Block copolymer electrolytes for rechargeable lithium batteries

Authors

  • Wen-Shiue Young,

    1. Department of Chemical and Biomolecular Engineering, University of Delaware, Newark, Delaware
    Current affiliation:
    1. The Dow Chemical Company, Spring House Technology Center, Spring House, Pennsylvania
    Search for more papers by this author
    • Wen-Shiue Young and Wei-Fan Kuan contributed equally to this work

  • Wei-Fan Kuan,

    1. Department of Chemical and Biomolecular Engineering, University of Delaware, Newark, Delaware
    Search for more papers by this author
  • Thomas H. Epps, III

    Corresponding author
    1. Department of Chemical and Biomolecular Engineering, University of Delaware, Newark, Delaware
    Search for more papers by this author

ABSTRACT

Ion-conducting block copolymers (BCPs) have attracted significant interest as conducting materials in solid-state lithium batteries. BCP self-assembly offers promise for designing ordered materials with nanoscale domains. Such nanostructures provide a facile method for introducing sufficient mechanical stability into polymer electrolyte membranes, while maintaining the ionic conductivity at levels similar to corresponding solvent-free homopolymer electrolytes. This ability to simultaneously control conductivity and mechanical integrity provides opportunities for the fabrication of sturdy, yet easily processable, solid-state lithium batteries. In this review, we first introduce several fundamental studies of ion conduction in homopolymers for the understanding of ion transport in the conducting domain of BCP systems. Then, we summarize recent experimental studies of BCP electrolytes with respect to the effects of salt-doping and morphology on ionic conductivity. Finally, we present some remaining challenges for BCP electrolytes and highlight several important areas for future research. © 2013 Wiley Periodicals, Inc. J. Polym. Sci., Part B: Polym. Phys. 2014, 52, 1–16

INTRODUCTION

The demand for safe and economical energy storage devices with high energy densities and fast charge/discharge rates is ever-growing,[1-7] and electrochemical devices such as rechargeable batteries and fuel cells are promising solutions in these clean and sustainable energy storage systems. One application example is rechargeable (secondary) batteries that have been used in various portable electronic devices and vehicles. Lithium batteries are ideal for these applications due to several desirable features including high energy densities, low self-discharge rates, high open circuit potentials, and minimal memory effects.[7, 8] Typically, a rechargeable lithium battery is composed of several electrochemical cells to provide the required voltage and capacity. Each cell consists of two electrodes, anode and cathode, and an electrolyte system. Depending on the anode materials and electrolyte systems, the lithium batteries can be divided into several different categories.[9] Lithium ion batteries, which typically use an anode material such as graphite, tin, or silicon to host the lithium ion, are the most common.[8] The operating principle of a typical lithium ion battery cell is illustrated in Figure 1. Lithium ions travel through a electrolyte system between two intercalation compounds, such as carbonaceous material (anode) and lithium cobalt oxide (cathode).[8] The electrolyte system usually requires a separator membrane and a gel-like (or liquid-like) ion-conducting medium that provide mechanical integrity and ion-conducting properties, respectively. Typically, lithium ion batteries can achieve conductivities on the order of 10−2 S/cm at room temperature, a 300–500 charge/discharge cycle life, and an electrochemical stability window within the 0–4 V range.[10-12] Note that a carefully selected electrolyte system is necessary to avoid thermal runaway reactions and dendrite formation during repeated charge–discharge cycles.[13-17]

Figure 1.

Operation schematic for an idealized rechargeable lithium ion battery: During the discharging process, lithium ions migrate through the electrolyte system, ion-conducting medium, and separator membrane, from anode to cathode (left image), while the transport direction is reversed during the charging process (right image).

Since the first rechargeable lithium battery (lithium ion battery) was commercialized by Sony Corporation in 1991, significant efforts have focused on improving the battery's life cycle and safety.[2, 8] One prevalent approach is to reduce the amount of organic solvent in the electrolyte system. Though organic solvents offer improved ion transport compared to a dry (solvent-free) electrolyte system, solvent usage renders the system thermally and electrochemically unstable.[18] To reduce solvent usage while maintaining the ionic conductivity, lithium polymer batteries using gel-type polymer electrolytes were commercialized.[8] This less-volatile polymer electrolyte system allows the batteries to be fabricated in various configurations and has been used in devices including cellular telephones and laptop computers.[8] However, inappropriate operation of lithium polymer batteries still may cause delamination of cell materials, leading to reduced battery life, cell expansion, or even fires. These safety concerns have promoted interest in solid-state lithium batteries with a solvent-free composition.[2, 9, 19]

Solid-state lithium batteries use either a dry polymer electrolyte system (e.g., poly(ethylene oxide) (PEO)) or a metal-oxide electrolyte system (e.g., lithium lanthanum titanium oxide or lithium phosphate).[19, 20] These solvent-free electrolytes are thermally and electrochemically stable compared to the traditional liquid or gel-like electrolyte systems and provide sufficient ionic conductivities (>10−4 S/cm at 50 °C);[21] thus, lithium metal possibly can be used as anode material to provide much higher energy density than current anode materials (Table 1). However, when lithium metal serves as anode, the possibility of dendrite formation is increased by non-uniform electrochemical deposition of lithium during subsequent charging.[22] To inhibit the growth of dendrites, electrolytes with high elastic moduli are suggested.[23] More specifically, by incorporating the elastic behavior of the lithium electrode and polymer electrolyte into a kinetic model, Monroe and Newman demonstrated that the electrolyte shear modulus should be above 7 GPa to inhibit dendrite formation.[23] This kinetic model described the deformation of electrolyte membrane upon dendrite formation as well as the stress distribution on the membrane during dendrite growth.

