Direct ESR detection and spin trapping of radicals generated by reaction of oxygen radicals with sulfonated poly(ether ether ketone) (SPEEK) membranes

Polymer Degradation and Stability 94 (2009) 1779–1787

Contents lists available at ScienceDirect

Polymer Degradation and Stability journal homepage: www.elsevier.com/locate/polydegstab

Direct ESR detection and spin trapping of radicals generated by reaction of oxygen radicals with sulfonated poly(ether ether ketone) (SPEEK) membranes Mariana Pinteala a, b, *, Shulamith Schlick a a b

Department of Chemistry and Biochemistry, University of Detroit Mercy, 4001 West McNichols, Detroit, MI 48221, USA ‘‘Petru Poni’’ Institute of Macromolecular Chemistry, Aleea Grigore Ghica Voda 41A, 700487 Iasi, Romania

a r t i c l e i n f o

a b s t r a c t

Article history: Received 12 December 2008 Received in revised form 10 June 2009 Accepted 12 June 2009 Available online 21 June 2009

The stability of membranes under the strong oxidizing conditions in fuel cells is one of the major challenges in the development of fuel cells based on proton exchange membranes (PEMs). This study is centered on the determination of the susceptibility to degradation of SPEEK membranes exposed to OH radicals, using both direct ESR and spin trapping with 5,5-dimethyl-1-pyrroline-1-oxide (DMPO). In order to achieve a complete picture on SPEEK degradation, two types of experiments were performed: 1. UV irradiation at 77 K of SPEEK membranes swollen by aqueous solutions of H2O2; 2. UV irradiation of SPEEK membranes swollen by aqueous solutions of H2O2 in the presence of DMPO as a spin trap. UV irradiation without oxygen of SPEEK at 77 K in acid or basic form in the presence of H2O2/H2O produced phenoxyl radicals as the predominant radicals detected by direct ESR or spin trapping methods. At pH 4, the oxygen radicals produced phenyl radicals as the predominant species detected by spin trapping methods. The hydroperoxyl radical, as DMPO/OOH adduct, was detected only when the DMPO/OH adduct was absent. The appearance of phenyl and phenoxyl radicals provides the evidence that OH radicals react with the aromatic ring of SPEEK or leading to the scission of its ether bridge. Ó 2009 Elsevier Ltd. All rights reserved.

Keywords: Sulfonated poly(ether ether ketone) (SPEEK) Electron spin resonance (ESR) 5,5-Dimethyl-1-pyrroline-1-oxide (DMPO) Spin trapping Free radicals

1. Introduction Sulfonated poly(aryletherketone) membranes have been developed for use in hydrogen and direct methanol fuel cells (DMFCs). The most commonly used membranes have an ether–ether–ketone sequence, EEK, as in the case of sulfonated poly(ether ether ketone) (SPEEK) membranes. Reaction of PEEK with sulphuric acid leads to sulfonation of the O-phenyl-O units (Structure 1) and to membranes with a broad range of solubilities, conductivities, mechanical properties, and degree of swelling by solvents [1–8]. The degree of sulfonation (DS) dictates the solubility of the membrane: for DS>30%, SPEEK ionomers are soluble in dimethylformamide (DMF), dimethylsulfoxide (DMSO), and N-methylpyrrolidone (NMP); solubility in methanol occurs in membranes for DS>70%; and in hot water for DS z 100% [1,3]. The advantages of using SPEEK membranes in DMFCs are reduced swelling by methanol when compared to Nafion

* Corresponding author. ‘‘Petru Poni’’ Institute of Macromolecular Chemistry, Aleea Grigore Ghica Voda 41A, 700487 Iasi, Romania. Tel.: þ40 232 217454; fax: þ40 232 211299. E-mail address: [email protected] (M. Pinteala). 0141-3910/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.polymdegradstab.2009.06.009

membranes as a reference, reduced methanol crossover, and improved electrode kinetics [1,2,4,5]. The reactions of OH radicals with aromatic compounds showed that the addition of OH radical to the phenyl ring is more favorable than that of hydrogen abstraction, giving a cyclohexadienyl type radicals due to the combined ortho- or para-activation to alkyl- (R-) or RO-substituents and the meta directing effect of SO3 group (arrows in Structure 1) [8]. In addition, the pH can have an important effect on the type of radicals. Furthermore, these radicals could be studied by ESR methods. Spin trapping is an analytical technique employed in the detection and identification of short-lived free radicals. Spin trapping involves the addition of radical to a nitrone spin trap resulting in the formation of a spin adduct, a nitroxide-based persistent radical, that can be detected using electron spin resonance (ESR) spectroscopy. The spin adduct usually yields a distinctive ESR spectrum characteristic of a particular free radical that is trapped. Electron spin resonance (ESR) methods are used in our laboratory for the study of stability and degradation processes in PEMs, because this technique is sensitive and specific for the detection of radical species [9–13]. In the laboratory, oxygen radicals can be generated by two major methods: (a) The Fenton reagent, where

1780

M. Pinteala, S. Schlick / Polymer Degradation and Stability 94 (2009) 1779–1787

2. Experimental section

SO3H O (

O

C

2.1. Materials

O )

n

Structure 1. Chemical structure of SPEEK; arrows represent the favorable positions where the oxygen radicals can attack the phenyl rings due to the combined effects of ortho-, para- activation to alkyl- (R-) or RO-substituents and the meta directing effect of SO3H function.

the major step is H2O2 þ Fe(II) / Fe(III) þ OH þ HO [14]. (b) Photolysis of aqueous solutions of H2O2, and the main expected reactions are [15]: H2 O2 /2OH

(1)



OHDH2 O2 / OOHDH2 O

(2)



OHDH2 O/O2 LDH3 OD

(3)

