Chemical composition and formation mechanisms in the cathode-electrolyte interface layer of lithium manganese oxide batteries from reactive force field (ReaxFF) based molecular dynamics

Abstract

Lithium manganese oxide (LiMn₂O₄) is a principal cathode material for high power and high energy density electrochemical storage on account of its low cost, non-toxicity, and ease of preparation relative to other cathode materials. However, there are well-documented problems with capacity fade of lithium ion batteries containing LiMn₂O₄. Experimental observations indicate that the manganese content of the electrolyte increases as an electrochemical cell containing LiMn₂O₄ ages, suggesting that active material loss by dissolution of divalent manganese from the LiMn₂O₄ surface is the primary reason for reduced cell life in LiMn₂O₄ batteries. To improve the retention of manganese in the active material, it is key to understand the reactions that occur at the cathode surface. Although a thin layer of electrolyte decomposition products is known to form at the cathode surface, the speciation and reaction mechanisms of Mn²⁺ in this interface layer are not yet well understood.

To bridge this knowledge gap, reactive force field (ReaxFF) based molecular dynamics was applied to investigate the reactions occurring at the LiMn₂O₄ cathode surface and the mechanisms that lead to manganese dissolution. The ReaxFFMD simulations reveal that the cathode-electrolyte interface layer is composed of oxidation products of electrolyte solvent molecules including aldehydes, esters, alcohols, polycarbonates, and organic radicals. The oxidation reaction pathways for the electrolyte solvent molecules involve the formation of surface hydroxyl species that react with exposed manganese atoms on the cathode surface. The presence of hydrogen fluoride (HF) induces formation of inorganic metal fluorides and surface hydroxyl species. Reaction products predicted by ReaxFF-based MD are in agreement with experimentally identified cathode-electrolyte interface compounds. An overall cathode-electrolyte interface reaction scheme is proposed based on the molecular simulation results.

Keywords

lithium manganese oxide batteries / reactive force field (ReaxFF) / cathode-electrolyte interface layer / molecular dynamics

Introduction

Lithium manganese oxide (LiMn₂O₄) is the primary cathode material being exploited for high power and high energy density applications like electric vehicles [1] and renewable energy storage [2]. LiMn₂O₄ cathodes possess several advantages, such as low cost, non-toxicity, and simple preparation process [3] as compared to alternative cathode materials. However, there are well-documented problems with capacity fade of lithium ion batteries containing LiMn₂O₄ due to active material loss, caused by dissolution of divalent manganese from the cathode surface and deposition of dissolved manganese onto the anode [4]. The LiMn₂O₄ spinel consists of manganese ions in two oxidation states (Mn³⁺ and Mn⁴⁺) in equal proportions to give a net manganese oxidation of+ 3.5 [5]. Conventional wisdom holds that trivalent manganese ions are unstable to acid attack and are lost from the cathode surface into the electrolyte following charge disproportionation through the Hunter reaction [4]

\mathrm{2}{\mathrm{Mn}}^{\mathrm{3}+}\to {\mathrm{Mn}}_{\mathrm{solid}}^{\mathrm{4+}}+{\mathrm{Mn}}_{\mathrm{solution}}^{\mathrm{2+}}

The dissolution of divalent manganese from the surface of the cathode into the electrolyte leads to a decrease in the effective amount of cathode material available for lithium intercalation. Furthermore, deposition of dissolved manganese onto the anode decreases the Li content of the anode by consuming electrons.

A key to improving the capacity and service life of LiMn₂O₄ based cells is to understand the reactions that occur at the surface of the cathode leading to dissolution of manganese. The cathode-electrolyte interface has only recently been considered as a potential impediment in the long-term viability of Li-ion cells [6]. Advances in imaging techniques such tunneling electron microscopy have revealed the presence of an extremely thin (2 to 5 nm thickness [7]) layer of compounds on the cathode surface. Electrochemical impedance spectroscopy measurements have confirmed that this film causes increased cell impedance [8], implying that the compounds formed at the cathode-electrolyte interface are poor ion conductors, thus limiting the diffusivity of Li⁺ through the layer [9]. The cathode-electrolyte interface is speculated to consist of a layer of electrolyte oxidation products and cathode active material salts formed by reduction of manganese [10]. Ex situ analysis of the cathode surface by infrared spectroscopy suggests the presence of polyethylene carbonate (PEC), perhaps deposited by oxidative polymerization of ethylene carbonate (EC) in the electrolyte. However, this finding could not be confirmed by X-ray photoelectron spectroscopy due to overlap between the PEC and polyvinylidine fluoride binder peaks [11], which yielded an inconclusive spectroscopic analysis of the cathode surface. Oxidation of other common electrolyte solvents, such dimethyl carbonate (DMC) and fluoroethylene carbonate (FEC) have been speculated but remains unconfirmed. Chemical characterization of the cathode-electrolyte interface is challenging because the concentrations of the interface compounds are below the detectable limits of nuclear magnetic resonance spectrometry and gas chromatography/mass spectrometry instruments [11]. Although manganese deposition on the anode surface [12] provides compelling evidence of manganese dissolution from the cathode, the speciation and reactions responsible for this at the cathode solid-electrolyte interface (SEI) layer are not yet well understood.

