High Ion Content Siloxane Phosphonium Ionomers with Very Low Tg

Article pubs.acs.org/Macromolecules

High Ion Content Siloxane Phosphonium Ionomers with Very Low Tg Siwei Liang,† Michael V. O’Reilly,‡ U Hyeok Choi,⊥ Huai-Suen Shiau,§ Joshua Bartels,† Quan Chen,† James Runt,† Karen I. Winey,*,‡ and Ralph H. Colby*,† †

Department of Materials Science and Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802, United States ‡ Department of Materials Science and Engineering, University of Pennsylvania, Philadelphia, Pennsylvania 19104-6272, United States § Department of Chemical Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802, United States ⊥ Functional Composites Department, Korea Institute of Materials Science, Changwon, 642-831, Korea S Supporting Information *

ABSTRACT: Polysiloxane phosphonium single-ion conductors grafted with oligomeric PEO and with ion contents ranging from 5 to 22 mol % were synthesized via hydrosilylation reaction. The parent Br− anion was exchanged to F− or bis(trifluoromethanesulfonyl)imide (TFSI−). X-ray scattering data suggest ion aggregation is absent in these phosphonium ionomers, which contributes to low glass transition temperatures (below −70 °C) with only a weak dependence on both ion content and counteranion type. Conductivities weakly increase with ion content but exhibit a strong dependence on anion type. The highest conductivity at 30 °C is 20 μS/cm for dry neat ionomer, with the TFSI− anion, consistent with its relatively delocalized negative charge and large size that weaken interactions between TFSI− and the phosphonium cation.

1. INTRODUCTION Anion exchange ionomers (AEI) have been widely used in many areas, such as water purification, antimicrobial agents, desalination, and alkaline fuel cell membranes.1−4 Recently, potential applications of AEIs for energy storage and conversion have prompted the study of ion conduction in AEIs.1,5−8 Ammonium salts were the first cations investigated in hydroxide exchange fuel cell membranes.1,9 However, because of poor chemical and thermal stability of ammonium salts, alternate salts such as phosphonium and imidazolium have attracted increasing attention.3,5−7,10 Phosphonium salts are more promising for AEI applications than ammonium because phosphorus is more inclined to delocalize electrons than nitrogen due to its empty 3d orbital.11,12 The interesting charge distributions of ion pairs for tetrabutylphosphonium fluoride and tetrabutylammonium fluoride equilibrated in Gaussian 09 are compared in Figure 1. The N atom of the ammonium cation carries a partial negative charge of −0.5e, where e is the elementary charge, with the four α-carbons being positive (+0.24e to +0.35e). In contrast, the P atom of the phosphonium cation has +1.6e charge, with the four α-carbons being negative (−0.19e to −0.36e). The lower electronegativity of P (2.06) relative to C (2.5) allows the positively charged P of phosphonium to be shielded by partially negative carbons, providing for less interaction between a phosphonium cation and many neutralizing anions,11 whereas the stronger electronegativity of N (3.07) makes it negative in ammoniums and imidazoliums, with positive α-carbons that © XXXX American Chemical Society

bind to anions. In addition to phosphonium salts exhibiting naturally weaker ionic interactions, Zhou and Blumstein13 compared phosphonium and ammonium salts having nearly identical structures and concluded that the phosphonium salts demonstrated better thermal and chemical stability. These differences in electronic structure make tetraalkylphosphoniums quite distinct from ammoniums, causing differences in ion aggregation, glass transition temperature (Tg), and chemical stability that are discussed below. Long and co-workers14−16 have synthesized a series of copolymers and polyurethane ionomers based on phosphonium salts. The resulting ionomers were reported to be stable above 300 °C. Gu et al.5 prepared hydroxide exchange membranes for fuel cells with a phosphonium-based ionomer that showed superior conductivity and chemical stability compared to ammonium- and imidazolium-based ionomers. Moreover, many phosphorus-containing materials have proven to be fireretardant.17 Owing to weaker ionic interactions, phosphonium salts synthesized from tributylphosphine, or phosphines with longer alkyl groups, are ionic liquids.18 Polysiloxane-based ionomers are promising ion conductors, owing to their highly flexible backbone that imparts low Tg and therefore respectable room temperature conductivity. While siloxane has low polarity, synthetic versatility enables a wide Received: January 20, 2014 Revised: May 28, 2014

A

dx.doi.org/10.1021/ma5001546 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

