Synergistic Increase in Ionic Conductivity and ... - ACS Publications

Jul 7, 2015 - preferred polymer of these two is SMS. However, the high glass transition temperature. (Tg) of PMMA limits the usable SMS polymer concen...
0 downloads 4 Views 3MB Size
Article pubs.acs.org/Macromolecules

Synergistic Increase in Ionic Conductivity and Modulus of Triblock Copolymer Ion Gels Boxin Tang,† Scott P. White,† C. Daniel Frisbie,*,† and Timothy P. Lodge*,†,‡ †

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

ABSTRACT: Ion gels formed with ABA triblock polymers and ionic liquids (IL) have recently attracted significant attention. Because of their high ionic conductivity, high capacitance, and good mechanical integrity, ion gels prepared from triblock polymers of polystyrene-b-poly(methyl methacrylate)-b-polystyrene (SMS) and polystyrene-b-poly(ethylene oxide)-b-polystyrene (SOS) and an IL 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([EMI][TFSI]) have been successfully applied as the dielectric layer in thin film transistors. However, water absorption can negatively affect the stability of the dielectric layer and lead to electrical breakdown. Consequently, the preferred polymer of these two is SMS. However, the high glass transition temperature (Tg) of PMMA limits the usable SMS polymer concentration in order to ensure comparable ionic conductivity to that of SOS ion gel; this constraint limits the modulus of the gel to about 103 Pa. In this work, we developed a new ABA triblock ion gel system using poly(ethyl acrylate) (PEA) as a low Tg and hydrophobic midblock. The low Tg of the midblock ensures the ionic conductivity of the resulting ion gels is comparable to that of SOS ion gels at polymer concentrations up to 50 wt %, which is a significant improvement relative to the currently used SMS ion gels. Additionally, by decreasing the size of the midblock at constant polymer concentration, the modulus and ionic conductivity of the ion gels increase synergistically. This interesting and counterintuitive effect reflects the concurrent increase in the number density and chain stretching of midblocks, accompanied by a net reduction in midblock concentration within the conducting phase. We demonstrate that electrolyte gated transistors (EGTs) made with SEAS ion gels have improved stability under ambient humid conditions in comparison to those made with SOS ion gels.



modulus (ca. 103 Pa) make such ion gels applicable as dielectric layers in EGTs.35−37 The reversible physical cross-links of the ion gels also make solution processing a viable fabrication pathway. The electrolyte dielectric layer of an EGT needs to have a high ionic conductivity to ensure rapid response under high frequency operation, while maintaining a sufficiently high modulus so that the dielectric layer can be strong enough to support metal gate electrodes and withstand stresses induced by roll-to-roll fabrication processes.38−40 The ionic conductivity of an ion gel is influenced by the Tg of the conducting phase. The lower the Tg of the conducting phase, the lower the resistance to ionic motion and the higher the conductivity. The conductivity, in turn, decreases with polymer loading, primarily due to the increase in Tg of the conducting phase. To date, the two ABA triblock polymer systems most thoroughly investigated for ion gel applications have been poly(styrene-bethylene oxide-b-styrene) (SOS) and poly(styrene-b-methyl methacrylate-b-styrene) (SMS).35,36 The Tg of [EMI][TFSI] is as low as −87 °C,41 which is much lower than that of either of

INTRODUCTION Ionic liquids (IL) are room temperature molten salts comprising bulky ions. Because of their nonvolatility, nonflammability, high electrochemical stability, and high ionic conductivity,1−4 ILs have been widely used as electrolyte components in electrochemical actuators,5−8 supercapacitors,9−13 lithium ion batteries,14−16 electrochemiluminescent displays,17,18 and electrolyte gated transistors (EGTs).19−23 The variety of component ion combinations gives a large degree of freedom in tailoring this class of materials for specific applications. The liquid nature of an IL limits its application in a solidstate material. To solve this problem, gelators can be mixed with ILs to provide mechanical integrity. Common categories of gelators include chemically cross-linked homopolymers24−29 and physically cross-linked block polymers.30−34 The latter approach is a particularly versatile method, as it allows for readily tuned morphologies and properties through variations of block length and sequence. We have previously reported ion gels prepared by dispersing an ABA triblock polymer in a hydrophobic, highly ionically conductive IL, [EMI][TFSI].35,36 In this system, the B midblock is soluble in the IL while the A end blocks are insoluble and self-assemble into micellar crosslinks. The high ionic conductivity (ca. 10−2 S/cm) and modest © XXXX American Chemical Society

Received: April 28, 2015 Revised: June 19, 2015

A

DOI: 10.1021/acs.macromol.5b00882 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 1. Cartoon illustration of the influence of shortening midblock size.

