UC Berkeley

The next generation of rechargeable batteries must have significantly improved gravimetric and volumetric energy densities while maintaining a long cycle life and a low risk of catastrophic failure. Replacing the conventional graphite anode in a lithium ion battery with lithium foil increases the theoretical energy density of the battery by more than 40%. Furthermore, there is significant interest within the scientific community on new cathode chemistries, like sulfur and air, that presume the use of a lithium metal anode to achieve theoretical energy densities as high as 5217 W·h/kg. However, lithium metal is highly unstable toward traditional liquid electrolytes like ethylene carbonate and dimethyl carbonate. The solid electrolyte interphase that forms between lithium metal and these liquid electrolytes is brittle which causes a highly irregular current distribution at the anode, resulting in the formation of lithium metal protrusions. Ionic current concentrates at these protrusions leading to the formation of lithium dendrites that propagate through the electrolyte as the battery is charged, causing it to fail by short-circuit. The rapid release of energy during this short-circuit event can result in catastrophic cell failure.

Polymer electrolytes are promising alternatives to traditional liquid electrolytes because they form a stable, elastomeric interface with lithium metal. Additionally, polymer electrolytes are significantly less flammable than their liquid electrolyte counterparts. The prototypical polymer electrolyte is poly(ethylene oxide). Unfortunately, when lithium anodes are used with a poly(ethylene oxide) electrolyte, lithium dendrites still form and cause premature battery failure. Theoretically, an electrolyte with a shear modulus twice that of lithium metal could eliminate the formation of lithium dendrites entirely. While a shear modulus of this magnitude is difficult to achieve with polymer electrolytes, we can greatly enhance the modulus of our electrolytes by covalently bonding the rubbery poly(ethylene oxide) to a glassy polystyrene chain. The block copolymer phase separates into a lamellar morphology yielding co-continuous nanoscale domains of poly(ethylene oxide), for ionic conduction, and polystyrene, for mechanical rigidity. On the macroscale, the electrolyte membrane is a tough free-standing film, while on the nanoscale, ions are transported through the liquid-like poly(ethylene oxide) domains.

Little is known about the formation of lithium dendrites from stiff polymer electrolyte membranes given the experimental challenges associated with imaging lithium metal. The objective of this dissertation is to strengthen our understanding of the influence of the electrolyte modulus on the formation and growth of lithium dendrites from lithium metal anodes. This understanding will help us design electrolytes that have the potential to more fully suppress the formation of dendrites yielding high energy density batteries that operate safely and have a long cycle life.

Synchrotron hard X-ray microtomography was used to non-destructively image the interior of lithium-polymer-lithium symmetric cells cycled to various stages of life. These experiments showed that in the early stages of lithium dendrite development, the bulk of the dendritic structure was inside of the lithium electrode. Furthermore, impurity particles were found at the base of the lithium dendrites. The portion of the lithium dendrite protruding into the electrolyte increased as the cell approached the end of life. This imaging technique allowed for the first glimpse at the portion of lithium dendrites that resides inside of the lithium electrode.

After finding a robust technique to study the formation and growth of lithium dendrites, a series of experiments were performed to elucidate the influence of the electrolyte’s modulus on the formation of lithium dendrites. Typically, electrochemical cells using a polystyrene – block¬ – poly(ethylene oxide) copolymer electrolyte are operated at 90 °C which is above the melting point of poly(ethylene oxide) and below the glass transition temperature of polystyrene. In these experiments, the formation of dendrites in cells operated at temperatures ranging from 90 °C to 120 °C were compared. The glass transition temperature of polystyrene (107 °C) is included in this range resulting in a large change in electrolyte modulus over a relatively small temperature window. The X-ray microtomography experiments showed that as the polymer electrolyte shifted from a glassy state to a rubbery state, the portion of the lithium dendrite buried inside of the lithium metal electrode decreased. These images coupled with electrochemical characterization and rheological measurements shed light on the factors that influence dendrite growth through electrolytes with viscoelastic mechanical properties.

Next, the morphology of lithium dendrites formed upon many charge and discharge cycles were compared to the morphology of those grown upon a continuous charge using a combination of X-ray and electron microscopy techniques. When cycled, the lithium dendrite morphology consisted of multiple interconnected lithium globules that amassed to form a structure that punctured the electrolyte causing the cell to fail by short-circuit. When charge is passed in only one direction until the samples fails by short-circuit, the dendrite morphology is markedly different. Instead of observing a multi-globular morphology, a single lithium-filled globule encased in a polymer sac expands until it touches the counter-electrode. These blunt structures formed in solid polymer electrolytes are in stark contrast to the needle-like morphologies observed in lithium dendrites formed in liquid electrolyte systems.

Time-resolved hard X-ray microtomography was used to monitor the internal structure of a symmetric lithium-polymer cell during galvanostatic polarization. The microtomography images were used to determine the local rate of lithium deposition, i.e. local current density, in the vicinity of a dendrite growing through the electrolyte. Measurements of electrolyte displacement enabled estimation of local stresses in the electrolyte. At early times, the current density was maximized at the dendrite tip, as expected from simple current distribution arguments. At later times, the current density was maximized at the dendrite perimeter. We show that this phenomenon is related to the local stress fields that arise as the electrolyte is deformed. The local current density, normalized for the radius of curvature, decreases with increasing compressive stresses at the lithium-polymer interface. To our knowledge, our study provides the first direct measurement showing the influence of local mechanical stresses on the deposition kinetics at lithium metal electrodes.