Volume Deformation of Large-Format Lithium Ion Batteries under Different Degradation Paths

Lithium ion batteries experience volume deformation in service, leading to a large internal stress in modules and potential safety issues. Therefore, understanding the mechanism of volume deformation of a lithium ion battery is critical to ensuring the long-term safety of electric vehicles. In this work, the irreversible and reversible deformation of a large-format lithium ion battery under four degradation paths, including cycling at − 5°C/1 C, 55°C/1 C and 25°C/4 C, and storage at 55°C/100% state of charge, are investigated using laser scanning. The reversible deformation decreases while the irreversible deformation increases as batteries age, following a linear trend with the state of health. The mechanism behind irreversible deformation is investigated using incremental capacity analysisandscanningelectronmicroscopy.Theirreversibledeformationofthebatterycycledat25°C/4Candstoredat55°Cbecomesextremelylargebelow80%stateofhealth,mainlybecauseoftheadditionaldepositlayersontheanodeandincreasedgasproduction,respectively.Mechanicalcalculationsshowthehugestressintheagedmodules.Properspacersbetweenbatteriesaresuggestedtoreducesuchdamage.Thisstudyisvaluableforunderstandingthemechanicalsafetyofbatterymodules.©TheAuthor(s)2019.PublishedbyECS.ThisisanopenaccessarticledistributedunderthetermsoftheCreativeCommons This study analyzes the mechanisms of reversible and irreversible deformation of large-format lithium ion batteries under different degradation paths. A commercial large-format pouch battery is used for the degradation test with four paths including cycling at − 5°C/1 C, 55°C/1 C and 25°C/4 C, and storage at 55°C/100% SOC. The degra- dation mechanism is identiﬁed through ICA and scanning electron microscopy (SEM) characterization. The reversible and irreversible deformations are analyzed based on the identiﬁed degradation mecha-nism.Mechanicalcalculationsareperformedtoshowthestressinagedmodules,andpossiblecountermeasuresaresuggested.Thisworkwillgiveacomprehensiveunderstandingofthedeformationmechanismofagedcommercialbatteriesandprovideusefulguidanceontheme-chanicaldesignforbatterymodules,forexample,ontheselectionofspacers.

Lithium ion batteries have become the most reliable energy storage media 1-3 owing to their various advantages such as high energy density, low self-discharge rates, and wide operating temperature range. 4 These batteries need to be connected in series and parallel to form modules that meet the specific power or energy requirements; 1 and restraints should be added to the battery modules to ensure mechanical integrity and sufficient electrical contact. 5 However, as these batteries age, they undergo significant deformation; large volumes of gas may also be produced, 6 resulting in increased stresses, 7 increased impedance, 8 breakage of aluminum packaging or hard cases, and even safety issues. 9 Volume deformation of a lithium ion battery may be caused by lithium intercalation and de-intercalation during cycling, 10 thermal deformation, 11 irreversible deformation of the anode or cathode, 12,13 and gas formation. 4,6 Among these, deformation caused by lithium (de-)intercalation and heat is regarded as reversible, while deformation caused by other reasons is considered irreversible. Reversible deformation can be defined as the change in the thickness of the battery during cycling, 14 while irreversible deformation can be defined as an increase in the thickness of an aged battery compared with its new counterpart at a certain state of charge (SOC). 15 The mechanism of reversible deformation of batteries has been explored through experiments and models. Atomic force microscopy (AFM) and X-ray diffraction (XRD) have been used to monitor the volume deformation of particles on the electrode and establish various look-up tables for volume deformation under different voltages. [16][17][18][19] Different theoretical model 11,[20][21][22] and phenomenological models 5,23,24 have been proposed to depict the deformation and force curves of batteries based on the measured look-up tables at the particle level or cell level. However, previous studies have primarily focused on the reversible deformation of new batteries. The reversible deformation of aged batteries may change significantly due to electrolyte decomposition, electrode particle cracking, or other side reactions. 