How the presence of electric fields alters the miscibility of mixtures has been studied since the 1960s with conflicting reports on both the magnitude and direction of the shift in the phase separation temperature Ts. Theoretical understanding of the phenomenon has been hampered by the lack of experimental data with unambiguously large shifts in Ts outside of experimental error. Here, we address these concerns by presenting data showing that uniform electric fields strongly enhance the miscibility of polystyrene (PS)/poly(vinyl methyl ether) (PVME) blends. Reliable shifts in Ts of up to 13.5 ± 1.4 K were measured for electric fields strengths of E = 1.7 × 107 V/m in a 50/50 PS/PVME mixture. By using a sensitive fluorescence method to measure Ts, the PS/PVME blend can be quenched back into the one phase region of the phase diagram when the domains are still small allowing the blend to be remixed such that Ts can be measured repeatedly on the same sample. In this manner, highly reproducible Ts values at non-zero and zero electric field can be ascertained on the same sample. Our results agree with the vast majority of existing experimental data on various mixtures finding that electric fields enhance miscibility, but are opposite to the one previous study on PS/PVME blends by Reich and Gordon [J. Polym. Sci.: Polym. Phys. Ed.17, 371 (1979)] reporting that electric fields induce phase separation, a study which has been considered anomalous in the field.

Mixtures of different polymers are extensively utilized to achieve blended material properties not available in single component systems. However, the naturally poor miscibility of polymers stemming from their macromolecular character leave the phase behavior of polymer blends extremely sensitive to enthalpic interactions and external perturbations. Techniques which can externally control and manipulate the phase behavior of miscible systems, without altering chemistry on a molecular level, have great practical benefits. One such possible mechanism is the use of electric fields with isolated reports showing that the presence of uniform fields can cause shifts in the phase separation temperature Ts for various mixtures.1–11 However, at present, there is extensive debate2 and limited understanding of how uniform electric fields influence the compatibility of polymeric mixtures, or even small molecules, with one of the main limitations stemming from the lack of experimental data with unambiguously large shifts in Ts outside of experimental error.

Here, we present results demonstrating that the presence of uniform electric fields strongly enhance the miscibility of polystyrene (PS)/poly(vinyl methyl ether) (PVME) blends. This work follows from our previous study12 in which we presented a fluorescence technique for measuring the phase separation temperature Ts of PS/PVME blends. In this method, an aromatic fluorophore is covalently bonded to the non-polar PS component such that upon phase separation a sharp increase in fluorescence intensity is observed as the dye separates from the more polar PVME component. This sensitivity arises from local fluorescence quenching of the dye when in the proximity of PVME,13 likely the result of a weak hydrogen bond forming between the ether oxygen of PVME and the aromatic hydrogens of the dye.12 A similar such weak hydrogen bond between PVME and PS is believed to be responsible for making this blend miscible.14 In the present work, this fluorescence technique is used to investigate the change in Ts due to the presence of electric fields. We show that the presence of uniform electric fields strongly enhances the miscibility of PS/PVME polymer blends. Focusing on a 50/50 PS/PVME blend composition, Ts is found to increase by over 13 K for electric field strengths of 1.7 × 107 V/m. These results contradict the one previous report on PS/PVME blends by Reich and Gordon,4 but agree with the vast majority of studies on how electric fields affect the miscibility of mixtures.1–3,5–9 We discuss possible reasons for this discrepancy.

Only a few experimental results have been published over the past several decades, with no clear consensus in the size of the shift in phase separation temperature Ts, or even whether electric fields consistently enhance mixing or induce phase separation. The first experimental results on the subject were published by Debye and Kleboth,1 who studied a small molecule mixture of nitrobenzene/isooctane. They reported that electric fields of E = 0.45 × 107 V/m enhance mixing, causing Ts to shift up by 0.015 K for this upper critical solution temperature (UCST) type mixture. Later, Debye and Kleboth's results were verified with greater accuracy by Orzechowski8 for the same mixture. Beaglehole5 reported shifts of ΔTs = 0.08 K towards enhanced mixing under fields of E = 0.03 × 107 V/m in solutions of cyclohexane/aniline. However, Early,6 who later worked on the same mixture with similar magnitude of electric fields as Beaglehole, reported not seeing shifts in Ts at all. Early suggested Beaglehole's results could be explained by Joule heating from current conduction through the sample. Similar concerns have also been expressed about the Debye and Kleboth results.6,15

Wirtz and Fuller7 studied three different solutions: small molecule mixtures of nitrobenzene/n-hexane, and polymeric solutions of PS/cyclohexane and poly(p-chlorostyrene)/ethylcarbitol. They saw enhanced mixing in all three systems, reporting shifts of ΔTs = 0.02 K for E = 0.1 × 107 V/m in nitrobenzene/n-hexane, ΔTs = 0.04 K for E = 0.05 × 107 V/m in PS/cyclohexane, and ΔTs = 0.03 K for E = 0.05 × 107 V/m in poly(p-chlorostyrene)/ethylcarbitol. Much bigger shifts, up to ΔTs = 1.5 K for E = 0.85 × 107 V/m, have also been recently reported in poly(styrene-block-isoprene) (SI) solutions by Schoberth et al.9 towards enhanced mixing.

All the above mentioned experimental data show enhanced compatibility, lower UCST or higher lower critical solution temperature (LCST) behavior, in the presence of uniform electric fields. However, there are two research groups who have reported the opposite. According to Reich and Gordon,4 who reported the largest shifts with ΔTs = 54 K for E = 2.72 × 107 V/m in PS/PVME polymer blends, and Lee et al.,10,11 who also showed large shifts of ΔTs = 18 K for E = 0.9 × 107 V/m in poly(vinylidene fluoride) (PVDF)/poly(butyl acrylate) (PBA) and ΔTs = 2.5 K for E = 0.7 × 107 V/m in PVDF/poly(methyl methacrylate) (PMMA), electric fields strongly reduce compatibility. Note, however, these last two blend systems by Lee et al.10,11 involving PVDF are unique in that PVDF is strongly piezoelectric. PVDF, having a negative piezoelectric coefficient, undergoes volume contraction when in the presence of electric fields.16,17 As blend miscibility and Ts are very sensitive to many perturbations such as shear, pressure, temperature, etc.,18 it would be difficult to distinguish between the effects of mechanical stress in PVDF due to the presence of electric fields and other possible electric field effects leading to shifts in Ts. Volume changes on blending have historically been particularly challenging to account for theoretically.19 For example, White and Lipson recently showed that individual component volume changes associated with thermal expansion is a strong factor affecting blend miscibility.20 Thus, for the present discussion we exclude consideration of these piezoelectric PVDF systems. This leaves only one outlying study, by Reich and Gordon on PS/PVME blends,4 not reporting that the presence of electric fields enhance miscibility.

