**2. Test Protocol Description**

The commercial Li-ion 18650 cell format was chosen for this protocol. Tests were performed on INR18650-30Q cells (from Samsung DI Co., Yongin, South Korea). The cell chemistry is a blend of graphite/silicon (approximately 2–3% Si) in the negative electrode and a blend of NCA (LiNi0.89Co0.10Al0.01O2) and NCO (Ni0.8Co0.2) in the positive electrode. Performance of each cell batch was measured at 1C: capacity was 3000 mAh and energy density was 230 Wh/kg. Cells were aged using a battery electric vehicle (BEV) protocol with representative ageing cycles at various temperatures and calendar ageings according to the international standard "IEC 62660-1". The temperature tests were (−20 ◦C, 0 ◦C, 25 ◦C, 45 ◦C) and the calendar ageing was realised at 45 ◦C at constant voltage (CV) at 4.2 V and at 45 ◦C in an open circuit voltage (OCV). Ageing profiles contained current pulses in charge or discharge that simulated car breaking and acceleration. For each condition 15 cells were aged on a battery tester (*PEC*® SBT 05250). The choice of these conditions was guided by the IEC 62660 standard.

During the test, the state of health (SOH) was followed via an electrical performance measurement realised at 25 ◦C every 28 days for cells aged by cycling and every 6 weeks for calendar ageing on a battery cycler. This electrical test allowed us to track the evolution of some relevant characteristics of the cell: capacity, internal resistance [6,7], and nominal voltage. The internal resistance was evaluated by the current pulse method (10 s, I discharge max = 15 A).

In order to identify the favoured ageing mechanism in each condition, aged cells were dismantled in a glove box under argon (>99.999%) for post-mortem analyses. Scanning electronic microscopy (SEM) and electrochemical analyses were performed on both positive and negative electrodes recovered from the aged cells. Glow discharge-optical emission spectroscopy (GD-OES) and solid state 7Li NMR were realised in addition on the negative electrodes to clearly identify the main ageing mechanisms. Indeed, different studies [8–10] have shown that the NCA positive electrode has no significant changes after ageing. The only suspect ageing mechanism on such NCA active material is the transition metal (Ni, Co, Al) dissolution during high-temperature ageing.

To help in the comprehension of mechanisms of ageing, electrochemical impedance spectra (EIS) measurements were carried out on cells at 60% SOC. EIS was performed using a VMP3 multi-potentiostat (Biologic Science Instruments) equipped with an EIS option. The frequency range used was from 1 mHz to 500 kHz, starting from high frequency to low frequency in a logarithmic scan. The alternating voltage amplitude was set to 5 mV at 25 ◦C.

Finally, an accelerated rate calorimetry (ARC) test was performed. An ARC test is a technique that allows measurement of the self-heating of a cell. The cell is heated in 5 ◦C steps. After a 35 min break, if the heat rate *β* is not over 0.02 ◦C/min, a new step is started. As soon as the heat rate reaches over 0.02 ◦C/min, the so-called "onset temperature" is identified.

In this study, only the beginning of self-heating was studied. Tests were stopped at 90 ◦C in order not to damage the cell and allow further investigations. Two cells per ageing condition were tested. A fresh cell was also used as a reference in order to compare the different safety behaviours.
