Ab initio studies on the proton dissociation and infrared spectra of sulfonated poly(ether ether ketone) (SPEEK) membranes.

PCCP View Article Online

Published on 27 November 2013. Downloaded by UTSA Libraries on 09/09/2014 16:16:25.

PAPER

Cite this: Phys. Chem. Chem. Phys., 2014, 16, 1041

View Journal | View Issue

Ab initio studies on the proton dissociation and infrared spectra of sulfonated poly(ether ether ketone) (SPEEK) membranes Yuan-yuan Zhao,*a Eiji Tsuchida,b Yoong-Kee Choe,b Tamio Ikeshoji,ab Mohammad Abdul Bariquea and Akihiro Ohiraa SPEEK is known to possess high proton conductivity at high water content, being comparable with other popular membranes used in fuel cells, such as Nafion and sulfonated polyethersulfone (SPES). However, much less is known about its fundamental properties, including the status of proton dissociation and spectral features. In this work, the properties of two model molecules of SPEEK, M1 (20 atoms), M2 (50 atoms) and their hydrated systems, M1 + nH2O and M2 + nH2O (n = 1–9), have been investigated using static electronic structure calculations and the ab initio molecular dynamics (MD) method. Optimized structures for all of the systems and the trajectories of M1 + nH2O (n = 3–6) at finite temperatures have been computed using density functional theory at the B3LYP level of theory. Proton dissociation has been discussed in detail, especially for n = 3 and n = 4. In addition, the infrared spectra of SPEEK and its

Received 26th July 2013, Accepted 4th November 2013 DOI: 10.1039/c3cp53146e

hydrated systems have been studied using a combination of theory and experiment. The characteristic bands of SPEEK and the surrounding water clusters have been assigned with emphasis on their relationship with the degree of proton dissociation. We have found that the hydronium ion stretching modes, which appear in the 2000–3000 cm

www.rsc.org/pccp

1

region in static electronic structure calculations, are not observed experi-

mentally. This discrepancy is explained by the stationary structure and the temperature effect.

1 Introduction Polymer electrolyte membrane (PEM) fuel cells are receiving increasing attention from the developers of portable and stationary power sources, which are employed in ‘‘green’’ vehicles and power generation equipment. In PEM fuel cells, there are two important ingredients involved, PEMs and electrodes. PEMs are not only utilized to separate the anode and the cathode but also to transfer the proton and screen the small gas molecules.1–4 At present, Nafion is one of the most popular membranes used in fuel cells because of its high proton conductivity and excellent chemical, thermal and mechanical stability. However, drawbacks such as high production cost and low operating temperature (o80 1C) have limited its wide application.5–8 Therefore, development of a novel and excellent substitute for Nafion is an urgent issue. Several series of promising membranes have been presented, such as ionic liquid–polymer membranes and copolyimide membranes.9–11

a

Fuel Cell Cutting-Edge Research Center (FC-Cubic), 2-3-26 Aomi, Koto-ku, Tokyo 135-0064, Japan. E-mail: [email protected] b Nanosystem Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba Central 2, Umezono 1-1-1, Tsukuba 305-8568, Japan

This journal is © the Owner Societies 2014

In particular, the aromatic polymer family is one of the most attractive candidates because of their low cost, low gas permeability and good chemical and thermal stability.12–15 Sulfonated poly(ether ether ketone) (SPEEK), as a member of this polymer family, has been synthesized and extensively investigated experimentally.16–20 It is also currently very popular for forming composite membranes with Nafion, hydrocarbon membranes and nanotubes to improve the performance of fuel cells.21–24 Therefore, a thorough understanding of its fundamental properties at the electronic and atomic level of theory is necessary for its applications under realistic conditions with high reliability. As is well known, proton conductivity is a key property of these membrane materials. Many experimental studies on the proton conductivity of SPEEK membranes have been reported.25–28 In particular, some aromatic PEMs have been found to have high proton conductivities that are comparable with Nafion at high water content.15,29–32 However, an ideal membrane for fuel cells should also possess high proton conductivities at low water content. Because pure SPEEK membranes exhibit relatively low proton conductivities at low water content, much effort, e.g., forming a composite membrane with other polymers, has been made to improve its proton conductivity. Generally, the proton conductivity of a membrane is related to two properties. One is the status of proton dissociation4,29,30,32 and the other is the

Phys. Chem. Chem. Phys., 2014, 16, 1041--1049 | 1041

View Article Online


Paper

mechanism of proton transfer.33 Therefore, extensive understanding regarding such properties at an atomistic level is crucial to improve the proton conductivity of SPEEK. Herein, we will employ ab initio electronic structure calculations and ab initio MD simulations to investigate the proton dissociation of SPEEK at various hydration levels. Spectroscopy is also an effective means to study structures and dynamical properties of the target system. The IR spectra of SPEEK membranes have been measured and reported by several groups.16–18,23 However, the assignment for the vibrational bands is still controversial. Here, we will discuss the IR spectra of SPEEK in detail and explain the status of the water cluster surrounding the polymer based on the calculated and measured IR spectra. This paper is organized as follows. In Section 2, we describe the computational and experimental details used herein. In Section 3, we report the optimized structures and IR spectra of model molecules of SPEEK with different hydration numbers. We also discuss proton dissociation and the characteristic vibrational bands of SPEEK membranes, as well as the IR spectra and role of the water cluster surrounding the SPEEK polymer. Section 4 concludes.

