By Kyle Proffitt
February 5, 2024 | Researchers from the John A. Paulson School of Engineering and Applied Sciences at Harvard University have created new solid-state batteries boasting 12-minute charge time and 80% capacity retention after 6,000+ charge-discharge cycles. While these numbers are impressive, the fundamental advance actually comes from understanding and controlling physical forces at the anode to enable improvements in density and performance. The research was published in Nature Materials January 8, 2024 (DOI: 10.1038/s41563-023-01722-x).
Compared with traditional liquid electrolyte lithium-ion batteries, solid-state batteries offer several potential advantages. They can provide greater energy density, enabling lighter, smaller battery packs. Safety is increased by avoiding the flammable liquid electrolyte. Additional expectations include faster charging, longer lifetime, and better tolerance to high and low temperatures.
Despite those promising characteristics, several factors have stymied solid state battery advances, and some of these problems converge at the anode. When charging a lithium ion battery, the lithium ions migrate to the anode, where two primary options await. They can incorporate into the material there, often intercalating between layers (also known as alloying, or lithiation), or they can recombine with electrons, becoming reduced to lithium metal that deposits in or on the surfaces, a process referred to as plating. Both processes can be harnessed effectively, but the dynamics and associated effects of each present unique challenges.
If the primary anode activity of the lithium is intercalation, swelling is a complication. For instance, silicon, although providing ~10x capacity for lithium ions compared with graphite, swells as much as 300% in this process. This can be particularly problematic in a solid state cell, as the expansion can initiate cracks or pulverize silicon particles and render them ineffective. This is why silicon was explored and then generally dismissed as a good anode material starting back in 1976 (except for high-temperature applications), although newer efforts involving nanostructures have given silicon new anode life.
Using lithium metal at the anode provides a similar capacity for additional lithium, but the lithium only plates in this case. This lithium plating presents challenges because uneven layers can lead to the formation of dendrites, spikes that can puncture the separator between cathode and anode, creating short circuits and inactivating the battery. Dendrites also reduce battery efficiency, as the stacked-up lithium is not as effectively stripped on battery discharge—if it does not recycle, this portion of the lithium is effectively dead. If lithium plating is used to store energy, then special designs must be employed to ensure smooth, even layering.
Interestingly, it’s actually a function of the rigidity and constriction in a solid cell that the authors exploit in their new design. The researchers, led by Professor Xin Li, created a composite of micrometer-sized silicon and graphite to protect lithium metal in the anode. For a traditional lithium ion battery’s graphite anode, the theoretical energy density is 400 mAh/g. Silicon and lithium metal anodes both have approximately 10 times this capacity, but the composite silicon-graphite (SiG) anode used by Li’s group showed an energy density in preliminary experiments greater than 5,600 mAh/g. The researchers used a range of x-ray microscopic methods and specialized experiments to determine that this was ultimately possible because the lithium only alloyed with the micrometer-sized silicon particles to a depth of ~65 nm, after which plating began. In other words, it was a little bit of the best of both worlds. The round silicon particles accepted some lithium, first forming a thin skin, but not so much that particles were pulverized.
Competition Between Lithiation and Plating
But what prevented deeper silicon lithiation? Essentially, the particle size and the local contact with solid electrolytes disallowed deep penetration and favored an early switch to plating. At a point of contact between a silicon particle and the solid electrolyte, lithiation can begin. However, this reaction induces a local strain in the immediate vicinity (from expansion), limiting the free movement and intercalation of additional lithium ions. Around a curved silicon particle, the process continues until the particle is effectively coated. Then plating begins, filling in voids between particles. “Li metal plating is forced to occur through the dynamic interaction between lithiation and plating at those shallow lithiation nanosites at the Si particle surface,” the authors explained in their report. They refer to this material characteristic as its “constriction susceptibility”—how much local strain affects lithium incorporation in the vicinity.
But an important thing happens compared with other lithium metal plating designs. The dynamic interaction between lithiation and plating “helps more homogeneously distribute plating current densities at the anode,” the authors stated. As a result, the deposition is even, and dendrites are less likely to form. The competition between lithium intercalation and lithium plating apparently enables plating to occur evenly and lay down in thick layers. This is unlike batteries with inert scaffolding at the anode to encourage lithium metal deposition in the crevices; according to the authors, their approach actually creates an “active” 3-dimensional scaffold with more evenly distributed current density.
Room for More Anode Materials
Li and colleagues took the findings further to explore the constriction susceptibility of 59,524 potential anode materials in the Materials Project using high-throughput computational methods. The researchers focused on calculation of a parameter called the critical modulus, or Kcrit—essentially a stress level above which no more lithiation can occur. Materials with a low Kcrit are likely to behave similarly as silicon, limiting lithiation during reaction-induced strain. By considering this value in combination with a material’s ability to host lithium and then plotting that combination against the voltage needed to push lithium into these materials, the researchers suggest a range of additional materials with constriction susceptibility that could be exploited. They identified over 11,000 known materials maximizing these parameters, including enrichment in Mg, Si, Ge, Sn, In, Al, and Ag along with some specific combinations.
The researchers turned back to readily available silicon for proof-of-principle battery designs, creating both coin cells and pouch cells with desirable properties. The pouch cell they created was similar in size to the face of a smaller Apple Watch and was created using an NMC83 cathode, contrasting layers of solid electrolytes, and a lithium metal anode protected with the new SiG composite. It accomplished 12-minute charge time, >6,000 cycle longevity (80% capacity retention), and an energy density of 218 Wh/kg. Although this energy density falls short of commercially available lithium ion batteries, the authors are hopeful that small design improvements will boost this value.
In line with the protocols of many laboratories working toward solid-state batteries amenable to rapid charging, the authors’ results were obtained using a stack pressure 250 times greater than atmospheric pressure to ensure good contact between electrodes and solid electrolyte. Reducing this pressure to 50 times atmospheric pressure reduced cycling stability to 2800 cycles. It is not entirely clear how such pressures can be maintained outside of a laboratory setting, but the authors note that different manufacturing methods can create denser layers to mitigate this requirement and that reducing separator thickness and increasing cathode loading (how much lithium) will improve overall performance. Of note, the same researchers previously reported solid-state batteries that charged in 3 minutes and survived 10,000 cycles using contrasting layers of solid electrolytes (the same electrolyte stacking approach was preserved here), but this was performed with smaller coin cells and twice the stack pressure. The current work has been licensed to Adden Energy, a Harvard spinoff company co-founded by Li and three additional alumni, and they have reportedly already scaled the technology to a smartphone-sized pouch battery.