Comparing solely the absolute maximum magnitude of the Ts shifts observed in the experiments listed above, there appears, at first glance, to be a trend of electric fields causing smaller Ts shifts in small molecule systems and larger shifts in polymeric systems. However, when the relative change of ΔTs/E2 between published results is compared, this trend disappears. The one agreement between all experimental results, and theoretical approaches, is that the shift in Ts is proportional to the square of electric field, E2. Comparing relative ΔTs/E2 changes among the available experimental literature makes it clear that no correlation exists between the relative size of the shift in Ts as a function of electric field and the system molecule size. The Debye and Kleboth1 and Orzechowski8 studies, both in solutions of nitrobenzene/isooctane, showed ΔTs/E2 = 0.08 × 10−14 Km2/V2, which is almost 1000 times smaller than ΔTs/E2 = 87 × 10−14 Km2/V2 reported by Beaglehole5 in cyclohexane/aniline. For Wirtz and Fuller's data,7 ΔTs/E2 = 2 × 10−14 Km2/V2 in nitrobenzene/n-hexane, ΔTs/E2 = 16 × 10−14 Km2/V2 in PS/cyclohexane, and ΔTs/E2 = 12 × 10−14 Km2/V2 in poly(p-chlorostyrene)/ethylcarbitol. For the Schoberth et al.9 data in SI solutions, ΔTs/E2 = 2.1 × 10−14 Km2/V2, while the Reich and Gordon4 data in PS/PVME give ΔTs/E2 = 7.3 × 10−14 Km2/V2. And included simply for completeness, the Lee et al.10,11 studies give ΔTs/E2 = 22 × 10−14 Km2/V2 in PVDF/PBA and ΔTs/E2 = 4.6 × 10−14 Km2/V2 in PVDF/PMMA. Hence, there appears to be little if any discernable trend among the different systems with the ΔTs/E2 shifts varying from 0.08 to 87 (×10−14 Km2/V2).

More confusion arises when considering theoretical expectations of electric field effects on the phase separation temperature Ts. The typical theoretical approach uses a thermodynamic argument of adding an electrostatic free energy term to the free energy of mixing,1–3,5,7,10,11,15,21 for polymers this is typically written as an extension of the classic Flory-Huggins equation:2,7,10,11

\begin{eqnarray}\Delta G_m &=& k_B T\left\{ {\frac{{\phi _A \ln \phi _A }}{{v_A N_A }} + \frac{{\phi _B \ln \phi _B }}{{v_B N_B }} + \frac{{\phi _A \phi _B \chi }}{{v_{ref} }}} \right\}\nonumber\\&& - \frac{1}{2}\varepsilon _0 \varepsilon (\phi)E^2,\end{eqnarray}
ΔGm=kBTϕAlnϕAvANA+ϕBlnϕBvBNB+ϕAϕBχvref12ɛ0ɛ(ϕ)E2,
(1)

where ΔGm is the system's total free energy of mixing per unit volume. The term within brackets on the right side is the sum of the entropy and enthalpy of mixing per unit volume, where ϕA and ϕB are volume fractions and νA and νB are monomer volumes of the two blend components with degree of polymerization NA and NB; while νref is a reference volume of monomer size and χ is the empirical interaction parameter.22,23 In polymers, the interaction parameter χ represents a monomer-monomer interaction because flexible polymer chains are intertwined molecularly at the monomer level. For example, criticality occurs at χcN = 2 for a 50/50 polymer blend with equal chain lengths NA = NB = N, instead of the typical χc = 2 for a small molecule mixture (N = 1). In Eq. (1), νA NA and νB NB represent the molecular volume of each polymer, with νref accounting for the fact that νA and νB are likely not the same size. The last term in Eq. (1) is the additional contribution to ΔGm due to the presence of a uniform electric field E, the free electrostatic energy density of the system, where ɛ0 is the absolute permittivity of vacuum and ɛ(ϕ) is the composition dependent dielectric constant.24,25 In Eq. (1), we include a negative sign in front of the electrostatic energy term, as this is the most accepted treatment;2,3,5,10,11,15 however, there exists discussion among theoretical studies, whether a plus sign should be used instead.1–3,7,15 The negative sign is applicable to dielectrics for the case of a constant applied potential (our present experiments), where the work done by an external power supply to maintain a constant voltage is included in the total free energy of the system.2,24,25 When a positive sign is used, the sample (a capacitor setup with dielectric medium) is considered to have a constant charge on its plates, such that the external voltage supply is not included in the total free energy.

Considering the standard conditions for stability |$\frac{{\partial ^2 \Delta G_m }}{{\partial \phi ^2 }} = 0$|2ΔGmϕ2=0 and criticality |$\frac{{\partial ^3 \Delta G_m }}{{\partial \phi ^3 }} = 0$|3ΔGmϕ3=0, and the empirical form for the interaction parameter |$\chi = A + \frac{B}{T}$|χ=A+BT (where empirical parameters A > 0 and B < 0 for LCST and A < 0 and B > 0 for UCST type phase diagrams), it follows from Eq. (1) that

\begin{equation}\frac{{T_s (E) - T_s (0)}}{{T_s (0)}} = \frac{{\varepsilon _0 v_{ref} E^2 }}{{4kB}}\frac{{\partial ^2 \varepsilon (\phi)}}{{\partial ^2 \phi }}.\end{equation}
Ts(E)Ts(0)Ts(0)=ɛ0vrefE24kB2ɛ(ϕ)2ϕ.
(2)

Here, Ts(E) and Ts(0) are the phase separation temperatures measured with and without electric field defining ΔTs(E) = Ts(E) – Ts(0) as the shift in phase separation temperature due to the electric field. As written, this is the most commonly used form, however, formally it is only correct to first order.2 As we describe in more detail in the discussion, recent work by Orzechowski et al.26 has argued that the next higher order term may become dominant at high field strengths, demonstrating quantitative agreement between theory and experiment for electric field shifts in the nitrobenzene/n-octane system. Assuming that |$\frac{{\partial ^2 \varepsilon (\phi)}}{{\partial ^2 \phi }} > 0$|2ɛ(ϕ)2ϕ>0, as is typically seen for mixtures of polar and non-polar components,1,3,5,8 and using the negative sign for the electrostatic energy term in Eq. (1), it follows from Eq. (2) that a shift towards a reduction of compatibility on the order of only a few mK is predicted for the presence of moderate to strong fields.2,3,5,15,21 Thus overall, theoretical predictions are in conflict with the vast majority of experimental results.1,2,4–11,15 As discussed above, all experimental results, except the single study by Reich and Gordon4 (and excluding the piezoelectric PVDF data by Lee et al.10,11), find that electric fields enhance mixing with significantly larger shifts in Ts than theory predicts.