Article

Figure 1. Electronic charge distribution of ion pairs in the gas phase for (a) tetrabutylammonium fluoride and (b) tetrabutylphosphonium fluoride, calculated by Gaussian 09 using the B3LYP/6-31+G* basis set. Light green denotes a positive charge, and red denotes a negative charge, with brighter red indicating stronger negative charge. The atomic charges of F, N, P, and the four α-C (with their hydrogens summed into their charge) are given next to the atoms. The charges in the absence of anions are in parentheses. Tetrahydrofuran (THF) from EMD Chemicals was refluxed over sodium metal before use. 2.2. Synthesis of 2,5,8,11-Tetraoxatetradec-13-ene (PEO3). PEO3 was synthesized following the report of West et al.31 To a mixture of NaH (1.44 g, 60% in mineral oil) dispersed in 20 mL of dry THF was added to a solution of tri(ethylene glycol) methyl ether (4.7 mL, 0.03 mol) in 150 mL of THF dropwise at ice-bath temperature. The mixture was stirred for 3 h before being transferred into a solution of allyl bromide (3.58 g, 0.03 mol) in 20 mL of dry THF. The mixture was allowed to react overnight to complete the reaction. The produced NaBr was vacuum filtered, and the volatiles were removed by a rotovap. The yellowish liquid was then purified by vacuum distillation to yield 5.5 g of colorless liquid (90%). 1H NMR (in d6-acetone), δ (ppm) 5.85 (m, 1H, CH), 5.2 (s, (d, cist H of = CH2), 5.1 (d, trans H of = CH2), 3.95 (d, 2H, CC−CH2), 3.8−3.6 (m, 8H, OCH2CH2O), 3.35 (s, CH3). PEO7 and PEO16 were synthesized by a similar method using poly(ethylene glycol) methyl ether with average molecular weights of 300 and 750 g/mol. 2.3. Synthesis of Allyltributylphosphonium Bromide (ATPB). To a pre-degassed three-neck flask were added allyl bromide (3.58 g, 30 mmol) and tributylphosphine (3.6 g, 16 mmol). The entire mixture was allowed to stir at room temperature for 10 h before being diluted with diethyl ether (50 mL). The mixture was filtered, and the solid was washed with diethyl ether to afford the product as a white powder (5 g, 87%). 1H NMR (in d6-acetone), δ (ppm) 0.96 (t, 9H, −CH3), 1.51 (m, 6H, −CH2−), 1.73 (m, 6H, −CH2−), 2.57 (m, 6H, CH2P), 3.64 (q, 2 H, CH2−CHCH2), 5.41 (dd, 1H, trans H of = CH2), 5.65 (dd, 1H, cis H of = CH2), 5.95 (m, 1H, CH). 31P NMR (d6-acetone) δ (ppm) 35.5 (s). 2.4. Synthesis of PSPE Ionomers. Generally 2 g of PMHS was added into a predried flask equipped with a condenser. The desired molar amounts of ATPB and vinyl PEOx were charged into the flask followed by 20 mL of anhydrous CH3CN and several drops of Pt[dvs] catalyst solution. The reaction mixture was stirred at 90 °C. The completion of the reaction was judged by 1H NMR. The mixture was condensed, and the residue was dissolved in DI water and dialyzed against ultrapure water for at least 1 week to remove catalyst, unreacted monomer, and any ionic impurities, with final dialyzate conductivity below 1 × 10−6 S/cm. These ionomers with Br− counterion were then dried in a vacuum oven at 80 °C for 24 h. The ionomers with TFSI− were prepared by dialysis in DI water with an over 50-fold excess of LiTFSI salt. The ionomers with F− and OH− were prepared by passing an aqueous solution through a column packed with anion exchange resin. Our siloxane ionomers with OH− counterions are not stable in water (under basic conditions) presumably because the siloxane bonds hydrolyze.32 2.5. NMR. 1H and 31P NMR spectra were recorded on a Bruker AM 300 M spectrometer with deuterated acetone as the solvent. The

variety of ions and polar side groups to be grafted to the backbone via hydrosilylation reactions. Anionic polysiloxanebased ionomers have been studied extensively as lithium conductors.19−22 We have recently attached bulky tetraphenylborate anions to a polysiloxane backbone to synthesize a series of novel ionomers with very low activation energy for the conducting ions.23 Polysiloxane ionomers with side chains incorporating ammonium salts have been reported,24−27 and conductivities as high as 10−5 S/cm have been reported for ionomers neutralized by I−. Long and co-workers14−16 studied the morphology of several types of phosphonium ionomers. For random copolymer ionomers, wide-angle and small-angle X-ray experiments showed no indicators of ion aggregation.14 Cheng et al.16 found that in triblock copolymer ionomers trioctylalkyl chain substitution on the phosphonium cation protects the charge from aggregation. In polyurethane phosphonium ionomers, Williams et al.15 observed microphase separation into ion-rich and ion-poor domains, while TEM imaging suggested ion-rich domains on length scales larger than 10 nm. Parent et al.28 observed elastomeric behavior consistent with the formation of ionic aggregates in isobutylene-based phosphonium bromide ionomers. In that case though, the polymer matrix poorly solvates ions, promoting ionic aggregation. Herein, we report the synthesis of allyltributylphosphonium bromide monomer and phosphonium-containing, oligomeric PEO grafted siloxane ionomers. These unique single-ion conductors are neutralized with three different conducting counteranions: F−, Br−, and bis(trifluoromethanesulfonyl)imide (TFSI−). X-ray characterization shows no evidence of ion clustering. The phosphonium ionomers with the F− anion display conductivity as high as 10−6 S/cm, which makes our phosphonium ionomers potential electrolytes for novel fluoride-ion batteries.29,30

2. EXPERIMENTAL SECTION 2.1. Materials. Allyl bromide, diethyl ether, toluene, dichloromethane, ethyl acetate, and anhydrous acetonitrile were purchased from VWR and used without further purification. Sodium hydride (60% in mineral oil), tri(ethylene glycol) methyl ether, poly(ethylene glycol) methyl ether (300 and 750 g/mol), platinum divinyltetramethyldisiloxane complex (Pt[dvs]) (3% in xylene) catalyst, tributylphosphine, and polymethylhydrosiloxane (PMHS, Mn = 1700−3200 g/mol) were purchased from Aldrich and used as received. B


Macromolecules

Article

Scheme 1. Synthesis of 2,5,8,11-Tetraoxatetradec-13-ene (PEO3), Allyltributylphosphonium Bromide (ATPB), and Siloxane Random Copolymer Ionomers PSPE-nBr(3)

when there was no longer any peak at ∼4.7 ppm, which is assigned to the Si−H group. The resulting ionomers are named PSPE-nA(x), where n is the mol % of phosphonium salt, A is the anion type, and x is the degree of polymerization of the oligomeric PEO side chain. The compositions of these phosphonium ionomers were determined by the ratio of the integrated areas of the peaks at 0.6 ppm (h) and 2.6 ppm (d) assigned to SiCH2CH2CH2O and PCH2CH2CH2CH3, respectively (Figure 2). The 31P NMR spectrum of the monomer