The cathode-electrolyte interface was found to have a layered structure similar to the anode-electrolyte interface through argon ion sputtering experiments [13], but its composition and electrochemical characteristics differ markedly from the anode SEI layer. Experiments have shown that the amount of Mn²⁺ present in the electrolyte increases with cycle number [14], suggesting that the cathode SEI layer formed during the initial cycling does not prevent further decomposition of electrolyte, in contrast to the anode SEI layer. The dynamic behavior of the cathode SEI layer is attributed to organic radicals and ions formed via electrolyte oxidation, which further combine with solvent molecules in a chain reaction [15]. Hence, the cathode SEI layer can limit battery performance by consuming electrolyte continuously as cycling proceeds. The LiMn₂O₄ surface is also highly susceptible to acid attack by HF produced via hydrolysis of LiPF₆ salt by trace water present in the electrolyte [16]. While HF is a putative trigger for Mn dissolution, the reaction mechanisms by which this occurs remain unknown. To improve battery performance and cycle life, a better understanding is therefore sought of the reactions that occur at the cathode surface that cause the loss of active material.

Various strategies have been adopted to improve cathode stability under cycling and decrease active material dissolution; these strategies purportedly involve the formation of a passivating cathode SEI layer [17]. The most popular approach is cathode modification via active material doping or cathode surface coating, where in Mn sites in the LiMn₂O₄ spinel are doped with other transition metals such as Ni [18], Fe [18], Co [19], Mg [20], Zn[20], or Al [21] to reduce the number of Mn³⁺ cations that are susceptible to acid attack. Spinel doping with Ni has proven especially successful for increasing the overall electrochemical performance of the cathode and reducing the Mn dissolution rate. However, nickel-doped spinel exhibits capacity loss after prolonged cycling due to surface film formation [22]. Another somewhat effective modification is the application of a surface coating to protect the cathode active material, either by altering the chemistry at the cathode-electrolyte interface, or by acting as a physical barrier between the active material and electrolyte. Al₂O₃ is the most commonly used coating for transition metal based cathodes as it is known to scavenge HF, reducing the acidity of the cathode surface, and it also reduces electrolyte decomposition at the cathode surface [23]. Surface coatings, however, have low lithium ion conductivity, necessitating the optimization of coating thickness in order to act as an effective barrier to electrolyte oxidation while still conducting Li ions [24, 25]. The use of electrolyte additives is another strategy to suppress electrolyte decomposition on the anode surface. Although a few compounds such as vinylene carbonate have been reported as potentially effective additives to reduce electrolyte degradation on LiFePO₄ cathodes [26], these results could not be replicated on other cathode chemistries [27].

