DINITRILE-BASED LIQUID ELECTROLYTES

Cross-reference to Related Applications

This application claims the benefit of United States Provisional Patent Applications USSN 60/924,445 filed May 15, 2007 and USSN 60/929,977 filed July 20, 2007, the entire contents of both of which are herein incorporated by reference.

Field of the Invention

The present invention relates to liquid electrolytes in lithium-based electrochemical devices, particularly Li-ion batteries.

Background of the Invention During the last ten years, primary and secondary (rechargeable) lithium batteries have been the object of considerable research and development. The aim has been to develop low cost batteries, with a large energy content and good electrical performance. With this in mind, a large number of battery designs have been developed to comply with the requirements for various applications mostly in the field of portable electronics and power tools. However, the usage of lithium ion cell technology in larger scale batteries for transportation or stationary storage will require significant improvements in safety and further cost reductions. The composition of the electrolyte in a lithium ion cell is a key determinant of the overall safety of the electrochemical cell. The electrolytes currently used in lithium ion cells for portable electronics or power tools are highly volatile, very flammable, corrosive and highly toxic. In particular, the lithium salt currently favoured, LiPF ₆ , produces HF on contact with even trace amounts of water. Clearly, there is a need for new solvents with good solubility for lithium salts that are less toxic and less corrosive.

However, resolving all of these issues while still meeting the demanding electrochemical requirements for usage as an electrolyte in a lithium-based electrochemical cell is no small feat. A suitable electrolyte solvent must be able to solubilize a significant amount of electrochemically stable lithium salt and form a solution with good ionic conductivity over a broad temperature range. Furthermore it must be, at least kinetically, stable in contact with the highly reducing potential at the anode and the highly oxidizing potential at the cathode. To be commercially viable, it must also be inexpensive and easy to handle.

The focus to date has been on Li-ion batteries that use a liquid electrolyte comprising lithium salt, typically LiPF ₆ , in a combination of two or more organic carbonate solvents such as ethylene carbonate, dimethyl carbonate and/or diethyl carbonate. Historically other lithium salts have also been investigated including, amongst others: LiAsF ₆ , LiBF ₄ , LiCIO ₄ and Li(CF ₃ SO ₂ ^N. Other liquid solvents have been considered for lithium electrochemical cells including, amongst others, propylene carbonate, butyrolactone and dimethoxyethane. In general, it is necessary to combine one non-aqueous co-solvent with high volatility and low viscosity with another non-aqueous co-solvent having high viscosity but lower volatility to meet the requirements for ionic conductivity over a reasonable temperature range and to achieve required electrochemical stability at both the anode and cathode interfaces.

Organic solvents such as dimethyl carbonate (DMC), ethyl methyl carbonate (EMC), and diethyl carbonate (DEC) are the most commonly used components for the ionic liquid electrolytes currently used in lithium secondary batteries [1]. Though these solvents provide good charge-discharge capacity and cycle life they have the problem of low flash points: 17°C for DMC, 23°C for EMC, and 33°C for DEC. Further, conventional electrolytes like those based on the carbonate family, although working successfully in commercial Li-ion batteries for consumer electronics, cannot be used above 4.8 V versus the potential of lithium due to decomposition of the solvent. A novel electrolyte system having lower volatility, higher flash point, greater electrochemical stability and good solubility for non-corrosive and low toxicity lithium salts is needed to improve the safety of lithium-based electrochemical devices. Without such improvements lithium ion technology will never be suitable for the larger scale batteries needed for stationary storage or for hybrid or electric vehicles.

There is little known about lithium electrolytes with high anodic stability and previous work [12] on sulfone-based electrolytes showing anodic stability up to 5.8 V is one of the few rigorous studies in this field. This work used commercially-available or synthesized sulphone solvents to make lithium battery electrolytes and demonstrated that most of the solvents did not work either due to their physical characteristics (high melting point) or their incompatibility with the graphite anode.

In addition to the issues discussed above, the current state-of-art electrolyte of LiPF ₆ dissolved in organic carbonate solvents has disadvantages in low- temperature and high-temperature environments. At high temperature, the thermal instability of LiPF ₆ , which can thermally decompose to HF and PF ₅ , is believed to be the main cause for the poor performance of lithium-ion batteries. At low temperature,

the high viscosity of ethylene carbonate, which is a major component in the solvent mixture of state-of-art electrolyte, restricts the use of electrolyte to above -2O ⁰ C. These factors restrict the operation of lithium ion batteries to be between -2O ⁰ C and 60 ⁰ C. Succinonitrile has been previously used as a solid plastic crystal electrolyte in lithium battery applications. Succinonitrile exhibits plastic crystal formation at temperatures between -40 ⁰ C and 58°C and, in neat form, is consequently a solid over this temperature range. For lithium battery applications, high ionic conductivities have been reported for solid electrolytes based on succinonitrile (NC-CH ₂ -CH ₂ -CN, abbreviated as SCN) in its plastic crystal phase doped with certain lithium salts [2, 3].

Further, Canadian patent application 2,435,218 [4] discloses the use of lithium titanate anodes in electrochemical cells comprising a succinonitrile solid plastic crystal electrolyte. For electrochemical cells incorporating succinonitrile, it was believed that lithium metal, and therefore lithium containing materials having an electrochemical potential similar to lithium metal, could not be used as the anode due to the possibility of reactivity between -CN group and lithium metal [2], resulting in polymerization of the succinonitrile and consequent blockage of ionic conductivity. Commonly owned International patent publication WO 2007/012174 [6] discloses that lithium-based anodes having a potential within about 1.3 V of lithium metal may be used with succinonitrile-based solid plastic crystal electrolytes.

United States Patent 7,226,704 issued June 5, 2007 in the name of Panitz et al. [11] discloses electrolytes for lithium ion batteries based on the use of lithium bioxalato borate Li(C ₂ O ₄ ) ₂ B [often abbreviated as LiBOB]. This patent suggests that dinitriles may be used in the solvent system, however, no attempts were made to actually use dinitriles, especially with anodes having potentials approaching that of lithium metal. As previously mentioned [2], it has been generally thought in the art that succinonitrile (a dinitrile) would polymerize at lithium potentials resulting in polymerization of the dinitrile and consequent blockage of ionic conductivity.