Table 1. Energy Densities of Anode Materials for Lithium Batteries
Anode MaterialsFully Charged StateEnergy Densitya (mAh/g)
  1. a

    Based on per gram of anode material at fully charged state.[1]

LithiumLi3,860
Carbon (graphite)LiC6372
TinLi22Sn5790
SiliconLi22Si52,010

Although metal-oxide electrolyte systems typically exhibit higher elastic moduli than dry polymer electrolyte systems, metal-oxide electrolyte systems usually are limited to thin-film lithium battery applications due to material processability considerations.[19] Considering the mass production of large battery systems, dry polymer electrolyte systems are more adaptable to continuous processing than metal-oxide electrolyte systems and can be implemented easily in commercial battery cell designs at low fabrication cost. However, the current dry polymer electrolytes tend to suffer from either insufficient ionic conductivities or subpar mechanical strength.

To improve the mechanical strength of the polymer electrolytes while maintaining high ionic conductivities, block copolymer (BCP) electrolytes, containing both well-defined conducting pathways and a sturdy supporting matrix, have been proposed.[24-27] These BCP electrolytes typically are based on the ion-complexation (salt-doping) behavior of ion conducting domain and the inherent nanoscale phase separation in BCPs.[18, 28] However, the complexation of salts with the solvating blocks of the BCPs can change the properties of the individual polymer domains and the overall copolymer morphology, thus impacting the ionic conductivity and mechanical strength of the nanostructured BCP electrolytes.[29-31]

During the last several years, a few reviews have been published that highlight efforts related to BCPs for solar cells, fuel cells, and hybrid materials applications.[32-37] In this review, we focus our discussion on experimental investigations of BCP electrolytes for solid-state lithium batteries. First, we give a brief summary of the underlying principles of homopolymer electrolyte systems, which provides a basis for understanding the ion transport properties in, and material selection for, the conducting domains in BCP electrolytes. Then, we describe the effects of salt-doping and morphology on the ion transport of copolymer electrolytes. Finally, we discuss new approaches that provide opportunities to overcome current challenges in nanostructured lithium battery electrolyte systems.

HOMOPOLYMER ELECTROLYTES

Ion-Solvating Polymers

To design BCP systems for solid-state lithium battery electrolytes, it is important to understand the ion solvating and ion transport properties of homopolymer electrolyte systems. Polymers with sequential polar groups, such as [BOND]O[BOND], [DOUBLE BOND]O, [BOND]S[BOND], [BOND]N[BOND], [BOND]P[BOND], C[DOUBLE BOND]O, and C[DOUBLE BOND]N, may dissolve lithium salts and form polymer:salt complexes.[18, 20] Further, to facilitate the dissociation of inorganic salts in polymer hosts, the lattice energy of the salt should be relatively low and the dielectric constant of the host polymer should be relatively high.[18, 20] PEO and poly(propylene oxide) (PPO) are examples of two homopolymers commonly combined with lithium salts for use in dry polymer electrolyte systems.[18] The ionic conductivity of salt-complexed PEO in the amorphous state is considerably higher than that of the comparable salt-complexed PPO due to the lower dielectric constant and the methyl groups hindering effect of PPO.[18, 20]

Typically, ionic conductivity is proportional to the effective number of mobile ions, the elementary electric charge, and the ion mobility.[20] The effective number of mobile ions (free ions) depends on the degree of salt dissociation in the polymer host. However, the degree of dissociation usually decreases with increasing salt concentration. In many PEO-based systems, the optimal salt concentration that maximizes the concentration of free ions (i.e., maximum conductivity) is located between oxygen to lithium ratios ([EO]:[Li]) of 25:1 and 8:1.21,38 A high lithium ion transference number, the ratio of the charge transported by lithium ions to the total transported charge, also facilitates high energy density in a battery and reduces the formation of concentration gradients in the electrolyte.[39-42]

Crystallization of PEO:Salt Complexes

PEO has been studied extensively as a promising candidate for polymer electrolytes due to its low glass transition temperature (Tg), which promotes charge transport, and good ability to dissolve lithium salts.[2, 18, 20, 21, 43] The ion transport in PEO:salt complexes is assisted by segmental motion of PEO chains in the amorphous state. However, PEO:salt complexes exhibit crystalline phases at certain salt-doping ratios, and it has been noted that the formation of crystallites at ∼60 °C decreases the ionic conductivity[18, 38, 44] except for some crystalline phases such as P(EO)6:LiAsF6, P(EO)6:LiSbF6, and P(EO)6:LiPF6 pointed out by Bruce and coworkers.[45-47] Nevertheless, as poor conductivities of PEO-based electrolytes due to crystalline phases at room temperature have limited their applications, several approaches have been reported to reduce PEO crystallinity.[48-59]

One approach incorporated noncovalent interactions to prevent PEO crystallization.[55-57] Zhang et al. utilized electrostatic interactions between positively and negatively charged PEO-containing ionomers (i.e., PEO+ and PEO) to form an amorphous complex.[56] Their lithium salt-doped complexes showed the absence of a sharp drop in conductivity over the experimental temperature range [Fig. 2(a)] (a sharp drop of conductivity in the temperature ramp study typically indicates the formation of crystallites) and had improved mechanical strength compared to PEO homopolymers.[56] Another approach employed PEO, poly(bisphenol A-co-epichlorohydrin) (PBE), and poly(vinyl ethyl ether) (PVEE) ternary blends to reduce the crystallinity of PEO.[58] Moreover, the addition of lithium ions in this PEO/PBE/PVEE system enhanced the miscibility of the polymers and induced a larger decrease in the crystallinity. The ionic conductivity for 60/25/15 (PEO/PBE/PVEE) ternary blends was as high as 10−3 S/cm (20 wt % LiClO4) at room temperature.[58]

Figure 2.

(a) Ionic conductivity profiles for PEO+/PEO complexes of varied [EO]:[Li] ratios. (b) Ionic conductivity profiles for PEO-mimetic polypeptoids (Nme: n = 1; Nde: n = 2; Nte: n = 3; n: numbers of PEO repeat units) at [Li]/[EO] = 0.085. The solid lines are VFT fits. [(a) Adapted from Ref. [56], with permission from American Chemical Society; (b) Adapted from Ref. [62], with permission from American Chemical Society.]

