Influence of Vinylene Carbonate and Fluoroethylene Carbonate on Open Circuit and Floating SoC Calendar Aging of Lithium-Ion Batteries

Abstract

The purpose of this study was to investigate the calendar aging of lithium-ion batteries by using both open circuit and floating current measurements. Existing degradation studies usually focus on commercial cells. The initial electrolyte composition and formation protocol for these cells is often unknown. This study investigates the role of electrolyte additives, specifically, vinylene carbonate (VC) and fluoroethylene carbonate (FEC), in the aging process of lithium-ion batteries. The results showed that self-discharge plays a significant role in determining the severity of aging for cells without additives. Interestingly, the aging was less severe for the cells without additives as they deviated more from their original storage state of charge. It was also observed that the addition of VC and FEC had an effect on the formation and stability of the solid electrolyte interphase (SEI) layer on the surface of the carbonaceous anode. By gaining a better understanding of the aging processes and the effects of different electrolyte additives, we can improve the safety and durability of lithium-ion batteries, which is critical for their widespread adoption in various applications.

1. Introduction

Lithium-ion batteries (LIBs) have become a ubiquitous power source in today’s world, powering everything from mobile devices to electric vehicles. However, one of the key challenges of LIBs is their limited cycle and calendar life, which can significantly affect their performance and safety over time. Calendar aging, in particular, is a major concern for LIBs, as it can cause irreversible degradation of the battery’s capacity and increase the risk of safety hazards [1,2,3].

Calendar aging studies have been performed by many authors to model the aging behavior or to elucidate the mechanisms responsible for the reduced lifetime of LIBs [4,5,6,7,8,9,10]. The majority of calendar aging studies have been performed on commercial cells with graphite anodes and various cathode materials such as Li(NiMnCo)O₂ (NMC), Li(NiCoAl)O₂ (NCA), LiMn₂O₄ (LMO), LiFePO₄ (LFP), and LiCoO₂ (LCO). In contrast to known active materials, electrolytes and their additives are usually not specified. Ethylene carbonate (EC) is used as a solvent in LIBs, which exhibits a high dielectric constant to dissolve the inorganic salt [11]. Due to the relatively high melting point of EC (36 °C), it is mixed with linear aliphatic carbonates such as dimethyl- (DMC), diethyl- (DEC), and/or methyl ethyl-carbonate (EMC). This mixture is combined with lithium-ion conductive salts such as LiPF₆ to form the electrolyte [11,12,13]. Calendar aging refers to the idle time of the battery, so the thermodynamic stability of the components governs the aging process [14]. Previous studies have shown a strong influence of temperature and the state of charge on the aging behavior. Ecker et al. [5] state an accelerated aging of NMC111/graphite cells for both high temperature and state of charge (SoC). This can be explained by the increased reaction rates induced by higher temperatures and the instability of the electrolyte, due to electrode/electrolyte reactions at high cell potentials [4]. The work of Schmalstieg et al. [10] used the same cells as Ecker et al. [5] at elevated temperatures and with a focus on high SoC resolution of 12 different SoCs in the range of 0 to 100%. Their study confirmed an increased capacity loss and resistance increase with higher storage SoC. Keil et al. [8] performed calendar aging tests on cells with different cathodes and graphite anodes. Their work revealed a non-linear influence of the SoC on calendar aging, resulting in plateau regions with a largely constant capacity fade. Dubarry et al. [4] reviewed calendar aging of graphite anodes with respect to various cathode materials. They found accelerated degradation for temperatures above 50 °C and storage SoCs greater than 60%. In addition, cells with NMC cathodes appeared to suffer greater degradation than those with NCA cathodes. Hahn et al. [7] studied $50.8$ $\mathrm{A}$$\mathrm{h}$ NMC111 automotive cells and analyzed the capacity loss and volume change of the full cell during calendar aging. Their approach also considered the capacity loss due to the initial SEI growth during formation.

Calendar aging studies can be performed by charging or discharging the cells to the desired voltage [6] or SoC [15] and storing them at a certain temperature without compensation of voltage changes. In contrast to these so-called open circuit voltage (OCV) storage tests, the voltage can be kept at a fixed voltage level (floating SoC) [5]. Thus, voltage changes due to side reactions or self-discharge can be compensated [16]. Since most of the calendar aging studies used only one of the two methods, a direct comparison is difficult. Käbitz et al. [17] report significant differences only for 100% SoC.