All of the aforementioned strategies have been designed based on indirect evidence of increased cell performance, without knowing the changes in actual surface chemistry induced by cathode or electrolyte modifications. To devise more efficient cathode improvement strategies, greater insight into the composition and dynamics of the cathode electrolyte interface is essential. Considering the inherent difficulty of characterizing the composition of the cathode electrolyte interface layer using experimental techniques, computational methods are a resource to probe otherwise inaccessible nanoscale phenomena. Ogawa et al. [28] used quantum chemical molecular dynamics to study the transport of Li ions at the cathode electrolyte interface and determined that Li ions are solvated by EC molecules at the cathode surface. This study however did not consider differences in the chemical composition of the cathode-electrolyte interface and bulk electrolyte. Density functional theory methods have been used to study the oxidation of isolated solvent molecules and salt anions [29, 30] that do not explicitly interact with the cathode. Leung [31] conducted the first study on EC decomposition on LiMn₂O₄ surfaces, revealing the generation of surface hydroxyl species (and possibly H₂O molecules) via H abstraction from EC as the primary reaction mechanism on the surface of the cathode. While this study provided invaluable insights into the chemical nature of the cathode surface, it is important to investigate the reactions occurring at the cathode surface in the presence of the complete electrolyte. Among the important research questions yet to be answered are the products of the oxidative decomposition of the solvent molecules; the fate of acidic species (HF) on the cathode surface; and the complete reaction mechanism generating the cathode electrolyte interface layer. In this work, reactive force field based molecular dynamics is therefore applied to investigate the reactions occurring at the LiMn₂O₄ cathode surface and the mechanisms of manganese dissolution.

Methods

The ReaxFF model is employed to conduct molecular dynamics simulation of the surface-electrolyte interface between a lithium manganese oxide cathode and a nonaqueous electrolyte with a force field that allows for changes in bond order and bond length during the course of the simulation. The ReaxFF methodology is described in full in previous work [32–34] and was recently extended to model pairwise interactions of manganese [35] and fluorine [36] with carbon, hydrogen, oxygen, and lithium. The developed reactive force field is used to investigate the cathode-electrolyte interface layer in electrochemical cells with LiMn₂O₄ cathode, ethylene carbonate/dimethyl carbonate electrolyte, LiPF₆ salt, and 5% fluoroethylene carbonate additive. To probe manganese dissolution due to acid attack, HF is introduced into the simulation cell. Conventional wisdom holds that HF is generated within the battery when LiPF₆ molecules decompose and react with trace water molecules present in the electrolyte according to the following reactions [16]

{{\displaystyle \mathrm{L}\mathrm{i}\mathrm{P}\mathrm{F}}}_{\mathrm{6}}\leftrightarrow \mathrm{L}\mathrm{i}\mathrm{F}+{{\displaystyle \mathrm{P}\mathrm{F}}}_{\mathrm{5}}

{{\displaystyle \mathrm{P}\mathrm{F}}}_{\mathrm{5}}+{{\displaystyle \mathrm{H}}}_{\mathrm{2}}\mathrm{O}\to {{\displaystyle \mathrm{P}\mathrm{O}\mathrm{F}}}_{\mathrm{3}}+\mathrm{H}\mathrm{F}

The focus of this work is to investigate the fate of the HF molecules on the cathode surface and not simulate the formation reactions of HF. Hence, in the interest of reducing the complexity of the study, the simulations assume the product of LiPF₆ decomposition (HF) to be present in the battery without simulating the actual decomposition mechanisms. This assumption overlooks the effects of other LiPF₆ decomposition products on the cathode electrolyte interface. The inactive role of these decomposition products has not been proved or speculated but this assumption simplifies the system without compromising the purpose of this study.

The formation of the cathode electrolyte interface layer was simulated using a cathode half-cell composed of a LiMn₂O₄ crystal with its 100 surfaces exposed to the electrolyte consisting of 208 EC molecules, 480 DMC molecules, 20 FEC molecules and 60 LiPF₆ molecules. The LiMn₂O₄ crystal measuring 3.5 nm × 3.5 nm × 3.5 nm is located at the center of the simulation box with dimensions of 10.5 nm × 3.5 nm × 3.5 nm, the rest of the simulation box is filled with the electrolyte as shown in Fig. 1. This configuration helps with maintaining periodic boundary conditions in all three directions without any discontinuities. Bulk electrolyte character is maintained at distances greater than 2 nm from the electrode surfaces even after 2 ns, the maximum time period in this study. The equilibrium electrolyte density in the simulation cell is 1.26 g/mL with 3:7 mol ratio [37] of ethylene carbonate (EC) and dimethyl carbonate (DMC), 1 mol/L lithium hexafluorophosphate and 5% fluoroethylene carbonate.

The simulation cells were maintained at 330 K under the NVT ensemble using the Nosé Hoover thermostat with a temperature rescaling constant of 10 fs. The starting configuration of the electrolyte molecules was generated using the PACKMOL [38] code to make sure short range repulsive forces do not disrupt the simulation during the first time step. Each MD simulation was initiated from an energy-minimized structure and was equilibrated to the simulation temperature for 100 ps prior to the production run for 2 ns with a 0.2 fs time step.

