Quadrupled Cycle Life of High‐Voltage Nickel‐Rich Cathodes: Understanding the Effective Thiophene‐Boronic Acid‐Based CEI via Operando SHINERS

Increasing the cell voltage of lithium‐ion batteries (LIBs) is a straightforward approach to increasing their capacity and energy density. However, state‐of‐the‐art cathode materials like LiNixMnyCo1‐x‐yO2 (NMC) suffer from severe failure mechanisms at high operating voltages, significantly degrading the performance and cycle life of the cells. Notably, an effective cathode electrolyte interphase (CEI) mitigates these failure mechanisms. Nevertheless, a deep understanding of the formation mechanisms and properties of the CEI is necessary to tailor effective interphases. This study introduces a promising electrolyte additive for high operating voltage NMC811||graphite cells. Implementing an optimized concentration of 3‐thiophene boronic acid (3‐Thp‐BOH) significantly enhances the cells' performance and reduces capacity fading, resulting in a quadrupled cycle life and a six‐times higher accumulated specific energy. Operando shell‐isolated nanoparticle‐enhanced Raman spectroscopy (SHINERS) is employed to shed light on the formation mechanism and molecular composition of CEI during cell operation, proving that the presence of the additive results in the formation of a complex 3‐Thp‐BOH‐based polymeric CEI on the NMC811 surface. The CEI investigation is additionally supported by scanning electron microscopy and energy dispersive X‐ray analysis and highly accurate quantum chemistry modeling of the suggested polymerization mechanisms.


Introduction
Lithium-ion batteries (LIBs) are, in terms of specific capacity and energy density, the most effective energy-storage system currently available and are already widely used in a broad field of applications, from portable electronics to electric vehicles (EVs) or grid storage. [1][2][3][4][5][6] However, especially for EVs, critical parameters of LIBs like energy density, safety, and cost have to be improved to meet increasing market demands. [2,3,7] Two straightforward approaches to achieving high specific energies are to increase the operating voltage of LIB cells and the implementation of high specific capacities. [2,3,8] Nonetheless, nextgeneration high-capacity cathode materials, such as Ni-rich LiNi x Mn y Co 1-x-y O 2 (NMC, x ≥ 0.8), suffer from thermodynamic instabilities during operation at high cell voltages. This may lead to oxygen evolution and the dissolution of transition metals into the electrolyte, resulting in significantly reduced cell cycle life and performance. [3][4][5]7,9,10] In addition, oxidative degradation of the electrolyte occurs at high potentials, resulting in the loss of active lithium and the formation of a decomposition layer on the cathode surface known as cathode electrolyte interphase (CEI). [2,3,5,8,9] The CEI increases the cells' internal resistance and limits the Li-ion conductivity, greatly influencing their electrochemical performance and cycle life. [3,5,7] Significant research efforts have already been devoted to address these challenges. A promising approach is the use of coating agents, which modify the surface of the cathode active material to suppress the transition metal dissolution and decomposition reactions of the electrolyte. [3,8,11,12] Another parallel approach takes advantage of implementing film-forming additives to ensure the formation of an effective CEI, which mitigates electrolyte decomposition and inhibits transition metal exposure to the electrolyte while securing facile Li-ion transfer at the interphase. [3,4,6,8,[13][14][15][16] Promising studies have already been published by Park et al. and Dong et al., who introduced diphenyl diselenide and lithium difluoro(oxolato)borate as CEI-forming additives for high-voltage cells with NMC811 cathodes, suppressing the www.advancedsciencenews.com www.advenergymat.de failure mechanisms related to transition metal dissolution. [17,18] However, successful and target-driven implementation of CEIforming additives requires a deep understanding of the interfacial processes, reaction potentials, composition, and aging mechanisms of the formed CEI. [9] In an ideal case, understanding these properties is realized by operando investigations under real cell operating conditions, ruling out the external influence of, for example, experimental cell configuration or interphase contamination/destruction during sample preparation. Unfortunately, suitable techniques to perform operando investigations and, therefore, studies, especially on the investigation of the CEI, are scarce. Liu et al. used scanning electrochemical microscopy in combination with X-ray photoelectron spectroscopy (XPS) to characterize the CEI on LiMn 2 O 4 electrodes in aqueous electrolytes. In this study, the formation of an amorphous CEI layer comprising of -(CF) n -species as well as small amounts of Li 2 CO 3 , MnF 2, and LiF was shown. [19] Attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy was employed to investigate the surface layer on NMC electrodes in non-aqueous electrolytes by de Villers et al. Based on the spectroscopic results, the formation of a carbonate-species-dominated surface layer was proposed. [20] Using a similar approach, Kanamura et al. employed FTIR spectroscopy to characterize the CEI on LiCoO 2 electrodes in contact with a propylene carbonate-based electrolyte. In this work, organic carboxylates were identified as the main components of the surface layer. [21] Raman spectroscopy is a promising technique for conducting operando investigations and to overcome several downsides of other interphase characterization techniques. [22][23][24][25] Conventional Raman spectroscopy is well-known in modern battery research. To this point, it has mainly been used to