2 Method 2.1

Quantum chemical calculations

Simulations of intact PEMs by ab initio electronic structure calculations are always difficult because of the large size and complicated nature of the system. Paddison et al. suggested an effective approach to study the properties of PEMs theoretically using a simple model representation.34–36 Here, we selected two model molecules of SPEEK, M1 and M2, to perform the simulations. The chemical representations of the SPEEK polymer and its model molecules are shown in Fig. 1. M1 is a simple model that includes only 20 atoms and the basic ingredients of SPEEK, i.e., –SO3H and C–O–C groups. The hydrogen atom and the –CH3 group are used to terminate the main chain of M1. This model is mainly used to study the local structures near the sulfonate group. M2 includes 50 atoms and is a more realistic model with a longer main chain, identical to a monomer of SPEEK, where all of the main functional groups, –SO3H, C–O–C and CQO, are included. The phenyl group and the H atom are used to terminate the main chain of M2. Calculations were performed using density functional theory (DFT), employing the B3LYP hybrid functional.37,38 Preliminary calculations for

Fig. 1 The chemical structures of (a) SPEEK, and its model representations, (b) M1 and (c) M2.

1042 | Phys. Chem. Chem. Phys., 2014, 16, 1041--1049

PCCP

hydrated models, M1 + nH2O (n = 1–9) and M2 + nH2O (n = 1–9), were carried out at the B3LYP/6-31+G* level of theory. Most stable structures for each n were further refined at the B3LYP/ 6-311++G(2d,2p) level of theory. Vibrational frequency calculations were performed to ensure that the optimized structures are minima as well as to simulate the IR spectra of SPEEK. To estimate the effects of basis set superposition errors, additional calculations were conducted with the inclusion of the counterpoise (CP) method39 for M1 + 3H2O at the B3LYP/6-311G**, B3LYP/6-31+G* and B3LYP/6-311++G(2d,2p) levels of theory and for M2 + 3H2O at the B3LYP/6-311++G(2d,2p) level of theory. In what follows, this approach is called static electronic structure calculations. 2.2

Ab initio MD simulations

The ab initio atom-centered density matrix propagation (ADMP) approach40–42 was employed to simulate the structural and dynamical properties of SPEEK at finite temperatures. To keep the computational cost manageable, only the small model M1 was used in the ADMP simulations. The trajectories of M1 + nH2O (n = 3–6) were obtained using the B3LYP hybrid density functional with the 6-31+G* basis set. A time step of 0.1 fs was applied in these simulations. Trajectories were started from the local minima obtained by geometry optimization. Because of the small system size and isolated boundary conditions used in our simulations, the obtained trajectories included unreasonable ones, in which some of the water molecules moved too far to contribute to the proton dissociation. In particular, unless all three Os (oxygen atoms in the sulfonate group) atoms were hydrogen bonded to Hw (hydrogen atoms in the water molecule) atoms, the proton could not be dissociated regardless of the hydration number. Therefore, we first excluded unreasonable trajectories whose Os Ow (the oxygen atoms in H2O and H3O+ molecules) distribution is extended beyond 6 Å, and used the remaining 1–3 trajectories of length 10 ps for each hydration number. We excluded the first 4 ps for equilibration and used the last 6 ps for analysis. The average temperature was approximately 300 K in all runs. All of the calculations in this work were carried out using the Gaussian09 program package.43 Molden44 and VMD45 were used to visualize the structures and dynamics of the studied systems. 2.3

FTIR–ATR spectra

The Fourier transform infrared (FTIR) spectra of the SPEEK membranes were recorded using a PerkinElmer Spectrum 100 FTIR spectrometer system operating with a deuterated triglycine sulfate (DTGS) detector and KBr beam splitter for the range 400–4000 cm 1 in transmission mode. Each spectrum was recorded with a spectral resolution of 4 cm 1 and 256 scans (5 s per scan) were averaged per spectrum. The water vapor was produced in a specially designed temperature- and pressure-controlled water vessel system and was passed to the sample chamber through a transfer tube by a dry N2 flow into the vessel. SPEEK was synthesized by the sulfonation of a commercial PEEK polymer with concentrated sulfuric acid.46 The SPEEK polymer was casted from N-methyl-2-pyrrolidone (NMP) solutions onto a flat


View Article Online

PCCP

glass substrate and dried at 80 1C under vacuum. The membrane was immersed in 0.5 M sulfuric acid overnight for complete protonation, then rinsed with deionized water and dried under ambient conditions. The thickness was ca. 50 mm (SPEEK) and the ion exchange capacity (IEC) was 1.40 meq g 1, as determined by back titration.

3 Results and discussion Published on 27 November 2013. Downloaded by UTSA Libraries on 09/09/2014 16:16:25.

3.1

Proton dissociation

3.1.1 Optimized structures of M1. The selected stable structures of the hydrated systems, M1 + nH2O (n = 1–4, 6, 9), calculated at the B3LYP/6-311++G(2d,2p) level of theory, are shown in Fig. 2, where r1 and r2 denote the distances between the proton and its donor (or acceptor) oxygen atoms. For convenience, we have named several atoms in the model molecules. In addition to Os, Hw, and Ow, the proton initially bonded to a sulfonate group is named Hp. Then, r1 is the distance of Os Hp and r2 is the distance of Ow Hp. The listed distances show that, with the increase of the hydration number from 1 to 3, r1 increases, while r2 decreases, which indicates the degree of proton dissociation. The proton is found to move gradually from the sulfonate group to the water cluster with the increase in hydration number. In addition, two important changes in the distances are observed as n is increased from 3 to 9. One is observed on going from n = 3 to n = 4. The distance r1 jumps from 1.06 Å for n = 3 to 1.41 Å for n = 4. The distance r2 decreases from 1.46 Å for n = 3 to 1.06 Å for n = 4. These changes in r1 and r2 indicate a transformation of covalent and hydrogen bonds for Os–Hp Ow from n = 3 to n = 4. In other words, the proton is detached from the sulfonate group to form a hydronium ion with the nearest water molecule when M1 is hydrated by four water molecules. Another change is observed on going from n = 5 to n = 6. The distance r1 increases from

Fig. 2 The optimized structures and interatomic distances (in Å) of M1 and its hydrated systems M1 + nH2O (n = 1–4, 6, 9) calculated at the B3LYP/ 6-311++G(2d,2p) level of theory. Red, gray, yellow and white spheres represent oxygen, carbon, sulfur and hydrogen atoms, respectively.


