Controlling Crystallinity in Graft Ionomers, and Its Effect on Morphology

Apr 7, 2013 - The polystyrene graft chains were subsequently sulfonated to different degrees to provide three series of polymers with controlled ion e...
0 downloads 9 Views 2MB Size
Article pubs.acs.org/cm

Controlling Crystallinity in Graft Ionomers, and Its Effect on Morphology, Water Sorption, and Proton Conductivity of Graft Ionomer Membranes Ami C. C. Yang,† Rasoul Narimani,‡ Zhaobin Zhang,†,§ Barbara J. Frisken,‡ and Steven Holdcroft*,† †

Department of Chemistry and ‡Department of Physics, Simon Fraser University, Burnaby, British Columbia V5A 1S6, Canada S Supporting Information *

ABSTRACT: To gain insight into the role of crystallinity and morphology on proton transport through solid polymer electrolytes, we synthesized graft copolymers, poly(vinylidene difluoride-co-chlorotrifluoroethylene)-g-polystyrene [P(VDF-co-CTFE)-g-PS], consisting of a hydrophobic, fluorous backbone and styrenic graft chain of varied length (DPstyrene = 39, 62, and 79), by graft atom transfer radical polymerization (ATRP). The polystyrene graft chains were subsequently sulfonated to different degrees to provide three series of polymers with controlled ion exchange capacity (IEC). The crystallinity and morphology of solution-cast membranes were examined by XRD and TEM, respectively. The grafting of the parent side chain is found to hinder crystallization of the fluorous backbone and the impact of the degree of sulfonation of the side chain on the crystallinity of the polymer is dependent on the graft length: No impact is found for medium and long graft lengths, but for short graft length copolymers (PS39), the degree of crystallinity in the sulfonated membranes is twice that of the unsulfonated membrane. A phase-separated morphology consisting of 2−5 (±1) nm ion-rich domains is observed for all of the graft copolymers. These graft copolymers allow access to very high IEC membranes (>3 mmol/g), which are insoluble in water. The shorter graft length series, P(VDF-co-CTFE)-g-SPS39, swells less in the intermediate IEC range (312 000 g/mol), low polystyrene graft density (∼0.3 mol %) and large graft length (∼120 styrene units) yielded membranes that exhibited long-range lamella/cylinder ionic channels imbedded in the hydrophobic PVDF matrix. These materials offer improved resistance to water swelling at high IEC, with comparable proton conductivity to Nafion 112 up to 120 °C and high RH. Graft copolymers based on high molecular weight PVDF, higher polystyrene graft density (1.4−2.4 mol %) and shorter graft length (14−21 styrene units) resulted in a disordered cluster-network morphology which exhibited lower water swelling and less sensitivity to RH. In complementary work, Tsang et al. investigated graft copolymers of [P(VDF-co-CTFE)-g-PS] using a higher graft density, resulting in higher PS to VDF molar ratios.56 Copolymers possessing shorter graft side chains yielded larger and purer ionic domain clusters because of the higher degree of sulfonation (DS) compared to other analogues with similar IEC. For higher IEC membranes, short graft copolymers were found to possess larger ionic aggregates due to the high DS but a lower number density of clusters; the ionic domains are therefore more isolated, which strengthened the cohesivity of the surrounding hydrophobic matrix, resulting in lower water uptake.

Scheme 1. Synthetic Scheme for P(VDF-co-CTFE)-g-SPS

previously.56 These series of graft copolymers were designed to possess a lower CTFE content compared to analogous materials studied by Tsang et al.,56 in order to achieve a lower number density of graft side chains, thereby possessing a larger sequence length between grafts and promoting higher crystallinity of the fluorous backbone. The macroinitiator [P(VDF-co-CTFE)] (Scheme 1a) was used to prepare three series of copolymers possessing different lengths of graft side chains (Scheme 1b). Graft copolymers were sulfonated to different degrees and solution cast to yield membranes with varying IEC (Scheme 1c). Four factors associated with this model polymer system are addressed: (1) the effect of the polymer architecture (ion content, graft chain length, and graft density) on the degree of crystallinity; (2) the effect of polymer architecture on the size and nature of the ionic domains; (3) the correlation of crystallinity and water sorption, and their combined effects on proton conductivity; (4) the properties of high IEC membranes on proton conductivity under conditions of low RH.



