Degradation-Safety Analytics in Lithium-Ion Cells and Modules Part II. Overcharge and External Short Circuit Scenarios

Cell manufacturers specify an operating voltage window (VW) based on several factors of which the chemistry of the electrode couple is one of them. This voltage window represents an optimal range as provided by the manufacturer's specification sheet, in which a margin is included for safety while providing the optimum cycle and calendar life. The upper limit provided by the manufacturer is set to ensure that the cathode material does not release lithium ions beyond its cyclable lithium amount which can lead to cathode structure destabilization, especially in metal oxide cathode materials, the electrolyte is stable, and the anode does not experience deterioration effects such as lithium plating, dendrite formation and excessive growth of the solid electrolyte interphase (SEI). Thus, overcharge is defined as the phenomenon in which the cell is charged beyond the manufacturer's upper cutoff voltage or charged at very high currents. In this study we look at the overcharge condition of charging beyond the manufacturer's upper cutoff voltage.

Overcharging can also lead to significant gas generation within the cell before driving the cell into a thermal runaway. In prismatic form factors, and particularly in cells with thin cases or with soft-pouch, the generation of gas within the cell causes swelling of the cell and can force separation between the electrodes, effectively limiting ion transfer and interrupting the charge/discharge process. ¹³ The geometry of cylindrical cells prevents electrode separation even if gas generation occurs. ^14,15 Cell designers have developed designs that include a central mandrel that directs the flow of gases generated inside the cell and a mechanical current interrupt device (CID) for cylindrical cell designs. When activated due to increased internal pressure, CIDs physically and irreversibly disconnect the cell from the external circuit rendering the cell fail-safe.

The extent of degradation caused by overcharging depends on several factors such as the rate of charge, the environmental temperature, extent of overcharge, cell type, size, and chemistry. Leising et al. studied the effect of charge rate for overcharge in commercial prismatic LCO/graphite cells. ¹⁶ They observed that at low rates of overcharge, swelling, temperature and internal resistance increased at a lower rate compared to cells overcharged at higher rates. Cathode was also identified as the main source of heat generation confirming that overcharge is a cathode-dominated process, which is consistent with DSC studies by Zhang et al. ^2,17 Studies have also been carried out to compare the stability of different electrode materials under overcharge conditions. LCO was found to be a more thermally unstable cathode material compared to LMO when paired with graphite as anode material. ¹⁸ Similarly, in the work by Larsson et al., LIBs with LFP-based cathodes show more stability compared to cobalt-rich cathode materials. ¹

The studies listed above share a common characteristic, all of them were performed on fresh cells. The goal of this work is to fill the knowledge gap on the interaction between the characteristics induced by aging and those induced by overcharge. In Part I, the aging effect on cylindrical cells was presented. ¹⁹ This paper will cover the overcharge characterization tests on cells aged to various levels and compare them to the results obtained by overcharging fresh cells. This was accomplished by overcharging cells that had been aged up to 10, 15 and 20% capacity fade (CF). The overcharge tests were also carried out on fresh and aged modules. The cells and modules were cycled to different voltage ranges, operating temperatures, and realistic discharge conditions under the aging part (Part I) of the study. ¹⁹ This study characterized the electrochemical, thermal, and morphological changes induced by the overcharge test and was used to elucidate if aging is a benign or a malign factor with respect to the safety aspects. Results for external short test are briefly discussed.

Despite the fact that a battery is discharged at variable currents during its use in an EV, there are few publications in the literature analyzing this effect. Panchal et al. investigated the large batteries subjected to different drive cycle profiles at different environmental temperatures. ²⁰ Their analysis focused on the long range performance but not the actual degradation mechanisms. The drive cycle test emulates the discharge pattern of a highway-driven EV Fig. 12a, that travelled 10.26 miles in 765 s. Unlike the CC discharge protocol, the drive cycle (DC) test shows the variable loads demanded by the vehicle when accelerating. The positive current spikes correspond to the charge induced by periodic regenerative breaking. Assuming the vehicle is used every day, a discharge pattern with intermittent continuous stops was created as shown in Fig. 12b. The test was conducted on cells from the same lot as those used for the aging and overcharge test in single cells and modules. From a fully charged condition, a group of three drive cycles is applied followed by a 2-h rest. Then another group of three drive cycles were applied followed by a 4-h rest. The cell was then charged fully using the CCCV protocol to 4.2 V at a C/2-rate and kept at constant voltage of 4.2 V until the current fell to 50 mA. It should be noted here that the drive cycle profile consists of pulses of charge and discharge. The cycling process was repeated for a week at three conditions, T = 10 °C, 25 °C, and 40 °C, emulating the performance of the vehicle under different thermal environments. At the end of the week, cell capacity and internal resistance were measured at ambient room temperature following the conditioning test as described in Part I. ¹⁹ The weekly test pattern was repeated until the cell capacity, as determined by the full capacity checks, decreased by 20% its initial capacity. Three single cells were aged under each thermal environment.