Scheme 1. Synthetic Route to SEAS Triblock Polymer

the IL-soluble midblocks. The higher Tg of M (ca. 100 °C) compared to O (ca. −60 °C) implies that SOS is favored over SMS in terms of conductivity. Note that the “obstruction” effect of the impermeable polystyrene cross-links has been shown to reduce the conductivity only modestly.35 In general, the modulus of the ion gel increases with polymer loading as well as with the number of effective network strands at fixed polymer loading. Because of a significantly lower side-group mass and resulting higher number of network strands per unit volume, SOS ion gels have higher modulus than SMS ion gels at a given polymer loading. However, PEO is hygroscopic, and water is known to degrade the performance of an organic semiconductor such as poly(3-hexylthiophene) (P3HT) over time.42−44 Consequently, SMS ion gels have been more widely used in EGT applications, which, in turn, limits the tunability of the ion gel in terms of ionic conductivity and modulus.35 Here, we adopted a new midblock, poly(ethyl acrylate) (PEA), that has a low Tg of −24 °C to approach the low Tg of PEO, yet it is at least as hydrophobic as PMMA. The relative ease of RAFT over anionic polymerization is another advantage for SEAS. Furthermore, as shown in Figure 1, we highlight the roles of the midblock chain stretching and reduction of midblock concentration in the conducting phase and show that decreasing the midblock molecular weight of the poly(styreneb-ethyl acrylate-b-styrene) (SEAS) ion gels simultaneously increases both modulus and ionic conductivity at fixed polymer loading. SEAS ion gels outperform SMS gels with respect to both mechanical properties and ionic conductivity. Furthermore, EGTs with SEAS ion gels exhibited excellent stability in

an ambient testing environment and were significantly more stable than EGTs using SOS ion gels.



EXPERIMENTAL SECTION

Materials Synthesis. The ionic liquid [EMI][TFSI] was synthesized following a previously reported procedure.45 Four SEAS triblock polymers with various midblock sizes were synthesized using reversible addition/fragmentation and chain transfer (RAFT) polymerization reactions of the PEA blocks from a PS−CTA−PS difunctional macroinitiator, following a procedure adapted from the literature.46 The synthesis route is illustrated in Scheme 1. Styrene, S,S′-di(1phenylethyl) trithiocarbonate, ethyl acrylate and 2-butanone were used as received (Sigma-Aldrich). Styrene was reacted with S,S′-di(1phenylethyl) trithiocarbonate, a symmetric difunctional chain transfer agent, to yield a narrowly distributed PS-CTA-PS. The molecular weight of the PS end blocks was determined by size exclusion chromatography (SEC) to be 4 kg/mol. Then, the resulting PS macroCTA was reacted with ethyl acrylate for various lengths of time (1, 1.5, 3, and 4 h) to yield SEAS triblock polymers with different midblock molecular weights. The molecular characterization results are shown in Table 1. Dispersities and molecular weights of the triblock polymers

Table 1. Molecular Characterization of PS−PEA−PS Triblock Polymers

B

polymer

Mn,PS (kDa)

Mn,PEA (kDa)

Đ

f PEA,bulk

SEAS(4-23-4) SEAS(4-36-4) SEAS(4-60-4) SEAS(4-150-4)

8 8 8 8

23 36 60 150

1.05 1.04 1.04 1.14

0.71 0.80 0.87 0.94

DOI: 10.1021/acs.macromol.5b00882 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 2. Ionic conductivity comparison of (a) SOS(3-35-3) vs SEAS(4-23-4) and (b) SMS(17-86-17) vs SEAS(4-23-4). The vol % of the corresponding components are identical between SMS(17-86-17) and SEAS(4-23-4) ion gels at the same wt % loading. Squares indicate SEAS(4-234) ion gels; triangles indicate SOS(3-35-3) ion gels; circles indicate SMS(17-86-17) ion gels. Black indicates 10 wt %, red indicates 20 wt %, green indicates 30 wt %, blue indicates 40 wt %, and purple indicates 50 wt %. determined by the value of G′ where G″ exhibits a local minimum. The modulus of the remainder of the gel samples was determined from the high-frequency plateau. EGT Device Fabrication and Testing. Source, drain, and gate electrode contacts were patterned on a silicon wafer by photolithography with 5 nm of chromium and 30 nm of gold. The semiconductor channel width to length ratio was 100. The semiconductor, P3HT (Rieke Metals Inc.), was dissolved in chloroform/terpineol (9:1 weight ratio) at a concentration of ca. 1 mg/mL. The polymer concentration in the ion gels was controlled to be 10 wt %, and the gel was dissolved in ethyl acetate at 10 wt %. The gate electrode, poly(3,4-ethylenedioxythiophene) poly(styrenesulfonate) (PEDOT:PSS) (H.C. Starck) in a 1.5 wt % aqueous solution, was dissolved in ethylene glycol at 6 wt %. The P3HT, ion gel, and PEDOT:PSS were printed, in that order, on silicon wafers with patterned electrode contacts using a commercially available aerosol ink jet printer (Optomec, Inc.) and then annealed at 120 °C for 30 min. The EGTs were kept in a sealed chamber with dry-air flow, and the humidity was kept at 10% RH in between electrical measurements. Measurements were done in ambient air (∼40% R.H.) using Keithley 2400 and Keithley 2611B electrometers.

