CA2149853A1 - Pre-graphitic carbonaceous insertion compounds and use as anodes in rechargeable batteries - Google Patents

Abstract

Carbonaceous insertion compounds and methods for preparation are described wherein the compounds comprise a highly disordered, impurity free, hard pre-graphitic carbonaceous host. Carbonaceous insertion compounds can be prepared which have large reversible capacity for lithium yet low irreversible capacity and voltage hysteresis. Such insertion compounds can be prepared by simple pyrolysis of suitable carbohydrate precursors at an appropriate tempera-ture. These insertion compounds may be suitable for use as high capacity anodes in lithium ion batteries.

Description

-PRE-GRAPHITIC rA~RONACROUS INSERTION COMPOUNDS AND USE AS
ANODES IN ~r~A~r.~ART.~ BATTERIES

FIELD OF THE lNV~. ~ lON

The invention pertains to the field of carbonaceous materials and, in particular, to pre-graphitic carbonaceous insertion materials. Additionally, the invention pertains to the field of rechargeable batteries and, in particular, to rechargeable batteries comprising carbonaceous anode materials.

RAC~G~OUND OF THE lN V~. ~ lON

The group of pre-graphitic compounds includes carbon-aceous materials that are generally prepared at low tem-peratures (eg: less than about 2000C) from various organic sources and that tend to graphitize when annealed at higher temperatures. There are however both hard and soft pre-graphitic carbon compounds, the former being difficult to graphitize substantially even at temperatures of order of 3000C, and the latter, on the other hand, being virtually completely graphitized around 3000C.
The aforementioned set of compounds has been of great interest for use as anode materials in lithium-ion or rocking chair type batteries. These batteries represent the state of the art in small rechargeable power sources for consumer electronics applications. These batteries have the greatest energy density (Wh/L) of conventional rechargeable systems (ie. NiCd, NiMH, or lead acid bat-teries). Additionally, lithium ion batteries operate around 3~ volts which is often sufficiently high such that a single cell can suffice for many electronics applica-tions.
Lithium ion batteries use two different insertion compounds for the active cathode and anode materials.
Insertion compounds are those that act as a host solid for the reversible insertion of guest atoms (in this case, lithium atoms). The structure of the insertion compound 21~8S3 host is not significantly altered by the insertion. In a lithium ion battery, lithium is extracted from the anode material while lithium is concurrently inserted into the cathode on discharge of the battery. The reverse processes occur on recharge of the battery. Lithium atoms travel or ~rock" from one electrode to the other as ions dissolved in a non-aqueous electrolyte with the associated electrons travelling in the circuit external to the battery.
The two electrode materials for lithium ion batteries are chosen such that the chemical potential of the inserted lithium within each material differs by about 3 to 4 electron volts thus leading to a 3 to 4 volt battery. It is also important to select insertion compounds that reversibly insert lithium over a wide stoichiometry range thus leading to a high capacity battery.
A 3.6 V lithium ion battery based on a LiCoO2/pre-graphitic carbon electrochemistry is commercially available (produced by Sony Energy Tec.) wherein the carbonaceous anode can reversibly insert about 0.65 Li per six carbon atoms. (The pre-graphitic carbon employed is a disordered form of carbon which appears to be similar to coke.) However, the reversible capacity of lithium ion battery anodes can be increased by using a variety of alternatives mentioned in the literature. For example, the crystal structure of the carbonaceous material affects its ability to reversibly insert lithium (as described in J.R. Dahn et al., "Lithium Batteries, New Materials and New Perspec-tives", edited by G. Pistoia, Elsevier North-Holland, pl-47, (1993)). Graphite for instance can reversibly incor-porate one lithium per six carbon atoms which correspondselectrochemically to 372 mAh/g. This electrochemical capacity per unit weight of material is denoted as the specific capacity for that material. Graphitized carbons and/or graphite itself can be employed under certain conditions (as for example in the presentation by Matsushita, 6th International Lithium Battery Conference, 2149~3 Muenster, Germany, May 13, 1992, or in U.S. Patent No.
5,130,211).
Other alternatives for increasing the specific capac-ity of carbonaceous anode materials have included the addition of other elements to the carbonaceous compound.
For example, Canadian Patent Application Serial No.

2,098,248, Jeffrey R. Dahn et al., 'Electron Acceptor Substituted Carbons for Use as Anodes in Rechargeable Lithium Batteries', filed June 11, 1993, discloses a means for enhancing anode capacity by substituting electron acceptors (such as boron, alnm;nllm, and the like) for carbon atoms in the structure of the carbonaceous compound.
Therein, reversible specific capacities as high as 440 mAh/g were obtained with boron substituted carbons.
Canadian Patent Application Serial No. 2,122,770, Alfred M.
Wilson et al., 'Carbonaceous Compounds and Use as Anodes in Rechargeable Batteries', filed May 3, 1994, discloses pre-graphitic carbonaceous insertion compounds comprising nanodispersed silicon atoms wherein specific capacities of 550 mAh/g were obtained. Similarly, specific capacities of about 600 mAh/g could be obtained by pyrolyzing siloxane precursors to make pre-graphitic carbonaceous compounds containing silicon as disclosed in Canadian Patent Applica-tion Serial No. 2,127,621, Alfred M. Wilson et al., 'Carbonaceous Insertion Compounds and Use as Anodes in Rechargeable Batteries', filed July 8, 1994.
Recently, practitioners in the art have prepared carbonaceous materials with very high reversible capacity by pyrolysis of suitable starting materials. At the Seventh International Meeting on Lithium Batteries, Extended Abstracts Page 212, Boston, Mass. (1994), A.
Mabuchi et al. have demonstrated that pyrolyzed coal tar pitch can have reversible specific capacities as high as 750 mAh/g at pyrolysis temperatures about 700C. K. Sato et al. in Science 264, 556, (1994) disclosed a similar carbonaceous material prepared by heating polyparaphenylene at 700C which has a reversible capacity of 680 mAh/g. S.