Comb-branched polymers with short PEO side chains that eliminate PEO crystallization, such as poly(oligo(oxyethylene) methacrylate) (POEM), also have been studied.[49, 50, 52, 60] Completely amorphous systems could be produced if the length of PEO side chains was restricted.[49, 61] For example, Cowie and Sadaghianizadeh examined lightly crosslinked comb-branched polymers with various PEO side chain lengths.[61] Their results showed that good conductivity (10−4 S/cm at 303 K) was obtained with five PEO repeat units, and no evidence of crystallinity was detected. In other work, Sun et al. employed a series of comb-like peptoid homopolymers with PEO side chains of varying length to explore the effect of side chain length on thermal and electrical behaviors.[62] Their PEO-mimetic polypeptoids did not crystallize, and the Tg of the polypeptoids decreased with increasing PEO side chain length. The length of PEO side chain also dominated the ionic conductivity in this system [Fig. 2(b)]. The solid lines through the data points in Figure 2(b) are fits to a Vogel–Fulcher–Tammann (VFT) model (eq (1)), which typically is used to fit the ionic conductivity, σ, of dry homopolymer electrolytes.[62]

display math(1)

In this expression, A, B, R, T, and T0 are pre-exponential factor, pseudoactivation energy, gas constant, actual temperature, and reference temperature, respectively. T0 usually is related to the glass transition temperature of the electrolyte. As the PEO-mimetic polypeptoids did not crystallize, the VFT equation was employed to fit the entire data set over the temperature window in Figure 2(b).[62]

Molecular Weight of Ion-Solvating Polymers

As ion mobility relies on chain motion,[20, 46] decreasing the Tg of polymer:salt complexes increases the chain motion and the ionic conductivity.[21, 63] Shi et al. investigated cation mobility in PEO hosts with different molecular weights and found that the cation mobility decreased as the molecular weight of PEO increased until a critical molecular weight of 3200 g/mol was reached.[64] Similarly, Bruce et al. incorporated 1000 g/mol PEO into their polymer electrolyte systems to obtain a fast local segmental motion.[46] However, benefits of high ionic conductivity in low-molecular-weight PEO electrolyte systems were countered by a decrease in mechanical strength. Therefore, to overcome this limitation and obtain high ionic conductivity while maintaining sufficient mechanical strength, BCPs containing both a conducting block and a rigid block are employed.

BLOCK COPOLYMER ELECTROLYTES FOR SOLID-STATE LITHIUM BATTERIES

Block Copolymers

BCPs consist of chemically dissimilar polymer segments or blocks that are covalently bound. BCPs provide the opportunity to design materials with attractive transport and mechanical properties based on their ability to self-assemble into periodic structures with domain spacings on the order of 10 nm. The self-assembly of BCPs is governed by the Flory–Huggins interaction parameter (χ), the number of repeat units per polymer chain (N), the block volume fractions (fA = 1 − fB), and the chain architecture (e.g., linear, star, graft, etc.).[65] Theoretical phase diagrams for a linear AB diblock copolymer and a linear ABC triblock copolymer calculated using self-consistent field theory (SCFT) are shown in Figure 3.[66, 67] There are five common morphologies for AB diblock copolymers including body-centered cubic spheres or close-packed spheres (S or Scp), hexagonally-packed cylinders (hex), double gyroid (G), orthorhombic network (O[70]), and lamellae (lam). In addition to morphologies in the diblock copolymer system, an alternating gyroid (GA) is shown in the triblock copolymer system.[67-70] Several reports have shown that network morphologies, such as the gyroid, exhibit superior mechanical properties relative to their one- (1D) and two-dimensional (2D) counterparts (e.g., hex and lam) due to the continuity of the mechanical supporting domain.[70-72] The nanoscale phase separation of BCPs allows one to obtain desirable mechanical, chemical, and electrical properties for ion-conducting membrane applications, as described in more detail in the following sections.

Figure 3.

Theoretical phase diagrams of (a) a linear AB diblock copolymer and (b) a linear ABC triblock copolymer calculated using SCFT. L, lam; C, hex; G, gyroid; S, sphere; Scp, close-packed sphere; O70, orthorhombic network; GA, alternating gyroid; D, disordered. [(a) Reproduced from Ref. [66], with permission from American Chemical Society; (b) Reproduced from Ref. [67], with permission from American Chemical Society.]

Block Copolymer Electrolytes Overview

Early work on BCP electrolytes for lithium batteries was reported by Giles et al. in 1987.24 Poly(ethylene glycol methyl ether) (MPEG) was grafted on the polybutadiene block of a poly(styrene-b-butadiene-b-styrene) (PS-PB-PS) triblock copolymer to form a (PS-P(B-g-MPEG)-PS) graft copolymer. These LiCF3SO3-doped PS-P(B-g-MPEG)-PS materials with various MPEG lengths and salt concentrations were amorphous over the temperature range studied, except at the lowest doping ratio ([EO]:[Li] = 50:1).[24, 25] The highest ionic conductivity obtained at 20 °C was about 10−5 S/cm. Later, in 1989, Khan et al. reported an ABA triblock copolymer electrolyte system with a PS middle block and POEM end blocks (POEM-PS-POEM).[73] This triblock copolymer with the ion-conducting POEM block as the matrix for ion transport exhibited an ionic conductivity of 1.2 × 10−6 S/cm at 25 °C for an [EO]:[LiClO4] ratio of 17:1. However, morphological and mechanical analyses were not reported in either Giles' or Khan's work.