6 For a 2.28 Ah battery cycled at 25°C/1 C, the reversible deformation at a certain location of first increased and then decreased with the state of health (SOH). 14 The increase was ascribed to higher utilization at that location during the early degradation stage, which may increase the local capacity. While the decrease was due to loss of lithium inventory and resulting less lithium ions intercalated into the graphite sheets. 14 Researchers have also conducted various experiments to reveal the mechanism of the irreversible deformation. The 2.28 Ah batteries cycled at 25°C/1 C, 40°C/1 C, and 25°C/0.5 C showed that the irreversible deformation increased with SOH, following a linear trend. 14 The irreversible deformation for the battery cycled at 25°C/1 C was mainly ascribed to lithium plating on the anode surface. 14 The 15.7 Ah batteries cycled at 25°C and different C rates exhibited thicker deposit layers on the anode for higher C rates. 12 A lithium layer of 5 μm on the anode was observed for a 2.5 Ah battery cycled at −22°C/1 C. 25 The size of graphite particles increased while the d-spacing remained unchanged for the 1 Ah batteries cycled at 30°C and different C rates. 26 The increased irreversible deformation may result in a separation between the electrode layers in a free state 27 or severe stress rise in a constrained state. 28 Therefore, irreversible deformation in turn has a great impact on battery performance. 8,29 However, previous studies mainly focused on small batteries and one kind of degradation path. They were insufficient to reflect the different conditions encountered by commercial EV batteries in service, such as winter or summer driving, summer storage, and fast charging. 1 Moreover, the irreversible deformation growth was given on electrode or particle level rather than cell level, which made it difficult to evaluate possible hazards in real modules. Therefore, the reversible and irreversible deformation of large-format commercial lithium ion batteries degraded under different paths and their resulting influence on real modules should be further explored.
The deformation of aged batteries is strongly related to the degradation mechanism. 12,14,25,30 Battery degradation can originate from loss of lithium inventory (LLI), loss of active material (LAM), and increased internal resistance (IIR). 31-33 LLI can be induced by repeated cracking and thickening of SEI films, lithium plating under high rate cycling, or low temperature cycling. 4,12 LAM on the cathode can result from side reaction between cathode and electrolyte, lattice oxygen release, transition metal dissolution, chemo-mechanical breakdown, and phase transformation. 34 LAM on the anode can arise from cracking of anode particles or exfoliation of anode material from the current collector due to stress discontinuity. 4 IIR can be brought by electrolyte consumption and the growth of the deposit layers, such as SEI films on the anode and cathode electrolyte interface (CEI) films on the cathode. 4,6,35 The degradation mechanism of lithium ion batteries can be analyzed noninvasively through the cyclic voltammetry (CV), incremental capacity analysis (ICA), and differential voltage (DV) methods. [31][32][33]36 This study analyzes the mechanisms of reversible and irreversible deformation of large-format lithium ion batteries under different degradation paths. A commercial large-format pouch battery is used for the degradation test with four paths including cycling at −5°C/1 C, 55°C/1 C and 25°C/4 C, and storage at 55°C/100% SOC. The degradation mechanism is identified through ICA and scanning electron microscopy (SEM) characterization. The reversible and irreversible deformations are analyzed based on the identified degradation mechanism. Mechanical calculations are performed to show the stress in aged modules, and possible countermeasures are suggested. This work will give a comprehensive understanding of the deformation mechanism of aged commercial batteries and provide useful guidance on the mechanical design for battery modules, for example, on the selection of spacers.

Experimental
Twelve commercial large-format pouch batteries with a nominal capacity of 24 Ah were used in this study. Each battery has a LiNi 1/3 Co 1/3 Mn 1/3 O 2 (NCM111) cathode and a graphite anode; the dimensions of each battery are 200 mm × 150 mm × 7.7 mm.