Since, it could reasonably be anticipated that a dinitrile in liquid form would be much more reactive with metallic lithium or other lithium containing materials at similar electrochemical potential, dinitriles, especially succinonitrile, was not previously used as an electrolyte solvent or co-solvent in liquid electrolytes for lithium-based electrochemical devices.

There remains a need in the art for improved liquid electrolytes combining good ionic conductivities with broader potential windows and improved safety characteristics.

Summary of the Invention

Surprisingly, it has now been found that liquid electrolytes comprising a dinitrile provide good ionic conductivities, a stable electrolyte interface and better thermal stability over a broader potential window than similar electrolytes without the dinitrile.

According to one aspect of the invention, there is provided a liquid electrolyte comprising liquid dinitrile and an ionic salt.

According to another aspect of the invention, there is provided a use of a liquid electrolyte having liquid dinitrile and an ionic salt in an electrochemical device.

According to yet another aspect of the invention, an electrochemical device is provided comprising: a liquid electrolyte having liquid dinitrile and an ionic salt; an anode; and, a cathode.

Surprisingly, the liquid electrolyte comprising a dinitrile shows good thermal stability, high ionic conductivity, a very wide electrochemical stability window and/or good compatibility with lithium metal, without polymerization of the dinitrile at lithium ion potentials despite the dinitrile being in the liquid phase. Electrochemical devices of the present invention also have a large voltage differential between the anode and cathode leading to the delivery of higher energy density, while maintaining taking advantage of the dinitrile's relative non-flammability and its non-corrosiveness.

Advantageously, dintrile may be incorporated as a co-solvent with other liquid solvents to yield a liquid with good solubility for lithium salts of low toxicity and corrosiveness such as lithium bis-trifluoromethanesulphonylimide Li(CF ₃ SO ₂ ^N

[often abbreviated as LiTFSi] or lithium bioxalato borate Li(C ₂ O ₄ J ₂ B [often abbreviated as LiBOB].

Advantageously, liquid electrolytes of the present invention have a window of electrochemical stability on the oxidation side, versus LiVLi ⁰ , of about 4.6 V or greater, or 4.75 V or greater, or preferably greater than 5, or more preferably 5.1 V or greater, or 5.2 V or greater, or 5.3 V or greater, or 5.4 V or greater, or 5.5 V or greater, or 5.6 V or greater, or 5.7 V or greater, for example up to 6 V or higher.

Advantageously, liquid electrolytes of the present invention are thermally stable at temperatures at least 10 ⁰ C higher, preferably at least 2O ⁰ C higher, more preferably at least 3O ⁰ C higher, even more preferably at least 4O ⁰ C higher, than the maximum temperature at which similar electrolytes that do not contain dinitrile are thermally stable.

Preferred dinitriles for use in the present invention comprise aliphatic dinitriles, for example succinonitrile (SCN, NC(CH ₂ ) ₂ CN), glutaronitrile (GLN, NC(CHz) ₃ CN), adiponitrile (ADN, NC(CH ₂ ) ₄ CN), pimelonitrile (NC(CH ₂ ) ₅ CN) and suberonitrile (NC(CH ₂ ) ₆ CN) or mixtures thereof. Succinonitrile, glutaronitrile, adiponitrile or any mixture thereof is preferred, especially succinonitrile, adiponitrile or a mixture thereof, more especially succinonitrile or a mixture of succinonitrile and another dinitrile. The volatility and flash points of such dinitriles are attractively low for improved safety characteristics. For example, adiponitrile has a boiling point (b.p.) of 295 ⁰ C, a flash point (f.p.) of 159 ⁰ C, and a melting point (m.p.) of about 2 ⁰ C. The liquid electrolyte may comprise neat dinitrile or may comprise a mixture of dinitrile and a co-solvent. The amount of dinitrile in the liquid electrolyte is preferably in a range of from about 1% v/v to about 100% v/v, more preferably about 2-100% v/v, even more preferably about 25-100% v/v. The amount of dinitrile may be affected by the solubility of the particular ionic salt used in the electrolyte. Where ionic salts do not have great solubility in the dinitrile, the amount of dinitrile is preferably sufficiently low to prevent precipitation of the ionic salt. In such cases, the amount of dinitrile is preferably in a range of about 10% v/v to about 90% v/v, more preferably about 16-80% v/v, even more preferably about 25-75% v/v.

The liquid electrolyte preferably comprises a solution of dinitrile and a co- solvent. The co-solvent may be organic, inorganic or a mixture thereof. Preferred co-solvents include, for example, organic carbonates, lactones, sulfones, nitriles, ethers or mixtures thereof. The co-solvent may be, for example, dimethyl carbonate (DMC), diethyl carbonate (DEC), ethyl methyl carbonate (EMC), ethylene carbonate (EC), propylene carbonate (PC), methyl propyl carbonate (MPC), dimethyl formamide (DMF), tetrahydrofuran (THF), 2-methyl tetrahydrofuran, 2-chloromethyl tetrahydrofuran, methyl formate, methyl acetate, γ-butyrolactone (BL or γ-BL), acetonitrile (ACN), 3-methoxypropionitrile (MPN), tetramethylene sulfone ((CHj) ₄ SO ₂ ), dimethyl sulfoxide (DMSO), tetraethylsulfonamide (TESA), dimethyl sulfite, sulfolane (SL), 1 ,3-dioxolane, dimethoxyethane (DME), sulfur dioxide (SO ₂ ), thionyl chloride (SOCI ₂ ), sulfuryl chloride (SO ₂ CI ₂ ) or a mixture thereof. In keeping

with amounts previously states, volume ratio of dinitrile to co-solvent is preferably from 1 :99 to 99:1 , more preferably from 1 :9 to 9:1 , even more preferably from 1 :5 to 5:1.

