Evaluation of Alternative Lithium Salts for Li Ion Batteries With SiO x ‐Containing Anodes: Characteristic Failure Mechanisms and Different Impacts of the Fluoroethylene Carbonate Additive
Anindityo Arifiadi, Jaroslav Minar, Ankita Das, Linus Voigt, Dominik Voigt, Marc Vahnstiege, Julius Buchmann, Feleke Demelash, Peng Yan, Simon Wiemers‐Meyer, Sascha Nowak, Frank Glorius, Martin Winter, Johannes Kasnatscheew
TL;DR
This paper explores how different lithium salts and additives can improve the cycle life of lithium-ion batteries with silicon oxide anodes.
Contribution
The study reveals a novel 'self-healing' mechanism via FEC and LiDFOB that reactivates lost capacity in SiOx anodes.
Findings
Replacing LiPF6 with LiBOB or LiDFOB improves cycle life in SiOx-containing anodes.
FEC suppresses failure cascades and reduces active lithium loss.
LiDFOB enables active lithium generation through oxidation of soluble SEI species.
Abstract
Incorporating silicon‐based active materials, e.g., SiO x , into the negative electrodes can increase the gravimetric/volumetric energy of Li ion batteries. Nevertheless, SiO x shortens cycle life due to the large volume expansion during charge/discharge cycling. The mechanical stress on the solid electrolyte interphase (SEI) necessitates continuous SEI repair, which accelerates active lithium loss (ALL) over cycling. In this work, the impact of common lithium salts is investigated in electrolytes for LiNi0.5Co0.2Mn0.3O2 (NCM 523) || 10%SiO x ‐graphite Li ion pouch cells. The end‐of‐life (EOL) with LiPF6 can be enhanced by anode passivation via fluoroethylene carbonate (FEC), which not only decreases ALL but also suppresses failure cascades (e.g., electrode crosstalk) initiated by HF over the course of SiO x reactions with LiPF6. Though ALL remains similar when adding FEC to lithium…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4
FIGURE 5
FIGURE 6
FIGURE 7
FIGURE 8| Salt | Solvent | Additive | Name |
|---|---|---|---|
| 1 M LiFSI | EC:EMC3:7 by weight | — | LiFSI |
| 1 M LiTFSI | — | LiTFSI | |
| 1 M LiBF4 | — | LiBF4 | |
| 2 wt% FEC | LiBF4 + FEC | ||
| 1 M LiPF6 | — | LiPF6 | |
| 2 wt% FEC | LiPF6 + FEC | ||
| 0.6 M LiBOB | — | LiBOB | |
| 2 wt% FEC | LiBOB + FEC | ||
| 1 M LiDFOB | — | LiDFOB | |
| 2 wt% FEC | LiDFOB + FEC |
- —Bundesministerium für Bildung und Forschung10.13039/501100002347
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvancements in Battery Materials · Advanced Battery Materials and Technologies · Fiber-reinforced polymer composites
Introduction
1
Achieving lithium‐ion batteries (LIBs) with even higher gravimetric/volumetric energy requires development on the active material level [1]. The abundant silicon is considered a promising material to replace the state‐of‐the‐art graphite anode due to its larger theoretical specific capacity of ≈3500 versus 372 mAh g^−1^ and capacity densitiy ≈2194 versus 719 mAh cm^−3^ [2, 3], but is limited by a severe volume change (≈300%) during the (de)lithiation process that results in particle cracking, pulverization, electronic contact loss and consequently severe capacity fade [4, 5]. The intertwined continuous rupture and reformation of the solid–electrolyte interphase (SEI, = SEI repair) throughout cycling, leading to continuous active lithium loss (ALL), is considered the predominant reason for capacity degradation at full cell level [6]. As a result, Si is utilized in smaller amounts in LIB anodes (≈5–10 wt%) in form of Si‐ or SiO_ x _‐graphite composite electrodes [1, 7, 8]. This relatively small amount of Si can relevantly increase anode capacity and enable thinner anodes. As a result, energy density and fast charging performance can be improved [9].
The addition of extra active lithium prior cell assembly during a so‐called pre‐lithiation approach and/or structural design of Si (e.g., nano wires, SiO_ x _ composite) to compensate ALL and better withstand the volume changes are common strategies of current R&D [10, 11, 12, 13, 14, 15], though an effective SEI remains a vital criteria for application in order to prevent continuous ALL in the first place. In general, an ideal SEI layer is formed in the initial cycle(s) with minimal ALL and is electronically insulating while ionically conductive [16]. For Si, elasticity is additionally required to withstand the volume changes [17, 18], which poses an extra challenge for electrolyte R&D [5, 19, 20, 21, 22, 23]. Electrolyte additives such as vinylene carbonate (VC) or fluoroethylene carbonate (FEC) are found to form an SEI that can better accommodate the large volume expansion of Si due to the formed polymeric species [22, 24, 25]. Also, employing FEC as co‐solvent or 1,1,2,2‐tetraethoxyethane as a single solvent has been shown to be beneficial for increasing the flexibility of the SEI [22, 23, 26]. Replacing LiPF_6_ with lithium bis(fluorosulfonyl)imide (LiFSI), lithium bis(trifluoromethanesulfonyl)imide (LiTFSI), lithium bis(oxalato)borate (LiBOB) or lithium difluoro(oxalato)borate (LiDFOB) can also be beneficial, due to the more anion‐dominated, thus inorganic‐enriched SEI formation. Obviously, the presence of both organic and inorganic species is essential and complicates the SEI R&D [21, 27, 28].
The impact of different lithium salts on the electrochemical performance of Si || Li coin cells has been reported by Zheng et al*.* [21]. Although meaningful insights on salt decomposition mechanism and its corresponding SEI chemistry are presented, the insights may not be valid in practical full cells, i.e., with a limited lithium inventory [29, 30]. In these regards, this work investigates the impact of different lithium salts in 200 mAh wounded pouch cells with LiNi_0.5_Co_0.2_Mn_0.3_O_2_ (NCM 532)‐based cathode and SiO_ x ‐graphite composite anode (10 wt% SiO x _), focusing on relevant cell performance metrics, i.e., cycling stability, resistances, self‐discharge, gassing, electrolyte decomposition, electrode crosstalk, and SEI characteristics.
