WO1996024956A1 - Nonaqueous electrolyte secondary battery - Google Patents

Abstract

Composite intercalation electrodes are disclosed which include a mixture of homogeneous graphitic-carbon particles, homogeneous non-graphitic carbon particles, and binders as necessary. By choosing a particular type of non-graphitic carbon and adjusting the ratio of graphitic to non-graphitic carbons, composite electrodes can be formulated to have high capacities, low electrode potentials, and other desirable properties of graphite, and, at the same time, have discharge profiles in which the electrode potential varies significantly with the degree of intercalation. The homogeneous non-graphitic carbon particles are obtained by pyrolysis of one of the following materials: coke, petroleum residues, acrylic resins (e.g., polyacrylates, polyacylonitrile, polymethylacrylonitrile, etc.), polydivinyl benzene, polyvinyl chloride, furfuran resins (e.g., polyfurfural, polyfurfurol, furfural-phenol, etc.), phenolic resins, resorcinol-formaldehyde resins, polyimide resins, and cyclic hydrocarbons containing at least two rings (e.g., naphthalene, anthracene), and various benzene derivatives.

Description

NONAOUEOUS ELECTROLYTE SECONDARY BATTERY

Description

TECHNICAL FIELD

The present invention relates to carbon-based electrodes for use in electrochemical energy storage devices. More particularly, the invention relates to intercalation electrodes containing a mixture of both graphitic and substantially non-graphitic carbons.

BACKGROUND ART

There is an increasing demand for rechargeable cells of higher energy density and specific energy due to the increasing demand for portable electronic equipment. In order to meet this demand, various type of rechargeable cells have been developed including improved nickel- cadmium aqueous batteries, various formulations of aqueous nickel metal hydride batteries, and, most recently, nonaqueous rechargeable lithium cells.

Although rechargable lithium metal cells (cells employing a lithium anode) have high energy densities and specific energies, they have not gained wide-spread acceptance because they have a poor cycle life, rate, and safety characteristics. As a result, several groups have developed rechargeable battery systems based on carbon anodes which intercalate lithium. Cells based upon such carbon intercalation systems are commonly referred to as the "lithium ion," "lithium rocking chair," or "lithium intercalation" cells. Although such cells have a lower theoretical energy density than lithium metal cells, they are inherently safer and more rechargeable due to their intercalation energy storage mechanism.

Not surprisingly, various carbon-based materials have been proposed for use in the anodes in lithium intercalation cells. Unfortunately, most of these materials have one or more disadvantages as discussed below. The proposed anode materials can be roughly divided into two classes: (1) graphite (or "graphitic" carbons), and (2) less ordered carbons (or "non-graphitic" carbons). U.S. Patent No. 4,423,125 issued December 27, 1983 to S. Basu is one patent describing a cell including a lithium intercalated graphite anode. Various patents describing non- graphitic carbon electrodes will be discussed below.

Graphite intercalation electrodes have relatively high capacities and very negative potentials (corresponding to high cell voltages). Specifically, lithium intercalates in graphite to the theoretical maximum of one lithium ion to every six carbon atoms (corresponding to a capacity of 372 mA*hr/gm carbon). Further, the majority of the lithium/graphite intercalation in graphite occurs at an open circuit potential spanning a range of less than 100 mV, centered near 175 mV versus lithium metal. While graphite electrodes have both high capacity and low voltage, they suffer from significant technical problems. For example, graphite undergoes exfoliation when intercalated with lithium from certain desirable electrolyte systems (most notably systems containing propylen carbonate). Various alternative electrolytes have been proposed to overcome this problem. For example, U.S. Patent No. 4,980,250 issued to Y. Takahashi et al. describes the use of 1,3 dioxolane with LiAsFό as a dissolved salt. U.S. Patent No. 5,028,500 issued to Fong et al. describes a procedure of intercalating graphitic with lithium in a propylene carbonate/ethylene carbonate electrolyte at an elevated temperature. Further, U.S. Patent 5,130,211 issued to Wilkinson et al. discusses the addition of "sequestering agents" (e.g., glymes, crown ethers, or cryptands) to a propylene carbonate based electrolyte to reduce the tendency for graphite to undergo exfoliation.

While the above approaches have mitigated the problem of exfoliation, other problems associated with graphite-based intercalation electrodes remain. For example, graphite based cells can generally be expected to perform poorly at high discharge rates. This is because graphite intercalation electrodes have, as noted, a nearly constant open-circuit potential versus the level of intercalation (i.e., they exhibit small variations in voltage with state of charge). (See Fig. 2, curve 3 of this disclosure, for example.) This lack of change in potential with state of charge causes under-utilization of the electrode material, limiting the rate, energy, and cycling performance of cells. This view is supported by a phenomena which was recently reported by T. Fuller et al. (J. Electrochem. Soc., 1, 114 (1994) incorporated herein by reference for all purposes). This paper describes a finite element model used to analyze the polarization of a hypothetical lithium-ion cell having an intercalation anode containing carbon derived from a coke and a metal oxide cathode containing LiMn2θ4.

The coke-based electrode employed in the model was a non-graphitic carbon exhibiting an open circuit potential profile which varies by 1.2 volts with state of charge (unlike graphite).

Therefore, the anode would not be expected to limit the high rate capabilities of the modeled coke electrode system. On the other hand, the LiMn2θ4 cathode's open circuit potential profile is substantially invariant with the state of charge, thus presenting the possibility of poor high rate performance. The simulations of Fuller et al. showed that this was indeed the case. Specifically, an increase in concentration overpotential near the separator/oxide interface limited the high rate discharge (in their model cell) because of the depletion of lithium ions in the electrolyte. Further, the analysis of the current distribution in the porous cathode showed the importance of the metal oxide's open-circuit potential's relatively small rate of change with the state of the overall cell performance. Typical cathode materials proposed for use in lithium ion batteries (e.g., LiMn2θ4, LiCoθ2, LiNiθ2, and atomic mixtures of Mn, Co, and Ni oxides) all have substantially invariant open-circuit potentials over a wide range of state of charge. Graphite's open-circuit potential profile also varies very little with state of charge. Therefore, it can be expected that combining a graphite anode with a standard oxide cathode would result in a cell with a significantly reduced rate capability.

As mentioned, various non-graphitic carbons have been used as lithium intercalating anodes. While these materials do not suffer from the problem of a flat discharge profile (and can be utilized as anodes in lithium-ion batteries), they generally have relatively lower capacities and higher average discharge voltages (vs. a lithium metal reference electrode) than would be optimally desirable. (It should be understood that low voltage, high capacity anodes coupled with high capacity, high voltage anodes result in the highest voltage and energy cells.) Generally, non- graphitic carbons are produced by the pyrolysis of certain organic materials in an inert atmosphere at temperatures under 2000° C. Exemplary patents describing various non-graphitic intercalation electrodes include U.S. Patent No. 4,668,595 issued to A. Yoshino et al., U.S. Patent No. 4,702,977 issued to Hiratsuka et al. (describing a carbonaceous material obtained by carbonizing a "pseudographitic structure"), U.S. Patent No. 4,863,814 issued to M. Mohri et al., U.S. Patent No. 4,945,014 issued to Miyabayashi et al., and U.S. Patent No. 4,959,281 issued to Nishi et al. In addition, the above-noted U.S. Patent No. 5,028,500 describes the use of a particulate carbon in which substantially every particle includes two phases intimately admixed, with the first phase having a higher degree of graphitization than the second phase. The above patents recite a litany of physical properties used to characterize the various carbon materials they employ as anodes.

U.S. Patent No. 5,093,216 issued to Azuma et al. describes a method of increasing the capacity of non-graphitic carbon by adding a phosphorous containing precursor to furfuran resins, phenolic resins and oxygen crosslinked petroleum cokes. U.S. Patent No. 5,358,802 issued to Mayer et al. describes phosphorous doping of several polymeric precursors of carbon which substantially increases the discharge capacity of the resulting carbon material. However, it has been determined that, although the lithium storage capacity can be substantially increased via phosphorous doping, the average discharge voltage is also substantially higher than that of both graphite and undoped non-graphitic materials. Thus, cells containing such doped non-graphitic anodes will have lower-than-desirable average battery voltages and energies.