The fate of HF on the cathode surface was studied by running a duplicate simulation as described above and introducing HF molecules into the simulation after 1.5 ns of cathode electrolyte interface layer formation. The HF molecules were introduced at random locations within a distance of 10 Å from the surface of the anode. HF molecules are introduced into the cathode electrolyte interface layer at a rate of 0.5 molecules/ps with a total of 10 molecules introduced into the simulation cell. After the introduction of HF molecules, the simulations were run for an additional 500 ps.

The ReaxFF based MD simulations were implemented using the LAMMPS parallel molecular dynamics code [39]. Molecular analysis of the simulations, which is to identify SEI layer compounds, was performed using a bond order cutoff of 0.2 as implemented in a home built MATLAB code [32].

Results

This work sets out to investigate the reaction mechanisms that result in the formation of the cathode-electrolyte interface layer. The reaction pathways of electrolyte molecule oxidation will be elucidated, the role of HF in the dissolution reaction will be determined and finally, the overall reaction mechanism on the surface of the cathode will be discussed.

Electrolyte solvent decomposition mechanisms

The ReaxFF based MD simulations of the cathode-half cell reveal a layer, 10 Å in thickness, consisting mainly of electrolyte solvent oxidation products formed on the cathode surface after 2 ns of simulation at 330 K. Figure 2 shows the primary compounds identified on the cathode surface. The ReaxFF model predicts a highly organic cathode-electrolyte interface consisting of compounds very similar to those speculated by other researchers [11, 40–42]. At least three different oxidation products of EC were found in the simulation cell after 2 ns depending on the reaction pathway followed. The oxidation of EC also triggers polymerization, yielding large organic molecules as observed by other studies on EC oxidation [40]. While EC is most susceptible to oxidation on the cathodes surface, oxidation products of DMC and FEC are also present in the layer.

Figure 3 identifies the primary oxidation pathways of the three solvent molecules. The oxidation pathway of the solvents has been found to be very similar to the pathway reported by Leung [31], proceeding via alkyl hydrogen abstraction and consequent lattice oxygen removal. The surface lattice oxygen atoms take up the hydrogen atoms from the alkyl groups of the solvent molecules resulting in a surface hydroxyl group and an organic radical adsorbed onto the cathode surface. After about 150–600 ps (depending on the reacting solvent molecule; see Fig. 2), the organic radical pulls out a lattice oxygen atom which is made available due to the reduced oxidation state of the surface manganese atoms in the presence of a nearby hydroxyl species. The removal of lattice oxygen results in the formation of organic aldehyde groups, ester groups and polycarbonate species in the case of EC oxidation. DMC and FEC are found to be oxidized to only one type of product with no evidence of polymerization. The organic radicals formed after H abstraction from EC often react with available surface hydroxyl groups resulting in the formation of alcohol groups. Unlike the anode electrolyte interface layer, no inorganic carbonates or organic lithium salts have been predicted by the ReaxFF model in the absence of salt decomposition; this is in good agreement with experimental evidence [43]. The oxidation reaction time for the solvent molecules ranges from 70 ps to 650 ps and increase in the order of FEC<EC<DMC.

The oxidation reaction pathways for all the electrolyte solvent molecules reveal the formation of surface hydroxyl species. The removal of oxygen from the cathode lattice exposes manganese atoms to the bulk electrolyte, which react with the surface hydroxyl species to form manganese hydroxide that is no longer a part of the lattice structure. The hydroxyl molecules are further, found to form water molecules that are likely to hydrolyze the LiPF₆ salt yielding HF. The formation of hydroxyl species and removal of lattice oxygen atoms result in manganese dissolution from the cathode into the electrolyte. The role of lattice oxygen vacancy formation on manganese dissolution has been speculated by previous experimental work using XAS analyses on cathode surfaces [44]. ReaxFF predictions of lattice oxygen removal during oxidation of electrolyte solvent molecules confirms the role played by oxygen vacancies on manganese dissolution and provides further insights into the reaction mechanisms that lead to lattice oxygen removal.