spectra were used to confirm chemical structure and quantify the targeted theoretical ion contents. 2.6. Thermal Analysis. Glass transition temperatures were determined by differential scanning calorimetry (DSC) in the second heating scan with 10 K/min heating and cooling rates, using a TA Q2000 calibrated with the melting point of methanol (except for the PSPE-22F(3) sample for which a TA Q100 was used). TGA experiments were conducted under a nitrogen atmosphere at 10 K/ min heating rate over a temperature range from 25 to 800 °C. 2.7. Dielectric Relaxation Spectroscopy. For dielectric/ conductometric measurements, samples were sandwiched between two polished brass electrodes with 50 μm silica spacers under 24 h to remove water absorbed from the atmosphere. The liquid ionomers were then loaded into 0.7 mm borosilicate glass capillaries and sealed. X-rays are generated by a Nonius FR-591 rotating anode generator operating at 40 kV and 85 mA that emits Cu Kα radiation (λ = 0.154 nm). The flight path is evacuated, and the beam is focused by Osmic Max-Flux optics. The triple pinhole collimated X-ray beam is scattered at a sample-to-detector distance of 11 or 54 cm and collected by a Bruker Hi-Star two-dimensional multiwire detector. Isotropic two-dimensional intensity data were integrated and converted to 1D plots with Datasqueeze33 analysis software. Sample spectra were corrected for background scattering and transmission by an empty capillary.

Figure 2. Representative 1H NMR spectrum of the phosphonium ionomer with Br− anion with ion content n/(m + n) = 0.05, the inset showing the 31P NMR spectrum (top left). 1H NMR and 31P NMR spectra for the allyltributylphosphonium monomer and another ionomer are shown in Figures S1 and S2, and 1H NMR spectra for all five Br phosphonium ionomers are shown in Figure S3.

3. RESULTS AND DISCUSSION 3.1. Synthesis and Ion Exchange. Scheme 1 shows the synthesis of the phosphonium monomers, oligomeric PEO monomers, and siloxane-based ionomers. No solvent was involved in the synthesis of allyltributylphosphonium bromide (ATPB). The ATPB was prepared under “dry” conditions with very good yield (90%), which provides an economical and facile avenue for the preparation of phosphonium-based ionic liquids. The polymer synthesis reaction was monitored by proton NMR spectroscopy, and completion of the reaction was confirmed

ATPB shifts downfield from −32 ppm for tributylphosphine34 to around 35 ppm (see the inset of Figure 2 and the inset of Figure S1 in the Supporting Information), consistent with the literature.16 After the hydrosilylation reaction, the 31P NMR spectra of the ionomers (see the inset of Figure S2) display a single peak at 33.4 ppm, nearly identical to that of the ATPB monomer, suggesting intact phosphonium salts after the chemical reaction. C


Macromolecules

Article

Table 1. Physical Properties of Nonionic PEO-Grafted Siloxanes and Siloxane Phosphonium Ionomers composition (intended) sample

a

PSPE-0(3) PSPE-0(7) PSPE-0(16) PSPE-5Br(3) PSPE-8Br(3) PSPE-11Br(3) PSPE-22Br(3) PSPE-5TFSI(3) PSPE-8TFSI(3) PSPE-11TFSI(3) PSPE-5F(3) PSPE-8F(3) PSPE-11F(3) PSPE-22F(3)

anion

n

m

none

0 0 0 5 8 11 22 5 8 11 5 8 11 22

100 100 100 95 92 89 78 95 92 89 95 92 89 78

Br−

TFSI−

F−

ion content (nm−3) expectation based on Br salt parentb

0.116 0.183 0.228 0.116 0.183 0.228 0.441

NMRc

DSC Tg (°C)

conductivity at 30 °C (μS/cm)

0.115 0.183 0.228 0.441 0.116 0.183 0.228 0.116 0.090 0.16 0.15

−86 −72 −69 −83 −82 −80 −86 −81 −81 −80 −80 −83 −82 −73d

0.56 0.75 0.68 1.44 10.9 31.2 21.2 0.19 0.2 0.17 0.74

a

Number after the dash indicates mol % of phosphonium relative to PEOx side chains. Number in parentheses indicates the PEOx side chain degree of polymerization. bValues are based on the analysis of NMR results of that same ionomer with Br− as the counterion. For ionomers with different counteranions, the ion contents are assumed to be the same; found to be valid for TFSI− but not for F− above 5 mol %. cValues calculated from NMR were determined by the ratio of integrated area of the peaks at 0.6 and 2.6 ppm. dFigure 9c suggests this Tg value is about 10 K too high for PSPE-22F(3).

Aqueous solutions of PSPE-5Br(3), 8Br(3), 11Br(3), and 22Br(3) are cloudy and colloid-like. When the Br anions were replaced by TFSI anions, the solutions turned more turbid, partially consistent with Ye and Elabd’s observation10 that imidazolium ionomers with bromide anions are water-soluble, while the same ionomers with TFSI anions are insoluble in water. While the oligomeric PEO side groups promote water miscibility (the PEO3−siloxane homopolymer is fully soluble and dissolves readily in water), the hydrophobic butyl groups of phosphonium explain the turbidity. Ionomers exchanged to OH− were unstable in aqueous solution due to the decomposition reactions between OH− and the ionomer in strong basic conditions. When Br− is replaced by F− (ion exchange was conducted by passing the ionomer aqueous solution through a column packed with anion exchange resin charged with NaF), it was found that the ionomers were unstable in aqueous solution for a long time. Some ionic groups were lost, as confirmed by proton NMR (see Table 1), which might be explained by the weak acidity of HF with pKa ∼ 3.1. Consequently, F− forms relatively stable HF in aqueous solution, thereby leaving the solution slightly basic. In Figure 3, the charge distribution of one repeat unit with the ionic group is calculated, and it is found that the αcarbon connecting to the polysiloxane backbone is the most negative of the four α-carbons (−0.37e compared to ∼−0.20e) and hence the most susceptible to cleave. This is consistent with the NMR result showing decreased intensity of the phosphonium group but no new peak identified for the phosphonium fluoride ionomers with higher ion content. 3.2. Glass Transition Temperature and Thermal Stability. Table 1 shows DSC Tgs of the phosphonium ionomers with different anions and varying ion content. For each anionic counterion, as ion content increases, Tg is nearly independent of ion type and only mildly increases with ion content (see also the graphic accompanying the abstract). Cheng et al.16 observed the same behavior for their phosphonium ionomers with ion contents up to 21 mol %, with Tg = −47 °C, only 4 K above the Tg of their lowest ion