One or more ionic salts are present in the liquid electrolyte. The ionic salt is preferably a lithium salt. Some examples of suitable ionic salts are lithium bioxalato borate salt (Li[CaO ₄ J ₂ B) sometimes abbreviated as LiBOB, lithium bis- trifluoromethanesulphonylimide (Li(CF ₃ SO ₂ ) ₂ N) sometimes abbreviated as LiTFSI, lithium bis-perfluoroethylsulphonylimide (Li(C ₂ F ₅ SO ₂ ) ₂ N), lithium difluoro(oxalato)borate (LiC ₂ O ₄ BF ₂ ) sometimes abbreviated as LiODFB, lithium tetrafluoroborate (LiBF ₄ ), lithium hexafluorophosphate (LiPF ₆ ), LiPF ₃ (CF ₂ CF ₃ ) ₃ , lithium thiocyanate (LiSCN), lithium triflate (LiCF ₃ SO ₃ ), lithium tetrafluoroaluminate (LiAIF ₄ ), lithium perchlorate (LiCIO ₄ ) and mixtures thereof. The ionic salt may be incorporated into the electrolyte in any suitable amount, for example, in an amount of from 1-20 mol%, more preferably in an amount of from 2-17 mol% or from 2-15 mol% or from 2-12 mol%.

The anode preferably has a potential within about 2.0 V of lithium metal, more preferably within about 1.6 V of lithium metal, even more preferably within about 1.3 V of lithium metal. It is especially remarkable that the anode potential can be within 1.3 V of lithium metal without inducing polymerization of the dinitrile. The anode preferably comprises a Li-containing material, for example lithium metal, a lithium alloy, lithium intercalated into hard or soft carbon (e.g. lithium intercalated into graphite), lithium intercalated into an oxide, a nitride or a phosphide, lithium inserted into a compound or composite by displacement, or a mixture thereof. Compounds and composites in which lithium may be inserted may comprise, for example, Sn compounds, Sb compounds, Al compounds, transition metal oxides, transition metal nitrides or transition metal phosphides (e.g. Cu ₂ Sb, CoSb ₃ , SnFe ₂ , Sn ₅ Cu ₆ , Mn ₂ Sb, tin oxide, silicon oxide, cobalt oxide, iron oxide, titanium oxide, copper oxide, Cu ₃ P, FeP ₂ , FeP, NiP ₂ , NiP ₃ , and Li ₂₆ Co ₀₄ N). Alloys of lithium may comprise, for example, lithium alloyed with Si, Sb, Al, Bi, Sn and/or Ag. Anode materials may be used alone or in combination with other materials. For example, lithium alloys may be used alone or in combination with carbon and/or other metals (e.g. Ni, Mn, Cr, Cu, Co). In one embodiment, anode materials may comprise a lithium titanate, for example Li ₄ Ti ₅ O ₁₂ .

The cathode may be any material suitable for use as a counter-electrode in an electrochemical device where the electrolyte is a liquid electrolyte with an ionic

salt. The cathode may comprise an insertion compound comprising lithium ions reversibly or non-reversibly inserted into an atomic framework. The atomic framework may comprise, for example, a single metal oxide, a mixed metal oxide, a single metal phosphate, a mixed metal phosphate, a single metal vanadate or a mixed metal vanadate. The metal is preferably one or more first row transition metals. Examples of suitable cathode materials include LiCoO ₂ , Li(Ni ₁ Co)O ₂ , LiMn ₂ O ₄ , Li(Mn _{0 5} Ni ₀₅ )O ₂ , Ui _+X (Mn ₁ Ni) _1-X O ₂ , Li _1+X (Mn ₁ Ni ₁ Co) _Vx O ₂ , LiNi _{0 5} Mn _{1 5} O ₄ , LiFePO ₄ , LiFe _{0 5} Mn _{0 5} PO ₄ , LiMnPO ₄ , LiNiPO ₄ , V ₂ O ₅ and mixtures thereof.

Electrochemical devices include, for example, electrochemical cells (e.g. batteries), fuel cells, electrochromic devices, supercapacitors and chemical sensors.

The present invention is particularly well suited to commercial lithium battery applications such as rechargeable batteries for portable electronics, power tools and electric vehicles (e.g. hybrid electric vehicles (HEVs) and plug-in hybrid electric vehicles (PHEVs)). Brief Description of the Drawings

In order that the invention may be more clearly understood, embodiments thereof will now be described in detail by way of example, with reference to the accompanying drawings, in which:

Fig. 1 is a graph depicting variation in log of conductivity (S/cm) as a function of temperature ( ⁰ C) for compositions of 4 mol% LiBOB, 4 mol% LiBF ₄ , 4 mol%

LiTFSI, and 2 mol% LiBOB + 8 mol% LiTFSI in neat succinonitrile (SCN);

Fig. 2 is a graph depicting variation in log of conductivity (S/cm) as a function of temperature ( ⁰ C) for a 1 molar concentration of LiTFSI in a 1 :1 mixture by volume of dimethyl carbonate (DMC) and succinonitrile (SCN); Fig. 3 a graph depicting variation in log of conductivity (S/cm) as a function of molar concentration of LiTFSI in a 1 :1 mixture by volume of ethylene carbonate (EC) and succinonitrile (SCN) at temperatures of 2O ⁰ C and -20 ⁰ C;

Fig. 4 a graph depicting variation in log of conductivity (S/cm) as a function of molar concentration of LiBOB in a 1 :1 mixture by volume of propylene carbonate (PC) and succinonitrile (SCN) at temperatures of 2O ⁰ C and -2O ⁰ C;

Fig. 5 is a graph depicting cyclic voltammograms obtained at 40 ⁰ C and 5O ⁰ C of 4 mol% LiBOB in neat succinonitrile (SCN) electrolyte using metallic lithium as blocking electrode and stainless steel as working electrode at scan rate of 10 mV-S ^"1 ;

Fig. 6 is a graph depicting a cyclic voltammogram obtained at 2O ⁰ C of 0.8 M LiBOB in a 1 :1 mixture by volume of propylene carbonate (PC) and succinonitrile