The formation of radical species from the attacked substrate, for example R, can occur by H-abstraction, OH þ RH / H2O þ R; the polymer-derived R radicals can initiate a degradation cascade. The presence of the oxygen radicals was confirmed in our laboratory by direct ESR, using a combination of UV irradiation and ESR measurements at low temperature, typically 77 K, followed by gradual annealing of irradiated samples for short intervals (z3 min) above 77 K. In this way ESR signals from the OH, OOH and O2 radicals have been detected [15]. We present a study of SPEEK membranes exposed to oxygen radicals generated by UV irradiation of aqueous solutions of H2O2. As will be seen below, the presence of radicals was detected by direct ESR; additional details on the type and relative intensity of the various radical species were obtained by spin trapping with 5,5-dimethyl-1-pyrroline-1-oxide (DMPO). In most cases the ESR spectra of spin adducts exhibit hyperfine splittings (hfs) from the 14N and Hb nuclei. Spin trapping is based on scavenging of short-lived radicals by spin traps and formation of more stable adducts, typically nitroxide radicals [16–21]. The DMPO as spin trap was chosen because the hfs of the DMPO adducts are sensitive to the nature of the trapped radicals R [22]; moreover, DMPO was used extensively for the detection of oxygen radicals such as OH, OOH and O2, as well as carbon-centered radicals (CCRs) [11,12,18–36]. The clear distinction between these two groups of free radical adducts is based on the large difference in the hfs of the Hb protons: typically 10–15 G for oxygen adducts, and >20 G for carbon-centered radical adducts, respectively. Spin trapping by DMPO is often complicated by the presence of two types of radicals directly derived from the spin trap: the product of DMPO oxidation, 5,5-dimethyl-2-oxopyrroline-1oxyl (DMPOX) with aN ¼ 7.2 G and aHg ¼ 4.2 G (two protons) [11,20,27,32], and the radical formed by cleavage of the N–C bond and ring opening, with aN z 15 G [28]. In the presence of more than one radical, the selectivity of the spin trap is also an important issue. In a recent paper the kinetics of trapping ethanol-derived radicals and OH radicals was compared [23]. The goal in the spin trapping experiments is to translate these splittings into an identity of a radical. The spin trapping data base is a useful resource [37]. It should be mentioned that the hfs (aN, aHb or aHg) of DMPO adducts can be slightly different for a given adduct, due to the local polarity or the conformation of spin trap adduct [22,23,37]. The goal of our experiments was to study the behavior in presence of oxygen radicals of a membrane that is known to have low chemical stability (SPEEK), and to compare it with the behavior of Nafion as a reference, which has been studied in our laboratory by exposure to the Fenton reagent, and by UV irradiation of H2O2 in the presence of Fe(II), Fe(III), and Cu(II) counterions [10,11,13].

The SPEEK membrane (2 meq sulfonic groups/g) was a gift from T. Fuller of the GM Fuel Cell Center in Honeoye Falls, NY, and was prepared by sulfonation of PEEK from ICI by Scientific Polymer Products, Ontario, NY. Solutions of SPEEK in dimethylacetamide (18% by weight) were cast on a glass plate and dried in air for 16 h and at 80  C for at least 1 h. The coated film was released from the glass plate by immersion in water, and the free-standing film was then allowed to dry in air. Prior to the ESR experiments, the membrane was dried under vacuum for 24 h at 35  C. The DMPO as the spin trap (99.97%) was purchased from Aldrich–Sigma Chemical Company and used without further purification. Dimethylsulfoxide (DMSO, 99.9%) from Aldrich– Sigma Chemical Company, the buffer solution from Orion Research, and hydrogen peroxide (3% w/v) from Fisher were used as received.

2.2. Sample preparation Samples were prepared from deionized and double distilled water (or buffer solution) and H2O2, and dried membrane pieces of typical size 5  15 mm2. The dry membrane was soaked in the ESR sample tube for 4 h with the H2O2 solution in the refrigerator to prevent H2O2 decomposition; the supernatant was then removed, and the samples were immersed in liquid nitrogen prior to UV irradiation at 77 K. For the spin trapping experiments, the DMPO solution in the appropriate solvent was added prior to or after irradiation. Fully neutralized membranes were prepared by immersion in KOH aqueous solution for 24 h, followed by drying in the ESR sample tube for 24 h at 30–35  C under vacuum. When the irradiation was performed on membrane solutions, samples consisting of SPEEK dissolved in DMSO (or DMSOd6) (1% w/v), aqueous solution of H2O2 (3% w/v), and DMPO solution in appropriate solvent were bubbled with nitrogen for 5–10 min. The concentrations

UV/Annealed/ESR K

K 220 / 77

77/

200 / 77

77/

180 / 77

77/

165 / 77

77/

155 / 77

77/

100 / 77

77/

- / 3240

3260

3280

3300

3320

K

77/

3340

77

/ 77

165 / 77 3360

Magnetic Field (G) Fig. 1. Experimental (at 77 K) ESR spectra of UV-irradiated SPEEK membrane in acid form swollen by 3% H2O2/H2O. The sample was annealed at indicated temperatures for 10 min. The spectra reveal the presence of OH and OOH radicals, as predominant radicals, until approximately 160 K. At about 160 K, the signal from the new species dominated the ESR spectrum.

M. Pinteala, S. Schlick / Polymer Degradation and Stability 94 (2009) 1779–1787

2.3. ESR measurements

SPEEK/H2O2/UV Experimental Simulated

3275

3300

3325

1781

3350

Magnetic Field / G Fig. 2. Experimental (at 77 K) and simulated ESR spectra of UV-irradiated SPEEK membrane in acid form swollen by 3% H2O2/H2O. The sample was annealed at 155 K for 10 min.

of DMPO and the H2O2/DMPO molar ratios will be presented in the Results and Discussion section. UV irradiation was performed at 77 K or ambient temperature with a low-pressure mercury source (Mineralight Model PCQX1), equipped with four Model 50 053 tubes (Ultra–Violet Products, San Gabriel, CA). The irradiation time varied in the range 5–80 min. UV irradiation at ambient temperature was accomplished by using the flat cell. The dried membrane (30  2 mm) was positioned in the cell, and the H2O2 solution and the DMPO solution in the buffer (pH ¼ 4) were added prior to irradiation. Samples consisting only of the H2O2 solution, or mixture of H2O2 and DMPO solutions, were used as controls, and irradiated at 77 K for 30 min.