characterize bulk processes of electrode materials and electrolytes. [22,26] However, conventional Raman spectroscopy fails to characterize the formation of (nanometric thin) interphases due to its intrinsic limitations. This is mainly caused by the very low amount of material present in the interphase, and intrinsically low Raman scattering cross-section of the interfacial species. [22,23] This shortcoming of conventional Raman spectroscopy can be overcome by the use of near-field Raman spectroscopy, realized through the introduction of plasmonic active nanoparticles (NPs) onto the surface of the probed electrode. Upon laser excitation, these NPs create a strong near-field in their proximity, increasing the efficiency of the Raman scattering process by several orders of magnitude. This technique is known as shell-isolated nanoparticle-enhanced Raman spectroscopy (SHINERS). [22,24,27,28] Our previous study already showed that SHINERS is a powerful tool for investigating the CEI in high-voltage LIB cells. [29] In addition, several other studies have realized the potential of this technique to characterize the interphase on different anode materials. [22,23,[29][30][31][32] In this study, we employed operando SHINERS to investigate the CEI on NMC811 electrodes in high-voltage NMC811||graphite cells in the presence of a promising thiophene-based additive. Thiophene and some of its derivatives have been employed successfully as film-forming additives for high-voltage applications, improving the cells' performance by forming a polythiophene-based interphase. [33][34][35][36][37] 3-thiophene boronic acid (3-Thp-BOH) proved to significantly improve the electrochemical performance and the long-term cycle life of the corresponding cells. In two very recent studies by Ren et al. and Zhou et al., 3-Thp-BOH and the very similar 2-Thp-BOH were already introduced as promising electrolyte additives  for NMC622||Li-metal and LNMO||graphite cells. [37,38] In our study, the improved electrochemical performance was achieved by forming an effective CEI, mitigating the failure mechanisms to a large extent, such as transition metal dissolution or oxidative electrolyte decomposition. Operando SHINERS caught the formation of a stable 3-Thp-BOH-based polymeric film on the cathode surface at 2.8 V versus Li|Li + . In addition, the formation of a copolymer product of ethylene carbonate (EC) and 3-Thp-BOH was detected only at potentials above 4.3 V versus Li|Li + . The electrochemical and spectroscopic results were complemented with highly accurate G4MP2 level quantum chemistry (QC) calculations, providing deeper insights into the reaction mechanism at the interface.

Electrochemical Characterization
Galvanostatic cycling experiments were performed to characterize the effect of 3-Thp-BOH as an electrolyte additive for highvoltage NMC811||graphite cell chemistry. Figure 1 shows specific discharge capacities, impedances, and a mean charge/discharge voltage hysteresis analysis (ΔVA) of cells containing the baseline electrolyte and the electrolyte with an optimized 0.20 m concentration of 3-Thp-BOH. The complete set of galvanostatic cycling data for different additive concentrations and the corresponding analysis results can be found in Figure S2, Supporting Information.
In Figure 1, it is clearly shown that the addition of 3-Thp-BOH to the baseline electrolyte significantly influences the cell performance. The obtained specific discharge capacities of the considered cells and their corresponding Coulombic efficiencies (CEs) are displayed in Figure S2a, Supporting Information. A steep capacity fading was observed for the cell containing the baseline electrolyte starting right after the initial cycles. In addition, after about 450 charge/discharge cycles, the rate of fading drastically increased, indicated by the sharp decay of the cells' discharge capacity. In contrast, the cells with additive-containing electrolytes cycled for over 1750 cycles, showing a significantly lower capacity fading and higher discharge capacities. It is worth noting that in the first 100 cycles, the capacity of the cells containing the baseline electrolyte slightly exceeded the capacity of the cells containing 3-Thp-BOH. A correlative trend in the CEs of the cells containing the considered electrolytes was observed. For the first cycle, a CE of 89% was obtained for the cells containing the baseline electrolyte, while a CE of 85% was observed for the cells with 3-Thp-BOH (see Figure S2b, Supporting Information). The CE of the cells with both considered electrolyte formulations significantly increased in the following cycles, leveling out at > 99.5% at the 10 th cycle. Nevertheless, it was observed that for the initial seven cycles, the CE of the cells with baseline electrolyte is higher compared to the cells with the additive-containing electrolyte. For the continuous galvanostatic cycling, the cells with additivecontaining electrolytes showed slightly higher CEs. The calculated state-of-health (SoH) of the investigated systems shown in Figure S2c, Supporting Information mirrors the specific discharge capacities almost completely. It was observed that starting The spectra were recorded during the charging half-cycle in the range of 100 000 to 0.01 Hz. e) Accumulated specific energy plotted for the baseline electrolyte (black) and the baseline electrolyte + 0.20 m 3-Thp-BOH (blue) versus the state-of-health. The accumulated energy was calculated based on the discharge energy.
from the first cycle, the cells with additive-containing electrolytes showed reduced capacity fading. While the cells containing the baseline electrolyte reached a SoH of 50% after 500, the cells with the additive reached 82% of their initial capacity after 500 and 69% after 1750 charge/discharge cycles.