2149~5~

Yata et al, Synthetic Metals 62, 153 (1994) also disclose a similar material made in a similar way. These values are much greater than that of pure graphite. The aforementioned materials can have a very large irreversible capacity. Additionally, the voltage versus lithium of all the aforementioned materials has a significant hysteresis (ie. about 1 volt) between discharge and charge (or between insertion and extraction of lithium). In a lithium ion battery using such a material as an anode, this would result in a similar significant hysteresis in battery voltage between discharge and charge with a resulting undesirable energy inefficiency.
It is not understood why the aforementioned carbon-aceous materials have very high specific capacity. All were prepared at temperatures of about 700C and are crystalline enough to exhibit x-ray patterns from which the parameters doo2, Lc, a, and La can be determined. (The definition and determination of these parameters can be found in K. Kinoshita, "Carbon - Electrochemical and Physicochemical Properties", John Wiley & Sons 1988.) Also, all have substantial amounts of incorporated hydrogen as evidenced by H/C atomic ratios that are greater than 0.1, and often near 0.2. Finally, it appears that pyrolyz-ing at higher temperature degrades the specific capacity substantially with a concurrent reduction in the hydrogen content. (In the aforementioned reference by Mabuchi et al, pyrolyzing the pitch above about 800C results in a specific capacity decrease to under 450 mAh/g with a large reduction in H/C. Similar results were found in the aforementioned reference by Yata et al.) The prior art also discloses carbonaceous compounds with specific capacities higher than that of pure graphite made from precursors that form hard carbons on pyrolysis.
However, the very high specific capacities of the aforementioned materials pyrolyzed at about 700C were apparently not attained. A. Omaru et al, Paper #25, Extended Abstracts of Battery Division, p34, Meeting of the -Electrochemical Society, Toronto, Canada (1992), disclose the preparation of a hard carbonaceous compound containing phosphorus with a specific capacity of about 450 mAh/g by pyrolyzing polyfurfuryl alcohol. The polyfurfuryl alcohol in turn had been prepared from the monomer polymerized in the presence of phosphoric acid. In Japanese Patent Appli-cation Laid Open number 06-132031, Mitsubishi Gas Chemical disclose a hard carbonaceous compound comprising 2.4~
sulfur with a specific capacity of about 500 mAh/g. These hard carbonaceous compounds have additional elements incorporated therein and have all been pyrolyzed at suffi-cient temperature such that they contain little hydrogen (ie. the H/C atomic ratio is substantially less than 0.1).
These hard carbonaceous compounds neither exhibited the very high specific capacities nor the same serious hysteresis in voltage of the aforementioned materials pyrolyzed at about 700C.
Additionally, other high capacity carbonaceous materials have recently been prepared which show high capacity for lithium and little or no voltage hysteresis.
In Paper 2B05 at the 35th Battery Symposium in Nagoya, Japan, Nov. 14-16, 1994, Y. Takahashi et al. describe materials with reversible capacities of about 480 mAh/g, but do not give the details of their preparation. In paper 2B09 at the same Symposium, N. Sonobe et al. describe hard carbon materials made from petroleum pitch with reversible capacities near 500 mAh/g. The synthesis procedure therein was not given.
In Canadian Patent Application Serial No. 2,138,360, Y. Liu and J. Dahn, titled 'Pre-Graphitic Carbonaceous Insertion Compounds and Use as Anodes in Rechargeable Batteries', filed Dec. 16, 1994, carbonaceous insertion compounds also having high capacity for lithium and little voltage hysteresis were disclosed. Therein, the carbon-aceous insertion compounds comprised a pre-graphitic carbonaceous host wherein i) the empirical parameter R, determined from an x-ray diffraction pattern and defined as 21~98S3 the {002} peak height divided by the background level, is less than about 2.2; ii) the H/C atomic ratio is less than about 0.1; and iii) the methylene blue absorption capacity of the pre-graphitic carbonaceous host is less than about 4 micromoles per gram of host. These carbonaceous inser-tion compounds were prepared by pyrolyzing suitable organic precursors. Specifically shown in the Examples were insertion compounds prepared from different epoxy precur-sors.
Additionally, Canadian Patent Application Serial No.
(unassigned), by U. von Sacken, Q. Zhong, T. Zheng and J.
Dahn, titled 'Phenolic Resin Precursor Pre-Graphitic Carbonaceous Insertion Compounds and Use as Anodes in Rechargeable Batteries', filed April 5, 1995, discloses carbonaceous insertion compounds having R less than 2.2, H/C atomic ratio less than 0.1, and methylene blue absorp-tion capacity less than 4 micromoles per gram which were prepared from phenolic resin precursors.
Japanese patent application laid open number 06-089721 discusses the high capacity advantages of hard disordered carbons in terms of the parameters PO (the fraction of stacked carbon), nave (the number of graphene sheets per stack), and SI (the stacking index). Therein, SI is defined by the height of the {002} peak relative to the local background and is related to the aforementioned parameter R. When the local background is relatively flat and/or if the {002} peak is relatively large compared to the background, SI is approximately equal to 1-(1/R).
Therein, carbonaceous compounds having values of SI below 0.76 were claimed and the examples provided had a minimum SI of 0.67. Using the approximate conversion formula, these values correspond to R of 4.2 and 3.0 respectively.
Reversible capacities for lithium up to 460 mAh/g were obtained. However, voltage curves (and hence hysteresis characteristics) and irreversible capacities were not reported. Additionally, discussion and data regarding hydrogen contents after pyrolysis and surface area access-ible to non-aqueous electrolyte were not provided.

SUMMARY OF THE lN VL llON

This invention comprises novel carbonaceous insertion compounds with a high reversible capacity for alkali metal insertion, methods of preparing said insertion compounds, and the use of said novel insertion compounds as electrode materials in electrochemical devices in general. The alkali metal can be lithium and, in such a case, the insertion compound can have a low irreversible capacity and a small voltage hysteresis between insertion and extrac-tion.
Lithium-carbonaceous insertion compounds of the invention comprise a pre-graphitic carbonaceous host and lithium atoms inserted therein. The empirical parameter R, as determined from an x-ray diffraction pattern of the host and defined as the {002} peak height divided by the back-ground level, is less than about 2.2. The H/C atomic ratio of the host is less than about 0.1. The pre-graphitic host has a surface area accessible to non-aqueous electrolyte that is sufficiently small such that the irreversible capacity is less than about a half that of the reversible capacity, and preferably less than about a third that of the reversible capacity. The non-aqueous electrolyte can be a solution comprising ethylene carbonate and diethyl carbonate.
Electrochemical methods are preferably used to deter-mine reversible and irreversible capacities after which an accessible surface area can be deduced. However, other physical characteristics can be used to estimate the accessible surface area. For example, methylene blue absorption capacity and BET (a standard nitrogen adsorption test) surface area provide such estimates. When the methylene blue absorption capacity of the carbonaceous host is less than about 4 micromoles per gram of host or when 214985~

the surface area of the carbonaceous host as determined by BET is less than about 300 m2/gram, the accessible surface area can be sufficently small to meet the capacity require-ments.
Suitable carbonaceous hosts can be rendered unsuitable by relatively mild oxidation without overly dramatic effects on methylene blue absorption. The BET surface area may increase substantially but still be in a range con-sidered acceptable in principle. However, a mildly oxi-dized carbonaceous host can comprise enough surface oxygen such that more than 5~ by weight is lost after pyrolyzing at about 1000C under inert gas. Thus, suitable carbon-aceous hosts preferably have not been oxidized after preparation. Suitable carbonaceous hosts typically lose less than about 5~ by weight under such inert pyrolysis conditions.
The pre-graphitic carbonaceous host can be prepared by pyrolyzing a carbohydrate precursor at a temperature above about 800C, thereby predominantly removing hydrogen from the precursor. However, the pyrolysis temperature cannot be too high in order that the empirical parameter R, determined from an x-ray diffraction pattern of the host and defined as the {002} peak height divided by the back-ground level, is less than about 2.2, and preferably less than about 2. Along with other previously mentioned advantages, such hosts can have relatively large tap density, often exceeding 0.7 g/ml. Alkali metal atoms can be inserted into the host thereafter by conventional chemical or electrochemical means to make insertion com-pounds of the invention.
Suitable pre-graphitic hosts can be prepared by pyrolyzing a carbohydrate precursor or a carbohydrate containing precursor. Such a carbohydrate precursor can be selected from the group consisting of sugar, starch, and cellulose or substances containing these materials.
Specifically, the carbohydrate precursor can be sucrose, 214g853 -starch, or the cellulose in red oak, maple, walnut shell, filbert shell, almond shell, cotton or straw.
The pyrolysis can be performed at a temperature in the range from about 900C to about 1100C for about an hour.
It can be advantageous to attain the pyrolysis temperature quickly, for example by ramping at a rate of about 25C per minute.
It can be advantageous to precarbonize the carbohy-drate by washing with an acid (such as concentrated sulfuric acid) before pyrolysis.
Compounds of the invention can be used as a portion of an electrode in various electrochemical devices based on insertion materials (eg. supercapacitors, electrochromic devices, etc.). A preferred application for these com-pounds is use thereof as an electrode material in a bat-tery, in particular a non-aqueous lithium ion battery comprising a lithium insertion compound cathode; a non-aqueous electrolyte comprising a lithium salt dissolved in a mixture of non-aqueous solvents; and an anode comprising the carbonaceous insertion compound of the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 shows the definition of R on an almost featureless x-ray diffraction pattern of a pre-graphitic carbon in the region around the {002} peak.