Figure 3. Structure of a simple model of the ionic repeat unit with F− counterion, optimized at the B3LYP/6-3+G* level in Gaussian 09. The atomic charges of F, P, and the four α-carbons (with hydrogens summed into the charge on each C) are given next to the atoms. The atomic charges of Si and O for the siloxane backbone are also displayed.

content. As will be shown in the next section, this is very likely a consequence of negligible ion aggregation in our ionomers, thus minimal limitation of chain segmental motion by physical ionic cross-links, consistent with a detailed dynamics study of the phosphonium bromide ionomers.35 It was reported10 that when counter-anions were exchanged from Br− to TFSI−, the Tg of imidazolium ionomers decreased substantially, owing to a plasticizing effect of TFSI and much weaker ionic interactions between TFSI and imidazolium. While for the siloxane ionomers studied in this paper, it is noted that at ion contents of 11 mol % or lower the siloxane ionomers with TFSI counterion exhibit similar Tgs but superior conductivity to the ionomers containing Br− or F− anions. The backbone of our phosphonium ionomers is polysiloxane, a highly flexible polymer chain, which endows our ionomers with D


Macromolecules

Article

lower Tgs than typical ionomers having C−C backbones. In contrast, unpublished results from our group show polymethacrylates with PEO5 side chains have Tg = 226 K, 40 K higher than PSPE-0(3). The importance of the siloxane backbone can also be seen by comparing Tg of PSPE-0(3) that is ∼1/4 siloxane backbone to PSPE-0(16) that is ∼1/10 siloxane backbone having 17 K higher Tg. In contrast, carbon backbone polymers (such as acrylates and methacrylates) have Tg decrease as flexible side chains are made longer. We have recently shown that the molar volume, Vm, of the side group (including the counterion) controls the Tg in this class of ionomer:36 Tg decreases rapidly with increasing Vm and Tg becomes insensitive to Vm in the large Vm limit. In the current study, Tgs of the phosphonium ionomers typically vary in a small range from −80 to −70 °C. The insensitivity of Tg to ionic content and type of counterion suggests that the allyltributylphosphonium ion is sufficiently large that all the ionomer samples are in the large Vm limit. Tg ≈ −80 °C for phosphonium siloxane ionomers is considerably lower than Tg ≈ −52 °C for imidazolium acrylate and methacrylate ionomers in the large Vm limit.36 This difference is attributed to the far more flexible polysiloxane chain backbone in comparison to polyacrylate and polymethacrylate backbones. The thermal stability of our phosphonium ionomers is similar to the phosphonium ionomers reported by Long et al.14−16 There is no significant weight loss at temperatures up to 300 °C in TGA, regardless of the counterion. Weight loss measurements for PSPE ionomers with Br− counterions are shown in Figure 4. After dielectric spectroscopy measurement (red line in Figure 4), with over 1 h at 120 °C under N2, these ionomers maintain thermal stability.

Figure 5. X-ray scattering of PSPE-0(x) at 125 °C, constructed by splicing wide and intermediate angle scattering data, where x represents the PEO side chain degree of polymerization. Curves are shifted vertically for clarity.

is ∼0.44 nm (2π/q), which is consistent with the spacing found for interchain separation in crystalline PEO.37 The peak II position is independent of the random copolymer composition. As the PEO side chain length increases from 3 to 7 to 16, corresponding to 72, 78, and 86 vol % PEO, respectively, peak I weakens and shifts to lower angle and we assign peak I to siloxane backbone−backbone separation. Galin and Mathis38 determined the interaction parameter (χ) between polydimethylsiloxane and poly(ethylene oxide) to be large and positive for PDMS−PEO−PDMS triblock copolymers. Because of the strong incompatibility of the siloxane backbone and the PEO side chains in our system, the peaks at q = 1.5−4.0 nm−1 indicate the typical spacing between backbones containing side chains.39−44 The electron densities for siloxane and PEO are 310 and 371 e−/nm3, respectively, which provide sufficient contrast to observe the separation between siloxane backbones. Lengthening the oligomeric PEO side chain from 3 to 16 effectively dilutes the siloxane backbones, contributing to the broadening and loss of intensity of peak I. Peak I shifts from q = 3.9 nm−1 for x = 3 to q ≈ 1.6 nm−1 for x = 16 (Table 2). The dTable 2. Characteristics of Nonionic Oligomeric PEO Grafted Siloxanes sample PSPE-0(3) PSPE-0(7) PSPE-0(16)

q vol % PEOa (nm−1) 72 78 86

3.9 2.4 1.8

d (nm)

bond lengths per side chain

N (bonds/6.7)

bN1/2 (nm)

1.6 2.6 3.5

14 26 53

2.1 3.9 7.9

1.6 2.2 3.1

Figure 4. TGA weight loss of the phosphonium monomer and PSPE ionomers with Br− anion. Each sample is fully dry, and data were taken under a heated stream of dry nitrogen. Note that the primary degradation onset of 350 °C is ∼425 K above Tg.

a

3.3. Morphology of Oligomeric PEO Grafted Siloxanes. Before presenting the morphology of ionic PEO grafted siloxanes, we consider carefully three nonionic PEO grafted siloxane polymers. Figure 5 shows X-ray scattering of three PSPE-0(x) polymers, where x represents the degree of polymerization of the PEO side chain: 3, 7, and 16. The dominant feature in the wide angular regime is a broad amorphous halo (labeled peak II) centered at q ∼ 15 nm−1 at 125 °C, corresponding to the amorphous PEO side chain to side chain separation. This amorphous interchain PEO spacing

spacings (d = 2π/q) are comparable to the end-to-end length of a Gaussian PEO side chain, which are also listed in Table 2 along with the number of bonds and the number of Kuhn segments N in the PEO side chain. The calculation of the endto-end distance used a Kuhn length of b = 1.1 nm and Flory characteristic ratio 6.7.40 As number of Kuhn segments increases, the side chain approaches random walk statistics, and siloxane backbone distance is expectedly shorter than the side chain’s contour length. Previous studies on poly(n-alkyl methacrylates),41−45 poly(alkylene oxides),42 and poly(n-alkyl

Vol % PEO was approximated using the bulk densities of PEO and siloxane: ρPEG = 1.13 g/mL, ρsilox = 1.00 g/mL.