Capacity test and degradation test.-Capacity test.-A capacity test was carried out to evaluate the standard capacity of new and aged batteries. This test was conducted at 25°C (±0.5°C) in a thermal chamber using a battery cycler from Neware Corporation. All batteries were rested for more than 6 h to ensure thermal equilibrium prior to the capacity test. A thermal couple with an accuracy of 0.5°C was attached to the positive tab to detect the temperature during the capacity test and subsequent tests, because the temperature of the positive tab is always higher than the other parts based on many studies. [37][38][39] The voltage and current profile of the capacity test is presented in Fig. 1. The batteries were first charged with constant current (CC) at 1/3 C up to 4.2 V and then constant voltage (CV) at 4.2 V until the current fell below C/20. After a rest period of 1 h, the test batteries were discharged with constant current (CC) at 1/3 C to the cutoff voltage of 2.7 V. The batteries were cycled between 2.7 and 4.2 V for three times, and the discharge capacity in the third cycle was regarded as the standard capacity. The SOH of each battery was calculated as the ratio of the standard to nominal capacities.
Degradation test.-The degradation test was performed on 4 groups in a thermal chamber according to Table I. Group 1 to 4 corresponded to cycling at −5°C/1 C, 55°C/1 C, 25°C/4 C, and storage at 55°C/100% SOC, respectively. The charge process followed the CCCV protocol up to 4.2 V, while the discharge process followed the CC protocol down to 2.7 V. The rest period between the charge and discharge processes was set to 5 min. The degradation test was stopped after certain number of cycles or storage days, as indicated in Table I. Then, the temperature was adjusted to 25°C. The batteries were rested for 6 h after the temperature switch to ensure thermal equilibrium of the whole experimental set-up.

Measurement of battery volume deformation.-
The entire experimental set-up for battery volume deformation measurement (Fig. 2) was placed in a thermal chamber to ensure an environmental temperature of 25°C (±0.5°C).
A commercial laser camera from Gocator with an accuracy of 5 μm was used. The laser camera was attached to a linear axis parallel to the y axis of the defined coordinates. A laser line emitted from the laser camera could scan along the linear axis with a width of 100 mm. The laser line recorded the z-coordinate every 0.1 mm, resulting in 1000 points in total. The tested battery was placed on a flat base plane and connected to the cycler. The long edge of battery was parallel to the y axis whereas the short edge was parallel to the x axis. The base plane was set under the tested battery as z = 0 mm prior to the test. Therefore, the z-coordinate detected by the camera was regarded as the local thickness of the tested battery. The deformation measurement won't be affected by the thermal deformation of the test rig since the environmental temperature is constant and the emitted laser line generate no heat.
The deformation measurement could be divided into the following two groups: Reversible deformation.-Reversible deformation was obtained by focusing on the battery surface center (Fig. 2b) during the capacity test at a specific SOH. The laser line was emitted and recorded with a frequency of 1 Hz. Such frequency was enough to obtain an accurate result for a cycling rate of 1/3 C. The reversible deformation of the 1000 points was averaged to eliminate the stochastic error at certain locations and represent the real reversible deformation of the entire battery. Due to the large surface of the battery, limit of the memory space and speed of the linear axis, we use the reversible deformation at the center line to represent the whole battery. But the reversible deformation on two other locations 80 mm away from the center line showed almost the same magnitude, which is presented in the Appendix.
Irreversible deformation.-The irreversible deformation was defined as the increase in thickness of the aged batteries compared with the new batteries at 100% SOC. The deformation measurements started after the batteries were charged to 100%SOC using CCCV protocol and rested for 6 h 25°C. The procedure was to ensure that the batteries were thermally uniform. The irreversible deformation was obtained by scanning the entire surface of the batteries. A black line was drawn to separate the battery surface into two identical parts ( Fig. 2c) because the laser line was shorter than the short edge of the battery. The two parts would be scanned sequentially to cover the  SEM characterization.-SEM characterization were conducted to analyze the surface morphology of the cathode and anode before and after degradation. The batteries were discharged with 1/3 C current to 2.7 V (0% SOC) before and after the degradation test, and then dismantled in an argon filled glove box (H 2 O and O 2 content < 0.1 ppm). The harvested electrode pieces were rinsed by dimethyl carbonate (DMC) for 2 h to remove the lithium salt, and then dried for 1 h in the glove box. The washing process would not affect the surface morphology of the sample. All the samples were sealed in a container in the glove box and transferred to the SEM vacuum chamber from JSM-IT300. Images with an amplification factor of 500 were captured.