Figure 2 shows a cross-sectional view of a conven-tional lithium ion spiral-wound type battery.
Figure 3 shows the powder x-ray diffraction profiles for the directly pyrolyzed sucrose samples (numbers 1, 2, 4, 5, 6, and 7) of the Inventive example. The data pres-ented has been offset sequentially by 500 counts for clarity.

21~9853 Figure 4 shows the powder x-ray diffraction profiles for samples pyrolyzed at 1000 C from starch and cellulose precursors (numbers 14, 15, 16, 17, and 18). The data have been offset sequentially by 500 counts for clarity.

Figures 5a and b show the voltage versus capacity plots for the second cycle for representative batteries comprising sample numbers 8, 2, 10, 11, and 12 pyrolyzed between 700C and 1100C. Figure 5a is a magnified view of a portion of Figure 5b. The onset of lithium plating during discharge and the termination of lithium stripping during charge is indicated by the vertical lines for sample 8 in Figure 5a. The data has been offset sequentially for clarity by 0.05V in Figure 5a and by O.lV in 5b.
Figures 6a and b show the voltage versus capacity plots for the second cycle for representative batteries comprising sample numbers 2, 18, 14, 16, and 15 pyrolyzed at 1000C. Figure 6a is a magnified view of a portion of Figure 6b. The data has been offset sequentially for clarity by 0.05V in Figure 6a and by O.lV in 6b.

Figure 7 shows the capacity versus cycle number for the two batteries containing electrodes made from sample number 8.

Figure 8 shows the capacity versus cycle number for one of the two batteries containing electrodes made from sample number 14.
Figure 9 shows the capacity versus cycle number for the two batteries containing electrodes made from sample number 18.

Figure 10 compares the voltage profiles of the first two cycles of the batteries comprising sample number 8 and a previously pyrolyzed sample of phenolic resole resin.

Figure 11 shows the differential capacity versus voltage during charging of the two batteries of Figure 10.

Figure 12 shows the x-ray diffraction patterns of several oxidized samples from the Illustrative Examples re - burnoff.

Figures 13a (magnified view) and b show the voltage versus capacity plots for the second cycle of representa-tive batteries from the Illustrative Examples re burnoff.

Figures 14a and b show plots of the intensity versusscattering angle and ln (intensity) versus q2 respectively for the samples from the Illustrative Examples re small angle scattering.

DET~TT-Rn DESCRIPTION OF THE S~ lC
EMBODIMENTS OF THE lNv~L.llON

Insertion compounds of the invention comprise hard pre-graphitic carbonaceous hosts having very poorly stacked graphene layers, little hydrogen content, and a small surface area accessible to common non-aqueous electrolyte solutions. This type of insertion compound was the subject of the aforementioned Canadian Patent Application Serial No. 2,138,360 (hereinafter referred to as CA 2,138,360), the disclosure of which is incorporated herein by refer-ence. Similarly, this type of insertion compound was the subject of the aforementioned Canadian Patent Application Serial No. (unassigned), by U. von Sacken et al., titled 'Phenolic Resin Precursor Pre-Graphitic Carbonaceous Insertion Compounds and Use as Anodes in Rechargeable Batteries', filed April 5, 1995 (hereinafter referred to as CA 2,146,426), the disclosure of which is incorporated herein by reference.
The carbonaceous hosts of said compounds can be derived from pyrolysis products of suitable precursors.

214~8~3 Suitable precursors are those that can be pyrolyzed such that little hydrogen remains (ie. such that the H/C atomic ratio is less than about 0.1) and yet such that the host does not graphitize to such an extent that the empirical parameter R as determined by x-ray diffraction pattern exceeds about 2.2. In CA 2,138,360, R was defined as the {002} graphite peak height divided by the background level.
R provides a convenient empirical means of quantifying the degree of graphitization of these compounds which have almost featureless x-ray diffraction patterns. In the instant application, R has been given a slightly different definition. Instead of determining the ratio at the m~l ml~m peak height including the background (as in CA
2,138,360), the ratio in this application is determined at the m~l mllm peak height excluding the background. Figure 1 illustrates the new definition of R on a representative, almost featureless x-ray diffraction pattern of a pre-graphitic carbon in the region around the {002} peak. A
tangential line is drawn below the {002} peak to exclude the background. The point where a parallel line just intersects the peak defines the position of the maximum peak height excluding the background. Although the defini-tion of R has changed, the magnitude of the change is small for practical purposes at this time. Values determined earlier by the previous definition for R are about the same as would be determined by the new definition. The latter definition, however, is believed to be a more rigorous definition.
As mentioned in CA 2,138,360, this type of insertion compound can have a high reversible capacity for alkali metal insertion. When the alkali metal is lithium, it was demonstrated that such insertion compounds additionally can have a low irreversible capacity and a small voltage hysteresis between insertion and extraction. It appears necessary for the carbonaceous host to have a small surface area accessible to common non-aqueous electrolyte solutions in order to obtain these additional advantages. As men-- 21~98S3 tioned in CA 2,146,426, this is especially important for application in lithium ion batteries. Electrolyte reac-tions that consume lithium occur at the anode surface in such batteries. Thus, use of an anode having a large surface area accessible to electrolyte results in substan-tial irreversible capacity loss and electrolyte loss.
These losses are avoided if the anode surface is not accessible to the electrolyte. As mentioned in CA
2,138,360, the surface area of the carbonaceous hosts accessible to common non-aqueous electrolytes is not directly measurable. However, it can be inferred to a certain extent by the observed irreversible capacity of the lithium-carbon insertion compound. Desirably, the access-ible surface area is such that the irreversible capacity is less than about half that of the reversible capacity for practical application in lithium ion batteries. Preferab-ly, the irreversible capacity is smaller still, being less than about a third that of the reversible capacity.
In CA 2,138,360, a method for estimating the access-ible surface area was employed that was based on the absorption of methylene blue. The electrolyte-accessible surface area was often sufficently small if the methylene blue absorption capacity of the carbonaceous host was also found to be less than about 4 micromoles per gram of host.
However, recently compounds have been synthesized that meet the methylene blue criterion yet still appear to have unacceptably large electrolyte-accessible surface area.
This is demonstrated in the illustrative examples to follow.
The BET method is a conventional way of measuring surface area accessible to nitrogen. This too provides a means of estimating the electrolyte-accessible surface area of the carbonaceous host. The electrolyte-accessible surface area can be sufficently small when corresponding BET surface area values are as high as about 300 m2/gram of host. However, hosts might conceivably have larger BET