The manufacturer-recommended voltage window for the cathode/graphite commercial cells is 2.7 to 4.2 V. ²¹ Accordingly, the cell is considered to be overcharged when it is charged above 4.2 V. Prior to the overcharge abuse test, the cells are fully charged to 4.2 V at 1C-rate using a constant current—constant voltage (CCCV) protocol with a cutoff current of 50 mA. The cell is then overcharged using a constant current (CC) of 1C-rate to a 12 V limit for a maximum of six hours or until an off-nominal event happens. The test protocol includes the recording of the cell temperature at the rate of 1 Hz until the cell temperature returns to ambient room temperature. The same overcharge test conditions were used for the single cells that underwent aging using the drive cycle protocol.

The overcharge test on the cells is conducted using a module tester (Arbin system, BT-ML-40V-20A-A). Cell temperature is measured on the cell surface by means of a J-type thermocouple (Omega, TJ72-CPSS-116U-6) taped on the cell surface in the axial direction.

The external short test was conducted on fully charged cells with a 50 mΩ load. Low impedance external shorts are characterized by a load resistance that is equal to or less than the internal resistance of the test article. The average internal resistance, at 50% SOC, of the 25 fresh cells tested was 44 mΩ. The internal resistance of the aged cells was lower than 70 mΩ. For consistency, a 50 mΩ load was chosen for the short circuit test. The short circuit load was maintained for 3 h. The sampling rate for the first 3 s was 1 kHz to capture the initial current spike produced at the application of short circuit. The same external short test conditions were used for the single cells that underwent aging using the drive cycle protocol.

In a similar fashion to the single cells, the modules are fully charged at 1C-rate prior to the overcharge test. The fully charged module is overcharged using a 9 A current up to a 108 V limit for a maximum of six hours or until an off-nominal event happens. Temperature and voltage measurements of the banks are collected along with the module voltage.

After the overcharge and external short tests, the single cells undergo a destructive physical analysis (DPA) in an argon-filled glovebox. The step-by-step procedure is described in the Appendix A of Part I. ¹⁹ This is a sensitive task that requires additional precautions as the electrodes may undergo a short circuit during the disassembly process causing excessive heat that can result in a fire. Representative samples of the electrodes and separators are harvested and analyzed via Scanning Electron Microscopy (SEM) (Hitachi S4800) and Energy Dispersive X-ray Spectroscopy (EDS) (Oxford X-MaxN 80) techniques.

The protocol used to overcharge the cells is exemplified in Fig. 1a, corresponding to the N₁₅OV₃ cell. The preparation test consists of two full charge/discharge cycles followed by a full charge to ensure that the cell is fully charged at the beginning of the overcharge test. Internal resistance is also measured at 50% SOC during the discharge of the second cycle from the preparation test. The idea was to compare internal resistance before and after the overcharge abuse test. However, this was not possible since the internal resistance of the cell after the overcharge test, was too large (> 1 MΩ) indicating that the overcharged cells were internally electrically disconnected.

Zoom In Zoom Out Reset image size

A detailed view of the overcharge test highlighted in green in Fig. 1a is shown in Fig. 1b. During the overcharge test, the voltage increases but never reaches the upper cutoff voltage set at 12 V. Instead, the cell voltage reaches a maximum value (∼ 5.3 V) and then decreases even though the cell is charging. The tip of the curve is a characteristic feature of the overcharge test and indicates that the cell will fail soon. During failure, the CID activates, see Fig. 1b from Part I, ¹⁹ and electrically disconnects the cell. The CID activation is said to be a fail-safe mode and is reflected as the black vertical line in Fig. 1b. The voltage peak is only plotted as an indicator of cell shut down but is neglected for post-processing analysis. As the CID disconnects, the cell resistance increases infinitely and, consequently, the cell voltage reaches the compliance voltage set on the test equipment. Once the CID is activated, the voltage and current drop to zero.