During the past decade, the mechanical and morphological properties of salt-doped BCPs have attracted increased interest, and several studies provided greater understanding of the effects of salt-doping on ionic conductivity, mechanical strength, and morphologies.[26, 27, 29, 30, 74-78] These studies demonstrated the advantages of using an ordered BCP system over corresponding random copolymers, disordered BCPs, and homopolymers.[26, 79, 80] For example, a lithium bis(trifluoromethanesulfonyl)imide (LiTFSI)-doped poly(styrene-b-(styrene-g-ethylene oxide)-b-styrene) (PS-P(S-g-EO)-PS) triblock copolymer electrolyte (in which the copolymer was symmetric, and the PEO weight fraction was 60%) (Fig. 4) exhibited an elastic modulus of ∼109 dyn/cm2 between 0 and 100 °C, which was much higher than the corresponding elastic modulus of PEO homopolymer electrolytes.[80] This elastic modulus was close to the shear modulus required to inhibit macroscopic dendrite formation (estimated by Monroe and Newman ∼7 GPa23), and thereby likely preventing short-circuiting of the battery. Additionally, the ionic conductivity of this graft copolymer electrolyte was ∼2 × 10−5 S/cm at room temperature (and ∼10−4 S/cm at 60 °C).[80] In recent years, PS-PEOs have become the most attractive candidates for BCP electrolytes due to the combination of their appealing performance in the batteries. PS-PEO systems can achieve the ionic conductivity on the order of 10−3 S/cm ([Li]/[EO] = 0.085) at 90 °C,[81] and reach a shear modulus on the order of 108 Pa at 90 °C.[27] Also, PS-PEO electrolyte cells showed a good cycle life with retention of 80% of their initial capacity after 300 cycles at 90 °C and an electrochemical stability window up to 3.7 V.[17] However, low room temperature conductivity resulting from the crystallization of PEO block has limited the application of PS-PEO electrolytes. Table 2 lists several examples of BCP electrolytes, detailing the reported polymer ionic conductivities and shear moduli. Although the shear moduli of most current BCP electrolyte systems are less than 7 GPa, conductivity and mechanical strength are decoupled in BCPs, thus enabling the potential future increase in shear modulus without compromising conductivity. In this work, we aim to provide an overview of ion conduction in BCP electrolytes for rechargeable lithium batteries, and we refer the reader to additional sources for studies on battery performance.[17, 37, 82]

Figure 4.

Dynamic modulus data of (a) PEO homopolymer, (b) PS homopolymer, and (c) LiTFSI-doped PS-P(S-g-EO)-PS electrolyte. (Reproduced from Ref. [80], with permission from the Electrochemical Society.)

Figure 5.

(a) Order-disorder transition temperature, TODT, and (b) the domain spacing at the corresponding TODT for LiCF3SO3-doped PS-PEO as a function of salt-doping ratio. (Reproduced from Ref. [96], with permission from American Chemical Society.)

Table 2. Examples of Block Copolymer Electrolytes
Polymers [Conducting Materials]Chemical StructureConductivity (S/cm) Temperature (°C)Modulus (Pa) Temperature (°C)Ref.
Poly(styrene-b-(styrene-g-ethylene oxide)-b-styrene) [LiTFSI]
image
2 × 10−5251083080
Poly(ethylene-b-ethylene oxide) [LiCF3SO3]
image
2 × 10−56010612030
Poly(styrene-b-ethylene oxide) [LiTFSI]
image
10−3901089027,81
Poly(styrene-b-oligooxyethylene methacrylate-b-styrene) [LiClO4]
image
2 × 10−43075
Poly(styrene-b-ethylene oxide-b-styrene) [BMIm-PF6]
image
1.1 × 10−326.51031083
Poly(styrene-b-methyl methacrylate) [EMI-TFSI]
image
2.5 × 10−46084
Poly(methyl methacrylate-b−1-[(2-methacryloyloxy)ethyl]-3-butylimidazolium hydroxide) [OH]
image
3.2 × 10−26085
Poly(styrene trifluoromethanesulfonylimide of lithium-b-ethylene oxide-b-styrene trifluoromethanesulfonylimide of lithium) [Li+]
image
1.3 × 10−56086
Poly(styrene-b−4-vinylbenzyl alkylimidazolium bis(trifluoromethanesulfonlyl)imide) [TFSI]
image
10−66087
Tetraethylene glycol dodecyl ether [LiClO4]
image
3 × 10−43088

Charge Carrier and Salt-Doping Effects

The most common method to introduce mobile lithium ions into a polymer electrolyte system is to blend lithium salts into solvating polymers. The salts/polymer combinations usually are selected using several aspects including polymer:salt complexation, the degree of dissociation in the polymer:salt complex, and the electrochemical and thermal stabilities of the salt/polymer mixture.[18, 89] An efficient complexation between the lithium salt and the polymer gives opportunities for loading more salts (mobile ions) into the system, and a high degree of dissociation gives the system more free ions. In order to determine the effect of the degree of dissociation, Bannister et al. compared two lithium salts of anionic polymers and showed that polymer electrolytes with lithium salts derived from stronger acids exhibited higher conductivities.[90] Conversely, a strong ion-pairing interaction between cations and weak acids can restrict the number of mobile cations, which leads to a low conductivity. Similar results also have been noted in liquid electrolyte systems.[91]

Most salt-doping studies involving PEO (and PEO-like) copolymers were investigated in the low salt concentration regime, analogous to [EO]:[Li] ratios ranging from 50:1 to 6:1. Several studies have shown that salt-doping can change the morphologies and thermal properties of BCPs.[18, 25-27, 31, 92-95] For example, significant increases in the order-disorder transition temperature (TODT) were found in LiCF3SO3-doped poly(methyl methacrylate-b-oligooxyethylene methacrylate) (PMMA–POEM),[26] LiClO4-doped poly(styrene-b-isoprene-b-ethylene oxide) (PS-PI-PEO),[29] LiClO4-doped poly(isoprene-b-styrene-b-ethylene oxide) (PI-PS-PEO),[74] and LiCF3SO3-doped poly(styrene-b-ethylene oxide) (PS-PEO) systems relative to their neat polymer analogs (Fig. 5).[96] In these systems, increases in TODT and domain spacing, along with morphological changes, were attributed to an increase in the effective Flory-Huggins interaction parameter (χeff) as a function of salt concentration. The χeff was used to estimate the change in phase behavior upon salt-doping.