E


Macromolecules

Article

glutamates)46 with similar comblike molecular architectures also show a backbone−backbone spacing peak. The length scale of this scattering feature is correlated to the number of bonds in the side chain. Backbone spacings of our PEO-grafted siloxanes are plotted against the number of bonds per side chain in Figure 6, along with literature values culled for various

Figure 6. Comparison of backbone spacing in amorphous polymers with flexible side chains, plotted against the number of bonds per side chain.

amorphous polymers. The backbone spacing shows a nonlinear dependence on side chain length, as previously reported. Backbone monomer molecular weight, side chain polydispersity, and backbone/side chain compatibility account for minor fluctuations around the trend observed in Figure 6. Overall, our PSPE-0(x) siloxane polymers at 125 °C follow the observed behavior for polymers with amorphous side chains with respect to how backbone−backbone separation depends on the side chain length. 3.4. Morphology of Oligomeric PEO Grafted Siloxane Phosphonium Ionomers. Ionomers neutralized with Br−, TFSI−, and F− show nearly identical X-ray scattering in Figure 7, with three scattering features, two of which were previously observed in the nonionic polymers in Figure 5. The high-angle peak, q = 15 nm−1, is primarily the amorphous halo from the PEO side chains (peak II). The peak position (q ∼ 3.7 nm−1) and intensity of the siloxane backbone-to-backbone spacing (peak I) are also constant at all ion contents and counterion types because these copolymers have the same PEO side chain (PEO3) with similar molar volume as the phosphonium side chains with various counterions. Electron densities for phosphonium salts are calculated to be 358, 377, and 419 e−/nm3 for F, Br, and TFSI, respectively, based on densities approximated by Ye and Shreeve.47 Because these electron densities are close to that of PEO, contrast between these polymers is nominally independent of ion type. Peak III is intermediate between peak I and II, only seen in the ionomers and strongest at highest ion contents (Figure 7). Thus, we assign peak III to a shorter local backbone−backbone spacing created when ion pairs on two different chains form a quadrupole by dipole−dipole attraction. Interestingly, there are no explicit scattering contributions from more extensive ionic aggregation. The suppression of ionic aggregates is attributed to the bulkiness of the phosphonium cations and charge shielding caused by its butyl segments. The absence of physical ionic cross-links and weak

Figure 7. X-ray scattering of PSPE-nA(3) at 25 °C, where the anionic counterion A = (a) Br−, (b) TFSI−, or (c) F−. Curves are shifted vertically for clarity.

quadrupole binding energy are consistent with Tg remaining low in all of these phosphonium ionomers. Despite the lack of explicit ionic aggregation, we expect that these ionomers will show ion pair-to-ion pair scattering between q = 3−5 nm−1, as estimated by assuming the ion pairs are randomly distributed in the ionomer. Phosphonium bromide pairs, for example, will scatter ca. q ∼ 3 nm−1 at 5 mol % phosphonium salt and shift to ca. q ∼ 5 nm−1 at 22 mol % phosphonium salt. Broad contributions from ion pair-to-ion pair and quadrupole-toquadrupole scattering account for the increase in scattering intensity of PSPE ionomers between q = 3−8 nm−1 as ion content increases. Consequently, the morphology may be interpreted as a PEO matrix containing siloxane backbones, well-distributed ion pairs, and some association of ion pairs to form quadrupoles that result from dipole−dipole attraction of two nearby ion pairs. F


Macromolecules

Article

opposite side of the P is more negatively charged than the other three (Figure 3). The dielectric constant for the F− ionomers obeys the Onsager equation but with a much larger effective dipole of 15 D, suggesting a mixture of separated pairs and contact ion pairs for the F− ionomers, while the ionomers with Br− and TFSI− counterions have nearly exclusively contact pairs. Thus, these static dielectric constants are also suggesting little or no ion aggregation, consistent with X-ray measurements and DSC Tg. 3.6. Ionic Conductivity. The phosphonium single-ion conductors with different counterions and ion contents shows a weak dependence of ionic conductivity on ion content in Figure 9. It is well-known that ion conduction in polymers is usually

Figure 8 compares X-ray scattering for PSPE-11A(3) for all three neutralizing anions at 25 and 125 °C. Thermal expansion

Figure 8. X-ray scattering of PSPE-11A(3), where A = Br−, TFSI−, or F−. Closed symbols (●) are data at 25 °C, and open symbols (○) are data at 125 °C. Curves are shifted vertically for clarity, and samples are thermally reversible.

causes the amorphous halo (peak II) to shift to slightly lower q at 125 °C. Most notably in all three samples, the scattering intensity of peak I increases in intensity at elevated temperatures relative to the amorphous carbon halo for which no significant change in scattering intensity is expected. Upon cooling to 25 °C the scattering patterns are fully recovered. Overall, the PSPE-11A(3) copolymers do not exhibit substantial morphology changes across the temperature range of 25−125 °C, which is important as we explore the transport properties as a function of temperature. 3.5. Static Dielectric Constant. The dependence of static dielectric constant εs (the low frequency value of permittivity prior to electrode polarization) on temperature and ion content, for all the phosphonium ionomers with different counterions, obeys the Onsager equation.48 The effective dipole associated with each phosphonium is compared with the dipole of the contact ion pair in the gas phase from Gaussian 09 using the B3LYP/6-31+G* basis set in Table 3. For ionomers with

Figure 9. Temperature dependence of DC conductivity of phosphonium ionomers with different ion content, normalized by DSC Tg for three counterions: (a) bromide, (b) TFSI, (c) fluoride. The conductivity data for PSPE-22F(3) suggest that the DSC Tg reported in Table 1 is 10 K too high (the real Tg may be −83 °C for PSPE-22F(3), and that is the only sample for which DSC Tg had to be measured on a different instrument than all other samples).