On the other hand, the temperature increases due to the electrochemical reactions and degradation processes induced by the overcharge test, Fig. 1b. Initially, the temperature rises due to the irreversible heat generation caused by the increased intercalation of lithium ions into the anode. ^13,22 Then the temperature increases at a faster rate due to side reactions caused by lithium plating and the degradation of the electrolyte and electrodes. Once the cell shuts down electrically, the temperature continues increasing for a short time, as side reactions continue to occur internal to the cell even after the external supply of current stops. ²

In order to investigate the aging effect on the overcharge response, the voltage curves for the fresh (red line) and aged cells are put together in Fig. 2a. In this figure, only the overcharge portion is plotted for the fresh as well as the aged cells in the normal and reduced VW.

Zoom In Zoom Out Reset image size

The overcharge capacity (Q_OV) indicated in Fig. 2a corresponds to the additional capacity induced only by the overcharge part of the test. In this case, no significant difference is observed with the cells aged to various levels of capacity fade. All aged cells experience CID activation at approximately the same time and overcharge capacity, Q_OV = 0.45 Ah, corresponding to a 113.5% SOC. On the other hand, the fresh cell can be overcharged up to Q_OV = 0.8 Ah, corresponding to a 123.5% SOC. There are two factors that contribute to the difference observed in the response of the cells. The first one is based on the lithium inventory and the second is based on the gas evolution internal to the cell. In the fresh cell, the amount of lithium ion inventory from the cathode that can intercalate into the anode is larger than that for the aged cells since the entire negative electrode is available and it can accommodate some of the additional lithium transferred due to the overcharge. And unlike the fresh cell, aging degradation reduces the amount of lithium ions in the cathode inventory in the aged cells, not only due to displacement of the lithium ions in the cathode structure but also due to its loss to sites inside the anode as well as inside the thickened SEI. As it was explained in our previous paper, CID activation occurs due to an increase in internal pressure caused by the release of gases. ^19,23 In cells fitted with a CID, an additive, lithium carbonate (LiCO₃) is added to cause activation of the CID above a certain overvoltage condition but before the cell becomes unsafe. Above a certain voltage the lithium carbonate decomposes and releases carbon dioxide gas which provides the pressure to activate the CID. Our observation in the overcharge tests indicate that due to other side reactions that occur during the aging process, there were other gases present in the cells and hence the CID activation occurs much earlier in the aged cells compared to the fresh cells. In Part I, it was shown that as the cell ages, its internal resistance increases. ^13,19 The internal resistance for cells aged under normal VW is greater than that of cells aged in reduced VW. Accordingly, during overcharge, the initial voltage rise for cells aged in normal VW is slightly greater than that of cells aged in the reduced VW. However, at the end of the overcharge test no difference was obvious between those cycled in the normal voltage range and those cycled with the reduced voltage range.

The thermal response of the aged cells during the overcharge test is shown in Fig. 2b. The maximum cell temperature of 80 °C occurs in the fresh overcharged cell. The maximum temperature for aged cells is only 55 °C and takes place in the cell aged in the reduced VW with 10% capacity fading. The minimum temperature response occurs in the two cells, N₂₀OV₁ and R₂₀OV₁, with 20% capacity fading. It is important to highlight how the maximum temperature occurs after the CID activates and the cell fails safe. This indicates that side reactions within the cell continue even after the applied current has been removed. In the cylindrical cells that were studied, the activation of the CID prevents the catastrophic results typically observed with an unprotected cell design. In general, it was observed that the higher the capacity fade, the lower the thermal response of a cell to an off-nominal condition. However, the presence and activation of the CID in the cylindrical cells tested limited the catastrophic events under this off-nominal condition.