Fate of HF on the cathode surface

HF in the lithium ion cells is believed to be a byproduct of LiPF₆ hydrolysis due to the presence of trace water molecules. But, the results from this study show that water molecules can be formed on the cathode surface during electrolyte oxidation. This finding implies that even in the absence of water as an impurity, HF can be formed by the hydrolysis of LiPF₆ salt via water molecules generated through electrolyte oxidation on the cathode surface. Therefore, HF is an inevitable part of the battery chemistry and considering that HF is a putative trigger for manganese dissolution, its fate on the cathode surface is an essential part of the cathode-electrolyte interface chemistry. To investigate its fate on the cathode surface, HF is introduced into the simulation cell after 1.5 ns. A total of 10 HF molecules are introduced within 10 Å from the cathode surface and the cathode interface layer composition is analyzed after 500 ps.

The introduction of HF into the simulation cell results in a consequent rise in the number of surface hydroxyl species on the cathode. Figure 4 shows the primary compounds identified in the cathode electrolyte interface after 500 ps of simulation in the presence of HF molecules. The number of Mn(OH)₂ molecules also increases in the presence of HF. As can be seen in Fig. 2, the manganese atoms that form Mn(OH)₂ are extracted from the cathode lattice structure on bond formation with the surface hydroxyl species. The Mn-OH species observed in the simulation remain adsorbed on the cathode surface, but it is highly likely that manganese atoms are solvated by the electrolyte compounds leading to the dissolution of manganese at longer timescales. The fluoride anions from HF mostly end up as fluorinated salts of manganese (MnF₂) and lithium (LiF), which are the only inorganic salts present on the cathode surface, as observed experimentally [45]. Organic radicals formed after alkyl hydrogen abstraction from electrolyte solvent molecules also react with HF to yield fluorinated organic solvents.

The presence of HF results in an increased amount of surface hydroxyl species and the formation of inorganic salts on the surface of the cathode. While the hydroxyl species react with manganese atoms displacing the metal atoms from the cathode lattice, the fluoride anions react with metal atoms to form inorganic salts which may help passivate the cathode surface in the long run. ReaxFF analysis yields insights into the acid attack mechanism occurring on the cathode surface and reveals that scavenging hydroxyl molecules formed on the surface of the cathode may lead to the formation of a stable cathode-electrolyte interface layer.

Cathode-electrolyte interface layer formation reaction mechanism

Based on the ReaxFF MD analysis of the LiMn₂O₄ cathode electrolyte interface layer, the formation reaction scheme can be illustrated as shown in Fig. 5 [16]. Electrolyte solvent molecules undergo alkyl hydrogen abstraction and oxidation via the removal of lattice oxygen atoms on the cathode surface. This oxidation process results in the formation of organic aldehydes, esters and alcohols that compose the cathode electrolyte interface layer. The oxidation reaction also results in the production of surface hydroxyl species that react with manganese atoms displacing them from the lattice. The hydroxyl species also form water molecules. The manganese hydroxide molecules and water molecules are likely to be released into the electrolyte. The formation of water molecules is likely to contribute to the continuous degradation of the cathode electrolyte interface layer due to salt hydrolysis that yields HF molecules [46], further explaining the non-passive nature of the cathode electrolyte interface layer. The presence of HF in the electrolyte adds to the formation of surface hydroxyl species and yields metal fluorides. Manganese fluoride, lithium fluoride and fluorinated solvent molecules can be found in the cathode electrolyte interface layer. The components of the cathode electrolyte interface layer identified by this work (excluding the compounds produced due to salt decomposition) are similar to those speculated by previous studies [13, 47]. This work adds to previously established cathode-electrolyte interface reaction schemes [48, 49] by providing detailed to reaction mechanism. The ReaxFF analysis also establishes that the formation of hydroxyl species is central to the formation and non-passivating nature of the cathode-electrolyte interface layer.

Conclusions

The formation of the cathode-electrolyte interface layer was investigated using ReaxFF based molecular dynamics. The MD simulations reveal that the cathode-electrolyte interface layer is mainly composed of oxidation products of electrolyte solvent molecules such asaldehydes, esters, alcohols, polycarbonates and organic radicals. The electrolyte solvent molecules undergo oxidation by removing lattice oxygen atoms and producing surface hydroxyl species. In the presence of HF, increased number of hydroxyl species, metal fluorides and fluorinated organic solvent molecules were formed. Manganese dissolution was aided by the removal of lattice oxygen atoms that expose manganese atoms to the bulk electrolyte, and the formation of surface hydroxyl species that react with the exposed manganese atoms and displace the atoms from the cathode lattice.