Spectra were recorded using a Bruker X-band EMX spectrometer operating at 9.7 GHz with 100 kHz magnetic field modulation, and equipped with the Acquisit 32 Bit WINEPR data system version 3.01 for acquisition and manipulation and with the ER 4111 VT variable temperature units. The microwave frequency was measured with a Hewlett Packard 5350B frequency counter. The absolute value of the magnetic field was calibrated with Cr(III) in a single crystal of MgO (g ¼ 1.9796). The ESR settings for UV irradiation experiments at 300 K were as follows: sweep width 80 G, microwave power typically at 20 mW, time constant 160 ms, conversion time 320 ms, 5-8 scans, 1024–2048 points, modulation amplitude 0.5–0.8 G, and receiver gain 5.02  104. The settings for the UV irradiation experiments in the temperature range 77–290 K were as follows: sweep width 1800 or 500 G, microwave power typically at 2 mW, time constant 40.96 ms, conversion time 81.92 ms, 4–10 scans, 1024–2048 points, modulation amplitude 1–3 G, and receiver gain 1–5.02  104. A quartz dewar was used for recording ESR spectra at 77 K. Annealing at temperatures above 77 K was performed with the Bruker ER 4111 VT variable-temperature system mounted outside the microwave cavity. ESR spectra in the rigid limit were simulated using SimFonia (Bruker) with manual parameter optimization. Solution spectra were simulated with the program WinSim (NIEHS/NIH) [38]; the software provided an automatic fit to the experimental spectrum and determined the relative intensity of each component in the case of a superposition of several components. 3. Results and discussion In this section we will present and interpret results obtained in two types of direct ESR and spin trapping experiments: 1. UV

Table 1 Hyperfine splittings of radicals and DMPO adducts in the SPEEK system. System

Adduct

ESR Parameters

Hyperfine splittings, G aN

aHb

Relative intensity, % aHg



UV-irrad at 77 K of SPEEK (acid form)/H2O2/H2O, annealed at 155 K, 10 min.

OPh g values: 2.0012; 2.0074; 2.0130 giso ¼ 2.0072 AH:5,5,5 G (1H) AH:3,7,7 G (2H)

Dominant

Fig. 2 (spectra at 77 K, in water) UV-irrad at 77 K of SPEEK (acid form)/H2O2/H2O, annealed in the presence of DMPO

DMPO/OPh DMPO/OOH DMPO/Ph DMPO Degrad

13.4 13.7 15.7 13.9

8.3 12.1 22.6 1.1

1.5 0.8

60 24 6 10

DMPO/OPh DMPO/OOH DMPO/Ph DMPO Degrad

13.1 14.1 15.5 14.4

9.1 12.3 22.1

1.7 0.9

15 24 51 10

DMPO/OH DMPO/Ph DMPO/OPh DMPO Degrad

14.5 16.0 (12.7) 14.5

14.4 22.6 (7.2) 1.0

DMPO/OPh DMPO/OH DMPO/OOH DMPO/Ph DMPO Degrad

(13.8) 14.9 (14.3) 15.5 14.6

(9.2) 14.6 (12.9) 22.1 1.0

Fig. 4 (spectra at 300 K, in DMSO) UV-irrad at 77 K of SPEEK (pH 4)/H2O2/H2O, annealed in the presence of DMPO

Fig. 5 (spectra at 300 K, in DMSO) UV-irrad at 77 K of SPEEK (pH 4)/H2O2/H2O/DMPO

1.0

(1.8)

44(34)a 30(27) 0(28) 26(10)

Fig. 6B (spectra at 260 K, in water) UV-irrad at ambient temperature of SPEEK (pH 4)/H2O2/H2O/DMPO

Fig. 8 (spectra at 300 K, in water) a b

(1.6) (0.8)

0(26)b 33(0) 0(25) 17(29) 50(20)

Regular numbers are for 10 mM DMPO and represent hfs and the % intensity; numbers in brackets are for 50 mM DMPO and represent hfs and the % intensity. Regular numbers are hfs and the % intensity immediately after irradiation; numbers in brackets are hfs and the % intensity after 60 min subsequent to UV irradiation.

1782

M. Pinteala, S. Schlick / Polymer Degradation and Stability 94 (2009) 1779–1787

irradiation at 77 K of SPEEK membranes swollen by aqueous solutions of H2O2. 2. UV irradiation of SPEEK membranes swollen by aqueous solutions of H2O2 and DMPO as the spin trap.

signals of OOH radicals remained together with additional paramagnetic species. At about 160 K, the signal from the new species dominated the ESR spectrum. As the temperature was increased, a relatively narrow single line appeared and it could be observed a dramatic decrease in its intensity after 200 K. In Fig. 2 we present the ESR spectrum at 77 K after annealing for 10 min at 155 K following the UV irradiation of SPEEK in acid form, and the corresponding simulated spectrum. The magnetic parameters deduced from the simulated spectrum, Table 1, suggest the formation of phenoxyl radicals, taking into consideration data obtained for phenoxyl radical derived from p-methoxybenzensulfonic acid (proton hfs of 8.04 G and 4.57 G) and from 1-hydroxy-4-methoxybenzensulfonic acid or 2,5-dihydroxybenzensulfonic acid (proton hfs 2.64 G and 3.03 G) [8]. Additional results were obtained for o-hydroxyl-m,m’-dicarboxy-phenoxyl radicals (proton hfs 5.95 G and 3.77 G) [39]. It should be mentioned that the same result was obtained by simulation of spectrum at 200 K. We note that the simulation of the ESR spectrum in Fig. 2 is not perfect; in particular, the low field shoulder in the experimental spectrum was not reproduced, suggesting that additional species may be present.

