Permeability of Rubbery and Glassy Membranes of Ionic Liquid Filled

Nov 20, 2015 - It is of interest to determine how sensitive the rate of transport is to temperature, particularly for membranes in the vicinity of the...
0 downloads 11 Views 3MB Size
Article pubs.acs.org/JPCB

Permeability of Rubbery and Glassy Membranes of Ionic Liquid Filled Polymersome Nanoreactors in Water Soonyong So,† Letitia J. Yao,‡ and Timothy P. Lodge*,†,‡ †

Department of Chemical Engineering & Materials Science and ‡Department of Chemistry, University of Minnesota, Minneapolis, Minnesota 55455, United States S Supporting Information *

ABSTRACT: Nanoemulsion-like polymer vesicles (polymersomes) having ionic liquid interiors dispersed in water are attractive for nanoreactor applications. In a previous study, we demonstrated that small molecules could pass through rubbery polybutadiene membranes on a time scale of seconds, which is practical for chemical transformations. It is of interest to determine how sensitive the rate of transport is to temperature, particularly for membranes in the vicinity of the glass transition (Tg). In this work, the molecular exchange rate of 1-butylimidazole through glassy polystyrene (PS) bilayer membranes is investigated via pulsed field gradient nuclear magnetic resonance (PFG-NMR) over the temperature range from 25 to 70 °C. The vesicles were prepared by the cosolvent method in the ionic liquid 1-ethyl-3methylimidazolium bis(trifluoromethylsulfonyl) imide ([EMIM][TFSI]), and four different polystyrene-b-poly(ethylene oxide) (PS-PEO) diblock polymers with varying PS molecular weights were examined. The vesicles were transferred from the ionic liquid to water at room temperature to form nanoemulsion solutions of polymer vesicles in water. The exchange rate of 1-butylimidazole added to the aqueous solutions was observed under equilibrium conditions at each temperature. The exchange rate decreased as the membrane thickness increased, and the exchange rate through the glassy membranes was three to four times slower than through the rubbery polybutadiene membranes under the same experimental conditions. These results demonstrate that the permeability through nanosized membranes depends on both the dimension and chemistry of membrane-forming blocks. Furthermore, the exchange rate was investigated as a function of temperature in the vicinity of the Tg of PS-PEO membranes. The exchange rate, however, is not a strong function of the temperature in the vicinity of the membrane Tg, due to a combination of the nanoscopic dimension of the membrane, and some degree of solvent plasticization.



INTRODUCTION Just as with transport through cell membranes, the bilayer core membrane plays a key role in polymersome nanoreactors, in terms of permeability and selectivity of molecules for reactions. Permeability (P, m2/s) is a function of the diffusion coefficient in the membrane (D) and the partition coefficient (H) of the diffuser between the membrane and the surrounding fluid, as shown by the equation, P = DH.1 One method for tuning permeability is by controlling the membrane fluidity, defined as the inverse viscosity of the bilayer membranes, in order to change the diffusivity of the penetrating molecules (D).2 Lower fluidity therefore corresponds to lower permeability. For example, Yu et al. reported pH-sensitive sandwiched membranes from ABC triblock polymers composed of polystyrene (PS), poly(2-diethylaminoethyl methacrylate) (PDEA) and poly(ethylene oxide) (PEO).3 The protonation of amino groups on the PDEA block by a pH change from 7.94 to 6.98 induced a change in size, which introduced swollen pathways through the sandwiched layers, accelerating the permeability of protons from the outer medium by 3 orders of magnitude. Another method for permeability control by D is formation of physical defects for the permeation through pores © 2015 American Chemical Society

in the membrane. For instance, Kim et al. embedded stimuliresponsive boronic acid-containing block copolymers into PSPEO polymersomes, and by the removal of selectively soluble embedded boronic acid-containing block copolymers at high pH, increased the transportation rate of 6,8-difluoro-4methylumbelliferyl octanoate.4 To control the fluidity of the membrane core, the choice of the core block with an appropriate glass transition temperature (Tg) is a possible approach.5 The fluidity of the membrane (∼1/η) should increase significantly above Tg. For example, it is well established that the inverse viscosity (1/η) of PS shows a strong increase with T above Tg ≈ 100 °C.6 This temperatureinduced fluidity change can be advantageous for the polymersome nanoreactor system, especially for a system involving phase transfer between two immiscible phases, such as 1,2polybutadiene-b-poly(ethylene oxide) (PB-PEO) polymersomes in water and an ionic liquid (IL), 1-ethyl-3methylimidazolium bis(trifluoromethylsulfonyl)imide Received: August 29, 2015 Revised: November 6, 2015 Published: November 20, 2015 15054