Results and Discussion
2
Cell Formation and Electrochemical Performance
2.1
Figure 1a–f presents the first cycle cell voltage versus capacity graphs, as well as differential capacity versus voltage plots of LiNi_0.5_Co_0.2_Mn_0.3_O_2_ (NCM 532) || 10%SiO_ x ‐graphite cells with electrolyte formulations defined and named in Table 1 . Figure 1d shows that cells with LiPF_6 in ethylene carbonate (EC)‐ethyl methyl carbonate (EMC) mixture exhibit a reduction peak in the 1st cycle at 3.0 V (≈0.75 V vs. Li|Li^+^), suggesting EC reduction [32, 33]. This peak remains unchanged in cells with LiFSI and LiTFSI, rendering their limited participation in solid electrolyte interphase (SEI) formation. Moreover, these salts are impractical for application as cells with lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) do not deliver discharge capacity (Figure S1a) and cells with lithium bis(fluorosulfonyl)imide (LiFSI) rapidly fade (Figure 2a). This phenomenon occurs due to the anodic dissolution of the aluminum current collector that causes foil disintegration and consequently electrical disconnection [34, 35].
Cell voltage versus capacity, and their corresponding differential capacity versus voltage plots of cells with (a,d) LiFSI and LiTFSI, (b,e) LiBF4 and LiBF4 + FEC, and (c,f) LiBOB, LiBOB + FEC, LiDFOB, and LiDFOB + FEC electrolytes compared to LiPF6 and LiPF6 + FEC electrolytes. Note that the charge capacities in (a–c) are much lower than the nominal cell capacity (200 mAh) due to the limited cut‐off voltage (3.5 V) in the first formation cycle. The negative electrode potentials in (d–e) are estimated from three‐electrode electrochemical experiments of electrodes extracted from the pouch cells used in this study [31]. (g) Comparison of LUMO and HOMO levels of electrolyte components relevant for this study, obtained from DFT calculations. In the case of Li salts, multiple scenarios are considered, i.e., under Li coordination and decoordinated conditions.
(a) Normalized discharge capacity versus cycle number plots of cells with (a) LiFSI and LiTFSI, (b) LiBF4 and LiBF4 + FEC, and (c) LiBOB, LiBOB + FEC, LiDFOB, and LiDFOB + FEC electrolytes compared to LiPF6 and LiPF6 + FEC electrolytes. (d) Capacity endpoint plot for charge and discharge of cells in (c).
The addition of 2 wt% fluoroethylene carbonate (FEC) to the LiPF_6_ electrolyte results in an additional plateau in the voltage profile (Figure 1b) and decreases the reduction peak to ≈2.8 V cell voltage (≈0.95 V vs. Li|Li^+^; Figure 1e), pointing to FEC reduction and contribution to SEI formation [5]. It is notable that density functional theory (DFT) calculations show that the lowest unoccupied molecular orbital (LUMO) level of FEC is slightly higher than that of EC (Figure 1g). This emphasizes that the LUMO energy levels offer only a rough indication of reducibility, with compromised accuracy when the differences are small. A more accurate comparison of reducibility can be obtained by calculating the thermodynamic redox potentials, where the reduction potential for FEC is found to be higher than EC [36].
The FEC‐derived SEI consumes less charge during formation (Figure 3a) and enhances the cycle life from 183 to 305 cycles (until 70% state‐of‐health (SOH); Figure 2c; Figure 3b). It is also notable that FEC‐derived SEI results in longer cycle life compared to vinylene carbonate (VC)‐derived SEI when the respective additives are added in an equimolal amount (Figure S1b) [5, 19]. Moreover, FEC is oxidatively more stable than VC [37], as indicated by the lower level of highest occupied molecular orbital (HOMO) (Figure 1g).
(a) Charge consumption for SEI formation (0th cycle), (b) cycle number at 70% SOH, (c) 3rd cycle (0.5 C) discharge capacities, (d) ionic conductivity of selected electrolytes at a temperature range of −20–60°C, (e) voltage drop after the first 0.5 C CC‐CV charging, and (f) polarization growth of cells with respective electrolytes.
The larger EC reduction peak at ≈3.0 V with LiBF_4_ compared to LiPF_6_ points to higher amount of EC reduction (Figure 1e), [38] which even remains high when FEC is added, finally requiring more capacity for SEI formation (Figure 3a). Thus, the SEI formed under the presence of LiBF_4_ is not only less efficient (higher capacity consumption) but also less effective (lower cycle life) compared to that of LiPF_6_ (Figures 2b and 3b).
Figure 1c shows the 0th cycle cell voltage versus capacity profiles of cells with salts that are involved in SEI formation process, i.e., lithium bis(oxalato)borate (LiBOB) and lithium difluoro(oxalato)borate (LiDFOB), displaying plateaus from the anion reduction process. The corresponding differential capacity versus voltage plots in Figure 1f show peaks at 1.9 and 2.1 V, corresponding to ≈1.85 and 1.75 V versus Li|Li^+^, originating from the reduction of BOB^‐^ and DFOB^‐^ anions, respectively, hinting at the formation of anion‐derived SEI [27, 39]. The earlier reduction of BOB^‐^ anion is in agreement with the DFT calculation, where its LUMO level is calculated to be lower than that of DFOB^‐^ anion (Figure 1g). While LiBOB and LiBOB + FEC electrolytes consume less charge for SEI formation compared to EC‐reduction in a LiPF_6_ electrolyte (Figure 3a), LiDFOB and LiDFOB + FEC electrolytes consume even more. Nevertheless, in both cases, cycle life is enhanced (Figures 2c and 3b).
Interestingly, adding LiDFOB as an additive (instead of conducting lithium salt) in an equimolal amount as FEC to an LiPF_6_‐based electrolyte results in worse cycle life compared to LiPF_6_ + FEC electrolyte (Figure S4). Thus, the beneficial impact can be concluded to stem from continuous LiDFOB decomposition, i.e., continuous oxidation of SEI species over the course of inverse crosstalk.
The 3rd cycle discharge capacities of LiDFOB and LiBOB electrolytes at 0.5 C are shown in Figure 3c. Cells with LiBOB‐based electrolytes deliver less capacity compared to LiPF_6_ and LiDFOB, despite the similar ionic conductivities between LiBOB‐ and LiDFOB‐based electrolytes (Figure 3d), suggesting other factors to be more relevant [40, 41]. Indeed, cells with LiBOB‐based electrolytes experience larger voltage drops after the 0.5C constant current‐constant voltage (CCCV) charging step compared to those with LiPF_6_ and LiDFOB (Figure 3e), pointing to higher interphase resistances [42, 43]. These cells also have the largest polarization growth (Figure 3f), despite their higher cycle life compared to LiPF_6_, hinting at ALL rather than polarization as the predominant origin of fading. Adding FEC to the LiBOB electrolyte decreases the polarization growth, suggesting the vital role of fluorine for effective SEI and/or cathode electrolyte interphase (CEI) [44], which also might cause the superior performance of LiDFOB compared to LiBOB (Figure 3e).