EXPERIMENTAL SECTION

Materials. Vinylidenedifluoride (VDF, Aldrich, 99+%), chlorotrifluoroethylene (CTFE, Aldrich, 98%), potassium persulfate (K2S2O8, Allied Chemical, reagent grade), sodium metabisulfite (Anachemia, anhydrous, reagent grade), pentadecafluorooctanoic acid (Aldrich, 96%), 2,2-dipyridyl (bpy, Aldrich, 99+%), 1,2-dichloroethane (DCE, Caledon, reagent grade), N-methyl-2-pyrrolidone (NMP, Aldrich, anhydrous, 99.5%), sulfuric acid (Anachemia, 95−98%, ACS reagent), 1936

dx.doi.org/10.1021/cm4005932 | Chem. Mater. 2013, 25, 1935−1946

Chemistry of Materials

Article

a N2-purged vial. The solution was cooled in a 10 wt % CaCl2 ice bath and 95−97% sulphuric acid was injected into the mixture. The resultant acetyl sulfate was immediately transferred to the polymer solution kept at 40 °C. Samples with different degrees of sulfonation were periodically extracted and precipitated in a mixture of ethanol/ hexanes (50:50 by vol.). The precipitate was washed with Millipore deionized water until the pH of the residual water was 7. The sulfonated polymers were dried under vacuum at 60 °C overnight. The degree of sulfonation DS was determined from 1H NMR spectra (see the Supporting Information, Figure S3) and calculated from the areas under the peaks according to

cuprous chloride (CuCl, Aldrich, 99%), and acetic anhydride (Aldrich, 99.5%) were used as received. Copper(II) chloride (CuCl2, Aldrich, 99.999%) was purified according to the literature.57 Styrene (St, Aldrich, 99+%) was washed with aqueous 5 wt % NaOH and water, dried overnight with MgSO4, distilled over CaH2 under reduced pressure, and stored under N2 at −20 °C. Synthesis of Fluorous Macroinitiators. The macroinitiator P(VDF-co-CTFE) was prepared by emulsion copolymerization of vinylidene difluoride (VDF) and chlorotrifluoroethylene (CTFE). A mixture of 100 mL of water, 0.40 g of KPS, 0.29 g of Na2S2O5, and 0.04 g of pentadecafluorooctanoic acid was added to a 160 mL pressure vessel (Parr Instruments) equipped with a 600 psi pressure relief valve and a magnetic stir bar.46 A VDF and CTFE monomer mixture of predetermined composition was introduced into the reactor to give a constant pressure of 300 psi at 60 °C. The polymerization was carried out for 1.0−1.5 h. The resulting polymer latex was coagulated by freezing, followed by washing with water and ethanol repeatedly. Crude polymer was purified by repeated dissolution in THF and precipitation in ethanol. The sample was dried at 80 °C under vacuum for 24 h. The composition of VDF and CTFE in the fluorous macroinitiator was determined by 19F NMR spectroscopy using a 400 MHz Varian MercuryPlus spectrometer. Chemical shifts were measured with respect to trichlorofluoromethane (CFCl3). The monomer ratio of VDF to CTFE was determined by integrating the areas of the peaks in the 19F NMR spectra according to eqs 1 and 2

S + S2 + 3S3 − S4 P(VDF) 3 = 1 × P(CTFE) S5 + S6 + 2(S4 − S3) 2

DS(%) =

1+

P(VDF) P(CTFE)

(1)