DOI: 10.1021/acs.jpcb.5b08425 J. Phys. Chem. B 2015, 119, 15054−15062

Article

The Journal of Physical Chemistry B ([EMIM][TFSI]).7,8 In the biphasic system of water and [EMIM][TFSI], the polymersomes prefer to reside in the aqueous phase below the transfer temperature (Tt), which is around 75 °C for PB-PEO and PS-PEO polymersomes, whereas above the Tt the polymersomes move into the IL phase. The permeability of the membrane can be effectively controlled by temperature, since the permeability depends on the state of the membrane between the glassy and rubbery limits. To this end, it should be possible to use a membrane with a Tg range of 25 °C < Tg < Tt. For the bulk polymer membrane in this Tg range, a good candidate is a random copolymer of glassy and rubbery polymers (e.g., a statistical copolymer of PS and PB). However, in dilute micelle solution, it has been shown that Tg of the PS block itself in PS-PEO spherical micelles is lower than the bulk Tg because the PS core is on the order of nanometers and some level of solvent can contribute to reducing Tg.9 For example, the onset transition temperature was around 45 °C for PS-PEO micelles having a 20 kg/mol PS block, whereas the bulk Tg was around 100 °C. Membrane porosity is also a candidate to control molecular transport, as in the examples described above. However, it leads to leakage of the interior liquid and active species, and it is difficult to have reversible porosity in a self-assembled membrane, which is necessary for recycling nanoreactors. Therefore, in this study, the fluidity control approach is explored with four different PS-PEO block polymers, which are expected to have a membrane Tg in the desirable range, 25 °C < Tg < Tt. The molecular exchange rate of 1-butylimidazole through PS bilayer membranes is investigated via the pulsed field gradient nuclear magnetic resonance (PFG-NMR) technique. Even though the system is at equilibrium, if the tracer molecules exchange their location between different diffusing environments on the experimental time scale, the exchange rate can be obtained.10 It is also possible to extract permeation information under nonequilibrium conditions, such as in the presence of a concentration gradient,11 but this is more useful for low permeation rates; the measurement should be completed before equilibrium is attained.12 Herein, relatively rapid permeation of the tracer molecules is measured under equilibrium conditions over the temperature range of 25−70 °C, and compared with permeation rate through rubbery PB membranes of PB-PEO polymersomes.

OH. The number-average molecular weight (Mn) and dispersity (Đ) of the PS-OH polymers and the block polymers were characterized by 1H NMR spectroscopy and size exclusion chromatography (SEC), respectively (see Supporting Information). The Mn, Đ, and the PEO volume fraction (f PEO) of the synthesized PS-PEO polymers are listed in Table 1. The Table 1. Characteristics of PS-PEO Diblock Polymers polymer

MPSa (kg/mol)

MPEOb (kg/mol)

Đc

f PEOd

SO(10−2) SO(14−2.5) SO(18−3.6) SO(27−4)

9.8 14.1 18.4 27.3

2.0 2.6 3.6 4.3

1.01 1.04 1.01 1.01

0.16 0.14 0.15 0.13

a,b Number-average molecular weight of the ω-hydroxyl polystyrene (PS-OH) and PEO blocks obtained from 1H NMR spectroscopy. c Dispersity (Mw/Mn) of PS-PEO diblock polymers from SEC. d Volume fraction of PEO block in PS-PEO diblock polymers calculated by using Mn of each block and the bulk densities of polymers (ρPS = 1.05 g/cm3, ρPEO = 1.13 g/cm3).15,16

hydrophobic PS block length was systematically varied from 10 000 to 27 000 g/mol with relatively short PEO blocks (f PEO < 0.2) for polymersome preparation, and the four different block polymers, SO(10−2), SO(14−2.5), SO(18−3.6), and SO(27−4), were named based on Mn,PS and Mn,PEO (in kDa). Ionic Liquid Synthesis. Equimolar amounts of 1-ethyl-3methylimidazoium bromide ([EMIM][Br], Io-Li-Tec, 99%) and lithium bis(trifluoromethylsulfonyl)imide ([Li][TFSI], 3M, > 98%) were mixed with water (30 wt %) at 70 °C for 24 h. The water immiscible IL product was isolated by a separation funnel, and washed with water three times to remove LiBr salts from the IL phase. The IL was further purified by passage through a column of aluminum oxide (Brockmann I, activated, standard grade, Aldrich). Finally, the IL was dried under vacuum for 3 days, and kept in a house-vacuum desiccator. Nanoemulsion-like Polymersome Solution Preparation. Polymersomes were prepared through the cosolvent method. The weighed polymers were dissolved in dichloromethane, then the IL, [EMIM][TFSI], was added to prepare a 0.5 wt % polymersome solution in the IL after cosolvent removal. The volume ratio of dichloromethane:[EMIM][TFSI] was 60:40 for SO(14−2.5), SO(18−3.6), and SO(27−4), but 30:70 for SO(10−2). The volume ratio of solvents was determined based on a previous study of the phase transfer behavior of PS-PEO polymersomes in the biphasic system of water and [EMIM][TFSI].14 The cosolvent dichloromethane was removed under N2 atmosphere and vacuum at room temperature. The polymersomes were transferred to the aqueous phase with gentle stirring after an equal volume of D2O was added to the IL solution, and the aqueous phase was removed for the permeability study. Note that the polymersome dimensions are independent of temperature over the range relevant to this report. NMR Spectroscopy. NMR spectroscopy was used to determine the molecular permeability through the PS membranes with a Bruker Avance III 500 MHz NMR spectrometer equipped with a 5 mm Triple Resonance Broadband (TBO) PFG probe. For the NMR study, 50 mM 1-butylimidazole as the tracer molecule was added to the polymersome solutions in D2O at room temperature, and the samples were stirred for at least 24 h before the measurement to ensure equilibrium. As shown previously, equilibration across