were determined by size exclusion chromatography (SEC) in THF at room temperature on an Agilent Technologies 1260 Infinity system using a multiangle light scattering detector (Wyatt Technology Dawn Heleos-II), and 1H NMR spectroscopy (Varian Inova 500). The samples are designated SEAS(4-xx-4), where the numbers in parentheses indicate the block molecular weights in kDa. The last column in Table 1 shows the volume fractions of PEA in the bulk polymers. Both polymers and IL were dried under vacuum (∼100 mTorr) at 80 °C for 48 h and stored in a desiccator before use. Preparation of Ion Gels. SEAS polymer and IL were combined in a predetermined ratio in a scintillation vial and codissolved in dichloromethane. The mixture was kept under nitrogen gas flow for 24 h before being further dried under vacuum (∼100 mTorr) at 80 °C for 48 h to remove all of the cosolvent. Determination of Tg. Differential scanning calorimetry measurements of ion gels were carried out using a TA Instruments Q1000 DSC with liquid N2 cooling capability. A sample of 5−10 mg was contained in a hermetically sealed aluminum pan. The samples were initially heated up to 160 °C and annealed for 5 min to erase prior thermal history and then cooled rapidly to −150 °C. Subsequently, they were heated to 160 °C at 10 °C/min. Glass transition temperatures were acquired from the second heating curves. Determination of Ionic Conductivity. Impedance spectroscopy was conducted using a two-point probe on a Solartron 1255B frequency response analyzer connected to a Solartron SI 1287 electrochemical interface. Samples were initially sandwiched in a Teflon spacer of 4 mm diameter and 2 mm thickness between two stainless steel electrodes. Then, the assembly was annealed at 80 °C under vacuum to ensure uniformity of the sample prior to measurement. During a measurement, the sample temperature was varied from 30 to 150 °C in 10 deg intervals and was controlled by a custom-built heating stage with a temperature feedback loop monitored by an independent thermocouple placed next to the sample. The ionic conductivity, σ, was calculated as σ = l/Ra, where l is the sample thickness, a is the superficial area of the sample, and R is the bulk resistance determined from the real part of the impedance, Z′, in the high-frequency plateau. Viscoelastic Properties. Rheological measurements were performed on a Rheometric Scientific ARES-2 rheometer using 25 mm parallel plates, with a gap spacing of ca. 1 mm. At each temperature, the sample was annealed for ∼15 min before measurement, and the gap spacing was adjusted for thermal expansion of the tool set and the sample. The strain was set to 2% to ensure that frequency sweep measurements remained within the linear regime. Time−temperature superposition was used to generate an approximate master plot for each sample. The modulus for 10 wt % polymer concentration ion gel samples and SEAS(4-150-4) samples of various concentrations was



RESULTS AND DISCUSSION Ionic Conductivity of SEAS, SOS, and SMS Ion Gels. The electrical impedance behavior of SOS and SMS ion gels has been previously reported.35,36 The temperature dependence of the conductivity can be approximated using the Vogel− Tammann−Fulcher (VTF) equation:24 ⎛ B ⎞ σ = σ0 exp⎜ − ⎟ ⎝ T − T0 ⎠

(1)

where σ0 is a prefactor, B is a constant related to the entropic barrier for ion conduction, and T0 is the Vogel temperature, which is correlated with the Tg of the conducting phase. Because of the low Tg of PEO, the ionic conductivities of the SOS ion gels are significantly higher than those of the SMS ion gels at and above 20 wt % polymer. However, the hygroscopic nature of PEO makes it a less favorable candidate for the dielectric layer of an EGT. As a result, SMS ion gels are currently preferred in EGT applications. However, the Tg of PMMA is about 100 °C while that of PEO is −60 °C. The large increase in Tg of the block polymer midblock in SMS ion gels limits the applications of these ion gels to lower polymer concentrations as the dielectric layer of rapid-switching EGTs.47 C

DOI: 10.1021/acs.macromol.5b00882 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 3. (a) Time−temperature superposition (tTS) master curves of dynamic storage moduli G′ (closed squares) and loss moduli G″ (open circles) of SEAS(4-23-4) ion gel of 10 wt % polymer. (b) tTS master curve of dynamic storage moduli G′ (closed squares) and loss moduli G″ (open circles) of SEAS(4-23-4) ion gel of 20 wt % polymer. The reference temperature is 30 °C.

exhibit terminal behavior, which indicates that the ion gel is thermoreversible and becomes liquid at elevated temperatures. The characteristic time implied by the crossover (ca. 105 s at 30 °C) reflects sufficient pullout of the PS blocks to allow flow. Starting at 20 wt %, no crossover of G′ and G″ is observed on the master plot, indicating the gels are elastic solids up to at least 160 °C. Instead, two distinctive plateaus are observed in G′. As previously, we attribute the higher frequency plateau to the network of the midblocks (which may or may not be entangled) anchored by frozen micellar cores, and the partial relaxation at lower frequency to the internal relaxation of PS blocks within their cores above the PS glass transition.35 In this case, we take the high frequency/room temperature plateau to be G for application purposes. Figure 4 shows the modulus of the four SEAS ion gels at 30 °C, with loadings from 10 to 50 wt % polymer. It is clear from the figure that for a particular kind of ion gel the modulus increases as polymer loading increases. Meanwhile, at a given polymer concentration, the modulus increases significantly as the midblock size decreases from 150 to 23 kDa. Particularly for