In addition to the above described difficulties inherent in known anode materials, some cells suffer from the additional problem of being fabricated from pre-intercalated or lithiated carbons. For example, the above mentioned U.S. Patent No. 4,423,125 issued to S. Basu describes a method of making lithium intercalated graphite (by placing graphite in molten lithium under an inert atmosphere). The resulting lithium intercalated graphite anode is then combined with a cathode, an electrolyte, and a separator to form a cell. A similar approach of using prelithiated material, but using a non-graphitic carbon instead of graphite has been described in the above mentioned U.S. Patent No. 4,980,250 issued to Y. Takahashi et al. In either case, the lithiated carbon is unstable in air, and therefore a cell using the above described materials must be handled completely in an inert atmosphere environment, increasing units costs and raising safely concerns.

Accordingly, there is a need for carbon-based lithium intercalation electrodes which overcome many of the problems associated with prior graphitic and non-graphitic carbon electrodes.

DISCLOSURE OF THE INVENTION

The present invention is directed to composite electrodes including mixtures of homogeneous graphitic carbon particles, homogeneous non-graphitic carbon particles, and binder as necessary. Such electrodes can be formulated to have high capacities, low electrode potentials, and other desirable properties of graphite, and, at the same time, have discharge profiles in which the electrode potential varies significantly with the degree of intercalation. Thus, lithium ion cells employing the composite electrodes of this invention will perform well at high rates of discharge. Further, some electrodes of this invention have been found to have unexpectedly high capacities which are, in fact, greater than that of either the homogeneous graphitic carbon particles or homogeneous non-graphitic carbon particles that make up the composite electrode. Such electrodes will be discussed further below in the Examples section of the application.

The term "mixture" is used herein in the sense commonly employed in the chemical arts. Thus, a mixture of carbons in accordance with this invention is composed of distinct chemical species, and, in theory, can be separated by physical means. Typically, the mixtures of this invention will include "particles" of graphitic carbon interspersed with "particles" of non-graphitic carbon. Various forms of both these carbon particle types may be employed in the electrodes of this invention. Such particles may each assume various shapes such has fibers, plates, spheres, crystallites, etc. In each case, the morphology of the particle may be somewhat smooth, rough, jagged, porous, fractious, etc. Still further, the size and size distribution of the particles may vary widely from dust at one extreme to large continuous structures at the other extreme. In the latter case, the carbon form making up the large structure will be interspersed with smaller particles of the other carbon form.

In preferred embodiments, the homogeneous graphitic carbon particles will have an average particle size of between about 0.5 to 50 μm, and more preferably between about 5 and 20 μm. Further, these preferably have highly anisotropic morphology, such as a plate-like structure. In further preferred embodiments, the homogeneous non-graphitic carbon particles will have an average particle size of between about 0.1 to 100 μm, more preferably between about 1 to 50 μm, and most preferably between about 5 to 20 μm. In still further preferred embodiments, the homogeneous non-graphitic carbon particles are obtained by pyrolysis of one of the following materials: coke, petroleum residues, acrylic resins (e.g., polyacrylates, polyacylonitrile, polymethylacrylonitrile, etc.), polydivinyl benzene, poly vinyl chloride, furfuran resins (e.g., polyfurfural, polyfurfurol, furfural-phenol, etc.), phenolic resins, resorcinol-formaldehyde resins, polyimide resins, and cyclic hydrocarbons containing at least two rings (e.g., naphthalene, anthracene), and various benzene derivatives. Copolymers including monomers of the above- listed polymers and/or other monomers may also be used. The non-graphitic carbon particles obtained from such products may be doped with a dopant element selected from the elements in groups HI and IV of the periodic table. It has been observed that doping with phosphorous generally improves the capacity of the electrode but increases its potential (thus decreasing the cell potential).

One general advantage of this invention is the flexibility it affords in preparing anodes having desirable discharge characteristics such as a high capacity and a potential that slopes with state of intercalation. The flexibility is provided because various forms of two different materials (graphitic carbon and non-graphitic carbon) can be selected and mixed in appropriate ratios to obtain the desired performance. For many electrode systems, it has been observed that a carbon electrode including at least about 25 weight percent homogeneous graphitic carbon particles provides good performance. Still better performance may be achieved with mixtures including at least about 50 weight percent (more preferably about 75 weight percent) homogeneous graphitic carbon particles. Of course, the actual ratio may vary quite a bit depending upon the carbon constituents of the mixture and the desired properties of the electrode. Generally, the mixture should be chosen such that the resulting electrode has an open circuit potential of that varies by at least about 0.25 volts from a fully charged state in which the electrode is fully intercalated to a state of charge at about 90% deintercalation. Further, the mixture should provide a relatively high volume average composite density electrode, e.g., at least about 1.2 g/cc. High density electrodes permit fabrication of cells having greater quantities of energy for a given cell volume. Further, because of their better mechanical integrity, they may also have longer cycle lives in some cells. Specifically, some cells employing prior carbon anodes had to be used in a rolled compressed form (know as the "jelly roll" design) in order to exhibit good cycle life. The electrodes of this invention do not require such cell designs to deliver good cycle life.

One aspect of the invention is directed to a lithium ion cell which includes the following elements: (1) a container, (2) an anode including a mixture of homogeneous graphitic carbon particles and homogeneous non-graphitic carbon particles, the anode being capable of intercalating lithium during charge and deintercalating lithium during discharge, (3) a cathode capable of reversibly taking up lithium on discharge and releasing lithium on charge, and (4) an electrolyte conductive to lithium ions, wherein the anode, cathode, and electrolyte are provided within the container. Preferably, the lithium ion cells are constructed with carbon anodes which are initially provided in a deintercalated state. As noted above, there are special problems associated with handling pre-intercalated anodes. The cathode employed in the cell preferably includes one of lithium nickel oxide, lithium cobalt oxide, lithium manganese oxide (LiMnθ2 or LiMn2θ4), or a chemical or physical mixture of two or more of these materials. The electrolyte may include one more of the following: propylene carbonate, ethylene carbonate, 1, 2-dimethoxyethane, 1,2- diethoxyethane, γ-butyrolactone, tetrahydrofuran, 2-methyltetrahydrofuran, 1,3-dioxolan, 4- methyl 1,3-dioxolan, diethyl ether, sulfolane, acetonitrile, propionitrile, glutaronitrile, dimethyl carbonate, diethyl carbonate, anisole, and mixtures thereof. The electrolyte may further include one or more of die following salts: lithium bis-trifluoromethane sulfonimide (Li(CF3SO2)2N), LiAsF6, LiPF6, LiBF4, LiB(C6H5)4, LiCl, LiBr, CH3SO3Li, and CF3SO3Li. In particularly preferred embodiments, the electrolyte includes a mixture of ethylene carbonate and dimethyl carbonate and dissolved Li(CF3SO2)2N (about 0.5 to 1 M) and dissolved LiAsFό or LiPFδ

(either of which is present in a concentration of about 0.1 to 0.4 M). The total concentration of Li(CF3SO2)2N and LiAsFδ or LiPF6 should not exceed about 1M.

These and other features of the present invention will be presented in more detail in the following specification of the invention and in the figures.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 is an illustration of an experimental apparatus employed to test the carbon-based intercalation electrodes of the present invention.

Fig. 2 is a graph displaying voltage (versus a lithium reference electrode) as a function of fractional lithium deintercalation for three electrodes: (1) a graphite electrode, (2) a non-graphitic electrode (made from pyrolyzed phenolic resin), and (3) a composite electrode containing graphiti and non-graphitic (phenolic) carbons.

Fig. 3 is a graph of voltage (versus a lithium reference electrode) as a function for fractional deintercalation of three electrodes: ( 1 ) a graphite electrode, (2) a non-graphitic carbon electrode (made from pyrolyzed polyfurfuryl alcohol), and (3) a composite electrode containing graphitic and non-graphitic carbons.

Fig.4 is a graph of voltage (versus a lithium reference electrode) as function of fractional deintercalation for three electrodes: (1) a graphite electrode, (2) a non-graphitic carbon electrode (made from pyrolyzed resorcinol-formaldehyde resin), and (3) a composite electrode made from the graphitic and non-graphitic carbons.

Fig. 5 is a graph of voltage (versus a lithium reference electrode) as a function of fractiona deintercalation for three electrodes: ( 1 ) a graphite electrode, (2) a non-graphitic carbon electrode (made from pyrolyzed PAN fibers), and (3) a composite electrode made from the graphitic and non-graphitic carbons.

Fig.6 is a graph of voltage (versus a lithium reference electrode) as function of fractional deintercalation for three electrodes: (1) a graphite electrode, (2) a non-graphitic carbon electrode (made from pyrolyzed CIPS PAN), and (3) a composite electrode made from the graphitic and non-graphitic carbons.