ΔVA is a tool to estimate the internal resistance of LIB cells. [29,39] As shown in Figure 1b, both cells containing the considered electrolyte formulations showed a continuous increase in resistance with ongoing galvanostatic cycling. However, it was observed that the resistance of the cells containing the baseline electrolyte increased significantly faster with ongoing galvanostatic cycling, reaching a ΔV of 850 mV after 500 cycles. The cells with the additive showed a ΔV of 330 mV after 500 cycles and a ΔV of 610 mV after 1750 cycles. To support the findings of the ΔVA, operando EIS experiments were performed in the cells containing the baseline electrolyte and the electrolyte with an optimized concentration of 0.20 m 3-Thp-BOH (Figures 1c,d). For comparison, the 1 st , 50 th , and 100 th cycle of cells with both electrolyte formulations were chosen. It was found that the cell with the baseline electrolyte showed a significantly lower impedance for the first cycle than the cell with the additive-containing electrolyte. For the 50 th cycle, a similar trend was observed. However, for the 100 th cycle, it was found that the cells with the baseline and the additive-containing electrolyte show a similar impedance, being in very good agreement with the results of the ΔVA. The ΔVA in Figure 1b demonstrates that around the 100 th cycle, the cells of both investigated electrolytes exhibit similar internal resistances.
The overall electrochemical performance of the cells with the investigated electrolyte formulations can be fairly compared by plotting the accumulated specific energy versus SoH, as this includes the relative contribution of all effects on the galvanostatic cycling, such as initial internal resistance, failure mechanisms, and capacity fading (Figure 1e). It is evident that the cells with the additive-containing electrolyte significantly outperform the cells with the baseline electrolyte. For galvanostatic cycling until 50% SoH, a quadrupled cycle life and a six-times higher accumulated specific energy for the cells with an optimized concentration of 3-Thp-BOH were obtained. Even for galvanostatic cycling until 80% SoH, the cells with the additive-containing electrolyte show distinct advantages over the baseline electrolyte counterparts, resulting in a tripled cycle life and a 3.5 times higher accumulated energy. In addition, the effect of the utilized additives' influence on the electrochemical cycling performance was compared with LiPO 2 F 2 , an additive known for its benefits in high-voltage applications. [40,41] As can be seen in Figure S3, Supporting Information, the cells with 3-Thp-BOH containing electrolyte show significantly lower capacity fading during the performed galvanostatic cycling experiments. Figure 2 exhibits the calculated differential capacity analysis (dQ/dV) of the first three charge and discharge cycles of the cells with the baseline electrolyte (a) and with the additivecontaining electrolyte (b). It was observed that the peaks are slightly shifted towards higher potentials for the cells with the additive-containing electrolyte. This shift can be attributed to increased internal resistance for the cells with the additivecontaining electrolyte, as already observed in the EIS experiments and ΔVA. Nevertheless, both dQ/dV plots are dominated by features that can be assigned to the lithiation of graphite and phase transitions of the NMC811 lattice structure. For the cells with the baseline electrolyte, the corresponding peak for the lithiation of graphite is located at a voltage of 3.5 V. The peak at 3.6 V can be ascribed to the H1-M phase transition, while the peaks at 3.9 and 4.1 V can be attributed to the M-H2 and H2-H3 phase transitions of the NMC811 active material, respectively. [42] For the cells with the additive-containing electrolyte, the corresponding peaks are observed at 3.6, 3.7, 3.9, and 4.1 V. The comparison of the dQ/dV plots of the first cycle of the cells with the two considered electrolyte formulations is depicted in Figure S4, Supporting Information. Despite great similarities in both plots, additional peaks can be observed at a voltage of 3.0 and 4.2 V for the cells with the additive-containing electrolyte. It is assumed that this peak can be attributed to the electrochemical reaction of 3-Thp-BOH, as it is not present for the baseline system. It can be concluded that most of the additive molecules already react during the first cycle, as the peak in the dQ/dV plot disappeared in the second cycle.
In addition to the electrochemical analysis, comparing the voltage profiles of the cells with the baseline and the additivecontaining electrolytes revealed interesting insights into interfacial Li-ion kinetics. As shown in Figure 2c,d, both voltage profiles show similar characteristics; however, minor differences were observed for the additive-containing cells, especially within the first cycle. For example, the beginning of the plateau, starting at about 3.5 V for the cell containing the baseline electrolyte, is shifted to higher voltage by approximately 100 mV for the additive-containing electrolyte counterpart. For higher additive concentrations, it was observed that the plateau is shifted to even higher voltages. This shift is probably due to increased overvoltage caused by additional internal resistance in the presence of 3-Thp-BOH. The additional resistance is another indicator that the consumption of the additive occurs before a voltage of 3.5 V is reached, resulting in the formation of a more resistive interphase, matching well with the findings from the dQ/dV plots, as displayed in Figure S4, Supporting Information. Another difference is the slope of the voltage versus time between 4.1 and 4.5 V. For the cells with the additive-containing electrolyte, a significantly lower incline was observed. The less steep incline can be accounted for a charge-consuming reaction of 3-Thp-BOH with electrolyte solvent molecules or other additive molecules. However, following the work of Stolz et al., this region is mainly dominated by material transport effects. [43] Consequently, the lower incline in the voltage profile for the cells with the additive-containing electrolyte could also indicate a decreased Li-ion transport out of the NMC811 bulk material caused by the interphase formation. For the subsequent cycles, no shift in the initial plateau at 3.5 V was observed. However, for the additivecontaining cells, the discussed decreased incline in the voltage profile is still present, indicating the ongoing influence of the formed interphase on the Li-ion transport. A more detailed discussion of the influence of the different investigated additive concentrations on the dQ/dVs and the voltage profile can be found in Figure S5, Supporting Information.