3.1. UV irradiation at 77 K of SPEEK membrane swollen by aqueous solutions of H2O2 In this part we will describe both direct detection of ESR spectra, and spin trapping by DMPO that was added after the UV irradiation. Samples were annealed 10 min at the specific temperature above 77 K, and ESR spectra were measured at 77 K. ESR spectra at 77 K consisted of a broad feature with no evident structure. Annealing above 77 K led to better resolution and to a significant decrease of intensity above 200 K (Fig. 1). The spectra in Fig. 1 reveal the presence of OH and OOH radicals. The OH and OOH radicals were detected at 77 K following irradiation. The ESR parameters, measured directly from spectra, were in excellent agreement with the values measured in the radiolysis of pure water or the other matrix [15]. Fig. 1 showed that the OH radicals began to decay by annealing at approximately 100 K and then disappeared all together at about 120 K, while the

O H 3C H 3C

SO3H

Ph N

SPEEK

H

O

DM PO

(DMPO/Ph)

O O

HO H

6 1 5

H

4

H 2O2/h υ

SO3H

O H

3C

SO3H

1

O 5

2

4

2

O

H

H

(1) ( .Ph)

O

6

O2

3

H O

chain scission

+ O O

O

HOO H H

6 5

SO3H

1

.O 6

2 4

3C

-H2O

H

O

+ Oxidation Product (yellowing)

5

SO3H

1 2 4

H

SO3H

O

3

H

(2) ( .OPh)

H

H

6 5

1 2 4

O

O

O

3

H

(4) ( .OPh)

O H 6 5

SO3H

1

(3) ( .OPh)

2 3

4

H

H O

DMPO

OPh

H 3C H 3C

N

H

O (DMPO/OPh) Scheme 1. Proposed degradation mechanism of SPEEK in the presence of hydroxyl radicals generated by UV irradiation of hydrogen peroxide based on results presented in this study and on refs 6-8.

M. Pinteala, S. Schlick / Polymer Degradation and Stability 94 (2009) 1779–1787

Experimental Simulation specie 1 (Phenoxyl radical)

Simulation specie 2 (DMPO degradation)

Sum specie 1 (75 %) and specie 2 (25 %) 3275 3300 3325 3350 3375 3400 3425

Magnetic Field (G) Fig. 3. ESR experimental (300 K) and powder simulation spectra of radicals presented in Fig. 2 trapped by DMPO. The magnetic parameters used in the simulated spectra are: Species Hb Hg N Hb 1: gzz ¼ 2.0023 AN zz ¼ 31 G Azz ¼ 9 G Azz ¼ 1.5 G DH ¼ 7 G gyy ¼ 2.0061 Ayy ¼ 5 G Ayy ¼ 9 G Hg Hb Hg ¼ 1.5 G DH ¼ 7 G gxx ¼ 2.0088 AN Ayy xx ¼ 5 G Axx ¼ 9 G Axx ¼ 1.5 G DH ¼ 9 G giso ¼ 2.0058 b g N Hb Hg aN ~ 13.6 G aH H ¼ 9 G aH ¼ 1.5 G Species 2: gzz ¼ 2.0023 Azz ¼ 32 G Azz ¼ 1.08 G Azz ¼ 0.53 G N Hb Hg DH ¼ 7 G gyy ¼ 2.0061 Ayy ¼ 5 G Ayy ¼ 1.08 G Ayy ¼ 0.53 G DH ¼ 7 G gxx ¼ 2.0088 AN xx ¼ 5 G b b g Hg AH xx ¼ 1.08 G Axx ¼ 0.53 G DH ¼ 9 G giso ¼ 2.0058 aN ~ 14 G aH ¼ 1.08 G aH ¼ 0.53 G.

As we mentioned, the addition of oxygen radicals to the phenyl ring is more favourable than hydrogen abstraction, and leads to the formation of cyclohexadienyl type radicals [6–8,13,39–41], due to the combined effects of ortho-, or para-activation to alkyl- (R-) or RO-substituents and the meta directing effect of SO3H function (arrows, Structure 1). Finally, the resulting O2 (eq. (3)) can produce phenoxyl radicals (OPh) a process that is strongly pH dependent [6–8,13]. All these transformations for the species with

–SO3H substituted aromatic rings can give scission of membrane chain [8]; alternately, complete degradation of the aromatic ring can be achieved [7], as seen in Scheme 1. We note that in Scheme 1 there are three possible structures for phenoxyl radicals, and we cannot specify which structure is reflected in the ESR spectrum shown in Fig. 2. It was assumed that only one principal radical species was present, with any other species being too low in concentration to distort the simulation. The principal values of the g tensor are characteristic of oxygen – centered radical (OCR) (Table 1), and the coupling constants were in agreement with the literature and the DFT calculations [15]. In Fig. 2, it may be observed, that the rendering of broad line of experimental spectrum made difficult to obtain a best fit of simulated spectrum. This broad line is due to species present in a small concentration. The structures assumed from the simulated spectra at about 160 K are of the phenoxyl radicals type [8,15] (2,3,4 structures from Scheme 1); the same result was obtained by simulation of spectrum at 200 K. In order to clarify the nature of radicals present at and above 155 K, an aqueous solution of 100 mM DMPO was added to the SPEEK sample and the temperature was slowly increased to ambient temperature during 30 min. The supernatant was then removed and the membrane was dried in vacuum for 1 h at 25  C. From simulated powder spectrum of DMPO adducts (Fig. 3) it may be deduced that the experimental spectrum could be interpreted by assuming the superposition of at least two sets of powder nitroxyl radicals according with Refs. [25] and [37]. According to these references, the principal values of the g tensor and the coupling constants of simulated species (Fig. 3) are characteristics for DMPO/oxygen–carbon-centered radical (OCR) and DMPO degradation, with the contribution to the simulated spectrum of 75% and 25%, respectively. The broad line in the experimental spectrum may be due to the presence of small amounts of the other species. The dry membrane was dissolved in DMSO, which is a solvent for the membrane; the resulting ESR spectrum measured at 300 K is presented in Fig. 4. The experimental spectrum was fitted by a superposition of signals from three DMPO adducts and a radical product of DMPO degradation. The radicals were identified on the basis of their hyperfine coupling constants and the ratio of the 14N and Hb splittings, aN/aHb, Table 1. One adduct was attributed to the DMPO adduct of the phenoxyl radical, DMPO/OPh (60% of the total intensity) and was

Experimental

Exp Sim DMPO/OPh

1783

Simulated

60 % DMPO/Ph

DMPO/OH

24 %

DMPO degr.