Paper

1.56 Å to 3.26 Å and r2 increases from 1.02 Å to 1.08 Å, which means that the proton is completely dissociated and further transferred to the center of the water cluster. Our results show that four and six are two key hydration numbers for the dissociation of a proton in SPEEK. In a recent computational study on the proton dissociation of various imide and sulfonic acid ionomers, Clark et al.32 included the CP corrections to minimize the effects caused by basis set superposition error in their calculations. They observed spontaneous proton dissociation upon addition of only three waters at the B3LYP/6-311G** level of theory for several aromatic sulfonic acid molecules. To investigate the effect of CP corrections in our case, we have repeated the geometry optimization for M1 + 3H2O at the B3LYP/6-31+G*, B3LYP/6-311G** and B3LYP/6-311++G(2d,2p) levels of theory, including the CP corrections. The calculated results show that the obtained energies for a configuration where the proton is not dissociated (hereafter referred to as the neutral configuration) are about 3–4 kJ mol 1 lower than the energy where the proton is dissociated (hereafter referred to as the ionic configuration) at the B3LYP/6-31+G* and B3LYP/ 6-311++G(2d,2p) levels of theory. Such a trend does not differ from that determined from the method without the CP corrections. However, the energy of the neutral configuration computed at the B3LYP/6-311G** level of theory is found to be about 3 kJ mol 1 higher than that of the ionic configuration. Therefore, diffuse functions are found to play a nonnegligible role in the geometry optimizations and total energies, while the CP corrections affect the results obtained at the level of theory employed in the present study only to a very small extent. The energy difference between the neutral and ionic configurations is too small to determine if proton dissociation occurs at n = 3. It can be concluded that hydration by four waters is sufficient for proton dissociation. 3.1.2 Optimized structures of M2. To estimate the size effect of the model molecules, the larger model M2, which is equal to the monomer of SPEEK, was also investigated at the B3LYP/6-311++G(2d,2p) level of theory. The optimized structures are shown in Fig. 3, together with the distances of Os Hp (r1) and Ow Hp (r2). The results show a similar trend in comparison with those of M1. However, one difference is found on going from n = 2 to n = 3 for the first jump of the proton, in contrast to that found on going from n = 3 to n = 4 in the case of M1. The corresponding distance r1 increases from 1.01 Å to 1.45 Å, while the distance r2 decreases from 1.59 Å to 1.05 Å. This indicates that the proton is detached from the sulfonate group at n = 3. The second jump of the proton is also found on going from n = 5 to n = 6. The distance r1, 1.53 Å, increases to 3.22 Å, indicating that the proton is transferred to the center of the water cluster, while r2 only slightly decreases from 1.03 Å to 1.01 Å. When comparing the results of M2 with those of M1, the first jump point of the proton is found to be a major difference, and thus it is worth paying attention to the state of the proton at n = 3. We also optimized the neutral structure for n = 3 at the same level of theory. The energy is only slightly (2.1 kJ mol 1) higher than that of the ionic configuration. Virtually the same energy difference (2.1 kJ mol 1) is


View Article Online


Paper

Fig. 3 The optimized structures and interatomic distances (in Å) of M2 and its hydrated systems M2 + nH2O (n = 1–4, 6, 9) calculated at the B3LYP/6-311++G(2d,2p) level of theory.

obtained when the CP correction is included. The small energy difference between the neutral and ionic configurations of M2 + 3H2O agrees with that observed in M1 + 3H2O, except for the sign. This result is also consistent with the trend reported in a previous study on SPES (2.5 kJ mol 1), while both are much lower than that for Nafion (10.1 kJ mol 1).15 A transition state connecting the ionic and neutral configurations for M2 + 3H2O has been obtained and the potential energy surface (PES) is shown in Fig. 4. We find that the energy barrier is calculated to be 0.4 kJ mol 1 for the forward reaction and 2.5 kJ mol 1 for the reverse reaction. Such a small energy barrier for both directions indicates that the two states of M2 + 3H2O, ionic

Fig. 4 Schematic representation of the potential energy profile of proton dissociation for M2 + 3H2O calculated at the B3LYP/6-311++G(2d,2p) level of theory. The values (in kJ mol 1) are energies relative to the neutral structure.


PCCP

and neutral, may transform frequently in both directions. For n = 4, the ionic configuration was calculated to be more stable compared with the neutral configuration. Therefore, we have concluded here that (1) the size of the model molecule has a slight effect on the proton dissociation, and (2) the proton may be weakly detached from the sulfonate group when hydrated by three waters, while the proton is readily detached when hydrated by four waters. 3.1.3 ADMP calculations. To determine the proton dissociation properties of SPEEK in more detail, we have analyzed the trajectories of ADMP simulations for M1 + nH2O (n = 3–6), calculated at the B3LYP/6-31+G* level of theory. Radial distribution functions (RDFs) for Os Ow and Os H (Hw and Hp) are shown in Fig. 5. Fig. 5(a) shows RDFs between Os and Ow for M1 + nH2O (n = 3–6), which show two main peaks for all systems, indicating two hydration shells. For the system hydrated by three waters, the first shell is located at approximately 2.75 Å, which indicates that the state of water and sulfonate is given by SO3H OH2 or SO2 H2O and the proton is undissociated. For the other three systems, the first shell is shown by a broader peak, located between 2.5 Å and 2.8 Å. The peak at approximately 2.5 Å corresponds to the formation of SO3 H OH2 associated with proton dissociation, while the peak at approximately 2.8 Å corresponds to SO2 H2O. Another large peak for all four systems, observed at approximately 4.0 Å away from sulfonate, arises from a second hydration shell.