The cathode-electrolyte surface compounds predicted by the ReaxFF model are in good agreement with the experimentally observed products. The ReaxFF analysis yielded insights into the acid attack mechanism occurring on the cathode surface, establishing that the generation of hydroxyl species is central to the formation and non-passivating nature of the cathode-electrolyte interface layer as well as the issue of manganese dissolution. The results of this work imply that developing strategies to scavenge hydroxyl molecules formed on the surface of the cathode will help in the formation of a stable cathode-electrolyte interface layer. Furthermore, this work also developed a detailed formation reaction pathway for the cathode-electrolyte interface layer with insights into the exact chemical compounds that can be found on the cathode surface. This information will be especially useful to future studies looking to bridge gaps between computational and experimental techniques.

References

[1] Scrosati B, Garche  J. Lithium batteries: status, prospects and future. Journal of Power Sources, 2010, 195(9): 2419–2430 

[2] Dunn B, Kamath  H, Tarascon J M . Electrical energy storage for the grid: a battery of choices. Science, 2011, 334(6058): 928–935

[3] Desilvestro J, Haas  O. Metal oxide cathode materials for electrochemical energy storage: a review. ChemInform, 1990, 137(1): 5C–22C

[4] Vetter J,Novák P, Wagner M R, Veit C,Moller   K C,Besenhard  J O,WinterM, Wohlfahrt-Mehrens M, Vogler C, Hammouche A.Ageing mechanisms in lithium-ion batteries. Journal of Power Sources, 2005, 147(1): 269–281

[5] Rodriguezcarvajal J ,  Rousse G ,  Masquelier C ,  Hervieu M . Electronic crystallization in a lithium battery material: columnar ordering of electrons and holes in the spinel LiMn₂O₄. Physical Review Letters, 1998, 81(21): 4660–4663

[6] Amine K, Chen  C H, Liu  J, Hammond M ,  Jansen A ,  Dees D, Bloom  I, Vissers D ,  Henriksen G . Factors responsible for impedance rise in high power lithium ion batteries. Journal of Power Sources, 2001, 97(01): 684–687

[7] Xu B, Fell  C R, Chi  M, Meng Y S . Identifying surface structural changes in layered Li-excess nickel manganese oxides in high voltage lithium ion batteries: a joint experimental and theoretical study. Energy & Environmental Science, 2011, 4(6): 2223–2233 

[8] Chen C, Liu  J, Amine K . Symmetric cell approach and impedance spectroscopy of high power lithium-ion batteries. Journal of Power Sources, 2001, 96(2): 321–328 

[9] Edström K, Gustafsson  T, Thomas J O . The cathode-electrolyte interface in the Li-ion battery. Electrochimica Acta, 2004, 50(2–3): 397–403

[10] Balbuena P B, Wang  Y H. Lithium-ion Batteries: Solid-electrolyte Interphase. USA: Imperial College Press, 2004

[11] Yang L, Ravdel  B, Lucht B L . Electrolyte reactions with the surface of high voltage LiNi_0.5Mn_1.5O₄ cathodes for lithium-ion batteries. Electrochemical and Solid-State Letters, 2010, 13(8): A95–A97 

[12] Zhan C, Lu  J, Jeremy K A ,  Wu T, Jansen  A N. Mn(II) deposition on anodes and its effects on capacity fade in spinel lithium manganate-carbon systems. Nature Communications, 2013, 4(9): 2437

[13] Edstrom K, Gustafsson  T, Thomas J O . The cathode–electrolyte interface in the Li-ion battery. Electrochimica Acta, 2004, 50(2–3): 397–403 

[14] Doh C H, Lee  J H, Lee  D J, Jin  B S, Moon  S I. The quantitative analyses of the dissolved manganese in the electrolyte of  Li/LiMn₂O₄ cell using by ion chromatography. Bulletin of the Korean Chemical Society, 2009, 30(10): 4–7

[15] Demeaux J, Caillon-Caravanier  M, Galiano H ,  Lemordant D ,  Claude-Montigny B . LiNi_0.4Mn_1.6O₄/electrolyte and carbon black/electrolyte high voltage interfaces: to evidence the chemical and electronic contributions of the solvent on the cathode-electrolyte interface formation. ECS Transactions, 2012, 41(31): 65–78