One approach to mitigate calendar aging is to modify the battery’s electrolyte by adding various additives. These electrolyte additives can improve the battery’s performance by enhancing its electrochemical stability, preventing the formation of unwanted reaction products, and reducing the negative impact of side reactions [18,19,20,21]. However, the effect of these additives on the calendar aging of lithium-ion batteries is not well understood.

The electrochemical reduction of the electrolyte co-solvent ethylene carbonate (EC) during the initial formation cycles of the battery cell is crucial for the development of the passivating interphase [22]. Two widely used additives are vinylene carbonate (VC) and fluoroethylene carbonate (FEC), which are known to improve the performance of lithium-ion batteries. Several studies have reported that the addition of small amounts of VC or FEC to the electrolyte results in a reduction of SEI thickness, an improvement in cell capacity retention, and an increase in interphase temperature stability, while retaining the positive character of the SEI layer [21,23,24,25,26].

This paper aims to investigate the influence of the electrolyte additives VC and FEC on the calendar aging of lithium-ion batteries by OCV and floating SoC storage tests. The study focuses on evaluating the performance and aging behavior of batteries containing different electrolyte additives under various storage SoCs at 25 °C. The findings of this research could provide valuable insights into long-term reliability of lithium-ion batteries, regarding the influence of the two commonly used additives VC and FEC.

2. Materials and Methods

2.1. Cell Information

The high energy pouch cells used in this study were manufactured by an in-house battery production line and have a nominal capacity of $5.2$ $\mathrm{A}$$\mathrm{h}$ and a nominal voltage of $3.7$ $\mathrm{V}$, with voltage limits of $3.0$ $\mathrm{V}$ and $4.2$ $\mathrm{V}$. The anode material is graphite and the cathode material is Li(Ni_0.6Mn_0.2Co_0.2)O₂ (NMC622). Utilized electrode materials were processed in an in-house manufacturing plant. Active material for positive electrodes, i.e., NMC622, was coated onto aluminum foil (thickness: 15 $\mathrm{\mu }$$\mathrm{m}$) with conductive carbon and polyvinylidene difluoride in a ratio of 93/2/3 $\mathrm{wt}\%$, respectively. For negative electrodes, graphite (as active material), sodium carboxymethyl cellulose, and styrene butadiene rubber (ratio of 95.0/0.5/1.5 $\mathrm{wt}\%$, respectively) were coated onto copper foil (thickness of 10 $\mathrm{\mu }\mathrm{m}$). N-methyl-2-pyrrolidone was used as a processing solvent for positive electrodes, while the negative electrodes were processed with water. For NMC622, the chosen mass loading was 13.5 mg cm⁻², representing an aerial capacity of 2.3 mA h cm⁻². The chosen mass loading for graphite based electrodes was 7.9 mg cm⁻², resulting in an aerial capacity of 2.8 mA h cm⁻². Electrode sheets were calendared and dried under reduced pressure (120 °C; $1\times {10}^{-3}$ mbar) for four days. To separate the electrodes, a Mitsubishi OZ-S30 separator (polyester non-woven separator with ceramic coating) was used. The base electrolyte is constituted of 1 M LiPF₆ in LP57 (EC:EMC = 3:7 $\mathrm{wt}\%$) and serves as a reference for the following aging tests. In order to investigate the influence of electrolyte additives on the calendar aging behavior, 5 $\mathrm{wt}\%$ of either VC or FEC was added by the supplier (Solvionic, Toulouse, France) as two electrolyte variations. This amount was chosen to gain better understanding of the additives, although higher amounts of VC can have detrimental effects [27]. The cells were filled and sealed in a dry room with a dew point of $-70$ °C. An overview over the specific parameters of the cells in this study is given in

Table 1. Cell parameters for the pouch cells used in the aging studies.