(SCN) using metallic lithium as blocking electrode and stainless steel as working electrode at scan rate of 10 mV-S ^"1 ;

Fig. 7 is a graph depicting the first two galvanostatic charge-discharge cycles of a carbon/1 M LiPF ₆ in 1 :1 ethylene carbonate:dimethyl carbonate/LiCoO ₂ cell cycled at a current of 6.45 mA/g of LiCoO ₂ at 20 ⁰ C;

Fig. 8 is a graph depicting the first two galvanostatic charge-discharge cycles of a carbon/1 M LiTFSI in 1 :1 ethylene carbonate:succinonitrile/LiCoO ₂ cell cycled at a current of 13.40 mA/g of LiCoO ₂ at 2O ⁰ C;

Fig. 9 is a graph depicting discharge capacity retention at 2O ⁰ C as a function of cycle number for two lithium ion cells each with a carbon anode and LiCoO ₂ cathode. One cell having an electrolyte comprised of 1 M LiTFSI in a 1 :1 mixture by volume of ethylene carbonate (EC) and succinonitrile (SCN) was cycled at a current density of 13.40 mA/g of LiCoO ₂ . A second cell shown for comparison having electrolyte comprised of 1 M LiPF ₆ in a 1 :1 mixture by volume of ethylene carbonate (EC) and dimethyl carbonate (DMC) was cycled at a current density of 6.45 mA/g.;

Fig. 10 is a graph depicting the first three galvanostatic (C/8 rate) charge- discharge cycles of a cell with a graphitic carbon cathode, a metallic lithium anode and an electrolyte containing 0.8 M LiBOB in a 1 :1 mixture by volume of propylene carbonate (PC) and succinonitrile (SCN) cycled at 2O ⁰ C; Fig. 11 is a graph depicting the first galvanostatic (C/12 rate) charge- discharge cycle of a Li/0.8 M LiBOB in 1 :1 propylene carbonate^uccinonitrile/ Li _{1 2} Mn ₀₄ Ni _{0 3} Coo i0 ₂ (LMNCO) cell cycled at 20 ⁰ C;

Fig. 12 is a graph depicting the retention of discharge capacity for a Li/0.8M LiBOB in 1 :1 by volume PC:SCN/Li., _! ,Mn ₀₄ Ni _{0 3} Co ₀ iO ₂ (LMNCO) cell cycled at 2O ⁰ C galvanostatically at C/12 rate;

Fig. 13 is a graph depicting first two galvanostatic (C/24 rate) charge- discharge cycles of a Li ₄ Ti ₅ 0 ₁₂ /SCN-2%LiB0B+8%LiTFSI/Lii 2Mn ₀₄ Ni ₀ 3Co ₀ 1O ₂ (LMNCO) cell cycled at 4O ⁰ C on the first cycle and at 2O ⁰ C on the second;

Fig. 14 is a graph depicting the retention of charge and discharge capacities mAh/g of Li _{1 2} Mn _{0 4} Ni _{0 3} Co ₀ ^O ₂ as a function of cycle number for the same cell as in

Fig. 13 cycled at 4O ⁰ C on the first cycle and at 2O ⁰ C on the second and subsequent cycles;

Fig. 15 is a graph illustrating the cyclic voltammetry scan of an electrolyte comprising 1 M LiTFSI in ADN; Fig. 16 is a graph illustrating the cyclic voltammetry scan of an electrolyte comprising 1 M LiTFSI in a 1 :1 mixture by volume of MPN and ADN;

Fig. 17 is a graph illustrating the cyclic voltammetry scan of an electrolyte comprising 1 M LiTFSI in a 1 :1 mixture by volume of γ-BL and ADN;

Fig. 18 is a graph illustrating cyclic voltammetry scan of an electrolyte comprising 1 M LiBOB in a 1 :1 mixture by volume of γ-BL/ADN;

Fig. 19 is a graph illustrating cyclic voltammetry scan of an electrolyte comprising 1 M LiTFSI in a 1 :1 mixture by volume of EC and ADN;

Fig. 20 is a graph illustrating conductivity as a function of concentration of LiTFSI in various electrolyte solutions; Fig. 21 is a graph illustrating conductivity as a function of temperature of 1 M

LiTFSI in various electrolyte solutions;

Fig. 22 is a graph illustrating cyclic voltammetry of Al wire in an electrolyte comprising 1 M LiTFSI in ADN;

Fig. 23 is a graph illustrating cyclic voltammetry of an Al wire in an electrolyte comprising 1 M LiTFSI in a 1 :1 mixture by volume of γ-BL and ADN;

Fig. 24 is a graph illustrating cyclic voltammetry of an electrolyte comprising 1 M LiTFSI in a 1 :1 mixture by volume of MPN and ADN;

Fig. 25 is a graph illustrating cyclic voltammetry of an Al wire in an electrolyte comprising 1 M LiTFSI and 0.1 M LiBOB in a 1 :1 mixture by volume of γ-BL and ADN;

Fig. 26 is a graph illustrating cyclic voltammetry of an Al wire in an electrolyte comprising 1 M LiTFSI in a 1 :1 mixture by volume of EC and ADN;

Fig. 27 is a graph illustrating cycling performance of a MCMB/1 M LiTFShO.1 M LiBOB in 1 :1 EC:ADN/LiCoO ₂ cell at C/12;

Fig. 28 is a graph depicting a cyclic voltammogram obtained of 1 M LiBOB in a 1 :1 mixture by volume of ethylene carbonate (EC) and adiponitrile (ADN); Fig. 29 is a graph depicting aluminum current collector corrosion behavior of an electrolyte comprising 1 M LiBOB in the ethylene carbonate (EC)/ethyl acetate (EA)/γ-butyrolactone (BL) solvent system of example 4 from US patent 7,226,704;

Fig. 30 is a graph depicting aluminum current collector corrosion behavior of an electrolyte of the present invention comprising 1 M LiBOB in a solvent system similar to the one from example 4 from US patent 7,226,704 except that adiponitrile