10 %

DMPO/Ph

6%

3330

3340

3350

3360

3370

3380

3390

3400

24%

DMPO/OH DMPO/OPh

15%

DMPO degr.

10%

3310 3320

51%

3320

3330

3340

3350

3360

3370

3380

3390

3400

Magnetic Field (G)

Magnetic Field (G) Fig. 4. Experimental and simulated ESR spectra of DMPO adducts from Fig. 3 as DMSO solution at 300 K.

Fig. 5. Experimental and simulated ESR spectra of DMPO adducts obtained after UV irradiation of SPEEK at 77 K in the presence of H2O2 in a buffer solution (pH ¼ 4). ESR experimental spectrum at 300 K was acquired after dissolving the DMPO adducts in DMSO.

1784

M. Pinteala, S. Schlick / Polymer Degradation and Stability 94 (2009) 1779–1787

assigned on the basis of its hfs and aN/aH ratio [37]. In addition to this adduct, the DMPO/OOH [11,19,22,23,42] and DMPO/Ph (Ph ¼ phenyl radical) [26,33,34,37] adducts, as well as the product of DMPO degradation [11,20,27,32], are present, as seen in Table 1. Similar results were obtained for SPEEK with 100% degree of neutralization, but the presence of the DMPO/Ph was not detected. In conclusion, for membranes in the acid form and for fully neutralized membranes, both direct ESR spectra and the spin adducts are dominated by the phenoxyl radical [6–8,13], whose formation is described in Scheme 1. The same adducts were also obtained in SPEEK membranes at pH 4, but their relative intensities are different (ex.: from spin trapp results there are DMPO/C(Ph) [26,33,34] as the main product and DMPO/OPh [19], DMPO/OH [11,22,23,37] and DMPO degradation [32] as secondary products), as shown in Fig. 5 and Table 1. Similar pH dependences were also reported in model compounds [8]. The appearance of DMPO/OH adduct instead DMPO/OOH requires an explanation: the DMPO in the presence of OOH radical produces DMPO/OOH adduct which is short-lived adduct and is reduced to DMPO/OH adduct [27].

SPEEK/H2O2/DMPO/UV

A

Phenyl adduct

260 K 260 K

3.2. UV irradiation of SPEEK/H2O2/H2O in presence of DMPO (pH 4) 3.2.1. UV irradiation at 77 K The temperature variation of ESR spectra obtained after UV irradiation at 77 K is presented in Fig. 6A; black and red tracings represent results in the presence and absence of SPEEK, respectively. At 100 K the signals are broad, but signals from OH radicals appear and are indicated by arrows. At 160 K the doublet characteristic of  OOH radicals is shown, due to the proton splitting of z8 G. We notice the stabilization of this radical in the sample that contained the membrane, a result that is in agreement with those obtained in Nafion exposed to the Fenton reagent [11]. The ESR spectra of samples with and without SPEEK are different at 230 or 240 K. The 1:2:2:1 quartet representing the DMPO/OH adduct is clearly seen in the absence of the membrane at 240 K. At 230 K in the presence of SPEEK, an ESR spectrum typical for a nitroxide radical near the rigid limit (extreme separation of 66 G) is clearly seen. ESR spectra at 260 K are better resolved. In the presence of the membrane the additional DMPO/Ph adducts are indicated by arrows. In Fig. 6B we present the ESR spectrum at 260 K and the corresponding simulation for the sample containing the SPEEK membrane. The experimental spectrum was simulated by a superposition of signals from three adducts that were identified on the basis of their hfs (Table 1). The six-line ESR spectrum characterized by aN ¼ 16.0 G, aHb ¼ 22.6 G was assigned to the DMPO/Ph adduct [26,33,34,37]. The quartet represents the DMPO/OH adduct [11,22,23,42], and the triplet is a nitroxide formed by DMPO degradation [11,20,27,32]. These results can be interpreted in terms

230 K 240 K Exp 4

.OOH 160 K 160 K

Sim 4

100 K .OH

100 K 3275

3300

3325

3350

3375

3400

Exp 3

3425

Magnetic Field / G Sim 3

SPEEK/H2O2/DMPO/UV

B

Exp Exp 2

Sim

Adducts

Sim 2 44%

DMPO/OH 30% Exp 1

DMPO/Ph 26% DMPO Degrad 3330

Sim 1 3340

3350

3360

3370

3380

3390

Magnetic Field / G Fig. 6. A. ESR spectra as a function of temperature obtained after 30 min of UV irradiation at 77 K of H2O2 at pH 4 (red traces) as reference, and of the same sample after addition of SPEEK (black traces). Both experiments were carried out in the presence of DMPO (10 mM) as the spin trap; molar ratio H2O2]/[DMPO] ¼ 1/10. Fig. 6B. ESR experimental and simulated spectra at 260 K of DMPO adducts obtained after UV irradiation of SPEEK from Fig. 6A.