[16] Gulbinska M K . Catalytic materials and processes in secondary lithium-ion batteries. New & Future Developments in Catalysis, 2013: 479–498

[17] Jow R T, Xu  K, Borodin O ,  Ue M. Electrolytes for Lithium and Lithium-ion Batteries. New York: Springer,  2014

[18] Amine K, Tukamoto  H, Yasuda H ,  Fujita Y . Preparation and electrochemical investigation of LiMn_2−xMe_xO₄ (Me: Ni, Fe, and x = 0.5, 1) cathode materials for secondary lithium batteries. Journal of Power Sources, 1997, 68(2): 604–608

[19] Banov B, Todorov  Y, Trifonova A ,  Momchilov A ,  Manev V . LiMnCoO₄ cathode with enhanced cycleability. Journal of Power Sources, 1997, 68(2): 578–581

[20] Gummow R J, de Kock  A, Thackeray M M . Improved capacity retention in rechargeable 4 V lithium/lithium-manganese oxide (spinel) cells. Solid State Ionics, 1994, 69(1): 59–67 

[21] Myung S T, Komaba  S, Kumagai N . Enhanced structural stability and cyclability of Al-doped LiMn₂O₄ spinel synthesized by the Eemulsion drying method. Journal of the Electrochemical Society, 2001, 148(5): A482–A489

[22] Bhaskar A, Mikhailova  D, Kiziltas-Yavuz N ,  Nikolowski K ,  Oswald S ,  Bramnik N N ,  Ehrenberg H. 3d-Transition metal doped spinels as high-voltage cathode materials for rechargeable lithium-ion batteries. Progress in Solid State Chemistry, 2014, 42(4): 128–148 

[23] Chen Z, Qin  Y, Amine K ,  Sun Y K . Role of surface coating on cathode materials for lithium-ion batteries. Journal of Materials Chemistry, 2010, 20(36): 7606–7612

[24] Li C, Zhang  H P, Fu  L J, Liu  H, Wu Y P ,  Rahmb E ,  Holze R ,  Wu H Q . Cathode materials modified by surface coating for lithium ion batteries. Electrochimica Acta, 2006, 51(19): 3872–3883 

[25] Sahan H, Goktepe  H, Patat S . A novel method to improve the electrochemical performance of LiMn₂O₄ cathode active material by CaCO₃ surface coating. Journal of Materials Science and Technology, 2011, 27(5): 415–420 

[26] Wu H C, Su  C Y, Shieh  D T, Yang  M H, Wu  N L. Enhanced high-temperature cycle life of LiFePO₄-based Li-ion batteries by vinylene carbonate as electrolyte additive. Electrochemical and Solid-State Letters, 2006, 9(12): A537 

[27] Eom J Y, Jung  I H, Lee  J H. Effects of vinylene carbonate on high temperature storage of high voltage Li-ion batteries. Journal of Power Sources, 2011, 196(22): 9810–9814

[28] Ogawa T, Miyano  M, Suzuki Y ,  Suzuki A ,  Tsuboi H ,  Hatakeyama N ,  Endou A ,  Takaba H ,  Miyamoto A . A theoretical study on initial processes of Li-ion transport at the electrolyte/cathode interface: a quantum chemical molecular dynamics approach. Japanese Journal of Applied Physics, 2010, 49(4): 04DP11–04DP11–6 

[29] Tasaki K. Solvent decompositions and physical properties of decomposition compounds in Li-ion battery electrolytes studied by DFT calculations and molecular dynamics simulations. Journal of Physical Chemistry B, 2005, 109(7): 2920–2933 

[30] Xing L, Li  W, Wang C ,  Gu F, Xu  M, Tan C ,  Yi J. Theoretical investigations on oxidative stability of solvents and oxidative decomposition mechanism of ethylene carbonate for lithium ion battery use. Journal of Physical Chemistry B, 2009, 113(52): 16596–16602

[31] Leung K. First-principles modeling of the initial stages of organic solvent decomposition on Li_xMn₂O₄ (100) surfaces. Journal of Physical Chemistry C, 2012, 116(18): 9852–9861

[32] Chenoweth K, van Duin  A C T, Persson  P, Cheng M J ,  Oxgaard J ,  Goddard W A  III. Development and application of a ReaxFF reactive force field for oxidative dehydrogenation on vanadium oxide catalysts. Journal of Physical Chemistry A, 2008, 112(37): 14645–14654