Several methods have been adopted to estimate χeff in salt-doped BCPs. Young et al. calculated χeff values in salt-doped PS-PEO systems based on the relationship between domain spacing (d) and χ in the strong segregation regime, dχ1/6 (χeff are calculated for each salt-doped sample relative to the interaction parameter of the neat sample, χPS-PEO).[92, 97] Values for χeff were found to exhibit a linear relationship with salt concentration, and the slope (m) of the linear regression increased as the Lewis acidity of the anion increased [Fig. 6(a)].[92] Wang et al. obtained values for χeff in a LiCl-doped poly(styrene-b-methyl methacrylate) (PS-PMMA) system using small-angle neutron scattering and Leibler's mean field theory.[98, 99] Their results showed that χeff and the segmental lengths of both PS and PMMA increased upon formation of PMMA-LiCl complexes, and χeff became less temperature-dependent compared to neat samples due to the increase in χS and decrease in χH upon salt-doping (χ = χS + χH /T, in which χS and χH are entropic and enthalpic terms of χ, respectively, and T is temperature).[98] Wanakule et al. estimated χeff in a LiTFSI-doped PS-PEO system by comparing the experimental copolymer compositions (i.e., χeffN and fPEO:salt) to the theoretical diblock copolymer phase diagram (e.g., (χeffN)ODT = 10.5 at fPEO:salt = 0.5) at the order-disorder transitions.[100] Interestingly, the slope (m) of χeff versus salt concentration data from a LiTFSI-doped PS-PEO system was much smaller than the slope from other salt-doped PS-PEO systems reported by Young et al.[92] Wanakule et al. later found that the size of the anion might affect the m value, in which their hypothesis was based on a theory of ion solvation for binary polymer blends developed by Wang.[101, 102] Although the experimental m values were consistently lower than those calculated from the theory, the trends in the experimental results were captured by the theory [Fig. 6(b)].[101] The lack of agreement between theory and experimental results was possibly due to discrepancies between the dielectric constants for PEO and PS used in the theoretical predictions versus in the experimental setup (temperature and salt concentration), as well as the validity of the assumptions inherent in the strong segregation theory analysis employed in the experimental work.[92, 101]

Figure 6.

(a) χeff versus salt concentration ([Li]/[EO]) at 120 °C in various salt-doped PS-PEO systems. The slope of χeff equation decreases from PS-PEO:LiAsF6 (□) > PS-PEO:LiClO4 (Δ) > PS-PEO:LiCF3SO3 (○). (b) The slope of χeff versus salt concentration: experimental versus theoretical. [(a) Reproduced from Ref. [92], with permission from American Chemical Society; (b) Reproduced from Ref. [101], with permission from American Chemical Society.]

As discussed previously, the ionic conductivity depends strongly on the nature of salt, the solvating polymer, and the salt concentration in the polymer. However, the geometry of the conducting domains and the local environment surrounding the ions also are important in BCP electrolyte systems. Gomez et al. studied the distribution of lithium ions in lamellae-forming PS-PEO using energy-filtered transmission electron microscopy [Fig. 7(a)].[103] Lithium ions were localized in the middle of the PEO lamellae [Fig. 7(a)], and the ratio of the layer thickness of lithium ions to the layer thickness of PEO (dLi/dPEO) decreased as the segregation strength increased [Fig. 7(b)]. Using self-consistent field theory, local stress field calculations indicated that this stress interfered with the ability of PEO near block interfaces to interact with salts and thus decreased the salt concentration near the block interface (where the conductivity is expected to be low).[103] Interestingly, the ionic conductivities of the BCP electrolytes (σ) improved with decreasing dLi/dPEO [Fig. 7(c)]. As χN (molecular weight) increased, the normalized ionic conductivity increased. This phenomenon demonstrated the fundamental difference in the mechanism of ion transport in homopolymers and BCPs. In homopolymer electrolyte systems, the molecular weight has no significant effects on ionic conductivity above a critical molecular weight (∼3200 g/mol).[64] In BCP electrolyte systems, Balsara and coworkers found that the conductivity of the LiTFSI-doped PS-PEO increased with the increasing molecular weight of PEO block, MPEO, in the high molecular weight regime (MPEO > 10,000 g/mol) due to chain stretching effects.[27, 81] Also, the conductivity of the LiTFSI-doped PS-PEO in the low molecular weight regime (MPEO < 10,000 g/mol) slightly decreased as a function of MPEO which resulted from the dependence of conductivity on the Tg of the PS block because all ions were in contact with both PS and PEO domains.[104]

Figure 7.

(a) Lithium ion distribution in a LiTFSI-doped PS-PEO system imaged by energy-filtered transmission electron microscopy. Red represents the oxygen locations in PEO, blue represents the lithium ions in the salt, and gray represents PS. (b) Lithium to PEO thickness ratios with respect to the copolymer segregation strength (χN) calculated via TEM and SCFT. (c) Lithium to PEO thickness ratios and normalized conductivity with respect to χN. Reproduced from Ref. [103], with permission from American Chemical Society.