To investigate the actual role of the aging effect in thermal response, the heat generated by cells during overcharge is calculated using the same fitting method described in Part I. The temperature profile going from the beginning of the overcharge test to the point when the cell temperature reaches its maximum value is used to perform the data fitting. The part where the temperature begins to drop is not considered in the analysis since at this point convection dominates the heat transfer phenomena and the cells are no longer a threat. The energy balance, the assumptions, the physical properties of the cell and number of fitting terms used to obtain the heat generation, see Figs. 2c, 2d, are the same as those reported in Part I. The regions defined by the voltage profile are identified in the temperature and heat generation profiles based on the extent of overcharge. The contrast allows to directly correlate the electrochemical and thermal results. It is important to note how the transition in voltage, indicated by the markers in Fig. 2, corresponds to the inflection points in the heat generation curves, Figs. 2c, 2d. Although cell temperature and temperature rise do not show the aging effect in a clear way, the overcharge heat generation profiles do. When comparing the heat generated by the cells aged in normal VW with that of the aged cell in reduced VW, Figs. 2c, 2d, no significant difference in trend or magnitude is detected. The heat generated during overcharge is still relatively high compared to that generated during the normal charge and discharge process. To put it in perspective, the maximum heat generated during the discharge on the aging test, regardless of the VW, does not exceed 250 mW, as shown in Fig. 6 of Part I. During overcharge, aged cells reach a value of 5 W in ∼8 min while the fresh cell generates 5 W in ∼12.5 min. The higher rate of heat generation in aged cells is a combination of lack of intercalation sites that leads to the formation of lithium dendrites and the occurrence of side reactions leading to an increase in internal resistance. ¹³ It is worth noting that the cells aged to 10% capacity loss generated the most heat compared to the fresh cells and the heat generated falls as the cell ages as observed for the cells aged to 15% and 20% capacity loss. This may be attributed to the loss of capacity with cycle life which leads to lesser energy content in the cells and less temperatures generated under the overcharge condition. However, the temperatures are still higher than that obtained for the fresh cells and this may be due to the degradation of the electrolyte and electrodes with aging and the formation of gaseous side products.

At the end of the overcharge test, the cells undergo a destructive physical analysis as shown in Fig. 3. This operation must be carried out with care since the cells are still in an overcharged condition. And even within the glovebox, minimal contact between the electrodes can cause a thermal ramp up. The inert environment inside the glovebox helps prevent any possible fire. The detailed procedure for opening the cell is described in the appendix A section of Part I. ¹⁹ The graphite and cathode electrodes, and the separator harvested from the overcharged cells are shown in Figs. 4 and 5, respectively. In both figures, the electrodes labeled as 0% CF correspond to the overcharged fresh cell (F₀₀OV₁). The left and right columns correspond to the overcharged cells from the normal and reduced VW electrodes, respectively.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The DPA of the overcharged fresh cell showed that the electrolyte, which is generally translucent, was brown in color. This discoloration may be due to the degradation induced by the overcharge test, (see Figs. 3a, 3b). The electrochemical stability of the electrolyte is compromised by the high voltage. ²⁴ In the aged cells, the ceramic coating on the separator that faces the cathodes displayed a characteristic brownish-yellow coloration, see Fig. 5, which may be attributed to the degradation of the electrolyte that impregnates the separator. ¹⁹ A second feature found during the DPA in the overcharged fresh cells is the bubbling of the electrolyte, see Fig. 3a. This could be due to the escape of gases from the gas pockets in the jelly roll, and it is also highly likely to be the reaction of lithium plating with the small amount of oxygen present inside the glovebox. Excess electrolyte and electrolyte bubbling were not observed with the aged cells due the decomposition and degradation that occurs with the electrolyte during the cycling process.

Disassembly of the cell header is a key part of the destructive analysis of the overcharged cells in order to understand the influence of the built-in protection mechanisms and confirm the internal electrical disconnection caused by CID activation. Figures 3c, 3d show the cell header after the external seal has been removed. In a fresh cell, the physical connection of the inner and outer CID disks is visible in the center of the assembly as shown in Fig. 3c, where they are physically connected in one spot providing the electrical connection from the positive electrode to the cell terminal. During overcharge, the internal pressure exceeds the pressure activation limit and physically and irreversibly disconnects the cell, as shown in the Fig. 3d. Unlike the fresh cell, the two discs of the CID are electrically disconnected due to the physical movement of the top disc caused by the increased internal gas pressure.