(ADN) replaced the γ-butyrolactone (BL);

Fig. 31 is a differential scanning calorimetry (DSC) plot of 1 M LiBOB in the ethylene carbonate (EC)/ethyl acetate (EA)/γ-butyrolactone (BL) solvent system of example 4 from US patent 7,226,704, and of 1 M LiBOB in an ethylene carbonate (EC)/adiponitrile (ADN) solvent system in accordance with the present invention;

Fig. 32 is a graph depicting the retention of discharge capacity for a lithium intercalated into graphite/various electrolytes/LiCoO ₂ electrochemical cell cycled at 2O ⁰ C galvanostatically at C/12 rate; and,

Fig. 33 is a graph depicting the retention of discharge capacity for a Li/various electrolytes/LiCoO ₂ cell cycled at 2O ⁰ C galvanostatically at C/12 rate.

Detailed Description of the Invention

Preparation of Liquid Phase Electrolytes Incorporating Dinitriles

The solvating properties and high flash points of dinitriles, especially succinonitrile, make them useful for liquid electrolytes for lithium secondary batteries.

dinitriles can be used neat or dissolved in a number of liquid solvents having properties suitable for use as electrolyte solvents for lithium ion batteries and lithium secondary batteries. The following examples demonstrate that usage of dinitriles as a low cost solvent or co-solvent in liquid electrolytes for lithium batteries yields electrolytes with unexpectedly good ionic conductivity, a very broad window of electrochemical stability and good thermal stability when compared to similar electrolytes not containing dinitrile. The usefulness of a dinitrile, especially succinonitrile, as a solvent or co-solvent in lithium secondary batteries was evaluated in electrochemical cells with either metallic lithium or graphitic carbon anodes.

The electrolytes were prepared by dissolving the appropriate amount of an ionic salt into neat dinitrile or into a solvent mixture of dintrile plus a co-solvent. The electrolyte solutions were mixed well and if necessary heated until complete dissolution.

Conductivity measurements were performed using the impedance spectroscopy technique. The electrolyte solutions were poured into a two-platinum- electrode conductivity cell with a cell constant of 0.96. The frequency was swept between 100 Hz and 1 MHz using a HP frequency analyzer. The temperature was varied between -20°C and 80 ⁰ C allowing 20 min for thermal equilibration.

Cyclic voltammograms were collected with a platinum microelectrode (25 μm) (for electrochemical window) or aluminum wire (100 μm) (for Al corrosion) and a silver wire as pseudo reference. The true potential was established with butyl- ferrocene (Aldrich).

Battery investigations were carried out with coin-type cells. Cathode and anode materials were prepared by mixing 85:5:5:5 (w/w) ratios of active material, graphite, super S carbon black and polyvinylidene difluoride binder dissolved in N- methyl pyrrolidinone, respectively. The resulting paste was applied to an aluminum foil current-collector and then was dried, first at room temperature and then at 150°C under vacuum for two days. A Celgard™ separator (30 μm thickness) was put between electrodes and soaked with the electrolyte. The cells were assembled in an Ar-filled glove box at room temperature. Cell performance was evaluated by galvanostatic experiments carried out on a multi-channel Arbin battery cycler. The cells were first charged and then discharged at constant current density between two potential limits set depending on the choice of electrodes used. All the electrolyte

preparation and handling as well as assembling of electrochemical cells were performed in an Ar-filled glove box.

Example 1: Conductivity in Neat Succinonitrile (SCN)

Fig. 1 shows temperature dependency of the conductivity of 4 mol% LiBOB in succinonitrile, 4 mol% LiBF4 in succinonitrile, 4 mol% LiTFSI in succinonitrile and 2 mol% LiBOB with 8 mol% LiTFSI in succinonitrile. Conductivities for the four electrolytes are quite different at room temperature when the succinonitrile is a solid plastic crystal. Above 49 ⁰ C, the succinonitrile is melted and the conductivities of the four electrolytes are both higher and achieve approximately the same value. It is apparent that liquid succinonitrile provides an excellent medium for conductivity for a liquid electrolyte having a variety of lithium-based ionic salts.

Example 2: Conductivity in Mixtures of Succinonitrile (SCN) and Co-solvent

Temperature dependence of the conductivity of 1 M LiTFSI in a 1 :1 mixture by volume of dimethyl carbonate (DMC) and succinonitrile is shown in Fig. 2. The conductivity at ambient temperature compares favourably with prior studies for

LiTFSI in conventional non-aqueous solvents [8, 9]. However, the new formulation offers huge improvements in low temperature conductivity in comparison to LiTFSI in 1 :1 EC-DMC which freezes by -2O ⁰ C [8]. Fig. 3 depicts the dependence of the conductivity of LiTFSI in 1 :1 mixture by volume of ethylene carbonate (EC) and succinonitrile as a function of salt concentration at 2O ⁰ C and at -2O ⁰ C. Adequate conductivity for application in lithium secondary batteries is achieved at both temperatures. The new formulation increases the flash point, leading to a safer battery, while improving the low temperature conductivity. A similar improvement in the flash point can be achieved with other suitable lithium salts. Fig. 4 is a graph depicting the dependence of the conductivity on the concentration of LiBOB salt added to a 1 :1 by volume mixture of propylene carbonate (PC) and succinonitrile at 2O ⁰ C and -20 ⁰ C. This high flash point, non-corrosive, electrolyte formulation achieves acceptable conductivities for application in lithium secondary and lithium ion batteries, with conductivities comparable to that of LiBOB in conventional non- aqueous solvents [10].