- 21~9853 surface areas and still have sufficently small electrolyte-accessible surface area.
Recently, it has been discovered that carbonaceous hosts of the invention which have been slightly oxidized can have significantly increased electrolyte-accessible surface area without exceeding or significantly exceeding the aforementioned methylene blue absorption or the BET
criteria. Thus, oxidizing represents a means of ruining otherwise suitable carbonaceous hosts. Also, oxidizing represents a means for fine tuning the characteristics of the hosts such that the electrolyte-accessible surface area of the host cannot be adequately distinguished by the aforementioned methods of estimating. Such oxidation results in the formation of surface oxides which can be subsequently removed by pyrolyzing at high temperature (eg.
1000C) under inert gas. Under these circumstances, a weight loss of 5~ or more can be indicative of a host ruined by oxidation. Conversely, a weight loss of less than about 5~ after oxidation can be indicative of a suitable carbonaceous host. This is demonstrated in the illustrative examples to follow.
As is known to those skilled in the art, the capacity values of lithium carbonaceous insertion compounds can vary depending on the choice of non-aqueous electrolyte employed. Certain choices might always result in large irreversible capacity values. A solvent mixture known in the art to be associated with low irreversible capacities comprises ethylene carbonate and diethyl carbonate.
Electrolytes based on this solvent mixture can be suitable for evaluating the electrolyte-accessible surface area.
Various precursors can be pyrolyzed to provide the aforementioned type of carbonaceous hosts that have a high reversible capacity for alkali metal insertion. In CA
2,138,360, it was shown that epoxy precursors were suitable precursors. In CA 2,146,426, it was shown that phenolic resin precursors were suitable. Now, carbohydrates have also been found to be suitable precursors.

The McGraw-Hill Dictionary of Scientific and Technical Terms, McGraw-Hill, Inc., New York, defines a carbohydrate as any of the group of organic compounds composed of carbon, hydrogen and oxygen, including sugars, starches and celluloses.
The carbohydrate precursors of the subject invention encompass all carbohydrates composed of carbon, hydrogen and oxygen.
The sugars can comprise monosaccharides (simple sugars), disaccharides (more complex sugars including sucrose, the common table sugar), and polysaccharides, the latter comprising the entire starch and cellulose families.
Starch is a polymer of ~-D-glucose while cellulose is a polymer of ~-D-glucose. The glucose rings in cellulose have a different relative orientation than in starch. Isomers or compounds with such orientation differences can behave radically differently in biochemical processes. However, in inorganic processes, such differences may not matter.
For example, the physical characteristics and electrochemi-cal behaviour of insertion compounds prepared by pyrolyzingdifferent isomers would likely be the same.
The use of carbohydrates as a precursor offers certain advantages over epoxy and/or phenolic resin precursor options. CA 2,138,360 discloses that the pyrolysis of epoxy novolac resins can give product yields near 30~.
However, such resins can cost about $5 per pound. CA
2,146,426 discloses that the pyrolysis of phenolic resins can give product yields near 60~. Such phenolic resins can cost less, being about $1 per pound. However, the pyrolysis process generates substantial amounts of tarry residue which is difficult to dispose of and may be carcinogenic.
Naturally occurring carbohydrates are attractive precursors because they are plentiful and relatively inexpensive. For example, oak (predominantly consisting of cellulose) can cost about $0.08 per pound. Even with a pyrolysis yield of 20~, this can result in a cost for the `~ 21498S~