3330

3340

3350

3360

3370

3380

3390

Magnetic Field (G) Fig. 7. ESR spectra at 260 K obtained after 30 min of UV irradiation at 77 K of SPEEK containing the following DMPO concentrations: (1) 0.870 mM; (2) 10 mM; (3) 50 mM; (4) 180 mM; the molar ratio [H2O2]/[DMPO] ¼ 1/10 in pH ¼ 4 buffer solution.

M. Pinteala, S. Schlick / Polymer Degradation and Stability 94 (2009) 1779–1787

1785

Table 2 The ESR splitting constants, relative intensities, and the DMPO spin adducts deduced from the simulated spectra shown in Fig. 7. Spectrum

Adduct/splitting constants (G)/contribution of adduct (%) DMPO/C(Ph)

DMPO/OH

DMPO degradation

DMPO/OR

aN

ab

ag

aN

ab

ag

aN

ab

ag

aN

ab

ag

1

16.00

23.00

–

14.90

14.50

–

14.89

1.01 1.00

–

–

–

–

2

15.97

22.55

–

14.61

14.30

–

14.50

1.01 1.01

–

–

25.0%

50.0%

44.0% 3

15.69

4

15.97

25.0%

44.0% 23.21

–

14.86

23.00

–

14.61

26.8%

30.0% 14.44

–

15.24

14.30

–

16.16

34.7%

11.0%

–

– 0.74 1.17

–

12.70

0.35 0.35

–

12.7

10.0%

40.0%

3.2.1.1. Competition of DMPO. In the spin trapping experiments we have noticed that the relative intensities of the various adduct are sensitive to the DMPO concentration in the range 0.87–50 mM. When the concentration of DMPO is high (>50 mM) or low (<0.87 mM), the signal from the DMPO/Ph adduct (DMPO/carbon– centered radical (CCR) adduct) is either weak or absent [36] (Fig. 7, Table 2). This result indicates that when DMPO concentration is low the initially formed radicals (OH) will consume all DMPO while the new radicals formed are not trapped. On the other hand, at high concentration of DMPO, where the trapping of OH radical predominates, oxygen evolution and the DMPO/CCR adduct cannot be detected. This is consistent with the fact that the OOH radical and the CCRs cannot be formed when most of the OH radicals are trapped before they attack the aromatic ring from membrane (Spectrum 4, Fig. 7) [18] or DMPO adduct by himself can react with carbon-centered radicals to yield neutral species (Scheme 2) [43]. It should be underlined that by increasing the amount of DMPO from 0.87 to 50 mM, the contribution of DMPO/CCR (DMPO/Ph) adduct was lowered and by contrast, the contribution of DMPO/OR and DMPO/OOH adducts was greater than before. 3.2.2. UV irradiation at ambient temperature in the presence of DMPO These experiments were possible by positioning the swollen membrane in the flat cell. Immediately after UV irradiation, signals from DMPO degradation and the DMPO/OH adduct were detected. ESR spectra obtained at 60 min after UV irradiation, shown in Fig. 8, indicate the presence of signals corresponding to DMPO/Ph, DMPO/ OOH and DMPO/OPh adducts. The relative intensity of the DMPO/

7.23

1.79

7.28

1.74

28.5%

40.0%

of the mechanism presented in Scheme 1. The DMPO/OOH adduct, which was expected for pH lower than 7.7 [11,19,22,42], is not present at this temperature.

–

9.0%

OH adduct was 33% immediately after irradiation and decreased to z0% after after 60 min subsequent to UV irradiation (Table 1). UV irradiation using the flat cell at room temperature for 30 min of SPEEK/H2O2/H2O at pH 4 yielded different signals in different ratios as a function of DMPO concentration: (1) for low DMPO concentration (0.8 mM), the main adduct was DMPO/OH; (2) at higher DMPO concentration (10 mM), the DMPO/Ph and DMPO/ OPh adducts appeared; (3) for irradiation times between 0 and 60 min, signals from both DMPO/OH and DMPO/OOH were detected. In addition, the assignment of the DMPO/OOH adducts instead DMPO/O2 requires an explanation. In much of the spin trapping literature these two adducts are giving the same parameters, as seen, for example, in ref 44. The rate constants for the reaction of DMPO with OOH and with O2 radicals were determined by fitting the variation of Kapp with pH; Kapp is the apparent constant for the reaction of DMPO with OOH/O2 [44]. According to this model, trapping of OOH predominates below pH ¼ 7.7, and trapping of O2 above this pH value. The assignment of the DMPO/ OOH adduct followed most data given in the literature [11,19,22,42]. The nature of carbon-centered radical needs to be discussed. In Scheme 1 we proposed a degradation mechanism of the SPEEK membrane by UV in the presence of H2O2 that does not take into

SPEEK/H2O2/DMPO/UV Exp

Sim Adducts

29%

MPO/Ph 25% MPO/OOH 26%

R

R

.C(Ph)

MPO/OPh 20% MPO Degrad

N

H

N

H 3330

O.

O C(Ph)

Scheme 2. Formation of diamagnetic DMPO species.

3340

3350

3360

3370

3380

Magnetic Field / G Fig. 8. ESR experimental (300 K, using flat cell) and simulated spectra of UV irradiation of the DMPO/SPEEK/H2O2/buffer solution (pH 4) system. Conditions: 10 mM DMPO; H2O2 aqueous buffer solution (pH 4); molar ratio [H2O2]/[DMPO] ¼ 1/10, irradiation at 300 K for 30 min. The spectrum was acquired after 60 min subsequent to UV irradiation.