[33] Chenoweth K, van Duin  A C T, Goddard  W A. ReaxFF reactive force field for molecular dynamics simulations of hydrocarbon oxidation. Journal of Physical Chemistry A, 2008, 112(5): 1040–1053

[34] Bedrov D, Smith  G D, van Duin  A C T. Reactions of singly-reduced ethylene carbonate in lithium battery electrolytes: a molecular dynamics simulation study using the ReaxFF. Journal of Physical Chemistry A, 2012, 116(11): 2978–2985

[35] Reddivari S, Lastoskie  C M, van Duin  A C T. A reactive force field for manganese oxide reduction by methane. Physical Chemistry Chemical Physics, (in press)

[36] Reddivari S. Electrode-electrolyte interface layers in lithium ion batteries using reavtive force field based molecular dynamics. Dissertation for the Doctoral Degree. Ann Arbor: University of Michigan, 2016

[37] Borodin O, Smith  G D. Quantum chemistry and molecular dynamics simulation study of dimethyl carbonate: ethylene carbonate electrolytes doped with LiPF₆. Journal of Physical Chemistry B, 2009, 113(6): 1763–1776 

[38] Martínez L, Andrade  R, Birgin E G ,  Martinez J M . PACKMOL: A package for building initial configurations for molecular dynamics simulations. Journal of Computational Chemistry, 2009, 30(13): 2157–2164

[39] Plimpton S. Fast parallel algorithms for short-range molecular dynamics.  Journal of Computational Physics, 1993, 117(1): 1–19

[40] Eriksson T, Andersson  A M, Bishop  A G, Gejke  C, Gustafsson T ,  Thomas J O . Surface analysis of LiMn₂O₄ electrodes in carbonate-based electrolytes. Journal of the Electrochemical Society, 2002, 149(1): A69–A78 

[41] Aurbach D, Markovsky  B, Salitra G ,  Markevich E ,  Talyossef Y ,  Koltypin M ,  Nazar L ,  Ellis B ,  Kovacheva D . Review on electrode-electrolyte solution interactions, related to cathode materials for Li-ion batteries. Journal of Power Sources, 2007, 165(2): 491–499 

[42] Matsui M, Dokko  K, Kanamura K . Dynamic behavior of surface film on LiCoO₂ thin film electrode. Journal of Power Sources, 2008, 177(1): 184–193 

[43] Aurbach D, Ein-Ely  Y, Zaban A . The surface chemistry of lithium electrodes in alkyl carbonate solutions. Journal of the Electrochemical Society, 1994, 141(1): L1–L3 

[44] Carroll K J, Qian  D, Fell C ,  Calvin S ,  Veith G M ,  Chi M, Baggetto  L, Meng Y S . Probing the electrode/electrolyte interface in the lithium excess layered oxide Li_1.2Ni_0.2Mn_0.6O₂. Physical Chemistry Chemical Physics Pccp, 2013, 15(26): 11128–11138

[45] Aurbach D, Zaban  A, Gofer Y ,  Ely Y E ,  Weissman I ,  Chusid O ,  Abramson O . Recent studies of the lithium-liquid electrolyte interface electrochemical, morphological and spectral studies of a few important systems. Journal of Power Sources, 1995, 54(1): 76–84 

[46] Lux S F, Lucas  I T, Pollak  E, Passerini S ,  Winter M ,  Kostecki R . The mechanism of HF formation in LiPF₆ based organic carbonate electrolytes. Electrochemistry Communications, 2012, 14(1): 47–50 

[47] Simmen F, Hintennach  A, Horisberger M ,  LippertT, Novák  P, Schneider C W ,  Wokaun A . Aspects of the surface layer formation on Li_1+xMn₂O_4−δ during electrochemical cycling. Journal of the Electrochemical Society, 2010, 157(9): A1026

[48] Wang E, Ofer  D, Bowden W ,  Iltchev N ,  Moses R ,  Brandt K . Stability of lithium ion spinel cells. III. improved life of charged cells. Journal of the Electrochemical Society, 2000, 147(11): 4023–4028

[49] Wang R, Li  X, Wang Z ,  Guo H. Manganese dissolution from LiMn₂O₄ cathodes at elevated temperature: methylene methanedisulfonate as electrolyte additive. Journal of Solid State Electrochemistry, 2016, 20(1): 19–28

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg