Silicon has proven to have a great potential as anode material for lithium ion batteries due to its high theoretical capacity, however; there are several obstacles that need to be overcome in order to make it a commercially viable option. Two of the main issues stem from the fact that silicon undergoes a large volume change during lithiation and delithiation [1]. This makes forming a stable solid electrolyte interphase (SEI) difficult, resulting in a continuous loss mechanism of electrolyte and lithium as SEI is formed and broken each cycle. Uneven expansion and contraction causes the silicon to fracture, both exposing new surface on which more SEI might form, as well as electrically disconnecting material from the electrode, rendering it inactive.

In this work we investigate the use of amorphous silicon nitride as an alternative to pure silicon anodes. Silicon nitride is believed to form lithiated silicon and lithium nitride during the initial lithiation [2]. We hypothesize that, under the right circumstances, the resulting material end up having small and interconnected domains of the two phases, complementing each other: Lithium nitride, which is known to be a good lithium ion conductor, enhances lithium transport, while the silicon would, in addition to store lithium, contribute to the electronic conductivity of the electrode. In a sufficiently fine structure, the silicon domains would be able to expand and contract with little to no fracturing, owing to dimensional stabilization, short diffusion distances, and fast lithium diffusion in the lithium nitride around the domains.

We explore this theory using a thin film electrode system. For this purpose, a-SiN_x:H thin films are deposited on copper foil substrates using plasma enhanced chemical vapor deposition (PECVD) with silane (SiH₄) and ammonia (NH₃) as precursors. PECVD allows the composition of the films to be varied by changing the ratio of the precursor gases in the plasma. Ellipsometry and transmission electron microscopy (TEM) are used to determine the thickness, composition, structure and quality of the pristine films. Post cycling characterization is conducted in order to investigate the conversion reaction and measure the expansion during the initial lithiation, as well as to monitor long term structural changes during cycling. Electrochemical tests are conducted in 2032 coin cells, with a lithium metal counter electrode, using a commercial electrolyte with 5% FEC and 1% VC, and cycling is conducted between 50 mV and 1 V.

Three 40 nm thin film electrodes cycled at C/6 for 6 cycles and at 1C for 260 cycles had an average capacity of 1200 mAh/g on the first 1C cycle, and only experienced a slight increase in capacity to 1340 mAh/g during the next 150 cycles. While this is notably lower than the capacity of silicon, it is still above the point where further increasing the anode capacity results in only a negligible increase in the total cell specific capacity. The cycling stability in the remaining cycles were excellent, and after finishing the 260 cycles the cells still retained an average capacity of 1330 mAh/g.