1786

M. Pinteala, S. Schlick / Polymer Degradation and Stability 94 (2009) 1779–1787

consideration the activation of carbonyl groups upon UV irradiation, which could lead to hydrogen atom abstraction from different types of compounds, for example, poly(vinyl alcohol), and to a carbon-centered radical [45,46]. However, in our case, the phenyl groups in SPEEK are more reactive to addition reactions with OH or O2 radicals than to hydrogen abstraction [6–8], and yield the carbon-centered radical (1) indicated in Scheme 1. 4. Conclusions UV irradiation at 77 K of SPEEK in the acid form or fully neutralized in the presence of H2O2/H2O resulted in phenoxyl radicals as the predominant radicals detected by direct ESR and spin trapping methods; phenyl radical formation was observed at pH 4, when the irradiation was performed at 77 K or at room temperature. We assume that the formation of phenoxyl radical is due to water elimination from the peroxyl radical or to scission of the polymer chain, as shown in Scheme 1. The hydroperoxyl radical, as the DMPO/OOH adduct, was detected only when the DMPO/OH adduct was absent, as in Nafion [11], suggesting that the OH radical is formed first, and further reacts to form the other radicals, via reactions (2) and (3) above [18]. With increasing temperature, the carbon-centered radical adduct appears, as seen in Nafion membrane at 240 K [11]. The only possible source of this adduct is the phenyl group from SPEEK. Data obtained by spin trapping were compared with the results of direct ESR detection. This comparison led to the conclusion that these two methods provide complementary information on the nature of the radicals present. Acknowledgements This study was supported by the Polymers Program of the National Science Foundation and by the General Motors Fuel Cell Activities Program. We are grateful to Timothy Fuller of GM for the gift of SPEEK membranes. References [1] Li L, Zhang J, Wang Y. Sulfonated poly(ether ether ketone) membranes for direct methanol fuel cell. J Membrane Sci 2001;226:159–67. [2] Kreuer KD. On the development of proton conducting polymer membranes for hydrogen and methanol fuel cells. J Membrane Sci 2001;185:29–39. [3] Bishop MT, Karasz FE, Russo PS, Langley KH. Solubility and properties of a poly(aryl ether ketone) in strong acids. Macromolecules 1985;18:86–93. [4] Silva VSF, Silva VB, Reissner R, Vetter S, Mendes A, Madeira LM, et al. Nonfluorinated membranes thickness effect on the DMFC performance. Sep Sci Technol 2008;43:1917–32. [5] Lee JK, Li W, Manthiram A. Sulfonated poly(ether ether ketone) as an ionomer for direct methanol fuel cell electrodes. J Power Sources 2008;180:56–62. [6] Panchenko A. DFT investigation of the polymer electrolyte membrane degradation caused by OH radicals in fuel cells. J Membrane Sci 2006;278: 269–78. [7] Mitov S. Ph.D thesis: in-situ radiation grafting of polymer films and degradation studies of monomers for application in fuel cell membranes. Stuttgart: Institut fur Physikalische Chemie der Universita¨t Stuttgart; February 2007. [8] Hu¨bner G, Roduner E. EPR investigation of HOradical initiated degradation reactions of sulfonated aromatics as model compounds for fuel cell proton conducting membranes. J Mater Chem 1999;9:409–18. [9] Bosnjakovic A, Schlick S. Nafion perfluorinated membranes treated in Fenton media: radical species detected by ESR spectroscopy. J Phys Chem B 2004;108: 4332–7. [10] Kadirov MK, Bosnjakovic A, Schlick S. Membrane-derived fluorinated radicals detected by electron spin resonance in UV-irradiated Nafion and Dow ionomers: effect of counterions and H2O2. J Phys Chem B 2005;109: 7664–70. [11] Bosnjakovic A, Schlick S. Spin trapping by 5,5-dimethylpyrroline-N-oxide in Fenton media in the presence of Nafion perfluorinated membranes: limitations and potential. J Phys Chem B 2006;110:10720–8. [12] Danilczuk M, Coms FD, Schlick S. Fragmentation of fluorinated model compounds exposed to oxygen radicals: spin trapping ESR experiments and implications for the behavior of proton exchange membranes used in fuel cells. Fuel Cell 2008;8:436–52.