In the overcharged fresh cells, uniform lithiation is observed on both sides of the graphite electrode, see Fig. 4. Minimal degradation is found in the electrode areas closest to the center of the cell on one of the sides of the electrode. The gold color is a characteristic feature of the fully lithiated graphite, LiC₆. ^25,26 The aging test creates uneven lithiated areas, especially in the center of the electrode. When the cell is overcharged, the non-uniformities exacerbate and lead to the creation of the darker areas in the 20% CF cells for both VWs. The dark area is graphite that was not fully lithiated. The reason behind the non-uniform lithiation is induced by the cathode degradation as shown in Fig. 5 and the transfer of the ceramic coating from the separator to the cathode or vice versa that prevents that area of the cathode from being electrochemically active. The charring observed in the separator facing the anode material indicates that there is internal cell heating that occurs even though the CID activates and protects the cells from a catastrophic failure. The blackish color on the separator is observed on the overcharged cells but not in the cells that have undergone only capacity fade aging. The two common features in all the overcharged negative electrodes are the detachment of the graphite active materials near to the electrode edges and the presence of scattered dark spots in the center of the electrode, as shown in the 10% CF cell aged in the normal VW.

The cathode electrodes obtained from the overcharged cells are shown in Fig. 5. The two images in the first row corresponds to the overcharged fresh cell. The degradation found on them is directly attributed to the overcharge test since the cell is a fresh cell. Detachment of the cathode active material from the electrode as well as detachment of the alumina (Al₂O₃) ceramic coating from the separator are the two main features found on the electrodes for the aged cells subjected to the overcharge test. In all cases, there was either cathode active material stuck on the separator or the ceramic coating from the separator stuck on the cathode electrode. The extent of degradation induced by the overcharge test was different for the cells aged under the different voltage windows. The cathode electrodes aged in the normal VW exhibit more degradation than the ones aged in the reduced VW one, no matter the CF percentage. This may be due to the higher level of degradation observed with the cells cycled in the normal VW which is also reflected in the faster degradation of capacity (Part I). ¹⁹ In both cases, the extreme delithiation and more specifically the high voltage (E_Cell > 4.30 V) can lead to the dissolution of transition metals (e.g., nickel and cobalt) and its migration to anode which has been reported in Part I of this study. ^19,27 The higher voltages lead to a destabilization of the cathode, a change in the structure of the cathode active material and oxygen release. ²⁸ The apparent difference in degradation on the two sides of the same electrode is simply an effect of the jelly roll winding.

Representative micrographs of overcharged graphite and cathode electrodes are shown in Fig. 6, respectively. The composition for each of the samples, obtained using EDS, is tabulated in Tables I and II.

Zoom In Zoom Out Reset image size

In Fig. 6, the graphite particles (dark background) exhibit some deposits on the surface. Based on the composition of the electrodes and the information gathered from the Fig. 4, it can be concluded that the deposits are from the electrolyte salts and lithium. It is well known that lithium plating will occur during the overcharge test since the high cell voltage drives the anode potential to a negative value. ^29,30 Lithium plating can occur as amorphous or dendritic forms, but gets quickly converted to the oxide form, Li₂O on exposure to air which is reflected as a higher oxygen percentage in the EDS analysis. ³¹ The EDS indicates that the components found on the overcharged graphite electrodes, for both VW in Table I, shows a significant increase in the percentage of oxygen, phosphorous and fluorine and a reduction of the carbon percentage due to the deposits, when compared to that found in a fresh cell. Overall, there is no significant difference in the amount of oxygen present between the electrodes aged in both VWs indicating that the quantity of lithium deposited is similar irrespective of the voltage window used for aging.

In Fig. 6, the overcharged cathode samples (columns 3 and 4) with the representative spherical agglomerated particles (diameter ≈ 10–12 μm) are shown. The sample from the fresh cell that was overcharged, F₀₀OV₁, shows the inner layers of the electrode after the active material got detached due to the extreme delithiation. The dark background corresponds to the aluminum current collector and the small particles are crumbled agglomerates. Overcharge magnifies the stress that was induced by cycling leading to cracking of the active material particles. For both VWs, the greater the CF, the greater the degradation observed. The residues of the alumina ceramic coating attached to the sample increases the aluminum and oxygen percentages. According to these results, one can conclude that the overcharge is a cathode dominated phenomena.