Fig. 5

The radial distribution function between (a) Os and Ow, (b) Os and H.


View Article Online


PCCP

Fig. 5(b) shows RDFs for Os H (Hw and Hp) in the 0.8–4.0 Å region and a close-up of the 1.1–2.3 Å region. There are several main peaks for the distribution of H atoms observed in these RDFs. Here, we discuss only those less than 2.0 Å. For all four systems (n = 3–6), there is one common peak at approximately 1.0 Å, which indicates that the state of the proton is undissociated and still bonded to sulfonate. This peak was also observed in our previous study on the hydrated SPES system,15 while no peak was found for Nafion because of the complete dissociation of the sulfonate groups.7 The distributions at approximately 1.3–1.6 Å indicate the formation of a hydrogen bond between the dissociated proton and sulfonate, which is only observed for the systems hydrated by four, five and six waters. For the system hydrated by three waters, the distribution of Os H goes to zero at about 1.2 Å, which indicates that Os–Hp–Ow may form but no proton dissociation occurs. Integration of the RDFs for Os H gives the probability of dissociation for each hydration number. If we take 1.2 Å as the threshold distance of the O–H covalent bond, 15% of the proton is dissociated when hydrated by four waters, 61% when hydrated by five waters, and 75% when hydrated by six waters. Therefore, the results calculated by ab initio MD simulations are in good agreement with our static electronic structure calculations. In summary, the proton in the SPEEK models is found to remain undissociated when the sulfonate group is hydrated by three waters, while the proton is dissociated more readily when hydrated by four or more waters. When hydrated by six waters, the proton has a 75% probability of being dissociated. Although our hydration number may not be completely equal to l, defined as the number of water molecules per sulfonic group in bulk systems, our results still give an intuitive understanding of the status of proton dissociation for SPEEK. 3.2

IR spectrum

The IR spectra of the model systems, M1, M2 and their hydrated systems M1 + nH2O and M2 + nH2O (n = 1–9) were obtained by static electronic structure calculations (B3LYP/6-311++G(2d,2p)). We also measured the FTIR spectra of the SPEEK membrane experimentally at room temperature under the relative humidities (RHs) of 38% and 100%. Some of the calculated and measured IR spectra are presented in Fig. 6, where the calculated IR spectra of M1, M2 and their hydrated systems are shown by black lines, and the measured IR spectra of SPEEK by colored lines. All of the calculated vibrational frequencies shown here are unscaled. Here, we will separate the vibrational bands into two regions for discussion, the low-wavenumber region (800–2000 cm 1) and the high-wavenumber region (2000–4000 cm 1). 3.2.1 IR spectrum in the 800–2000 cm 1 region. As shown in Fig. 6, in the 800–2000 cm 1 region, the calculated IR spectra of M2 and its hydrated systems show quite good agreement with those from experiment. However, the calculated IR spectra of M1 and its hydrated systems are in substantial disagreement with the experimental results. Indeed, it is not difficult to understand that the lack of main molecular fragments in the small model M1, e.g., carbonyl and phenyl groups, is responsible for the lack of the corresponding characteristic vibrations, which


Paper

Fig. 6 Comparison of the absorption bands calculated at the B3LYP/ 6-311++G(2d,2p) level of theory (black lines) and the experimental FTIR spectra for the polymer (colored lines).

generally appear in the low-wavenumber region (800–2000 cm 1). This result suggests that the size effect of model molecules has to be carefully considered when simulating the IR spectra of polymer membranes. Compared with the experimental results, the calculated unscaled vibrational bands exhibit a slight blue shift of 1–3%. The vibrational modes of M2 and its hydrated systems have been assigned and the results are given in Table 1. According to the assignments, the vibrations in the 800–2000 cm 1 region mainly originate from the vibrations of the model molecule, except for the bending mode of water molecules. It is not easy to make clear assignments for all of the bands because most of them are coupled to each other. Here, we only made assignments for those characteristic vibrations of CQO, OQSQO, C–O–C and the bending modes of H2O. The feature in the calculated spectra at approximately 1700 cm 1, slightly depending on the studied systems (1706 cm 1 for M2, 1700 cm 1 for M2 + 3H2O, 1698 cm 1 for M2 + 6H2O), is likely to be dominated by contributions from the stretching mode of CQO, while the corresponding peak was found at 1651 cm 1 (ref. 17) and 1653 cm 1 (ref. 18) in previous experiments. In our measured spectra, it is observed at about 1650 cm 1 as a small shoulder of the bending modes of H2O. Features at around 1214–1282 cm 1 are assigned to the C–O–C stretching mode coupled with the bending mode of C–H. Corresponding experimental bands appear at 1253 cm 1 and 1224 cm 1. These two vibrations are not so sensitive to the change in hydration number. However, the calculated


View Article Online

Paper

PCCP

Table 1 The assignments for calculated IR absorption bands (in cm 1) of M2 + nH2O (n = 0, 3, 4, 6) and experimental IR absorption frequencies of SPEEK in the 800–2000 cm 1 region

M2

M2 + 3H2O

M2 + 4H2O

M2 + 6H2O

Assignment

Exp.a

Exp.

1706 — 1198 — 1237, 1219 1376 1164

1700 1660, 1699 1462 — 1284, 1242 1279 1125

1698 1659, 1686 — 1439, 1734, 1778 1280, 1244 1195 1114

1698 1645, 1666 — 1427, 1713,1769 1282, 1242 1193 1125

ns(CQO) d(Ow–H) d(Os–H) + d(C–H) d(H3O+/H5O2+) d(C–H) + ns(C–O–C) nas(OQSQO) + d(Os–H) ns(OQSQO) + d(Os–H)

1650 1642 — — 1224, 1253 1162 1080

1651b, 1653c — — 1711d — 1274, 1079, 1023e 1255, 1080, 1020b


a

This work.

b

Ref. 17. c Ref. 18.

d

Ref. 47. e Ref. 23.