Currently, such ion gels are not mechanically robust enough for advanced manufacturing methods such as roll-to-roll printing and metal vapor deposition. Instead, PEA has a Tg of −24 °C, which is much lower than that of the PMMA. Figure 2 shows a comparison of the temperature-dependent ionic conductivity for SEAS(4-23-4) with SOS(3-35-3) (Figure 2a) and SMS(1786-17) (Figure 2b) ion gels. The temperature-dependent ionic conductivity of pure [EMI][TFSI] is also plotted.41 Since PEA has a Tg much closer to that of PEO than PMMA, the ionic conductivities of SEAS(4-23-4) ion gels are quite comparable to those of SOS(3-35-3) ion gels, as shown in Figure 2a. On the other hand, the ionic conductivity comparison between SEAS(4-23-4) and SMS(17-86-17) is shown in Figure 2b. Despite the fact that the chain lengths of the two polymers are quite different, the volume fractions of the midblocks, end blocks, and IL are approximately the same, so that the comparison of ionic conductivity between the two gels is appropriate. SEAS(4-23-4) ion gels exhibit a much smaller reduction in ionic conductivity as polymer loading increases, in comparison to SMS(17-86-17) ion gels. For example, the ionic conductivity of the SEAS(4-23-4) ion gel is twice from that of a SMS(17-86-17) ion gel at a concentration of only 20 wt %. At 50 wt % polymer, the ionic conductivity of the SEAS(4-23-4) ion gels is more than 1 order of magnitude higher than that of the corresponding SMS ion gels. Modulus Comparison of SEAS, SMS, and SOS Ion Gels. The rheological behavior of SOS and SMS ion gels has been studied in detail.35 When the ABA triblock polymer is in a midblock selective solvent, the midblock is solvated by solvent while the end blocks are insoluble and self-assemble into micelles, forming a network structure. For the lower three midblock sizes, the ion gels exhibit BCC packing above 20 wt % polymer, which is confirmed by small-angle X-ray scattering shown in the Supporting Information. At 20 wt % polymer, Nagg is typically above 100 and increases with polymer concentration due to the screened corona crowding,48 and decreases with midblock size as corona crowding increases with corona block size.48 As shown in Figure 3, the SEAS ion gels behave similarly to the ion gels previously studied. At 10 wt % polymer, the time−temperature superposition (tTS) master plot shows a single plateau. At low frequency, G′ and G″ cross over and

Figure 4. Modulus vs polymer wt % comparisons of SEAS with various midblock molecular weights at 30 °C. Red triangles represent SEAS(423-4), green represent SEAS(4-36-4), navy blue represent SEAS(4-604), and purple represent (4-150-4). D

DOI: 10.1021/acs.macromol.5b00882 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules the lower molecular weight midblocks, the moduli easily exceed 10 kPa. According to rubber elasticity theory, the plateau modulus G can be estimated as

G = υkBT

Table 2. Number of Entanglements per Polymer Chain, Calculated Modulus Based on Eq 5, Measured Modulus of the Ion Gels with Various Midblock Sizes, and the Inferred fd2/⟨h0⟩ Factor Values

(2)

G (kPa)

where kB is Boltzmann’s constant and υ is the number of elastically effective strands per unit volume. In the unentangled regime and for an ideal network, υ = cNA/Mtriblock, where c is the concentration of polymer in the gel (w/v), NA is Avogadro’s number, and Mtriblock is the molecular weight of the triblock polymer. Thus, eq 2 can be written as49,50 G=f

cNART d 2 M triblock ⟨h0 2⟩

polymer SEAS(4-23-4)

SEAS(4-36-4)

(3)

where 0 ≤ f ≤ 1 is the bridging fraction, indicating how many of the midblock chains are elastically effective, R is the gas constant, d is the average distance between PS micelle crosslinks, and ⟨h02⟩ is the mean-square end-to-end distance for an equilibrium solution of the midblock polymers with a molecular weight equal to the average molecular weight between PS crosslinks. To account for the effect of midblock entanglements, an estimate of the number of entanglements per molecule, ne, for a polymer in a good solvent is given by49,51 ne = Neϕp1.25 =

M 1.25 ϕ Me p

SEAS(4-60-4)

SEAS(4-150-4)

(4)

where Ne is the number of entanglements per polymer chain in a homopolymer melt, Me is the entanglement molecular weight, and ϕp is the polymer volume fraction. As an example, for 20 wt % SEAS(4-60-4), the volume fraction of the midblock is 21% and ne is estimated to be ∼1. In this case, the average bridging midblock is separated into two elastically effective strands, which will increase the calculated modulus by at most a factor of 2 in comparison to the calculated result from eq 3. Thus, eq 3 can be modified as G=f

cNART d 2 (ne + 1) M triblock ⟨h0 2⟩

polymer loading (wt %)

ne

calcd

exptl

fd2/⟨h0⟩

10 20 30 40 50 10 20 30 40 50 10 20 30 40 50 10 20 30 40 50

0.12 0.28 0.46 0.65 0.86 0.26 0.60 1.0 1.3 1.7 0.48 1.1 1.6 2.5 3.1 1.3 3.0 4.8 6.6 8.5