Polystyrene (unlabeled), with molecular weight of Mw = 101.3 kg/mol and polydispersity index of Mw/Mn = 1.04, was purchased from Scientific Polymer Products, and used as the matrix (neat) PS for all studies. Pyrene-labeled PS (designated as PS* in Fig. 1(a)) was synthesized by free radical polymerization from styrene in the presence of trace levels of 1-pyrenylmethyl methacrylate, as described in our previous publication.12 The fluorescent monomer 1-pyrenylmethyl methacrylate was purchased from Polysciences and used as received. PS* of Mw = 86.8 kg/mol and Mw/Mn = 1.65, with a label content of 1.93 mol.%, was used for the electric field measurements, while the initial zero-field measurements demonstrating the protocol used to remix the blend (Fig. 2) used a PS* of Mw = 76.7 kg/mol and Mw/Mn = 1.70, with label content of 0.33 mol.%. No significant change in remixing time was observed for this small change in molecular weight. PVME (shown in Fig. 1(b)) was purchased from Scientific Polymer Products and washed prior to use by dissolving in toluene and precipitating into heptane 9 times. The resulting PVME used had a molecular weight of Mw = 80 kg/mol, Mw/Mn = 2.5. Molecular weights were determined by gel permeation chromatography done with tetrahydrofuran (THF) as eluent relative to PS standards, and PVME values determined using universal calibration with Mark-Houwink parameters a = 0.739 and k = 13.5 × 10−3 ml/g.27 Fluorophore label content was measured using UV-visible absorbance spectroscopy in high-performance liquid chromatography (HPLC) grade THF.

FIG. 1.

Chemical structure of (a) pyrene-labeled polystyrene (designated as PS*) and (b) poly(vinyl methyl ether) (PVME). Schematic of the sample geometry, (c) top view and (d) side view, placing the PS/PVME blend within a parallel-plate capacitor formed by two ITO-coated quartz slides. A 25 μm thick Kapton spacer is used to define the blend thickness and isolate the edges of the samples from dielectric breakdown.

FIG. 1.

Chemical structure of (a) pyrene-labeled polystyrene (designated as PS*) and (b) poly(vinyl methyl ether) (PVME). Schematic of the sample geometry, (c) top view and (d) side view, placing the PS/PVME blend within a parallel-plate capacitor formed by two ITO-coated quartz slides. A 25 μm thick Kapton spacer is used to define the blend thickness and isolate the edges of the samples from dielectric breakdown.

Close modal
FIG. 2.

Fluorescence intensity as a function of temperature for a (40/10)/50 (PS/PS*)/PVME blend measured on heating at 1 K/min. (a) Curves (0)–(6) are collected one after another within the same sample, where between each measurement the blend is quenched back into the one phase region and remixed (see text for details). Panel (b) is the data from panel (a) with the curves (0)–(6) vertically shifted for visual clarity. A short vertical black bar denotes the measured phase separation temperature Ts, identified as the intersection of two linear fits to the data above and below phase separation,12 as illustrated on curve (2): Ts(0) = 102.4 °C, Ts(1) = 107.7 °C, Ts(2) = 108.1 °C, Ts(3) = 107.7 °C, Ts(5) = 106.9 °C, and Ts(6) = 106.5 °C.

FIG. 2.

Fluorescence intensity as a function of temperature for a (40/10)/50 (PS/PS*)/PVME blend measured on heating at 1 K/min. (a) Curves (0)–(6) are collected one after another within the same sample, where between each measurement the blend is quenched back into the one phase region and remixed (see text for details). Panel (b) is the data from panel (a) with the curves (0)–(6) vertically shifted for visual clarity. A short vertical black bar denotes the measured phase separation temperature Ts, identified as the intersection of two linear fits to the data above and below phase separation,12 as illustrated on curve (2): Ts(0) = 102.4 °C, Ts(1) = 107.7 °C, Ts(2) = 108.1 °C, Ts(3) = 107.7 °C, Ts(5) = 106.9 °C, and Ts(6) = 106.5 °C.

Close modal

Polymer blends were prepared by dissolving both PS and PVME in toluene to produce solutions with 18 wt.% total polymer content. Blend compositions of 50/50 PS/PVME were made by combining 40 wt.% neat PS, 10 wt.% pyrene-labeled PS*, and 50 wt.% PVME, that is (40/10)/50 (PS/PS*)/PVME, producing samples with <0.2 mol.% total fluorescent dye content. Solutions were poured into the center of a wide glass dish with flat bottom and dried in a fume hood at room temperature for 24 h, followed by annealing in a vacuum oven at 80 °C for at least 24 h to remove residual solvent. Disks were cut from the dry films having masses of 5.8 ± 0.2 mg and diameters slightly smaller than 5/8 in. These disks were placed between two 1 × 1 in. quartz slides (SPI Supplies) coated with conducting indium tin oxide (ITO) as electrodes. To prevent dielectric breakdown in the air around the edges of the sample, as the dielectric strength of air is approximately an order of magnitude less than that of the polymer, a 25 μm (0.001 in. nominal) thick Kapton sheet was used as a spacer between the ITO electrodes, which extended laterally beyond the ITO electrodes. A 5/8 in. diameter hole was punched in the Kapton spacer to accommodate the polymer blend sample. Figures 1(c) and 1(d) give a schematic of the sample geometry. Once assembled, the samples were placed under vacuum at 80 °C and pressed with a 2.2 kg weight for 3–5 days, until the polymer blend filled in the form of the Kapton hole. This process achieved uniform film thicknesses of 24 ± 2 μm on average across the polymer blend dictated by the thickness of the Kapton sheet. The thickness of each sample used for electric field measurements was determined (to within ±2 μm) by using an optical microscope (Leica DMIRB inverted microscope) to focus on the top and bottom of the optically transparent samples and recording the micrometer scale of the microscope. Finally, electrical wires were glued to the corners of the ITO layers using silver paste in order to apply constant voltage to the sample using a DC Agilent Technologies N5752A high-voltage power supply, up to 440 VDC to achieve electric field strengths of 1.8 × 107 V/m. All samples were stored under vacuum at room temperature prior to measurement to avoid moisture uptake.

Steady-state fluorescence measurements were carried out using a Photon Technology International QuantaMaster fluorimeter. Sample temperature was controlled to an accuracy of ±0.3 K using an Instec HCS402 heater. Fluorescence measurements used a front-face geometry with an angle of incidence of 28°. Excitation and emission slits were set to 4.00 nm (ex) and 4.25 nm (em) bandpass, with an excitation wavelength of 324 nm. During the time-based measurements, the samples were heated at a rate of 1 K/min, while simultaneously measuring the fluorescence intensity during a 3 s time window every 30 s, at a wavelength of 379 nm. This procedure follows the method we have previously established for measuring the phase separation temperature Ts of PS/PVME blends under zero electric field.12 After each measurement, samples were quenched to below the phase separation temperature at the rate of 40 K/min using the liquid nitrogen cooling capability of the Instec heater.

To reliably determine the shift in phase separation temperature Ts due to the presence of electric fields, defined as ΔTs(E) = Ts(E) – Ts(0), we considered it necessary to measure both Ts(E) and Ts(0) on the same sample. Small differences in sample-to-sample variability lead to larger variability in Ts values across different samples than within the same sample. In our previous study,12 we found that the standard error for the Ts values, when measured across different samples of PS/PVME blends, was as high as ±1.5 K. Such differences are associated with slight variability in precisely mixing the same composition of the blend each time, environmental conditions affecting moisture uptake in PVME as the polymer is hydroscopic, or from inhomogeneities in composition formed during initial casting of the blend, as has been previously reported.28–30 Because the total error in the shift ΔTs(E) = Ts(E) – Ts(0) will be the sum of the errors of the individual measurements of Ts(E) and Ts(0), it was important to minimize the variability between measurements with and without electric field as much as possible. For the present study, we found that highly reproducible Ts values, typically ±0.7 K, can be obtained when Ts is measured within the same sample while remixing the blend in between measurements. It was first shown by Bank et al.31 that the thermally induced phase separation process of PS/PVME polymer blends is reversible. We demonstrate that 50/50 PS/PVME blends can be phase separated and subsequently remixed over and over again, while repeatedly measuring the same Ts value. Figure 2 shows data where the same Ts value has been measured 5 times to within ±0.7 K.