Concluding from the results of the electrochemical characterization and analysis, it can be stated that the addition 3-Thp-BOH in an optimized concentration of 0.20 m to the baseline electrolyte significantly improves the cells' electrochemical performance and reduces the capacity fading. Based on the literature and our previous study, we assume that the additive forms an effective surface layer on the NMC811 cathode via oxidative electropolymerization and the formation of a polythiophenic-based structure. The formation of this interphase explains the observed lower CE during the initial cycles. Also, the higher resistance in the initial cycles can be traced back to the additional interphase formation by the additive. The increased resistance also accounts for the lower capacity in the initial cycles, compared to the cells containing the baseline electrolyte, as higher resistance results in higher overpotentials. Nevertheless, despite increased resistance and lower capacity in the initial cycles, it is clear that the addition of 3-Thp-BOH in an optimized concentration results in the formation of a significantly more effective CEI, compared to the baseline electrolyte, increasing the performance and cycle life of the cells drastically.

Theoretical Calculations
To evaluate the assumed formation of a polythiophenic-based structure via electropolymerization of 3-Thp-BOH and to gain additional insights into the plausible reaction mechanisms, highly accurate QC calculations were performed. Note that a semideprotonated state of the additive molecules was assumed for the QC calculations. For the oxidation potential of a single semideprotonated 3-Thp-BOH molecule, a value of 4.2 V versus Li|Li + was calculated. This value is lower than for a single thiophene molecule (5.1 V vs Li|Li + ) and significantly lower than the oxidation potentials for a single EC molecule (5.7 V vs Li|Li + ), indicating 3-Thp-BOH oxidation and subsequent polymerization before oxidative degradation of the electrolyte carbonate compounds. This significant difference in the calculated oxidation potentials can be explained by the negatively charged boronic acid group, easing the oxidation, while the aromatic ring can stabilize the negative charge. Nevertheless, as calculations on single molecules do not grant additional insights into actual mechanisms, the free energy profile of the assumed electropolymerization was computed. For the reaction, two slightly different mechanisms were considered (see Figure 3). The first one (A) is similar to the polymerization mechanism of thiophene, which was considered in our previous work, where the oxidized additive is deprotonated by a PF 6 − anion. [29] Additionally, a second mechanism (B), where the proton is intramolecularly transferred to the semi-deprotonated boronic acid group instead of the deprotonation via an external anion, was suggested. Please note that the calculated energies of the reaction mechanisms in Figure 3 in electron volts (eVs) are equivalent to potentials in volts, as only one electron is transferred per elementary reaction step.
In the initial step, a 3-Thp-BOH molecule is oxidized with a free energy of ΔG = 5.59 eV, forming a radical (I). The radical can form a dimer on encountering another 3-Thp-BOH molecule with a free energy difference of 0.63 eV (II). In the next step, it was assumed that the dimer is deprotonated to stabilize the molecule (III). For deprotonation via a PF 6 − anion (A), this step occurred with a slight increase in the free energy of 0.57 eV. For deprotonation via intramolecular proton transfer (B), however, a significant gain in the free energy of 2.09 eV was calculated. This gain in energy can be explained by the stabilization of the dimer due to the equalized charge distribution. Furthermore, due to the deprotonating base in close proximity, one can expect fast kinetics of this step. Subsequently, the radical dimer is further oxidized (IV). For this step, free energies of ΔG = 4.14 and 4.77 eV were calculated for mechanisms A and B, respectively. Both are smaller compared to the free energy calculated for the first oxidation. This indicates a cascade-like development of the polymerization once the first dimer has formed, since the energy barrier of the oxidation is smaller for the dimer than the monomer. Following the second oxidation, the dimer is deprotonated again (V). For this step, a slight energy barrier of ΔG = 0.27 eV for deprotonation via an external anion was found. In comparison, the deprotonation via intramolecular proton transfer resulted in a gain in free energy of 2.79 eV. The total free energy of the calculated reaction mechanisms amounts to ΔG = 11.20 eV (5.60 eV per transferred electron) for mechanism A and ΔG = 6.57 eV for mechanism B (3.29 eV per transferred electron). These values equal potentials of 4.20 V and 1.89 V versus Li|Li + for the A and B polymerization reaction mechanisms, respectively. Based on the calculated free energy profile for both mechanisms, it was proposed that mechanism B is more likely to occur in the investigated system, explaining the observed phenomena.
Nevertheless, it has to be taken into account that the potential of the initial oxidation of the 3-Thp-BOH molecule is significantly higher than the potential of the total reaction. Additionally, not all monomers are necessarily deprotonated under experimental conditions, which may result in a mixed polymerization mechanism. Moreover, deprotonated monomers may coordinate to lithium ions, which slightly increases the oxidation potential as the cation's presence favors the negatively charged state of the additive. Therefore, it was suggested that the polymer formation only occurred at a low rate for low cell voltages and with an increased intensity at a higher state of charge, where the cell voltage equals the oxidation potential of the single 3-Thp-BOH molecule. This suggestion also fits the results of the electrochemical investigation, which indicated interphase formation at cell voltages before 3.5 V, as discussed for Figure 2.