It is clear that the magnitude and thermal dependence of χeff in salt-doped BCP systems can be affected by the solvating polymer, anion size, and salt acidity. However, the morphology of the BCP also may be affected by the increase in χeff upon salt-doping. As mentioned in the previous discussion of work by Gomez et al., a non-uniform salt distribution in the PEO domain was found at higher segregation strengths, implying a strong repulsion exists between the PS and the PEO:salt complex.[103] To minimize the system overall free energy, the PEO:salt complex is forced to segregate from the PS domain, resulting in chain stretching. The balance between salt distribution and chain stretching may cause substantial changes in morphology. For example, network-forming PI-PS-PEO and PS-PI-PEO systems, in which PEO is the minority phase, transformed to core–shell cylinders upon salt-doping, even though accounting for the increases of χeff,SO, χeff,IO, and volume fraction of PEO domain would lead one to expect the formation of two- or three-domain lamellae.[74]

Morphology Effects

While the morphology effects on mechanical properties of BCPs have been well-studied and understood,[71, 72] the morphology effects on the ionic conductivities of BCP electrolytes are much more complicated, and further studies are needed. The typical thickness of electrolyte membranes is around 10–100 µm, and a BCP generally forms several grains across this thickness. As a result, when lithium ions travel between electrodes, ion conduction can be divided into two categories, intra-grain and inter-grain transport. Intra-grain transport describes conduction within grain boundaries, while intergrain transport indicates the connectivity of conducting channels across grain boundaries. For intra-grain ion transport, the dimensionality of the conducting pathway and the domain orientation are key factors. For example, in a lamellae-forming system, the lamellar conducting pathways are 2D, and only two-third of the domains (on average) positively contribute to the ion transport (as illustrated in Fig. 8).[27, 105]

Figure 8.

Effect of domain orientation on ion transport in a lamellae-forming system. The blue and red layers represent conducting and nonconducting domains, respectively.

Figure 9.

(a) Three-dimensional reconstructed TEM image of grain boundary morphology in a cylinder-forming PS-PI diblock copolymer system. (b) Highlighted cylinders from (a) show continuous (A) and discontinuous (B and C) cylinders at grain boundaries. (Reproduced from Ref. [106], with permission from John Wiley and Sons.)

Figure 10.

(a) Calculated interfaces of a lamellae-forming BCP across the grain boundary. (b) Simulated images of thin sections for the grain boundaries at various tilt angles, ϕ. (Reproduced from Ref. [108], with permission from American Chemical Society.)

For inter-grain ion transport, the overall ionic conductivity of BCP electrolytes is lowered by bends in the conducting channels (which may cause discontinuities in conducting domains) or reductions in the size of conducting channels across the grain boundaries. For example, Jinnai et al. found that the tortuosity and continuity of domains in a cylinder-forming sample were strongly affected by the grain orientation angle (Fig. 9).[106] When the inter-grain orientation angle was ∼90°, the majority of the cylindrical domains bent at the grain boundary to avoid intersecting each other (i.e., the domains were discontinuous). Furthermore, the reduction of cross-sectional area of conducting channels (i.e., decrease in conductivity) could be found at grain boundaries in lamellae-forming samples (Fig. 10).[107, 108] This phenomenon of domain discontinuity across grain interfaces illustrated the importance in obtaining long-range order of copolymer nanodomains as well as the advantages of three-dimensionally (3D) conducting pathways (e.g., the continuity of the matrix for hex is not (or less-) affected by the grain orientation angle).

Typically, the ionic conductivity, σ, in the BCP system can be expressed as eq (2).

display math(2)

in which σmax and ϕ represent the ionic conductivity of the corresponding homopolymer electrolyte and the volume fraction of the conducting phase, respectively; α represents the morphology effects in a randomly oriented case, and τ represents the effects of the grain boundaries. Considering a case of randomly oriented conducting cylinders for which the minority phase is the conducting domain. Only one-third of the domain orientations contribute to the ion transport direction between two electrodes; thus, α is one-third. In similar analysis, α is 1 for the case of a conducting matrix for the same nanostructure,[27, 105] so the morphology effects can be eliminated when the conducting domain is the majority and continuous phase allowing ion transport in all direction.[78] The cross-sectional area of conducting pathways at grain boundaries also should be considered. This boundary effect, τ, mainly is attributed to the discontinuity of the conducting pathways at grain boundaries; [107-110] however, the effect is mitigated to a large extent by 3D continuous pathways. These 3D continuous pathways have larger cross-sectional areas across grain boundaries relative to 1D and 2D pathways.

The morphology effects on the ionic conductivity and mechanical strength were demonstrated by Wiesner et al. using LiCF3SO3-doped poly(ethylene-b-ethylene oxide) (PE-PEO) dendrons, for which PEO was the minority domain.[30] The elastic modulus and ionic conductivity profiles as a function of temperature for two PE-PEO dendrimer materials are shown in Figure 11.30 Thermal transitions [order–order transitions and order–disorder transitions (ODT)s] during heating were determined via step changes in elastic modulus [Fig. 11(a,b)] and further confirmed by SAXS and differential scanning calorimetry (DSC), respectively. The nanostructures and phase transitions found in these salt-doped dendrimers were ideal for studying the morphology effect of conducting paths because they represented 0D (mc), 1D (hex), 2D (lam), and 3D (continuous cubic [cc]) conducting pathways. The conductivities of salt-doped samples are shown in Figure 11(c). As one would expect, the micelle phase exhibited the poorest conductivity and had no clear change across the ODT. Contrary to the micelle-to-disordered phase transition, the conductivity increased by an order of magnitude during the hex-to-cc transition, which was much larger than the threefold increase expected in going from 1D to 3D. This result implied that the morphology effect on the ionic conductivity is not only due to the geometry of conducting pathways but also due to the continuity at the grain interfaces. A similar result was obtained by Young and Epps using LiClO4-doped PS-PEOs that formed hexagonally perforated lamellae (hpl), lam, and hex (PEO matrix) morphologies.[78] Their studies showed that samples with 3D conducting pathways (hpl and hex) demonstrated much higher ionic conductivities than those with 2D conducting pathways (lam) even after correcting for the domain orientation and molecular weight (Fig. 12). They also suggested that the continuity of conducting channels across the gain boundary between BCP domains was an important factor contributing to overall conductivity.

Figure 11.