Results and Discussion
Degradation mechanism analysis.- Fig. 3 shows the capacity degradation of the batteries under four paths. The batteries cycled at −5°C/1 C experience the most severe capacity loss, degrading to 77% SOH in only 60 cycles. The batteries cycled at 55°C/1 C preserve 76% SOH after 1260 cycles, whereas the batteries cycled at 25°C/4 C main-tain 79% SOH after 1800 cycles. The batteries stored at 55°C/100% SOC maintain 80% SOH after 180 days.
The incremental capacity (IC) curve can be derived by applying the probability function method (PDF) to the constant current cycling data. 31,36,[40][41][42] The voltage current data is divided into multiple small intervals. The dQ/dV can be calculated using Eq. 1.
Where n is the number of data points in the corresponding interval, I is the constant current, f is the sampling frequency of the cycler, and V is the width of the corresponding interval. The voltage interval width is 5 mV. This method is much easier to implement and more robust compared to differentiating the voltage directly.
IC curves of the batteries under the four degradation paths obtained from C/3 discharging are compared in Fig. 4. The curve for new batteries is also presented as reference. The integral of the IC curves represents capacity of the batteries, while the peaks of the IC curves correspond to phase transformation of the electrode material. The left shift of the peak represents an increase in internal resistance.
As shown in Fig. 4a, there were four peaks denoted as I, II, III, and IV for new batteries corresponding to the phase transformation processes of the electrode material. To be specific, Peak I at 3.4 V and Peak III at 3.7 V were related to the phase transformation processes of the graphite anode, whereas Peak IV at 4.1 V corresponds to the phase transformation process of the NCM111 cathode. 31,33 Peak II at 3.59 V can be ascribed to a combination of both the graphite anode and the NCM111 cathode. 33 For batteries cycled at −5°C/1 C (Fig. 4a), Peak II drops much more significantly than other peaks. This may be related to the severe LLI due to lithium plating. Peak I and III degrade slightly, indicating a little LAM on the graphite anode. Peak IV remains constant, revealing that LAM does not occur on the NCM111 cathode. The positions of  each peak do not change above 85% SOH, indicating that there is no obvious increase in resistance.
For batteries cycled at 55°C/1 C (Fig. 4b), Peak IV exhibits a severe decline, indicating significant LAM on the NCM111 cathode. The LAM on the cathode substantially contributes to degradation under a high-temperature cycle according to Ref. 43. Peak II and Peak III drop considerably, whereas Peak I does not change much. This may be a result of LLI and no LAM on the graphite anode. All the peaks move to the left side, indicating drastic increase in resistance.
For batteries cycled at 25°C/4 C (Fig. 4c), Peak II and Peak III drop significantly, showing severe LLI. Peak I and Peak IV do not vary significantly, indicating that no obvious LAM on both the graphite anode and the NCM111 cathode is observed. The positions of all the peaks remain unchanged, indicating that there was no increase in resistance.
For batteries stored at 55°C/100% SOC (Fig. 4d), Peak I, II, and III all degrade, which is a result of LAM on the graphite anode. Peak II drops more than other peaks, indicating that there is also LLI in these batteries. Peak IV remains almost the same, showing that there is no LAM on the NCM111 cathode. The peaks do not exhibit any shift, indicating that there was not much increase in resistance.
