UC Berkeley

It is becoming increasingly clear that the next generation of rechargeable batteries for the emerging clean energy landscape will require electrolytes that are fundamentally different from those used in today’s lithium-ion batteries. In addition, it has been shown that the most promising approach to significantly increase the energy density of rechargeable batteries is through the implementation of the Li metal anode. Current electrolytes consist of mixtures of organic carbonates and a lithium salt. They pose safety concerns due to their flammability and are also incompatible with Li metal. Polymer electrolytes, which are mixtures of polymers and a lithium salt, are both less flammable than the organic solvents used currently and have been shown to be compatible with Li metal. Although linear polymers are able to exert some stress on the battery electrodes, which is essential to enabling rechargeable batteries with Li metal anodes, their viscoelastic nature prevents them from enduring stress in the long-time limit. One approach to improving the mechanical properties of polymer electrolytes is to use microphase separated block copolymers, which allows for decoupling of the ionically conducting and mechanically reinforcing properties.

The phase behavior of block copolymers, wherein two chemically distinct homopolymer chains are covalently bound, is dependent on two properties: segregation strength, χN, and copolymer composition, f_A. Segregation strength is the product of the Flory-Huggins interaction parameter, χ, and the overall chain length, N; it combines both enthalpic contributions from the monomer-monomer interactions, as well as entropic contributions which are directly related to chain length through the configurational entropy. The copolymer composition is measured in terms of the volume fraction of “block A”, f_A, and dictates the degree of curvature needed to minimize the areal contact between the unlike phases. For a given copolymer composition, when a copolymer has a segregation strength less than (χN)_odt, the copolymer melt will form a homogeneous disordered mixture. When the same copolymer has χN≥(χN)_odt, the copolymer will microphase separate into an ordered morphology such as body center cubic spheres, hexagonally packed cylinders, gyroid phases, or lamellae. The addition of salt to block copolymers allows them to conduct ions, but significantly alters their thermodynamics and resulting phase behavior. It is well-known for symmetric block copolymers, wherein the volume fractions of the two blocks are equal to 0.5, that the addition of salt will increase the segregation strength of the copolymer. For example, a copolymer that forms a disordered morphology in the salt-free state will undergo a disorder-to-order transition and form an ordered lamellar phase once a critical salt concentration is reached. In the simplest case, the increase in χ is linear with respect to salt concentration.

It is well-known that ion transport in polymer electrolytes is linked to the segmental motion of the polymer backbone. At short times, the polymer repeat units can be modeled as beads linked together by springs that are characterized by a friction coefficient, ζ. Motion in this regime is known as Rouse dynamics. At longer times, the motion of a segment of a polymer chain is influenced by the presence of neighboring chains that form entanglement constraints represented by tubes of diameter, d. As time progresses, the polymer chain must reptate through the entanglement constraints until it is fully free of its tube and can undergo self-diffusion. The motion of the polymer chain while it is entirely or partially constrained by its tube is known as reptation. Previous studies of dynamics in the Rouse regime have shown that the monomeric friction coefficient decreases as salt concentration increases in polymer electrolytes due to the coordination between the polymer backbone and the cations. In fact, for some polymer electrolytes, the ionic conductivity can be explained entirely by the segmental motion of the polymer. However, no studies have been conducted on polymer electrolytes at time-scales that correlate with reptation.

The block copolymer electrolyte used in this research is a well-studied model system: polystyrene-block¬-poly(ethylene oxide) mixed with lithium bis(trifluoromethanesulfonyl) imide (LiTFSI). SEO/LiTFSI was chosen as the model system because it is well-known that the Li salt preferentially segregates into the poly(ethylene oxide) (PEO) domains and the electrochemical properties of PEO/LiTFSI have been fully characterized. The objective of this dissertation is to develop a molecular-level understanding of the effect of salt on block copolymer self-assembly, thermodynamics, and dynamics. Specifically, I will use theories developed for salt-free polymer systems and apply them to data collected on block copolymer electrolytes to determine how the presence of salt changes polymer behavior. In cases where the original theories cannot explain the new data, I will derive new theories to describe the observed phenomenon.

In order to conduct these studies, a library of SEO copolymers with precise molecular weights and a wide range of copolymer compositions was synthesized. The first section describes the methods used to synthesize the SEO copolymers with precise molecular weights and compositions and low polydispersities.