Elastic modulus (G′) profiles of neat and LiCF3SO3-doped PE-PEO samples: 1 (a) and 2 (b). Sample 2 had a longer PEO block than Sample 1. (c) Ionic conductivity profiles of LiCF3SO3-doped PS-PEO (1 and 2). k, crystalline lamellae; mc, micelle (sphere); hex, hexagonal cylinder; lam, lamellae; cc, continuous cubic; dis, disordered; dec, decomposition. (Reproduced from Ref. [30], with permission from the American Association for the Advancement of Science.)

Figure 12.

Normalized conductivities of LiClO4-doped PS-PEO BCPs with lam, hpl, and hex morphologies. Reproduced from [78], with permission from American Chemical Society.

In an attempt to eliminate grain boundary effects, Li et al. probed a poly(ethylene oxide-b-(methacrylate-g-azobenzene)) (PEO-PMA(Az)) system with aligned LiCF3SO3-doped PEO cylinders.[111] The in-plane and through-plane ionic conductivities of electrolyte films demonstrated the advantages of aligned conducting domains. The AFM micrographs, top views and cross-sectional views, of the aligned samples at different doping ratios are shown in Figure 13(a–f). The PEO cylinders were perpendicularly aligned on the silicon wafers, and the samples with high salt-doping ratio (4:1) exhibited poor long-range order compared to 120:1 and 20:1 [EO]:[Li] salt-doping ratios. The conductivity profiles [Fig. 13(g)] measured in the directions of perpendicular and parallel to the substrate showed the in-plane conductivity (σ) was dramatically lower than the through-plane conductivity (σ).[111] Also, the through-plane conductivity dropped to the same order of magnitude as the in-plane conductivity when the sample was brought above its isotropic transition temperature (unaligned).[111] This work provides an alternative approach to prevent the loss of the conductivity caused by the discontinuity of conductive pathways at the BCP grain boundaries.

Figure 13.

(a–f) AFM images of perpendicularly aligned LiCF3SO3-doped PEO-PMA(Az) samples, in which the PEO:LiCF3SO3 domains are cylinders. (g) Ionic conductivity profiles of PEO-PMA(Az):LiCF3SO3 complexes as a function of temperature. Through-plane (open symbols); in-plane (filled symbols). (Reproduced from Ref. [111], with permission from American Chemical Society.)

Figure 14.

Room temperature through-plane (red filled circle) and in-plane (violet filled circle) ionic conductivity of PS-PMMA/IL compared to predicted conductivity (solid line) for perfect lamellae aligned in the direction of transport. The schematics illustrate the idealized BCP morphologies and the directions of measured ion transport. L, lam morphology; C, hex morphology. (Reproduced from Ref. [84], with permission from Elsevier.)

Ionic Liquids in Block Copolymers

Recently, BCPs containing ionic liquids (ILs) have drawn attention as alternative electrolytes for lithium batteries.[82, 84, 112-125] ILs are room temperature molten salts that are composed mostly of organic ions, offering several advantages including reduced flammability, low vapor pressure, good thermal stability, low toxicity, and good ionic conductivity.[7, 9, 122, 126, 127] In 2007, He et al. reported a method for preparing ion gels through the self-assembly of PS-PEO-PS triblock copolymers at high IL, 1-butyl-3-methylimidazolium hexafluorophosphate (BMIm-PF6), compositions for polymer electrolytes.[27] The conductivity from the work by He et al. reached about 1.1–1.6 × 10−3 S/cm at room temperature; however, the triblock copolymers exhibited modest impact on the conductivity, and the conductivity decreased with the increasing viscosity similar to neat ILs. More recently, Simone and Lodge studied the concentrated solutions of PS-PEO using IL, 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide (EMI-TFSI), as the solvent, and they found that conductivity increased with increasing ion concentrations and MPEO.[118] Moreover, their work demonstrated that PS-PEO/IL mixtures with continuous conducting domains (PEO matrix in hex morphologies) have significantly higher conductivity than those with discontinuous conducting domains (e.g., lam morphologies). A similar result was reported in a PS-PMMA/IL (EMI-TFSI) system by Gwee et al.[84] They investigated PS-PMMA/IL with various IL compositions (0–50 wt %) and found that the morphology and domain orientation strongly impacted ionic conductivity. High through-plane conductivities were obtained when the morphologies underwent phase transitions from a discontinuous conductive domain (anisotropic lam morphology) to a continuous conductive domain (anisotropic hex with PEO matrix).[84] Also, when the lamellae-forming PS-PMMA/IL was orientated, the in-plane conductivity (parallel to the lamellar sheets) was considerably higher than the through-plane (perpendicular to the lamellar sheets) conductivity (Fig. 14).[84] Recently, Elabd and Winey studied polymerized ionic liquid (PIL) diblock copolymers to explore relationship between the morphology and the ionic conductivity.[85, 120, 121] Their work showed that the ionic conductivity of PIL BCPs were higher than their random copolymer and homopolymer analogs due to the effects of local confinement and connectivity of conducting ions in nanoscale BCP domains.[85, 121] They also indicated that 3D network morphologies in PIL BCPs resulted in higher ionic conductivities in comparison to lam and hex morphologies.[120] Although the conductivity of the nanostructured polymer electrolytes can be improved using ILs with continuous and ordered conducting pathways, studies related to aligning BCP/ILs morphological domains are limited. Note: Most of the conducting species discussed in the IL literature are not lithium ions, which are needed for lithium batteries; however, the interesting transport properties in IL BCP systems present an intellectual synergy that can lead to a rational design of novel BCP electrolytes.

CHALLENGES AND OUTLOOK

Significant work has been performed to improve the fundamental understanding of salt-doped BCP electrolytes. As described herein, BCP electrolytes can be considered attractive candidates for lithium batteries due to their desirable mechanical strength and their high ionic conductivity (similar to corresponding homopolymer electrolytes). However, several hurdles must be overcome to facilitate their usage in various battery systems.