The protocol used to short circuit the cells is exemplified in Fig. 7a, corresponding to the cell N₁₅EX₄. The cells are fully charged at the beginning of the external short test. The voltage dropped from 4.2 V to approximately 100 mV in a few seconds (Fig. 7b). The cell voltage and temperature remained constant for almost three hours. The end of the test is marked by the drop in voltage and temperature. The current flowing through the external load was recorded at 1 kHz rate in order to understand the initial current spike when the short circuit load is applied (Fig. 7c). The maximum current peak was 70 A and occurred in the aged R₂₀EX₂ cell and the lowest peak current was 61 Amps and occurred in cell N₂₀EX₂. The peak current values for the rest of the fresh and aged cells were between these two values and no trend was observed with the capacity fade of the cell. Figure 7d shows the cell temperature during the external short test. The temperature recorded ranged from 63 °C for cell R₂₀EX₂ to 70 °C for R₁₅EX₄ cell. The presence of the PTC internal to the cell in the header area protects the cell from external short circuits and reduces the temperature achieved by the externally shorted cells. The cells could be cycled after undergoing the short circuit test. No significant trend was observed in the peak current or temperatures recorded. There did not appear to be a relationship between the cell's response to the external short test and the percentage of capacity fade.

Zoom In Zoom Out Reset image size

Externally shorted cells were subjected to post-mortem analysis similar to that performed on overcharged cells. The fresh cells subjected to the external short test showed severe degradation with meltdown like features on the anode electrode (Figs. 7e and 7f). This was observed in spite of the PTC activation in the fresh cells. However, no further degradation was found in the electrodes from the externally shorted cells with respect to the electrodes from the only aged cells reported in Part I. ¹⁹ For that reason, electrodes pictures are not included here.

To characterize the overcharge behavior of the cells configured into modules, it was necessary to age the modules before subjecting them to the overcharge test. The aging test results for the modules and their banks are shown in Fig. 8. In Fig. 8a the discharge capacity trend for the two modules subjected to the aging test is shown as well as a schematic of the module design used. Like single cells, the capacity of the module aged under normal VW decreased faster than that of the aged one in reduced VW. It takes 100 (normal VW) and 135 (reduced VW) cycles to reduce the modules' capacity by 20% with respect to the initial capacity. In the case of modules, a lower number of cycles were obtained since modules were aged at a 1C-rate which was twice that used for the single cells.

Zoom In Zoom Out Reset image size

Due to the lack of a cell balancing protocol, it was observed that the cell banks in the module that underwent the charge and discharge cycle life aging under the normal VW, deviated in voltage which was more obvious during the discharge step, see Fig. 8d. The imbalance resulted in bank 9 going below 0 V since the voltage was controlled at the module level and not at the individual bank level. Previous work has shown that an overdischarge condition degrades the electrolyte and release gases that makes the cells swell. In some cases, internal shorting can occur due to a combination of the dissolution of the copper current collector during overdischarge and redeposition on the electrodes and separator during subsequent charge. Other observations included lithium deposits on the anode electrodes and significant electrode degradation. ^32,33 All these degradation mechanisms can cause internal temperature rise and temperatures of up to 164.2 °C was recorded on the overdischarged bank, Fig. 8b. It is important to note that the bank temperature rise occurred only when the bank went below 0.5 V. Before that cycle the temperature was almost the same than the rest of the banks. This overdischarge and high temperature condition caused the test equipment to stop the test. The test was restarted after electrically disconnecting bank 9 which shows the bank temperatures slowly rising again and the test was stopped at that point and bank 9 was electrically disconnected for the rest of the aging test. The extreme overdischarge condition was not observed in the module aged under the reduced VW, see Fig. 8f and the bank temperatures did not rise above the 70 °C. Only a slight overdischarge was observed at the end of the cycle life aging test of up to 20% capacity loss, but the bank voltage never went below 2 V. A very slight overcharge was found in some of the banks from the module aged under the normal VW, Fig. 8c, but not in the reduced one, Fig. 8e. In general, aging causes voltage deviation between banks but the use of the reduced VW may help in lessening the deviation.

The module resistance was also measured in a similar way to that of the single cells, i.e., with a 1.5 C pulse for 100 ms except that it was measured every 25 cycles. The internal resistance of the module initially decreased in the first 25 cycles and then grows showing the well-known tendency of cells to increase in internal resistance with aging.