EXPERIMENTAL SECTION Block Polymer Synthesis. Styrene (99.9%), ethylene oxide (≥99.5%), n-butyllithium (2.5 M in hexane), di-n-butylmagnesium (1.0 M in heptane), and sec-butyllithium (1.4 M in cyclohexane) were purchased from Aldrich. PS-PEO block polymers were synthesized via sequential anionic polymerization.13 Styrene and ethylene oxide were purified twice in flasks with n-butyllithium and di-n-butylmagnesium, respectively. The PS block was synthesized at 40 °C for 4 h with styrene initiated by sec-butyllithium in anhydrous cyclohexane. Next, purified ethylene oxide was added to the reactor, and the mixture was reacted for 24 h, followed by the addition of deoxygenated methanol to end-cap polystyryl anions with a hydroxyl terminal group (PS-OH). The molecular weight of PS was controlled by varying the molar ratio of styrene and secbutyllithium. A short PEO block was grown from a PS-O− macroinitiator (prepared by addition of potassium naphthalenide) with ethylene oxide in tetrahydrofuran (THF) at 45 °C for 24 h. The PEO block synthesis was terminated by adding excess deoxygenated methanol in the same manner as for PS15055

DOI: 10.1021/acs.jpcb.5b08425 J. Phys. Chem. B 2015, 119, 15054−15062

Article

The Journal of Physical Chemistry B

Figure 1. 1H NMR spectra of SO(14−2.5) solution in D2O (a) before adding 1-butylimidazole, and (b) after adding 50 mM 1-butylimidazole. As with [EMIM], 1-butylimidazole shows two sets of chemical shifts. Distinguishable peaks of 1-butylimidazole are highlighted in (b). Acidic protons (vi) of [EMIM] and (e) of 1-butylimidazole at around 8.5 ppm cannot be seen due to hydrogen−deuterium exchange.

the membrane can be obtained within several minutes.8,12 The measurement temperature was increased in 10 °C increments from 25 to 65 °C by using a temperature controller on the instrument. The actual temperature was calibrated as shown in Supporting Information Figure S3 with the chemical shift separation between the −OH and −CH2− protons of ethylene glycol.17 The actual temperatures of the samples were 25, 36, 47, 58, and 69 °C, and at each temperature, the NMR samples were equilibrated for at least 30 min before the measurements. The “ledbpgp2s” pulse sequence (longitudinal eddy current delay experiment using bipolar gradients acquired in 2D)18 was used for the PFG-NMR experiments for the translational diffusion of probing molecules. The translational diffusion (D) is associated with the intensity (I) attenuation of resonant nuclei in the PFG-NMR experiments, as depicted in eq 1: I = exp(−γ 2δ 2G2D(Δ − δ /3)) I0

As shown in Figure 1b, 1-butylimidazole added to the solution also shows proton peaks that are slightly upfield shifted in the SO(14−2.5) polymersomes. For example, “ae” and “ai” represent protons at the “a” positions in the exterior and interiors of the polymersomes, respectively. Two sets of peaks were also found in the other PS-PEO polymersomes. Both the formation of vesicles from four different PS-PEO block polymers, and the facile permeation of 1-butylimidazole through the glassy PS membranes, were therefore simply demonstrated through comparative 1H NMR spectroscopy of pre-encapsulated [EMIM] (Figure 1a) and postadded 1butylimidazole (Figure 1b) with the IL-filled polymersomes, respectively. Molecular Exchange Rate through Glassy Membranes. The rate of molecular exchange of the tracer molecules was determined by PFG-NMR. As an example, the normalized echo-attenuated intensity of poly(ethylene glycol) (PEG) (Mn = 200 g/mol, Aldrich) in IL-saturated D2O is shown in Figure S4, for various values of Δ and δ. When there is no diffusion barrier during a diffusion time, Δ, the echo-decays fall on eq 1 as shown in Figure S4, and for PEG D = 4.1 × 10−10 m2/s, as obtained from the slope of the decay. However, the diffusion behavior of tracer molecules in a polymersome solution is not simple translational diffusion. Unless the exchange rate is very fast, or the diffusion time is very short, the molecular motion cannot be described by a single decay, since the molecules are both inside and outside the vesicles, and they also exchange through the membranes. Tracer molecules in confined vesicles (typically 100 ms, Dapp crosses over

Na′ = 1 − Nb′

(5)

Figure 3. Experimental data at 25 °C and fitted echo curves of the proton from 1-butylimidazole in the polymersome solutions of (a) SO(10−2), (b) SO(14−2.5), (c) SO(18−3.6), and (d) SO(27−4) with Δ = 150, 250, 350 ms and fixed δ = 8 ms. The data were fitted using the Kärger exchange model. 15057

DOI: 10.1021/acs.jpcb.5b08425 J. Phys. Chem. B 2015, 119, 15054−15062

Article

The Journal of Physical Chemistry B Table 2. Permeation Rate from the Kärger Exchange Model polymer SO(10−2) SO(14−2.5) SO(18−3.6) SO(27−4) BO(9−3)

De (m2/s)a

Di (m2/s)a

−10

−13

6.1 6.1 6.1 6.0 5.8

× × × × ×

10 10−10 10−10 10−10 10−10

7.2 2.1 2.4 1.3 2.2

× × × × ×

10 10−12 10−12 10−12 10−12

Nib

Neb

1/τi (1/s)c

1/τe (1/s)c

0.09 0.08 0.07 0.08 0.07

0.91 0.92 0.93 0.92 0.93

1.5 1.0 0.7 0.3 2.5

0.14 0.09 0.06 0.02 0.18

a

De and Di represent the diffusion coefficients (m2/s) of a tracer molecule in the external aqueous medium and in the interiors of polymersomes, respectively. bNi and Ne stand for the mole fractions of probe molecules in the encapsulated and free spaces (Ni + Ne = 1), respectively. cτi and τe are the average residence times at each phase, and the inverse of the residence times represent the rate of escape (1/τi) and entry (1/τe), respectively.