(2)

where S# is the integrated intensity of peak, # (see the Supporting Information, Figure S1). Graft-ATRP of Styrene. P(VDF-co-CTFE)-g-PS was synthesized by graft atom transfer radical polymerization. The Cl sites of the macroinitiator were used to initiate ATRP of styrene. P(VDF-coCTFE)-g-PS was synthesized by dissolving the fluorous macroinitiator P(CTFE-co-VDF) ([macroiniator] = 8.63 × 10−3 M, 1.1071 g) in Nmethylpyrrolidone (40 mL) in a dry round-bottom flask. Bipyridine ([pby] = 0.16 M, 3.0029 g), styrene ([M]o = 1.44 M, 20 mL), catalyst CuCl ([CuCl] = 5.25 × 10−2 M, 0.6430 g), and CuCl2, ([CuCl2] = 5.87 × 10−3 M, 0.0892 g) were added to the flask, and sealed with a septum and degassed by three freeze−pump−thaw cycles to remove oxygen and water. The reaction mixtures were heated in an oil bath under nitrogen at 110 °C for a total reaction time of 24 h. Polymer mixtures were collected after 8, 16, and 24 h reaction and precipitated in methanol to yield polymers with different graft lengths. Polystyrene homopolymers were removed by rinsing repeatedly with cyclohexane. The polymer precipitates were filtered and dried under vacuum at 60 °C. GPC, 19F, and 1H NMR spectra were used to determine the degree of polymerization of styrene and the molecular weights of the graft copolymers. 1H NMR spectra were recorded in d6-acetone using a 600 MHz Varian Inova spectrometer. By comparing the ratio of protons attributed to VDF and to PS units, the average degree of polymerization (DP) of the PS chain was calculated according to

DP =

water uptake =

Wwet − Wdry Wdry

× 100% (5)

where Wwet and Wdry represent the wet and dry weight of the membrane. The water content was calculated both as a mass and a volume percentage of water in the wet membrane water content (wt%) =

Wwet − Wdry Wwet

× 100%

(6)

The ion exchange capacity (IEC) was determined by titration. Membranes were first equilibrated in 2 M NaCl for at least 4 h to release the protons, which were titrated with 0.001 M NaOH to a phenolphthalein end point. After titration, the membranes were immersed in 2 M HCl to reprotonate the sulfonic sites. The membranes were dried under vacuum at 70 °C overnight. Finally, the membranes were cooled in a desiccator before their dry weights were measured. IEC (mmol/g) was calculated as

(2nVDF)(integral PS) ((5nCTFE)(%Cl reacted)(integral VDF)

(4)

where b and c represent integrals of peaks “b” and “c”, respectively. Titration was also performed to measure the ion content and to confirm successful sulfonation. Gel Permeation Chromatography. GPC was used to estimate the molecular weights of the macroinitiator and graft copolymers in DMF, 0.01 M LiBr at 80 °C. The GPC was equipped with three styragel high temperature (HT) columns with pore sizes of 1 × 103 Å, 1 × 104 Å, 1 × 105 Å (HT3, HT4, HT5, respectively) manufactured by American Polymer Standards Corporation, a Waters 1515 isocratic HPLC pump, a Waters 2414 differential refractometer, and a Waters 2487 dual UV absorbance detector (λ = 254 nm). PS standards were used for calibration. Membrane Preparation, Water Uptake, and IEC. Membranes were prepared by dissolving graft copolymers in N, N-dimethylacetamide and casting on a Teflon sheet. The films were dried at room temperature and further dried under vacuum at 60 °C overnight. The membranes were soaked in 2 M HCl overnight to convert them to their protonic form. The membranes were rinsed with Millipore deionized water several times to wash excess acid from the membranes, and stored in Millipore deionized water. The thickness of the membranes was ∼100 μm. The water uptake was calculated as the percentage increase in mass over the dry weight of the membrane