SEM images for the anode and cathode under a magnification factor of 500 are presented Fig. 5 and Fig. 6, respectively.
As shown in Fig. 5, fresh batteries and aged batteries show no differences in the cathode morphology under a magnification factor of 500, indicating the degradation of the batteries didn't result from deposit layers on the cathode. This is consistent with the aforementioned degradation mechanism derived from IC curves except for batteries cycled at 55°C/1C, whose degradation may originate from CEI growth or lattice distortion and can't be distinguished by SEM. Moreover, CEI growth or lattice distortion won't lead to much deposit layers.
In Fig. 6, the anode morphology of the fresh battery, battery cycled at 55°C/1C, battery stored at 55°C/100%SOC looked similar, showing no deposit layers under a magnification factor of 500. However, mossy deposit layers were observed on the anode from the battery cycled at −5°C/1C in Fig. 6b. We interpreted this as lithium plating induced by low temperature cycling. 30 Additional deposit layers covered a large area of the anode from the battery cycled at 25°C/4C in Fig. 6d. This may be a result of SEI decomposition under high rate cycling and resulting more electrolyte decomposition. 44 Such layers would lead to LLI as derived from IC curves.
Reversible deformation.-The reversible deformation and voltage profile within one 1/3 C charge and discharge cycle are shown in Fig. 7. The reversible deformation first increases during the constant current charge process and starts to decline slowly during the constant voltage period. Such relaxation continues in the rest period with a smooth decline of the reversible deformation. It further decreases during the discharge process and recovers to the initial state at the end of the rest period. The deformation curve in one full cycle is asymmetrical owing to the different phase transformation path of intercalation and de-intercalation. 45,46 The deformation relaxation at the end of the charge and discharge process can be induced by both lithium concentration relaxation 47 and thermal effects. 11 The reversible deformation is extracted from the beginning of discharging to the end of the rest period to avoid the effect of such relaxation, marked as point A and B in Fig. 7. The battery temperature change during the capacity test is very small, as indicated in the infrared figures in the appendix. The maximum temperature rise is below 1.6°C for charging and below 1.1°C for discharging for all the batteries. The 2C pulse excitation test following Ref. 11 showed that batteries used in this study expansion 3.97 μm with a temperature rise of 1.0°C. Thus, thermal expansion of the cell material is negligible in this study and the detected reversible deformation can be regarded as the pure intercalation induced deformation.  As presented in Fig. 9, new batteries, batteries cycled at −5°C/1 C and cycled at 55°C/1 C exhibit a flat surface, while batteries stored at 55°C/100% SOC and batteries cycled at 25°C/4 C become quite non-uniform. For batteries cycled at 25°C/4 C, the thickness of the edge part is larger than that of the middle part. Moreover, the entire battery becomes harder, as revealed through an appearance inspection. This was because of the larger current distribution on the edge part under higher rates, which resulted in thicker deposit layers. 48 For batteries stored at 55°C/100% SOC, the large irreversible deformation in the middle is originated from gas production due to electrolyte decomposition. 49 The average thickness is obtained based on the thickness of the entire surface. The irreversible deformation under different degradation paths is shown in Fig. 10.
The irreversible deformation of batteries gradually increases as batteries age, which is similar to the result of Ref. 14. Most points exhibit a linear relationship between SOH and irreversible deformation, with a RMSE of 176 μm and an R square of 0.81. Two exceptional points at approximately 80% SOH correspond to the batteries stored at 55°C/100% SOC and cycled at 25°C/4 C. The irreversible deformation of the battery cycled at 25°C/4 C becomes extremely large below 80% SOH, mainly because of additional deposit layers on the anode. The battery stored at 55°C below 80% SOH also experiences a significant irreversible deformation increase due to increased gas production. This is consistent with the inhomogeneous result shown in Fig. 9.