Example 3: Cyclic Voltammetry in Neat Succinonitrile (SCN)

Referring to Fig. 5, electrochemical stability window of a 4 mol% LiBOB in neat SCN electrolyte was measured by cyclic voltammetry at 4O ⁰ C and 5O ⁰ C with a

scan rate of 10 mV/s in an electrochemical cell. A stainless steel working electrode was separated from a lithium metal disk that served as both the reference and counter electrodes by a sheet of micro-porous separator Celgard™ 3501 impregnated with the electrolyte. At 4O ⁰ C, after lithium stripping at 0.36 V and lithium deposition at -0.48 V, no onset voltage was observed for anodic and cathodic currents even at 6 V versus LiVLi ⁰ . The same behavior is observed at 5O ⁰ C (melting of the SCN with 4 mol% LiBOB occurs at 49 ⁰ C as shown by differential scanning calorimetry (DSC)), except that the current densities for the deposition and the stripping of lithium are increased by 2.7 orders of magnitude. It is evident that melting of the SCN does not adversely affect the window of electrochemical stability.

Example 4: Cyclic Voltammetry in a Mixture of Succinonitrile (SCN) and Co-solvent

The potential window of electrochemical stability of 0.8M LiBOB in a 1 :1 mixture of propylene carbonate (PC) and succinonitrile was measured in an electrochemical cell with a stainless steel working electrode and a metallic lithium counter electrode at a scan rate of 10 mV/s. The room temperature cyclic voltammogram (Fig. 6) shows lithium deposition and stripping at potentials near LiVLi ⁰ . However at the upper voltage range, there is no onset of significant anodic or cathodic current until potentials above 5.5 volts. These results indicate that the combination of LiBOB salt with co-solvents containing succinonitrile provides an exceptionally wide window of electrochemical stablitiy that will allow the elelctrolyte's usage with cathodes having potentials of up to 5 volts versus LiVLi ⁰ . The power output of a battery increases with the cell's potential so increasing the cell's voltage provides a substantial advantage in power output.

Example 5: Electrochemical Performance Using Succinonitrile (SCN)

In order to compare the performance of the liquid electrolytes containing succinonitrile and non-corrosive lithium salts with conventional electrolytes two electrochemical cells were assembled and cycled at room temperature (2O ⁰ C). The first cell was made with conventional liquid electrolyte containing 1M LiPF ₆ in a 1 :1 mixture of ethylene carbonate and dimethyl carbonate, a graphitic carbon anode and a LiCoO ₂ cathode. The first two charge-discharge cycles at a current density of 6.45 mA/g of LiCoO ₂ (-C/20 rate) are shown in Fig. 7. A second cell was fabricated with a carbon anode and a LiCoO ₂ cathode and a liquid electrolyte comprising LiTFSI in 1 :1 ethylene carbonate and succinonitrile. The first two charge-discharge cycles at 2O ⁰ C and a current density of 13.4 mA/g of LiCoO ₂ are depicted in Fig. 8. A comparison of

Figs. 7 and 8 demonstrates that the electrolyte containing LiTFSI and succinonitrile has comparable capacity and somewhat reduced polarization, even at double the current density, in comparison to the conventional electrolyte with LiPF ₆ in EC-DMC. A comparison of the retention of discharge capacities for these two cells is provided in Fig. 9. Even though the lithium ion cell with the 1 M LiTFSI in 1 :1 EC-SCN is being cycled at about twice as high current density, the discharge capacity at 30 cycles is the same as for the cell with 1 M LiPF ₆ in 1 :1 EC-DMC. The new electrolyte based on LiTFSI and succinonitrile meets or improves the electrochemical performance while increasing the cell's safety through the utilization of solvents with higher temperature flash points and lithium salts that are less toxic or corrosive.

To further demonstrate the utility of liquid electrolytes comprising succinonitrile and non-corrosive lithium salts, two other electrochemical cells were assembled containing an electrolyte comprised of 0.8 M LiBOB in a 1 :1 mixture by volume of propylene carbonate (PC) and succinonitrile (SCN). The first of these cells contained a metallic lithium anode and a graphitic carbon (MCMB) anode. The first three galvanostatic (C/8 rate) charge-discharge cycles taken at 2O ⁰ C are shown in Fig. 10. The cell showed no evidence of excessive loss of capacity to SEI formation even though there was no ethylene carbonate in the electrolyte. This demonstrated that a stable SEI can be formed on graphitic carbons with a liquid electrolyte containing LiBOB and succinonitrile. The second electrochemical cell was fabricated with a metallic lithium anode and a Li _{1 2} Mn ₀₄ Ni _{0 3} Co _{0 1} O ₂ (LMNCO) cathode. The cell was cycled between 2.5 and 4.6 volts. The first galvanostatic (C/12 rate) charge- discharge cycle (Fig. 11) further illustrates the stability of this electrolyte to high charge potentials. The graph of the retention of discharge capacity as a function of cycle number (Fig.12) shows that even on extended cycling the cell shows excellent coulombic efficiency and capacity retention.

To further demonstrate that the utility of liquid electrolytes comprising succinonitirle and an ionic salt is not limited to electrolytes with co-solvents. A cell was assembled with a Li ₄ Ti ₅ O ₁₂ anode, a Li _{1 2} Mn ₀₄ Nio ₃ Cθo ^ ₂ (LMNCO) cathode and an electrolyte comprised of 8 mol% LiTFSI and 2 mol% LiBOB in neat succinonitrile.

DSC studies have shown that at this composition the electrolyte has a melting point of 21 ⁰ C. The first two galvanostatic (C/24 rate) charge-discharge cycles between 1 and 3.2 volts of this cell are shown in Fig. 13. The first charge-discharge cycle was taken at 40 ⁰ C (well above the electrolyte's melting temperature) and the second and subsequent cycles were at 2O ⁰ C (slightly below the electrolyte's melting

temperature). Fig. 14 is a plot of the retention of charge and discharge capacities as mAh/g of LH ₂ Mn ₀₄ Ni _{0 3} Co _{0 1} O ₂ as a function of cycle number which clearly illustrates that this succinonitrile electrolyte provides excellent capacity retention and coulombic efficiency even when charged to high potentials. The upper voltage limit of 3.2 volts versus Li ₄ Ti ₅ Oi ₂ is equivalent to about 4.7-4.8 volts versus Li ⁺ /Li°.

