Flame Retardant Incorporation into Lithium-Ion Batteries

The use of Lithium-ion batteries (LIB’s) in commercial electronics such as computers and cell phones has expanded in recent years. LIB technology offers higher energy density, lower self-discharge as well as higher operating voltage vs. other rechargeable battery technologies. However, the natural flammability of standard LIB carbonate based electrolyte along with risk of thermal runaway poses safety concerns. Thus, the research and development of nonflammable alternative electrolyte mixtures for standard LIB’s is of high interest to researchers. To that end, Organophosphate containing Flame retardant (FR) compounds are being investigated as they possess natural fire suppressing qualities. LIB utilization in large platform applications, such as electric vehicles (EV’s) and aerospace designs has stimulated interest in higher energy density electrode materials such as Si. However, the practical use of Si does bring with it challenges related to the enormous volume changes which take place during cycling. The use of LIB’s for large high energy applications raises elevated safety concerns relating to thermal runaway. Detailed investigations relating to the benefit, cycling performance, and effect on the solid electrolyte interphase (SEI) upon FR incorporation into LIB’s with various anodes with/without SEI film stabilizing agents will be presented. SEI composition and structural changes upon FR incorporation are analyzed via surface analysis techniques including SEM and XPS.


Chapter 1 -Introduction Background
Lithium-ion batteries have fast become the preferred energy storage option for consumer electronics including laptop computers, and smartphones. In addition, Li-ion batteries are being utilized for large-scale applications such as hybrid and electric vehicles (EV's) as well as aerospace platforms. As compared to other battery systems such as NiZn, NiMH, and NiCd, Li-ion batteries offer lower self-discharge, superior operating voltage, wide operating temperature range, and higher gravimetric and volumetric energy density. 1,2 The early use of Li metal anodes in rechargeable cells met with cell safety issues relating to Li dendrite accumulation on the surface of the Li during repeated cycling. The constant buildup of dendrites led to puncturing of the separator material and therein safety issues stemming from internal short circuit. The safety issues relating to use of Li metal as an anode led to the study of anode materials which allow for the reversible intercalation and deintercalation of Li ions. 2,3 The current standard for Li-Ion batteries ( Fig. 1) 2,4 The physical properties of these solvents are detailed in Table 1. 2,5 These mixtures are by their very nature flammable and thus the safety risks associated with thermal runaway pose concern.

Li-Ion battery Safety Issues
Li-Ion cell thermal runaway events can stem from any one of several causes including excessive heat buildup within cell, and/or cell overcharge/cell overdischarge. Internal short circuit via metallic dendrite accumulation as a result of poor manufacturing quality is also a trigger for a thermal runaway event. During an overcharge, significant heat within the cell leads to break down of the protective SEI (solid electrolyte interphase) film layer and separator material. The destruction of the SEI layer exposes the bulk electrode material now at states of extreme voltage and heat triggers conversion of the electrolyte into flammable gases. This over-delithiation of the cathode leads to failure of the cathode structure as well as the generation of oxygen and further heat evolution. Flammable gases build up also results in excess internal cell pressure. This process often leads to venting and subsequent ignition upon exposure to air as well as possible flame ignition inside the cell. Today, the majority Li-Ion cell models are fitted with safety relief valves/vents to reduce the possibility of explosion. However, the threat of thermal runaway upon individual Liion cells threatens fire spread to surrounding cells and therein the overall safety of the outside payload . 2,5,6

Flame Retardant Incorporation
The possibility of thermal runaway poses significant safety threats most especially to large scale high power/energy applications such as electric and aerospace vehicles. These safety concerns have prompted researchers to investigate the feasibility of Flame retardant (FR) cosolvent/additive incorporation into standard Liion electrolyte. Organophosphate containing compounds are now being studied for their natural fire suppressing qualities. [5][6][7] Many research groups have reported using 4-Isopropyl Phenyl Diphenyl Phosphate (IPPP) 8 , Diphenyloctyl phosphate (DPOF) 9 , Triphenyl Phosphate (TPP) [10][11][12][13][14] , and Dimethyl methylphosphonate (DMMP) 7,15-17 as FR cosolvents/additives for Li-Ion cells. The advantage through the use of these additives is to perfect a viable nonflammable alternative which offers comparable electrochemical performance to standard electrolyte mixtures. The origination of these flame mediating qualities is thought to stem via radical scavenging and therein halting of combustion or though char layer formation. 2,5-7 Chapter 2 of this dissertation discusses the FR benefits and electrochemical effects of the incorporation of Triphenyl phosphate (TPP) into Li-Ion batteries with standard Graphite anodes.