The electrochemical and thermal results of the overcharge induced test on the modules are shown in Figs. 9 and 10. The voltage of the modules, Fig. 9, shows that the fresh module takes longer to be overcharged compared to the aged modules. Irrespective of the voltage window the module was cycled for cycle life aging, in an overcharge condition both aged modules failed almost at the same time and in a similar manner. During the overcharge test, the fresh module experienced a complete thermal runaway; and the aged modules experienced benign CID activations around 80 V. Thus, modules with fresh cells experience thermal runaway, while those with aged cells do not.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The bank voltages and temperatures for each of the modules is shown in Fig. 10. During the overcharge of the fresh module, banks 2 through 8 underwent thermal runaway and caught fire as shown in Fig. 11b. Only bank 1 did not go into thermal runaway but was still overcharged up to 5.15 V. The voltage at which the CID activation in the cells occurred seemed to be consistently close to 5.15 V. This peak voltage on the banks corresponds to point D in Fig. 2. Upon activation of the CID, the voltage on banks 2 through 8 spiked above 15.6 V. Regarding the temperature of the banks, Fig. 10e, the maximum temperature recorded was 2299 °C in bank 2 whereas the lowest temperature was 352 °C in bank 1. The high temperatures recorded were due to the fire caused by the complete thermal runaway of all the cells as shown in Fig. 11b and the readings above 2000 °C on module 2 is due to the temperature reading exceeding the thermocouple and data collection capability.

Zoom In Zoom Out Reset image size

During the overcharge of the module aged under normal VW, none of the banks went into thermal runaway. The maximum voltage that the banks reached, before the CID activation, was 5.20 V. Analysis of the thermal response, Fig. 10e, showed that the maximum temperature was 88.0 °C and it was observed in bank 4. The rest of the banks reached temperatures no higher than 60 °C.

During the overcharge of the module aged under reduced VW, none of the banks went into thermal runaway. Similar to the module aged under the normal voltage window, the maximum voltage recorded on the banks before the CID activation, was 5.20 V. Upon activation of the CID, the voltage on the banks reaches the compliance voltage of the equipment but the maximum voltage attained by each bank varied, as the voltage limit set on the equipment is based on the total module voltage. Analysis of the thermal response, Fig. 10f, showed that the maximum temperature recorded was 56.0 °C observed in bank 9. The rest of the banks reached temperatures no higher than 60 °C.

The overcharge test results showed that the two aged modules did not go into thermal runaway but the fresh one did. No visible degradation was found in the aged modules at the end of the overcharge test, see Figs. 11c, 11d. This indicates that off-nominal conditions such as overcharge do not cause a catastrophic effect on the aged modules as they do on the fresh cells. This could be attributed to the reduced energy content in the modules that reduces the temperatures achieved keeping the off-nominal condition benign. The early activation of the CIDs in the aged cells could have also led to a faster shutdown before the temperatures got too hot to cause a catastrophic failure. It should be kept in mind that these modules were aged to more than 20% cf One cannot conclude that aging to any level of capacity fade will result in the same benign behavior.

The drive cycle depicting the demand from a vehicle while driving in a highway at an ambient temperature of T = 10 °C is shown in Fig. 12a. Two sets of three drive cycles were applied emulating driving a car for almost an hour followed by a full charge, as shown in Fig. 12b. The full cycling profile was repeated for a whole week, see Fig. 12c. After that the cell capacity and internal resistance were measured using the conditioning test protocol reported in Part I ¹⁹ and results are reported in Figs. 12d and 12e. These periodic capacity checks were used to determine when greater than 20% capacity loss was attained by the cells. Voltage response to the variable current showed a strong dependence on the operating temperature with much higher voltage changes in the cells aged at low temperatures. The discharge capacity, Fig. 12c, exhibited the characteristic aging decay similar to the conventional CC discharge profile but the extent of decay varied with the environmental temperature that the cells were tested in. Cells that underwent the drive cycle profiles at a low temperature, 10 °C, reached 20% CF in an average of 17 weeks. Low temperature increases the internal resistance of the cell rapidly due to the higher viscosity of the electrolyte at the lower temperatures which then promotes lithium plating and a fast rate of capacity reduction. ³⁴ As the ambient temperature was increased to 25 °C, the cells could be cycled for an average of 29 weeks before 20% CF was obtained. As the ambient temperature was further increased to 40 °C, the performance and cycle life of the cells slightly decreased to an average of 22 weeks to obtain 20% cf By increasing the operating temperature, the transport and kinetic processes are improved, but the thickening of the SEI layer and the degradation of the organic solvents of the electrolyte increase significantly leading to a greater loss in performance. ^35–37 In the case of internal resistance, its value increased in all cells as they aged. However, no evident trend was found regarding the effect of the environmental temperature.