OQSQO and Os–H vibrations show a strong dependence on the hydration number. We assign features at 1193 cm 1 and 1125 cm 1 to the asymmetric and symmetric OQSQO stretching modes, respectively, for M2 + 6H2O, which were observed at 1274 cm 1, 1079 cm 1 and 1023 cm 1 by Ye’s group23 and at 1255 cm 1, 1080 cm 1 and 1020 cm 1 by Xing’s group.17 In our experimental measurement, the corresponding vibrations are observed at 1162 cm 1 and 1080 cm 1. In addition, the bending bands of water are found in the 1600–1700 cm 1 region in our calculations and a corresponding experimental peak is observed at around 1642 cm 1. The bending mode of Os Hp, before the proton dissociation, is assigned at 1198 cm 1 for M2, 1396 cm 1 for M2 + 2H2O, and 1462 cm 1 for M2 + 3H2O. In contrast, after proton dissociation, the wagging of H3O+/H5O2+ appears at 1439 cm 1 for M2 + 4H2O and at 1427 cm 1 for M2 + 6H2O. The features at 1734 cm 1, 1778 cm 1 for M2 + 4H2O and 1713 cm 1, 1769 cm 1 for M2 + 6H2O are assigned to the scissoring of H–O–H in H3O+/H5O2+, which was observed at 1711 cm 1 by Zanderighi’s group47 in their FTIR study of Nafion. 3.2.2 IR spectrum in the 2000–4000 cm 1 region. Vibrational features of protonated water clusters have been extensively discussed in the past because of the ubiquity of aqueous acids in chemical and biological systems.48–53 In our study, bands observed in the 2000–4000 cm 1 region mainly consist of the O–H stretching vibrations. While a broad band is obtained from experiment in this region, the spectra from static electronic structure calculations show several discrete peaks. This is because a bulk system is studied in experiments, while only a cluster is considered in simulations. As shown in Fig. 6, hydrated M1 and M2 have similar absorption peaks in the 2000–4000 cm 1 region, which suggests that the size of the model has little effect on the spectra in this region. The assignments for the obtained bands of M2 and its relatives in the 2000–4000 cm 1 region are illustrated in Table 2, which shows that it mainly consists of three types of vibrations with different

Table 2

wavenumber ranges. To understand them intuitively, we have classified and named them: the stretching vibrations of Ow–Hdang (dangling hydrogen atoms in water molecules), the stretching vibrations of Ow–Hb (H-bonded hydrogen atoms in water molecules) and the stretching vibrations of the hydronium ion. As shown in Table 2, the stretching vibrations of Ow–Hdang mostly occupy the higher wavenumber region, around 3800 cm 1. In contrast, features at approximately 3300–3600 cm 1 are contributed by the stretching modes of Ow–Hb. The weak bands in the 3170–3240 cm 1 region are assigned to the stretching vibrations of C–H in methyl and phenyl groups. In addition, the Os Hp stretching vibration was observed at 3723 cm 1 for M2, 3192 cm 1 for M2 + 1H2O, 2870 cm 1 for M2 + 2H2O and 2448 cm 1 for M2 + 3H2O. This mode shows a large red shift with an increase in hydration number for the undissociated systems. For the dissociated systems, the stretching vibration of the hydronium ion is observed in the lower wavenumber region between 2000 and 3000 cm 1 by the static electronic structure calculations, while no corresponding peak was found in our experimental measurement. This issue will be discussed in more detail below. In the previous theoretical and experimental studies on the IR spectra of water clusters, several groups have noticed and discussed the bands of O–H stretching vibrations in the 2000–3000 cm 1 region. For instance, Iyengar et al.50,51 performed a theoretical study on a protonated 21-water cluster and emphasized that the instantaneous configuration of the water cluster, in the presence of Zundel and Eigen complexes, plays an important role in those bands between 2000 and 3000 cm 1. Park et al.49 observed the bands at 2771/2783 cm 1 in Car–Parrinello MD simulations of H3O+(H2O)3 at 50–150 K. In Headrick et al.’s report,48 the band at 2665 cm 1 was observed in the IR spectra measurement of H+(H2O)4 at low temperature. According to our assignments shown in Table 2, the vibrational bands observed between 2000 and 3000 cm 1 are dominated by the stretching

The assignments for the calculated IR absorption bands (in cm 1) of M2 + nH2O (n = 0, 1, 3, 4, 6) in the 2000–4000 cm

1

region

M2

M2 + 1H2O

M2 + 3H2O

M2 + 4H2O

M2 + 6H2O

Assignment

3723 3228–3190 — — —

3192 3237–3170 3878, 3703 — —

2448 3239–3170 3882–3843 3598–3292 2448

— 3231–3169 3877–3874 3452–3206 2921, 2667, 2396

— 3226–3169 3883–3744 3558–3349 3057, 2801, 1983

n(Os–H) n(C–H) n(Ow–Hdang) n(Ow–Hb) n(H3O+/H5O2+)



View Article Online

PCCP

Paper

Table 3 The relative energies, the Hp Od and Hp Oa distances and the corresponding stretching vibrations for different structures of M2 + nH2O (n = 4–6) calculated at the B3LYP/6-311++G(2d,2p) level of theory