Figure 2(a) plots the fluorescence intensity as a function of temperature for a (40/10)/50 (PS/PS*)/PVME blend measured on heating at 1 K/min. Curves (0)–(6) were collected one after another within the same sample, while cycling up and down through the phase diagram from the mixed state to the unmixed state. As demonstrated in our previous study,12 phase separation in the PS/PVME blends is characterized by a sharp increase in fluorescence intensity resulting from the elimination of fluorescence quenching by the polar PVME component as the non-polar PS component, and covalently attached fluorophore, segregate into domains. The small decrease in fluorescence intensity with increasing temperature observed prior to phase separation is commonly observed for fluorophores, resulting from the increase in internal conversion or other nonradiative processes with higher thermal energy as has been previously discussed.12 Because fluorescence quenching is a very local phenomenon (few nanometer),32 we are able to measure the onset of phase separation when the domains are still very small. Previous work by Halary et al.33 using a similar fluorescence method demonstrated that fluorescence is as sensitive as small angle neutron scattering, and significantly more sensitive than light scattering, at identifying phase separation. Thus, use of fluorescence allows us to identify, measure, and halt phase separation when the domain sizes are still sufficiently small to be easily remixed. Between each fluorescence measurement heating ramp, which starts from a temperature of 80 °C at a heating rate of 1 K/min and continues till approximately 5–8 °C after the sharp increase in fluorescence intensity indicative of phase separation12 is observed, the blend is rapidly quenched at 40 K/min to a temperature 5–10 °C below the measured Ts and held in the one-phase region for several hours to remix the blend. As described below, we found that this remixing time needed to be at least 18 h after the initial first ramp, but could be reduced to 2–3 h between subsequent ramps.

Figure 2(b) shows the curves (0)–(6) vertically offset for clarity, where for each curve, the measured Ts value is identified by a vertical bar. The Ts values are determined from the intersection of linear fits to the data above and below the sharp increase in intensity signifying phase separation,12 as illustrated for curve (1). For the data shown in Fig. 2(b), the phase separation temperature values for curves (0)–(6) are: Ts(0) = 102.4 °C, Ts(1) = 107.7 °C, Ts(2) = 108.1 °C, Ts(3) = 107.7 °C, Ts(5) = 106.9 °C, and Ts(6) = 106.5 °C. Note, we chose not to identify a Ts value for curve (4) because there was insufficient data after phase separation for a viable linear fit to be made. These data were collected by the following measurement protocol. After the sample preparation procedure described in the Experimental methods section, which includes annealing the blend under vacuum at 80 °C for 3–5 days, the samples were transferred to the fluorimeter heater and equilibrated at 80 °C. Fluorescence intensity was then collected while simultaneously heating the blend at a rate of 1 K/min, yielding curve (0). Heating was stopped at 108 °C (about 6 °C above Ts(0) = 102.2 °C) and the sample temperature quenched down to 97 °C (about 5 °C below Ts(0)), followed by an anneal for 22 h. Next, a second fluorescence heating ramp was made yielding curve (2) with Ts(1) = 107.7 °C, notably ∼5 °C higher than Ts(0). In this case, heating was stopped at 113 °C and the sample quenched, followed by an anneal at 97 °C for 2 h. Subsequent measurements repeated this protocol for curve (1), yielding curves (2)–(6) with Ts values for the different ramps (excluding the very first) all within a standard error of only ± 0.7 °C, with an average of Ts(1–6) = 107.4 °C.

From extensive study of many samples, we found that the phase separation temperature Ts is strongly dependent on the history and thermal treatment of the sample, with the very first phase separation temperature Ts(0) measured after blend casting consistently lower than subsequent measures of Ts after remixing the blend by thermal annealing in the one-phase region. Previous studies have shown that the extent of PS/PVME blend homogeneity after solvent casting can be quite variable depending on casting conditions. The extreme example of this are the reports that PS/PVME blends are only formed in the mixed state when cast from aromatic solvents such as toluene and benzene, but phase separate on casting from chlorinated solvents such as chloroform and trichloroethylene.14,29,31 This effect has been identified as depending on whether or not the solvent can form C–H to O hydrogen bonds with PVME, displacing the weak hydrogen bonds formed between the aromatic hydrogens of PS and the ether oxygen of PVME that make this blend miscible.14 Davis et al.34 has demonstrated that extended annealing of trichloroethylene cast PS/PVME blends at temperatures between 60 and 100 °C can remix these blends. Because of these dependencies on casting conditions, most studies on PS/PVME blends include varying amounts of sample annealing at elevated temperatures within the one-phase region.28–31 In the present study, we similarly wanted to ensure that our Ts measurements were always starting from the same mixed state. We found that the measured Ts value increased with longer annealing time at Ts(0) –5 °C (around 95 °C) up until 18 h of annealing, after which the measured Ts saturated at a value of Ts(1) typically 5–10 °C above Ts(0). Later remixing of the blend after heating to only 5–8 °C above Ts required only 2–3 h at Ts(0) –5 °C to achieve identical and stable Ts values for the subsequent ramps. Thus, in our study, all the samples were annealed for 18–24 h at Ts(0) –5 °C after obtaining the first phase separation temperature Ts(0), with subsequent annealing times of 2–3 h at Ts(0) –5 °C between ramps. This procedure resulted in highly reproducible Ts values within a standard error of typically ±0.7 °C for Ts(1), Ts(2), Ts(3), etc.

From optical microscopy of the PS/PVME blends (using a 63× objective lens with 0.70 numerical aperture), we found the morphologies after phase separation to exhibit uniformly sized domains (∼1 μm in size) characteristic of spinodal diffusion, while prior to phase separation and after remixing the blends were featureless. We can estimate the time required to remix such PS/PVME blends using literature values for the diffusion coefficient D, where the diffusion time |$\tau = \frac{{x^2 }}{{6D}}$|τ=x26D, assuming three-dimensional diffusion for our 24 μm thick samples. For molecular weights Mw = 105 kg/mol, Mw/Mn = 1.06, for PS and Mw = 99 kg/mol, Mw/Mn = 2.10, for PVME (comparable to our molecular weights in the present study), Jabbari and Peppas found that D = 4.2 × 10−14 cm2/s at 85 °C and D = 1.1 × 10−12 cm2/s at 105 °C.35 Such diffusion coefficients suggest that domains sizes of x ≈ 1 μm will interdiffuse in ∼10 h at 85 °C or ∼30 min at 105 °C, consistent with the annealing times used in the present study.