High Capacity Si Anodes
The higher energy and power requirements of large scale platforms such as electric automobiles and space vehicles have prompted the development of anodes with higher energy density. Si anodes are of keen interest as an anode material due to their significant theoretical specific capacity advantage (3579 mAh/g) vs. standard Graphite anodes (372 mAh/g). 18 The practical use of Si anodes has been wrought with challenges relating to the immense volume variations (3-4 fold) that occur between their charged and discharged states resulting in substantial internal mechanical stresses. These physical stresses lead to loss of electrical contact between the Si anode active material and the Cu current collector. The surface variations of Si anodes during repeated cycling also leads to breakdown of the protective SEI and continual reformation. This continuous SEI formation results in large initial irreversible capacity loss, poor capacity stability and over the long term shorter cell life. 18,19 Many research groups have been investigated thin-film Si anodes as well as Siinactive composite materials with decreased Si particle size and alternative binder. [20][21][22][23] These efforts are directed towards mediation of the enormous mechanical strains associated with repeated cycling of Si anodes. The cycling benefits offered via the use of SEI film stabilizing additives such as lithium bis(oxalato)borate (LiBOB) and fluoroethylene carbonate (FEC) have also been explored by various groups. 24

Introduction
Lithium-ion battery technology in recent years has proven itself as a dependable energy storage medium for commercial consumer electronics. Li-ion batteries offer higher operating cell voltage, higher energy density, longer cycle life and lower selfdischarge. These advantages make Li-ion cells superior to other rechargeable systems such as Ni-MH and Ni-Cd. Safety issues however remain a concern with today's Liion batteries since the electrolyte is typically a blend of ethylene carbonate (EC) with ethyl methyl carbonate (EMC) with a lithium salt, such as lithium hexafluorophosphate (LiPF 6 ). These electrolyte solutions are flammable and a risk of thermal runaway is a concern. The main causes of Li-ion cell thermal runaway are attributed to both internal short via metallic dendrite accumulation and/or cell overcharge leading to destabilizing over-deliathiation of the cathode. [1][2][3][4][5] The potential for thermal runaway has led to efforts to reduce the fire risk and the  Preparation of the electrolyte and coin cell assembly was performed in a pure Argon atmosphere glove box with a water content < 5 ppm. Cells were constructed and cycled between 4.1 V and 3.0 V using an Arbin BT4010 battery cycler at 60 °F (15.5 °C).
The cycling protocol followed an initial formation cycling schedule with the first cycle at a C/20 current rate, followed by C/10 during cycles two and three, and C/5 for cycles four and five. Nominal cycling was conducted at C/5 current rate for an additional 30 cycles. The cells were then opened in an Ar filled glove box after a total of 35 cycles. Electrodes were extracted and rinsed three times with dimethyl carbonate (DMC) to remove residual salts. The rinsed electrodes were then vacuum dried overnight prior to surface film and morphological examination.

Self-Extinguishing Time (SET)
The self extinguishing time (SET) of electrolyte with increasing TPP concentration is summarized in TPP does not show appreciable decrease in flammability. However, incorporation of 10% -15% TPP results in a significant reduction in SET (9 s). SET experiments were also conducted on solvent blends without LiPF 6 and the trends in flammability reduction were very similar.

Flash Point (FP)
The flash points (FP) of solvent blends incorporating TPP are provided in alter the composition of the vapor phase above the cup. The significant differences in the quantity of reduced flammability when comparing SET and FP data with added TPP suggest that the development of additional straightforward flammability measurements would be beneficial. This also supports that the primary flame retarding action of the triphenyl phosphate is dependent upon its decomposition, either due to the formation of radical scavenging species or the formation of a thermal barrier of char, which would not be as significant in the flash point test.

Ionic Conductivity
The ionic conductivity of 1.

