Identification and quantification of ageing mechanisms in Li-ion batteries by Electrochemical impedance spectroscopy .

The transportation sector reported almost a quarter of global CO 2 emissions. Thus, efforts to decarbonize this sector are essential to achieving net zero emission goals. Among the actions to mitigate the effects of climate change in favor of decarbonization, lithium-ion electric vehicle market has expanded over the past years because of both scientific advances and encouraging public policies. The analysis of ageing in lithium-ion batteries is essential to ensure optimal performance and determine the end of useful life for that purpose. The degradation of lithium-ion batteries is a complex multi-causal process. The ageing mechanisms could be grouped mainly into three degradation modes: Loss of Conductivity (CL), Loss of Active Material (LAM) and Loss of Lithium Inventory (LLI). Ageing battery or state of health was tracked based on capacity and power. Through the state of health (SoH) the degradation of the battery is quantified based on the decrease in capacity. However, the definition of SoH does not include an indication of the underlying deterioration mechanisms causing the degradation. Combined with electrochemical impedance spectroscopy (EIS), degradation modes can be identified and quantified non-destructively with the aim of find the correlation between their evolution with SoH. This paper proposes a method to identify and quantify the ageing mechanisms in commercial 18650 NMC lithium-ion batteries over time using the EIS technique. The EIS spectra were fitted to the equivalent electrical circuit. Through the variation of the calculated parameters with time, the main mechanisms responsible for the degradation are identified, associating the Rohm increases with CL, Rsei and Rct with LLI and Rw with LAM respectively. Correlation between degradation modes with SOH were reported.


Introduction
Mitigation of climate change is a key pressing problem we are facing today. Addressing this problem will require decarbonization of energy systems, particularly, those related to mobility. Achieving zero emissions in the transportation sector confronts us with new related challenges which involves investigation and development in energy electrochemical storage in batteries.
The use of Lithium-ion batteries in electrical vehicles is expanding very fast. However, it remains a considerable amount of research regarding degradation mechanisms and ageing modes with regards transport applications to predict remaining lifetime based on the development of battery models [1][2][3][4][5][6][7].
Ageing mechanisms are usually classified into three degradation modes named as: Loss of conductivity (CL), Loss of Lithium Ion (LLI) and Loss of Active Material (LAM) [8][9]. Loss of Lithium inventory denotes lithium ions consumption due to decomposition reactions, SEI layer growth and lithium plating. Loss of active material refers to the fact that active materials of anode and cathode are no longer available due to structural disarranging, particle cracking, graphite exfoliation, loss of electrical contact due to passivation surface layers. Characterization of ageing effects is very important meanwhile the quantification of the evolution of each degradation mode over the time has the potential to predict the remaining capacity and the amount of energy stored in the battery for each time. Several electrochemical techniques have been widely explored in investigations to identify and quantify degradation mechanisms [10][11][12][13][14][15][16][17]. Previous works [15][16][17][18] have demonstrated the potential of Electrochemical Impedance Spectroscopy (EIS) to identify ageing mechanisms in LIBs implementing this technique online as well as offline. Electrochemical Impedance Spectroscopy is a robust electrochemical technique which allows to recognize and track the evolutions of lithium-ion battery degradation processes. The advantage of EIS for characterization of ageing mechanisms is the large amount of information it can provide for understanding degradation modes.
Concerning NMC cathode chemistry, numerous outcomes were stated in literature. Several authors [19][20][21][22][23] addressed the increase of ohmic resistance to electrolyte decomposition and attributed to SEI layer growth the increase of impedance in the mid-frequency range of the spectra. Jalkanen et al. [24] also ascribed the increase of ohmic resistance to electrolyte decomposition as well as assigned impedance increase in the mid-frequency region to processes such as: lithium plating, mechanical cracking, loss of cathode active material contact and SEI layer growth. Temperature effect was also studied by Buchberger et al. in [25]. At room temperature impedance rise in the Mid-frequency region of the spectra has been assigned to electrolyte decomposition and dissolution of metals from NMC cathode. The usual procedure followed to evaluate impedance spectra of Li-ion cells is by electric equivalent circuits model track by a Complex Nonlinear Least Square fitting [26]. Sun et al. [27] refers to first and Second-Order Equivalent Circuit Models. To analyze the spectroscopy response, these authors divide the impedance spectrum into four zones according to frequency regions, identifying the characteristic elements of each of them according to what is detailed below.
Model resistances are associated with some electrochemical phenomena and can be used as indicators for evaluating the state of health (SoH)of the battery and its evolution with time operating under different conditions [28]. An ohmic resistance (Rohm) in serie with an inductor (L) characterize the region in the spectra at ultra-high frequencies. The high-frequency region is characterized by the SEI structure. It is denoted by a SEI layer resistance (Rsei) connected with a constant phase element (CPEsei) in parallel. On the other hand, in the medium-frequency region a charge transfer resistance (Rct) is depicted while the low-frequency region is characterized by a Warburg element Zw. Both are connected in series in the model and in parallel with a CPE (CPEdl), representing the effect of the double-layer capacity.
The detection and quantification of ageing and degradation modes in li ion batteries during real-daily operation is still a goal in battery commonplace applications. Ageing and degradation mechanisms often imply complex processes in operating conditions, being affected by conduction cycles, temperature effects, resting time between cycles, etc. So, it is real a difficult task quantifies ageing and degradation effects to predict how much loss of capacity will occur under a different regime. In this work, EIS is applied to study degradation mechanisms on 18650 commercial NMC cells. Fitted parameters were correlated with the SOH.