To accurately determine the shift in phase separation temperature when in the presence of an electric field, ΔTs(E) = Ts(E) – Ts(0), we first use the remixing protocol described above to establish the zero-field Ts value for a given blend, then apply the electric field and measure how much this Ts value has shifted. We can also subsequently remove the electric field and recover the same zero-field Ts value as measured previously. In this fashion, we provide clear evidence showing that the miscibility of PS/PVME blends is substantially and reversibly altered when in the presence of electric fields. Figure 3 substantiates this for a single sample by plotting the fluorescence intensity of a (40/10)/50 (PS/PS*)/PVME polymer blend as a function of temperature at zero electric field [traces (1), (2), and (4)] and under an external electric field of E = 1.4 × 107 V/m [trace (3)]. Curves (1)–(4) collected on heating at 1 K/min are measured one after another within a single sample following the same protocol as used to collect the data in Fig. 2, with the exception of applying an external electric field across the sample during the measurement of curve (3) in Fig. 3. Note, the very first heating ramp, corresponding to curve (0) in Fig. 2, is omitted in Fig. 3 because it is always anomalous. In Figure 3, the phase separation temperature prior to applying the electric field was measured twice, giving Ts(1)(0) = 94.0 °C and Ts(2)(0) = 93.2 °C, then an electric field of E = 1.4 × 107 V/m was applied shifting the phase separation temperature to Ts(3)(E) = 102.2 °C, and finally the electric field was removed shifting the phase separation temperature back to Ts(4)(0) = 92.8 °C, very close to its original value. The zero field Ts values for this sample are on average Ts(0) = 93.3 ± 0.6 °C. Thus, Figure 3 clearly demonstrates that the presence of the electric field increases the phase separation temperature of the PS/PVME blend by 8.9 K for a field strength of E = 1.4 × 107 V/m.

FIG. 3.

Fluorescence intensity as a function of temperature for a (40/10)/50 (PS/PS*)/PVME polymer blend with curves (1)–(4) collected one after another within the same sample following the procedure described for the data in Fig. 2 where the blend is quenched and remixed between each measurement (note the very first heat, curve (0), is omitted). Curves (1) and (2) establish Ts(0) at zero electric field: Ts(1)(0) = 94.0 °C and Ts(2)(0) = 93.2 °C. For curve (3), an external electric field of E = 1.4 × 107 V/m is applied, shifting the phase separation temperature up to Ts(3)(E) = 102.2 °C. Finally, the electric field is turned off and curve (4) shows that the same zero field value, Ts(4)(0) = 92.8 °C, is recovered. For this electric field strength, the shift in the phase separation temperature ΔTs(E) = Ts(E) – Ts(0) = 8.9 K, relative to the average zero field value |Ts(0)| = 93.3 °C measured for curves (1), (2), and (4).

FIG. 3.

Fluorescence intensity as a function of temperature for a (40/10)/50 (PS/PS*)/PVME polymer blend with curves (1)–(4) collected one after another within the same sample following the procedure described for the data in Fig. 2 where the blend is quenched and remixed between each measurement (note the very first heat, curve (0), is omitted). Curves (1) and (2) establish Ts(0) at zero electric field: Ts(1)(0) = 94.0 °C and Ts(2)(0) = 93.2 °C. For curve (3), an external electric field of E = 1.4 × 107 V/m is applied, shifting the phase separation temperature up to Ts(3)(E) = 102.2 °C. Finally, the electric field is turned off and curve (4) shows that the same zero field value, Ts(4)(0) = 92.8 °C, is recovered. For this electric field strength, the shift in the phase separation temperature ΔTs(E) = Ts(E) – Ts(0) = 8.9 K, relative to the average zero field value |Ts(0)| = 93.3 °C measured for curves (1), (2), and (4).

Close modal

Note that if Joule heating were present, frequently a concern with such electric field measurements, it would manifest as a decrease in Ts, opposite to the direction in Ts shift we observe due to changes in miscibility of the blend when in the presence of an electric field. During our Ts measurements, we observe no evidence of current flow through the samples and a current limit is set on the power supply to cut out if dielectric breakdown occurs at very high field strengths. For example, application of electric field strengths higher than 1.8 × 107 V/m caused electrical shortage in the samples, accompanied by current (200 mA or higher) through the samples with fast Joule heating. We also mention that for a fixed temperature, we do not observe any change in the pyrene fluorescence emission spectrum, in either shape or overall intensity, when in the presence of electric fields, as compared to the pyrene spectra collected under zero field. From studies of pyrene doped in poly(methyl methacrylate) (PMMA) films,36 any change in fluorescence intensity with electric field is expected to be less than 0.1% for our maximum field strengths, resulting from a slight change of the molecular polarizability of the excited state of pyrene.

Similar measurements to those described for Fig. 3 were carried out on a number of different samples to characterize the shift in miscibility at varying electric field strengths. Figure 4 plots the fluorescence intensity for (40/10)/50 (PS/PS*)/PVME polymer blends as a function of temperature where traces are given for different electric field values: E1 = 0.94 × 107 V/m, E2 = 1.4 × 107 V/m, and E3 = 1.8 × 107 V/m, with a zero electric field trace also provided for reference. The temperature axis has been referenced to the individual Ts(0) values measured for each sample. This allows explicit comparison of the shift in Ts with electric field, ΔTs(E), regardless of any small variability in Ts(0) between different samples. The curves for different E are superimposed atop each other in Fig. 4(a) where the intensity of the curves have been specifically matched at TTs = –30 K, while in Fig. 4(b), the curves have been vertically offset for clarity and a short vertical bar is used to denote the location of Ts(E) for each curve. The data clearly show a progression of larger shifts in Ts with increasing electric field: Ts(E1) = 4.1 K at E1 = 0.94 × 107 V/m, Ts(E2) = 8.9 K at E2 = 1.4 × 107 V/m, and Ts(E3) = 11.9 K at E3 = 1.8 × 107 V/m.

FIG. 4.

Fluorescence intensity as a function of temperature for (40/10)/50 (PS/PS*)/PVME polymer blends, where the temperature axis has been referenced to the individual zero-field Ts(0) values for each sample, enabling explicit comparison of the ΔTs(E) shift for different electric field strengths: ΔTs(E) = 4.1 K for E1 = 0.94 × 107 V/m, ΔTs(E) = 8.9 K for E2 = 1.4 × 107 V/m, and ΔTs(E) = 11.9 K for E3 = 1.8 × 107 V/m. Panel (b) shows the same data from (a) vertically shifted for clarity with a vertical black bar designating the Ts(E) value for each curve. A zero electric field curve (black), corresponding to the same sample as E2, is included for reference.

FIG. 4.

Fluorescence intensity as a function of temperature for (40/10)/50 (PS/PS*)/PVME polymer blends, where the temperature axis has been referenced to the individual zero-field Ts(0) values for each sample, enabling explicit comparison of the ΔTs(E) shift for different electric field strengths: ΔTs(E) = 4.1 K for E1 = 0.94 × 107 V/m, ΔTs(E) = 8.9 K for E2 = 1.4 × 107 V/m, and ΔTs(E) = 11.9 K for E3 = 1.8 × 107 V/m. Panel (b) shows the same data from (a) vertically shifted for clarity with a vertical black bar designating the Ts(E) value for each curve. A zero electric field curve (black), corresponding to the same sample as E2, is included for reference.