Cyclic Voltammetry
The cyclic voltammogram (CV) of the BL1 electrolyte and the electrolyte with 10 % TPP and 15 % TPP are shown in Figure 2. During the first potential sweep of the BL1 electrolyte, no reduction peaks are observed above 0.5 V vs. Li. The first potential sweep of electrolytes containing 10 % TPP and 15 % TPP contains a reduction peak at 1.8 V which is not present during the subsequent second and third potential sweeps. In addition, the current intensity of the peak increases with increasing concentration of TPP. This indicates that TPP is reduced on the anode surface but does not adversely affect the formation of a stable anode SEI.

Electrochemical Performance of Cells with Triphenyl Phosphate
Lithium ion coin cells containing an MCMB anode and LiNi 0. 8  indicating that the incorporation of TPP into the electrolyte does not adversely impact the initial specific capacity. It should be noted that the small observed differences in reversible capacity are not entirely attributable to electrolyte effects, but rather owing to cell to cell variability (i.e., < 5% variation in electrode weights). As shown in After completing the formation cycling and electrochemical characterization of the cells (discussed in the sections below), the cells were subjected to low temperature discharge rate characterization. This testing consisted of charging the cells at room temperature and discharging the cells at -20°C at various rates. As illustrated in Table   4, a noticeable decrease in the discharge rate capability was observed at low temperature with increasing TPP content. This is partly attributed to a decrease in the conductivity of the electrolyte solutions with increasing TPP content. As discussed below, the decreased rate capability is also attributed to decreased lithium intercalation/de-intercalation kinetics at the interfaces, since increased film and charge transfer resistances are observed during the measurement of electrochemical kinetics parameters.

Tafel Polarization Measurements of Three-Electrode Experimental Cells with Electrolytes Containing Triphenyl Phosphate
To determine the lithiation/de-lithiation kinetics of both the anode and the From the exchange current densities listed in Table 5, it is clear that the anode kinetics are nearly comparable for the baseline electrolyte (0.51 mA/cm 2 ) and the electrolyte with 10% TPP content (0.57 mA/cm 2 ). However, decreased kinetics were observed at the MCMB anode when utilizing an electrolyte with 15% TPP (0.46 mA/cm 2 ). These results suggest that increasing TPP content results in interfacial surface films that impede the lithium kinetics and is also accompanied by decreased ionic conductivity of the electrolyte, which will be reflected by decreased limiting currents. In contrast to the anode kinetics which are not significantly altered even though TPP contributed to the SEI, the cathode kinetics are noticeably reduced upon incorporation of TPP into the electrolyte, for example from 1.  Figure 6. In addition to resulting in reduced ionic conductivity, these results support the contention that the TPP is being incorporated into the cathode surface films as well, which is supported by the ex-situ analysis of the electrode harvested from the coin cells discussed in the section below.