product that is about 5 times less than the corresponding cost for phenolic resin derived product. Additionally, carbohydrate precursors can lead to a product with a high tap density which is needed for high volumetric energy density in lithium ion batteries. Finally, carbohydrate precursors can result in less tarry residue per gram of carbon produced than do phenolic resin precursors.
We have discovered that pyrolyzing suitable carbohy-drate precursors, and carbohydrate containing precursors, above 800C can provide pre-graphitic carbonaceous hosts which have low H/C atomic ratios (<0.1). Additionally, pyrolyzing at temperatures such that R is below 2.2 pro-vides for hosts with very high specific capacities for lithium. The specific capacity for lithium increases as R
decreases. As shown in the examples to follow, pyrolysis products can be prepared with R values less than about 2 that have large reversible capacities. These products also have methylene blue absorption values less than 4 micro-moles per gram and BET values less than 300 m2/gram and do not exhibit large irreversible capacities nor severe hysteresis in voltage upon insertion or extraction of lithium. Tap densities as high as 0.7 g/ml can also be achieved.
The pyrolysis should be performed under a controlled atmosphere to prevent formation of undesired oxides of carbon. A suitable reaction system could consist of a reaction tube (quartz for example) installed in a conven-tional tube furnace wherein the tube has sealed inlet and outlet connections for purposes of controlling the atmos-phere therein. The carbohydrate precursor/s could thus bepyrolyzed in the reaction tube under an inert gas flow or even under reduced or elevated pressure.
The electrolyte-accessible surface area of the pyrolyzed product should be relatively small. In general therefore, it is undesirable to oxidize the precursor during pyrolysis as this would be expected to result in an increase in this area. Since the by-product gases of pyrolysis include unwanted oxidizing gases, it is desirable to remove these quickly.
Ramping the furnace temperature relatively quickly to the pyrolysis temperature and minimizing the pyrolysis period is also generally desirable in order to minimize graphitization of the product. Also, it can be advantage-ous to precarbonize the carbohydrate prior to pyrolysis at a low temperature. A means for so doing is to wash the carbohydrate with a strong acid which is subsequently rinsed away.
The aforementioned product has no alkali metal inserted as prepared. Alkali metal atoms, in particular Li, can be inserted thereafter via conventional chemical or electrochemical means (such as in a lithium or lithium ion battery).
Generally, powdered forms of such compounds are used in electrode applications and thus a grinding of the pyrolyzed product is usually required. A variety of embodiments, in particular various battery configurations, are possible using electrode material prepared by the method of the invention. Miniature laboratory batteries employing a lithium metal anode are described in the examples to follow. However, a preferred construction for a lithium ion type product is that depicted for a conven-tional spiral-wound type battery in the cross-sectional view of Figure 2. A jelly roll 4 is created by spirally winding a cathode foil 1, an anode foil 2, and two micro-porous polyolefin sheets 3 that act as separators.
Cathode foils are prepared by applying a mixture of a suitable powdered (about 10 micron size typically) cathode material, such as a lithiated transition metal oxide, possibly other powdered cathode material if desired, a binder, and a conductive dilutant onto a thin aluminum foil. Typically, the application method first involves dissolving the binder in a suitable liquid carrier. Then, a slurry is prepared using this solution plus the other powdered solid components. The slurry is then coated 2149~j3 uniformly onto the substrate foil. Afterwards, the carrier solvent is evaporated away. Often, both sides of the aluminum foil substrate are coated in this manner and subsequently the cathode foil is calendered.
Anode foils are prepared in a like manner except that a powdered (also typically about 10 micron size) carbon-aceous insertion compound of the invention is used instead of the cathode material and thin copper foil is usually used instead of aluminum. Anode foils are typically slightly wider than the cathode foils in order to ensure that anode foil is always opposite cathode foil.
The jelly roll 4 is inserted into a conventional battery can 10. A header 11 and gasket 12 are used to seal the battery 15. The header may include safety devices if desired. A combination safety vent and pressure operated disconnect device may be employed. Figure 2 shows one such combination that is described in detail in Canadian Patent Application No. 2,099,657, Alexander H. Rivers-Bowerman, ~Electrochemical Cell and Method of Manufacturing Same', filed June 25, 1993. Additionally, a positive thermal coefficient device (PTC) may be incorporated into the header to limit the short circuit current capability of the battery. The external surface of the header 11 is used as the positive terminal, while the external surface of the can 10 serves as the negative terminal.
Appropriate cathode tab 6 and anode tab 7 connections are made to connect the internal electrodes to the external terminals. Appropriate insulating pieces 8 and 9 may be inserted to prevent the possibility of internal shorting.
Prior to crimping the header 11 to the can 10 in order to seal the battery, electrolyte 5 is added to fill the porous spaces in the jelly roll 4.
Those skilled in the art will understand that the types of and amounts of the component materials must be chosen based on component material properties and the desired performance and safety requirements. The compounds prepared in the Examples to follow can have somewhat -- 21498~ 3 increased irreversible capacity for lithium along with an increased reversible capacity over that of many typical commercial carbonaceous anode materials. Also, the highest tap density of the Example compounds is still somewhat lower than that of typical commercial anode materials.
This must be taken into account in the battery design.
Generally an electrical conditioning step, involving at least the first recharge of the battery, is part of the assembly process. Again, the determination of an appropri-ate conditioning step along with the setting of the batteryoperating parameters (eg. voltage, current, and temperature limits) would be required of someone familiar with the field.
Other configurations or components are possible for the batteries of the invention (eg. prismatic format). A
miniature embodiment, eg. coin cell, is also possible and the general construction of such cells is described in the laboratory coin cell examples to follow.
Without wishing to be bound by theory, adversely or otherwise, the inventors offer the following discussion regarding this type of hard carbonaceous host compound in order to explain how structural characteristics relate to the electrochemical characteristics and therefore what structural characteristics are desirable for electrochemi-cal applications. For overall simplicity, the followingdiscussion pertains to lithium insertion compounds.
However, where appropriate, similar comments apply for other alkali metals.
The presence of substantial hydrogen in carbonaceous materials of the prior art prepared by pyrolysis at low temperatures (between 550C and 750C) correlates with very high specific capacity but also with large hysteresis between insertion and extraction voltage. These effects may involve a binding of the inserted lithium and the hydrogen.
Hard carbonaceous materials having little hydrogen can still exhibit specific capacities exceeding that of graph-214~853 ite however. When poorly stacked graphene layers are present, it may be possible to adsorb lithium onto the surfaces of each side of the layers. These surfaces are found within the carbon particles, on the atomic scale. In graphite, the layers are well stacked in a parallel fashion and intercalation of lithium to a composition of LiC6 is possible (corresponding to about 370 mAh/g and one interca-lated layer of lithium per graphene sheet). In materials with poorly stacked layers, unshared lithium layers might possibly be found on each side of the graphene sheets, resulting in compositions up to almost Li2C6 (corresponding to about 740 mAh/g). Thus, the number of single layer graphene sheets in the carbonaceous material may be import-ant vis a vis specific capacity.
Information about the average number, N, of stacked graphene sheets in a carbon between serious stacking mistakes can be obtained by x-ray diffraction. This number N, multiplied by the average layer spacing is commonly given the name, Lc. It may therefore be desirable to make carbonaceous materials with N about 1 and with very small Lc (eg. less than about 5A). The {002} Bragg peak measured in a powder x-ray diffraction experiment is normally used to determine Lc and N. For N=l, there is no {002} peak since there are no stacked parallel graphene layers to create interferences. (Such a carbon sample might be thought of as having single graphene sheets arranged as in a ~house of cards'.) As N increases (beginning to stack the deck of cards), the {002} peak increases in height and decreases in width. Simultaneously, the background on the low angle side of the peak decreases, as N increases.
Herein, the empirical parameter R is used for purposes of describing such structures and can be used to distinguish the stacking order in very disorganized materials.
Materials having very small R values (about 1) would have N values near 1. Materials having R near 5 would have significantly larger N, possibly with N about 10. Thus, increases in R can be interpreted as increases in the average N in the sample.
The 'house of cards' structure of such disorganized carbons implicitly suggests the presence of significant voids or pores in the structure. The pore number, size, and shape (particularly of the openings) would be expected to relate to the ability of the single-layer sheets to absorb lithium on both sides and also to have an impact on the electrolyte accessible surface area. For instance, a relatively large number of single-layer sheets implies the existence of a relatively large number of 'pores' between sheets . The preferred pore size is large enough to allow lithium to adsorb on both sides yet not allow access to non-aqueous electrolyte (a size in the nanometer scale).
Pores can be bottle shaped having neck openings that are small enough to exclude electrolyte from the interior.
However, the same pores can still have interiors that are large enough to easily accommodate electrolyte. Samples having numerous such bottle shaped pores can therefore have either relatively large or small surface area depending on how it is measured. For example, if the pore opening is large enough to admit nitrogen but not methylene blue, then nitrogen can be adsorbed on the interior pore surfaces whereas methylene blue cannot. Additionally, minor differ-ences in the size of the pore openings can result indramatically different electrochemical results. Conceivab-ly, a sample could have enormous internal pore surface area (>>300 m2/g) as determined by BET that is inaccessible to the larger methylene blue molecules. If the effective size of the non-aqueous electrolyte is intermediate to that of nitrogen and methylene blue, such a sample might have either an enormous or a negligible electrolyte accessible surface area depending on minor differences in the size of the pore openings.
A possible means of gradually increasing pore size and openings thereof is by burning off small amounts by heating samples in an oxygen containing atmosphere. (Previous studies on active carbon (J.S. Mattson et al., Activated Carbon, Marcel Dekker Inc. NY, 1971 and F. Rodriguez-Reinoso et al., Chemistry and Physics of Carbon, Vol. 21, Edited by P. A. Thrower, pl) showed that both the sizes and shapes of pores can be manipulated by physical and chemical activation processes. Note however that most activated carbons are not acceptable host materials for electrochemi-cal lithium insertion because the pore sizes are too large (on the micrometer scale)). Thus, oxidizing may be a means for incrementally increasing both the interior pore surface area and the critical size of pore openings. Some results pertaining to this subject are shown and discussed in the Illustrative Examples to follow.
Small angle x-ray scattering has been widely used for the study of pore structure in carbons (see for example, H.
Peterlik et al., Carbon, 32 (1994) p.939). The presence of a substantial number of micropores results in substantial scattering of x-rays at small angle. Thus, carbonaceous hosts of the invention are expected to exhibit such scat-tering. Conversely, the absence of such scattering isindicative of the absence of micropores (as shown in an Illustrative Example to follow). Note that pores can be closed (ie. no openings) and materials comprising such pores will still show substantial x-ray scattering. Thus, carbonaceous hosts can be imagined that have more pore volume, lower R values, and more small angle scattering, yet less lithium capacity and less irreversible capacity than a comparable host if many pores are closed.
The Guinier theory and formulae (in A. Guinier, Small-angle scattering of X-rays, Wiley and Sons, NY, 1955) can be used to determine pore sizes from the small angle scattering intensity assuming homogeneous spherical pore sizes and randomly located pores. The radius Ro of the pores is related to the radius of gyration, Rgl by:
Rg = (3/5)1/2 R8 - 2149~ 3 The intensity, Iql at wavevector q is related to the radius of gyration by:

Iq =NV2 exp(-q2Rg2/3) where N is the number of pores and V is their volume. This theory therefore predicts a straight line relationship between ln (Iq) and q2. Although the aforementioned assumptions do not generally hold, such a straight line relationship was observed in the case of the following Inventive examples. This suggests that these examples comprise pores of approximately uniform size. Generally speaking, uniform pore sizes are preferred since sizes at the small extreme (ie. in the range of the normal inter-atomic distances) would contribute less to reversible alkali metal capacity, while sizes at the larger extreme (ie.,30A) would be more accessible to electrolyte leading to irreversible capacity (as shown in an Illustrative Example to follow).
A. Mabuchi et al., J. Electrochem. Soc., Vol. 142, No.4, April 1995, show radii of gyration values derived from small angle scattering data for mesocarbon microbeads containing substantial hydrogen. The effective pore sizes are relatively very large (Rg of approximately 37 A and up) and the compounds exhibit significant hysteresis in their voltage curve upon insertion/extraction of lithium.