System


M2-4w M2-4w M2-4w M2-4w

Erela/kJ mol

1

r(Hp Od)/Å r(Hp Oa)/Å Frequency/cm

(S1) 43.23 (S2) 34.94 (S3) 0.00 (S4) 51.86

1.010 1.029 1.042 1.046

1.601 1.532 1.472 1.474

2953 2613 2396 2349

M2-5w (S1) 25.82 M2-5w (S2) 0.00

1.012 1.036

1.600 1.521

2887 2485

M2-6w (S1) 30.01 M2-6w (S2) 0.00 M2-6w (S3) 20.03

1.039 1.060 1.073

1.477 1.443 1.415

2403 2160 1983

1

vibrations of the hydronium ion or Hp–Os. The main differences between the computational and experimental conditions are twofold. (a) The target system: a stationary cluster of a model molecule and water molecules is used in the calculation, while a mobile bulk system of the polymer and bulk water is used in the experimental measurement; and (b) the operating temperature: 0 K for the calculation, 296 K for the experiment. In the following, we study the influence of these differences on the wavenumber of the hydronium ion or Os–Hp stretching. First, we investigate the relationship between structural parameters and vibrational frequencies for some of the structures obtained by geometry optimization. Selected distances for Hp–Od (oxygen as a proton donor) and Hp–Oa (the closest oxygen as a proton acceptor) are listed in Table 3, together with the corresponding relative energies of these structures and the wavenumber of the hydronium ion stretching vibration. The results indicate that the vibrational frequencies are very sensitive to the distances Hp–Od and Hp–Oa. This phenomenon had been considered as an important feature of hydrogen bonding, but still attracts many arguments as to whether a red shift or a blue shift of the H–X (X = Od or Oa) stretching frequency is caused by an elongation of the H–X bond.54,55 Here, we take M2 + 4H2O as an example to study the relationship between the distances Hp–Od, Hp–Oa and the wavenumber of the hydronium ion stretching vibrations. For the system of M2 + 4H2O, four local minima, M2-4w (S1), M2-4w (S2), M2-4w (S3) and M2-4w (S4), have been obtained. M2-4w (S3), shown in Fig. 3, is the most stable configuration with the proton dissociated, while for the three other configurations, the protons are all undissociated. As shown in Table 3, the distance Hp–Oa is 1.601 Å, 1.532 Å, 1.472 Å and 1.474 Å for S1, S2, S3 and S4, respectively, which is decreasing except for a minor variation for S3 and S4. In contrast, the bond length of Hp–Od is increasing as follows: 1.010 Å, 1.029 Å, 1.042 Å and 1.046 Å for S1, S2, S3 and S4, respectively. The corresponding wavenumbers for the hydronium ion stretching vibrations are 2953 cm 1, 2613 cm 1, 2396 cm 1 and 2349 cm 1, respectively. These results suggest that the shorter distance of Hp–Oa and longer bond length of Hp–Od lead to a lower wavenumber of the hydronium ion stretching vibration, irrespective of the status of the energy and dissociation. The same tendency is observed for


the other two systems, M2 + 5H2O and M2 + 6H2O. Consequently, our results support a red shift of the H–X stretching frequency caused by elongation of the H–X bond. Moreover, quantitatively, the wavenumber of the stretching vibration of Hp–Od shows a nearly linear dependence on its distance. When the distance of Hp–Od is about 1.01 Å, its stretching is at about 3000 cm 1, while for distances of more than 1.03 Å, the band would be at approximately 2600 cm 1. If the distance is more than 1.06 Å, the band would be at around 2100 cm 1. These O–H stretching vibrations in this wavenumber region can only be observed in the IR spectra of a stationary cluster or measured at very low temperature, which could produce a variety of local minima with different O–H distances. In contrast, the bulk system used in experiments involves a huge number of molecules at room temperature, and thus the molecular vibrations are averaged out. Therefore, the temperature effect might be another reason that the hydronium ion vibrational bands in the 2000–3000 cm 1 region are not observed in experimental measurements. To estimate the influence of finite temperature effects on the vibrational spectra, we have calculated power spectra from the 6 ps trajectory of ADMP simulations mentioned in Section 2.2. The power spectra were computed from the Fourier transform of the velocity autocorrelation functions. The obtained spectra for M1 + 4H2O are shown in Fig. 7. We notice that no clear peak is observed in the 2000–3000 cm 1 region, in agreement with our experimental measurement. This result is also consistent with Iyengar’s report that the hydronium ion stretching in the 2000–3000 cm 1 region gets spread out during the finite temperature dynamical simulations.50,51 In addition, we have decomposed the spectra into contributions from different hydrogen species: HC (hydrogen of phenyl and methyl groups) and HW + HP (hydrogen of water and proton). In the 1600–1800 cm 1 region and the 3300–3800 cm 1 region, the vibrations mainly consist of the bending and stretching modes of O–H, respectively. Around the 2900–3200 cm 1 region, the vibrations are dominated by the stretching vibrations of C–H. These results are consistent with our assignments determined by static electronic structure

Fig. 7 The power spectra calculated at the B3LYP/6-31+G* level of theory employing ADMP simulations. HC, HW and HP represent hydrogen atoms in phenyl and methyl groups, water and proton, respectively.


View Article Online

Paper

calculations. Therefore, a combination of a quantum chemical method and an ab initio MD method is useful for studying the proton dissociation, spectroscopic properties or other properties of materials, regarding the accuracy and efficiency of calculations.


4 Conclusions We have presented herein a theoretical study on proton dissociation and IR spectra of SPEEK. A combination of two theoretical methods, static electronic structure calculation and ab initio MD (ADMP) simulation, has been used during the investigation. The calculations show that the hydration number of four for sulfonate is the threshold for proton dissociation for SPEEK. When hydrated by four waters, the proton is always readily detached, while it is more difficult when hydrated by only three waters. If we increase the hydration number to six, the probability of proton dissociation is increased to 75%. SPEEK is observed to possess a comparable proton dissociation property with other hydrocarbon membranes, such as SPES. In addition, the size of the model molecules is found to have minor effects on proton dissociation and strong effects on the spectroscopic properties. We also made assignments for the characteristic bands of the main chain of the SPEEK polymer and the water vibrations. Three types of water vibrations have been identified: Ow–Hdang, Ow–Hb and hydronium ion stretching, which would be helpful for understanding the state of water surrounding the polymer. While the hydronium ion stretching bands in the 2000–3000 cm 1 region are obtained from static electronic structure calculations, these bands are not clearly observed in the ab initio MD simulation and experimental measurements. This discrepancy is attributed to two factors: (a) stationary structures of the target system and (b) operating temperatures. In this paper, we focused on the status of water surrounding the sulfonate group of SPEEK and only one sulfonate group was considered for each system. The effect of the sulfonate aggregation and the effect of water molecules around hydrophobic or other hydrophilic groups on the proton conductivity and vibrational spectra deserve a further study.