Parameter	Value
Nominal capacity	$5.2$ $\mathrm{A}$$\mathrm{h}$
Capacity used for charge and discharge protocol	$5.0$ $\mathrm{A}$$\mathrm{h}$
Nominal voltage	$3.7$ $\mathrm{V}$
End of charge, voltage	$4.2$ $\mathrm{V}$
End of discharge, voltage	$3.0$ $\mathrm{V}$

. The aging study was performed with 8 cells per electrolyte composition. This results in a total number of 24 cells.

2.2. Formation

The calendar aging studies were carried out at different locations, with the use of different testing equipment. All electrochemical tests were performed with the exact same cycling parameters to ensure comparability. The formation cycle consisted of a charge step to the upper cut-off voltage of $4.2$ $\mathrm{V}$ and a discharge step to the lower cut-off voltage of $3.0$ $\mathrm{V}$. Both the charge and discharge current were 1 $\mathrm{A}$, which equals a C-rate of C/5. The temperature of the cells during formation was recorded with thermocouples type K. After the formation cycle, the cells were degassed and vacuum sealed in an argon-filled glove box (H₂O < $0.1$ ppm, O₂ < $0.1$ ppm). Within this process, gas samples were taken through a gas-tight syringe from the pouch bag of some cells and analyzed with a gas chromatography–mass spectrometry (GC-MS) system (Clarus 690 Arnel 5019, PerkinElmer LAS GmbH, Rodgau, Germany). The formation of the cells for the open-circuit storage tests was carried out using a BasyTec XCTS (BaSyTec GmbH, Asselfingen, Germany) test device. It was conducted inside a Binder KB 400 (Binder GmbH, Tuttlingen, Germany) temperature chamber at a temperature of 25 °C. The formation of the cells used in the floating SoC tests was carried out inside a Memmert (Memmert GmbH + Co. KG, Schwabach, Germany) temperature chamber with a Maccor 4200 battery cycler (Maccor Inc., Tulsa, OK, USA).

2.3. Test Setup

Prior to the calendar aging tests, the cells were cycled with a constant current of C/5 for 10 cycles to obtain a stable discharge capacity. This serves as begin of life (BoL) with the initial characterization test. The voltage–SoC characteristic for a C/10 discharge is shown in Figure 1. The three different electrolytes show very similar behavior and the corresponding voltages for the storage SoCs of 100%, 80%, 30%, and 5% are listed in

Table 2. SoC and corresponding voltages for varying electrolytes.

SoC	LP 57	LP 57 + 5% VC	LP 57 + 5% FEC
100%	$4.18$ $\mathrm{V}$	$4.18$ $\mathrm{V}$	$4.19$ $\mathrm{V}$
80%	$3.94$ $\mathrm{V}$	$3.94$ $\mathrm{V}$	$3.94$ $\mathrm{V}$
30%	$3.57$ $\mathrm{V}$	$3.56$ $\mathrm{V}$	$3.56$ $\mathrm{V}$
5%	$3.36$ $\mathrm{V}$	$3.36$ $\mathrm{V}$	$3.36$ $\mathrm{V}$

.

2.4. Characterization Tests

In order to quantify the aging and to gain insight into the degradation mechanisms, the cells were intermediately characterized every 4 months and compared to the values of the BoL test. This characterization is composed of a capacity check, a quasi-OCV cycle and pulse resistance measurements at two different SoCs (70% and 50%). The capacity test itself consists of a full charge with a rate of C/2, followed by a capacity check with a discharge rate of C/2. The values obtained from this capacity check are used as a measure of capacity fade in this publication. Additionally, cells are subjected to quasi-OCV charge and discharge cycles with a C-rate of C/10. Pulse tests at 1C current are performed at SoCs of 70% and 50% to determine the change in resistance during aging. Here the pulse resistance $R_{pulse}$ is defined as follows:

$$R_{pulse}={\displaystyle \frac{|U\left(t_0\right)-U(t_0+10\mathrm{s})|}{|I_{pulse}|}}$$

(1)

The pulse tests were performed in both charge and discharge directions to negate any influence on the SoC. With a duration of 15 $\mathrm{s}$, the change of SoC is less than 0.5%. The capacity measured during the capacity test was used to determine the storage SoC for the next storage period for OCV aging. Constant-voltage aging was always performed at the same voltages. After the cells were characterized, they were taken back to their storage chambers. The tests were designed to minimize the influence of the test while ensuring a sufficiently high resolution of the data.