Table 3. Dipoles of Contact Ion Pairs from Gaussian 09 Compared with Experimental Values from the Measured Static Dielectric Constant counterion −

F Br− TFSI−

gas phase dipole of contact ion pair (D)

dipole associated with each cation, from εs (D)

2.5 12.3 16.3

15.0 13.3 15.9

coupled to chain segmental motion,49 and this seems universally true for all ionomers based on PEO, so in Figure 9 temperature is normalized by DSC Tg. However, when phosphonium salt concentration increases from 5 to 22 mol %, Tg barely changes (see Table 1). X-ray scattering and static dielectric constant data provide no evidence for physical crosslinking via ionic aggregates at 25 or 125 °C, consistent with ionomer segmental dynamics being largely unaffected. Therefore, we observe only modest conductivity improvements at the highest ion contents, whereas ionomers usually show lower conductivity at high ion content because Tg usually increases strongly with ion content.50 Consistently low Tgs and ionic

Br− and TFSI− counterions, the dipole calculated from εs using the Onsager equation compares favorably with the calculated dipole for the contact ion pair in the gas phase, suggesting that these ions are mostly in the contact ion pair state. For the F− ionomers, Gaussian finds a very small dipole for the contact pair of 2.5 D, partly because the F− gets very close to P and shares electrons and partly because the α-carbon on the G


Macromolecules

Article

aggregation, consistent with the X-ray, static dielectric constant, and DSC Tg results reported here. If the counterion is exchanged for a more bulky, charge delocalized mobile species, the ionomer exhibits a substantial increase in net counterion transport. Conductivities of phosphonium ionomers with the same 11 mol % ion content, but different anion species are shown in Figure 11. The

conductivity measurements that are relatively insensitive to phosphonium composition suggest that the conductivity is dominated by segmental motion of the PEO side chains. Furthermore, conductivity varies smoothly with temperature, consistent with the absence of significant morphology changes across this temperature range (Figure 8). Even though the Tg increases slightly with ion content, conductivity of the phosphonium ionomers with F− mobile anions increase with ion content up to the highest ion content studied (22 mol % phosphonium) because ion hopping distances are shortened by the higher ion content and Tg barely changed. The conductivity is as high as 10−6 S/cm at room temperature, making this a promising material for the electrolyte separator in a fluoride-ion battery.29,30 Although not studied here, the conductivity of iodide salts of these ionomers is expected to be between Br− and TFSI−, suggesting that these phosphonium ionomers also have potential use as single-ion conductors for dye-sensitized solar cells. The data analysis methods of electrode polarization and derivative spectra that are commonly utilized for ionomers reveal the number density of simultaneous conductors, their mobility, and a segmental relaxation that involves ion rearrangements, termed the α2 relaxation. Figure 10 demon-

Figure 11. DC conductivities of siloxane phosphonium ionomers with different counteranions having ion content n/(m + n) = 0.11.

conductivities of those ionomers increase with increasing counterion size: F− < Br− < TFSI−. Since the morphologies of these ionomers are comparable for all ion types, the differences in conductivity stem from the weaker ionic interactions associated with larger counterions. Ye et al.10 studied imidazolium-based polymerized ionic liquid and found that the conductivity of ionomers with TFSI− anions was greater than those of ionomers with PF6− or BF4−. They attributed the difference to not only the size effect but also delocalized charge distribution and flexibility of the TFSI− anion.57 Our electrode polarization analysis yields activation energies (Ea) for the number density of simultaneous conductors for these counterions summarized in Table 4. Figure 10. DC conduction rates of siloxane phosphonium ionomers as a function of the product εsωα2. As observed for other ionomers,36,55 this BNN relation51−54 suggests a connection between ionic segmental relaxation and ion conduction.

Table 4. Conducting Ion Properties of Different Anions (A) in PSPE-11A(3) Siloxane Phosphonium Ionomers

strates that our phosphonium ionomers follow the Barton− Nakajima−Namikawa (BNN) relation,51−54 that ionic conductivity is proportional to the product of ion motion peak relaxation frequency (ωα2) and the static dielectric constant (εs). As Choi et al.55 and Fragiadakis et al.56 suggest, this indicates the ionic segmental relaxation controls ionic conductivity, as expected in these PEO−siloxane ionomers. A simple scaling treatment based on the Nernst−Einstein equation leads to the BNN relation for DC ionic conductivity.36 σDC = Bε0εsωα2 (1)

ionomer

TFSI

Br

F

Ea (kJ/mol) anion size (van der Waals radii, nm)58,59 ion pair energy (gas phase, kJ/mol)

9.4 0.326 284

14.2 0.195 369

18.3 0.136 481

The low Ea for TFSI− containing phosphonium ionomer is consistent with its highest conductivity, which might suggest that Ea is the key factor determining conductivity in our low-Tg phosphonium ionomers in the absence of ionic aggregation.