Acknowledgements We gratefully acknowledge financial support through the New Energy and Industrial Technology Development Organization (NEDO). Y-Y. Z. thanks Dr Makoto Yamaguchi for helpful discussions.

References 1 K. V. Kordesch and G. R. Simader, Chem. Rev., 1995, 95, 191–207. 2 B. C. H. Steele and A. Heinzel, Nature, 2001, 414, 345–352. 3 M. Winter and R. J. Brodd, Chem. Rev., 2004, 104, 4245–4270.


PCCP

4 K.-D. Kreuer, S. J. Paddison, E. Spohr and M. Schuster, Chem. Rev., 2004, 104, 4637–4678. 5 K. A. Mauritz and R. B. Moore, Chem. Rev., 2004, 104, 4535–4585. 6 J. A. Elliott and S. J. Paddison, Phys. Chem. Chem. Phys., 2007, 9, 2602–2618. 7 Y.-K. Choe, E. Tsuchida, T. Ikeshoji, S. Yamakawa and S. Hyodo, Phys. Chem. Chem. Phys., 2009, 11, 3892–3899. 8 M. Danilczuk, L. Lin, S. Schlick, S. J. Hamrock and M. S. Schaberg, J. Power Sources, 2011, 196, 8216–8224. 9 M. A. Hickner, H. Ghassemi, Y. S. Kim, B. R. Einsla and J. E. McGrath, Chem. Rev., 2004, 104, 4587–4612. 10 K. Yamazaki, G. Wang, M. Tanaka and H. Kawakami, J. Power Sources, 2012, 216, 387–394. 11 P. Kiatkittikul, T. Nohira and R. Hagiwara, J. Power Sources, 2012, 220, 10–14. 12 M. Rikukawa and K. Sanui, Prog. Polym. Sci., 2000, 25, 1463–1502. 13 S. J. Paddison, J. New Mater. Electrochem. Syst., 2001, 4, 197–207. ¨rissen, V. Gogel, J. Kerres and J. Garche, J. Power 14 L. Jo Sources, 2002, 105, 267–273. 15 Y.-K. Choe, E. Tsuchida, T. Ikeshoji, A. Ohira and K. Kidena, J. Phys. Chem. B, 2010, 114, 2411–2421. 16 J. Devaux, D. Delimoy, D. Daoust, R. Legras, J. P. Mercier, C. Strazielle and E. Nield, Polymer, 1985, 26, 1994–2000. 17 P. Xing, G. P. Robertson, M. D. Guiver, S. D. Mikhailenko, K. Wang and S. Kaliaguine, J. Membr. Sci., 2004, 229, 95–106. 18 L. Grosmaire, S. Castagnoni, P. Huguet, P. Sistat, ´bin and S. Deabate, Phys. M. Boucher, P. Bouchard, P. Be Chem. Chem. Phys., 2008, 10, 1577–1583. 19 V. D. Noto, M. Piga, G. A. Giffin and G. Pace, J. Membr. Sci., 2012, 390–391, 58–67. 20 N. Takimoto, L. Wu, A. Ohira, Y. Takeoka and M. Rikukawa, Polymer, 2009, 50, 534–540. 21 Q. Luo, H. Zhang, J. Chen, D. You, C. Sun and Y. Zhang, J. Membr. Sci., 2008, 325, 553–558. 22 M. L. D. Vona, A. D’Epifanio, D. Marani, M. Trombetta, E. Traversa and S. Licoccia, J. Membr. Sci., 2006, 279, 186–191. 23 Y. Ye, Y.-C. Yen, C.-C. Cheng, W.-Y. Chen, L.-T. Tsai and F.-C. Chang, Polymer, 2009, 50, 3196–3203. 24 J. Zhu, B. S. Shim, M. D. Prima and N. A. Kotov, J. Am. Chem. Soc., 2011, 133, 7450–7460. 25 B. Liu, G. P. Robertson, D.-S. Kim, M. D. Guiver, W. Hu and Z. Jiang, Macromolecules, 2007, 40, 1934–1944. 26 J. Wang, X. Yue, Z. Zhang, Z. Yang, Y. Li, H. Zhang, X. Yang, H. Wu and Z. Jiang, Adv. Funct. Mater., 2012, 22, 4539–4546. 27 A. Brunetti, E. Fontananova, A. Donnadio, M. Casciola, M. D. Vona, E. Sgreccia, E. Drioli and G. Barbieri, J. Power Sources, 2012, 205, 222–230. 28 G. Bahlakeh, M. Nikazar, M.-J. Hafezi, E. Dashtimoghadam and M. M. Hasani-Sadrabadi, Int. J. Hydrogen Energy, 2012, 27, 10256–10264. 29 N. Idupulapati, R. Devanathan and M. Dupuis, J. Phys. Chem. A, 2010, 114, 6904–6912. 30 M. Eikerling, S. J. Paddison and T. A. Zawodzinski Jr., J. New Mater. Electrochem. System, 2002, 5, 15–23.