2.4.1. (a) Open-Circuit Tests

The characterization tests for aging under OCV conditions were performed at 25 °C inside an Binder KB 400 (Binder GmbH, Tuttlingen, Germany) with a BioLogic BCS 815 battery cycler (Biologic, Seyssinet-Pariset, France). Prior to testing, the cells were allowed to thermally equilibrate for 2 $\mathrm{h}$ after removal from their storage locations to the temperature (25 °C) of the characterization test. After the BoL characterization tests, the cells were charged to their respective storage SoC and stored in a Nüve ES120 (Nüve, Ankara, Turkey) temperature chamber at 25 °C for 4 months before they were characterized again. The calendar aging studies under OCV conditions were conducted for a total of 16 months.

2.4.2. (b) Floating SoC Tests

The floating SoC measurements were conducted using a Neware (Neware Technology Limited, Shenzhen, China) device in a Memmert temperature chamber (Memmert GmbH + Co. KG, Schwabach, Germany).

2.5. Self-Discharge

During open circuit calendar aging, a lithium-ion battery degrades over time, even when not in use. This results in less usable capacity and an increase in internal resistance. It is important to differ between irreversible and reversible capacity losses. An irreversible loss of capacity results from the loss of cyclizable lithium bound in the SEI [28]. Reversible capacity loss refers to the loss of charge over time due to various factors such as chemical reactions and leakage currents, a phenomenon commonly referred to as self-discharge. Self-discharge can cause the battery’s SoC to decrease, resulting in degraded battery performance and reduced capacity. In the present calendar aging study under open-circuit conditions, the cells were not connected to any charging device and therefore the voltage decreased during storage. In contrast to these measurements, the floating SoC measurements were performed under constant-voltage conditions, where the cells are kept at a fixed voltage.

Similar to the assessment of self-discharge given by the work of Werner et al. [29], two different approaches were used. The first method uses the charge throughput, which is necessary to charge the cells to 100% after each storage period. The self-discharge ($S{D}_x$) of the aging period x is defined in Equation (2), as used by Werner et al. [29].

$$S{D}_x={\displaystyle \frac{C_{N,x-1}-(Q_{SoC,x-1}+Q_{100\%-SoC,x})}{C_{N,x-1}}}<0$$

(2)

The charge throughput $Q_{SoC,x-1}$ equals the amount of charge, which was used to charge to the desired storage SoC of the previous storage period. This is then added to charge throughput, which was needed to reach a full charge after the storage period. By dividing it by the cell capacity $C_{N,x-1}$, the self-discharge is normalized.

The second approach to measuring self-discharge is based on the change in voltage between subsequent storage periods. The voltage after the characterization test is compared with the voltage of the cells before the next check.

Self-discharge could not be analyzed for cells under constant voltage storage conditions.

3. Results and Discussion

3.1. Formation

For each electrolyte formulation, 8 cells were used, for a total of 24 cells. The formation procedure was carried out in the same way for all cells. The type of electrolyte used can affect the behavior of lithium ions and their interaction with the anode and cathode, resulting in different electrochemical reactions at different voltages [22]. A commonly used technique to improve battery performance is the addition of electrolyte additives. This method has been shown to increase battery life in both calendar and cycle terms, while reducing negative effects such as active lithium depletion during SEI formation and repair, electrolyte oxidation at the positive electrode, transition metal dissolution, and electrode damage [30,31].

Figure 2 shows an example of the first charge sequence during formation for the three different electrolyte combinations. The voltage curves are generally very similar. The magnifications in the potential window between 2 and $3.2$ $\mathrm{V}$ show deviating voltage curves at the beginning of the very first charge sequence (SoC < 2%). Both FEC and VC have a low Lowest Unoccupied Molecular Orbital (LUMO) level, which means that they are more easily reduced than EC at the interface between the anode and the electrolyte during the initial charging process [22]. These reduction processes appear as a change in the slope of the voltage vs. the SoC curve. They become even more apparent within a Incremental Capacity Analysis (ICA) of the formation process. The incremental capacity (IC) is calculated (Equation (3)) as the derivative of the capacity (dQ) with respect to the voltage (dV).