SEM and EDX Analysis
Scanning electron microscopy (SEM) was employed to gain visual information on the interphase formed on the surface of NMC811 electrodes taken from cells with the baseline and additive-containing electrolytes. In addition, energy dispersive Xray (EDX) analysis of the electrode taken from a cell with the additive-containing electrolyte was performed to unravel the elemental composition of the NMC811 surface chemistry, as shown in Figure 4.
It is clear that the interphases formed on the electrodes exhibit significant differences in their respective morphologies, depending on the considered electrolytes. For the NMC811 electrode taken from an NMC811||graphite cell with Li-metal as reference electrode and containing the baseline electrolyte (Figure 4b), no major visible interphase formation was observed on the surface of the active material. For the electrode taken from the corresponding cell with the additive-containing electrolyte, however, the formation of an interphase was evident, uniformly covering the whole surface of the electrode. As seen in Figure 4c and Figure S8g-i, Supporting Information, the formed interphase fills the NMC particle gaps, possibly reducing the loss of active material due to particle cracking.
The EDX analysis of the NMC811 electrode's surface (Figure 4d,e) illustrated that the additive-induced interphase contains a significant amount of sulfur and boron, equally distributed over the surface of the electrode. The findings of this analysis support the claim that 3-Thp-BOH polymerizes on the surface of the NMC811 electrode during galvanostatic cycling, forming an effective polythiophenic-based interphase. In addition, the EDX analysis suggests that the boronic acid group remains attached to the polythiophenic backbone, in agreement with the QC calculations and the operando SHINERS results discussed in the next section. The corresponding EDX spectrum and additional SEM images with different magnifications can be found in the Supporting Information.

Operando SHINERS of the NMC811|Electrolyte Interphase
Operando SHINERS measurements of the surface of the NMC811 electrode were performed to gather detailed informa-   tion on the composition of the additive-induced interphase and further unravel the CEI formation mechanisms. The SHINERS measurements were conducted in the NMC811||graphite threeelectrode optical cell with Li-metal as reference electrode to determine the exact potential of the NMC811 working electrode. Galvanostatic cycling of the optical cell was carried out with the baseline and the additive-containing electrolytes with an optimized concentration of 0.20 m 3-Thp-BOH. The SHINER spectra of the NMC811 interphase with the additive-containing electrolyte at the open circuit potential (OCP) and the upper cut-off potential of 4.5 V versus Li|Li + are displayed in Figure 5. A comparison of the SHINER spectra of the cell with the baseline electrolyte and the additive-containing electrolyte at 4.5 V versus Li|Li + is shown in Figure S11, Supporting Information.
The following paragraph only focuses on the most dominant and relevant bands observed in the obtained spectra. Additional spectra and a detailed peak assignment can be found in the Supporting Information.
The spectrum of the surface of the NMC811 electrode at the OCP is dominated by bands that can be assigned to the cathode active material or the electrolyte. The broad bands between 555 and 630 cm -1 can be attributed to different lattice vibrations of the NMC811 active material, whereas the bands around 890 and 935 cm -1 can be assigned to EC and EMC, respectively. [44] Fur-www.advancedsciencenews.com www.advenergymat.de thermore, two bands around 1345 and 1600 cm −1 were observed, which can be related to the D-and G-bands of carbonaceous materials present as conductive agent in the electrode. [45] After charging the cell to 4.5 V versus Li|Li + , significant differences were observed. For example, instead of the broad band assigned to the NMC811 active material before, two relatively sharp bands around 460 and 560 cm -1 were observed at 4.5 V versus Li|Li + . Even though these bands can be attributed to the NMC811 lattice vibrations, a phase change of the active material due to delithiation is evident. [44,46] The significant decrease in the intensity of the bands assigned to NMC811 lattice vibrations at high potential can be explained by interphase formation shielding the back scattering vibrational process of the NMC811 particles. Furthermore, the increased electronic conductivity of the delithiated electrode lowers the skin depth of the Raman scattering process. [47] In addition, other new bands were observed at 4.5 V versus Li|Li + , which can be assigned to species evolving during CEI formation. The bands in Figure 5, located around 650-735 cm −1 and around 780 cm −1 , can be assigned to C-C and C-S-C ring deformation vibrations, [48,49] while the bands around 845 and 1055 cm −1 can attribute to C-H deformation vibrations of polythiophene. [29,48] The band observed around 1225 cm −1 can be related to the stretching vibrations of the C-C bonds between the polymerized thiophene molecules. The most prominent bands in the spectra around 1425 and 1490 cm −1 can be assigned to C=C ring and anti-ring stretching of polythiophene. [29,48,49] The aforementioned bands are absent in the spectrum of the CEI on the NMC811 electrode in the cell containing the baseline electrolyte, charged galvanostatically to 4.5 V versus Li|Li + ( Figure S10, Supporting Information). The operando SHINERS results show that a 3-Thp-BOH-based polymer interphase is formed on the surface of the NMC811 electrode, being in perfect agreement with the results of the electrochemical characterization, the information obtained from SEM/EDX analysis, and the reaction mechanism considered for the QC calculations proposed in Figure 3.