View Article Online


PCCP

31 X. Wu, X. Wang, G. He and J. Benziger, J. Polym. Sci., Part B: Polym. Phys., 2011, 49, 1437–1445. 32 J. K. Clark, S. J. Paddison, M. Eikerling, M. Dupuis and J. T. A. Zawodzinski, J. Phys. Chem. A, 2012, 116, 1801–1803. 33 D. Marx, ChemPhysChem, 2006, 7, 1848–1870. 34 S. J. Paddison and J. T. A. Zawodzinski, Solid State Ionics, 1998, 113–115, 333–340. 35 S. J. Paddison and J. A. Elliott, J. Phys. Chem. A, 2005, 109, 7583–7593. 36 S. J. Paddison and J. A. Elliott, Phys. Chem. Chem. Phys., 2006, 8, 2193–2203. 37 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785–789. 38 A. D. Becke, J. Chem. Phys., 1993, 98, 1372–1377. 39 S. F. Boys and F. Bernardi, Mol. Phys., 1970, 19, 553–566. 40 S. S. Iyengar, H. B. Schlegel, J. M. Millam, G. A. Voth, G. E. Scuseria and M. J. Frisch, J. Chem. Phys., 2001, 115, 10291–10302. 41 H. B. Schlegel, J. M. Millam, S. S. Iyengar, G. A. Voth, G. E. Scuseria, A. D. Daniels and M. J. Frisch, J. Chem. Phys., 2001, 114, 9758–9763. 42 H. B. Schlegel, S. S. Iyengar, X. Li, J. M. Millam, G. A. Voth, G. E. Scuseria and M. J. Frisch, J. Chem. Phys., 2002, 117, 8694–8704. 43 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,


Paper

44 45 46 47 48

49 50 51 52

53 54 55

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, ¨ . Farkas, J. B. Foresman, S. Dapprich, A. D. Daniels, O J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian 09 Revision A.1, Gaussian Inc., Wallingford, CT, 2009. G. Schaftenaar and J. H. Noordik, J. Comput. Aided Mol. Des., 2000, 14, 123–134. W. Humphrey, A. Dalke and K. Schultenk, J. Mol. Graphics, 1996, 14, 33–38. T. Kobayashi, M. Rikukawa, K. Sanui and N. Ogata, Solid State Ionics, 1998, 106, 219–225. M. Laporta, M. Pegoraro and L. Zanderighi, Phys. Chem. Chem. Phys., 1999, 1, 4619–4628. J. M. Headrick, E. G. Diken, R. S. Walters, N. I. Hammer, R. A. Christie, J. Cui, E. M. Myshakin, M. A. Duncan, M. A. Johnson and K. D. Jordan, Science, 2005, 308, 1765–1769. M. Park, I. Shin, N. J. Singh and K. S. Kim, J. Phys. Chem. A, 2007, 111, 10692–10702. S. S. Iyengar, M. K. Petersen, T. J. F. Day, C. J. Burnham, V. E. Teige and G. A. Voth, J. Chem. Phys., 2005, 123, 084309. S. S. Iyengar, J. Chem. Phys., 2007, 126, 216101. J.-W. Shin, N. I. Hammer, E. G. Diken, M. A. Johnson, R. S. Walters, T. D. Jaeger, M. A. Duncan, R. A. Christie and K. D. Jordan, Science, 2004, 304, 1137–1140. M. Miyazaki, A. Fujii, T. Ebata and N. Mikami, Science, 2004, 304, 1134–1137. A. Allerhand and P. v. R. Schleyer, J. Am. Chem. Soc., 1963, 85, 1715–1723. S. A. C. McDowell and A. D. Buckingham, J. Am. Chem. Soc., 2005, 127, 15515–15520.


Air Redox Flow Battery.

Degradation mechanism of sulfonated poly(ether ether ketone) (SPEEK) ion exchange membranes under vanadium flow battery medium.

A new ab initio potential energy surface and infrared spectra for the Ar-CS₂ complex.

In vitro evaluation of bioactivity of chemically deposited hydroxyapatite on polyether ether ketone.

Ab initio calculation of the neutron-proton mass difference.

Ab initio molecular crystal structures, spectra, and phase diagrams.

N2 dissociation on W(110): An ab initio molecular dynamics study on the effect of phonons.

polyacrylonitrile acid-base blend membrane for vanadium redox flow battery application.

Anion recognition by uranyl-salophen derivatives as probed by infrared multiple photon dissociation spectroscopy and ab initio modeling.

Enhanced response of microbial fuel cell using sulfonated poly ether ether ketone membrane as a biochemical oxygen demand sensor.

Vibronic bandshape of the absorption spectra of dibenzoylmethanatoboron difluoride derivatives: analysis based on ab initio calculations.

Craniofacial reconstruction using patient-specific implants polyether ether ketone with computer-assisted planning.

Shifting from hydrogen bond network to π-π stacking: a key mechanism for reversible thermochromic sulfonated poly(ether ether ketone).

High-strength, surface-porous polyether-ether-ketone for load-bearing orthopedic implants.

Ab initio molecular orbital and density functional studies on the ring-opening reaction of oxetene.

FT-IR and FT-Raman spectra, normal coordinate analysis and ab initio computations of Trimesic acid.

Computation and interpretation of vibrational spectra on the structure of Losartan using ab initio and Density Functional methods.

Ab initio multiple cloning simulations of pyrrole photodissociation: TKER spectra and velocity map imaging.

Ab initio quantum mechanical simulations confirm the formation of all postulated species in ionic dissociation.

Ab initio RNA folding.

Resolving the anomalous infrared spectrum of the MeCN-HCl molecular cluster using ab initio molecular dynamics.

Infrared signature of micro-hydration in the organophosphate Sarin: an ab initio study.

Ion mobility spectrometry, infrared dissociation spectroscopy, and ab initio computations toward structural characterization of the deprotonated leucine-enkephalin peptide anion in the gas phase.

On-the-fly ab initio semiclassical dynamics: identifying degrees of freedom essential for emission spectra of oligothiophenes.