4. CONCLUSIONS Allyltributylphosphonium bromide (ATPB) has been successfully synthesized under a solvent-free condition. These phosphonium salts and vinyl PEOx oligomers have been attached to polysiloxane backbones as side chains to produce single-ion conductors. Parent Br− ions were exchanged for different anions (F− or TFSI−). The ionomers with TFSI−, Br−, or F− counterions are stable at 120 °C in dry nitrogen or vacuum. X-ray scattering indicates no ionic aggregation in these phosphonium ionomers, consistent with their very low Tg that only increases quite weakly with ion content and is insensitive

This empirical scaling correlation relates the two primary measurements of dielectric spectroscopy, the ionic conductivity σDC, and the static dielectric constant εs through the frequency of the dielectric loss maximum ωα2 for the segmental relaxation that involves rearrangement of ions. B is a dimensionless empirical number that seems to range from 0.3 for ionomers without ion aggregates to 30 for ionomers with strongly aggregated ions, while ε0 is the permittivity of vacuum. The value of B = 0.3 in Figure 10 suggests little or no ion H


Macromolecules

Article

(8) Ye, Y.; Elabd, Y. A., Chemical Stability of Anion Exchange Membranes for Alkaline Fuel Cells. In Polymers for Energy Storage and Delivery: Polyelectrolytes for Batteries and Fuel Cells; Page, K. A., Soles, C. L., Runt, J., Eds.; American Chemical Society: Washington, DC, 2012; Vol. 1096, pp 233−251. (9) Yamamoto, M.; Toi, K. Electrodialytic Method. JP Patent 61192311, 1986. (10) Ye, Y.; Elabd, Y. A. Polymer 2011, 52, 1309−1317. (11) Wang, S.-W.; Liu, W.; Colby, R. H. Chem. Mater. 2011, 23, 1862−1873. (12) Haav, K.; Saame, J.; Kütt, A.; Leito, I. Eur. J. Org. Chem. 2012, 2167−2172. (13) Zhou, P. G.; Blumstein, A. Polymer 1997, 38, 595−604. (14) Cheng, S.; Zhang, M.; Wu, T.; Hemp, S. T.; Mather, B. D.; Moore, R. B.; Long, T. E. J. Polym. Sci., Part A: Polym. Chem. 2012, 50, 166−173. (15) Williams, S. R.; Wang, W.; Winey, K. I.; Long, T. E. Macromolecules 2008, 41, 9072−9079. (16) Cheng, S.; Beyer, F. L.; Mather, B. D.; Moore, R. B.; Long, T. E. Macromolecules 2011, 44, 6509−6517. (17) Riley, D. J.; Gungor, A.; Srinivasan, S. A.; Sankarapandian, M.; Tchatchoua, C.; Muggli, M. W.; Ward, T. C.; McGrath, J. E.; Kashiwagi, T. Polym. Eng. Sci. 1997, 37, 1501−1511. (18) Fraser, K. J.; MacFarlane, D. R. Aust. J. Chem. 2009, 62, 309− 321. (19) Okamoto, Y.; Yeh, T. F.; Lee, H. S.; Skotheim, T. A. J. Polym. Sci., Part A: Polym. Chem. 1993, 31, 2573−2581. (20) Siska, D. P.; Shriver, D. F. Chem. Mater. 2001, 13, 4698−4700. (21) Snyder, J. F.; Hutchison, J. C.; Ratner, M. A.; Shriver, D. F. Chem. Mater. 2003, 15, 4223−4230. (22) Fujinami, T.; Buzoujima, Y. J. Power Sources 2003, 119−121, 438−441. (23) Liang, S.; Choi, U. H.; Liu, W.; Runt, J.; Colby, R. H. Chem. Mater. 2012, 24, 2316−2323. (24) Kang, J. J.; Fang, S. B. Polym. Bull. 2002, 49, 127−134. (25) Kang, J. J.; Li, W. Y.; Lin, Y. A.; Li, X. P.; Xiao, X. R.; Fang, S. B. Polym. Adv. Technol. 2004, 15, 61−64. (26) Sauvet, G.; Dupond, S.; Kazmierski, K.; Chojnowski, J. J. Appl. Polym. Sci. 2000, 75, 1005−1012. (27) Huang, Z. X.; Yu, Y. Z.; Huang, Y. J. Appl. Polym. Sci. 2002, 83, 3099−3104. (28) Parent, J. S.; Penciu, A.; Guillen-Castellanos, S. A.; Liskova, A.; Whitney, R. A. Macromolecules 2004, 37, 7477−7483. (29) Yazami, R. Fluoride Ion Electrochemical Cell. US20090029237 A1, 2009. (30) Yazami, R.; Hamwi, A. Fluorination of multi-layered carbon nanomaterials. US7794880 B2, 2010. (31) Zhang, Z. C.; Lyons, L. J.; Amine, K.; West, R. Macromolecules 2005, 38, 5714−5720. (32) Metkin, I. A.; Piotrovskii, K. B.; Yuzhelevskii, Y. A. Zh. Prikl. Khim. 1975, 48, 1108−1112. (33) Heiney, P. A. Comm. Powder Diffr. Newsl. 2005, 32, 9−11. (34) Xu, S. L.; Chen, R. S.; He, Z. J. J. Org. Chem. 2011, 76, 7528− 7538. (35) Chen, Q.; Liang, S.; Shiau, H.-s.; Colby, R. H. ACS Macro Lett. 2013, 2, 970−974. (36) Choi, U. H.; Mittal, A.; Price, T. L.; Gibson, H. W.; Runt, J.; Colby, R. H. Macromolecules 2013, 46, 1175−1186. (37) Wang, W.; Liu, W.; Tudryn, G. J.; Colby, R. H.; Winey, K. I. Macromolecules 2010, 43, 4223−4229. (38) Galin, M.; Mathis, A. Macromolecules 1981, 14, 677−683. (39) Beiner, M.; Huth, H. Nat. Mater. 2003, 2, 595−599. (40) Rubinstein, M.; Colby, R. H. Polymer Physics; Oxford University Press: Ithaca, NY, 2003. (41) Arbe, A.; Genix, A. C.; Colmenero, J.; Richter, D.; Fouquet, P. Soft Matter 2008, 4, 1792−1795. (42) Gerstl, C.; Brodeck, M.; Schneider, G. J.; Su, Y.; Allgaier, J.; Arbe, A.; Colmenero, J.; Richter, D. Macromolecules 2012, 45, 7293− 7303.