Fig. 7 is a graph of voltage (versus a lithium reference electrode) as a functional of fractional deintercalation for three electrodes: (1) a graphite electrode , (2) a non-graphitic carbon electrode (made from twice-fired phenolic resin), and (3) a composite electrode made from the graphitic and non-graphitic carbons.

Fig. 8 is a graph of voltage (versus a lithium reference electrode) as function of fractional deintercalation for three electrodes: (1) a graphite electrode , (2) a non-graphitic carbon electrode (made from twice-fired polyfurfuryl alcohol), and (3) a composite electrode made from the graphitic and non-graphitic carbons.

Fig. 9 is a graph of voltage (versus a lithium reference electrode) as a function of fractional deintercalation for three electrodes: (1) a graphite electrode , (2) a non-graphitic carbon electrode (made from twice-fired resorcinol-formaldehyde resin), and (3) a composite electrode made from the graphitic and non-graphitic carbons.

Fig. 10 is a graph of voltage (versus a lithium reference electrode) as a function of fractional deintercalation for three electrodes: ( 1) a graphite electrode , (2) a non-graphitic carbon electrode (made from twice-fired PAN fibers), and (3) a composite electrode made from the graphitic and non-graphitic carbons.

Fig. 11 is a graph of voltage (versus a lithium reference electrode) as a function of fractional deintercalation for a series of electrodes made from mixtures of graphite and CIPS PAN in varying ratios.

Fig. 12 is a graph comparing the cycleability of composite and non-composite electrodes.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

Carbon anode electrodes of this invention include a mixture of two or more distinctly different carbonaceous materials. At least one of these is a highly ordered pyrolytic or natural graphite, and at least one other of these is a less-ordered non-graphitic carbon derived from the pyrolysis of organic compounds. The combination of these components in lithium intercalation electrodes overcomes many of the traditional problems associated with intercalation electrodes made from either graphite or non-graphitic carbon alone. For many electrode systems, the carbon mixture should include at least about 25 weight percent homogeneous graphitic carbon particles, more preferably at least about 50 weight percent homogeneous graphitic carbon particles, and most preferably about 75 weight percent homogeneous graphitic carbon particles. Of course, the optimal ratios may vary quite a bit depending upon the carbon constituents of the mixture and the desired properties of the electrode. It is generally desirable that the mixture result in electrodes having a potential which varies significantly with state of charge (state of deintercalation). Preferably, the mixture should be chosen such that the resulting electrode has an open circuit potential of that varies by at least abou 0.25 volts from a fully charged state in which the electrode is fully intercalated to a state of charge at about 90% deintercalation. The pure graphite intercalation electrode generally varies by only about 180 mV during discharge.

In further preferred embodiments, the carbons and mixing ratios are chosen such that the mixture has a relatively high density electrode, e.g., at least about 1.2 g/cc. High density electrodes permit fabrication of cells having greater quantities of energy for a given cell volume. I is also believed that they may also have longer cycle lives in cell designs which do not compress the electrodes (e.g., jelly roll designs).

The above-described graphitic carbon component should be a highly-ordered pyrolytic or natural graphite having a particle size of between about 0.5 to 50 μm (more preferably between about 5 and 20 μm). Suitable graphitic materials are expected to be highly crystalline and rather homogeneous. Thus, they will typically exhibit narrow or sharp X-ray diffraction peaks. Any dispersion (widening) observed in such peaks typically will be caused by intraparticulate scatterin due to the polycrystalline graphite's finite crystallite size parameter __. It should be noted that the diffraction patterns of most non-graphitic carbons exhibit significant dispersion because such carbons have a range of carbon-carbon interatomic distances. Graphite, in contrast, has rather narrow ranges of interatomic distances due to its highly crystalline structure. Graphites suitable for use in this invention typically will have Lς values of greater than about 100 A and interlayer d()02 spacings of around 3.34 A. In a preferred embodiment, the graphite used in this invention i a high purity natural graphite or a synthetic graphite having a high degree or anisotropic morphological structure similar to natural graphite and very good compressibility and electrical conductivity (e.g., SFG synthetic Graphites from Lonza Inc. of Fairlawn, NJ. Specific example of suitable materials include commercially available graphites such as Graphite KS (a round shaped particle) and Graphite T (having a flake-shaped particle with higher surface area) from Lonza Inc., or grade B6-35 or 9035 from Superior Graphite Co. of Chicago, 111. It is also within the scope of this invention to produce pure or relatively highly pure graphitic electrode materials b pyrolysis at temperatures of greater than about 2300°C.

Non-graphitic carbons of widely ranging properties may be employed in this invention. I general, the non-graphitic carbons should provide intercalation electrodes having sloping deintercalation profiles. The intercalation electrodes should also have a reasonably high capacity and a reasonably low voltage. The preferred particle size range is between about 0.1 to 100 μm, more preferably between about 1 to 50 μm, and most preferably between about 5 to 20 μm. A defining characteristic of the non-graphitic particles employed in this invention is their relative disorder in comparison to the graphitic carbon particles. Thus, for example, diffraction patterns of the non-graphitic particles will have relatively wide peaks, evidencing a spectrum of interatomic spacings. Although the non-graphitic materials used in this invention should be relatively homogeneous (in a materials sense), they may include localities having greater degrees of graphitic character. In general, while non-graphitic carbons are less ordered than graphite, they do contain a degree of crystalline order.

Sources of such non-graphitic carbon compounds include various petroleum and coke products and polymers including various acrylic resins (including polyacrylates, polyacylonitrile, polymethylacrylonitrile), polydivinyl benzene, polyvinyl chloride, furfuran resins (including polyfurfural, polyfurfurol, furfural-phenol, etc.), phenolic resins (including resorcinol- formaldehyde resins), polyimide resins, and cyclic hydrocarbons containing at least two ring structures (e.g., naphthalene, anthracene), and various benzene derivatives. Copolymers including monomers of the above-listed polymers and or other monomers may also be used. The non-graphitic carbon particles obtained from these may be doped with a dopant element selected from the elements in groups HI and IV of the periodic table. It has been observed that particularly high capacity electrodes can be made from non-graphitic carbons derived from doped polyfurfurol or doped polyacrylonitrile materials.

The non-graphitic carbon precursors are preferably pyrolyzed in an inert atmosphere (e.g., Ar, N2, He, Ne) or under vacuum at between about 600" to 2000°C, more preferably between about 750 and 1400°C, and most preferably between about 900 and 1150°C. In these temperature ranges, the resulting carbon material will not be highly graphitic. In a particularly preferred embodiment, the pyrolysis is conducted in a retort furnace under flowing nitrogen to maintain a positive pressure in the pyrolysis chamber. The pyrolysis time will generally be less than 8 hours, depending, of course, upon the pyrolysis temperature, amount of material being pyrolyzed, and the chemical composition of the carbon precursor.

In many cases, the carbon material resulting form pyrolysis will need to be ground and sieved before it is mixed with graphite. It has been found that for many pyrolysis products, a two stage grinding and sieving process works well. In such cases, the pyrolysis product is first ground and sieved and then reground and resieved. In preferred embodiments, the first grinding step is performed with a hammer mill, or, for small scale applications, with a mortar and pestle. The resulting particles are then sieved to a size of, for example, less than 300 μm. Next, a second grinding step is performed by attrition or ball milling, preferably under argon or other inert gas. Thereafter, the resulting particles are sieved to yield particles of 2 to 60 micrometers in diameter, for example.

Typically, non-graphitic carbon particles are prepared by a single pyrolysis step. However, in some preferred embodiments of this invention, the ground and sieved carbon particles are refired under the pyrolysis conditions that promote the formation of non-graphitic carbon. It has been found, for example, that by refiring a sample after grinding, the resulting electrode will exhibit significant reductions in irreversible capacity loss with a lower average voltage, while maintaining a good reversible capacity. Examples 18-31 below describe electrodes made from refired non-graphitic carbons.

In some preferred embodiments, the non-graphitic carbon precursor is combined with a phosphorous-containing compound prior to the pyrolysis. Examples of suitable phosphorous- containing compounds include phosphorus oxides such as phosphorous tetraoxide, phosphorous pentoxide, or phosphorous trioxide, phosphoric acid group materials such as ortho-phosphoric acid, meta-phosphoric acid, polyphosphoric acid (anhydrous phosphoric acid), and the various salts of phosphoric acid (e.g., Li3PO4, Na3PO4, (NH4)2HPO4, etc.). The resultant

"phosphorous doped" material (when observed without mixing with graphite to make a composite exhibits higher discharge capacities than undoped materials, but at a higher average open circuit voltage versus a lithium metal reference electrode.