The mechanism of the irreversible deformation under different degradation paths are revealed according to the ICA and SEM results presented above. For batteries cycled at −5°C/1 C, increase in irreversible deformation is most likely due to lithium plating, as revealed  by many studies. 1,6 This irreversible deformation increases considerably when the battery degrades to 81% SOH owing to mild gas production, as indicated in Fig. 9b. This is caused by the side reaction between the electrolyte and metallic Li. 6 For batteries cycled at 55°C/1 C, the irreversible deformation is relatively low compared with batteries degraded by other paths, even when the battery is degraded to 76% SOH. One possible reason is that the CEI-growth-induced impedance rise and the LAM on the cathode contributes more to the capacity degradation. 43 However, these inactive layers on the cathode or cracks of NCM111 particles will not result in large irreversible deformation such as that from SEI film growth or deposit layers on the anode. 50 For batteries cycled at 25°C/4 C, an increase in irreversible deformation may result from growth of the SEI and the additional deposit layers (Fig. 6d) induced by the side reaction of the electrolyte with the anode under a high local current density. 12 Such a process can lead to severe LLI as illustrated in Fig. 4c. For batteries stored at 55°C/1 C, gas production dominates the increase in irreversible deformation as revealed by visual observation; this causes a relatively larger increase in their thickness compared to batteries degraded by other paths (Fig. 10). All the batteries are stored at 100% SOC, causing severe thermal decomposition of the electrolyte in this highly active state. 49 This causes LLI, as shown in Fig. 4d.
For the batteries used in this study, the magnitude of irreversible deformation is 1000 μm, whereas the magnitude of reversible defor- mation is 100 μm. The maximum irreversible deformation is 3540 μm, which is equivalent to 45.3% of the original thickness. This can induce a large stress in the battery module with zero displacement boundary conditions. 7,8,22,51 The porosity of the relatively soft separators may further decrease and worsen the degradation process. 27,52 Internal stress calculation.-To evaluate possible hazards induced by the irreversible deformation of batteries and suggest feasible countermeasures, a model is proposed to calculate the stress variation within a module due to battery degradation.
In this study, all the batteries are in free state during the degradation test and experience an increase in irreversible deformation. However, if the batteries were placed in a module with fixed boundary, the stress would grow while the battery thickness remain unchanged. Such phenomenon can be simulated by virtually loading an aged battery until its thickness is restored to the initial value.
First, for simplicity, we consider only one battery in a module without any spacers (Fig. 11a) to represent one element in real modules. All the components of the module are regarded as springs connected in series, each with a spring constant of E i (ɛ)A. A is the area of the battery surface (200 mm × 150 mm), while E i (ɛ) is the elastic modulus in the thickness direction of each component as a function of its strain. The internal forces (denoted as F) and stresses (denoted as The overall displacement of the module (Fig. 11b) is obtained by superimposing the displacement responses of each part, where t i,0 is the initial thickness of each component as measured in Table II. Combining Eqs. 2 and 3, we can obtain We find S 0 = 133.67 μm by using the mechanical properties of the positive and negative electrodes, 53 the separator, 54 and the aluminum plastic film 55 in the literature. The fixed length within the module is assumed to be 7586.33 μm, as indicated in Fig. 11b. Then we ascribe the irreversible deformation of batteries in free state to the anode (Fig. 11c). This is valid for most batteries except for the batteries stored at 55°C, whose irreversible deformation mainly came from gas formation as discussed before. There are two main reasons for this assertion. Firstly, the cathode morphology of aged batteries changed little as presented in Fig. 5, indicating no observable deposit layers on the cathode. Secondly, many studies on degraded batteries through high rate cycling, 44,56 low temperature cycling 25,30 or high temperature cycling 43 has found that cathode thickness contributes little to the overall irreversible deformation. The overall compression displacement in the aged module is S iir +S 0 because of the constant displacement boundary condition (Fig. 11d). We assume that the mechanical properties of each component do not change after batteries degrade. Therefore, the internal force within the aged module with a spacer can be calculated using the following equation: where strain of the anode should include the irreversible deformation as ε an = S an t an,0 + S irr [6] Finally, the internal stress of the aged module is: In this study, polyurethane foams 57 and aluminum alloy plates 58 are considered as two candidate spacers. Fig. 12 shows the stress development in a module with different spacers when the irreversible deformation is 2000 μm. The internal stress in the module can reach 21.21 MPa without spacers between batteries. Such stress is equal to 64.93 tons of force considering a 200 mm × 150 mm surface. The strain of the separator will be 0.33, which will greatly increase the ion transport resistance of the separator. 