J. Appl. Electrochem. 38 (2008) 491–496. [6] L. Gubler, S.M. Dockheer, W.H. Koppenol, Radical (HO; H and HOO) formation and ionomer degradation in polymer electrolyte fuel cells, J. Electrochem. Soc. 158 (2011) B755–B769. [7] H. Hou, M.L.D. Vona, P. Knauth, Durability of sulfonated aromatic polymers for proton-exchange-membrane fuel cells, ChemSusChem 4 (2011) 1526–1536. [8] Q. Guo, P.N. Pintauro, H. Tang, S. O’Connor, Sulfonated and crosslinked polyphosphazene-based proton-exchange membranes, J. Membr. Sci. 154 (1999) 175–181. [9] J. Yu, B. Yi, D. Xing, F. Liu, Z. Shao, Y. Fu, H. Zhang, Degradation mechanism of polystyrene sulfonic acid membrane and application of its composite membranes in fuel cells, Phys. Chem. Chem. Phys. 5 (2003) 611–615. [10] A. Panchenko, H. Dilger, J. Kerres, M. Hein, A. Ullrich, T. Kazc, E. Roduner, In situ spin trap electron paramagnetic resonance study of fuel cell processes, Phys. Chem. Chem. Phys. 6 (2004) 2891–2894. [11] J. Maruyama, M. Inaba, Z. Ogumi, Effect of fluorinated alcohol on the kinetics of cathodic oxygen reduction at gold electrodes, Electrochim. Acta 45 (1999) 415–422. [12] A. Panchenko, DFT investigation of the polymer electrolyte membrane degradeation caused by OH radicals in fuel cells, J. Membr. Sci. 278 (2006) 269–278. [13] J. Lawrence, T. Yamaguchi, The degradation mechanism of sulfonated poly (arylene ether sulfone)s in an oxidative environment, J. Membr. Sci. 325 (2008) 633–640. [14] C.H. Park, C.H. Lee, M.D. Guiver, Y.M. Lee, Sulfonated hydrocarbon membranes for medium-temperature and low-humidity proton exchange membrane fuel cells (PEMFCs), Prog. Polym. Sci. 36 (2011) 1443–1498. [15] M. Kumara, S.J. Paddison, Side-chain degradation of perfluorosulfonic acids: an ab initio study, J. Mater. Res. 27 (2012) 1982–1991. [16] T. Tokumasu, I. Ogawa, M. Koyama, T. Ishimoto, A. Miyamoto, A DFT study of bond dissociation trends of perfluorosulfonic acid membrane, J. Electrochem. Soc. 158 (2011) B175–B179. [17] T.H. Yu, Y. Sha, W.-G. Liu, B.V. Merinov, P. Shirvanian, W.A. Goddard, Mechanism for degradation of Nafion in PEM fuel cells from quantum mechanics calculations, J. Am. Chem. Soc. 133 (2011) 19857–19863. [18] T.H. Yu, W.-G. Liu, Y. Sha, B.V. Merinov, P. Shirvanian, W.A. Goddard, The effect of different environments on Nafion degradation: quantum mechanics study, J. Membr. Sci. 437 (2013) 276–285. [19] L. Ghassemzadeh, T.J. Peckham, T. Weissbach, X. Luo, S. Holdcroft, Selective formation of hydrogen and hydroxyl radicals by electron beam irradiation and their reactivity with perfluorosulfonated acid ionomer, J. Am. Chem. Soc. 135 (2013) 15923–15932. [20] L. Jörissen, V. Gogel, J. Kerres, J. Garche, New membranes for direct methanol fuel cells, J. Power Sources 105 (2002) 267–273. [21] P. Xing, G.P. Robertson, M.D. Guiver, S.D. Mikhailenko, K. Wang, S. Kaliaguine, Synthesis and characterization of sulfonated poly(ether ether ketone) for proton exchange membranes, J. Membr. Sci. 229 (2004) 95–106. [22] Y.-K. Choe, E. Tsuchida, T. Ikeshoji, A. Ohira, K. Kidena, An ab initio modeling study on a modeled hydrated polymer electrolyte membrane, J. Phys. Chem. B 114 (2010) 2411–2421. [23] M.A. Hickner, H. Ghassemi, Y.S. Kim, B.R. Einsla, J.E. McGrath, Alternative polymer systems for proton exchange membranes (PEMs), Chem. Rev. 104 (2004) 4587–4612. [24] Y. Zhao, E. Tsuchida, Y.-K.C.T. Ikeshoji, M.A. Barique, A. Ohira, Ab initio studies on the proton dissociation and infrared spectra of sulfonated poly(ether ether ketone) (SPEEK) membranes, Phys. Chem. Chem. Phys 16 (2014) 1041–1049. [25] G. Hübner, E. Roduner, EPR investigation of HO radical initiated degradation reactions of sulfonated aromatics as model compounds for fuel cell proton conducting membranes, J. Mater. Chem. 9 (1999) 409–418. [26] S. Takamuku, P. Jannasch, Properties and degradation of hydrocarbon fuel cell membranes: a comparative study of sulfonated poly(arylene ether sulfone)s with different positions of the acid groups, Polym. Chem. 3 (2012) 1202–1214.