In preferred embodiments, the phosphorous is provided as a solution of about 0.1 to 5 percent phosphoric acid by weight in water, a ketone, or an alcohol. For example, phosphorous may be provided in a methanol solution containing dissolved phosphoric acid at a concentration of about 0.1-5 percent by weight. This solution is applied to a polymer or other carbonaceous substance which is subsequently dried before pyrolysis. Drying will leave an ortho-phosphoric acid residue on the polymer surface. During subsequent pyrolysis, the phosphoric acid decomposes, leaving phosphorous atoms which diffuse into the bulk polymer to give a phosphorous-doped carbon electrode material. In general, the dopant material should be provided in a solvent or other carrier which does not dissolve the polymer.

Although phosphorous is a preferred dopant, other dopant atoms may also be employed. In general, doping in the context of this invention refers to donor or acceptor dopants which are integrated into the carbon matrix. Preferably the donor or acceptor dopants are selected to be from group IHA (for acceptors) or from group VIA (for donors) of the periodic chart. Thus, suitable acceptor dopants are boron, aluminum, gallium, indium, and thallium, and suitable donor dopants include phosphorous, arsenic, antimony, and bismuth. In some cases, dopant atoms from other groups may be appropriate, such as, for example, sulfur. In particularly preferred embodiments, the dopant is phosphorous, boron, arsenic, or antimony. The graphitic and non-graphitic particles described above can be formed into an intercalation electrode by various techniques. Generally, they must be mixed with a binder to facilitate formation of an intercalation electrode. Preferably, a slurry mixture of the carbons is prepared by the addition of a solvent containing a dissolved polymer (the binder) which is substantially unreactive and insoluble in the electrolyte at the voltages which the anode experiences within the cell. Suitable binders include ethylene propylene diene monomer (with cyclohexane as a solvent) and polytetrafluoroethylene (Teflon™). In a particularly preferred embodiment, the polymer is polyvinylidene difluoride (PVDF, melting point 171° C) and the slurry solvent is dimethylformamide (DMF, boiling point 153°C).

In one relatively simple process for preparing electrodes, the carbon slurry is applied to a metal support which acts as a current collector for the completed electrode. Preferably, the slurry is first applied as a thin film onto a copper foil substrate, the solvent of the slurry is then evaporated, the temperature of the composite is then heated to the melting point of the polymer binder, and finally the composite is compressed onto the foil (e.g., by using a roll press). The resulting structure is then simply sized for use in an electrochemical cell, and optionally formatted or preprocessed in another manner to provide the desired physical-chemical properties of an electrode. Such procedures are well known to those in the skill of the art. In a further preferred embodiment, the composite electrode is reheated after the compression step to allow the polymer binder to melt a second time. The composite may then compressed onto the foil a second time (e.g., by using a roll press).

Various current collectors may be employed with the electrodes of the present invention. Preferably, the current collector is a metal foil, metal screen, or an expanded metal screen (or "Exmet" ™). If the current collector is a foil, adhesion of the composite carbon/binder mixture to the current collector may be enhanced by roughening the current collector's surface. Suitable methods of roughening the surface include mechanical roughening (e.g., with steel wool), chemical etching, and electrochemically etching, as are all known in the art. In a preferred embodiment, copper foil is chemically etched in a spray of 0.5 M aqueous solution of (NH4)2S2θ8 (ammonium persulfate) or Na2S2θ8 (sodium persulfate) at about 50°C for about 30 seconds to 1 minute.

After the intercalation anode has been prepared, it is assembled in a lithium intercalation cell. As is known in the art, the cell will include (1) a cell container, (2) an intercalation anode prepared as described above, (3) a cathode capable of reversibly taking up lithium on discharge and releasing lithium on charge, and (4) an electrolyte conductive to lithium ions. The cell should also include a separator between the anode and cathode.

No special modifications to conventional cell containers are required to fabricate cells from the anodes of this invention. Those of skill in the art will recognize the required properties of a cell container. It should be sized to compactly hold the various cell components and should be made of materials that are impervious to and chemically resistant to the other cell components at operating cell potentials.

The material used as the intercalation cell cathode should exhibit high capacity, good reversibility of lithium insertion, and a high average discharge voltage so as to achieve the largest possible energy of the cell. Such materials include, by way of example, lithium molybdenum sulfides, lithium molybdenum oxides, lithium vanadium oxides, lithium chromium oxides, lithiu titanium oxides, lithium tungsten oxides, lithium cobalt oxides, lithium nickel oxides, and lithium manganese oxides (e.g., LiMnθ2 and LiMn2θ4). In a particularly preferred embodiment of this invention, pure metal oxides (usually LiCoθ2, LiNiθ2, LiMnθ2 and/or LiMn2θ4) are combined with one another in certain ratios, combined with a conductive additive, a suspension thickener, and a solvent with a dissolved polymer, to produce a superior high voltage cathode with improved charge/discharge characteristics. Specifically, it is believed that such cathodes exhibit a discharge profile that is more sloped than their pure metal oxide counterparts.

An organic electrolyte for use in the cell may include any of various acceptable compounds and salts. Suitable organic electrolytes for use in intercalation cells include one or more of the following: propylene carbonate, ethylene carbonate, 1 ,2-dimethoxyethane, 1,2-diethoxyethane, γ- butyrolactone, tetrahydrofuran, 2-methyltetrahydrofuran, 1,3-dioxolan, 4-methyl 1,3-dioxolan, diethyl ether, sulfolane, acetonitrile, propionitrile, dimethyl carbonate, diethyl carbonate, anisole, and mixtures or combinations thereof. Suitable electrolyte salts include one or more of the following: lithium bis-trifluoromethane sulfonimide (Li(CF3SO2)2N or "HQ115" available from 3M Corp. of Minnesota), LiAsFό, LiPF6, LiBF4, LiB(C6H5)4, LiCl, LiBr, CH3SO3Li, and CF3SO3Li. In a preferred embodiment, the electrolyte includes a mixture of ethylene carbonate and dimethyl carbonate as the solvent together with HQ115, and LiAsF^ or LiPFό. In a particularly preferred embodiment, the electrolyte includes a mixture of ethylene carbonate and dimethyl carbonate and dissolved Li(CF3SO2)2N (about 0.5 to 1 M) and dissolved LiAsFό or LiPFό (either of which is present in a concentration of about 0.1 to 0.4 M). The total concentration of Li(CF3SO2)2N and LiAsFό or LiPFό should not exceed the solubility limit of lithium in the solvent. Thus, the total concentrations of these salts will generally be maintained below about 1M.

EXAMPLES

The following examples show that the composite electrodes of the present invention provide a unique combination of properties that are not found in electrodes made from either graphite alone or non-graphitic carbons alone. Most examples are designed to compare the capacity, voltage, and charge curve profile for three electrodes: (1) a graphite electrode, (2) a non- graphitic carbon electrode (made from a different carbon material in each example), and (3) a composite electrode made from a 50:50 mixture of graphite and the non-graphitic carbon. The graphs presented as Figs. 2-11 (each for a different non-graphitic carbon source) show electrode voltage versus fractional deintercalation for the three mentioned electrodes. In each graph, the data for the composite electrode was derived from an electrode made from a 50:50 mixture of graphite and the non-graphitic carbon used in the corresponding non-graphitic electrode.

The examples and corresponding data illustrate that generally the composite electrodes have voltage versus fractional deintercalation profiles that are intermediate in shape and magnitude between the profiles of the graphite electrode and the corresponding non-graphitic electrode. Thus, the composite electrodes had somewhat sloping charge profiles, suggesting that their performance at high discharge rates would be superior to that of graphite. Further, the composite electrodes generally had average voltages and capacities superior to those of the corresponding non-graphitic carbons. Most surprisingly, the capacities of the composite electrodes sometimes exceeded the capacities of both the graphite and non-graphitic electrodes.

The numerical results provided in the following examples (tables) are averages of two or more test runs. The corresponding graphical results in Figs. 2-11 depict only a single experimental run (giving a relatively higher capacity) chosen from among the set of runs used to provide the numerical results.