29 Such internal stress and strain values of the separator vary by less than 0.1% if aluminum alloy plates are used as spacers in the modules, as shown in Fig. 11. In contrast, the internal stress decreases to 11.19 MPa with a strain of 0.064 for the separator by using a 1 mm polyurethane foam plate between the batteries, which alleviates the rise in ion transport resistance. This is because the elastic modulus of the polyurethane foam is considerably smaller than that of a component such as the separator. Therefore, the foam absorbs most of the irreversible deformation. Moreover, the additional foam will not have a great impact on the mass energy density of the module owing to its low density (35-40 kg/m 3 ). 57 The above analysis illustrates the possible hazards to the battery performance and countermeasures for the irreversible deformation of  the aged batteries. The calculation can give some useful guidance for module design. For example, soft spacers like polyurethane foam plates are preferred to aluminum alloy plates.

Conclusions
In this study, the reversible and irreversible deformation of batteries with different degradation paths are measured by laser scanning. The origins of the reversible and irreversible deformation are revealed in detail through the degradation mechanism and deposit composition. The mechanical calculation shows that the stress generated in the module after aging is large without soft spacers. Therefore, some suggestions are given for the module design. The main conclusions in this study are as follows: The degradation mechanism of the battery varies with the different aging paths. The degradation of batteries cycled at −5°C/1 C is mainly caused by lithium plating. For the batteries cycled at 55°C/1 C, the LAM in the positive electrode plays an important role. SEI film growth and gas production dominate the degradation of the batteries stored at 55°C/100% SOC. The degradation of the batteries cycled at 25°C/4 C is mainly due to the LLI caused by the SEI films and additional deposit layers.
The reversible deformation gradually reduces with a declining SOH. This is because the amount of lithium ion transferred in one full cycle is directly related to the SOH despite the different degradation mechanisms.
The irreversible deformation of batteries with different aging paths increases with the decrease in SOH, following a linear trend for most points. Two exceptional points at approximately 80% SOH correspond to batteries stored at 55°C/100% SOC and batteries cycled at 25°C/4 C due to increased gas production and thicker deposit layers, respectively. For batteries cycled at −5°C/1 C, the increase in irreversible deformation is mainly due to lithium plating. For batteries cycled at 55°C/1 C, the irreversible deformation is relatively low because CEI-growth-induced impedance rise and the LAM in the cathode contributed more to capacity degradation than SEI film growth. For batteries cycled at 25°C/4 C, irreversible deformation is ascribed to the SEI film growth and additional deposit layers. For batteries stored at 55°C/1 C, gas production dominates the increase in irreversible deformation.
The irreversible deformation of the aged batteries can be of more than 45% of the full battery thickness, an order of magnitude higher than the reversible deformation. Therefore, this must be considered in the module design. The internal force in an aged module can surge to dozens of tons without soft spacers, leading to a strain of 0.4 for the separator. Then battery performance may degrade quickly. Such hazards can be avoided by introducing polyurethane foam with a suitable thickness or soft blocks as spacers. The stress and strain of the separator can be reduced by an order of magnitude and the possible ion transport resistance rise can be diminished. The future work will focus on implementing in the modules the countermeasures presented in this study by introducing spacers of different thickness and comparing the battery life under a constrained state.   Table AI shows the feasibility of detecting reversible deformation of batteries at the center line.