Bac~y-G~d information for the Examples The following examples are provided to illustrate certain aspects of the invention but should not be con-strued as limiting in any way.
Powder x-ray diffraction was used to characterize samples using a Seimens D5000 diffractometer equipped with a copper target x-ray tube and a diffracted beam mono-chromator. The diffractometer operates in the Bragg-Brentano pseudofocussing geometry. The samples were 214~53 made by filling a 2mm deep well in a stainless steel block with powder and levelling the surface. The incident slits used were selected so that none of the x-ray beam missed the sample in the angular range from 10 to 35 in scatter-ing angle. The slit width was fixed during the measure-ment. This ensured reproducibility in the measured values of R.
Where indicated, small angle x-ray scattering data was collected using the preceding diffractometer operating in transmission geometry. Samples were prepared by filling a rectangular frame, having kapton windows, with powder. The prepared samples were about 1.5 mm thick. The incident, antiscatter, and receiving slits were all set to their minimum values of 0.1, 0.1, and 0.1 mm respectively.
Minimum scattering angles of about 0.5 could be reached with this equipment, which corresponds to a wavevector q of about 0.035A-1. The intensity scattered at 2~=1 was measured and divided by the sample mass to get a relative measure of the number of pores in the samples. This value was denoted I1. Rg was determined using straight line fits to the small angle scattering data plotted as ln (inten-sity) versus q2 and the aforementioned formula.
Carbon, hydrogen, and nitrogen content was determined on samples using a standard CHN analysis (gas chromatographic analysis after combustion of the samples in air). The weight percents so determined had a standard deviation of +0.3~. In every case, the carbon content was greater than 90~ of the sample weight and the hydrogen content was less than 3~. The H/C atomic ratio was esti-mated by taking the ratio of the hydrogen and carbon weightpercentages and multiplying by 12 (the mass ratio of carbon to hydrogen). The oxygen content of the samples was not analyzed.
Where indicated, the absorption capacity for methylene blue (MB) was determined as generally described in CA
2,138,360. Herein, however, a lM methylene blue (MB) 214~853 titrating solution was used and the stepwise additions were not of constant magnitude.
In CA 2,138,360, it was mentioned that the adsorption of nitrogen proceeded for hours during conventional BET on these hard carbon products. Thus, the reliability of adsorption values was considered questionable. Herein, single point BET surface area measurements were made using a Micromeritics Flowsorb 2300 surface area analyzer.
Carbon samples were outgassed under inert gas for several hours at 140C before each measurement. The adsorption of nitrogen (from a 30~ nitrogen in helium mixture) at 77K on the samples was allowed to proceed for several hours.
Adsorption was considered complete when the thermal conduc-tivities of the gas stream before and after the sample were equal, indicating identical gas compositions. The amount of N2 adsorbed was determined by that which desorbed when the sample temperature was increased to room temperature. Two measurements were made for each sample and the results reported represent the average of the two desorptions. The measurements usually can be duplicated satisfactorily with an accuracy within +3~. Standard methods were used to calculate the specific surface area of the sample access-ible to N2 molecules.
Where indicated, tap densities were measured using a Quantachrome Dual Autotap device. Samples were placed in a 10 ml graduated cylinder and subjected to 500 standard taps.
Laboratory coin cell batteries were used to determine electrochemical characteristics of the samples including specific capacity for lithium. These were made with conventional 2325 size hardware and using the components and assembly procedure described in CA 2,146,426. The electrolyte used was therefore a lM solution of LiPF6 in ethylene carbonate and diethyl carbonate in a volume ratio of 30/70.
After construction, the coin cell batteries were thermostatted at 30 + 1C, and then charged and discharged 21~9853 using constant current cyclers with + 1~ current stability.
Data was logged whenever the battery voltage changed by more than 0.005 V. Currents were adjusted to be 18.5 mA/g of active material for the initial two cycles of the battery. Much of the discharge capacity of the example carbons is very close to the potential of lithium metal.
Thus, the special testing methods described in CA 2,146,426 were used to determine the full reversible capacity (ie.
involves cycling lithium in excess of the carbon capacity and thus plating/stripping lithium as well). Discharge and charge currents of 74 mA/g and 37 mA/g respectively were used thereafter for extended cycle testing between 2.0V and the onset of lithium plating.

Inventive ExAmple Carbonaceous hosts of the invention were prepared using a variety of carbohydrate precursors. Table 1 lists the precursors used, along with their source and morphol-ogY.

21~853 Table 1. List of carbohydrate precursors C~b-~_' M ' ' Supplier M~.P'-'~L~

Table Sugar (sucrose) Canada Safeway Powder SucroseBDH Inc.(Toronto), Reagent Powder grade Starch BDH Inc. (Toronto) Reagent Powder grade Walnut ShellsCanada Safeway Small pieces of shell sepqr~ted from the nut Filbert ShellsCanada Safeway Small pieces of shell sep from the nut Almond ShellsCanada SafewaySmall pieces of shell sepqr,qt~d from the nut Red Oak Reimer Hardwoods 1 cm3 chunks cut from furni-(AWo~rurd, B.C.) ture-grade lumber Maple Reimer Hardwoods 1 cm3 chunks cut from furni-(AWo~ruld, B.C.) ture-grade lumber Precursors (typically batches between 1 and 25 grams) were contained in nickel foil boats and placed within a stainless steel or quartz furnace tube. Prior to heating, the tube was flushed with argon (Ultra High Purity Grade -Linde) for 30 minutes to remove air. The samples were heated from room temperature to a desired pyrolysis tem-perature at a rate of 25C/min. They were held at the pyrolysis temperature for 1 hour. The furnace power was then turned off and the samples were cooled to near room temperature within the furnace tube under flowing argon (a process which took several hours). The samples were weighed before and after pyrolysis, so that the yield could be determined. Certain samples were pyrolyzed at tempera-tures of 1200C and higher. These samples were first 21~98~3 pyrolyzed to 1100C as in the preceding. Thereafter, pyrolysis was continued in a similar manner using a Centaur Series 10 furnace.
Some samples of table sugar (hereinafter denoted simply as 'sugar') were precarbonized by washing in excess concentrated sulfuric acid. About 50 grams of table sugar was first mixed with about 100 cc of concentrated sulfuric acid, added slowly. The resulting char was briefly crushed, washed with boiling water, and filtered to recover the solids. Rinsing was repeated until the filtrate gave the same pH (about 6) as the tap water used for rinsing.
The product was dried overnight at 110C overnight before pyrolysis. The carbon yield was calculated for these samples by the final carbon mass divided by the initial weight of table sugar. These samples are denoted as 'a-sugar' samples.
The pyrolyzed samples were ground to powder and analyzed as described in the preceding. Results of these measurements are tabulated in Table 2.