Adding FEC to the previous LiBOB‐based electrolyte improves cycle life, even though the amount of active lithium loss (ALL) remains similar, as indicated by similar discharge endpoint capacity (Figure 2d). Hence, the improvement of cycle life rather stems from the generation of active Li via oxidation reactions at the cathode, as indicated by the larger charge endpoint capacity (Figure 2d), which finally compensates ALL and counteracts capacity fading. This oxidizable species likely stem from the anode, over the course of inverse crosstalk [45], likely in form of soluble SEI species that is generated during SEI reformation. These species partly dissolves into the electrolyte, diffuse to the cathode, get oxidized, and “recover” the formerly apparently lost electron (=active Li). Electrolytes with LiDFOB behave similarly, but with an even less ALL, thus prolonging the cycle life toward 542 cycles. With FEC, ALL is further decreased and cycle life prolonged toward 801 cycles.
Self‐discharge and Gassing Behavior
2.2
Figure 4a presents the evolution of open circuit voltage (OCV) during storage experiments for 720 h at 60°C. Notably, the self‐discharge of cells with LiPF_6_ can be effectively suppressed by adding FEC (Figure 4b), showing the lowest self‐discharge among the investigated electrolytes. As shown in Figure 4c, FEC addition also lowers irreversible capacity loss compared to LiPF_6_ at subsequent cycles at 0.5 C after the storage experiment, hinting at an effective SEI; though these values can also be affected by other factors, as discussed next [6, 46].
(a) Open‐circuit voltage (OCV) versus time plot during storage experiment at 60°C, (b) ratio of self‐discharged versus remaining discharge capacity after 720 h storage experiment at 60°C, (c) ratio of irreversible capacity loss versus reversible discharge capacity after 720 h storage experiment at 60°C of cells with LiPF6, LiPF6 + FEC, LiBOB, LiBOB + FEC, LiDFOB, and LiDFOB + FEC electrolytes. (d) Overcharge experiments of the aforementioned electrolytes in LNMO || Li cells at 20°C, (e) volume change after 720 h storage experiment at 60°C, (f) volume change after formation, and 200 charge/discharge cycles of cells with the aforementioned electrolytes.
In cells with LiBOB, self‐discharge is suppressed compared to LiPF_6_, even more when FEC is added to LiBOB (Figure 4a), but is still worse compared to LiPF_6_ + FEC (Figure 4b), likely due to the limited impact of FEC compared to BOB anion in the SEI formation of LiBOB + FEC electrolyte (Figure 1c). Interestingly, Figure 4c shows that the “irreversible” capacity losses are higher in cells with the LiBOB compared to the LiPF_6_ electrolyte, likely due to the enhanced electrolyte decomposition and the formation of resistive interphases, indicated by the large increase of average charge voltage after the storage experiments (Figure S5a).
According to Figure 4a,b, cells with the longest cycle life, i.e., those with LiDFOB‐based electrolytes, exhibit the most severe self‐discharge, even after adding FEC, but also the lowest irreversible capacity losses compared to those with LiPF_6_‐ and LiBOB‐based electrolytes (Figure 4c). Self‐discharge via anode, e.g., reactions with electrolyte, would consume active lithium and increase irreversible capacities. Hence, the high self‐discharge of LiDFOB‐based cells rather points to oxidation reactions on the cathode, which would even imply a generation of active lithium, thus capacity, in accordance to the literature‐known principle of prelithiation via sacrificing additives [46, 47]. Indeed, overcharge experiments of a high voltage spinel active material, i.e., LiNi_0.5_Mn_1.5_O_4_ (LNMO) [48, 49], show the lowest oxidative stability of LiDFOB‐based electrolytes (Figure 4d) [50], as seen by potential plateaus already at <4.5 V versus Li|Li^+^. It is notable that traces of electrolyte oxidation can proceed at even lower electrode potentials, especially at higher temperatures and longer durations, which is the case during the high‐temperature storage experiment in Figure 4a [51]. Although self‐discharge is generally decreased at 20°C, LiDFOB‐based electrolytes are still worse than cells with LiPF_6_‐based electrolytes (Figure S5b).
The lower oxidative stability of LiDFOB‐based electrolytes correlates with the larger produced gas volume, likely CO_2_ [52], compared to LiPF_6_‐based electrolytes after formation cycles, 200 charge/discharge cycles, and the 720 storage experiment (Figure 4e,f). The increased gassing is also observed in cells with LiBOB‐based electrolytes compared to LiPF_6_‐based electrolytes after the storage experiment (Figure 4e) and 200 cycles (Figure 4f), despite their likely similar oxidative stabilities as LiPF_6_‐based electrolytes. This can be attributed to impurities in the LiBOB salt [53], cathode surface‐dependent chemical reactivity [53, 54], not detectable via electrochemical methods [55, 56] and/or via enhanced reduction of BOB^‐^ on anodes side, leading to higher irreversible capacity loss (Figure 4c). As the CO_2_ generation during the formation cycle is lower compared to pre‐lithiated cells, CO_2_ can be concluded to have less influence on the SEI composition [57].
Interestingly, while FEC decreases gas volume in LiPF_6_‐based cells, likely due to minimized EC reduction (Figure 1b), which would generate CO, C_2_H_2_, etc. [58]. The addition of FEC to LiBOB and LiDFOB electrolytes increases gas production (Figure 4e). This is likely due to the strong influence of FEC in accelerating the ring opening of the oxalato functional group bonded to the boron center in BOB and DFOB anions under oxidative conditions, which produces more CO_2_ [59]. Furthermore, the dehydrofluorination of FEC may also form HF and vinylene carbonate (VC). The former can be reduced to form H_2_ and the latter can be oxidized to form CO_2_ [60, 61]. For LiPF_6_, FEC defluorination can be particularly problematic as a co‐solvent, i.e., large amounts [62]. As FEC is only added in small amounts to the LiPF_6_ electrolyte in this study, the drawback is relevant.
Electrode–Electrolyte Interphase Investigation
2.3
Figure 5 presents the scanning electron microscopy (SEM) images of SiO_ x ‐graphite anodes extracted from cells with different electrolytes after 200 charge/discharge cycles at 0.5 C (Figure S6). The orange circles indicate the SiO x _ particles, identified via energy‐dispersive X‐ray spectroscopy (EDX; Figure S7–S9). With the LiPF_6_ electrolyte, the SiO_ x _ particle in Figure 5a is obscured by thick deposits, which can be attributed to SEI buildup [5, 19], or possibly high surface area lithium (HSAL) deposits, as indicated by the large discharge endpoint slippage (similar to the case of high‐voltage LIBs) before the SEM images were taken [35]. In contrast, practically no thick deposits are observed in the case of LiPF_6_ + FEC (Figure 5d), as well as other electrolytes (Figure 5b,c,e,f). Interestingly, a clear SiO_ x _ surface is also observed with LiBOB‐based electrolytes, despite the large voltage drop and cell polarization seen in Figure 3e,f [44].