Electrochemical Impedance Spectroscopy (EIS) Measurements of Three-Electrode Experimental Cells with Electrolytes Containing Triphenyl Phosphate
In an attempt to further understand the effect that triphenyl phosphate has upon the electrode/electrolyte interface, EIS measurements were performed on each individual electrode, as well as the full cell, by utilizing the reference electrode. In the interpretation of the data, an equivalent circuit consisting of a series resistance, R s , a parallel resistor-capacitor network (for film capacitance C f and film resistance R f ) in series for the high frequency relaxation loop, a resistor-capacitor parallel network in series for the low frequency relaxation loop, which is represented by a double-layer capacitance C dl in parallel with as series combination of charge transfer resistance R ct , and a Warburg impedance (w) representing the slow solid state diffusion of lithium ions in the bulk. [25][26][27][28] It is generally held that the high frequency relaxation loop is associated with the surface film between the electrolyte and the electrode, whereas the low frequency relaxation loop is correlated to the charge transfer resistance. These data were analyzed using the equivalent circuit described above and Z Simpwin software.
When EIS measurements were performed on the MCMB anodes after formation, as shown in Figure 7, a noticeable increase in the series resistance is observed with increasing TPP content, especially when 15% is added. This increase is primarily attributable to the decrease in ionic conductivity of the electrolyte solution, In general, there is an increase in the film and charge transfer resistances with addition of TPP, being again most dramatic for the electrolyte with 15% content (Table 6).
This suggests that TPP is altering the SEI film hindering facile lithium kinetics due to             volumetric energy density than Ni-Zn, NiCd and NiMH based battery systems. This translates into a longer running and more light-weight rechargeable system. 1,2 The integration of Li-ion battery systems into platforms with higher energy requirements such as hybrid and electric automobiles has led to the investigation of anode materials with superior energy density. To this end, Silicon has been investigated as a potential anode material due to the higher theoretical specific capacity (3579 mAh/g) compared to the traditional graphite anode (372 mAh/g). [3][4][5] The Organophosphates are of high interest due to their natural fire quelling attributes. 8,9 Compounds that have been studied by various groups include 4-Isopropyl Phenyl Diphenyl Phosphate (IPPP) 10 , Diphenyloctyl phosphate (DPOF) 11 , Triphenyl Phosphate (TPP) 9,12-16 , and Dimethyl methylphosphonate (DMMP) 8,[17][18][19] . Triphenyl phosphate (TPP) and Dimethyl methylphosphonate (DMMP) have both shown to be effective in reducing the flammability of standard electrolytes while offering a comparable electrochemical performance. 8,9,[12][13][14][15][16][17][18][19] Unfortunately, the incorporation of FR co-solvents into standard electrolyte mixtures has frequently resulted in poor capacity retention and poor low temperature performance. This has frequently been attributed to interference of the FR co-solvent with the formation of the anode solid electrolyte interface (SEI). This has prompted many groups to investigate the addition of SEI film stabilizing additives such as vinylene carbonate (VC) and lithium bis(oxalato)borate (LiBOB) in FR electrolytes. 8,9,[13][14][15] The use of Si anodes as a viable high capacity electrode material poses challenges due to the considerable volume changes (3-4 fold) between the charged and discharged states. 4 The enormous volume changes result in significant internal mechanical stress and subsequent loss of electrical contact between the current collector and Si active material. The overall high level of surface area changes leads to continual reformation of the SEI. This breakdown of the SEI allows for repeated exposure of the electrolyte with the bare electrode. The continuous SEI formation prevalent in the cycling of Si anodes can bring with it large irreversible initial capacity, poor long-term discharge capacity retention/stability and short cell life. In an effort to moderate the effects of the volume changes and resulting breakdown of the Si active material, many groups have focused on decreasing Si particle size within composite materials as well as pursuing thin-film Si anodes. [3][4][5] The addition of SEI film stabilizing additives such as VC, LiBOB, and fluoroethylene carbonate (FEC) has also been investigated. 5,20,21 The core focus of many research groups has been the incorporation of FR cosolvents into standard carbonate based lithium ion electrolytes in an effort to reduce flammability without sacrificing electrochemical performance. At the same time, much work has been directed towards the development of silicon anodes to improve the capacity of lithium ion batteries. This study focuses on the electrochemical performance and SEI properties of thin-film Si anodes cycled with flame retardant TPP and DMMP containing electrolytes with and without SEI film stabilizing LiBOB.

Experimental
Battery grade lithium hexafluorophosphate (LiPF 6 ), lithium bis(oxalato)borate  Table   1, in a pure Ar glove box. The Si anode functioned as the working electrode and the Li metal as the counter/reference electrode. A polyolefin separator was utilized and the coin cells were assembled and pressed under a load of 1000 psi.
Cells were cycled at constant-current charge and constant current discharge between 1.3V and 0.05V using an Arbin BT4010 battery cycler at 60 °F (15.5°C).
The coin cell cycling protocol followed a formation schedule consisting of one cycle at a C/20 current rate and two subsequent cycles at C/10. The cells were then cycled at a C/5 current rate for 52 cycles. Cycling performance was gathered and coulombic efficiency (cycling efficiency) as well as capacity retention was calculated.
Coulombic efficiency is defined as the ratio of discharge capacity or output of the cell to charge capacity or input. Capacity retention is defined as the ratio of discharge capacity at a particular cycle to the initial (1 st ) cycle discharge capacity.
Ex-situ analysis was conducted following the conclusion of the cycling schedule.
Cells were opened in a pure Argon atmosphere glovebox and the cycled Si anodes were extracted and rinsed with DMC to remove residual LiPF 6 salt. The rinsed electrodes were then vacuum dried overnight prior to surface analysis. Surface analysis of fresh and cycled Si electrodes was conducted using a JEOL Scanning Electron Microscope (SEM) in an Argon atmosphere chamber. Surface species characterization using X-Ray Photoelectron Spectroscopy (XPS) analysis was performed using a PHI 5500 system and Al Kα radiation. A hydrocarbon (C-H) contamination reference peak of 285 eV was used for spectral adjustment. Multipak versions 6.1 as well as XPS Peak 4.1 software were utilized for analysis and curve fitting of collected spectra respectively. Gaussian and Lorentzian functions were used for the least squares curve fitting during data processing.

