However, silicon anodes are problematic because the material’s volume changes by up to 300-400% upon the insertion and extraction of lithium ions during charge/discharge cycles. This results in pulverization and capacity fading, as well as design issues, since the expansion needs to be managed within the cell.

Researchers are exploring a variety of approaches to address those problems, including using composite electrodes of silicon and carbon, coated with a binder such as CMC (carboxymethylcellulose), to maintain particle-particle contact after many charge/discharge cycles; limiting the voltages over which the material is cycled, thereby limiting the extent of volumetric expansion by sacrificing some capacity, and using nanoparticles and nanowires to improve capacity retention.

Such work, note the authors in their paper, “may result in the commercialization of a next generation of Li-ion batteries with silicon negative electrodes in the near future.”

However, the exact nature of the processes that limit the cycle life has been poorly understood up to now.

The binary phase diagram of Li-Si system consists of four reported crystalline lithium silicides, from Li₁₂Si₇, to the increasingly lithium-rich phases Li₇Si₃, Li₁₃Si₄ and Li₂₁Si₅. However, at room temperature, silicon does not form any of these phases on electrochemical lithiation, but instead undergoes a crystalline to amorphous phase transition, forming a lithiated amorphous silicide. This phase then recrystallizes at deep discharge to form a metastable phase, Li₁₅Si₄, which is isostructural to the thermodynamic phase, Li₁₅Ge₄. This final composition provides the theoretical capacity of 3.75 Li per Si (3,572 mA hg^-1). Interestingly, lithiated nanoparticles of crystalline and amorphous thin films less than 2 µm thickness of silicon do not recrystallize at deep discharge. Unfortunately, due to the amorphous nature of the lithiated silicides, it is not possible to monitor all the structural changes that occur during lithium insertion/removal with conventional methods such as diffraction. The short-range order of the amorphous materials remains unknown, preventing attempts to optimize performance based on electrochemical-structure correlations.

—Key et al. (2009)

In their work, the researchers used a flexible plastic lithium-ion battery (LIB) design and a combination of static, in situ and MAS, ex situ 7Li and 29Si NMR spectroscopy as local structural probes in order to monitor/identify the changes in the short-range order that occur during the first discharge of crystalline silicon in a working LIB. They showed that their method can be used to capture changes that cannot be readily seen by ex situ methods.

They also found that one of the various lithium silicide phases formed during battery operation reacts with the battery electrolyte in a self-discharge process.

This excess-Li phase is extremely reactive in the electrolyte and the Li-Si cell “self-discharges”, leading to the loss of Li from this phase, and an accompanying increase in the open circuit voltage. The amorphous lithium silicides also appear to be reactive and the self-discharge mechanisms at lower stages of discharge are currently under investigation, along with a more detailed analysis of the changes that occur during the first charge and in subsequent discharge-charge cycles.

The self-discharge mechanism could represent a mechanism for capacity loss in Si batteries if they are discharged to too low voltages, which can occur if the batteries are cycled at too high rates. Furthermore, the reactivity of this phase may also be a safety concern. However, our preliminary experiments indicate that CMC inhibits this discharge process significantly, cells taking days to months to relax. These results indicate that the strong binding of CMC to silicon is important in preventing capacity loss due to side reactions with the electrolyte.

Finally, the results highlight the ability of in situ NMR methods to capture dynamic processes that are not straightforward to observe by using ex situ NMR methods. A wide range of different applications of the method to investigate different LIB electrode materials and intercalation/conversion processes, along with side-reactions such as those associated with SEI formation can be foreseen.

—Key et al. (2009)

Resources

• Baris Key, Rangeet Bhattacharyya, Mathieu Morcrette, Vincent Seznc, Jean-Marie Tarascon and Clare P. Grey (2009) Real-Time NMR Investigations of Structural Changes in Silicon Electrodes for Lithium-Ion Batteries. J. Am. Chem. Soc., Article ASAP doi: 10.1021/ja8086278