SEM images of negative electrodes extracted from cells with (a) LiPF6, (d) LiPF6 + FEC, (b) LiBOB, (e) LiBOB + FEC, (c) LiDFOB, and (f) LiDFOB + FEC electrolytes after 200 charge/discharge cycles. The orange circles indicate the SiO x particles.
Figure 6a,b present the relative atomic concentrations at the topmost layer of SEI and CEI obtained via X‐ray photoelectron spectroscopy (XPS). Compared to the LiPF_6_ electrolyte, the LiPF_6_ + FEC electrolyte forms an SEI that contains less inorganic LiF species and more polymeric species, which are both FEC decomposition products, in agreement with previous studies (Figure 6a) [63, 64, 65]. The FEC‐derived polymeric species is proposed to be crosslinked and exhibits an elastomeric property [66], offering higher robustness toward volume changes of SiO_ x _ particles. Consequently, it is assumed that H_2_O release from a parasitic reaction on SiO_ x _ surface is suppressed (Equation (1)) [67], thus decreasing LiF content from LiPF_6_ hydrolysis in the case of LiPF_6_ + FEC electrolyte (Figure 6a) [68, 69]. Furthermore, the suppression of LiPF_6_ decomposition also seems to influence CEI composition, as the LiF content is also lower with LiPF_6_ + FEC electrolyte (Figure 6b), showing SEI‐dependent CEI chemistry, known as “inverse crosstalk” [45, 70, 71].
Surface compositions of (a) SiO x ‐graphite and (b) NCM electrodes after formation cycles obtained via XPS. (c) Transition metal content of SiO x ‐graphite negative electrodes after 200 charge/discharge cycles obtained via ICP‐OES.
With LiBOB‐based electrolytes, a pronounced increase of organic species is observed in the SEI and CEI. In the SEI, lithium oxalates and oligomeric borates, which are LiBOB reduction products [27], contribute to the enhancement of the organic species, indicated by the increase of O—C=O species and the high B—O content, respectively (Figure 6a) [27, 44]. Surprisingly, despite the absence of fluorine in LiBOB, LiF is detected in the SEI and CEI for LiBOB electrolyte (Figure 6a,b), likely due to the chemical instability of polyvinylidene fluoride (PVDF) binder, which results in LiF formation and its diffusion toward the anode [72]. Introducing FEC to the LiBOB electrolyte slightly increases LiF, Li_2_CO_3_, and B—O content on the SEI derived from LiBOB + FEC electrolyte (Figure 6a), which can be correlated to the pronounced decrease of cell polarization growth (Figure 3f) and consequently enhancement of cycle life (Figure 3b).
With LiDFOB‐based electrolytes, a greater amount of LiF is detected in the SEI compared to those with LiBOB‐based electrolytes, which can be attributed to LiF generation during DFOB anion reduction [27]. Adding FEC to this electrolyte reduces LiF content and promotes polymeric species formation (Figure 6a), which may be attributed to the FEC‐induced polymerization of BF_2_ ^‐^ radicals from LiDFOB reduction that does not release LiF [59]. The resulting polymer likely enhances the SEI robustness compared to the “standard” crosslinked oligomeric borates which are proposed to be present in LiDFOB‐derived SEI [27].
The amounts of transition metals (TMs) deposited on the SiO_ x ‐graphite surface extracted from cells with various electrolytes after 200 charge/discharge cycles are presented in Figure 6c. It is notable that the different salt systems result in different TM deposit ratios. This can be attributed to the different TM coordination environment under the presence of different anions, which can affect cation transport properties [35, 73]. Regardless of the ratios, the largest amount of TM deposits is present on anodes from cells with LiPF_6 electrolyte, though adding FEC markedly reduces the amount of TM deposits. This is possibly due to the more robust FEC‐derived SEI that can afford better protection for SiO_ x _ surface against HF attack from the electrolyte, which may initiate an autocatalytic reaction that produces more HF and TM dissolution from NCM [67]. For cells with LiBOB and LiDFOB‐based electrolytes the lower amounts of TM deposits may stem from protective CEIs and/or robust SEIs that can protect the cathode from HF attack and prevent HF formation, respectively [50, 52, 74]. Adding FEC aids the protection of BOB‐derived SEI, lowering the amount of TM deposits on anodes for cells with LiBOB + FEC compared to LiBOB electrolyte. Employing LiDFOB electrolyte further suppresses the amount of TM deposits compared to LiPF_6_ and LiBOB electrolytes, pointing to improved SiO_ x _ protection by DFOB‐derived SEI. Nevertheless, adding FEC to the LiDFOB electrolyte slightly increases TM deposit amounts, possibly due to the dehydrofluorination reaction of FEC that produces HF [62].
Electrolyte Analysis
2.4
Figure 7a presents the gas chromatogram of electrolytes extracted after formation cycles. The LiPF_6_ electrolyte exhibits peaks that are identified to be dimethyl carbonate (DMC), diethyl carbonate (DEC), FEC, oligomers such as dimethyl‐2,5‐dioxahexane carboxylate (DMDOHC), ethylmethyl‐2,5‐dioxahexane carboxylate (EMDOHC), diethyl‐2,5‐dioxahexane carboxylate (DEDOHC), which are referred to as oxahexane carboxylates (OHCs) [5, 75, 76, 77]. DMC and DEC are produced from trans‐esterification reactions initiated by EMC reduction [75], whereas OHCs are generated from oligomerization involving EC reduction products initiated by EC ring opening [76, 77]. The presence of these compounds is an indirect proof of gas‐generating solvent reduction in the first charging step [78], as seen in the differential capacity plots in Figure 1c, which can be suppressed with FEC [5]. These products are also absent in electrolytes from cells with LiBOB‐ and LiDFOB‐based electrolytes.
Gas chromatogram of electrolytes extracted after (a) formation and (c) 200 charge/discharge cycles obtained via GC–MS. (b) Magnification of FEC peak in (a). In (a), the asterisk signs () indicate a measurement cut for dichloromethane, EMC, and EC. In (c), the two () and three () asterisk signs indicate a system peak of the GC–MS device and an intensity reduction due to switching from scan to selected ion monitoring (SIM) mode, respectively.
Figure 7b indicates a small amount of residual FEC in LiPF_6_ + FEC electrolyte after formation cycles, while in LiBOB + FEC electrolyte, the FEC amount appears to be higher. This indicates lower FEC consumption for SEI formation in the latter case, due to an anion‐dominated, i.e., BOB‐derived, SEI (Figure 1c). Interestingly, despite the similarly pronounced LiDFOB reduction during SEI formation, only a low amount of FEC is seen in LiDFOB + FEC electrolyte, possibly due to the participation of FEC in a polymerization reaction that involves BF_2_ ^‐^ radicals from LiDFOB reduction [59, 79]. Finally, after 200 charge/discharge cycles, similar amounts of FEC are detected in LiBOB + FEC and LiDFOB + FEC electrolytes (Figure 7c), which are higher compared to LiPF_6_ + FEC, pointing to higher FEC consumption for FEC‐derived SEI compared to anion‐derived SEI. At the same time, more DEC (from *trans‐*esterification reaction initiated by EMC reduction) is detected in LiPF_6_ + FEC, suggesting more reactions with electrolyte, i.e., less robust SEI compared to oxalato‐derived SEI.