Example 6: Cyclic Voltammetry in Neat Adiponitrile (ADN)

Fig. 15 shows the cyclic voltammetry scan of 1 M LiTFSI-ADN electrolyte on a Pt microelectrode (25 μm). The scan indicates that there is an electrochemical window of 6 V positive of LiVLi ⁰ within which there is no apparent oxidation or reduction currents. This value is much higher than the anodic stability of commercial electrolytes (LiPF ₆ : EC/DMC ~ 4.5 V) and slightly higher than that of reported sulphone-based ones (LITFShEMS ~ 5.9 V) [12].

Example 7: Cyclic Voltammetry in a Mixture of Adiponitrile (ADN) and Co-solvent

Figs. 16-19 show the cyclic voltammetry scans of ADN/co-solvent electrolyte mixtures. In all cases, the ratio of ADN:co-solvent was 1 :1 v/v. Scan rates were 20 mV/s and measurements taken at ambient temperature in all cases. In all cases, the concentration of ionic salt in the electrolyte was 1 M.

Referring to Fig. 16, in an LiTSFI-ADN electrolyte the addition of 3- methoxypropionitrile (MPN) as co-solvent decreased the anodic stability slightly to 5.7 V, compared to the 6 V for neat ADN .

Referring to Fig. 17, the addition of γ-butyrolactone (γ-BL) as co-solvent in LiTSFI-ADN electrolyte decreased the anodic stability even more significantly to 4.1 V rendering it less suitable for high voltage battery testing.

Referring to Fig. 18, lithium bis(oxalato)borate (LiBOB) salt was used as the ionic salt in an ADN-γ-BL electrolyte instead of LiTFSI. With LiBOB, the ADN-γ-BL electrolyte solution was found to have an anodic electrochemical window of stability of 5.1 V.

Referring to Fig. 19, the use of ethylene carbonate (EC) as a co-solvent was also investigated and we found that it had a similar cyclic voltammogram to that of neat ADN with good anodic and cathodic stabilities.

Example 8: Conductivity in ADN-based Electrolytes

The conductivity of LiTFSI-adiponitirle (ADN) electrolyte solution along with that of solutions containing other nitriles as co-solvents was measured as a function of temperature and concentration.

Referring to Fig. 20, the conductivity as a function of concentration of LiTFSI in adiponitrile, alone and with various co-solvents in 1 :1 by volume mixtures is shown. The conductivity was measured using the impedance spectroscopy technique between 100 Hz and 1 MHz in a two-platinum-electrode conductivity cell with a cell constant of 0.96. In general, all the solutions showed the characteristic maximum in conductivity. In the case of LiTFSI-adiponitrile electrolyte, the maximum occurs at around 1.25 M LiTFSI with an ionic conductivity of 2.9 mS/cm.

Referring to Fig. 21 , the conductivity of 1 M LiTFS l-ADN-based electrolyte solutions as a function of temperature are shown. The conductivity was measured using the impedance spectroscopy technique between 100 Hz and 1 MHz in a two- platinum-electrode conductivity cell with a cell constant of 0.96. Certain selected results are also summarized in Table 1. The conductivity of 1 M LiTFSI-adiponitirle (ADN) increased from 0.45 mS/cm at -20 ⁰ C to 1.7 mS/cm at 20°C reaching 8 mS/cm at 8O ⁰ C. In general, all the co-solvent mixtures studied had resulted in an increase in conductivity throughout the temperature range. Of particular interest is acetonitrile (ACN) which showed the highest reported conductivity of all the solutions. However its higher volatility is an issue for some applications. EC is also of interest due to its very good SEI formation properties on graphite and due to the good conductivities obtained.

Table 1

Example 9: Stability of ADN-based Electrolytes Against Aluminum Corrosion

The behavior of an aluminum current collector was evaluated in 1 M LiTFSI/adiponitrile-based electrolyte solutions. In some cases, neat adiponitrile (ADN) was used as the solvent, and in other cases, mixtures of ADN with various co- solvents (at 1 :1 v/v ratio) were used as the solvent. In one case, 0.1 M LiBOB was also present. Cyclic voltammetry (CV) using aluminum (Al) wire as a working electrode was used to measure stability. The results are shown in Figs. 22-26.

In general, the CV scans showed a hysteresis loop in the first cycle indicating pitting corrosion of aluminum that has diminished largely by the third cycle. The end of the reverse anodic scan, when it crosses the voltage axis, gives the repassivation potential (ER), characteristic of the electrolytes and the ability of its decomposition product to form a passivation layer to protect Al from further corrosion. All the electrolytes, except for EC/ADN showed an ER value of 4.7 V which is much higher than the 3.65 V [13] observed in LiTFSI solution in EC:DMC solvent system. The scans demonstrate that after the third cycle, passivation takes place and the electrolytes are capable of protecting Al up to voltages reaching as high as 5 V, as seen in the case of ADN.

Example 10: Electrochemical Performance Using Adiponitrile (ADN)

The cycling performance of a meso carbon micro bead (MCMB)/LiCoO ₂ battery assembled using 1 M LiTFSI in 1 :1 EC/ADN electrolyte, with 0.1 M LiBOB as an additive, at a C/12 rate is shown in Fig. 27. It shows an initial charge capacity of 145 mAh/g and discharge capacity of 108 mAh/g for the first cycle. On cycling the battery further, the capacity stabilizes around 100 mAh/g. The charge/discharge efficiency is above 99% after the 3rd cycle. Example 11: Comparison to US 7,226,704

Dinitrile-based electrolytes in accordance with the present invention were compared to an exemplified electrolyte system disclosed in US 7,226,704 to show that the present dinitrile-based electrolyte systems provide for electrochemical cells with improved electrochemical window of stability, improved specific capacities, improved resistance to aluminum corrosion, and improved thermal stability in comparison to similar systems that do not contain dinitrile.

Solubility of LiBOB in Adiponitrile:

The solubility of LiBOB in adiponitrile and in adiponitrile/ethylene carbonate mixtures was first assessed to ensure that LiBOB would be soluble in test electrolytes. Table 2 provides the maximum solubility of LiBOB in various solvent systems.