Figure 4. (a) Membrane thickness dependence of the escape rate (1/τi) of 1-butylimidazole from PS-PEO polymersomes at 25 °C: ■ refers to the data of SO(10−2) (d = 15 nm), SO(14−2.5) (d = 18 nm), SO(18−3.6) (d = 21 nm), and SO(27−4) (d = 26 nm), and red ★ refers to the data of BO(9−3) (d = 21 nm). (b) The escape rate (1/τi) versus 1/d. The solid line is a linear fit ((1/τi) = 42.03 − 1.34 × (1/d), R = 1).

1 Nb′ = − 2

1⎡ ⎢(Ni 4⎣

− Ne)(ke − k i) +

1 τe

+

1 τi

+

1 Te

+

1⎤ ⎥ Ti ⎦

where T1,i and T1,e are the spin−lattice relaxation times of a proton in the polymersome interior and exterior, respectively, and T2,i and T2,e are the spin−spin relaxation times in the polymersome interior and exterior, respectively. Di and De were obtained from the initial and final slopes of the echo decay curves in Figure 3, and T1 and T2 were experimentally obtained by T1 and T2 experiments.14 The values in SO(18−3.6) are summarized in Table S1. Given the parameters (Δ, δ, G, γ) and values (Di,e, T1,2), the remaining values, τi and Ni, were obtained by fitting with eq 3. Since four variables (τi,e and Ni,e) are correlated to each other, τi and Ne can be calculated from the relations, Ne = 1 − Ni and τe = (1 − Ni) × τi/Ni. The rates of escape (1/τi) and entry (1/τe) can be evaluated from the inverse residence times, which are directly related to permeance (1/τ ∼ p = P/d, where d is membrane thickness).21 Figure 3 shows echo-decays of both peaks from the interior and exterior (“ae”, “ai”) of 1-butylimidazole in Figure 1b at 25 °C. All decays are biexponential, and the final plateaus change depending on the diffusion time (Δ), indicating that molecular permeation contributes during the PFG-NMR measurement. If there were no exchange of molecules, the final plateau value would be solely determined by the mole fraction of molecules inside the vesicles, and the value would be independent of Δ. If there is significant exchange, the molecules in the exteriors, which have a low attenuated intensity in PFG-NMR due to rapid diffusion, can visit the interiors during Δ, but they contribute less to the intensity from the interiors compared to the molecules remaining inside during Δ. Therefore, with increased diffusion time Δ, the intensity contributed by the molecules in the confined polymersomes decreases, because

Q2 (6)

The parameters Q1 and Q2 are defined by ki, ke, τi, τe, Te, and Ti, Q1 =

1⎛ 1 1 1 1⎞ + + + ⎟ ⎜ke + k i + τe τi 2⎝ Te Ti ⎠

2 1 ⎛ 1 1 1 1⎞ 1 − + − ⎟ + Q2 = ⎜ke − k i + τe τi τeτi 2 ⎝ Te Ti ⎠

(7)

(8)

where ki and ke are rate constants defined by γ, δ, G, and Di, and De, k i = γ 2δ 2G2Di

(9)

ke = γ 2δ 2G2De

(10)

and the effective relaxation times, Ti and Te, are defined by eqs 11 and 12, Δ − δ⎞ 1 1 ⎛ 2δ ⎟⎟ = ⎜⎜ + Δ ⎝ T2,e Te T1,e ⎠

(11)

1 1 ⎛ 2δ Δ − δ⎞ ⎟⎟ = ⎜⎜ + Ti Δ ⎝ T2,i T1,i ⎠

(12) 15058

DOI: 10.1021/acs.jpcb.5b08425 J. Phys. Chem. B 2015, 119, 15054−15062

Article

The Journal of Physical Chemistry B

Figure 5. Experimental data at various temperatures and fitted echo curves of the proton from 1-butylimidazole in the polymersome solution of SO(18−3.6) with Δ = 150, 250, 350 ms. The data were fitted using the Kärger exchange model.

the study of Leson and co-workers, the exchange rate can decrease with vesicle size; however, the effect of size was insignificant for vesicles with hydrodynamic radii (Rh) larger than 60 nm.12 Therefore, in this study, the effect of Rh is assumed to be negligible. The concentration of polymersomes can also affect the permeation rate. Previously, we showed that the entry rate was proportional to the number of polymersomes due to the increased contact area with the diffusants.8 In this report, the concentration of polymers was held at 0.5 wt % for all samples, and thus the area differences were not considered. As the length of the PS block increases, the membrane thickness (d) also increases (Figure S6), and both the escape and entry rates of 1-butylimidazole decrease, as shown in Table 2 and Figure 4a. In Figure S7, the final plateau regions of SO(10−2) and SO(27−4) are compared to show the effect of exchange rate on the echo decay curves. As the diffusion time Δ increases, the polymersomes having higher permeation rate show larger plateau differences between Δ = 150 and 350 ms (Figure S7c). Membrane thickness dependent permeance has been observed in other polymer vesicles, such as PB-PEO,25 poly(2-vinylpyridine)-b-poly(ethylene oxide),12 and a series of poly(ethylene oxide)-b-poly(butylene oxide) copolymers.26 From the plot of 1/τi versus 1/d in Figure 4b, 1/τi is linear in 1/d indicating that the permeance of the polymersome