Close modal

Figure 5 plots the shift in phase separation temperature ΔTs(E) = Ts(E) – Ts(0) as a function of the square of the electric field strength for all the data collected in this study, using electric field strengths up to 1.8 × 107 V/m. The data consistently show an increase in the phase separation temperature Ts with increasing electric field strength. The magnitude of the Ts shifts is large, up to 13.5 K, and well outside our experimental error in ΔTs(E) of ±1.4 K by roughly an order of magnitude. Hence, it can be concluded that electric fields strongly enhance miscibility in PS/PVME blends. To quantify the magnitude of the shift, the data in Fig. 5 have been fit to a linear trend giving a relative change of ΔTs/E2 = (4.8 ± 0.4) × 10−14 Km2/V2. Such a value is consistent with those previously reported in the literature for different types of blends, although those values vary significantly between 0.08 and 87 (×10−14 Km2/V2).1–3,5–9 Our results also agree with the vast majority of the experimental data in the literature showing that the presence of electric fields enhances miscibility; however, as we will discuss below, we do disagree with the one previous study specifically on PS/PVME blends by Reich and Gordon,4 which has been considered anomalous in the field.2 

FIG. 5.

Shift in phase separation temperature ΔTs(E) = Ts(E) – Ts(0) as a function of square of electric field strength E2, where Ts(E) is the phase separation temperature in the presence of electric field and Ts(0) is the phase separation temperature under zero field, both measured on the same sample. The dashed line is a linear fit through the experimental data points with slope of ΔTs/E2 = (4.8 ± 0.4) × 10−14 Km2/V2.

FIG. 5.

Shift in phase separation temperature ΔTs(E) = Ts(E) – Ts(0) as a function of square of electric field strength E2, where Ts(E) is the phase separation temperature in the presence of electric field and Ts(0) is the phase separation temperature under zero field, both measured on the same sample. The dashed line is a linear fit through the experimental data points with slope of ΔTs/E2 = (4.8 ± 0.4) × 10−14 Km2/V2.

Close modal

The magnitude of our shift in phase separation temperature ΔTs(E) = 13.5 ± 1.4 K represents one of the largest absolute shifts ever reported (excluding the anomalous piezoelectric PVDF blends10,11 and the Reich and Gordon4 study to be discussed below). However, the relative magnitude of our shift, ΔTs/E2 = (4.8 ± 0.4) × 10−14 Km2/V2, is comparable to those previously reported in the literature for blends of small molecules and polymer solutions.1–3,5–9 As previously discussed,2,3,5,15,21 such observed shifts are much larger than those predicted based on Eq. (2) incorporating the standard electrostatic energy term for mixtures. However, a recent study by Orzechowski et al.26 suggests that the next order term in the free energy expansion may become dominant at high field strengths. Their expanded expression for the shift in phase separation temperature with electric field, replacing Eq. (2), is

\begin{equation}\frac{{\Delta T_s (E)}}{{E^2 }} = \frac{{\varepsilon _0 v_{ref} T_s (0)}}{{4kB}}\left[ {\frac{{\partial ^2 \varepsilon (\phi)}}{{\partial ^2 \phi }} - \frac{2}{{\varepsilon (\phi)}}\left( {\frac{{\partial \varepsilon (\phi)}}{{\partial \phi }}} \right)^2 } \right].\end{equation}
ΔTs(E)E2=ɛ0vrefTs(0)4kB2ɛ(ϕ)2ϕ2ɛ(ϕ)ɛ(ϕ)ϕ2.
(3)

The second term in the brackets accounts for the dielectric contrast between the components, suppressing concentration fluctuations parallel to the field direction37,38 and the formation of dielectric interfaces between domains during phase separation. The expression treats composition fluctuations as asymmetric, consistent with previous experimental observations of small-angle light scattering studies on polymer solutions in the presence of uniform electric fields that reported electric-field-induced remixing in the two-phase region.37 Our experimental fluorescence method is not sensitive to any particular orientation because the local fluorescence quenching is determined by the local composition and polarity of the material. However, the Orzechowski et al.26 theory predicts a difference in the expected magnitude of the ΔTs/E2 shift depending on the orientational dependence of the concentration fluctuations. Efforts to quantitatively evaluate Eq. (3) for the present PS/PVME system would require knowing the compositional dependence of the dielectric constant ɛ(ϕ). Although ɛ(ϕ) is not known at this time, we note that reasonable estimates for the various parameters in Eq. (3) do not provide quantitative agreement with our experimental results. Specifically, unless ɛ(ϕ) is particularly different than expected, it is hard to see how the second term in Eq. (3) would dominate the first. For the LCST-type phase diagram of PS/PVME, the interaction parameter B is negative, meaning that the second term in Eq. (3) must dominate for electric fields to enhance mixing. In essence, this term creates an energy penalty for the formation of dielectric interfaces oriented perpendicular to the field direction. Thus, this term always favors increased miscibility with increasing electric field strength, consistent with our experimental results.

It is well known that interfaces between two different dielectric media oriented perpendicular to the electric field direction are energetically unfavorable, resulting in alignment of domains along the field direction; an effect frequently exploited in pattern formation and to align block copolymer morphologies.21,39–43 Such an interface term could be particularly important in understanding polymer phase behavior in the presence of electric fields because chain connectivity may limit complete homogeneity of polymer blends even in the mixed phase. Although PS/PVME blends are regarded as miscible, based on a negative χ value23,28 and exhibition of a single glass transition temperature,30,44 several studies28–30,45–47 have described PS/PVME blends as being micro- or nano-heterogeneous at the segmental level (∼1–5 nm) when in the nominally one-phase mixed state. The experimental data presented here, with strong clear shifts in the phase separation temperature with electric fields, should prove useful for comparison with such theoretical efforts.

In 1979, Reich and Gordon4 published a study reporting very large decreases in the phase separation temperature Ts of PS/PVME blends when in the presence of strong electric fields. For E = 2.72 × 107 V/m, decreases of up to 54 K were reported for the one 33/67 PS/PVME blend composition investigated. PS/PVME exhibits an LCST-type phase behavior such that a decrease in Ts with electric field would imply reduced compatibility. This is in strong contrast to our results reported here in which we observe large increases in Ts of up to 13.5 K for E = 1.7 × 107 V/m. Our results would indicate that electric fields enhance mixing, which is consistent with the larger majority of reports on electric field miscibility effects.1–3,5–9 However, as we clearly see completely opposite effects relative to the one previous study on PS/PVME, we discuss here possible reasons for this discrepancy. To better understand this contradiction between the Reich and Gordon study and our results, it is important to examine the experimental approach used by Reich and Gordon and the context of their study in more detail.