Preliminary Experiments 1 (preparing a binder solution)

While several combinations of polymer binder and solvent can be used to fabricate electrodes in this invention, all examples described below use a binder solution prepared generally as follows. The solution was produced by combining 1 gm of polyvinylidene difluoride (Kynar ® grade 721 obtained from Atochem North America, Inc., Philadelphia, PA) with 10 cc of dimethylformamide (DMF), and heating the solution to about 40° C while stirring until the polymer dissolved and the solution became clear. This resulting solution is referred to herein as 10% w/v PVDF. It should be noted that the amounts of binder and solvent may be varied, but their ratio should remain approximately constant.

Preliminary Experiments 2 (preparing a current collector having good adhesion)

All carbon anode slurries described in the following examples were spread onto copper foil (nominally 0.0005 inch thick). The copper foil served as a current collector in the completed carbon electrodes. The slurries were swept over the foils using a threaded metal rod, and resulted in slurry films of approximately 0.006 to 0.015 inches in thickness. Subsequent drying and compression reduced the film mickness to typically 0.002 to 0.005 inches.

It was observed that copper foils that were roughened before application of the slurry provided good adhesion (i.e., the carbon material was unlikely to peel away from the copper backing when dried and subjected to compression). Several methods were used to roughen the surface of the copper foil prior to applying the slurry film. These included mechanical roughenin (e.g., with steel wool), chemical etching and electrochemically etching as known in the art. All o these techniques resulted in good adhesion. The best results were obtained by etching the copper foil in a 0.5 M aqueous solution of (NH4)2S2θ8 or Na2S2θ8 at about 30°C for about 1 minute.

All examples described below employed films etched with ammonium persulfate under these conditions.

Fig. 1 illustrates a cell 10 employed in the experiments described below. The cell includes a test tube 14 which together with a screw-in top 18 serves as the cell container. Screw in top 18 also provides the necessary electrical connections for a carbon working electrode 20, a lithium counter electrode 22, and a lithium reference electrode 24. The working electrode assembly 20 includes a porous nylon separator placed around both a carbon intercalation electrode 26 and a piece of copper expanded metal 30 (of a size substantially larger than that of the electrode). The cell contains 50 cm^ of an electrolyte 34 containing 0.5 M Li(CF3SO2)2N and 0.1M LiAsFό in a 50:50% (by volume) solution of dimethylcarbonate ("DMC") and ethylene carbonate ("EC"). The tests were run at room temperature (about 18° C).

In each example, the working electrode was deintercalated of lithium at a rate of 50 mA/g carbon until the potential reached 5 mV (versus the lithium reference electrode), after which the potential was maintained at that value for 4 hours. After these intercalation steps (both of which are "discharge" steps with respect to the lithium metal electrode), the working electrode was deintercalated at a constant current of 50 mA/gm carbon until the potential reached 2.0 V (the "charge" step). The amount of charge transferred during the discharge step is greater than that which was removed during the charge step. This difference represents the irreversible capacity loss associated with formatting lithium intercalation electrodes. The voltage versus fractional deintercalation (a fractional deintercalation of 1.00 would be associated with fully deintercalating graphitic LiC6 — requiring 372 mA*hr/gm carbon) is shown in the curves of Figs. 2-11.

It should be noted that in the experiments described herein, the intercalation steps were thermodynamically favored (as the counter electrode was lithium metal). Therefore both the constant current (50 mA/gm carbon) and constant potential (5mV vs. Li) intercalation procedures were "discharge" steps. In lithium ion cells, where the counter electrode is a lithium metal oxide, the opposite is true: i.e., the carbon intercalation step is a cell "charge" step because intercalation of lithium into the carbon electrode is thermodynamically unfavored or not spontaneous.

Comparative Example 1 (preparing graphite electrodes and testing all carbon electrodes)

As a comparative example, electrodes having graphite as their only carbon source were prepared and tested. The graphite used in these electrodes was SFG 15 (trade name) as received from Lonza Inc. of Fairlawn, NJ. SFG 15 is a synthetic graphite, having a highly anisotropic morphology (much like natural graphite), good compressibility, and good electrical conductivity. Typical characteristics of the material are an Lς value of greater than about 120 A, a BET surface area of about 8.8 m^/g, and a size distribution in which 95% of the particles are less than about 16 μm.

In this and subsequent examples, various quantities of graphite, non-graphitic carbon, solvent, and binder solution are recited. The magnitudes of these quantities were chosen to illustrate the ratios employed. The actual quantities used in the examples were typically greater than the listed values, but ratios of all materials agreed with the ratios presented below.

To form the electrodes, one gram of SFG 15 graphite was combined with 1.5 cc of 10% w/v PVDF binder solution and an additional 1.6 cc of DMF (to obtain an acceptable slurry viscosity). The resulting mixture was stirred in a beaker to form a slurry, a portion of which was then applied to a 1.5 inch wide and 12 inch long copper foil strip (etched using a 0.5M (NH4)2S2θ8 solution as described above). Specifically, the slurry was applied to one side of the foil strip over about a 6 inch length. The resulting film was dried by first blowing hot air on the foil and then heating the foil on a hot plate at about 200° C until the solvent was fully evaporated. The foil/film was then passed through a pair of compression roller which applied a total force of approximately 3000 lbs over the 1.5 in width of foil strip. The compressed electrode strip was then reheated on the hot plate to melt the PVDF binder, and passed through the compression rollers a second time. After drying, the electrode strip was composed of 87% graphite and 13% binder. From this, two square pieces (both the same size, approximately 1.4 cm^) were cut: one containing the electrode active material and the other containing only the etched foil without the film. Both pieces were weighed on a microbalance, and the difference yielded the mass of the film.

The voltage versus deintercalation for the SFG 15 electrode is shown repeatedly as curve 3 in Figs. 2-10. As can be seen from this curve, the graphite electrode has a rather flat voltage vs. deintercalation state profile (varying by less than 200 mV over nearly the entire deintercalation process). Comparative Example 2 (preparing non-graphitic electrodes from pyrolyzed phenolic resin)

An electrode was constructed as in comparative Example 1 except the carbon used was derived from the pyrolysis of a phenolic resin (grade 29217) instead of SFG 15 graphite. The resin was pyrolyzed at 1050° C in a nitrogen atmosphere for 3 hours. The resulting non-graphitic carbon was first coarsely ground using a mortar in pestle (in air) and sieved to less than 208 μm, then placed in an attrition mill and ground under an argon atmosphere, and finally sieved into particles of between 20-38 μm. Other size fraction were also obtained and electrodes were prepared from these particles. It was found that electrodes made from particles of sizes less than about 60 μm performed similarly to electrodes prepared from the 20-38 μm particles (with respect to reversible and irreversible capacity). A slurry was prepared from the non-graphitic carbon according to the procedure described in comparative Example 1 , except that no additional DMF was required to reach an acceptable slurry viscosity. The electrode was dried and compressed as in comparative Example 1 to produce a pyrolyzed phenolic resin electrode. In addition, the electrode was weighed and tested as in comparative Example 1. The voltage versus fractional deintercalation (complete deintercalation of LiCό = 1.00) for this electrode is shown as curve 1 of

Fig. 2. As can be seen, the voltage profile is much more sloped than that of the graphite electrode. In addition, the capacity is substantially lower than that of the graphite electrode.

Example 3 (preparing composite electrodes from pyrolyzed phenolic resin)

An electrode was constructed as in comparative Example 2 except that the carbon used was a mixture of 0.5 g of the carbon described in comparative Example 2 and 0.5 g of the graphite described in comparative Example 1. Also additional DMF (0.8 cc) was added to reach an acceptable slurry viscosity. The electrode was dried and compressed as in comparative Example 1 to produce a composite pyrolyzed phenolic resin/graphite electrode. Thereafter, the composite electrode was weighed and tested as in Example 1. The voltage versus fractional deintercalation (complete deintercalation of LiCό = 1.00) for this electrode is shown as curve 2 of Fig. 2. As can be seen, the discharge profile is more sloped than that of the graphite electrode and the capacity and voltage are superior to those of the pyrolyzed phenolic resin electrode.

Comparative Example 4 (preparing non-graphitic electrodes from pyrolyzed polyfurfuryl alcohol resin) An electrode was constructed as in comparative Example 2 except that the carbon was derived from pyrolysis of polyfuifural alcohol resin instead of phenolic resin. Polyfurfural alcohol ("PFA" from Ucar, Inc. Lawrenceburg, TN) was obtained by mixing 200 cc of furfural alcohol with 8 cc of 75% H3PO4, and heating the mixture for about 2 hours at about 170° C while covered and stirring until the mixture polymerized and thickened. Note that the phosphoric acid acts as a dopant precursor to produce a phosphorous doped end product. This polymer mixture was pyrolyzed, ground, sieved, mixed, applied, dried, compressed, weighed, and tested in the same manner as in comparative Example 2. The voltage versus fractional deintercalation (complete deintercalation of LiCό = 1 00) for the resulting pyrolyzed PFA carbon electrode is shown as curve 1 of Fig. 3. As can be seen, this curve slopes significantly with state of deintercalation. Note also that this material has a characteristic double-hump shaped discharge curve associated with phosphorous-containing intercalation electrodes.

