Silicon in the anode is becoming increasingly common in modern Li-ion batteries. Particularly, in the automotive industry active materials with high capacities are of great interest, since volumetric energy density is a key design parameter due to limited available space in electric vehicles. Silicon with its high theoretical capacity of 3579 mAh g-1 and material abundance offers opportunities but also challenges. One of the most critical challenge is the high volume expansion of silicon upon lithiation. One approach to address this is to use silicon-oxide (SiOx) rather than pure silicon. In combination with other advantages (higher electrical conductivity and faster Li-diffusion) SiOx thus is a promising candidate to combine the advantages of high volumetric capacity yet lower volume expansion compared to pure Si. Also with new Ni-rich cathode materials such as NMC811 the energy density of lithium-ion cells could be increased markedly but lithiiation & delithiiation processes lead to a reversible volume change of the symmetric crystal structure.
Various studies have investigated the thickness change of lithium-ion cells with Gr/Si anodes. However, the correlation between volume expansion and thickness change of the anode was not further investigated, also they did not differentiate between the contribution of Si and graphite to the thickness and porosity change of the anode. Transition of cathode expansion to increasing full cell thicknesses (pouch cell), particularly for higher Ni contents, has not been confirmed yet. There are hardly any studies of the influence of the Ni-rich materials on cell thickness growth, especially in cells containing silicon. Previous studies have also mostly used small lab-scale pouch cells to investigate thickness change, which leaves a degree of uncertainty when transferring results onto much larger pouch-cells as they are common in automotive applications. To the best of our knowledge no single study exists that precisely investigates and validates the volume expansions and thickness changes on all macroscopic levels, starting with industry-standard state-of-the-art Gr/SiOx|NMC811 large-format pouch-cells down to electrode level and down to active material level. With this study we closed this gap by resolving the contribution of each electrode to the pouch-cell expansion and the contribution of both anode active materials to the anode expansion and NMC811 to cathode expansion.
The investigated cell was cycled under constant pressure and constant temperature. The measured reversible cell swelling showed great repeatability after some initial irreversible thickness change. At the beginning of discharge, the cell thickness first decreased and then increased again. Electrochemical dilatometry revealed that this increase in cell thickness was caused by a steep increase in cathode thickness at the beginning of discharge. In this case, the decrease of NMC volume counteracted the volume expansion of the Gr/SiOx anode, which is interesting for automotive battery and cell design. Pouch-cell thickness showed large hysteresis, which was caused by the voltage and volume hysteresis of SiOx in the anode. For closer inspection, the capacities of both anode active materials were determined using DVA. The capacities were then used to calculate the voltage curves of SiOx and graphite as a function of the anode SOC. The sum of both active material voltage curves showed great agreement with the actual half-cell measurements of the composite anode and allowed to calculate the SOC and the volume expansion of each active material as a function of anode SOC. The correlation between the volume expansion and the thickness change of the anode was investigated starting with the theory of constant pore volume and decreasing porosity. The resulting values overestimated the thickness change which indicated that the pore volume actually decreased rather than staying constant. Additional proportionality constants were introduced to account for this behavior and enabled a better fit. Additionally, the model suggested a large influence on the cell thickness by SiOx below 20% SOC. However, similar to previous findings, no such influence was actually measured. This indicated that the initial large volume expansion of SiOx was swallowed by pore volume. After integrating this phenomenon into the model a very good fit to the measured data was achieved. With this fit the it was possible to determine the exact proportions of the change in thickness of graphite and SiOx in relation to the total change in thickness of the anode.