1

%CTFE =

c /2 × 100% c /2 + b/3

IEC = (3)

VNaOHMNaOH Wdry

(7)

where VNaOH and MNaOH are the volume (mL) and molar concentration (mol/L) of NaOH solution, respectively. Water uptake, water content, and IEC values were calculated from the average of five membrane samples. Errors were calculated based on the standard deviation of experiments. The number of water molecules per sulfonic acid group (λ) was calculated according to

where nVDF and nCTFE refer to the units of VDF and CTFE in P(VDFco-CTFE). Integral signals of PS and VDF are shown in Supporting Information, Figure S2, and are labeled peak “d”, “e”, and “a”, respectively. Sulfonation of PS Grafts. Sulfonation was carried out in 1,2dichloroethane at 40 °C. The procedure was as follows.58 To a 50 mL three-neck flask equipped with a condenser were added 15 mL of 1,2dichloroethane and 0.6 g of P(VDF-co-CTFE)-g-PS, and the mixture was heated to 50 °C under N2 and stirred. Acetyl sulfate was prepared by injecting 1 mL of acetic anhydride and 3 mL of dichloroethane into

λ= 1937

moles H 2O water uptake· 10 = moles SO3H 18·IEC

(8)

dx.doi.org/10.1021/cm4005932 | Chem. Mater. 2013, 25, 1935−1946

Chemistry of Materials

Article

Table 1. Composition of Graft Copolymers: P(VDF-co-CTFE2.6 mol %)-g-PS ATRP reaction time (h)

% CTFE initiateda

graft number density (mol %)b

DPPSc

St/VDFd (mole ratio)

Mn, PSe × 10−5 [g/mol]

8 16 24

65 ± 1 65 ± 1 67 ± 1

1.68 ± 0.02 1.68 ± 0.02 1.72 ± 0.03

39 ± 3 62 ± 2 79 ± 4

67/100 104/100 136/100

1.29 2.06 2.66

Mn,

P(VDF‑co‑CTFE)‑g‑PS

[g/mol]

× 10−5

2.52 3.29 3.89

P(VDF-co-CTFE2.6 mol %) macroinitiator consisted of 2.6 mol % CTFE, estimated from 19F NMR. Mn, P(VDF‑co‑CTFE2.6 mol %),GPC = 1.23 × 105 g/mol, measured by GPC and calibrated with linear PS standards. aBased on 19F NMR. bGraft side chains per 100 units in P(VDF-co-CTFE) backbone. c Styrenes per graft chain, determined by the ratio of protons in VDF units and styrenic units using 1H NMR. dMoles of styrene divided by moles of VDF, obtained from both 1H and 19F NMR. eCalculated from DPPS. Proton Conductivity. Membranes were placed across Pt electrodes in a Teflon probe. Their in-plane conductivity was measured by AC impedance spectroscopy using a Solartron 1260 frequency analyzer, Nyquist plots were obtained by applying a frequency from 10 MHz to 100 Hz at 100 mV. The membrane resistance R (Ω) was determined by fitting the data to the standard Randles equivalent circuit. The proton conductivity σ (S/cm) was calculated from R and the sample geometry as

L σ= RA

times of 2 min were typically sufficient to achieve an adequate signalto-noise ratio. The degree of crystallinity in the sample xcr was determined by calculating the ratio of scattering due to the crystalline domains to the total scattering intensity60 ∞

xcr =

Wdry Vwet

IEC

t=

∫0 Itotal(q)q dq

=

Icr Icr + Iam

(12)

0.9λ Bcos θB

(13)



RESULTS AND DISCUSSION a. Effect of Graft Chain Length. Synthesis of P(VDF-coCTFE). CTFE was copolymerized with VDF to yield a

(10)

σ F[ − SO3H]

2

where t represents the grain size, λ is the wavelength of the radiation, B is the fwhm in radians, and θB is the position of the crystalline peak.