[13] Roduner E, Schlick S. In: Schlick S, editor. Advanced ESR methods in polymer research, ESR methods for assessing the stability of polymer membranes used in fuel cells. New Jersey: Wiley-Interscience, John Wiley & Sons, Inc; 2006. p. 197–228. [14] Walling C. Fenton’s reagent revisited. Acc Chem Res 1975;8:125–31. [15] (a) Bednarek J, Schlick S. Stability of radical intermediates in microscopically heterogeneous media. Photolysis of water adsorbed on silica gel studied by ESR and DSC. J Phys Chem 1991;95:9940–4; (b) Schlick S, Kevan L. Study of g-anisotropy associated with molecular motion in the triphenylmethylperoxy radical: an environmental probe. J Phys Chem 1979;83(26):3424–9; (c) Melamud E, Schlick S, Silver BL. ESR of tetralin peroxy-17O radical in a single crystal. J Magn Reson 1974;14:104–11; (d) Sevilla MD, Becker D, Yan M. The formation and structure of the sulfoxyl radicals RSO(.), RSOO(.), RSO2(.), and RSO2OO(.) from the reaction of cysteine, glutathione and penicillamine thiyl radicals with molecular oxygen. Int J Radiat Biol 1990;57(1):65–81. [16] Janzen EG. Spin trapping. Acc Chem Res 1971;4:31–40. [17] Lagercrantz C. Spin trapping of some short-lived radicals by the nitroxide method. J Phys Chem 1971;75:3466–74. [18] Harbour JR, Chow V, Bolton JR. An electron spin resonance study of the spin adducts of OH and HO2 radicals with nitrones in the ultraviolet photolysis of aqueous hydrogen peroxide solutions. Can J Chem 1974;52:3549–53. [19] Harbour JR, Hair ML. Detection of superoxide ions in nonaqueous media. Generation by photolysis of pigment dispersions. J Phys Chem 1978;82: 1397–9. [20] Motley C, Mason RP. In: Berliner LJ, Reuben J, editors. Nitroxide radical adducts in biology: chemistry, applications and pitfalls. Biological magnetic resonance, vol. 8. New York: Plenum Press; 1989. p. 489–546. [21] Rosen GM, Britigan BE, Halpern HJ, Pou S. Free radicals: biology and detection by spin trapping. New York: Oxford University Press; 1999. [22] Villamena FA, Locigno EJ, Rockenbauer A, Hadad CM, Zweier JL. Theoretical and experimental studies of the spin trapping of inorganic radicals by 5,5dimethyl-1-pyrroline N-oxide (DMPO). 1. Carbon dioxide radical anion. J Phys Chem A 2006;110:13253–8. [23] Villamena FA, Hadad CM, Zweier JL. Kinetic study and theoretical analysis of hydroxyl radical trapping and spin adduct decay of alkoxycarbonyl and dialkoxyphosphoryl nitrones in aqueous media. J Phys Chem 2003;107:4407–14. [24] Panchenko A, Dilger H, Kerres J, Hein M, Ullrich A, Kaz T, et al. In-situ spin trap electron paramagnetic resonance study of fuel cell processes. Phys Chem Chem Phys 2004;6:2891–4. [25] Haire DL, Janzen EG, Hrvoic I. ESR of new b-phosphorus nitroxides from aldoand keto-nitrones. Magn Reson Chem 1999;37:767–73. [26] Chignell CF, Kalyanaraman B, Sik R, Mason RP. Spectroscopic studies of cutaneous photosensitizing agents-II. Spin trapping of photolysis products from sulfanilamide and 4-aminobenzoic acid using 5,5-dimethyl-1-pyrroline1-oxide. Photochem Photobiol 1981;34:147–56. [27] Bilski P, Reszka K, Bilska M, Chignell CF. Oxidation of the spin trap 5,5dimethyl-1-pyrroline N-oxide by singlet oxygen in aqueous solution. J Am Chem Soc 1996;118:1330–8. [28] Hawkins CL, Davies MJ. Hypochlorite-induced oxidation of proteins in plasma: formation of chloramines and nitrogen-centred radicals and their role in protein fragmentation. Biochem J 1999;340:539–48. [29] Taniguchi H, Madden KP. DMPO-alkyl radical spin trapping: an in situ radiolysis steady-state ESR study. Radiat Res 2000;153:447–53. [30] Dikalov SI, Mason RP. Reassignment of organic peroxyl radical adducts. Free Radic Biol Med 1999;27:864–72. [31] Dikalov SI, Mason RP. Spin trapping of polyunsaturated fatty acid-derived peroxyl radicals: reassignment to alkoxyl radical adducts. Free Radic Biol Med 2001;30:187–97. [32] Brezova V, Valko M, Breza M, Morris H, Telser J, Dvoranova D, et al. Role of radicals and singlet oxygen in photoactivated DNA cleavage by the anticancer drug camptothecin: an electron paramagnetic resonance study. J Phys Chem A 2003;107:2415–25. [33] Kadiiska MB, De Costa KS, Mason RP, Mathews JM. Reduction of 1,3-diphenyl1-triazene by rat hepatic microsomes, by cecal microflora, and in rats generates the phenyl radical metabolite: an ESR spin-trapping investigation. Chem Res Toxicol 2000;13:1082–6. [34] Stasko A, Rapta P, Brezova V, Nuyken O, Vogel R. Photochemically and electrochemically initiated radical decomposition of sulfonium salts (A spin trap study). Tetrahedron 1993;49:10917–24. [35] Hanna PM, Chamulitrat W, Mason RP. When are metal ion-dependent hydroxyl and alkoxyl radical adducts of 5,5-dimethyl-1-pyrroline N-oxide artifacts? Arch Biochem Biophys 1992;296:640–4. [36] Jones CM, Burkitt MJ. EPR spin-trapping evidence for the direct, one-electron reduction of tert-butylhydroperoxide to the tert-butoxyl radical by copper(II): paradigm for a previously overlooked reaction in the initiation of lipid peroxidation. J Am Chem Soc 2003;125:6946–54. [37] Spin trapping database: http://tools.niehs.nih.gov/stdb/index.cfm. [38] The software WinSim can be accessed at http://www.niehs.nih.gov/research/ resources/software/tools/index.cfm. [39] Neta P, Fessenden RW. Hydroxyl radical reactions with phenols and anilines as studied by electron spin resonance. J Phys Chem 1974;78:523–9. [40] Behar D. Are positive ion radicals formed in pulse radiolysis of alkylbenzenesulfonates? J Phys Chem 1991;95:4342–7.

M. Pinteala, S. Schlick / Polymer Degradation and Stability 94 (2009) 1779–1787 [41] Shiga T, Kishimoto T, Tomita E. Aromatic hydroxylation catalyzed by Fenton’s reagent. An electron paramagnetic resonance study. 11. Benzoic acids. J Phys Chem 1973;77:330–7. [42] Rosen GM, Beselman A, Tsai P, Pou S, Mailer C, Ichikawa K, et al. Influence of conformation on the EPR spectrum of 5,5-dimethyl-1-hydroperoxy-1-pyrrolidinyloxyl: a spin trapped adduct of superoxide. J Org Chem 2004;69:1321–30. [43] Emanuel NM, Buchachenco LM. Chemical physics of polymer degradation and stabilization. Gist, Netherlands: VNU Science Press; 1987. p. 179.

1787

[44] Finkelstein E, Rosen GM, Rauckman EJ. Spin trapping of superoxides and hydroxyl radical: practical aspects. Arch Biochem Biophys 1980;200:1–16. [45] Horie K, Ando H, Mita I. Photochemistry in polymer solids. 8. Mechanism of photoreaction of benzophenone in poly (viny1 alcohol). Macromolecules 1987;20:54–8. [46] Korchev AS, Konovalova T, Cammarata V, Kispert L, Slaten L, Mills G. Radicalinduced generation of small silver particles in SPEEK/PVA polymer films and solutions: UV–Vis, EPR, and FT-IR studies. Langmuir 2006;22:375–84.