In addition to the bands assigned to the polymerized additive poly(Thp-BOH), the band around 865 cm −1 matches the reference spectrum of pure 3-Thp-BOH but not pure thiophene (Figures S16 and S17, Supporting Information). Therefore, it was assumed that this band can be assigned to the boronic acid group. As this band was only observed at higher potentials, combined with the results of the EDX analysis, it was further suggested that the boronic acid group stays attached to the thiophenicbackbone during the polymerization (Figure 6c). According to the proposed reaction mechanism in Figure 3, it was suggested that the boronic acid groups on the formed polymer are semideprotonated during further cell operation. Therefore, it is plausible that these surface groups affect the properties of the formed CEI. It was already shown that negatively charged surface groups can coordinate Li-ions to facilitate their transport, reducing the resistance of interphases. [50][51][52] It is assumed that the boronic acid groups enable a similar mechanism, contributing to the improved electrochemical performance of the cells cycled with the additive-containing electrolyte.
Additionally, all spectra of the operando SHINERS measurement are shown in Figure 6a to outline potential-dependent changes in the interphase composition. The formation of poly(Thp-BOH) was already observed at lower potentials (2.8 V vs Li|Li + ), indicated by an increase in the background intensity in the region between 1250 and 1650 cm −1 , as well as distinct bands with low intensities around 1225, 1425, and 1490 cm −1 . However, a significant increase in the background intensity was observed between 3.3 and 3.8 V versus Li|Li + . These results show clearly that a small portion of 3-Thp-BOH already electropolymerizes at potentials before 2.8 V versus Li|Li + , whereas most of the polymerization occurs between 3.3 and 3.8 V versus Li|Li + . This is in good agreement with the observations made from the dQ/dV plots, the voltage profile, as well as the calculated reaction potentials for the suggested polymerization mechanism. With increasing potential only minor changes in the spectra were observed, and most of the newly arising peaks also attribute to active vibration bands of poly(Thp-BOH). Interestingly, a distinct Raman active band around 1180 cm −1 was only observed at high potentials (shoulder at 4.3 V vs Li|Li + , distinct peak at 4.5 V vs Li|Li + in Figure 6a). [23] Following the literature, this band can be assigned to semicarbonates and indicates the formation of a copolymer consisting of EC and 3-Thp-BOH. It also has to be noted that the potential corresponding to the rise of the mentioned peak is in good agreement with the additional peak observed at 4.2 V in the dQ/dV plot of the additive-containing electrolyte (Figure 2b). In our previous work, we already discussed the possibility of forming an EC-thiophene copolymer using theoretical modeling. [29] Within this study, additional QC calculations were performed to investigate the EC and 3-Thp-BOH copolymer formation, following a similar reaction mechanism as discussed for Figure 3 (see Figure 6b). In the initial step, EC is oxidized, forming a radical cation, which is deprotonated by a PF 6 − anion (i). For this step, an oxidation potential of 5.7 V versus Li|Li + was calculated. This value is in good agreement with literature values in the range of 5.2-6.0 V versus Li|Li + . [53,54] In the next step, a semi-deprotonated 3-Thp-BOH molecule is added to the EC radical. For this step, a small energy barrier of 0.44 eV was calculated (ii), followed by a stabilizing energy gain of 0.72 eV (iii). Subsequently, the dimer can be oxidized again and react with other EC or 3-Thp-BOH molecules to form the proposed copolymer in Figure 6b. Due to the stabilizing energy gain, the overall reaction for the dimer formation is thermodynamically favorable. However, compared to the formation of poly(Thp-BOH), the copolymer formation is energetically not favored. Nevertheless, due to the abundance of EC molecules in the electrolyte, it is still plausible that the copolymer is formed in small amounts. Therefore, it is suggested that the copolymer possibly is a minor component of the formed CEI. This suggestion also agrees with the operando SHINERS results, which indicate that the copolymer formation was only observed at high potentials and with relatively lower peak intensities compared to the bands ascribed to poly(Thp-BOH).
Following the logic of forming the copolymer from EC and 3-Thp-BOH, a possible reaction of EMC and the additive has to be considered, since EMC is present in an even higher ratio in the electrolyte compared to EC. However, according to the work of Vatamanu et al., who found an increased concentration of EC at the surface of the electrode at higher cell voltages, while the concentration of the investigated linear carbonate decreased, it is suggested that the possible formation of a copolymer formed from EMC and 3-Thp-BOH can be neglected. [55] Time-dependent measurements were additionally performed to investigate the CEI stability and identify possible Figure 6. a) Operando SHINER spectra of the NMC811 electrode surface during galvanostatic cycling of an NMC811||graphite cell with Li-metal as reference electrode containing the baseline electrolyte + 0.20 m 3-Thp-BOH. All mentioned potentials are measured versus Li|Li + . b) Putative reaction mechanism of the formation of the copolymer comprising of EC and 3-Thp-BOH with energy differences ΔE (computed at 0 K) and free energy differences ΔG (computed at 298 K). For each step of the mechanism, the calculated free energies are shown. Note that all calculated energy values are given in electron volts. c) Schematic potential-dependent overview of the proposed CEI formation and its effect on the electrochemical performance of the investigated LIB cell.
decomposition products. For this purpose, an NMC811 electrode was charged to 4.5 V versus Li|Li + . After reaching the potential, a constant potential step was applied to hold the cell's potential for 60 min. Despite the high applied potential, no significant differences were observed in the SHINER spectra that could indicate interphase decomposition or the formation of other species. The recorded spectra can be found in Figure S13, Supporting Information. The SEM and EDX analysis already indicated an even distribution of poly(Thp-BOH) over the whole surface of the electrode. Nevertheless, the composition of the surface layer on the aged electrode was further investigated via SHINER mapping. As expected, the results of the SHINER mapping confirmed a uniform formation of poly(Thp-BOH) on the surface of the investigated sample ( Figure S14, Supporting Information).