13 30 49 72 99 11 26 47 72 102 8 22 37 70 102 5 18 39 67 101

8 47 77 127 175 10 32 57 85 104 7 19 33 44 64 0.9 3 6 10 12

0.59 1.6 1.6 1.8 1.8 0.9 1.2 1.2 1.2 1.0 0.88 0.82 0.89 0.63 0.63 0.17 0.17 0.16 0.15 0.12

the calculated modulus is not accurate. Also, it is not likely that all the chains in the ion gel are elastically effective. Thus, in Table 2, a higher fd2/⟨h02⟩ value suggests that the polymer chains are more strongly stretched, and thus the modulus will be higher, as illustrated in eq 5. As the midblock size becomes smaller, the chains are more stretched between PS cross-links and thus fd2/⟨h02⟩ becomes larger. As shown in Figure 5e, the inferred values of fd2/⟨h02⟩ become larger as the midblock chains become smaller. This effect is more significant at higher polymer concentrations due to the fact that bridging is presumably favored at high concentrations where f becomes close to 1. In this scenario, SEAS(4-23-4) shows the highest modulus because the midblock chains are more strongly stretched. On the other hand, midblock chains in the SEAS(4150-4) ion gels are not well extended, and may even be compressed, which could promote self-looping. Figure 6a provides a direct comparison of the modulus versus ionic conductivity for SEAS(4-23-4) and the SMS(17-86-17) ion gels. Because of the significantly lower Tg of the midblock, the SEAS(4-23-4) ion gels have a much smaller drop in ionic conductivity in comparison to that of the SMS(17-86-17) ion gels with increasing polymer content. Meanwhile, the smaller midblock of the SEAS(4-23-4) facilitates a stretched conformation that improves its mechanical strength. Consequently, there is simultaneous improvement of both ionic conductivity and modulus by changing from SMS(17-86-17) to SEAS(4-234) ion gels. As illustrated in Figure 6b, tuning of the midblock length results in a similar modulus of SEAS(4-23-4) in comparison to SOS(3-35-3) ion gels, over the entire range from 10 to 50 wt % polymer, which represents a significant improvement from SMS(17-86-17). Ionic Conductivity as a Function of SEAS Midblock Size. As shown above, SEAS(4-23-4) ion gels have a 4 times higher PS volume fraction, which is the nonconducting phase,

(5)

The calculated ne and modulus are listed in Table 2. The calculated modulus assumes that every polymer chain in the ion gel is elastically effective and that the midblock is in the melt state so that the average distance between cross-links is equal to the unperturbed end-to-end distance of the midblock. Consequently, the factor of fd2/⟨h02⟩ is assumed to be 1 in the calculation. As shown in Table 2 and Figure 5, the calculated modulus values agree very well with the experimental values for all but the SEAS(4-150-4) ion gels. In other words, values of fd2/⟨h02⟩ inferred from the comparison of experimental and calculated moduli lie between 0.5 and 2 for the three shorter midblock gels, which is quite satisfactory given the number of assumptions underlying eq 5. The calculated modulus does not depend strongly on midblock size at a given polymer loading due to the balance in block size, polymer concentration, c, and number of entanglements per chain, ne. However, as shown in Figure 4, the measured modulus increases monotonically as the midblock size becomes smaller at a given polymer loading. The midblock conformations are a possible source of this trend. Since the midblock is in a good solvent in the ion gel case, the assumption that the midblock is in the melt state for E

DOI: 10.1021/acs.macromol.5b00882 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 5. (a) Modulus of SEAS(4-150-4) ion gels (purple) with prediction. (b) Modulus of SEAS(4-60-4) ion gels (navy blue) with prediction. (c) Modulus of SEAS(4-36-4) ion gels (green) with prediction. (d) Modulus of SEAS(4-23-4) ion gels (red) with prediction. (e) fd2/⟨h02⟩ factor of SEAS ion gels of various midblock sizes.

hypothetical lattice system occupied by ion and obstacle (PS) sites, and the ions cannot pass through the PS sites. In this case, the ratio between gel conductivity and the conductivity of the homopolymer solution that has the same conducting phase composition can be expressed as

in comparison to SEAS(4-150-4). Since the PS domains only serve as a mechanical support and are impermeable to ions, it is important to know the effect of the increase in PS volume fraction on conductivity as the chain length decreases. Intuitively, an increase in the nonconducting phase fraction should decrease the overall ionic conductivity due to a smaller fraction of the conducting phase. However, we have previously shown that the PS fraction has a very modest effect on ionic conductivity by comparing ion gels of different polymer concentrations to the classical Mackie−Meares obstruction model.52 The model assumes that the ions move in a

σgel σsolution

=

1−ϕ 1+ϕ

(6)

where ϕ is the volume fraction of PS. The remarkable agreement of this theory and experiments on SOS and SMS ion F

DOI: 10.1021/acs.macromol.5b00882 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 6. (a) G vs ionic conductivity comparisons of SEAS(4-23-4) and SMS(17-86-17) at 30 °C. Sky blue arrows indicate the difference between SEAS(4-23-4) and SMS(17-86-17) at the same polymer loading. (b) G vs polymer wt % comparisons of SEAS(4-23-4) and SOS(3-35-3) and SMS(17-86-17).