Example 5 (preparing composite electrodes from pyrolyzed polyfurfuryl alcohol resin)

An electrode was constructed as in Example 3 except that the carbon employed a mixture of

0.5 g of the carbon described in comparative Example 4 and 0.5 g of the graphite described in comparative Example 1. A slurry was produced and mixed, applied, dried, compressed, weighed, and tested as in Example 3. The voltage versus fractional deintercalation (complete deintercalation of LiCό = 1.00) for the resulting pyrolyzed PFA carbon/graphite composite electrode is shown as curve 2 of Fig. 3. Note that the curve not only has a strongly sloping deintercalation profile, but it also evidences a capacity that significantly exceeds that of both the pyrolyzed PFA electrode and the graphite electrode (as the vertical part of the deintercalation curve occurs at a fractional deintercalation lying to the right of the corresponding part of the curves for graphite and pyrolyzed PFA electrodes). In fact, the composite electrode's capacity exceeds graphite's theoretical maximum intercalation (LiC6), implying that on average lithium is intercalated to a level beyond

LiC6-

Comparative Example 6 (preparing non-graphitic electrodes from pyrolyzed resorcinol formaldehyde resin)

An electrode was constructed as in comparative Example 2 except that the non-graphitic carbon was derived from the pyrolysis of resorcinol-formaldehyde ("RF") resin instead of phenolic resin. RF resin was obtained by dissolving 200 gm of resorcinol (1, 3 dihydroxybenzene) from Inspec Corp. of Pittsburg, PA with 270 cc of 37% formaldehyde from Aldrich Corp. The mixture was heated to about 40° C and slowly stirred while slowly adding 10 cc of 10% nitric acid in water. (The polymerization reaction is highly exothermic and fast so care must be taken during this procedure). The resulting resin was pyrolyzed, ground, sieved, mixed, applied, dried, compressed, weighed, and tested as in comparative Example 2. The voltage versu fractional deintercalation (complete deintercalation of LiQj = 1.00) for the resulting pyrolyzed RF carbon electrode is shown as curve 1 of Fig. 4.

Example 7 (preparing composite electrodes from pyrolyzed resorcinol formaldehyde resin)

An electrode was constructed as in Example 3 except that the carbon used was a mixture o 0.5 g of the carbon described in comparative Example 6 and the 0.5 g of the graphite described in comparative Example 1. A slurry was produced which was mixed, applied, dried, compressed, weighed, and tested as in Example 3. The voltage versus deintercalation (complete deintercalation of LiC6 = 1.00) for the resulting pyrolyzed RF resin carbon/graphite composite electrode is show as curve 2 of Fig. 4. Note that this curve has a sloping profile and evidences a significantly highe capacity than the corresponding non-graphitic electrode.

Comparative Example 8 (preparing non-graphitic electrodes from pyrolyzed polyacrylonitrile fibers)

An electrode was constructed as in comparative Example 2 except that the carbon derived from the pyrolysis of polyacrylonitrile (PAN) fiber (obtained from Courtaulds Corp., grade SP/15) was used instead of phenolic resin. This particular grade of fiber contains about 95% PA and 5% methylacrylate as a comonomer. This material was pyrolyzed, ground, sieved, mixed, applied, dried, compressed, weighed, and tested as in comparative Example 2. The voltage versu fractional deintercalation (complete deintercalation of LiCό = 1 00) for the resulting pyrolyzed

PAN-fiber carbon electrode is shown as curve 1 of Fig. 5.

Example 9 (preparing composite electrodes from pyrolyzed polyacrylonitrile fibers)

An electrode was constructed as in Example 3 except that the carbon used was a mixture o

0.5 g of the carbon described in comparative Example 8 and 0.5 g of the graphite described in comparative Example 1. A slurry was produced that was mixed, applied, dried, compressed, weighted, and tested as in Example 3. The voltage versus fractional deintercalation (complete deintercalation of LiCό = 1.00) for the resulting pyrolyzed ground PAN-fiber carbon/graphite composite electrode is shown as curve 2 of Fig. 5. Comparative Example 10 (preparing non-graphitic electrodes from pyrolyzed CIPS PAN)

An electrode was constructed as in comparative example 2 except that the non-graphitic carbon was derived from the chemically induce phase separation ("CIPS") of PAN powder (Polysciences Corp., Warrington, PA) instead of phenolic resin. The CIPS process involved the following steps: 1 ) the combining 80 gm of PAN powder with 500 cc of dimethylsulfoxide

("DMSO") solvent, 2) rapid stirring of the mixture as it is heated to about 130°C until all the PAN is dissolved, 3) adding the DMSO/PAN solution of 70% DMSO in water while the solution is being aggressively stirred in a blender causing the PAN to chemically precipitate from the solution, 4) filtering and centrifuging the CIPS PAN, 5) extracting the entrained solvents with water, and 6) drying the resulting product. It has been found that the CIPS PAN is generally porous (and fractious), and that carbon produced from pyrolyzed CIPS PAN generally provides better electrodes than carbon derived from PAN fibers, both when the PAN is doped with phosphorous and when it is not doped with phosphorous. The CIPS PAN of this particular example was placed in a solution of acetone containing 2% by weight of 75% phosphoric acid for 2 hours, and then was filtered and centrifuged to remove excess entrained solution prior to evaporating the acetone. This procedure allowed for the uniform application of phosphoric acid to the CIPS PAN. The dried material was pyrolyzed, ground, sieved, mixed, applied, dried, compressed, weighed, and tested as in comparative Example 2. The voltage versus fractional deintercalation (complete deintercalation of LiC6 = 1 00) for the resulting pyrolyzed, phosphorous doped CIPS-PAN carbon electrode is shown as curve 1 of Fig. 6. Note the high capacity, sloping voltage profile, and double hump characteristic of phosphorous doping.

Example 11 (preparing composite electrodes from pyrolyzed CIPS PAN)

An electrode was constructed as in Example 3 except that the carbon used was a mixture of 0.5 g of the CIPS-PAN carbon described in comparative Example 10 and 0.5 g of the graphite described in comparative Example 1. A slurry was produced that was mixed, applied, dried, compressed, weighed, and tested as in Example 3. The voltage versus fractional deintercalation (complete deintercalation of LiC6 = 1 00) for the resulting pyrolyzed, phosphorous doped CIPS- PAN carbon/graphite composite electrode is shown as curve 2 of Fig. 6. As can be seen, this curve has a sloping deintercalation profile and extremely high capacity (higher than electrodes fabricated from either graphite or the CIPs PAN).

A comparison of performance parameters for the electrodes of Examples 1-11 is given in table 1. Note that the average voltages were taken over the complete charge from 5mV to 2V and were corrected for ohmic losses. Table 1

Several different methods of grinding and firing CIPS PAN were investigated to determine the optimal conditions for high reversible capacity, small irreversible capacity loss, and low average voltage. Several examples and their test results illustrate the effect of treatment on performance.

Example 12 (preparing non-graphitic electrodes from pyrolyzed CIPS PAN)

CIPS PAN was produced as described in comparative Example 10, except that the PAN was not doped with phosphorous. A slurry was produced that was mixed, applied, dried, compressed, weighed, and tested as in Example 3. In this case, the resulting pyrolyzed, CIPS- PAN carbon electrode was placed into a test tube containing 50 cc of an electrolyte containing 0.5 M Li(CF3SO2)2 and 0.1M LiAsFδ in propylene carbonate (no EC or DMC was used). A lithium metal reference and counter electrode were also inserted into the solution as shown in Fig. 1. The test tube was closed via the screw in top discussed above with reference to Fig. 1. The test was conducted at room temperature (about 18°C). Because the conductivity of the propylene carbonate electrolyte is lower than that of the EC/DMC mixture used in Examples 1-11, the average discharge voltages tended to be higher.