To sum up the results of the performed SHINERS investigation, we suggest that at low potentials, a relatively small amount of poly(Thp-BOH) is formed on the NMC surface, which continuously grows with increasing potential. At potentials above 3.5 V versus Li|Li + , the interphase formation rate drastically increases as the electrode potential reaches values closer to the ones calculated for the initial oxidation ( Figure 3). With potentials around 4.3 V versus Li|Li + , small amounts of an EC-3-Thp-BOH copolymer are formed, contributing to the CEI composition. We further suggest that most of the interphase formation occurs within the initial charge half-cycle of the cell, as indicated by the dQ/dV analysis (Figure 3). In addition, no significant changes in the interphase composition were observed after holding the upper cut-off potential for a significant amount of time, indicating that no major interphase composition changes are expected with continuous galvanostatic cycling. Concluding from the results of the electrochemical and SHINERS investigations, we propose that the formed 3-Thp-BOH-based CEI effectively mitigates known failure mechanisms at high operational cell voltages, like oxidative electrolyte decomposition or transition metal dissolution and, therefore, enables significantly enhanced cycle life and performance of the considered cell chemistry (Figure 6c).

Conclusion
In this study, we introduced 3-Thp-BOH as a promising filmforming additive for high-voltage NMC811||graphite cells. The addition of an optimized concentration of the additive results in reduced capacity fading and a significantly increased the electrochemical performance and cycle life of the considered cells. Using different electrochemical analysis methods, we characterized key parameters of the formed interphase. Additionally, the composition of the interphase, as well as its formation mechanisms, were investigated using operando SHINERS. Based on the results of the SHINERS investigations, supported by SEM and EDX analysis, it was found that the formed interphase mostly consists of poly(Thp-BOH), formed via oxidative electropolymerization of the additive. In addition, we were able to prove that the boronic acid group stays attached to the polythiophenic-backbone and is responsible for the improved properties of the CEI. Highly accurate QC calculations supported the study and confirmed the polymerization mechanism of 3-Thp-BOH forming poly(Thp-BOH) and the mechanism of the EC-3-Thp-BOH copolymer formation.

Experimental Section
Electrolyte Formulation: For the conducted investigations, three sets of organic carbonate-based electrolytes were formulated. As baseline electrolyte, a mixture of ethylene carbonate and ethyl methyl carbonate (EMC) (E-Lyte Innovations, battery grade; 3:7 by weight) and 1.00 m LiPF 6 was used. For the second set of electrolytes, different concentrations (0.12, 0.20, and 0.30 m) of 3-thiophene boronic acid (Sigma Aldrich, 98%) were added to the baseline electrolyte. For the third set, 0.10 m LiPO 2 F 2 was added to the baseline electrolyte. The additives were dried for 24 h at 80°C under reduced pressure. The electrolytes were formulated and stored in an argon-filled glovebox (MBraun, H 2 O, and O 2 content < 1 ppm).
Cell Assembly: For galvanostatic cycling, commercially available NMC811||graphite pouch-cells from Li-FUN Technology (Hunan, China) with an NMC811 active material mass of 1.082 g were used. Before assembly, the cells were opened and dried overnight at 90°C under reduced pressure and thereafter filled with 700 μL of electrolyte in a dry-room (dew point < −50°C). To ensure sufficient wetting, the cell stack was massaged gently for 5 min. Afterward, the cells were sealed at 165°C for 5 s at 15% of the regular atmosphere using a GN-HS200V pouch-cell sealer (Gelon LIB Co.; Shangdong, China). For galvanostatic cycling, the assembled cells were attached to a specially designed holder, and a constant pressure of ≈ 2 bar was applied using a torque screwdriver. [56] To ensure reproducibility, four cells were assembled for each electrolyte formulation.
Galvanostatic Cycling: Galvanostatic cycling experiments of the assembled cells were performed using a Maccor 4000 battery tester. Before cycling, the cells were rested for 24 h to ensure sufficient wetting. Galvanostatic cycling was conducted in a voltage range from 2.8 to 4.5 V. For the cell formation, three charge/discharge cycles at C/10 followed by three charge/discharge cycles at C/3 with constant current/constant voltage (CCCV) steps were performed. Further galvanostatic cycling was conducted at a C-rate of 1C using CC cycling steps. The C-rate was calculated on the nominal capacity (200 mAh) of the utilized pouch cells.
Electrochemical Impedance Spectroscopy: Electrochemical impedance spectroscopy (EIS) was employed to characterize the interphase resistance. A potentiostat/galvanostat PGSTAT204 (Autolab compact-series, Metrohm) controlled by the NOVA 2.1 software (Metrohm) was utilized for EIS measurements. For EIS, NMC811||graphite Li-FUN pouch-cells were used and prepared as described above. The cells were galvanostatically cycled at a C-rate of 0.5C between 2.8 and 4.5 V. EI spectra were recorded during every charge half-cycle at a cell voltage of 3.5 V in a frequency range from 100 000 to 0.01 Hz.