Figure 7. (a) Tg of the conducting phase vs polymer concentration of SEAS(4-23-4) and SEAS(4-150-4). (b) Ionic conductivity vs polymer concentration of SEAS(4-23-4) and SEAS(4-150-4) at 30 °C. Black represents SEAS(4-23-4) and red represents SEAS(4-150-4).

loading increases, as shown in Figure 7a. The differences in Tg shown in Figure 7a are directly reflected in the ionic conductivity of the ion gels. Shortening the midblock length increases the ionic liquid content in the IL + midblock phase at the same triblock polymer content. As a result, the ionic conductivity of SEAS(4-23-4) ion gels is higher than that of the SEAS(4-150-4) ion gels at the same triblock polymer concentration, as shown in Figure 7b. Comparison of modulus versus ionic conductivity of SEAS ion gels made with the shortest and longest midblock is summarized in Figure 8. At a given polymer concentration, ion gels with the shortest midblock exhibit significantly higher modulus than that of the longest midblock ion gels, while by shortening the midblock, the fraction of ionic liquid in the conducting phase increases at fixed polymer loading. As a result, the ionic conductivity with shortest midblock length is higher than that of longest midblock ion gels at a given polymer loading. Thus, both modulus and ionic conductivity improve synergistically. Stability of EGTs Fabricated with SEAS Ion Gels. Absorbed water can negatively impact the performance and stability of organic electronic devices.42−44,53 A more hydrophobic midblock should improve stability by mitigating the effect of ambient water or humidity. To investigate this, EGTs were fabricated with SOS(3-35-3) and SEAS(4-23-4) ion gels to compare their stability on the time scale of 1 week. Each EGT consisted of gold source and drain electrodes, a

gels indicated that there is only a 50% drop in conductivity by 20 vol % PS, which corresponds to about 50 wt % triblock. Consistent with this, Watanabe et al. also showed that an SMS ion gel with 48 vol % of PS in bulk SMS(9-19-9) has much higher ionic conductivity in comparison to SMS with only 18 vol % of PS in bulk SMS(2-19-2) at a given polymer concentration.8 The ion gels formed by SMS(9-19-9) have a larger ionic liquid content in the conducting phase at the same polymer loading in comparison to those made with SMS(2-192), and the increase in IL content influenced ionic conductivity more significantly than the effect of PS obstruction. Since the ion diffusion is mainly affected by their interactions with the IL solvable midblock chains, PS obstructions only moderately affected ionic conductivity.8,36,52 To interpret the effect of midblock length on ionic conductivity, glass transition temperatures of SEAS(4-23-4) and SEAS(4-150-4) ion gels are plotted against polymer content in Figure 7a. Since the conducting phase is the vast majority, a single Tg of the homogeneous midblock plus IL domain is observed; the Tg of PS is not resolved due to the low fraction within the ion gel. As discussed previously, the volume fraction of PS is higher in the ion gels with a shorter midblock length at the same polymer loading, which results in a progressively higher IL content in the conducting phase as polymer loading increases, reaching 6 vol % of difference by 50 wt % polymer. Consequently, the difference in the conducting phase Tg with midblock length becomes larger as polymer G

DOI: 10.1021/acs.macromol.5b00882 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

SEAS(4-23-4) ion gels is superior to many other reported EGT devices.52,54 The saturation mobility, μsat, of the semiconductor layer versus time is also shown in Figure 9b for both SEAS(4-23-4) and SOS(3-35-3) EGTs. The μsat is normalized to that of the first day and calculated from eq 8. The error bars represented the standard deviation over five devices. The EGTs fabricated by SEAS(4-23-4) maintained over 90% of the mobility of the first day, while the mobility of the EGTs fabricated by SOS(3-35-3) dropped to ca. 60% by the seventh day. Thus, EGTs with SEAS(4-23-4) ion gel layer show improved stability to ambient humidity and are significantly more stable than those with SOS(3-35-3) ion gels.



SUMMARY



ASSOCIATED CONTENT

We investigated a new ABA triblock polymer ion gel system with a low Tg hydrophobic midblock. Because of the low Tg of the midblock polymer, the SEAS ion gels demonstrated a much higher ionic conductivity than the previously studied SMS ion gels at the same polymer concentration. The ionic conductivity of SEAS ion gels doubled that of the previously studied SMS ion gels at 20 wt % polymer and is more than 1 order of magnitude higher by 50 wt % polymer. Furthermore, by tuning the midblock length, both the ionic conductivity and the modulus of the SEAS(4-23-4) ion gels increased synergistically from those of SEAS(4-150-4) ion gels at a given polymer concentration. EGTs made with SEAS and previously investigated SOS ion gels have been compared for stability under ambient humidity. Because of the hydrophobic midblock PEA, the EGTs with SEAS ion gels showed significantly better stability over EGTs with SOS ion gels, while at least matching them in terms of modulus and ionic conductivity. Overall, this work highlights the considerable flexibility of block polymer based ion gels in terms of tailoring the mechanical and electrical properties for specific applications.

Figure 8. GN vs ionic conductivity for SEAS(4-23-4) (red) and SEAS(4-150-4) (purple). Blue arrows represent the difference between SEAS(4-23-4) and SEAS(4-150-4).

semiconductor thin film of P3HT (50 nm) covered by an ion gel layer of ca. 1 μm thickness, and finally a gate electrode of printed PEDOT:PSS. The transfer curves (drain current ID versus gate voltage VG) of SEAS ion gels repeated over 7 days are shown in Figure 9a. For positive VG, a p-type semiconductor (P3HT) is off and the current is low and independent of VG, but for negative VG the current sharply increases due to positively charged carriers induced in the P3HT film. This relation is W C iμsat (VG − VT)2 ID = (8) 2L where W is the width of the semiconductor channel, L is the length of the semiconductor channel, Ci is the specific capacitance of the ion-gel/semiconductor interface, μsat is the hole carrier mobility in the saturation regime, and VT is the threshold voltage that marks the transition between ON and OFF regimes. The EGTs fabricated by SEAS(4-23-4) show excellent stability, and the transfer curves overlap from day 1 to day 7. Since the devices have been tested in ambient air with ca. 40% relative humidity, the stable performance of EGTs with

S Supporting Information *

1 H NMR spectra, SEC and DSC traces, glass transition data, ionic conductivities and VFT fits, volume fraction data, representative SAXS profiles, and fit parameters. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.5b00882.