more molecules have been exposed to the exteriors with longer Δ. The plateau intensity also decreases as the molecules exchange more rapidly.22 As shown in Figure 3 and Figure S5, the difference of the final plateaus with all PS-PEO polymersomes between Δ = 150 and 350 ms is much smaller compared to the case with PB-PEO (BO(9−3), Mn,PB = 9.3 kg/mol, Mn,PEO = 2.5 kg/mol) polymersomes having rubbery PB membranes (Figure S5). This suggests that molecular exchange is limited by the glassy PS membranes, and the diffusion barrier, which is related to the fluidity of the membrane, is the main contribution for slower exchange, given the solubility parameters of PS (δPS), PB (δPB), and 1-butylimidazole (δBIm), 18.6 MPa1/2, 17.1 MPa1/2, and 21 MPa1/2, respectively.23,24 Using the Kärger exchange model (eq 3), the echo-decay curves were fitted as shown in Figure 3, and the fitting results are summarized and compared to BO(9−3) in Table 2. Before comparing the exchange rate as a function of PS block length, the effect of vesicle size should be considered. The vesicle size can be obtained using the Stokes−Einstein equation and the long time Di values, which represent the diffusion coefficient of the polymersome. The average Rh values for SO(14−2.5) and SO(18−3.6) are around 100 nm, while those of SO(10−2) and SO(27−4) are 270 and 160 nm, respectively. On the basis of 15059

DOI: 10.1021/acs.jpcb.5b08425 J. Phys. Chem. B 2015, 119, 15054−15062

Article

The Journal of Physical Chemistry B

Figure 6. (a) Arrhenius plot of the escape rate (1/τi) of 1-butylimidazole through PS-PEO polymersomes at 25, 36, 47, 58, and 69 °C. The NMR samples were equilibrated at each temperature for 30 min before the measurement. The arrows indicate the transition points that might be related to Tg of PS membranes. (b) Toluene diffusion (cm2/s) prediction with Williams−Landel−Ferry (WLF) equation in PS/toluene solution (wPS = 0.881); WLF parameters are from ref 34.

Temperature Dependent Exchange Rate in Glassy Membranes. It has been shown that membrane fluidity control can effectively reduce permeation by introducing the glassy PS instead of the rubbery PB membrane. The membrane fluidity control has also been studied by Eisenberg and coworkers. They studied the effects of plasticizer amounts on the permeability of molecules through the glassy membrane of polystyrene-b-poly(acrylic acid) polymersomes.27,31 The plasticizer dioxane could partition into the PS membrane, and the permeability increased as the dioxane content increased. The proton diffusivity increased by an order of magnitude as the dioxane content in the solution increased from 7 to 14 vol %. Herein, however, the membrane fluidity was controlled by changing the temperature in the range of 25 °C < T < Tt. The permeation rate changes were monitored in order to see the transition temperature of the PS membrane of the PS-PEO polymersome based on the Arrhenius plot of the escape rate (1/τi). The PFG-NMR experiments were conducted at different temperatures with SO(10−2) and SO(18−3.6). The echodecay curves of 1-butylimidazole in SO(18−3.6) solution are shown in Figure 5 as an example of the temperature-dependent echo-decay curve evolution. In both polymersome systems, the difference in the final plateaus increases, which indicates faster molecular exchange as the temperature increases. Due to noise, some data in Figure 5 panels b−d were truncated. Experimental data at different temperatures were also fit with the Kärger exchange model, and the resulting Arrhenius plot of 1/τi is shown in Figure 6a. It is clear that 1/τi increases as the temperature increases, and at 36 °C, 1/τi reaches 1.8 s−1 for SO(10−2), and 1.9 s−1 for SO(18−3.6) at 58 °C. These 1/τi values are comparable with the rubbery BO(9−3) (1/τi at 25 °C reaches 2.5 s−1). In the Arrhenius plot, there is a hint of a breakpoint, where the slope of 1/τi becomes slightly steeper at higher temperatures. Though more experiments are required to determine the relationship between the breakpoint highlighted by arrows in Figure 6a and Tg, it is interesting that the data points at 36 °C for SO(10−2) and 47 °C for SO(18−3.6)