Nanoparticle Synthesis: For SHINERS investigation, SiO 2 -coated Au-NPs with a size of 55 nm were synthesized following a procedure derived from Li et al. [57] The NPs were synthesized and characterized according to the procedure described in the previous work. [29] Raman Measurements: Raman measurements were conducted using a confocal Raman microscope (Horiba Scientific, LabRAM HR evolution, air-cooled CCD detector, grating of 600 g mm −1 ). The samples were excited by a red laser (633 nm) with a power output of 10.5 mW at the objective. A 10% filter later adjusted the power to 1.05 mW. The laser was focused by a 50X long-working distance objective (Carl Zeiss Microscopy, 9.2 mm, numerical aperture 0.5). Raman spectra were recorded by three integrations of 35 s. The Raman spectrometer, data acquisition, and analysis were handled using LabSpec 6.7.1.10 (Horiba Scientific). Before each measurement, the system was calibrated on the peak of crystalline silicon at 520.7 cm −1 .
For preparing the Raman samples, 12 mm electrodes were punched out of the NMC811 electrode sheets, and 10 mm electrodes were punched out of the graphite electrode sheets taken from the opened pouch cells. Afterward, the electrodes were dried overnight at 90°C under reduced pressure. For SHINERS investigations, the prepared NPs were drop cast by 5×50 μL of the organic NP solution onto the NMC811 electrodes. The electrodes were dried in a 110°C oven for 15 min and further dried overnight at 90°C under reduced pressure. The SHINERS measurements were carried out using an air-tight optical cell (ECC-Opto-Std, EL-CELL) with a glass window. For operando SHINERS investigations, a three-electrode system with NMC811 as the working electrode (WE), graphite as the counter electrode (CE), and Li-metal as the reference electrode (RE) was chosen to determine the potential of the monitored reactions precisely. Note that both electrodes had their active material facing the glass window. However, an Al-mesh was placed on the NMC811 electrode to ensure sufficient wetting and ion diffusion. One layer of Celgard 2500 (polypropylene) separated the electrodes to avoid short circuits. In the last step, the cell was filled with approximately 200 μL of electrolyte. Figure S1, Supporting Information shows a schematic depiction of the optical cell. All cells were assembled in an argon-filled glovebox.
For operando SHINERS measurements, galvanostatic cycling was performed using a potentiostat/galvanostat PGSTAT204 controlled by the NOVA 2.1 software. All galvanostatic cycling experiments were conducted in a potential range of 2.8-4.5 V versus Li|Li + and at a C-rate of C/3. For stability experiments, the cells were charged up to the upper cut-off potential of 4.5 V versus Li|Li + . Afterward, the potential was held by applying a constant potential of 4.5 V versus Li|Li + . Note that constant potential steps were added to the galvanostatic cycling procedure to avoid potential drops during the measurement.
To investigate the signal-enhancing effect of the utilized SHINERS technique, a measurement of the interphase formed on the surface of the NMC811 electrode was performed under the same experimental conditions, without the presence of NPs.
SEM and EDX Analysis: Further investigations of the surface of charge/discharge cycled NMC811 electrodes taken from an NMC811||graphite cell with Li-metal as reference electrode, containing the baseline electrolyte and the baseline electrolyte + 0.20 m 3-Thp-BOH, were performed by scanning electron microscopy (SEM) and energy dispersive X-ray (EDX) analysis. For SEM imaging, an Auriga electron microscope (Carl Zeiss Microscopy) with an accelerating voltage of 3 kV was used. EDX measurements were carried out at an accelerating voltage of 10 kV using an energy-dispersive X-ray detector (Oxford Instruments). Charge/discharge cycling of the SEM and EDX samples was performed in an air-tight optical cell using cyclic voltammetry (CV). Three charge/discharge cycles were performed in the potential range of 2.8-4.5 V versus Li|Li + at a scan rate of 150 μV s -1 . After CV-cycling, the electrodes were transferred in an argon-filled glovebox, washed three times with 100 μL EMC, dried under reduced pressure for 10 min, and were transferred into the microscope chamber via an air-tight vessel.
Theoretical Calculations: Quantum Chemistry (QC) calculations were performed at the accurate but computationally expensive G4MP2 level of theory (typical error of 1 kcal mol −1 or 0.04 V) using the Gaussian 16 package to compare the energetics of the polymerization mechanisms of 3-Thp-BOH and thiophene. [58,59] The molecular structures of educts, products, and intermediates were taken from the previous study and modified by hand via Avogadro if required. [29,60] The electrochemical oxidation potentials have been computed via the Nernst equation: for the respective putative oxidation reactions. ΔG ox equals the free energy difference in eV, n is the number of electrons transferred per elementary step, and F is Faraday's constant (26.801 A h mol -1 ). Please note that the sign of the equation above changes for the reduction reactions. A constant shift of −1.4 V was applied to relate the calculated values for the reactions with the Li|Li + scale, relevant to the experimentally obtained values. [53,54] To mimic the intramolecular environment in the electrolyte, the SMD implicit solvation model was employed in all calculations with built-in parameters for acetone, showing a similar dielectric constant as liquid carbonate electrolytes. [54,[61][62][63][64]

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.