Figure 9. (a) Overlay of day 1, day 4, and day 7 transfer curves of EGTs fabricated with SEAS(4-23-4) ion gels tested in ambient air moisture. The channel size is W/L = 100 and VD = −0.5 V. (b) Mobilities of EGTs fabricated with SEAS(4-23-4) and SOS(3-35-3) ion gels. The mobilities are normalized to the values of day 1, which is tested right after fabrication. H

DOI: 10.1021/acs.macromol.5b00882 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules



(26) Klingshirn, M. A.; Spear, S. K.; Subramanian, R.; Holbrey, J. D.; Huddleston, J. G.; Rogers, R. D. Chem. Mater. 2004, 16, 3091−3097. (27) Neouze, M. A.; LeBideau, J.; Gaveau, P.; Bellayer, S.; Vioux, A. Chem. Mater. 2006, 18, 3931−3936. (28) Matsumoto, K.; Endo, T. Macromolecules 2008, 41, 6981−6986. (29) Jana, S.; Parthiban, A.; Chai, C. L. L. Chem. Commun. 2010, 46, 1488−1490. (30) He, Y.; Boswell, P. G.; Buhlmann, P.; Lodge, T. P. J. Phys. Chem. B 2007, 111, 4645−4652. (31) He, Y.; Lodge, T. P. Chem. Commun. 2007, 2732−2734. (32) He, Y.; Lodge, T. P. Macromolecules 2008, 41, 167−174. (33) Noro, A.; Matsushita, Y.; Lodge, T. P. Macromolecules 2008, 41, 5839−5844. (34) Noro, A.; Matsushita, Y.; Lodge, T. P. Macromolecules 2009, 42, 5802−5810. (35) Zhang, S.; Lee, K.; Sun, J.; Frisbie, C. D.; Lodge, T. P. Macromolecules 2011, 44, 8981−8989. (36) Zhang, S.; Lee, K.; Frisbie, C. D.; Lodge, T. P. Macromolecules 2011, 44, 940−949. (37) Lee, K.; Zhang, S.; Lodge, T. P.; Frisbie, C. D. J. Phys. Chem. B 2011, 115, 3315−3321. (38) Perelaer, J.; Smith, P. J.; Mager, D.; Soltman, D.; Volkman, S. K.; Subramanian, V.; Korvink, J. G.; Schubert, U. S. J. Mater. Chem. 2010, 20, 8446−8453. (39) Strunskus, T.; Zaporojtchenko, V.; Behnke, K.; Bechtolsheim, C.; Faupel, F. Adv. Eng. Mater. 2000, 2 (8), 489−492. (40) Popoff, R. T. W.; Zavareh, A. A.; Kavanagh, K. L.; Yu, H. Z. J. Phys. Chem. C 2012, 116, 17040−17047. (41) Tokuda, H.; Ishii, K.; Susan, M. A. B. H.; Watanabe, M. J. Phys. Chem. B 2005, 109, 6103−6110. (42) Abdou, M. S. A.; Orfino, F. P.; Son, Y.; Holdcroft, S. J. Am. Chem. Soc. 1997, 119, 4518−4524. (43) Hoshino, S.; Yoshida, M.; Uemura, S.; Kodzasa, T.; Takada, N.; Kamata, T.; Yase, K. J. Appl. Phys. 2004, 95, 5088−5093. (44) Ficker, J.; Ullmann, A.; Fix, W.; Rost, H.; Clemens, W. J. Appl. Phys. 2003, 94, 2638−2641. (45) Bai, Z.; Lodge, T. P. J. Am. Chem. Soc. 2010, 132, 16265−16270. (46) Gu, Y.; Lodge, T. P. Macromolecules 2011, 44, 1732−1736. (47) Ha, M.; Seo, J. T.; Prabhumirashi, P. L.; Zhang, W.; Geier, M. L.; Renn, M. J.; Kim, C. H.; Hersam, M. C.; Frisbie, C. D. Nano Lett. 2013, 13, 954−960. (48) He, Y.; Li, Z.; Simone, P.; Lodge, T. P. J. Am. Chem. Soc. 2006, 128, 2745−2750. (49) Ferry, J. D. Viscoelastic Properties of Polymers, 3rd ed.; John Wiley & Sons: New York, 1980. (50) Seitz, M. E.; Burghardt, W. R.; Faber, K. T.; Shull, K. R. Macromolecules 2007, 40, 1218−1226. (51) Milner, S. T. Macromolecules 2005, 38, 4929−4939. (52) Mackie, J. S.; Meares, P. Proc. R. Soc. London 1955, A232, 498− 509. (53) McCulloch, I.; Heeney, M.; Bailey, C.; Genevicius, K.; MacDonald, I.; Shkunov, M.; Sparrowe, S.; Tierney, S.; Wagner, R.; Zhang, W.; Chabinyc, M. L.; Kline, R. J.; McGehee, M. D.; Toney, M. F. Nat. Mater. 2006, 5, 328−333. (54) Kim, S. H.; Hong, K.; Lee, K. H.; Frisbie, C. D. ACS Appl. Mater. Interfaces 2013, 5, 6580−6585.