Zoom In Zoom Out Reset image size

Three cells aged with the drive cycle, one from each environmental temperature set, were subjected to an overcharge test. The voltage and temperature response are shown in Figs. 13a and 13b. CID activation occurred first on the cell aged at 10 °C followed by the 25 °C cell around Q_OV = 0.5 Ah (114.7% SOC). The cell aged at 40 °C could be overcharged to Q_OV = 0.8 Ah (123.5% SOC) before the CID activation. This indicates that despite cycling them to the same extent of 20% CF, the degradation caused in the cells is different depending on the environmental temperature. A significant difference in response voltage is the peak voltage reached by the cell before failing. The maximum voltage achieved, decreased as the operating temperature increased. The maximum temperature of the cell, during overcharge, was also affected by the different environmental temperature conditions. The cell aged at 10 °C reached a maximum temperature of 54 °C and then decreased. On the other hand, the cell aged at 40 °C reached a maximum temperature of 98 °C.

Zoom In Zoom Out Reset image size

The results for the cells aged using the drive cycle protocols and then externally shorted with a 50 mΩ external load are shown in Figs. 13c and 13d. The maximum current recorded was 64 A, 64 A, and 58 A, for cells aged at environmental temperatures of 10, 25, and 40 °C, respectively. Due to the activation of the PTC present in the cell headers, the cell voltage for all the cells fell from 4.2 to 0.2 V in less than a minute. The low voltage due to the activated PTC is observed until most of the capacity from the cell is removed. The maximum temperature in all three cells was observed for approximately two hours. The maximum temperature recorded varied with the environmental conditions, for the cells aged at 10 °C, a maximum temperature of 68 °C was observed while the cell aged at 25 °C reached a maximum temperature of 74 °C. The end of the test was marked by a small voltage spike in the cell and a drop in temperature, Fig. 13c. This may be due to the resetting of the PTC which then drains the cell completely causing the complete drop in voltage.

The post-mortem analysis of the cells aged with the drive cycle is shown in Fig. 14. A major observation was the fact that in all three cases, the aged electrodes displayed a lack of lithiation in the center one-third of the cell in the axial direction (Fig. 13b). The effect of aging the cells with the drive cycle protocol manifests as a non-uniform lithiation in the center of the electrode. This degradation appears as small dots scattered throughout the anode, Fig. 14b with a more visible evidence of lithium plating on the edges, Figs. 14a–14c. The overcharge test on the aged cells aggravates the degradation in the regions already degraded by the aging test, Figs. 14d–14f. The silver color on the overcharged electrode aged at 10 °C is lithium plated throughout the whole electrode, Fig. 14d. Lithium deposition occurred in a more localized way in the 25 °C overcharged cell where it was found in the center and the edges of the electrode, Fig. 14e. The excessive presence of lithium dendrites could have internally shorted the cell and caused the dark charred appearance in the separator. In the case of the 40 °C overcharged cell, the anode active material had detached from the electrode, and charring and residuals of lithium plating was found on the separator, Fig. 14f. Overcharge worsens the detachment of the cathode material and ceramic coating from the separator as shown in Figs. 14g and 14h. In the 25 °C overcharged cell, the shorting caused in the anode side transferred to the cathode side causing the presence of small spots on the cathode electrode where all the active material had delaminated and exposed the aluminum current collector Fig. 14h. The fact that the electrodes exhibit localized degradation as they age with the drive cycle can be interpreted as electrochemical interference. By charging and discharging the cell with a frequently fluctuating current, certain regions of the electrode especially near the center are neither fully lithiated nor fully delithiated compared to the rest of the electrode. This caused the variation in the extent of lithiation observed in the anode electrodes. Post-mortem analysis of the externally shorted cells did not show any additional degradation to the one caused by the aging process.

Zoom In Zoom Out Reset image size

The external short circuit tests were conducted on the cells indicated in Part I as well in the cells aged with the drive cycle. ¹⁹ Despite the level of degradation or the operating temperature, the results did not show any significant difference between the cells that were fresh and those that underwent just aging as well as the aged cells that underwent the external short circuit tests hence they are not discussed in detail here. This may be due to the performance of the PTC in a similar manner irrespective of the aging process.