to counteranion. The low Tg is attributed to (1) the inherent flexibility of the polysiloxane backbone, (2) the presence of ionsolvating PEO side chains that facilitate ion conduction, (3) lack of ionic aggregation, and (4) the electronic structure of the phosphonium cation. The conductivities of phosphonium ionomers are enhanced by increasing anion size. The ionomers with TFSI− show the highest conductivity across the whole temperature range, owing to the largest size of TFSI− and weakest ionic interactions between TFSI− and the phosphonium cation attached to the polymer. Whereas conventional ionomers have Tg increase strongly with ion content,50 our tetraalkylphosphonium ionomers with high ionic content of 22 mol % only have Tg 13 K larger than that of their nonionic equivalent (PSPE-0(3) with Tg = −86 °C). Tg barely changing with ion content is very rare, only previously reported in Weiss’ study of sulfonated polystyrene with a series of alkylammonium counterions.60 Tetrabutylammonium counterions, quite similar in size to our phosphoniums, exhibit similar insensitivity of Tg to ion content, while even longer tail ammonium counterions actually act as plasticizers that lower Tg! This suggests a new direction for materials synthesis of low-Tg single-ion conductors for superior ambient ionic conductivity.

■

ASSOCIATED CONTENT

S Supporting Information *

1 H and 31P NMR spectra of phosphonium monomer and representative phosphonium ionomers. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail [email protected] (R.H.C.). *E-mail [email protected] (K.I.W.). Author Contributions

S.L. and M.V.O. made equal contributions. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The work is supported by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, under Contract DEFG02-07ER46409. The authors thank Janna Maranas and Karl Mueller at Penn State, Yossef Elabd at Drexel University, and Timothy Long at Virginia Polytechnic Institute and State University for helpful discussions.

■

REFERENCES

(1) Couture, G.; Alaaeddine, A.; Boschet, F.; Ameduri, B. Prog. Polym. Sci. 2011, 36, 1521−1557. (2) Brenner, D.; Oswald, A. A., Effect of Counterion Structure on Ionomer Properties: Quaternary Phosphonium Counterions in Sulfonated EPDM. In Ions in Polymers; Eisenberg, A., Ed.; McGill University: Montreal, 1980; Vol. 187, pp 53−66. (3) Ghassemi, H.; Riley, D. J.; Curtis, M.; Bonaplata, E.; McGrath, J. E. Appl. Organomet. Chem. 1998, 12, 781−785. (4) Worley, S. D.; Sun, G. In Polymeric Materials Encyclopedia; Salamone, J. C., Ed.; CRC: Boca Raton, FL, 1996; Vol. 1, p 550. (5) Gu, S.; Cai, R.; Luo, T.; Chen, Z. W.; Sun, M. W.; Liu, Y.; He, G. H.; Yan, Y. S. Angew. Chem., Int. Ed. 2009, 48, 6499−6502. (6) Lu, S.; Pan, J.; Huang, A.; Zhuang, L.; Lu, J. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 20611−20614. (7) Tang, D.; Pan, J.; Lu, S.; Zhuang, L.; Lu, J. Sci. China: Chem. 2010, 53, 357−364. I


Macromolecules

Article

(43) Floudas, G.; Štepánek, P. Macromolecules 1998, 31, 6951−6957. (44) Miller, R. L.; Boyer, R. F.; Heijboer, J. J. Polym. Sci., Polym. Phys. Ed. 1984, 22, 2021−2041. (45) Beiner, M.; Kabisch, O.; Reichl, S.; Huth, H. J. Non-Cryst. Solids 2002, 307−310, 658−666. (46) Watanabe, J.; Ono, H.; Uematsu, I.; Abe, A. Macromolecules 1985, 18, 2141−2148. (47) Ye, C. F.; Shreeve, J. M. J. Phys. Chem. A 2007, 111, 1456−1461. (48) Onsager, L. J. Am. Chem. Soc. 1936, 58, 1486−1493. (49) Ratner, M. A.; Shriver, D. F. Chem. Rev. 1988, 88, 109−124. (50) Eisenberg, A.; Kim, J. S. Introduction to Ionomers; Wiley: New York, 1998. (51) Barton, J. L. Verres Refract. 1966, 20, 328−335. (52) Nakajima, T. In 1971 Annual Report, Conference on Electrical Insulation and Dielectric Phenomena, National Academy of Sciences, Washington DC, 1972; National Academy of Sciences: Washington DC, 1972. (53) Namikawa, H. J. Non-Cryst. Solids 1974, 14, 88−100. (54) Namikawa, H. J. Non-Cryst. Solids 1975, 18, 173−195. (55) Choi, U. H.; Lee, M.; Wang, S.; Liu, W.; Winey, K. I.; Gibson, H. W.; Colby, R. H. Macromolecules 2012, 45, 3974−3985. (56) Fragiadakis, D.; Dou, S.; Colby, R. H.; Runt, J. J. Chem. Phys. 2009, 130, 064907. (57) Arnaud, R.; Benrabah, D.; Sanchez, J. Y. J. Phys. Chem. 1996, 100, 10882−10891. (58) Ue, M.; Murakami, A.; Nakamura, S. J. Electrochem. Soc. 2002, 149, A1385−A1388. (59) Jenkins, H. D. B.; Thakur, K. P. J. Chem. Educ. 1979, 56, 576− 577. (60) Weiss, R. A.; Agarwal, P. K.; Lundberg, R. D. J. Appl. Polym. Sci. 1984, 29, 2719−2734.

J


High Ion Content Siloxane Phosphonium Ionomers with Very Low Tg

Recommend Documents