where Vwet is the volume of the wet membrane (cm3). The effective proton mobility μeff (cm2/(V s)) in the wet membrane was estimated from the conductivity and the analytical acid concentration

μeff =



where Icr and Iam are the integrated values of the PVDF crystalline peak and the sum of the PS and PVDF amorphous peaks, respectively, and q = 4π/λ sin(θ) is the scattering wave vector in which 2θ is the scattering angle and λ is the wavelength of the incident beam. The average crystalline domain sizes was calculated according to the Scherrer equation61,62

(9)

where L (cm) is the length of the conductor, in this case the distance between the two electrodes and A (cm2) is the cross-sectional area of the membrane. The width of the membrane was measured with a Mitutoyo Digimatic caliper and the thickness was measured with a Series 293 Mitutoyo Quickmike caliper. Membranes were recast and their proton conductivities measured several times. Proton conductivity was calculated as the average of several measurements. The analytical acid concentration [−SO3H] (M) in wet membranes was determined using IEC values measured by titration,

[ − SO3H] =

∫0 Icr(q)q2dq

(11) 59

where F is the Faraday constant. For measurements under controlled RH, membranes were equilibrated at a predetermined RH between 95 and 55% RH in an ESPEC SH-241 environmental chamber at 25 °C. The membranes were also equilibrated at a predetermined temperature between 25 to 80 °C at 95% RH. Measurements of conductivity were taken inside the environmental chamber until 3 or more conductivity values were consistent over a 2 h period, this usually took ∼5 h. Gravimetric water vapor sorption techniques were used to obtain water vapor sorption at different RH (DVS-1000, Surface Measurement Systems, U.K.). Transmission Electron Microscopy. The membranes were stained by immersion overnight in a saturated lead acetate solution. The membranes were rinsed with water and dried under vacuum at room temperature, embedded in Spurr’s epoxy resin, and sectioned to produce 60−100 nm thick slices using a Leica UC6 microtome. These were collected on a copper grid and electron micrographs taken with a Hitachi H7600 TEM by applying an accelerating voltage of 100 kV. The domain sizes were measured according to the scales given on each TEM image using software Image J. Approximately 150 measurements were performed in order to obtain an average value of ionic domain size for each copolymer. Wide-Angle X-ray Scattering. Wide-angle diffraction measurements were performed on a Rigaku Rapid Access XRD with an image plate detector. The X-ray unit was operated at 46 kV and 42 mA using a nickel filter. The X-ray wavelength was 0.154 nm. Data was acquired in transmission mode under ambient conditions. Data acquisition

Figure 1. WAXS patterns of the P(VDF-co-CTFE2.6 mol %) and P(VDF-co-CTFE2.6 mol %)-g-PS79 series.

perfluoroalkane macroinitiator containing chlorine sites suitable for graft-ATRP. According to previously published methods,63 the monomer ratio of VDF to CTFE was determined by integrating areas of peaks in the 19F NMR spectra (see the Supporting Information, Figure S1). The macroinitiator was estimated to contain 2.6 ± 0.3 mol % CTFE and 97.4 ± 0.3 mol % VDF. 1938

dx.doi.org/10.1021/cm4005932 | Chem. Mater. 2013, 25, 1935−1946

Chemistry of Materials

Article

Figure 2. Degree of crystallinity for P(VDF-co-CTFE2.6 mol %) and P(VDF-co-CTFE2.6 mol %)-g-PS copolymers as a function of IEC.

Figure 4. Diameter and number density of the ion-rich domains vs DS.

atoms (−CF2−CF*[CH2CH(C6H5)]−CF2−) juxtapositioned to styrenic groups (see Figure S1 in the Supporting Information). A considerable reduction in intensity of the peaks at approximately −120 to −122 ppm (due to −CF2− CF2−CF*Cl-CH2−CF2) is also observed in the 19F NMR spectra of the graft copolymers with respect to the macroinitiator due to CFCl units being consumed during ATRP. By measuring the percentage reduction in the peaks at −120 to −122 ppm using peaks at −89.0 to −94.0 ppm (−CF2−CH2− CF2*−CH2−CF2-) as a reference, the % Cl sites initiated was estimated to be 65, 65, and 67 for the three graft copolymers subjected to 8, 16, and 24 h ATRP reaction time, respectively. From the mol % of CTFE in the macroinitiator (2.6 mol %) and the percentage of CTFE initiated, the graft number density of the copolymers was estimated to be 1.69 ± 0.03 mol %. Mn of the fluorous macroinitiator was ∼1.23 × 105 g/mol, and the PDI was 1.59 as estimated by GPC. GPC results demonstrate growth of the polystyrene graft chains with increasing ATRP reaction time. The graft copolymers exhibit a significantly different hydrodynamic volume than the linear polystyrene standards, thus a more accurate determination of molecular weight was obtained by NMR analysis. It was calculated that the macroinitiator contained 1828 VDF units and 49 CTFE units. By comparing the ratio of protons attributed to VDF and to polystyrenic units in the 1H NMR spectra, the average degree of polymerization of PS (DPPS) was calculated to be 39, 62, and 79 per graft chain for ATRP reaction times of 8, 16, and 24 h, respectively. These three parent graft copolymers are termed P(VDF-co-CTFE2.6 mol %)-gPS 39 , P(VDF-co-CTFE 2.6 mol % )-g-PS 62 , and P(VDF-coCTFE2.6 mol %)-g-PS79, respectively. The St/VDF (mol %) ratio was calculated using the total moles of styrene divided by total moles of VDF. Mn of P(VDF-co-CTFE2.6 mol %)-g-PS39, P(VDF-co-CTFE2.6 mol %)-g-PS62, and P(VDF-coCTFE2.6 mol %)-g-PS79 were calculated to be 2.53 × 105, 3.30

Figure 3. TEM images of P(VDF-co-CTFE2.6 mol %)-g-SPS membranes. P(VDF-co-CTFE2.6 mol %)-g-SPS39 (A) IEC = 1.12 mmol/g, DS = 18%; (B) IEC = 3.52 mmol/g, DS = 100%. P(VDF-co-CTFE2.6 mol %)-gSPS62 (C) IEC = 1.23 mmol/g, DS = 19%; (D) IEC = 4.05 mmol/g, DS = 100%. P(VDF-co-CTFE2.6 mol %)-g-SPS79 (E) IEC = 1.35 mmol/ g, DS = 23%; (F) IEC = 4.29 mmol/g, DS = 100%.

ATRP of Styrene. 19F and 1H NMR spectra of the macroinitiator and a typical graft copolymer are shown in the Supporting Information (Figure S1 and Figure S2). Both spectra provide evidence for the growth of polystyrene from the fluorous backbone: 1H NMR spectra of the graft copolymer show additional peaks at 6.40−7.40 ppm, corresponding to polystyrene (see Figure S2 in the Supporting Information). 19F NMR spectra of the graft copolymers exhibit similar signature peaks corresponding to the macroinitiator as well as an additional peak at −164 ppm corresponding to tertiary fluorine 1939

dx.doi.org/10.1021/cm4005932 | Chem. Mater. 2013, 25, 1935−1946

Chemistry of Materials

Article

Figure 5. (a) Water uptake, (b) proton conductivity, (c) analytical acid concentration, and (d) effective proton mobility as a function of IEC for: P(VDF-co-CTFE2.6 mol %)-g-PS39; P(VDF-co-CTFE2.6 mol %)-g-PS62, P(VDF-co-CTFE2.6 mol %)-g-PS79, and Nafion 117.

Figure 6. (a) Proton conductivity and (b) effective proton mobility as a function of λ for: P(VDF-co-CTFE2.6 mol %)-g-PS39; P(VDF-coCTFE2.6 mol %)-g-PS62, P(VDF-co-CTFE2.6 mol %)-g-PS79, and Nafion 117.

× 105, and 3.90 × 105 g/mol, respectively. Chemical compositions are summarized in Table 1. Sulfonation of Side Chains. 1H NMR spectroscopy was used to confirm sulfonation of the side chains and to determine the degree of sulfonation (DS), as illustrated by the 1H NMR spectra of the pristine graft copolymer P(VDF-coCTFE2.6 mol %)-g-PS79 and partially sulfonated graft copolymers (see Supporting Information, Figure S3). The pristine copolymer exhibits peaks at 6.5−6.8 ppm (peak “a”) and 6.9−7.4 ppm (peak “b”) corresponding to ortho- and meta/ para-protons of the phenyl ring, respectively. The partially sulfonated copolymers exhibit an additional peak at 7.6 ppm, which corresponds to protons adjacent to the sulfonated group (peak “c”). A greater intensity of peak “c” represents a higher DS in the copolymer. The degree of sulfonation for each

copolymer is summarized in the Supporting Information (Table S1). IEC measured by titration was compared to that derived from NMR data. A plot of IEC as a function of DS is shown in the Supporting Information, Figure S4. The titration and NMR −derived IEC data show a slight deviation, with the titration-IEC data being up to 10% less. However, this is construed as being within experimental error. It demonstrates that the acid groups in the membrane are accessible to water. Crystallinity. Wide-angle X-ray scattering (WAXS) was performed to measure the degree of crystallinity in the graft membranes. WAXS patterns for the macroinitiator, and unsulfonated P(VDF-co-CTFE2.6 mol %)-g-PS79 and its selected sulfonated copolymers are shown in Figure 1. WAXS patterns for P(VDF-co-CTFE 2 . 6 m o l % )-g-PS 3 9 and P(VDF-coCTFE2.6 mol %)-g-PS62 are shown in the Supporting Information 1940

dx.doi.org/10.1021/cm4005932 | Chem. Mater. 2013, 25, 1935−1946

Chemistry of Materials

Article

Scheme 2. Schematic Representation of Graft Copolymers Possessing Similar Weight Fraction of Polystyrene but Different Graft Number Density

Table 2. Chemical Composition of P(VDF-co-CTFE2.6 mol %)g-PS39 and P(VDF-co-CTFE5.8 mol %)-g-PS35 CTFE contenta (mol %)

graft densityb (mol %)

DPPSc

Mn, PSd × 10−5 [g/mol]

weight fraction of PSe(%)

2.6 5.8

1.7 2.3

39 35

1.29 3.64

51 54

a Calculated on the basis of 19F NMR. bNumber of PS graft side chains per 100 units of fluorous backbone, calculated from the mol % of CTFE content in fluorous backbone multiplied by the % of CTFE reacted during ATRP. cNumber of styrenes per graft chain. d Calculated from DPPS. eMn,PS/ Mn, total.

Figure 7. (a) Water content: lambda (λ) and (b) proton conductivity as function of RH at 25 °C. Samples were equilibrated under water vapor.

Figure 8. Proton conductivity versus temperature at 95% RH.

(Figure S5). The scattered intensity is shown as a function of the scattering angle 2θ (top axis) and the scattering wave vector q = 4π/λ sin(θ) (bottom axis), where λ = 0.154 nm. Four peaks are observed for each copolymer. The peak positioned at 2θ = 10° is associated with PS chain-to-chain order,64 and decreases as DS increases. This behavior is consistent with our previous work.56 The peaks observed at 2θ = 19 and 20° are due to amorphous and crystalline domains, respectively.65,66 A very broad fourth peak positioned around 2θ = 27° is visible for some samples and is due to the presence of water. Three Gaussians peaks were fit to the data in the q-range of 0.45−1.6 Å−1 (see the Supporting Information, Figure S6, for example). The degree of crystallinity of the fluorous macroinitiator (VDF-co-CTFE2.6 mol %)-g-PS and representative graft copolymers was calculated using eq 12, plotted in Figure 2, and

Figure 9. TEM images of selected P(VDF-co-CTFE)-g-SPS graft membranes. P(VDF-co-CTFE2.6 mol%)-g-SPS39 (A) IEC = 1.12 mmol/ g, DS = 18%. (B) IEC = 3.02 mmol/g, DS = 70%. P(VDF-coCTFE5.8 mol %)-g-SPS35 (C) IEC = 0.64 mmol/g, DS = 13%. (D) IEC = 2.48 mmol/g, DS = 59%.

tabulated in the Supporting Information, Table S2. A much lower degree of crystallinity ∼6.7, 6.4, and 5.6% is observed for P(VDF-co-CTFE2.6 mol %)-g-PS39, P(VDF-co-CTFE2.6 mol %)-gPS62, and P(VDF-co-CTFE2.6 mol %)-g-PS79, respectively, compared to the fluorous macroinitiator P(VDF-co-CTFE2.6 mol %) (24%). The introduction of graft side chains leads to lower 1941

dx.doi.org/10.1021/cm4005932 | Chem. Mater. 2013, 25, 1935−1946

Chemistry of Materials

Article

size is determined. These TEM images are similar to Nafion membranes which have been reported to possess 4 − 10 nm diameter ionic clusters,35,40,67−69 but dissimilar to ionic fluorous diblock copolymers of similar ionic to fluorous ratio, which exhibit lamellar-like morphologies.46,70−73 The average diameter and number density of the ion-rich domains are summarized in the Supporting Information, Table S3. The domain size and number density of the ion-rich domains vs DS are plotted in panels a and b in Figure 4, respectively, to demonstrate the trend. For a given series, the ion-rich domain size increases as DS increase. For example, in the P(VDF-co-CTFE2.6 mol %)-g-SPS39 series, the membrane having DS 18% possesses 2.5 ± 0.4 nm diameter ion-rich domains, whereas the completely sulfonated membrane possesses 3.8 ± 0.6 nm ion-rich domains. As DS is decreased, the sulfonic acid groups are separated to greater and greater extents by hydrophobic unsulfonated PS. Thus ionic aggregation, stabilized by the proximity of ion pairs and electrostatic forces,74 and influenced by the chain elasticity of the host polymer,75 is hindered by the need to expel hydrophobic PS. The result is that a lower DS confers smaller ionic aggregates. Moreover, DS also affects the number density of the ion-rich domains in these graft samples. For a given series, the number density of the ion-rich domains decreases as DS increases. We postulate that as DS is increased, the closer proximity of sulfonic acid groups along the PS side chain promotes ionic aggregation by stronger electrostatic interactions between ion pairs, leading to slightly larger, but fewer ion-rich domains. To promote meaningful comparison, we prepared the membranes chosen for illustration in Figure 3 from copolymers of different graft length but comparable DS, thus aiding the examination of the effect of graft length on morphology. At

Table 3. Ionic Cluster Size and 2D Cluster Number Density for P(VDF-co-CTFE2.6 mol %)-g-SPS39 and P(VDF-coCTFE5.8 mol %)-g-SPS35 image label

IEC (mmol/g)

A B

1.12 2.27

C D

0.64 2.48

DS (%)

ionic domain diameter (nm)

2D cluster number density (per 1 × 104 nm2)

P(VDF-co-CTFE2.6 mol %)-g-SPS39 18 2.5 ± 0.4 86 ± 5 53 3.5 ± 0.6 76 ± 2 P(VDF-co-CTFE5.8 mol %)-g-SPS35 13 2.2 ± 0.4 210 ± 20 59 3.3 ± 0.4 190 ± 20

degree of crystallinity (%) 12.6 ± 1.5 12.5 ± 0.6