$$IC={\displaystyle \frac{dQ}{dV}}\approx {\displaystyle \frac{\Delta Q}{\Delta V}}$$

(3)

Prior to the analysis, the raw voltage and capacity data were filtered to obtain fixed voltage intervals of 3 $\mathrm{m}$$\mathrm{V}$ [32,33,34]. In Figure 3, the ICA for the first charge cycle is shown. The plot was created with the same data as for the charge vs. capacity plot. The first peak corresponds to the reduction of FEC at around $2.7$ $\mathrm{V}$. Secondly, VC is reduced at around $2.9$ $\mathrm{V}$, shortly before the EC from the standard electrolyte. This is in good agreement with the measurements from Nie et al. [24], who investigated very similar electrolyte combinations in graphite half cells. In addition, the work of Delp et al. [35] compared the same baseline electrolyte with different additives on the first charge cycle. They observed the same order of peaks, with FEC cells first, followed by FEC-filled cells before the final reduction peak of the baseline electrolyte.

Gas samples were taken from two cells of each type during the degassing and resealing of the pouch cells. They were qualitatively analyzed with gas chromatography–mass spectrometry (GC-MS).

Table 3. Gas species detected after the formation cycle.

LP 57	LP 57 + 5% VC	LP 57 + 5% FEC
H2, C2H4, C2H6, CH4, CO, DMC, EMC, DEC	H2, CO2, C2H4, C2H6, CH4, CO, EMC	H2, CO2, C2H4, C2H6, CH4, CO

lists the gaseous species, which were found in the different cells.

All samples contain H₂, C₂H₄, C₂H₆, CH₄, and CO. While the exact reaction mechanisms are still under discussion, various authors have reported CO₂ for both VC- and FEC-containing electrolytes [22,23,24,36,37]. This work confirms CO₂ as a product of VC- and FEC-containing electrolytes during formation.

3.2. Initial State

After the formation process, all 24 cells were cycled for 10 cycles at a current rate of C/5 at a temperature of 25 °C to obtain stable cell properties for a BoL characterization. This characterization consists of capacity tests and pulse current measurements. The values for the discharge capacity C_C/2 and the pulse resistance R_pulse listed in

Table 4. Initial state and deviation of the cells.

Parameter	LP 57	LP 57 + 5% VC	LP 57 + 5% FEC
CC/2/Ah	5.02 ± 0.11	5.21 ± 0.08	5.17 ± 0.05
$R_{\mathit{pulse}}$/$\mathrm{m}\mathrm{\Omega }$	15.26 ± 1.86	20.65 ± 0.53	12.34 ± 0.50

are the initial values that will be used here to quantify aging.

In terms of discharge capacity, the cells with additives show superior performance. The cells without additives exhibit an initial capacity of (5.02 ± 0.11) Ah, in contrast to (5.21 ± 0.08) Ah for the cells containing VC and (5.17 ± 0.04) Ah for FEC cells. Furthermore, the overall deviation of additive-free cells is considerably higher. Regarding the pulse resistance R_pulse, the cells with FEC feature the lowest values. Cells containing VC have the highest pulse resistance. High concentrations of VC were shown to exhibit a higher charge transfer resistance, especially at the graphite anode [19]. This can increase the resistance of the battery to the flow of lithium ions and reduce the battery’s pulse power capabilities. On the other hand, FEC forms a thinner and more stable SEI layer, which allows for better lithium-ion transport and hence better pulse power performance. Burns et al. [19] conducted studies with varying VC content up to a maximum of 6%. Their work confirmed the beneficial effect of VC as an electrolyte additive, although they reported increased cell impedance for VC contents above 2% compared to the control electrolyte. The work of Hildenbrand et al. [27] used Bayesian optimization to improve the content of both VC and FEC as additives. They suggested a maximum of 5% of each additive for further study. Their studies also showed detrimental effects of VC concentrations above this threshold, while no such limitation was found for the use of FEC.