Y.-y. Zhao et al. / Journal of Membrane Science 487 (2015) 229–239

[27] S.M. Dockheer, L. Gubler, P.L. Bounds, A.S. Domazou, G.G. Scherer, A. Wokaunb, W.H. Koppenol, Damage to fuel cell membranes. reaction of HO with an oligomer of poly(sodium styrene sulfonate) and subsequent reaction with O2, Phys. Chem. Chem. Phys. 12 (2010) 11609–11616. [28] C. Perrot, L. Gonon, M. Bardet, C. Marestin, A. Pierre-Bayle, G. Gebel, Degradation of a sulfonated aryl ether ketone model compound in oxidative media (sPAEK), Polymer 50 (2009) 1671–1681. [29] K. Umezawa, T. Oshima, M. Yoshizawa-Fujita, Y. Takeoka, M. Rikukawa, Synthesis of hydrophilic–hydrophobic block copolymer ionomers based on polyphenylenes, ACS Macro Lett. 1 (2012) 969–972. [30] A. Yokoyama, R. Miyakoshi, T. Yokozawa, Chain-growth polymerization for poly(3-hexylthiophene) with a defined molecular weight and a low polydispersity, Macromolecules 37 (2004) 1169–1171. [31] E.E. Sheina, J. Liu, M.C. Iovu, D.W. Laird, R.D. McCullough, Chain growth mechanism for regioregular nickel-initiated cross-coupling polymerizations, Macromolecules 37 (2004) 3526–3528. [32] C. Lee, W. Yang, R.G. Parr, Development of the Colle–Salvetti correlationenergy formula into a functional of the electron-density, Phys. Rev. B 37 (1988) 785–789. [33] A.D. Becke, A new mixing of Hartree–Fock and local density-functional theories, J. Chem. Phys. 98 (1993) 1372–1377. [34] M. Cossi, V. Barone, R. Cammi, J. Tomasi, Ab initio study of solvated molecules: a new implementation of the polarizable continuum model, Chem. Phys. Lett. 255 (1996) 327–335. [35] E. Cancès, B. Mennucci, J. Tomasi, A new integral equation formalism for the polarizable continuum model: theoretical background and applications to isotropic and anistropic dielectrics, J. Chem. Phys. 107 (1997) 3032–3041.

239

[36] J. Tomasi, B. Mennucci, R. Cammi, Quantum mechanical continuum solvation models, Chem. Rev. 105 (2005) 2999–3093. [37] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery, . Jr, J.E. Peralta, F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, Ö. Farkas, J.B. Foresman, J.V. Ortiz, J. Cioslowski, D.J. Fox, Gaussian 09 Revision A.1, Gaussian Inc., Wallingford, CT, 2009. [38] S. Raoult, M.-T. Rayez, J.-C. Rayez, R. Lesclaux, Gas phase oxidation of benzene: kinetics, thermochemistry and mechanism of initial steps, Phys. Chem. Chem. Phys. 6 (2004) 2245–2253. [39] I.V. Tokmakov, G.-S. Kim, V.V. Kislov, A.M. Mebel, M.C. Lin, The reaction of phenyl radical with molecular oxygen: a G2M study of the potential energy surface, J. Phys. Chem. A 109 (2005) 6114–6127. [40] Y. Yao, J. Liu, W. Liu, M. Zhao, B. Wu, J. Gu, Z. Zou, Vitamin E assisted polymer electrolyte fuel cells, Energy Environ. Sci. 7 (2014) 3362–3370.