Table 2. Summary of cl .~r..~ istics of pyrolyzed ~;a~luOhrdl dte ~e~ , D
No. E~. ,or Pyrolysis Yield C H N H/C R Rg (A) I~ Tap MB SurfaceReversible~ D~Ie Temp. % wt. % wt % wt. %atomic (countsDensity(~nolesareH C .
(C) per~g) (~/cc) perg)(m2/~)(i'OmAh/~t20mAh/g) sugar 1100 12* 97.3 0.^8 0.96 0.034 .91 '.~ 19. 0.69 - 15 '37 ,'^4 13~, 141 2 sucrose 1000 8* 96.9 0.~2 O. ~ 0.05 .96 ' .:^ 16. 0.91 <2.9 31 r29. ' ~ 13, 137 sugar 1000 1:* ~'7 O.' 1 O. 7 0.063 .~ 18.4 - - 220 ~42, 3~ 20', 222 SUBar 900 1'~* 1l .5 O.~ 9 O.J;~0.074 . ' 4.8^ 11.9 0.78 - 58 '90, " 17', 8 sugar `00 1 * ~.2 0.~4 O."~ 0.12 1.76 4.74 10 0.8 - 20 624, 62^ 1"-, "1"
6 sugar ~00 1"* ~1,.8 1.41 0."1 0.18 1.'8 4 13.3 0.62 - ^50 690, 40 274, ~6 7 sugar 600 14* 92.5 2.28 0.1 0.3 1.~6 : .13 6.8 0.67 ~60 -6~, 790 4'5, 3:
a-sugar 1100 "7 9-.2 - 0.25 - 1.63 ' .01 15.8 - < 1.5 1.8 ' 6~, '67 ;4, 7' 9 a-su&ar 1000 O ~7 0.49 0.36 0.061 1.7 ' .27 16." - - 180 ~7, 460 1^0, 1~7 a-sugar 900 ^9 9' .4 0.55 0.42 0.0~9 1.~'~4. ~ 11.; - - 68 ~1, nOS 1 ', 8 11 a-sugar ~00 29 94.3 0.~: 0.2-~ 0.1" 1.6~ 4.' 10.' - - ~0 '~~, '66'`^', '`20 12 a-sugar ~00 30 91.4 .': 0."6 0.2 1.~ 4.: 8.6 - - ~:0 '~, '75 :~', :78 13 a-sugar 600 JO 92.9 ".~ 0.21 0.31 1.3 3.0~ 6.5 - - ~0 66', 706 '^1, 466 14 shrch 1000 1 91.7 O.' O. 4 0.068 1.8 5. 23.2 0.76 <2.5 30 49., ~96 1~6, 199 filbert 1000 23 - - - - 1.9" '. ' 22.7 0.63 - 180 412, '00 1 3, 19' 16 walnut 1000 2: - - 1.8' ' .~ 16 0.63 - 60 ~90, ~90 1~7~ 12 1' almond 1000 23 - - - - 2 5.9: 17 0.6 - 46 95, :-71 16-, 18' 1' i~oak 1000 18 - - - - 1.85 '.53 19.1 0.54 <3.5 13 '18, 515 14', 159 19 maple 1000 18 - - - - 1.98 5.58 28.2 0.56 - 63 497, 503 140, 127 maple 1100 18 - - - - 1.86 5.54 20.0 0.56 - 11 547, 524 -, 104 21 sugar 1200 11* - - - - 1.98 5.66 22.5 0.60 - 5.5 374, 379 71, 59 22 sugar 1400 11* - - - - 2.37 6.08 31.5 0.63 - 7.9 284, 296 35, 38 23 sugar 1600 11* - - - - 3.09 6.53 46.7 0.58 - 6.3 208, 210 31, 34 24 a-sugar 1200 28 - - - - 1.83 5.78 17.3 0.77 - 1.3 367, 368 38, 45 oo a-sugar 1400 28 2.02 5.95 21.8 0.76 1.2 280, 274 25, 25 C
26 a-sugar 1600 28 - - - - 2.48 6.46 33.2 0.73 - 1.2 198, 202 24, 25 C~J
27 shrch 1100 11 - - - - 1.88 5.59 29.6 0.71 - 4.9 523, 526 154, 150 28 shrch 1200 10 - - - - 2.13 6.06 42.6 0.65 - 3.1 337, 389 58, 54 29 shrch 1400 10 - - - - 2.40 6.21 42.8 0.58 - 3.8 277, 286 32, 30 shrch 1600 10 - - - - 2.89 6.66 65.4 0.55 - 3.3 212, 207 35, 32 31 oak 1100 19 - - - - 1.78 5.47 19.5 0.59 - 12.1 587, 538 115, 120 32 oak 1200 18 - - - - 2.02 5.94 30.9 0.55 - 4.8 334, 330 38, 60 33 oak 1400 18 - - - - 2.26 6.11 35.5 0.55 - 4.7 261, 270 33, 35 34 o~k 1600 18 - - - - 2.66 6.58 46.8 0.53 - 4.6 192, 193 30, 29 * - The yield for these samples was difficult to estimate due to boiling of the samples with 9C~:m' ' ~ ion outside the sarnple boat.
** - The; ~ for both (2) batteries tested are reported.

2149~3 Yields near 20~ were readily achieved using this simple pyrolysis method. The H/C ratio was less than 0.1 for heating temperatures above 800C. Tap densities up to 0.9 g/cc were obtained.
5Figure 3 shows the powder x-ray diffraction profiles of some pyrolyzed sucrose samples (numbers 1, 2 (BDH
source), 4, 5, 6, and 7) as a function of pyrolysis tem-perature. The {002} Bragg peak near 22is poorly formed in all these samples, indicating materials made up predomi-nantly of single carbon layers arranged somewhat like a ~house of cards'. The {100} and {110} Bragg peaks near 44 and 80 respectively can be used to estimate the lateral extent of the graphene sheets (this is the distance over which the sheets are more or less flat). The lateral dimension ranges from near lOA for the sample pyrolyzed at 600C to near 25A for the sample pyrolyzed at 1100C. The diffraction patterns for the samples made from acid-washed sugar (numbers 8-13) show similar features.
Figure 4 shows the x-ray diffraction profiles for the samples pyrolyzed at 1000C from starch and cellulose precursors. The patterns shown are for samples number 18, 17, 16, 15, and 14 from top to bottom in Figure 4. These patterns resemble one another and additionally resemble the pattern of sample number 2 in Figure 3, suggesting similar structural arrangements.
Higher pyrolysis temperature tends to produce smaller BET surface area. (However, samples number 3 and 9 have anomolously high surface areas.) During pyrolysis, the samples emit water, CO2, and other gases. If the argon flow rate is too small, these gases remain in the tube and oxidize the samples leading to high surface areas.
Laboratory coin cell batteries were constructed using these pyrolyzed samples as described in the preceding.
Figures 5a and b show the voltage versus capacity plots for the second cycle for representative batteries comprising samples number 8, 2, 10, 11, and 12 prepared between 700C
and 1100C. Samples number 8, 2 ,and 10 show large revers-- 2149~53 ible capacities and little voltage hysteresis. (Materials prepared at 800C and below can contain substantial hydrogen leading to significant hysteresis in the voltage plateaus.
Nevertheless, such carbons, if prepared cheaply enough, might be useful for some battery applications.) From the data in Table 2, irreversible capacities are seen to decrease as the pyrolysis temperature increases.
Samples number 3 and 9 have significantly less reversible capacity than does sample 2, prepared at the same tempera-ture. This might be attributed to differences in thesamples as evidenced by the larger surface area of samples

3 and 9 compared to sample 2.
Figures 6a and b show the voltage versus capacity plots for the second cycle for representative batteries made with sucrose, cellulose, and starch precursors pyrolyzed at 1000C . Data is shown for sample number 2 (for comparison), 18 (oak), 14 (starch), 16 (walnut shells) and 15 (filbert shells). Samples 2, 18, and 14 show excellent behavior, and it is likely that the performance of the other samples could be improved through changes to the pyrolysis process. Thus, pyrolyzed products made from oak, starch, and walnut shells gave similar behavior to that made from sucrose.
Some of the batteries underwent extended cycling as described in the preceding. Figures 7, 8, and 9 show the capacity versus cycle number for batteries containing electrodes of samples 8, 14, and 18 respectively. These batteries show little capacity loss upon cycling and retain cycling capacities near 500 mAh/g. The battery containing sample 14 (Figure 14) shows the poorest performance. This may be due to the large impurity content in the sample (as per Table 2, this sample is only 91.7~ carbon by weight).
Thus, carbohydrates in general, as defined earlier, can be used as precursors to prepare insertion compounds having excellent electrochemical characteristics by pyroly-zing at temperatures between about 800C and about 1200C.
Some differences were noticed between the samples prepared 214~3 from different carbohydrate precursors, but these may be due in part to the differing amounts of impurities in the naturally occurring sources. For example, the wood and shell samples comprise significant, varied amounts of lignin and/or oil.