membrane is a strong function of the membrane thickness. The membrane thickness normalized permeation rate (d/τi), proportional to the permeability (P), also decreases linearly as d increases, as shown in Figure S8. This result is consistent with the previous study in our group about the membrane thickness-dependent permeability of PB-PEO polymersomes.25 We proposed then that changes in solubility (H) induced permeability changes in nanometer scale polymer membranes. In addition to the solubility changes, in this study, D may increase as d decreases due to the plasticized glassy membrane by the diffusing molecules.27 Therefore, the material function P was not constant for the various PS membranes. It is of particular interest to compare the exchange rates through glassy PS versus rubbery PB membranes. As expected from the different final plateaus in Figure 3 and Figure S5, 1butylimidazole exchange through a glassy membrane is slower than through a rubbery membrane. In particular, as shown in Figure 4a, for the same membrane thickness, d = 21 nm (SO(18−3.6) and BO(9−3)), the exchange rate was about 3−4 times slower through the PS membrane at 25 °C. For comparison, Rein and co-workers measured the permeability of Ar and CH4 in bulk PB and PS.28 The diffusivity difference was the dominant factor for the permeability changes in PB and PS. Ar diffused 1000 times faster in PB, while the solubility, which is related to H, was of the same order at 20 °C. The exchange rate here is also faster in the rubbery PB membrane, but the exchange rate difference is not nearly as significant as the difference of the bulk PS and PB polymer films. Similar solubility parameters of the IL (δ[EMIM][TFSI] = 27.6 MPa1/2)29 and 1-butylimidazole (δBIm = 21 MPa1/2)24 to PS (δPS = 18.6 MPa1/2)23 suggest they may plasticize the PS membrane, and PEO coronas can act as hydrophobic moieties near membrane cores.30 Thus, we can expect that the PS membrane fluidity is enhanced in the nanosized membrane in contact with the IL, 1butylimidazole, and PEO coronas having low Tg, and hence the exchange rate is not reduced as much as might be expected when the membrane chemistry is changed from rubbery to glassy. 15060

DOI: 10.1021/acs.jpcb.5b08425 J. Phys. Chem. B 2015, 119, 15054−15062

Article

The Journal of Physical Chemistry B correspond well with the onset of the glass transition (∼45 °C) in SO(20−5) micelle solutions, as measured by fluorescence spectroscopy,9 and also the Tg of ultrathin PS films, which is 40 °C when the thickness is around 20 nm.32 At temperatures below the breakpoint (Tbreak), the apparent activation energy (Ea) is 10−20 kJ/mol, whereas above Tbreak, Ea increases to 40−50 kJ/mol. Assuming that the solubility of 1butylimidazole in the PS membranes does not change too much, the slope changes can be explained by the movement of the molecules through the free volume below Tg, where less activation energy is necessary compared to the penetration through the rubbery chains above Tg. Extra frozen free volume increases as temperature decreases further below Tg, and leads to higher than expected mobility of tracer molecules.33,34 For example, Frick et al. followed the diffusivity of Aberchrome 540 dye through the Tg in PS/toluene solutions by forced Rayleigh scattering. There was an abrupt change in the temperature dependence of the diffusivity at the Tg of PS/toluene, with a jump of Ea upon heating.34 They qualitatively explained the abrupt change by the Vrentas-Duda model. In this model, the free volume of a polymer/solvent mixture below Tg has a term to incorporate extra “trapped” free volume, which induces the activation energy changes at the Tg of the polymer mixture.35 They also reported the diffusivity of toluene (Dtoluene) from Tg to Tg + 20 °C in a concentrated PS/toluene solution, which had a Tg ∼ 35 °C. Using the Williams−Landel−Ferry (WLF) equation (Tr = 35 °C, log(Dr) = −7.939, c1 = 4.37, and c2 = 109.80 (°C), see Supporting Information), the temperaturedependent permeation rate through the PS membranes can be compared to other small molecule mobilities, as shown in Figure 6a,b. When the temperature increased from 47 to 69 °C, 1/τi increased by a factor of 3.4, which corresponds well with the prediction from the WLF equation, which is a factor of 4.0. Yasuda et al. also studied gas (He, CO2, O2, Ar, N2) permeability through poly(acrylonitrile-co-methyl acrylate) films (Tg = 65 °C).36 In the Arrhenius plot, the slope changed for all gases except He at the Tg of the polymer, from Ea ≈ 26 kJ/mol (below Tg) to Ea ≈ 63 kJ/mol (above Tg). Similar Arrhenius plots of diffusivity and permeability are observed for other small molecules.37,38 Across the Tg, however, an “on−off” switch for transport was not observed with the glassy membranes, although previous studies with bulk polymers showed significant fluidity changes across Tg.6,34,39 Abrupt changes in permeability can be realized when the ratio of the activation energies above and below Tg (λ = Ea,rubbery/Ea,glassy)35 is much less than 1. Frick et al. showed that λ of D was reduced with the amount of solvent in the PS film,34 which was anticipated by the Vrentas−Duda model.35 Thus, the exchange rates did not change significantly through the expected Tg of the membranes, presumably because the λ of the plasticized PS membranes is around 0.3, which is not much less than 1.

the glassy membrane could be tuned simply by increasing the temperature above the Tg of the membrane. An Arrhenius plot of the escape rate of 1-butylimidazole exhibited a similar trend as other permeation studies, with different apparent activation energies at high and low temperatures, and the transition point corresponded well to the onset of the glass transition of PSPEO micelles in the IL. These permeability studies will provide important information on designing polymersome reactors, because one key to the polymersome nanoreactor is its molecular transport behavior. Furthermore, membrane fluidity control can be applied to the delivery of actives (e.g., pharmaceutical or agricultural agents) at a specific temperature.



ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.5b08425. Size exclusion chromatography traces of PS-PEO block polymers, 1H NMR spectra of SO(14−2.5), temperature calibration curve for the NMR spectrometer with ethylene glycol, the echo-decays of poly(ethylene glycol) in [EMIM][TFSI] saturated D2O, the echo-decay curves of 1-butylimidazole in BO(9−3) solution, PS membrane thickness of various PS-PEO polymersomes, echo decays of SO polymersomes at 25 °C, thickness normalized escape rate, WLF fitting with toluene diffusivity (cm2/s) in PS/toluene solution, and effective relaxation times (PDF)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by the National Science Foundation through Award DMR-1206459. NMR instrumentation was supported by the Office of the Vice President of Research, College of Science and Engineering, and the Department of Chemistry at the University of Minnesota.