Table 2

Preparation of Test Electrolytes:

TABLE 1 of US 7,226,704 provides electrolyte compositions used in the patent. Example 4 from TABLE 1 of US 7,226,704 was chosen for comparison to the present invention as this example provided the best electrochemical properties. Example 4 from US 7,226,704 is an electrolyte comprising 0.77 mol/kg (~1 M) LiBOB in 29.7 wt% ethylene carbonate (EC), 38.0 wt% ethyl acetate (EA) and 17.0 wt% v- butyrolactone (BL).

Three test electrolytes of the present invention were prepared on the basis of

Example 4 from US 7,226,704 except that adiponitrile (ADN) was used in place of one or two of the co-solvents as follows: Example I: ADN instead of EA. Example II: ADN instead of BL. Example III: ADN instead of EA and BL.

Conductivity:

The conductivity of the electrolytes is shown in Table 3. Conductivity was measured using the impedance spectroscopy technique between 100 Hz and 1 MHz in a two-platinum-electrode conductivity cell with a cell constant of 0.96. The replacement of adiponitrile (ADN) decreased the conductivity as expected due to its high viscosity.

Table 3

Cyclic voltammetry (oxidative stability):

Electrolytes of the present invention containing adiponitrile (ADN) all showed a wide electrochemical stability extending to values in excess of 6 V (see Fig. 28 for

Example III), which is at least one volt more than that of Example 4 of US 7,226,704 which shows a stability of 5 V. This allows for the use of high voltage cathodes that could not be otherwise used in conventional electrolytes like Example 4 of US 7,226,704. Such high voltage cathodes that may now be used with electrolytes of the present invention include, for example, LiNiPO ₄ (5.3 V). Similar cyclic voltammograms were obtained for Example I and Example II. Cyclic voltammograms were measured using a Pt microelectrode (25 μm) and a silver wire as counter and pseudo reference electrodes with a scan rate of 20 mV/s.

Aluminum corrosion:

Corrosion of aluminum (Al) in electrolytes of Example 4 of US 7,226,704 and of Example Il was investigated using an aluminum wire (100 μm) as a working electrode and a silver wire as counter and pseudo reference electrodes with a scan rate of 10 mV/s. This a property of the salt and not the solvent, but the solvent can play a role in the inhibition of corrosion. LiTFSI is known to corrode Al current collectors and it has been found that dinitriles inhibit this corrosion by at least 1 V as shown in Example 9 above.

In respect of LiBOB salts, it is apparent from Figs. 29 and 30 that LiBOB, which was not known to corrode Al, shows some corrosion behavior. Although the current is very small, the electrolyte of Example Il shows a decrease in the corrosion current of at least one order of magnitude compared to the corrosion current in the electrolyte of Example 4 of US 7,226,704 (compare Fig. 29 to Fig. 30).

Thermal stability:

For differential scanning calorimetry (DSC) studies, a hermetically sealed pan was slowly cooled to -100 ⁰ C and then heated to 200 ⁰ C at a scan rate of 10°C/min. From Fig. 31 it is apparent that the some components of the electrolyte of Example 4 of US 7,226,704 start to evaporate at 150 ⁰ C, while the electrolyte of Example III shows little evaporation and good stability up to 200°C. Example I and Example Il have shown similar DSC results to Example 4 of US 7,226,704 due to the low boiling point of EA and BL.

Electrochemical cell performance:

Two types of electrochemical cells were constructed as previously described.

One type had an anode comprising graphite intercalated with lithium and a cathode comprising LiCoO ₂ . The other type has an anode comprising Li metal and a cathode comprising LiCoO ₂ . Electrochemical cells using an electrolyte in accordance with Example III can show higher specific capacities than those using the electrolyte of Example 4 of US 7,226,704 (see Figs. 32 and 33), especially when graphite intercalated with graphite is used as the anode.

Thus, the liquid electrolytes of the present invention combining ionic salts of low corrosivity and electrochemical stability suitable for usage in lithium secondary cells, with a dintrile-based solvent system achieve electrochemical performances equal to or better than conventional electrolytes while providing the benefits of increased potential window of electrochemical stability, better thermal stability and better resistance to aluminum corrosion.

References:

1. A. Hammami, N. Raymond, and M. Armand, Nature, 424, 635 (2003). 2. P.J. Alarco, Y. Abu-Lebdeh, A. Abouimrane, M. Armand, Nature Materials, 3,

476 (2004).

3. A. Abouimrane, Y. Abu-Lebdeh, P.J. Alarco and Michel Armand, J. Electrochem. Soc, 151 (7), A1028 (2004).

4. CA 2,435,218, Y. Abu-Lebdeh et al., Jan. 28, 2005. 5. WO 0115258, D.R. MacFarlane et al., Mar. 1 , 2001.

6. WO 2007012174, A. Abouimrane et al. , Feb. 1 , 2007.

7. DE 19829030, U. Lischka et al., 1999.

8. JT. Dudley, DP. Wilkinson, G. Thomas, R. LeVae, S. Woo, H. Blom, C. Horvath, M. W. Juzkow, B. Denis, P. Juric, P, Aghakian and J. R. Dahn, J. Power Sources, 59-82, 31 (1991 ).

9. J. M Tarascon and D. Guyomard, Solid State Ionics, 293-305, 69 (1994).

10. S.S. Zhang, Electrochemistry Communications, 1423-1428, 8 (2006).

11. US 7,226,704, J-C. Panitz et al., Jun. 5, 2007.

12. K. Xu and CA. Angell, J. Electrochem. Soc. 145 (4) p L70 (1998).

13. L. Peter and J. Arai, J. Appl. Electrochem, 29, 1053 (1999).

Other advantages which are inherent to the structure are obvious to one skilled in the art. The embodiments are described herein illustratively and are not meant to limit the scope of the invention as claimed. Variations of the foregoing embodiments will be evident to a person of ordinary skill and are intended by the inventor to be encompassed by the following claims.