Experimental
Life cycle tests and EIS measurements at room temperature were carried out to study the change of resistance values and analyze the corresponding degradation modes of lithium-ion batteries. The test samples are commercial 18650 NMC lithium-ion. The batteries studied were subjected to 800 charge/discharge cycles with a current intensity of 1 A and 3 A, respectively. Measurements were performed on the Gamry Interface 5000E™ potentiostat/galvanostat. EIS measurements were made at open circuit potential at 50% SOC at room temperature, sweeping a frequency range between 10kHz and 10mHz, with an amplitude of 5mV and taking 10 points per decade. A Second-Order Equivalent Circuit Model reported by Sun [24] was employed to fit experimental EIS values. The equivalent circuit model employed is divided into four sections, which is fully corresponded to the frequency division in the impedance spectrum shown in Figure 1.  EIS measurements represented as Nyquist plots are depicted in Figure 3.   Where the subscript n represents the different cycles during the ageing processes and subscript 1 corresponds to the initial cycle, the battery freshly state During the first cycles in ageing process only one semi-circle in the impedance curve is shown. Based on the overlapping of two semi-circles it can be concludes that time constants of the internal processes associated with LLI are similar.
During the ageing process, Rohm, Rct and Rsei show a strong increasing trend, with a clear turning point at cycle 400 ( Figure 4). These changes are consistent with the corresponding ageing mechanisms. Figure 5 depicts the evolution of Y0 with cycles. During ageing Y0 decreases, consequently lithium diffusion coefficient also decreases. After obtaining each resistance value in the ageing stage, the degradation modes can be quantified by the relative change of its associated resistances as it was stated in eq (1)(2)(3).
Combined with the mechanism analysis previously detailed, we use Rohm to represent CL, Rsei and Rct to represent LLI and Y0 to characterize LAM. Certainly, the ageing mechanisms included in each degradation mode cannot be completely described by these resistances. Observing the quantitative results, we found that, as the ageing process progresses, resulted values of the three modes show a rising tendency. Compared with CL and LAM, LLI shows the most extensive increase. The fast growth of LLI is mainly attribute to the continuous growth of SEI films, which is also considered as the principal explanation for capacity fade.
It is worthwhile noticing that LAM and Y0 evolutions show a linear correlation with the state of health of the battery in all the range. CL and LLI depict a linear behaviour for SOH values higher than 40%.

Conclusions
This paper identifies and quantifies the ageing mechanisms in commercial 18650 NMC lithium-ion batteries over time using the EIS technique. The EIS spectra were fitted to the second order equivalent electrical circuit model. Through the variation of the calculated parameters with time, the main mechanisms responsible for the degradation are identified, associating the Rohm increases with CL, Rsei and Rct with LLI and Rw with LAM respectively. The shapes of the impedance spectra show dissimilar transformation tendencies during the ageing process. During the first cycles, there are not two distinguishable semi-circles in the medium-high frequency region. The overlap of two semi-circles indicates the time constants of the internal processes are similar. The quantification methods presented not only ensure the corresponding relationship between degradation modes and resistance characteristics but also allow quantitative analysis under different conditions. It was verified by experiments that, LLI always dominates the ageing process. LAM depicts a linear correlation with SOH for all values, though CL and LLI show a linear behaviour for SOH values higher than 40%.