3.3. Self-Discharge

The mechanism of self-discharge in Li-ion batteries involves several chemical and electrochemical processes that can occur even when the battery is not in use [38]. One of the primary mechanisms of self-discharge is the reaction between the battery’s electrodes and the electrolyte [39]. Over time, lithium ions can slowly migrate between the battery’s electrodes, even when the battery is not in use. This can cause the battery’s voltage to drop as a result of the gradual loss of charge. Another mechanism of self-discharge is the formation of side reactions within the battery. These side reactions can be caused by impurities in the battery materials or by other chemical processes that can occur within the battery [38].

The studied calendar aging was partially performed under OCV conditions. In this case, the cells were not recharged during the storage periods and self-discharge of the cells was not prevented.

Table 5. Self-discharge during storage for the investigated storage conditions and electrolyte combinations.

SoC	LP 57 Max./Avg.	LP 57 + 5% VC Max./Avg.	LP 57 + 5% FEC Max./Avg.
100%	93%/79%	2.9%/2.1%	4.8%/3.8%
80%	72%/65%	0.6%/0.2%	2.3%/1.5%
30%	42%/32%	0%/−0.3%	1.9%/0.9%
5%	9.9%/7.8%	1.1%/0.4%	0.7%/0.2%

shows the maximum and average self-discharge rates during the storage period for all investigated storage conditions and electrolytes. The cells with pure LP57 suffered from severe self-discharge, which was most prominent at high SoCs. At an SoC of 100%, the maximum self-discharge was found to be 93%, which is attributed to a nearly complete discharge during a storage period of 4 months. For the cells stored at 30% and 5% SoC, the maximum self-discharge exceeded the actual state of charge, which can be explained by an increase in capacity during these storage periods, which will be further discussed in the following section.

The cells containing additives showed only moderate self-discharge. The cells with VC showed the lowest self-discharge. In general, the self-discharge was higher for higher SoCs.

Self-discharge in calendar aging experiments has been considered in previous studies [9,40]. These studies were performed using commercial Kokam cells with either NMC/graphite or NCA/graphite. The electrolyte and possible additives are not reported. It can be assumed that commercial batteries use sacrificial additives to improve their performance and therefore the results cannot be directly compared [22,41]. Zhang et al. [21] studied the self-discharge behavior of various electrolyte additives in NMC111/graphite cells with an unspecified electrolyte. They showed that over a storage time of 550 $\mathrm{h}$, both small amounts of VC and FEC reduce the self-discharge, with VC being superior in terms of self-discharge compared to FEC.

Overall, the self-discharge process in lithium-ion batteries is complex and can involve several different mechanisms. The gradual loss of charge due to the migration of lithium ions between the battery’s electrodes, the formation of side reactions, and the impact of temperature are factors that can contribute to a decrease in the cell voltage and an increase in the battery’s self-discharge rate over time [42]. Recent research by Buechele et al. [43] attributed the self-discharge to the redox shuttle molecule dimethyl terephthalate (DMT). Their measurements showed that the addition of additives such as VC or FEC are able to suppress the undesired redox shuttle of DMT. Furthermore, Adamson et al. [44] showed that DMT is a decomposition product of the adhesive polyethylene terephthalate (PET) tape, which is used to secure the cell stack.

The self-discharge has a strong influence on the storage SoC. Its impact on the capacity and pulse resistance will be discussed in the next sections.

3.4. Capacity

3.4.1. OCV Conditions

Figure 4 shows the change in actual capacity over a 16-month period for all SoCs at 25 °C. The influence of the SoCs is obvious, resulting in faster degradation for cells stored at high potentials such as SoC100 or SoC80. This is in good agreement with the literature, where it has been shown that an increasing SoC causes a higher capacity loss [4,5,17,40].