AUTHOR INFORMATION

Corresponding Authors

*E-mail [email protected] (C.D.F). *E-mail [email protected] (T.P.L). Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors thank Dr. Yuanyan Gu and Dr. Lucas D. McIntosh for experimental assistance and Dr. Hongchul Moon for helpful discussions. C.D.F. and T.P.L. acknowledge financial support from the Air Force Office of Scientific Research under Grant FA9550-12-1-0067.



REFERENCES

(1) Welton, T. Chem. Rev. 1999, 99, 2071−2083. (2) Galinski, M.; Lewandowski, A.; Stepniak, I. Electrochim. Acta 2006, 51, 5567−55801. (3) MacFarlane, D. R.; Forsyth, M.; Howlett, P. C.; Pringle, J. M.; Sun, J.; Annat, G.; Neil, W.; Izgorodina, E. I. Acc. Chem. Res. 2007, 40, 1165−1173. (4) Wasserscheid, P.; Welton, T. Ionic Liquids in Synthesis, 2nd ed.; Wiley-VCH: Weinheim, Germany, 2008. (5) Lu, W.; Fadeev, A. G.; Qi, B.; Smela, E.; Mattes, B. R.; Ding, J.; Spinks, G. M.; Mazurkiewicz, J.; Zhou, D.; Wallace, G. G.; MacFarlane, D. R.; Forsyth, S. A.; Forsyth, M. Science 2002, 297, 983−987. (6) Ding, J.; Zhou, D.; Spinks, G.; Wallace, G.; Forsyth, S.; Forsyth, M.; MacFarlane, D. Chem. Mater. 2003, 15, 2392−2398. (7) Imaizumi, S.; Kato, Y.; Kokubo, H.; Watanabe, M. J. Phys. Chem. B 2012, 116, 5080−5089. (8) Imaizumi, S.; Kato, Y.; Kokubo, H.; Watanabe, M. Macromolecules 2012, 45, 401−409. (9) McEwen, A. B.; Ngo, E. L.; LeCompte, K.; Goldman, J. L. J. Electrochem. Soc. 1999, 146, 1687−1695. (10) Sato, T.; Masuda, G.; Takagi, K. Electrochim. Acta 2004, 49, 3603−3611. (11) Ue, M.; Takeda, M.; Toriumi, A.; Kominato, A.; Hagiwara, R.; Ito, Y. J. Electrochem. Soc. 2003, 150, A499−A502. (12) Katakabe, T.; Kaneko, T.; Watanabe, M.; Fukushima, T.; Aida, T. J. Electrochem. Soc. 2005, 152, A1913−A1916. (13) Balducci, A.; Henderson, W. A.; Mastragostino, M.; Passerini, S.; Simon, P.; Soavi, F. Electrochim. Acta 2005, 50, 2233−2237. (14) Lewandowski, A.; Swiderska-Mocek, A. J. Power Sources 2009, 194, 601−609. (15) Ishikawa, M.; Sugimoto, T.; Kikuta, M.; Ishiko, E.; Kono, M. J. Power Sources 2006, 162, 658−662. (16) Kim, G. T.; Jeong, S. S.; Joost, M.; Rocca, E.; Winter, M.; Passerini, S.; Balducci, A. J. Power Sources 2011, 196, 2187−2194. (17) Moon, H. C.; Lodge, T. P.; Frisbie, C. D. J. Am. Chem. Soc. 2014, 136, 3705−3712. (18) Moon, H. C.; Lodge, T. P.; Frisbie, C. D. Chem. Mater. 2014, 26, 5358−5364. (19) Lee, J.; Panzer, M. J.; He, Y.; Lodge, T. P.; Frisbie, C. D. J. Am. Chem. Soc. 2007, 129, 4532−4533. (20) Cho, J. H.; Lee, J.; He, Y.; Kim, B. S.; Lodge, T. P.; Frisbie, C. D. Adv. Mater. 2008, 20, 686−690. (21) Cho, J. H.; Lee, J.; Xia, Y.; Kim, B.; He, Y.; Renn, M. J.; Lodge, T. P.; Frisbie, C. D. Nat. Mater. 2008, 7, 900−906. (22) Ha, M.; Xia, Y.; Green, A. A.; Zhang, W.; Renn, M. J.; Kim, C. H.; Hersam, M. C.; Frisbie, C. D. ACS Nano 2010, 4, 4388−4395. (23) Xia, Y.; Frisbie, C. D. In Organic Electronics II: More Materials and Applications; Klauk, H., Ed.; Wiley-VCH: Weinheim, Germany, 2012; pp 199−233. (24) Susan, M. A. B. H.; Kaneko, T.; Noda, A.; Watanabe, M. J. Am. Chem. Soc. 2005, 127, 4976−4983. (25) Seki, S.; Susan, M. A. B. H.; Kaneko, T.; Tokuda, H.; Noda, A.; Watanabe, M. J. Phys. Chem. B 2005, 109, 3886−3892. I

DOI: 10.1021/acs.macromol.5b00882 Macromolecules XXXX, XXX, XXX−XXX