Conclusion
3
In this work, the impact of various Li salts in ethylene carbonate (EC)/ethyl methyl carbonate (EMC; 3:7 by wt.) is investigated on application‐relevant metrics within LiNi_0.5_Co_0.2_Mn_0.3_O_2_ (NCM 523; single crystal) || 10%SiO_ x _‐graphite multi‐layer pouch cells. Si‐based active materials promise improved gravimetric/volumetric energies of Li ion batteries (LIBs) compared to state‐of‐the‐art (SOTA) anodes, which are based on graphite only (Figure 8a), but suffers from larger active lithium loss (ALL) during charge/discharge cycling over the course of solid electrolyte interphase (SEI) formation (Figure 8b), in particular due to the relative high volume changes of Si (Figure 8c). Although lithium bis(fluorosulfonyl)imide (LiFSI) and lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) are claimed to have a beneficial impact on Si anodes in Si || Li cells [21], they are not suitable in full cells as they trigger anodic Al dissolution from the cathode current collector, thus shortening the end‐of‐life (EOL) of the cells.
(a) Adding SiO x to graphite‐based anode can increase the gravimetric/volumetric energy. (b) However, the accompanying volume‐stress‐reasoned SEI re‐formation consumes active Li and decreases the capacity. (c) The reaction of SiO x and LiPF6 can trigger additional failure cascades by HF formation and additionally decrease capacity and cycle life. The passivation of SiO x can be an effective measure, e.g., via adding FEC. (d) While cells with LiBOB are fading over the course of ALL, the addition of FEC does not decrease ALL, but increases amount of soluble SEI, which can partly diffuse to the cathode (inverse crosstalk) and generate active lithium (AL) via oxidation, this way “reactivating” the apparently “wasted SEI” and lost capacity, consequently prolonging cycle life.
Cells with LiBF_4_‐ and LiPF_6_‐based electrolytes achieve an EOL (=70% capacity retention) at 125 and 183 cycles, respectively. Adding fluoroethylene carbonate (FEC) as an additive (2 wt%) to the LiPF_6_‐based electrolytes prolongs the EOL to ≈305 cycles. This can be correlated with an FEC‐derived solid electrolyte interphase (SEI), as indicated by an FEC reduction peak thatsuppresses EC reduction peak in the differential capacity plots, and lower FEC content in the electrolyte after formation cycles by means of gas chromatography coupled with mass spectrometry (GC–MS). Compared to other well‐known electrolyte additives such as vinylene carbonate (VC), the LiPF_6_ + FEC (=STD) electrolyte formulation can be regarded as SOTA, as among others, FEC is also more oxidatively stable [37], but especially because it is more effective in prolonging the EOL (250 cycles only for VC).
The effectiveness of FEC‐derived SEI can be correlated with a more polymeric and flexible nature of the SEI, as concluded from the enhanced organic‐based species by means of X‐ray photoelectron spectroscopy. This decreases ALL, irreversible capacity, self‐discharge, resistance build‐up, and even gassing. The lower gassing can be correlated with suppressed formation of oligomers over the course of suppressed EC reduction, which are also known to initiate gassing.
Additionally, an effective passivation of SiO_ x _ can be concluded to suppress chemical reactions with LiPF_6_, which can form HF and trigger additional failure cascades, i.e., transition metal (TM) dissolution from the cathode, electrode crosstalk, TM‐reasoned local resistance increase on anode, risk of Li plating, shuttling, additional gassing, and irreversible capacity (Figure 8c). In fact, without FEC, a severe amount of TM deposit is observed on the anode, detected by inductively‐coupled plasma optical emission spectroscopy (ICP‐OES). Interestingly, thick deposits are predominantly observed on SiO_ x _ as shown by scanning electron microscopy (SEM). This can be related to larger volume change, more SEI, thus intrinsically higher resistance of SiO_ x _ compared to graphite, higher overpotentials during charging, and finally more reductive conditions on SiO_ x _. This results in more side reactions and deposit amount, which can be attributed to a mixture of TMs, high surface area lithium (HSAL), and SEI. This likely enhances inverse crosstalk, e.g., shuttling of SEI products to the cathode [45], as seen by the sudden and severe increase in charge endpoint slippage after ≈100 charge/discharge cycles; a hint to parasitic oxidation reactions at the cathode, known from high voltage LIBs [35]. Given the enormously higher discharge endpoint slippage, the ALL is more severe, and in consequently, the cell rapidly fades.
Lithium bis(oxalato) borate (LiBOB) as an alternative to LiPF_6_ is predominantly more involved in SEI formation because LiBOB is reduced earlier compared to EC. The EOL is slightly prolonged for cells with LiBOB—compared to LiPF_6_‐based electrolyte (242 vs. 183 cycles), but the high initial resistance, as well as resistance rise throughout cycling, raises the risk of Li plating (SEM), finally leading to ALL‐reasoned fading, safety concerns, and gassing. Though the FEC additive in LiBOB‐based electrolyte is only partly involved in SEI formation, as it remains detectable (i.e., partly “unreacted”) even after 200 cycles (GC–MS), it still increases the F‐content in the SEI and prolongs the EOL toward 392 cycles. This can be attributed to decreased resistance and enhanced effectiveness of SEI, as seen by the suppressed voltage hysteresis and decreased irreversible capacity during self‐discharge experiments. Interestingly, the increased charge endpoint slippage, an indication for oxidative reactions at the cathode, hints at more intense inverse crosstalk phenomena, likely via diffusion of SEI products to the cathode, which is concluded to subsequently generate active Li via oxidation and recover the apparently lost capacity, consequently prolonging the EOL (Figure 8d). The altered SEI and enhanced inverse crosstalk can be correlated with the higher amount of evolved gas, which overall renders this electrolyte still disadvantageous for application.