Example 13 (preparing non-graphitic electrodes from pyrolyzed CIPS PAN)

CIPS PAN was produced as described in Example 12, except the initial grinding was done not by hand with a mortar and pestle, but rather with a hammer mill (Weber Bros, and White Metal Works, Hamilton, Mich.) operating in the air prior to grinding with the attrition mill under argon. A slurry was produced that was mixed, applied, dried, compressed, weighed, and tested as in Example 3 to produce a pyrolyzed, CIPS-PAN carbon electrode.

Example 14 (preparing non-graphitic electrodes from pyrolyzed CIPS PAN)

CIPS PAN was produced as described in Example 13, except that the final grinding with the attrition mill was done in air. A slurry was produced and then mixed, applied, dried, and compressed as in Example 3 to produce a pyrolyzed, CIPS-PAN carbon electrode. The electrode was then weighed and tested as in Example 3.

Examples 15-17 (preparing non-graphitic electrodes from twice-pyrolyzed CIPS PAN)

Carbons powders produced in Examples 12-14 were pyrolyzed under nitrogen for 3 hours at 1050 °C to produce the electrodes used as Examples 15, 16, and 17 respectively. Slurries were produced and then mixed, applied, dried, and compressed as in Example 3 to produce pyrolyzed, CIPS-PAN carbon electrodes. These were then weighed and tested as in Example 3.

Table 2 gives the results of electrochemical tests (performed as described in comparative example 1) performed on CIPS-PAN electrodes prepared as described in Examples 12-17. All tests were performed in a propylene carbonate electrolyte, and the average discharge voltages were corrected for ohmic losses. Table 2

The results given in Table 2 show that by refiring a sample after grinding, significant reductions in irreversible capacity loss can be obtained, while lowering the average voltage, and maintaining the reversible capacity. Because of this promising result, electrodes were fabricated from refired carbon powder that was initially prepared as described in Examples 2, 4, 6, 8, and 10. Also, non-graphitic/graphite composite electrodes were fabricated from these refired carbons.

Comparative Example 18 (preparing non-graphitic electrodes from twice-pyrolyzed phenolic resin)

Carbon powder, derived from phenolic resin as produced and described in comparative Example 2, was refired under nitrogen at 1050°C for 3 hours. The resulting powder was then used to make a slurry which was mixed, applied, dried, compressed, weighed, and tested as in comparative Example 2. The voltage versus fractional deintercalation (complete deintercalation of LiC6 = 1 00) for the resulting repyrolyzed, phenolic resin electrode is shown as curve 1 of Fig. 7

Example 19 (preparing composite electrodes from twice-pyrolyzed phenolic resin) An electrode was constructed as in Example 3 except that the non-graphitic carbon was refired phenolic resin carbon prepared as discussed in comparative Example 18. Specifically, a slurry was produced which was applied, dried, compressed, weighed, and tested as in Example 3. The voltage versus fractional deintercalation (complete deintercalation of LiCό = 1.00) for the resulting repyrolyzed phenolic resin/graphite composite electrode is shown as curve 2 of Fig. 7.

Comparative Example 20 (preparing non-graphitic electrodes from twice-pyrolyzed polyfurfuryl alcohol)

Carbon powder, derived from polyfurfural alcohol (PFA) produced as described in Example 4, was refired under nitrogen at 1050° C for 3 hours. The resulting powder was then used to make a slurry which was mixed, applied, dried, compressed, weighed, and tested as in comparative Example 4. The voltage versus fractional deintercalation (complete deintercalation of LiCβ = 1.00) for the resulting repyrolyzed PFA resin electrode is shown as curve 1 of Fig. 8

(again note the double hump characteristic of phosphorous doping).

Example 21 (preparing composite electrodes from twice-pyrolyzed polyfurfuryl alcohol)

A composite electrode was constructed as in Example 3 except that refired polyfurfural alcohol resin (PFA) carbon powder described in Example 20 was used as the non-graphitic carbon. Specifically, a slurry was produced which was applied, dried, compressed, weighed, and tested as in Example 3. The voltage versus fractional deintercalation (complete deintercalation of LiCβ = 1.00) for the resulting repyrolyzed PFA resin/graphite composite electrode is shown as curve 2 of Fig. 8.

Comparative Example 22 (preparing non-graphitic electrodes from twice-pyrolyzed resorcinol- formaldehyde resin)

Carbon powder, derived from Resorcinol-Formaldehyde resin produced as described in Example 6, was refired under nitrogen at 1050° C for 3 hours. The resulting powder was then used to make a slurry which was mixed, applied, dried, compressed, weighed, and tested as in Example 6. The voltage versus fractional deintercalation (complete deintercalation of LiCό = 1.00) for the resulting repyrolyzed resorcinol-formaldehyde resin electrode is shown as curve 1 of Fig. 9. Exa ple 23 (preparing composite electrodes from twice-pyrolyzed resorcinol-formaldehyde resin)

An electrode was constructed as in Example 3 except that refired Resorcinol-Formaldehyde resin carbon powder as described in Example 22 was used as the non-graphitic carbon. A slurry was produced which was applied, dried, compressed, weighed, and tested as in comparative Example 7. The voltage versus fractional deintercalation (complete deintercalation of LiC6 = 1.00) for the resulting repyrolyzed Resorcinol-Formaldehyde resin/graphite composite electrode is shown as curve 2 of Fig. 9.

Example 24 (preparing non-graphitic electrodes from twice-pyrolyzed PAN fibers)

Carbon powder derived from PAN carbon fibers produced as described in Example 8 was refired under nitrogen at 1050°C for 3 hours. The resulting powder was then used to make a slurry which was mixed, applied, dried, compressed, weighed, and tested as in Example 8. The voltage versus fractional deintercalation (complete deintercalation of LiC6 = 1.00) for the resulting repyrolyzed, ground, PAN-carbon fiber electrode is shown as curve 1 of Fig. 10.

Example 25 (preparing composite electrodes from twice-pyrolyzed PAN fibers)

An electrode was constructed as in Example 3 except that ground and refired PAN carbon fiber powder as described in Example 24 was used as the non-graphitic carbon. A slurry was produced which was applied, dried, compressed, weighed, and tested as in comparative Example 9. The voltage versus fractional deintercalation (complete deintercalation of LiCβ = 1.00) for the resulting repyrolyzed ground, PAN-carbon/graphite composite electrode is shown as curve 2 of Fig. 10.

Comparative Example 26 (preparing non-graphitic electrodes from twice-pyrolyzed CIPS PAN)

Carbon powder, derived from the CIPS PAN process as described in Example 10 was refired under nitrogen at 1050°C for 3 hours. The resulting powder was then used to make a slurry which was mixed, applied, dried, compressed, weighed, and tested as in Example 10. The voltage fractional deintercalation (complete deintercalation of LiCό = 1.00) for the resulting repyrolyzed, phosphorous doped CIPS-PAN electrode is shown as curve 1 of Fig. 11.

Example 27 (preparing composite electrodes from twice-pyrolyzed CIPS PAN) An electrode was constructed as in Example 3 except that ground and refired CIPS PAN carbon powder as described in Example 26 was used as the non-graphitic carbon. A slurry was produced which was applied, dried, compressed, weighed, and tested as in comparative Example 11. The voltage versus fractional deintercalation (complete deintercalation of LiCδ = 1.00) for the resulting repyrolyzed, phosphorous doped CIPS-PAN/graphite composite electrode is shown as curve 4 of Fig. 11.

Comparisons of single and refired non-graphitic electrodes' performances are shown in Table 3. Some of the data in this table is repeated here from Table 1 for ease of comparison. The deintercalation curve shapes for the electrodes of Examples 18 to 27 are generally similar to the corresponding curves of Examples 2-11. However, the twice fired electrodes tended to have slightly lower reversible capacities, higher irreversible capacities, and average voltages.

Table 3

Examples 28 to 31

Electrodes were constructed as in Example 3 except that (1) ground and refired 2% phosphorous doped CIPS PAN carbon powder as described in Example 26 was used, and (2) a range of differing concentrations of graphite were used to produce a series of CIPS PAN/graphite electrodes. Specifically, a series of slurries were produced which were applied, dried, compressed, weighed, and tested as in Example 3. Table 4 gives composition and the film density, reversible capacity, irreversible capacity, and average discharge voltage for this series of repyrolyzed phosphorous doped CIPS-PAN/graphite composite electrodes (note that the 0, 50, and 100% composites are repeats of Example 26, 27, and 1 respectively). Figure 11 shows a family of curves of voltage versus fractional deintercalation (LiCό = 1 00) for this series of electrodes. As above, the values in the table are numerical averages of multiple runs and the curves in the graph are chosen based on high capacity.