The cells at the lowest SoC (see Figure 4a) investigated show a very small capacity fade and during the first storage period a slight capacity increase can be observed for all electrolyte combinations. This can be attributed to the so-called “anode overhang effect”, which can be responsible for a reversible capacity gain or loss [45]. The reversible capacity change is due to a geometrically larger anode compared to the cathode. The cells used in this study have an anode that overlaps the cathode by 1 $\mathrm{m}$$\mathrm{m}$ on all sides. For a given lithium ion concentration gradient between the overhang and the part of the anode facing the cathode, lithium ions can migrate to or from the overhang, resulting in a measurable change in discharge capacity [46,47,48,49]. Prior to the start of the first storage period, the cells were at a SoC of 50% for a few days. Afterwards, the cells underwent BoL characterization and were charged or discharged to their storage SoCs. Therefore, the cells stored beneath this SoC exhibit a capacity increase, due to the overhang, and the cells stored at higher potentials suffer from a capacity loss. However, the aging processes overlap and are always dependent on the actual materials. The cells at SoC30 (see Figure 4b) show a higher capacity fade than those at SoC 5. One of the cells without additives shows a high capacity fade of about 10%, followed by an increase of almost 4%. After the second storage period, this explicit cell showed significant swelling of the pouch bag and was separated from the other cells and tested in a safe environment. The swelling disappeared after the characterization test and the cell was left in the test, in order to study the further gassing behavior. The other cell at SoC30 containing only LP57 showed superior performance and after one year of storage, the remaining capacity was nearly the same as at the beginning of the calendar aging tests. The cells with additives show quite similar capacity fade and the deviation between the cells is also very small. The cells stored at SoC80 (see Figure 4c) showed a higher degradation, with a capacity fade of more than 10% for the cells with additives and around 5% for those without additives. At the highest investigated SoC (see Figure 4d), which corresponds to a full charge, the cells with FEC show a significantly higher degradation. The VC cells show only small differences between SoC100 and SoC80. In general, the cells with additives show very similar degradation, while the pure LP57 cells show a larger scattering. The overall superior performance of the cells without VC and FEC is quite counterintuitive, as previous studies have reported superior performance due to improved SEI formation from the addition of VC and FEC [50]. In this particular study, the cells exhibit such a high self-discharge that the actual storage SoC deviates greatly from the values set at the beginning of each storage period. This explains the smaller capacity fade in the cells without additives.

3.4.2. Floating SoC Conditions

Under constant voltage conditions, self-discharge does not influence the SoC value of the cell, as any decrease in cell voltage is compensated by an external current. Hence, no statement on the self-discharge of these cells can be made. Furthermore, the temperature of the oven was lower during the initial test, affecting initial capacity and resistance values. For storage at $3.35$ $\mathrm{V}$ (see Figure 5a), an initial increase in capacity is observed, followed by a slow capacity fade for all cells, regardless of the electrolyte. However, the magnitude of this increase was higher in cells containing VC. In cells stored at $3.55$ $\mathrm{V}$ (see Figure 5b), a similar behavior is visible: VC-containing cells exhibit an initial increase in capacity, followed by a decay. When stored at the upper cutoff voltage of $4.2$ $\mathrm{V}$ (see Figure 5d), all cells suffer from substantial capacity fade. VC-containing cells exhibit the lowest fade, while FEC-containing cells, and one cell without additives loose over 10% of the initial capacity. Cells with VC stored at $3.9$ $\mathrm{V}$ (see Figure 5c), corresponding to roughly 80% SoC, exhibit some initial capacity gain. For all storage conditions investigated, we can conclude that cells with VC lose less capacity.

3.5. Resistance

3.5.1. OCV Conditions

In addition to the evaluation of capacity measurements, Figure 6 shows the increase in the pulse resistance in the pulse tests, which were carried out at 50% SoC. The pulse resistance measurements for 70% SoC show similar trends. Thus, in the following, only the measurements for SoC50 will be taken into consideration.

According to the qualitative analysis of the data, there are similar trends in the behavior of the battery over time with respect to its SoC. The data (see Figure 4) suggest that the aging process is more pronounced at higher potentials, such as SoC100 and SoC80. The pulse resistances show the same dependence as the capacity. The cells stored at SoC5 (see Figure 6a) show only minor differences among each other. The pulse resistance increased by 20% while the capacity remained constant. This can be explained by the simultaneous mechanisms of the migration of lithium ions from the anode overhang and degradation mechanisms. At SoC30 (see Figure 6b) and SoC80 (see Figure 6c), the cells with VC show the highest increase in pulse resistance. The cells with FEC increase by only 20% at SoC30. The cell that showed gassing during storage has a significantly higher increase in pulse resistance than the other cell under the same conditions. In general, VC is more stable than FEC at high potentials. The cells, which are charged to the upper cut-off voltage (see Figure 6d) show the highest resistance increase. This is due to the high potential, which is near the stability window of the electrolyte [17]. The cells without additives are self-discharging and are therefore only briefly exposed to the high potentials. When used with a carbonaceous anode, VC forms a stable SEI layer that is dense and uniform. This SEI layer is mainly composed of lithium alkyl carbonate species, which are formed through the reaction of VC with the lithium ions in the electrolyte. The lithium alkyl carbonate species have good mechanical strength and electrochemical stability, which help to protect the anode from further reactions with the electrolyte [22,51].