Lithium difluoro(oxalato)borate (LiDFOB) prolongs cycle life toward cycle no. 542, likely due to the DFOB‐derived SEI, though the cells suffer from the highest self‐discharge (SD) among the herein‐investigated salts. Nevertheless, the SD is more reversible as the full capacity can be regained after SD experiments. Hence, DFOB^‐^ oxidation and/or inverse crosstalk is concluded to SD the cathode, i.e., processes which do not suffer from ALL, rather generate AL. While DFOB^‐^ oxidation is likely due to the relatively low oxidation potential of the LiDFOB‐based electrolyte (≈4.4 V vs. Li|Li^+^), especially when combined with high temperature (60°C) and long duration (720h) conditions, the inverse crosstalk is concluded to be more pronounced during charge/discharge cycling at ambient conditions. The charge and discharge endpoint slippage is decreased by adding FEC, hinting at less inverse crosstalk but also at less ALL. Finally, the herein‐shown outstanding cycle life of cells with the LiDFOB + FEC electrolyte can be concluded to stem from the more effective SEI (low ALL) and an effective “self‐healing” mechanism (Figure 8d) originating from DFOB^‐^ oxidation and/or inverse crosstalk that adds active lithium to the cell and compensates ALL. However, similar to LiBOB, the gassing renders this electrolyte formulation impractical for cells with Si‐containing anodes paired with NCM‐based cathodes. This formulation, however, might differ for cells with lower charge cut‐off voltage, e.g., LiFePO_4_ || SiO_ x _‐Gr.
To conclude, this study offers an additional perspective for electrolyte design of Li ion batteries with Si‐containing anodes. While it is well‐known that the electrolyte should be able to form an ionically conductive SEI that is respectively flexible to withstand the Si volume change and decrease ALL, the SEI should be partially dissolvable, as the apparently lost species can recover the AL via oxidation over the course of inverse crosstalk, thus prolonging cycle life by compensating ALL. However, during storage, this can worsen self‐discharge, though in a reversible manner, requiring respective optimization. Furthermore, gassing should be avoided by a respective design of electrolyte, and the corresponding dissolved species and oxidation products.
Experimental Section
4
Electrolyte Formulation
4.1
Electrolyte mixtures were prepared by mixing 1 M (unless otherwise stated) of conducting salts, i.e., LiPF_6_ (battery grade; E‐Lyte Innovations GmbH), LiBF_4_ (battery grade; E‐Lyte Innovations GmBH), LiB(C_2_O_4_)2 (lithium bis(oxalato)borate; LiBOB; Sigma Aldrich), LiBF_2_(C_2_O_4_) (lithium difluoro(oxalato)borate; LiDFOB; 99%; Abcr GmbH), lithium bis(trifluoromethane)sulfonimide (LiTFSI; 99.5%; Solvionic), and lithium bis(fluorosulfonyl)imide (LiFSI; 99.5%; Solvionic), in an ethylene carbonate (EC) and ethyl methyl carbonate (EMC) solution mixture (3:7 by weight; battery grade; E‐Lyte Innovations GmbH). Electrolytes containing fluoroethylene carbonate (FEC; battery grade; BASF) were prepared by adding 2 wt% FEC to the previously mixed electrolytes. All electrolytes were prepared in an argon‐filled glovebox (MBraun) with H_2_O and O_2_ levels below 1 ppm.
Density Functional Theory (DFT) Calculations
4.2
All calculations were carried out within the framework of the DFT [80, 81], using the ORCA software package [82, 83, 84]. Frontier orbital energies were obtained with the 6‐311++G** Pople‐style basis set, which includes double polarization functions to enhance flexibility and double diffuse functions to accurately describe species with extended electron densities, such as anions. The B3LYP functional was employed to model the exchange–correlation potential [85, 86, 87, 88]. To account for solvation effects, the conductor‐like continuum polarization model (C‐PCM) as applied [89], with all molecules simulated in ethylene carbonate (ε = 90.5) [90]. Prior to the energy evaluations, all structures were fully optimized with respect to their geometry. Frontier orbitals and molecular structures were visualized using Avogadro [91].
Cell Setup
4.3
Wounded multi‐layer NCM 523 || SiO_ x ‐graphite pouch cells (LiFun Technologies) balanced for a UCV of 4.3 V, with a nominal capacity of 200 mAh, were used for electrochemical investigations [5, 19]. The positive electrode has a mass loading of 16.5 mg cm^−2^, consisting of 94 wt% NCM 523 single‐crystalline powder, 4 wt% carbon black, and 2 wt% poly(vinylidene difluoride) binder (PVDF). The negative electrode has a mass loading of 8 mg cm^−2^, consisting of 94.8 wt% active material (90 wt% artificial graphite, 10 wt% SiO x _), 1.4 wt% carbon black, 1.3 wt% carboxymethylcellulose (CMC), and styrene‐butediene rubber (SBR). Prior to electrolyte filling, the cells were dried at 80°C under reduced pressure (<0.05 mbar) for 12 h. Afterward, the cells were filled with 0.75 mL of electrolytes and sealed by a heat‐crimper at 165°C for 5 s at a relative pressure of −80 kPa using a vacuum sealer (GN‐HS200V, Gelon Lib Group). For electrochemical investigation, a custom cell holder capable of applying 2 bar of pressure were used [19].
Overcharge experiments were carried out in three‐electrode setup in T‐cells (Swagelok). LiNi_0.5_Mn_1.5_O_4_ (LNMO; >99%; Sigma Aldrich)‐based composite electrode (Ø12 mm) with an areal capacity of ≈0.36 mAh cm^−2^, made up of 90 wt% active material, 5 wt% carbon black (Super C65, Imerys Graphite & Carbon), and 5 wt% PVDF binder (Solef 5130, Solvay) coated on aluminum foil, was used as the working electrode. For the counter and reference electrodes, Ø12 and Ø8 mm lithium foil (500 μm; Gelon Lib Group) were used, respectively. The working and counter electrodes were separated by three layers of Ø12 mm polypropylene (PP) fiber separators (FS2190; Freudenberg) soaked in 120 μL of electrolyte, and the reference electrode was separated by two layers of the same separator soaked in 120 μL of electrolyte. Pouch cells were sealed and T‐cells were assembled in a dry room atmosphere with a dew point of at least −50°C (0.16% relative humidity).