REFERENCES

(1) Cussler, E. L. Diffusion: Mass Transfer in Fluid Systems, 3rd ed.; Cambridge University Press: New York, 2009. (2) Narkwiboonwong, P.; Tumcharern, G.; Potisatityuenyong, A.; Wacharasindhu, S.; Sukwattanasinitt, M. Aqueous Sols of Oligo(ethylene glycol) Surface Decorated Polydiacetylene Vesicles for Colorimetric Detection of Pb2+. Talanta 2011, 83, 872−878. (3) Yu, S.; Azzam, T.; Rouiller, I.; Eisenberg, A. “Breathing” Vesicles. J. Am. Chem. Soc. 2009, 131, 10557−10566. (4) Kim, K. T.; Cornelissen, J. J. L. M.; Nolte, R. J. M.; van Hest, J. C. M. A Polymersome Nanoreactor with Controllable Permeability Induced by Stimuli-Responsive Block Copolymers. Adv. Mater. 2009, 21, 2787−2791. (5) Meeuwissen, S. A.; Kim, K. T.; Chen, Y.; Pochan, D. J.; van Hest, J. C. M. Controlled Shape Transformation of Polymersome Stomatocytes. Angew. Chem., Int. Ed. 2011, 50, 7070−7073. (6) Plazek, D. J.; O’Rourke, V. M. Viscoelastic Behavior of Low Molecular Weight Polystyrene. J. Polym. Sci. A−2 Polym. Phys. 1971, 9, 209−243. (7) Bai, Z.; Lodge, T. P. Polymersomes with Ionic Liquid Interiors Dispersed in Water. J. Am. Chem. Soc. 2010, 132, 16265−16270.



SUMMARY In this study, the permeability of small molecules through polymersome membranes was controlled by modifying the chemical composition, from rubbery PB to near glassy PS, and by varying temperature. The polymersomes contained an ionic liquid interior and were dispersed in water. The molecular exchange rate was monitored by the PFG-NMR technique. By employing glassy PS membranes, the molecular exchange was reduced by a factor of about three for a given polymersome size and membrane thickness. However, the exchange rate through 15061

DOI: 10.1021/acs.jpcb.5b08425 J. Phys. Chem. B 2015, 119, 15054−15062

Article

The Journal of Physical Chemistry B

Methylimidazolium-Based Room Temperature Ionic Liquids. Chem. Commun. 2005, 3469−3471. (30) Sanada, Y.; Akiba, I.; Sakurai, K.; Shiraishi, K.; Yokoyama, M.; Mylonas, E.; Ohta, N.; Yagi, N.; Shinohara, Y.; Ameniya, Y. Hydrophobic Molecules Infiltrating into the Poly(ethylene glycol) Domain of the Core/Shell Interface of a Polymeric Micelle: Evidence Obtained with Anomalous Small-Angle X-ray Scattering. J. Am. Chem. Soc. 2013, 135, 2574−2582. (31) Wu, J.; Eisenberg, A. Proton Diffusion across Membranes of Vesicles of Poly(styrene-b-acrylic acid) Diblock Copolymers. J. Am. Chem. Soc. 2006, 128, 2880−2884. (32) Mattsson, J.; Forrest, J. A.; Börjesson, L. Quantifying Glass Transition Behavior in Ultrathin Free-Standing Polymer Films. Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top. 2000, 62, 5187−5200. (33) Ramesh, N.; Davis, P. K.; Zielinski, J. M.; Danner, R. P.; Duda, J. L. Application of Free-Volume Theory to Self Diffusion of Solvents in Polymers below the Glass Transition Temperature: A Review. J. Polym. Sci., Part B: Polym. Phys. 2011, 49, 1629−1644. (34) Frick, T. S.; Huang, W. J.; Tirrell, M.; Lodge, T. P. Probe Diffusion in Polystyrene/Toluene Solutions. J. Polym. Sci., Part B: Polym. Phys. 1990, 28, 2629−2649. (35) Vrentas, J. S.; Duda, J. L.; Ling, H.-C. Influence of the Glass Transition on Solvent Self-diffusion in Amorphous Polymers. J. Polym. Sci., Part B: Polym. Phys. 1988, 26, 1059−1065. (36) Yasuda, H.; Hirotsu, T. The Effect of Glass Transition on Gas Permeabilities. J. Appl. Polym. Sci. 1977, 21, 105−112. (37) Ehlich, D.; Sillescu, H. Tracer Diffusion at the Glass Transition. Macromolecules 1990, 23, 1600−1610. (38) Ashworth, A. J.; Wickham, M. D. Permeation of Dichloromethane, Nitroethane and Tetrachloroethylene through Poly(hexafluoropropylene-co-tetrafluoroethylene). J. Membr. Sci. 1987, 34, 225−240. (39) Karlsson, O. J.; Stubbs, J. M.; Karlsson, L. E.; Sundberg, D. C. Estimating Diffusion Coefficients for Small Molecules in Polymers and Polymer Solutions. Polymer 2001, 42, 4915−4923.