On the other hand, FEC forms a more porous and less stable SEI layer on the carbonaceous anode. This SEI layer is mainly composed of lithium fluoride and polymeric materials, which are formed through the reaction of FEC with the lithium ions in the electrolyte [22,51]. The polymeric materials in the SEI layer can be less stable and can contribute to the formation of a thicker SEI layer over time, which can reduce the performance and capacity of the battery.

3.5.2. Floating SoC Conditions

First, the measurements show a significant drop in relative resistance for some of the cells after the third characterization test, which can be explained by a change in temperature in the climate chamber from 20 °C to 25 °C. Due to the higher temperature, the internal resistance of the cells decreased.

Under constant voltage aging, cells without additives exhibit the highest relative resistance increase when considering the median of the tested cells at all conditions except 5%-SoC (see Figure 7). Under these conditions, the cells with 5% VC perform worse.

However, the spread of cells containing VC is very large. Two cells with VC show a high resistance increase at 5% SoC, while the third cell with VC at this SoC exhibits the lowest resistance increase of all cells. Cells with FEC exhibit lower resistance increase compared to cells without additives. Nie et al. [24] investigated the surface films with respect to electrolyte additives, and their work showed that the use of FEC results in a surface film that is rich in LiF and thinner than the SEI formed by their control electrolyte without additives.

A comparison with cells containing VC is challenging due to the large spread between these cells. Nevertheless, one explanation for the high impedance growth of VC cells may be the relatively high concentration of 5% VC [19,27]. A similar effect has not been reported for FEC. The difference in degradation between VC and FEC cells is due to the different composition and morphology of the SEI resulting from different reduction reactions and products [50]. The resistance increase of cells with FEC is approximately in the same range as the best-performing VC-containing cell.

Most of the studies on the effect of additives on the performance of LIBs have been conducted to investigate cycle life. Battery degradation cannot be assessed by cycle life alone because it always overlaps with calendar aging and both processes must be considered to improve battery performance and safety.

4. Conclusions

This study investigated the impact of electrolyte additives on the calendar aging of pouch cells with NMC622 cathodes and graphite anodes under different states of charge. Moreover, the two commonly used approaches of open-circuit and floating SoC calendar aging were compared for four different SoCs (SoC100, SoC80, SoC30, and SoC5). While there are several calendar aging studies on commercial lithium-ion batteries, most of them have focused on the electrode materials without detailed knowledge of the electrolyte and additives. Our studies aim to fill the gap in the literature between calendar aging on commercial cells and studies focused on the electrolyte system. In our case, both open-circuit and floating SoC calendar aging did not show similar degradation regarding the investigated electrolyte. The results showed that self-discharge played a critical role in determining the severity of aging in open circuit conditions for cells without additives. Interestingly, the aging was less severe for those cells without additives as they deviated from their original storage state of charge. Therefore the cells without additives were stored at significantly lower SoCs than the cells with VC and FEC. Under floating SoC conditions, the cell voltage is kept constant and the effects of self-discharge are mitigated. The floating SoC tests confirm better capacity retention for cells with electrolyte additives compared to the baseline electrolyte. In addition, cells with VC showed the least degradation for all SoCs tested.

For future research, a deeper understanding of the self-discharge mechanism is needed to better understand the impact of electrolyte additives on aging. It will also be interesting to investigate the effect of different temperatures on calendar aging, as this can significantly affect the stability of the battery. Overall, this study sheds light on the important role of electrolyte additives in battery aging, and highlights the need for further investigation to optimize the performance and durability of lithium-ion batteries.