Electrochemical Investigations
4.4
Electrochemical investigation of NCM 532 || 10SiO_ x _‐Gr pouch cells was performed using a Maccor 4000 battery testing system. After sealing, the cells were charged to 1 V and held at this voltage for 20 h to ensure adequate wetting. For solid electrolyte interphase (SEI) formation, the cells were charged to 3.5 V at a constant current (CC) of 10 mA (≈0.05 C), held at a constant voltage (CV) of 3.5 V for 1 h, and then discharged to 2.8 V with a CC of 10 mA (≈0.05 C). This cycle is considered cycle zero. The cells then underwent two charge/discharge cycles with upper cut‐off voltages (UCVs) of 4.3 V. For these two cycles, the cells underwent CC charging step at 40 mA (≈0.2 C), followed by a CV step until the current drops to <4 mA (≈0.02 C), and finally CC discharge step at 40 mA (≈0.2 C) to 2.8 V. After the formation cycles at 20 C, the cells were subjected to cycling test until reaching 70% state of health (SoH) at 20 C, 200 charge/discharge cycling at 20 C, or high‐temperature storage experiment at 60 C. For cycling tests, the cells were subjected to CC charging step at 100 mA (≈0.5 C), followed by a CV step until the current drops to <4 mA (≈0.02 C) or 300 mAh (≈1.5 C) of capacity is reached, and finally CC discharge step at 100 mA (≈0.5 C). At the last cycle, the cells were further discharged with a CV step at 2.8 V until the current dropped to <4 mA (≈0.02 C). For the storage experiment, the cells rested for 6 h at 60 C after formation cycles, charged with CC mode at 100 mA (≈0.5 C) to 4.3 V, followed by a CV step until the current drops to <4 mA (≈0.02 C). The cells were then rested for 720 h, and their open‐circuit voltage (OCV) was measured. Finally, the cells were discharged with CC mode at 100 mA (≈0.5 C) to 2.8 V. Afterward, the pouch cells were removed from the battery tester for gas volume measurement. Then, the pouch cells were reconnected to the battery tester at 60°C and subjected to one charge/discharge cycle at 20 mA (≈0.1 C) followed by four cycles at 100 mA (≈0.5 C), with a CV step at the top of charge until the current dropped below 4 mA (≈0.02 C) to initiate the discharge. The 0.5 C cycles were used to determine the reversible discharge capacity after the storage experiments. At least two pouch cells per testing parameter were used to validate reproducibility.
Overcharge experiments were carried out at room temperature, after letting the T‐cells rest for 3 h after cell assembly to allow for adequate electrolyte wetting. The cells were then charged with a CC mode (0.06 C, 1 C defined as 145 mAh g^−1^), and the potential of the working electrode was measured. Charging is terminated when 30 h of measurement time or 6 V is reached. At least two T‐cells per electrolyte mixture were used to validate reproducibility.
Gas Volume Measurements
4.5
The volume of gas formed inside the pouch cell was estimated by the Archimedes In Situ Gas Analyzer (AISGA) method [92]. The pouch cells were submerged in deionized water by hanging them to a balance (S256 Low Range – Force Sensors SMD3277−010, 10 g, Strain Measurement Device) with a wire hook through a ≈1 mm hole at the sealed edge of the pouch cell. The DAQami v4.2.1 software (Measurement Computing MC, 10 Commerce Way, Norton, MA 0 2766, USA) was used to acquire data and generate signals.
Electrolyte Conductivity Measurement
4.6
Electrolyte conductivity experiments were carried out under an inert atmosphere inside a glovebox (MBraun, H_2_O and O_2_ <1 ppm). Conductivity cells were filled with the various electrolyte formulations as previously described [93]. Cell constants were determined by using a 0.01 M solution of KCl in H_2_O at 20°C (VWR, known conductivity of 1.276 mS cm^−1^), averaged over five measurements. Disposable 2 mL Eppendorf Safe‐Lock Tubes were used as sample containers, and each was filled with 750 μL of electrolyte. Impedance measurements were conducted on a Metrohm Autolab/M204 potentiostat/galvanostat with 12 channels and an 8‐channel multiplexer for a total of 96 channels in the frequency range of 50–20 000 Hz using in‐house developed electrodes. The conductivity cells were placed in a temperature chamber (Memmert TTC256, 0.1°C temperature setting accuracy) and each temperature was held for 2 h prior to measurement for equilibration. The ionic conductivity of the electrolytes was measured in the temperature range of −30°C to 60°C in 10°C steps. Impedance spectra were fitted using a model specified with set parameters for resistors Rs and Rp, as well as for the constant phase element (CPE) with the Metrohm Nova software. Fitting was carried out after each additional measuring point by using the fitting model Rs(CPE − Rp). Electrolyte conductivity values were obtained from the quotient of the cell constant and the determined electrolyte resistance.
Scanning Electron Microscopy (SEM)
4.7
SEM (Carl Zeiss AURIGA; Carl Zeiss Microscopy GmbH). was employed to examine the surface morphology of SiO_ x _ graphite electrodes. EDX was performed at an accelerating voltage of 6 kV using an X‐MaxN 80 mm^2^ EDX detector (Oxford Instruments). INCA software was used to assess the spectra (Oxford Instruments). Cells were disassembled in a dry room atmosphere with a dew point of at least −50°C (0.16% relative humidity). Prior to SEM imaging, 1 mL of dimethyl carbonate (DMC) was used to wash the electrode. To avoid moisture contact, samples were transferred using an air‐tight sample holder to the SEM analysis chamber.
Gas Chromatography‐Mass Spectrometry (GC–MS)
4.8
GC–MS measurements were carried out using Shimadzu QP2010 Ultra single quadrupole (SQ), equipped with AOC‐5000 autosampler and standard non‐polar Supelco SLB‐5 ms column (30 m × 0.25 mm × 0.25 µm, Sigma Aldrich). 1 μl of diluted electrolyte sample (1:100 in dichloromethane; DCM) was injected to a heated GC injector (250°C). Helium (6.0, Westfalen gas, Germany) was used as carrier gas with a column flow of 1.16 mL min^−1^, purge flow of 3 mL min^−1^, and a split of 1:100. The event time in MS scan mode was 0.2 s. Other GC and MS parameters were set according to Grützke and Mönninghoff et al. [94, 95]. To increase the sensitivity for samples after 200 cycles, a split of 1:10 was used and the MS method was modified using selected ion monitoring of 29, 62, and 106 m/z (FEC characteristic fragments) in a time range of 6.5–7.2 min. Other parameters were set as described above.
Inductively Coupled Plasma‐Optical Emission Spectroscopy (ICP‐OES)
4.9
Two 12 mm electrodes were extracted from two different cells for each electrolyte formulation. Electrode samples were digested in reversed aqua regia using a microwave system (Multiwave 7000, AntonPaar, Graz, Austria) prior to analysis by ICP‐OES. The measurements were performed on an ARCOS (Spectro Analytical Instruments GmbH, Kleve, Germany) instrument, equipped with a scott‐style spray chamber, a crossflow nebulizer and an axial positioned plasma torch. The following emission lines were observed during analysis: Li (670.780 nm), Ni (221.648, 231.604, 232.003 nm), Co (228.616, 237.862, 238.892 nm), Mn (257.611, 259.373 nm), and Cu (224.700, 324.754, 327.396 nm). The method and further parameters were adapted from Vortmann‐Westhoven et al. and Evertz et al. [96, 97].