Table 4

A comparative example of the cycleability of single phase versus composite electrodes is shown in Fig. 12. Curve 1 of that figure corresponds to an electrode prepared in the same manner as Example 20 (single phase, PFA only), while curve 2 corresponds to a composite electrode prepared in the same manner as Example 21 (50% PFA, 50% graphite). This result demonstrates the composite electrode retains its capacity with cycling, but that the single phase electrode looses its capacity rapidly. Other results indicate that other composite materials cycled much better than the pure phase materials, retaining their capacity longer.

Although the foregoing invention has been described in some detail for purposes of clarity of understanding, it will be apparent that certain changes and modifications may be practiced within the scope of the appended claims. For instance, although the specification has primarily described a process for preparing electrodes for use in lithium ion cells, the carbon material disclosed herein may have other applications as well. For example, double layer capacitors and fuel cells may also employ the electrodes of this invention.

Claims

WHAT IS CLAIMED IS:

1 . A carbon based electrode capable of intercalating lithium, said electrode comprising a mixture of homogeneous graphitic carbon particles and homogeneous non-graphitic carbon particles.

2. The carbon based electrode of claim 1 wherein the open circuit potential of the electrode varies by at least about 0.5 volts from a fully charged state in which the electrode is fiilly intercalated to a state of charge at about 90% of deintercalation.

3. The carbon based electrode of claim 1 wherein the homogeneous graphitic carbon particles have an average particle size of between about 0.5 to 50 μm.

4. The carbon based electrode of claim 1 wherein the homogeneous graphitic carbon particles have an average particle size of between about 5 to 20 μm.

5. The carbon based electrode of claim 1 wherein the homogeneous non-graphitic carbon particles have an average particle size of between about 0.1 to 100 μm.

6. The carbon based electrode of claim 1 wherein the homogeneous non-graphitic carbon particles have an average particle size of between about 5 to 20 μm.

7. The carbon based electrode of claim 1 wherein the mixture includes at least about 25 weight percent homogeneous graphitic carbon particles.

8. The carbon based electrode of claim 1 wherein the mixture includes at least about 50 weight percent homogeneous graphitic carbon particles.

9. The carbon based electrode of claim 1 wherein the mixture includes about 75 weight percent homogeneous graphitic carbon particles.

10. The carbon based electrode of claim 1 further comprising a binder.

1 1. The carbon based electrode of claim 10 wherein the binder is selected from the group consisting of polyvinylidene difluoride, polytetrafluoroethylene, and ethylene propylene diene monomer.

12. The carbon based electrode of claim 1 wherein the homogeneous non-graphitic carbon particles are obtained by pyrolysis of a material selected from the group consisting of coke, petroleum residues, acrylic resins, polydivinyl benzene, polyvinyl chloride, furfuran resins, phenolic resins, resorcinol-formaldehyde resins, polyimide resins, cyclic hydrocarbons containing at least two ring structures, and benzene derivatives.

13. The carbon based electrode of claim 12 wherein the homogeneous non-graphitic carbon particles are obtained by pyrolysis of a material selected from the group consisting of polyacrylonitrile and polyfurfuryl alcohol.

14. The carbon based electrode of claim 12 wherein the homogeneous non-graphitic carbon particles are obtained by first pyrolyzing the material, then grinding, and then repyrolyzing.

15. The carbon based electrode of claim 1 wherein the homogeneous non-graphitic carbon particles are doped with a dopant element selected from the elements in groups IQ and IV of the periodic table.

16. The carbon based electrode of claim 15 wherein the dopant element is phosphorous.

17. The carbon based electrode of claim 1 wherein the homogeneous graphitic carbon particles have a morphology having a high degree of anisotropy.

18. The carbon based electrode of claim 1 wherein the mixture of homogeneous graphitic carbon particles and homogeneous non-graphitic carbon particles has a density of at least about 1.2 g/cc.

19. The carbon based electrode of claim 1 wherein the open circuit potential of the electrode varies by at least about 0.25 volts from a fully charged state in which the electrode is fully intercalated to a state of charge at about 90% deintercalation.

20. A lithium ion cell comprising:

(1) a cell container;

(2) an anode including a mixture of homogeneous graphitic carbon particles and homogeneous non-graphitic carbon particles, the anode being capable of intercalating lithium during charge and deintercalating lithium during discharge; (3) a cathode capable of reversibly taking up lithium on discharge and releasing lithium on charge; and

(4) an electrolyte conductive to lithium ions, wherein said anode, cathode, and electrolyte are provided within said cell container.

21. The lithium ion cell of claim 20 wherein the anode is initially provided in a deintercalated state.

22. The lithium ion cell of claim 20 wherein the mixture of homogeneous graphitic carbon particles and homogeneous non-graphitic carbon particles has a density of at least about 1.2 g/cc.

23. The lithium ion cell of claim 20 wherein the cathode includes LiNiθ2, LiCoθ2, LiMnθ2, LiMn2O4, or a mixture thereof.

24. The lithium ion cell of claim 20 wherein the electrolyte includes one or more of the following: propylene carbonate, ethylene carbonate, 1,2-dimethoxyethane, 1 ,2-diethoxyethane, γ- butyrolactone, tetrahydrofuran, 2-methyltetrahydrofuran, 1,3-dioxolan, 4-methyl 1,3-dioxolan, diethyl ether, sulfolane, acetonitrile, propionitrile, glutaronitrile, dimethyl carbonate, diethyl carbonate, anisole, and mixtures thereof.

25. The lithium ion cell of claim 24 wherein the electrolyte further includes one or more of the following salts: (LiCF35θ2)2N, LiAsF6, LiPF6, L-BF4, LiB(C6Hs)4, LiCl, LiBr, CH3SO3Li, and CF3SO3LL

26. The lithium ion cell of claim 25 wherein the electrolyte includes a mixture of ethylene carbonate and dimethyl carbonate and dissolved Li(CF3SO2)2 at about 0.5 to 1 M and dissolved LiAsFό or LiPFό at about 0.1 to 0.4 M, wherein the total concentration of Li(CF3SO2)2N and LiAsFό or LiPFό does not exceed the solubility limit of lithium in the solvent.

27. The lithium ion cell of claim 20 wherein the open circuit potential of the anode varies by at least about 0.25 volts from a fully charged state in which the electrode is fully intercalated to a state of charge at about 90% deintercalation.

28. A method of preparing a carbon based electrode capable of intercalating lithium, said method comprising the following steps:

(1) applying a slurry film to a current collector, the slurry containing a binder, a solvent for the binder, and a mixture of homogeneous graphitic carbon particles and homogeneous non-graphitic carbon particles; (2) evaporating the solvent from the slurry to leave a composite on the current collector; and

(3) compressing the composite on the current collector to form said carbon based electrode.

29. The method of claim 28 further comprising a step of heating the composite to a temperature sufficient to cause said binder to flow after the step of compressing.

30. The method of claim 29 further comprising a step of recompressing the composite after the step of heating the composite.

31. The method of claim 28 wherein the binder is selected from the group consisting of polytetrafluoroethylene, polyvinylidene difluoride, and ethylene propylene diene monomer.

32. The method of claim 31 wherein the binder is polyvinylidene difluoride and the solvent for the binder is dimethylformamide.

33. A method of preparing a carbon based electrode capable of intercalating lithium, the method comprising the following steps: mixing homogeneous particles of graphite with homogeneous particles of a non- graphitic carbon to form a composite carbon mixture; forming said carbon based electrode by affixing said composite carbon mixture to a current collector.

34. The method of claim 33 wherein the step of forming said carbon based electrode includes a step of applying a slurry film to the current collector, the slurry containing a binder, a solvent for the binder, and said composite carbon mixture.

35. The method of claim 33 further comprising a step of preparing the homogeneous particles of a non-graphitic carbon by pyrolyzing an organic material at a temperature of between about 600 and 2000° C to produce a non-graphitic carbon material.

36. The method of claim 35 wherein said step of preparing the homogeneous particles of a non-graphitic carbon includes the following steps: conducting a first grinding step on the non-graphitic carbon material, followed by a first sieving step; and conducting a second grinding step on the non-graphitic carbon material, followed by a second sieving step.

37. The method of claim 36 wherein said first grinding step is conducted with a hammer mill.

38. The method of claim 36 wherein said second grinding step is conducted with an attrition mill.

39. The method of claim 35 wherein further comprising the following steps: grinding the non-graphitic carbon material; and refiring said non-graphitic carbon material.