(8) So, S.; Lodge, T. P. Rate of Molecular Exchange through the Membranes of Ionic Liquid Filled Polymersomes Dispersed in Water. J. Phys. Chem. C 2014, 118, 21140−21147. (9) Mok, M. M.; Lodge, T. P. Temperature-based Fluorescence Measurements of Pyrene in Block Copolymer Micelles: Probing Micelle Core Glass Transition Breadths. J. Polym. Sci., Part B: Polym. Phys. 2012, 50, 500−515. (10) Pfeuffer, J.; Flögel, U.; Dreher, W.; Leibfritz, D. Restricted Diffusion and Exchange of Intracellular Water: Theoretical Modelling and Diffusion Time Dependence of 1H NMR Measurements on Perfused Glial Cells. NMR Biomed. 1998, 11, 19−31. (11) Förster, S.; Borchert, K. Polymer Vesicles. In Encyclopedia of Polymer Science and Technology; John Wiley & Sons: New York, 2005. (12) Leson, A.; Hauschild, S.; Rank, A.; Neub, A.; Schubert, R.; Förster, S.; Mayer, C. Molecular Exchange through Membranes of Poly(2-vinylpyridine-block-ethylene oxide) Vesicles. Small 2007, 3, 1074−1083. (13) Hillmyer, M. A.; Bates, F. S. Synthesis and Characterization of Model Polyalkane−Poly(ethylene oxide) Block Copolymers. Macromolecules 1996, 29, 6994−7002. (14) So, S.; Lodge, T. P. Interfacial Tension-hindered Phase Transfer of Polystyrene-b-poly(ethylene oxide) Polymersomes from a Hydrophobic Ionic Liquid to Water. Langmuir 2015, 31, 594−601. (15) Simone, P. M.; Lodge, T. P. Micellization of PS-PMMA Diblock Copolymers in an Ionic Liquid. Macromol. Chem. Phys. 2007, 208, 339−348. (16) Meli, L.; Lodge, T. P. Equilibrium versus Metastability: High Temperature Annealing of Spherical Block Copolymer Micelles in an Ionic Liquid. Macromolecules 2009, 42, 580−583. (17) Raiford, D. S.; Fisk, C. L.; Becker, E. D. Calibration of Methanol and Ethylene Glycol Nuclear Magnetic Resonance Thermometers. Anal. Chem. 1979, 51, 2050−2051. (18) Wu, D.; Chen, A.; Johnson, C. S. An Improved Diffusion Ordered Spectroscopy Experiment Incorporating Bipolar Gradient Pulses. J. Magn. Reson., Ser. A 1995, 115, 260−264. (19) Ochab-Marcinek, A.; Holyst, R. Scale-dependent Diffusion of Spheres in Solutions of Flexible and Rigid Polymers: Mean Square Displacement and Autocorrelation Function for FCS and DLS Measurements. Soft Matter 2011, 7, 7366−7374. (20) Kärger, J. NMR Self-diffusion Studies in Heterogeneous Systems. Adv. Colloid Interface Sci. 1985, 23, 129−148. (21) Cho, C.-H.; Hong, Y.-S.; Kang, K.; Volkov, V. I.; Skirda, V.; Lee, C.-Y. J.; Lee, C.-H. Water Self-diffusion in Chlorella sp. Studied by Pulse Field Gradient NMR. Magn. Reson. Imaging 2003, 21, 1009− 1017. (22) Linders, J.; Mayer, C.; Sekine, T.; Hoffmann, H. Pulsed-Field Gradient NMR Measurements on Hydrogels from Phosphocholine. J. Phys. Chem. B 2012, 116, 11459−11465. (23) Meli, L.; Santiago, J. M.; Lodge, T. P. Path-Dependent Morphology and Relaxation Kinetics of Highly Amphiphilic Diblock Copolymer Micelles in Ionic Liquids. Macromolecules 2010, 43, 2018− 2027. (24) Shannon, M. S.; Bara, J. E. Properties of Alkylimidazoles as Solvents for CO2 Capture and Comparisons to Imidazolium-Based Ionic Liquids. Ind. Eng. Chem. Res. 2011, 50, 8665−8677. (25) Bai, Z.; Zhao, B.; Lodge, T. P. Bilayer Membrane Permeability of Ionic Liquid-Filled Block Copolymer Vesicles in Aqueous Solution. J. Phys. Chem. B 2012, 116, 8282−8289. (26) Battaglia, G.; Ryan, A. J.; Tomas, S. Polymeric Vesicle Permeability: A Facile Chemical Assay. Langmuir 2006, 22, 4910− 4913. (27) Choucair, A.; Soo, P. L.; Eisenberg, A. Active Loading and Tunable Release of Doxorubicin from Block Copolymer Vesicles. Langmuir 2005, 21, 9308−9313. (28) Rein, D. H.; Baddour, R. F.; Cohen, R. E. Gas Solubility and Diffusion in a Polystyrene−Polybutadiene Block Copolymer. J. Appl. Polym. Sci. 1992, 45, 1223−1227. (29) Lee, S. H.; Lee, S. B. The Hildebrand Solubility Parameters, Cohesive Energy Densities and Internal Energies of 1-Alkyl-315062

DOI: 10.1021/acs.jpcb.5b08425 J. Phys. Chem. B 2015, 119, 15054−15062