Supporting Information
Additional supporting information can be found online in the Supporting Information section. Supporting Fig. S1: (a) Voltage vs. time profile of 200 mAh pouch ells with LiPF_6_, LiFSI, and LiTFSI based electrolytes during the formation cycles. Note that cells with electrolytes containing LiTFSI do not deliver any discharge capacity after the first charge to 4.3 V. It is shown previously that aluminum current collector dissolution causes cathode foil disintegration [1], which results in electrical disconnection. (b) Normalized discharge capacity vs. cycle number plot of cells with the LiPF_6_, LiPF_6_ + FEC, and LiPF_6_ + VC based electrolytes. Here, VC is added in an equimolal amount to FEC: of 2 wt%. Supporting Fig. S2: Nominal discharge capacity vs. cycle number plots of cells with (a) LiBF_4_ and LiBF_4_ + FEC, and (b) LiBOB, LiBOB + FEC, LiDFOB, and LiDFOB + FEC based electrolytes compared to LiPF_6_ and LiPF_6_ + FEC based electrolytes. Supporting Fig. S3: (a) Photographs of electrolytes with 0.6, 0.8, and 1 M of LiBOB, showing the limited solubility of LiBOB at concentrations >0.6 M. (b) Normalized discharge capacity vs. cycle number plots of cells with 0.6 (+FEC), 0.8, and 1 M of LiBOB compared to the electrolyte with 1 M LiPF_6_ (+FEC). (c) Differential capacity vs. voltage plots of electrolytes with 0.6, 0.8, and 1 M of LiBOB compared to the 1 M LiPF_6_ based electrolyte. Supporting Fig. S4: (a) Normalized discharge capacity vs. cycle number and (b) differential capacity vs. voltage plots of cells with 0.19 mmol_additive_ g^−1^ electrolyte added to 1 M LiPF_6_ in ethylene carbonate (EC)‐ethyl methyl carbonate (EMC) solvent mixture (3:7 by weight). Note that 0.19 mmol_additive_ g^−1^ electrolyte corresponds to 2 wt% FEC and the LiPF_6_ + LiDFOB‐FEC electrolyte contain an equimolar amount of LiDFOB and FEC. Supporting Fig. S5: (a) Average charge voltage before and after storage experiment at 60°C for 720 h and (b) open‐circuit cell voltage (OCV) vs. time plot during storage experiment at 20°C of cells with LiPF_6_, LiPF_6_ + FEC, LiDFOB, and LiDFOB + FEC electrolytes. Supporting Fig. S6: Photographs of SiO_ x ‐graphite negative electrodes extracted from cells with (a) LiPF_6, (b) LiPF_6_ + FEC, (c) LiBOB, (d) LiBOB + FEC, (e) LiDFOB, and (f) LiDFOB + FEC electrolytes after 200 charge/discharge cycles. The electrodes are partially damaged due to a prior centrifuging process to extract the electrolytes. Supporting Fig. S7: SEM images and their corresponding EDX map of SiO_ x ‐graphite negative electrodes extracted from cells with (a) LiPF_6, (b) LiPF_6_ + FEC electrolytes after 200 charge/discharge cycles. Supporting Fig. S8: SEM images and their corresponding EDX map of SiO_ x ‐graphite negative electrodes extracted from cells with (a) LiBOB, (b) LiBOB + FEC electrolytes after 200 charge/discharge cycles. (c) A representative EDX spectrum showing the overlapping peaks of boron and carbon, suggesting that the boron map in (a) and (b) may not be accurate. Supporting Fig. S9: SEM images and their corresponding EDX map of SiO x ‐graphite negative electrodes extracted from cells with (a) LiDFOB, (b) LiDFOB + FEC electrolytes after 200 charge/discharge cycles. Supporting Fig. S10: SEM images and their corresponding EDX map of areas on SiO x _‐graphite negative electrodes with white deposits shown in Figure S6. The high oxygen and fluorine content suggest that the deposits may be lithium deposits.
Author Contributions
Anindityo Arifiadi: conceptualization (lead), investigation (lead), methodology (lead), writing – original draft (lead). Jaroslav Minar: methodology (equal). Ankita Das: investigation (equal), Linus Voigt: investigation (equal). Dominik Voigt: investigation (lead). Marc Vahnstiege: investigation (equal). Julius Buchmann: investigation (equal). Feleke Demelash: investigation (equal). Peng Yan: investigation (supporting), Simon Wiemers‐Meyer: methodology (equal), writing – review and editing (equal). Sascha Nowak methodology (supporting). Frank Glorius: conceptualization (supporting), writing – review and editing (equal). Martin Winter: conceptualization (supporting); writing – review and editing (equal). Johannes Kasnatscheew: conceptualization (lead), supervision (lead), writing review and editing (lead).
Funding
This study was supported by Bundesministerium für Bildung und Forschung.
Supporting information
Supplementary Material
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1R. Schmuch , R. Wagner , G. Hörpel , T. Placke , and M. Winter , “Performance and Cost of Materials for Lithium‐Based Rechargeable Automotive Batteries,” Nature Energy 3 (2018): 267.
- 2R. Teki , M. K. Datta , R. Krishnan , et al., “Nanostructured Silicon Anodes for Lithium Ion Rechargeable Batteries,” Small 5 (2009): 2236.19739146 10.1002/smll.200900382 · doi ↗ · pubmed ↗
- 3P. Meister , H. Jia , J. Li , R. Kloepsch , M. Winter , and T. Placke , “Best Practice: Performance and Cost Evaluation of Lithium Ion Battery Active Materials with Special Emphasis on Energy Efficiency,” Chemistry of Materials 28 (2016): 7203.
- 4C. K. Chan , H. Peng , G. Liu , et al., “High‐Performance Lithium Battery Anodes Using Silicon Nanowires,” Nature Nanotechnology 3 (2008): 31.10.1038/nnano.2007.41118654447 · doi ↗ · pubmed ↗
- 5A. Ghaur , C. Peschel , I. Dienwiebel , et al., “Effective SEI formation via phosphazene‐based electrolyte additives for stabilizing silicon‐based lithium‐ion batteries,” Advanced Energy Materials 13 (2023): 2203503.
- 6F. Holtstiege , A. Wilken , M. Winter , and T. Placke , “Running Out of Lithium? A Route to Differentiate between Capacity Losses and Active Lithium Losses in Lithium‐Ion Batteries,” Physical Chemistry Chemical Physics 19 (2017): 25905.28926044 10.1039/c 7cp 05405 j · doi ↗ · pubmed ↗
- 7S. Chae , S.‐H. Choi , N. Kim , J. Sung , and J. Cho , “Integration of Graphite and Silicon Anodes for the Commercialization of High‐Energy Lithium‐Ion Batteries,” Angewandte Chemie International Edition 59 (2020): 110.30887635 10.1002/anie.201902085 · doi ↗ · pubmed ↗
- 8S. Solchenbach , C. Tacconis , A. Gomez Martin , et al., “Electrolyte Motion Induced Salt Inhomogeneity – a Novel Aging Mechanism in Large‐Format Lithium‐Ion Cells,” Energy & Environmental Science 17 (2024): 7294.