Low room temperature conductivity has been a concern for applicability of dry polymer electrolytes in lithium batteries (see Table 2). Although several studies on homopolymer electrolyte systems have been reported to improve the room temperature conductivity, most methods that were employed are not ideal for increasing ionic conductivity in the bock copolymer systems. The methods either required low molecular weight PEO,[104] which has been proven to provide low ionic conductivity in BCP systems, or involved complicated synthesis procedures that are not easily adapted to BCPs. Comb-like BCPs provide opportunities for high room temperature conductivity but the nonsolvating backbone decreases the volume fraction of the conducting domain.

Young et al. proposed a mixed-salt method to increase the room temperature ionic conductivity which suppressed the crystalline phase in the PS-PEO:LiTFSI:LiClO4 system ([EO]:[Li] = 6:1).[128] This reduction in crystallinity provided a simple and effective approach to obtain higher ionic conductivity at low temperature compared to single-salt PS-PEO systems.[128] Additionally, single-ion conducting electrolyte also enables the opportunity for obtaining high ionic conductivities for lithium batteries. This approach decreases the mobility of the anion groups (increases the transference number); thus, self-discharge and possible degradation at the electrode surface can be reduced. Recent work by Bouchet et al. demonstrated single-ion polymer electrolytes based on P(STFSILi)-PEO-P(STFSILi) [poly(styrene trifluoromethanesulfonylimide of lithium-b-ethylene oxide-b-styrene trifluoromethanesulfonylimide of lithium)].[86] The P(STFSILi)-PEO-P(STFSILi) exhibited reasonable ionic conductivity (1.3 × 10−5 S/cm at 60 °C) in single-ion polymer electrolyte systems, a high transference number (>0.85), and improved mechanical strength and electrochemical stability compared with corresponding PS-PEO-PS:LiTFSI materials.[86] The single-ion BCP electrolyte system shows ideal properties for lithium battery applications, and further studies in this area are warranted.

For BCP electrolytes to be implemented in commercial lithium battery applications, fast processing and well-ordered nanostructures also are ideal. However, high-molecular-weight materials, which provide high ionic conductivity and mechanical strength in BCP systems, decrease the overall materials' processability. From the standpoint of reducing TODT (increasing processability), recent efforts on tapered BCPs deserve attention.[129-131] Manipulating the interfacial composition in BCP systems provides a way to decouple the segregation strength (χeffN) between blocks from the copolymer molecular weight; thus, the TODT can be controlled independently from the molecular weight.[132, 133] Moreover, the effects of domain orientations on transport in a polymer electrolyte are notable.[134] Work by Weber et al. showed that samples, poly(styrene-b−4-vinylbenzyl alkylimidazolium bis(trifluoromethane sulfonlyl)imide) [PS-b-PVBn-(alkyl)ImTFSI]), prepared by solvent-casting and melt-pressing strongly affect the copolymer conductivities due to the long-range ordering of polymer morphologies (well-ordered solvent-cast film showed higher conductivity than poorly ordered melt-pressed film).[87] Additionally, Osuji and coworkers used magnetic field alignment to control the nanoscale structure in BCPs,[88, 135-138] and their highly aligned samples demonstrated an order of magnitude improvement of ionic conductivity compared to unaligned counterparts by reducing microstructural tortuosity.[88, 135, 136] Their work showed that the ionic conductivity of BCPs strongly is influenced by both the alignment of block domains and the domain orientation. Other techniques, such as electric field and shear alignment, also are worthy of further examination to achieve the long-range ordering and desirable grain orientations necessary for optimal ionic conductivity in BCP systems.[139-141]

Thus far, most reports for BCP electrolytes have used AB diblock and ABA triblock copolymer systems, in which each block was selected to provide mechanical strength or ion-solvating properties. However, it also may be of interest to use ABCBA pentablock, ABC triblock, brush copolymer, or other systems. Incorporating additional blocks between the mechanical supporting block and the ion-conducting block offers an opportunity to increase the chain mobility of solvating polymers near block interfaces, and may provide a better understanding of ion transport mechanisms in nanostructured polymer systems. As highlighted by the works in this review, BCP electrolytes are exceptionally promising materials for lithium battery applications, and we believe the rational design of novel BCP electrolytes will accelerate the development and the commercialization of sturdy and easily processed lithium batteries with high energy density and fast charge/discharge rates.

Acknowledgments

W.-S. Young and W.-F. Kuan were supported by AFOSR-PECASE (FA9550-09-1-0706). W.-F. Kuan was partially supported by NSF-DMR CAREER (DMR-0645586).

Biographies

  • Image of creator

    Wen-Shiue Young is currently a senior engineer at The Dow Chemical Company. He received a B.S. and M.S. in Chemical Engineering from National Taiwan University and a Ph.D. in Chemical and Biomolecular Engineering from University of Delaware (UD) under the direction of Thomas H. Epps, III. His research focused on the application of polymer separation technologies and the development of nanostructured materials for lithium batteries.

  • Image of creator

    Wei-Fan Kuan is currently pursuing his Ph.D. in Chemical and Biomolecular Engineering at UD under the direction of Thomas H. Epps, III. He received his B.S. in Chemical Engineering from National Taiwan University in 2008. His research focuses on designing, synthesizing, and characterizing novel polymeric materials for transport and ion conduction applications.

  • Image of creator

    Thomas H. Epps, III is the Thomas and Kipp Gutshall Professor of Chemical and Biomolecular Engineering at UD. He received a B.S. in Chemical Engineering (MIT, 1998) and a Ph.D. in Chemical Engineering (University of Minnesota, 2004) under the direction of Frank S. Bates. He has received several awards, including a DuPont Young Professor Award, PECASE Award, Air Force Young Investigator Award, and NSF CAREER Award, among others. His research focuses on the nanoscale assembly of soft materials for transport, templating, and delivery applications.