The research in this dissertation covers a large area of polymer science; therefore, several different experimental techniques were used. Coincidentally, all experimental techniques were centered around scattering applications of both X-rays and neutrons. The second section provides an introduction to the fundamentals of scattering as well as a brief review on the specific techniques used in these studies.

The phase behavior of SEO/LiTFSI was determined using small angle X-ray scattering (SAXS). We studied the effect of salt concentration, molecular weight, copolymer composition, and temperature on the morphology of SEO/LiTFSI in sections four and five. In SEO copolymers with a majority polystyrene phase, the addition of salt induced the formation of ordered morphologies. However, unique phase behavior was seen in one copolymer, with a PEO volume fraction, f_EO, of 0.20. In this sample, a reentrant phase transition was found such that the disordered copolymer first formed an ordered morphology, followed by a disordered morphology, followed by a different ordered morphology with increasing salt concentration. The first ordered morphology seen at a low salt concentration was a novel type of coexistence wherein two distinct lattices of the same lattice type was observed. The nature of this type of coexistence was further probed with electron tomography to visualize the lattice structure and resonant soft X-ray scattering (RSoXS) to quantify the volume of salt in each lattice.

The goal of section five was to use the morphology data gathered from the SAXS experiments and assemble large experimental datasets of the phase behavior of SEO/LiTFSI. For simplicity, the phase diagrams were constructed at a single temperature of 100 °C. A simple framework based upon of mean-field theory developed for salt-free block copolymers was used to create a phase diagram plotting copolymer morphology as a function of χN and the volume fraction of the salt-containing phase, f_(EO,salt). It was found that the phase behavior of SEO/LiTFSI is qualitatively similar to that of salt-free block copolymer systems. In addition, the effect of copolymer composition and salt concentration was examined on the domain spacing of SEO mixed with two different Li salts. Expressions for the domain spacing as a function of copolymer chain length, composition and salt concentration were developed for both the weak (χN≤10) and strong segregation regimes (χN>10).

The following section focuses on the quantification of the thermodynamics of SEO/LiTFSI. Through application of Leibler’s Random Phase Approximation, the Flory-Huggins interaction parameter of the neat, χ_(0,SC), and salt-containing, χ_(eff,SC), SEO copolymers were measured. An expression for χ_(eff,SC) was developed as a function of N,f_EO and salt concentration. It was then used in conjunction with mean-field theory to calculate the critical chain length for ordering, N_crit, as a function of copolymer composition and salt concentration. Two regimes of phase behavior emerged: at f_EO>0.27, the addition of salt stabilizes the ordered phase and at f_EO<0.27, the addition of salt stabilizes the disordered phase. The copolymer composition wherein segregation is independent of salt concentration (f_EO=0.27) was termed the “isotaksis point” and simple theoretical model to predict this composition was developed.

The preceding sections focused on the effect of salt on the morphology of the block copolymer electrolytes, which corresponds to length-scales on the order of magnitude of the domain spacing of the microphases, ~10 nm. The effect of salt concentration on the polymer conformation of the corresponding homopolymer electrolyte, PEO/LiTFSI, was determined using small angle neutron scattering (SANS), which corresponds to length-scales less than 1 nm. Kratky analysis of the SANS profiles was used to determine the statistical segment length as a function of salt concentration. All parameters in the analysis were determined independently. It was found that the effect of salt concentration in non-monotonic: at low salt concentrations, the chain sizes decreases before reaching a minimum at intermediate salt concentrations and then increasing at higher salt concentrations. This relationship is hypothesized to be related to the preferred environment of Li ions within PEO where each Li ion prefers to be coordinated by six ether oxygens. As salt concentration increases, the ether oxygens become saturated with Li ions, which are hypothesized to lead to the formation of ion clusters, which could increase the chain dimensions. Additional SANS studies were used to probe the nature of the ion clusters and found that the clusters had a characteristic size of 0.6 nm.

The final section of this dissertation focuses on the effect of salt concentration on the polymer chain dynamics of SEO/LiTFSI. Neutron spin echo (NSE) spectroscopy was used to measure the segmental dynamics in the Rouse regime (t≤10 ns) and the polymer chain dynamics in the reptation regime (t≥50 ns). In both regimes, the dynamics slow down as salt concentration increases. This is quantified through an increase in monomer friction seen at low times and a decrease in tube diameter seen at longer times. The NSE results were used to calculate a normalized longest molecular relaxation time and it was found that block copolymers will flow more slowly with increasing salt concentration. This is important in a battery operating at high current wherein the diffusion of salt ions in one direction must induce diffusion of polymer chains in the opposite direction.