Differential Chronopotentiometry
Differential chronopotentiometry analysis of Si/Li cells is consistent with the reduction of LiBOB prior to the reduction of EC, as previously reported (Fig. 1). 21 The DQ/DV results also reveal irreversible reduction of both TPP and DMMP at 1.  Table 2.
Cells cycled with the standard electrolyte have good first cycle efficiency (69 %) and discharge capacity after formation cycles (3100 mAh/g, 5th cycle), but have rapid capacity fade and low coulombic efficiency (93-96 %) during the next 54 cycles. In addition, the discharge capacity after formation cycles is higher (3150 mAh/g, 5th cycle), the cycling efficiencies for cycles 10-55 remain high (~98 %), and the capacity retention after 55 cycles is good (83-87 %). The results suggest that the best overall performance of the thin film Si electrodes is observed with flame resistant electrolytes with added LiBOB. In order to develop a better understanding of the source of performance changes as a function of changes in the electrolyte, ex-situ analysis of the surface of the thin film silicon electrodes was conducted.

Scanning electron microscopy (SEM)
SEM imaging of fresh thin-film Si anodes show a smooth surface while Si anodes extracted from Si/Li cells after 55 cycles have observable changes to the surface consistent with the formation of an SEI (Fig 3a-f XPS analysis was also conducted on electrodes extracted from cells after 55 cycles and compared to the electrodes after 5 cycles to develop a better understanding of the evolution of the SEI and the role of the SEI in capacity fade ( Figure 5, Table 4).       The practical commercial implementation of Si anodes has been wrought with challenges relating to the enormous volume changes (3-4 fold) which take place during cycling. 3 The enormous volume changes which take place with repeated cell charging and discharging results in substantial mechanical stress both upon the Si-current collector interface as well as internal stresses to the Si alloy structure.
This stress can lead to loss of electrical contact between the current collector and bulk Si active material as well as the loss of electrical contact between the individual Si particles. 3,20 Continual breakdown and subsequent reformation of the protective solid The focus of many investigations has been FR cosolvent/additive incorporation into standard Li-ion electrolyte in an effort to enhance safety through reduced flammability without sacrificing electrochemical performance. Simultaneously, there has been a push towards the development of Si composite anodes so as to produce a practical high capacity anode alternative for high power Li-ion battery applications. This investigation centers on the electrochemical performance of Sinanoparticle anodes cycled with TPP and DMMP containing electrolytes combined with SEI film stabilizing FEC additive.
electrodes were dried again in a vacuum oven at 110°C for 12 hrs. All electrodes contained approximately 60% Si, 20% of Carbon black (Super C), and 20% of PAA-CMC binder at a ratio of 1:1.
Cells were cycled at constant-current charge and constant voltage between 1.5 V and 0.05 V using an Arbin BT4010 battery cycler at 60 °F (16 °C). The coin cell cycling protocol followed a schedule consisting of a C/5 current rate for 50 cycles.
Coulombic efficiency is defined as the ratio of discharge capacity of the cell to charge capacity at a particular cycle. Capacity retention is defined as the ratio of discharge capacity at a particular cycle to the recorded maximum discharge capacity.

Results & Discussion
The cycling performance ramifications of TPP and DMMP FR cosolvent incorporation into standard electrolyte was studied in Si-nanoparticle/Li half cells and is shown along with coulombic efficiency in Fig.1. The 1st cycle coulombic efficiencies of cells after 50 cycles are shown in Table 2.
Cells cycled with standard electrolyte show an initial (1 st ) cycle efficiency of (88%) and a maximum capacity of (≥ 2800 mAh/g) prior to substantial fade.