We find it surprising that Reich and Gordon did not further continue their studies with electric fields. Before 1979, the year Reich and Gordon published their results, only one experimental study on the electric field effects of phase separation had been published. In 1965, Debye and Kleboth1 reported seeing minuscule shifts in Ts, up to 0.015 K for E = 0.45 × 107 V/m in nitrobenzene/isooctane mixtures towards enhanced compatibility. The relative change of ΔTs/E2 = 0.08 × 10−14 Km2/V2 for Debye and Kleboth's results is almost 100 times smaller than that reported by Reich and Gordon, ΔTs/E2 = 7.3 × 10−14 Km2/V2. Thus, the large 54 K shifts observed by Reich and Gordon were extremely significant, especially compared to the only other study on a similar subject published at the time.

In 1981, two years after Reich and Gordon published their study on electric field effects on Ts,4 another study from the same group reporting film thickness effects on Ts in PS/PVME blends was published by Reich and Cohen.48 Reich and Cohen found shifts in Ts with decreasing film thickness, below approximately 1 μm in thickness, with the effect found to be substrate dependent. For films on gold substrates, enhanced compatibility with decreasing thickness was observed, while for films on glass substrates, both enhanced and reduced compatibility was observed depending on blend composition.48 We believe these results by Reich and Cohen48 may be informative in explaining the observations by Reich and Gordon.4 

In their electric field study, Reich and Gordon4 reported that the exact thickness of the polymer layers used in their experiments were not known, but assumed to be approximately 1–2 μm in thickness. In order to accurately determine the electric field strength across the polymer layers without knowing the exact film thickness, a 140 ± 4 μm glass microscope cover-slip was sandwiched in between the two polymer coated, ITO-covered glass slides. In such a sample geometry, the additional 1–2 μm thick polymer layers could be taken to be insignificantly small compared to the total sample thickness, such that the electric field strength due to the applied voltage was primarily determined by the 140 μm thickness of the intervening glass cover slip. The inserted cover slip also had the added benefit of virtually eliminating concerns of dielectric breakdown.

Reich and Gordon prepared their polymer blends by dip-coating ITO-covered glass substrates into PS/PVME solutions of 10 wt.% total polymer content, where the molecular weights were reported to be 31.5 kg/mol for PS and 14.4 kg/mol for PVME (neither the polydispersity, nor whether these values were number or weight average were specified).4 In the later study by Reich and Cohen,48 an extensive correlation is included plotting the resulting film thickness obtained by dip coating PS/PVME blends from solutions of different concentrations, for weight average molecular weights of Mw = 36 kg/mol for PS and Mw = 10.7 kg/mol for PVME. The data indicate a strong dependence of film thickness on concentration. For 10 wt.% PS/PVME total polymer content, the resulting film thickness is expected to be 750 nm thick, with film thicknesses varying between 600 and 1000 nm for concentrations 9–11 wt.%.48 Reich and Cohen's results suggest that the polymer layers in Reich and Gordon's experiments may be thinner than the presumed 1–2 μm, and be within the regime, below ∼1 μm, where the phase separation temperature exhibits some film thickness dependence. There may also be considerable variation in film thickness from sample-to-sample depending on the concentration of the solutions used. Unfortunately, we cannot estimate how this might or might not have affected Reich and Gordon's reported shifts in Ts with electric field.

To the best of our knowledge, Reich and Gordon4 determined their shift in the phase separation temperature, ΔTs(E) = Ts(E) – Ts(0), by comparing measurements at zero and non-zero field on different samples. They quote errors for Ts to within 3 K, averaging measurements over multiple samples. Phase separation was measured using a light scattering setup with a helium-neon laser, ramping up in temperature at 1 K/min and identifying the cloud point temperature at which the phase separated domains became large enough to scatter the light. There is no mention of remixing samples, as such domain sizes may be too large to feasibly remix in a timely manner. The as-cast samples were given typical annealing treatments of 24 h at 70 °C under a vacuum of 1 Torr, prior to measurement. Thus, such Ts values would have corresponded to values on first heat, which we find to be the most variable because some preparation conditions are difficult to standardize precisely.

Despite all these possible factors, we are unable to ascertain why Reich and Gordon4 observed, reproducibly for their samples, a direction of Ts shift opposite to our results presented here. We do remark that if their samples, which were immersed in an oil bath, experienced local Joule heating, such an uncontrolled temperature increase would manifest as an apparent decrease in Ts that would be expected to increase with increasing electric field strength. Their sample design containing a glass cover-slip between the electrodes may have effectively eliminated any current flow within the sample, but it is conceivable that some current flow through the surrounding oil may have locally increased the temperature, undetectable in the larger oil bath reservoir.

We have developed an experimental protocol using fluorescence by which the phase separation temperature Ts of PS/PVME blends can be repeatedly remeasured on the same sample by iteratively heating this LCST-type blend to determine Ts, followed by a quench and subsequent anneal of the blend in the one-phase region of the phase diagram. This remixing protocol enables us to reproducibly measure Ts to within ±0.7 K on the same sample with and without the presence of strong uniform electric fields. We demonstrate conclusively that the presence of electric fields substantially enhances the miscibility of the blend by measuring large increases in Ts, significantly outside of experimental error by over an order of magnitude, and subsequently recovering the same original Ts at zero field after the electric field has been turned off. Note that any presence of Joule heating would result in a decrease in Ts, opposite in direction to the increase in Ts we observe when in the presence of an electric field. Our measured shifts ΔTs(E) of up to 13.5 K for electric field strengths of E = 1.7 × 107 V/m are some of the largest absolute shifts in Ts ever reported, although the relative magnitude of the shift ΔTs/E2 = (4.8 ± 0.4) × 10−14 Km2/V2 is comparable to values from previous studies on other blends. We do, however, contradict the one previous study on PS/PVME blends,4 which reported that electric fields induced phase separation. The ultimate reason for this discrepancy is unknown, with several possibilities being considered, although we point out that this study4 has long since been considered an outlier in the field.2 Our findings that electric fields strongly enhance mixing in PS/PVME blends is in agreement with the vast majority of existing experimental data on other blend systems. We believe our study will help bring coherence to the existing experimental data and provide large, unambiguous shifts in Ts(E) with electric field for theoretical predictions to be tested.

We acknowledge financial support from the National Science Foundation CAREER program (Grant No. DMR-1151646) and Emory University in the form of a URC grant. We also thank Professor Haskell Beckham at the Georgia Institute of Technology for gel permeation chromatography measurements.

1.
P.
Debye
and
K.
Kleboth
,
J. Chem. Phys.
42
,
3155
(
1965
).
2.
Y.
Tsori
,
Rev. Mod. Phys.
81
,
1471
(
2009
).
3.
K.
Orzechowski
, in
Polymers, Liquids and Colloids in Electric Fields: Interfacial Instabilities, Orientation and Phase Transitions
, edited by
Y.
Tsori
,
J. E.
Mark
, and
U.
Steiner
(
World Scientific Publishing
,
New Jersey
,
2009
), pp.
87
112
.
4.
S.
Reich
and
J. M.
Gordon
,
J. Polym. Sci.: Polym. Phys. Ed.
17
,
371
(
1979
).
5.
D.
Beaglehole
,
J. Chem. Phys.
74
,
5251
(
1981
).
6.
M. D.
Early
,
J. Chem. Phys.
96
,
641
(
1992
).
7.
D.
Wirtz
and
G. G.
Fuller
,
Phys. Rev. Lett.
71
,
2236
(
1993
).
8.
K.
Orzechowski
,
Chem. Phys.
240
,
275
(
1999
).
9.
H. G.
Schoberth
,
K.
Schmidt
,
K. A.
Schindler
, and
A.
Boker
,
Macromolecules
42
,
3433
(
2009
).
10.
J. S.
Lee
,
A. A.
Prabu
,
K. J.
Kim
, and
C.
Park
,
Macromolecules
41
,
3598
(
2008
).
11.
J. S.
Lee
,
A. A.
Prabu
, and
K. J.
Kim
,
Macromolecules
42
,
5660
(
2009
).
12.
A.
Kriisa
,
S. S.
Park
, and
C. B.
Roth
,
J. Polym. Sci., Part B: Polym. Phys.
50
,
250
(
2012
).
13.
J. L.
Halary
,
J. M.
Ubrich
,
J. M.
Nunzi
,
L.
Monnerie
, and
R. S.
Stein
,
Polymer
25
,
956
(
1984
).
14.
M. M.
Green
,
J. L.
White
,
P.
Mirau
, and
M. H.
Scheinfeld
,
Macromolecules
39
,
5971
(
2006
).
15.
A.
Onuki
,
Europhys. Lett.
29
,
611
(
1995
).
16.
H.
Kawai
,
Jpn. J. Appl. Phys.
8
,
975
(
1969
).
17.
A. J.
Lovinger
,
Science
220
,
1115
(
1983
).
18.
L. M.
Robeson
,
Polymer Blends: A Comprehensive Review
(
Hanser Gardner Publications
,
Cincinnati, Ohio
,
2007
).
19.
I. C.
Sanchez
and
M. T.
Stone
, in
Polymer Blends: Formulation and Performance
, edited by
D. R.
Paul
and
C. B.
Bucknall
(
Wiley
,
New York
,
2000
), Vol.
1
, pp.
15
53
.
20.
R. P.
White
and
J. E. G.
Lipson
,
Macromolecules
47
,
3959
(
2014
).
21.
K.
Amundson
,
E.
Helfand
,
X. N.
Quan
,
S. D.
Hudson
, and
S. D.
Smith
,
Macromolecules
27
,
6559
(
1994
).
22.
M.
Rubinstein
and
R. H.
Colby
,
Polymer Physics
(
Oxford University Press
,
2003
).
23.
H. B.
Eitouni
and
N. P.
Balsara
, in
Physical Properties of Polymers Handbook
, 2nd ed., edited by
J. E.
Mark
(
Springer
,
New York
,
2007
), pp.
339
356
.
24.
L. D.
Landau
,
L. P.
Pitaevskii
, and
E. M.
Lifshitz
,
Electrodynamics of Continuous Media
, 2nd ed. (
Elsevier Butterworth-Heinemann
,
Oxford
,
1984
).
25.
J. D.
Jackson
,
Classical Electrodynamics
, 3rd ed. (
John Wiley and Sons
,
New York
,
1999
).
26.
K.
Orzechowski
,
M.
Adamczyk
,
A.
Wolny
, and
Y.
Tsori
,
J. Phys. Chem. B
118
,
7187
(
2014
).
27.
B. J.
Bauer
,
B.
Hanley
, and
Y.
Muroga
,
Polym. Commun.
30
,
19
(
1989
).
28.
T. K.
Kwei
,
T.
Nishi
, and
R. F.
Roberts
,
Macromolecules
7
,
667
(
1974
).
29.
C. W.
Chu
,
L. C.
Dickinson
, and
J. C. W.
Chien
,
J. Appl. Polym. Sci.
41
,
2311
(
1990
).
30.
T.
Wagler
,
P. L.
Rinaldi
,
C. D.
Han
, and
H.
Chun
,
Macromolecules
33
,
1778
(
2000
).
31.
M.
Bank
,
J.
Leffingwell
, and
C.
Thies
,
J. Polym. Sci., Part A-2
10
,
1097
(
1972
).
32.
B.
Valeur
,
Molecular Fluorescence: Principles and Applications
(
Wiley-VCH
,
Weinheim
,
2001
).
33.
J. L.
Halary
,
J. M.
Ubrich
,
L.
Monnerie
,
H.
Yang
, and
R. S.
Stein
,
Polym. Commun.
26
,
73
(
1985
).
34.
D. D.
Davis
and
T. K.
Kwei
,
J. Polym. Sci., Part B: Polym. Phys.
18
,
2337
(
1980
).
35.
E.
Jabbari
and
N. A.
Peppas
,
Macromolecules
26
,
2175
(
1993
).
36.
S.
Umeuchi
,
Y.
Nishimura
,
I.
Yamazaki
,
H.
Murakami
,
M.
Yamashita
, and
N.
Ohta
,
Thin Solid Films
311
,
239
(
1997
).
37.
D.
Wirtz
,
K.
Berend
, and
G. G.
Fuller
,
Macromolecules
25
,
7234
(
1992
).
38.
I.
Gunkel
,
S.
Stepanow
,
T.
Thurn-Albrecht
, and
S.
Trimper
,
Macromolecules
40
,
2186
(
2007
).
39.
K.
Amundson
,
E.
Helfand
,
D. D.
Davis
,
X.
Quan
,
S. S.
Patel
, and
S. D.
Smith
,
Macromolecules
24
,
6546
(
1991
).
40.
T. L.
Morkved
,
M.
Lu
,
A. M.
Urbas
,
E. E.
Ehrichs
,
H. M.
Jaeger
,
P.
Mansky
, and
T. P.
Russell
,
Science
273
,
931
(
1996
).
41.
E.
Schaffer
,
T.
Thurn-Albrecht
,
T. P.
Russell
, and
U.
Steiner
,
Nature
403
,
874
(
2000
).
42.
H. G.
Schoberth
,
V.
Olszowka
,
K.
Schmidt
, and
A.
Boker
,
Adv. Polym. Sci.
227
,
1
(
2010
).
43.
C.
Liedel
,
C. W.
Pester
,
M.
Ruppel
,
V. S.
Urban
, and
A.
Boker
,
Macromol. Chem. Phys.
213
,
259
(
2012
).
44.
T. P.
Lodge
,
E. R.
Wood
, and
J. C.
Haley
,
J. Polym. Sci., Part B: Polym. Phys.
44
,
756
(
2006
).
45.
K.
Schmidt-Rohr
,
J.
Clauss
, and
H. W.
Spiess
,
Macromolecules
25
,
3273
(
1992
).
46.
H.
Shimizu
,
S.
Horiuchi
, and
T.
Kitano
,
Macromolecules
32
,
537
(
1999
).
47.
D.
Wang
and
H.
Ishida
,
Macromol. Chem. Phys.
208
,
2222
(
2007
).
48.
S.
Reich
and
Y.
Cohen
,
J. Polym. Sci., Part B: Polym. Phys.
19
,
1255
(
1981
).