Comparative ExAmples For purposes of comparison, the characteristics of sample number VII from CA 2,138,360 (a sample of DEN 438 epoxy novolac resin from DOW pyrolyzed at 1000C) and sample B1000 from CA 2,146,426 (a sample of phenolic resole resin #29217 from Occidental Chemical Co. pyrolyzed at 1000C) are reported in Table 3 below.

Table 3. Cl~ of C , '~ F p' Sample HJC R Rg Il MB SurfaceReversibleIrreversible 20Number (counts(llmoles area C . - ~Y Capacity per mg) per g) (m2lg) (mAh/g) (mAh/g) ~1 0.031.585.7* 14* < 4 217 570 150 (epoxy) B1000 0.041.37 5.5 10 - 235 560 200 (phenolic 25resin) * Obtained from another sample similar to VII.

Voltage curves of the first two cycles of the bat-teries comprising Inventive Example sample 8 and sampleB1000 from CA 2,146,426 are shown in Figure 10. (The B1000 sample was discharged and charged at 37 mA/g.) The curves are similar. Figure 11 compares the differential 214~85 3 capacity, measured during charging, of the two batteries of Figure 10. Within error, these are identical.

The insertion compounds of the invention can have the same physical and electrochemical characteristics as those prepared from pyrolyzed epoxy and/or phenolic resins.

Illustrative Examples re burnoff A first amount of DEN 438 epoxy novolac resin (from DOW Chemical) was cured with 20 weight ~ 4-aminobenzoic acid at 170C and pyrolyzed at 1000C to produce carbon-aceous material similar to sample number VII of the Com-parative Examples. Samples (about 1 gram each) were then oxidized to varying degrees in a furnace tube under a flow of extra dry air. This was accomplished by heating the samples at a rate of 10C/minute to different specific maximum temperatures (TmaX). The amount of carbon burned off was obtained by calculating the difference between the initial and final mass (accurate to + 0.1~).
Figure 12 shows the x-ray diffraction patterns of three of the preceding oxidized samples with varying weight ~ burned off. The intensity of the diffraction peaks decreases with ~ burnoff while the intensity at small scattering angles increases with ~ burnoff. The diffrac-tion peaks may be expected to decrease as the number of x-ray scatterers decreases. The increase in intensity at small angles is consistent with an increase in porosity of the sample. The In (intensity) versus q2 relationship was roughly linear in each case, and the derived values of Rg also suggest a small increase in pore size with ~ burnoff.
A second amount of DEN 438 epoxy novolac resin (from DOW Chemical) was cured with 20 weight ~ phthalic anhydride at 170C and then pyrolyzed at 1000C to produce carbon-aceous material similar to sample number VII of the Com-parative Examples. Samples (about 1 gram each) were then oxidized to varying degrees in a furnace tube under a flow 21498~3 of extra dry air. This was accomplished by heating the samples at a rate of 10C/minute to different specific maximum temperatures (TmaX). The amount of carbon burned off was obtained by calculating the difference between the initial and final mass (accurate to + 0.1%). Physical and electrochemical characteristics were determined as in the preceding Inventive Examples. Table 4 shows a summary of the values obtained. (The specific reversible and irre-versible capacities represent the average value determined from two test batteries.) This second set of pyrolyzed samples was then reheated at 1000C under argon to remove surface oxides. The weight loss after this reheating is also shown in Table 4. Where indicated, the specific capacities of the reheated samples were also determined.
The surface area as determined by BET increased markedly with burnoffs of only a few weight %. Also, there were significant differences noticed in the nitrogen adsorption kinetics. It took progressively less time for samples to fully adsorb nitrogen (from about 4 hours for sample I-1 down to less than 1 hour for sample I-8). By contrast, the MB absorption values did not increase sig-nificantly until after about 5% by weight was burned off.
As pore openings enlarge or as new openings are created, the rate and total amount of nitrogen adsorbed may be expected to increase. A corresponding increase in the amount of MB absorbed may be delayed until pore openings enlarge enough to admit the larger MB molecules.
Figures 13a (magnified view) and b show the voltage versus capacity plots for the second cycle of representa-tive batteries comprising the preceding samples. Samples I-1 and I-2 exhibit a low voltage plateau having substan-tial capacity (about 200 mAh/g). The capacity of this plateau decreases quickly with ~ burnoff and is virtually eliminated by about 5% burnoff. As shown in Table 4, the reversible capacity decreases initially with ~ burnoff, and then increases above about 10~ burnoff. Surface oxide 21~9~53 complexes formed during the oxidation process may account for this subsequent increase in capacity. Close ~m~ n-ation of the voltage plots for samples I-6 and I-8 in Figure 13b shows that this subsequent increase is associ-ated with lithium insertion near zero volts and extractionabove one volt (ie. with substantial hysteresis in the voltage curve). Such high hysteresis capacity is generally unsuitable for lithium ion battery applications.
The irreversible capacity increases with burnoff ~
approximately linearly with the BET surface area and beginning well before the MB absorption values start to increase. This suggests that the electrolyte molecules are accessing pore surfaces before the MB do (ie. electrolyte molecules are smaller than MB molecules).
The weight loss for the reheated samples is greatest for sample I-6 indicating that the amount of surface oxides is greatest for this sample. Qualitatively, this agrees with previous work in the literature wherein higher tem-perature oxidation results in fewer surface oxides. Upon reheating, both the reversible and irreversible capacity of the samples is reduced up to about 100 mAh/g, suggesting that the surface oxides play a role in both. The low voltage plateau, present in sample I-1, is not recovered after reheating, even for sample I-3 having only 1.2 ~
burned off by weight. Thus, even minimal oxidation can seriously, and irreversibly, degrade the performance of the compounds of the invention. The presence of surface oxides can indicate that such oxidation occurred. In turn, an observed weight loss upon heating a carbon sample under inert gas can indicate the presence of such surface oxides.
Additionally, the preceding illustrates the diffi-culties in quantifying electrolyte accessible surface areas using nitrogen or methylene blue molecules as substitutes for the electrolyte itself. If the latter is intermediate in size to the former, a sample can have an acceptable MB
absorption value but still not exhibit the advantages of the invention (eg. sample I-3). Conversely, an acceptable limit for the BET surface area is difficult to define since a sample can conceivably have an enormous internal surface area that is accessible by nitrogen but not electrolyte (Note that carbons having the advantages of the invention with BET surface areas as high as 212 m2/g were already made in CA 2,138,360. Sample I-3, with a 316 m2/g surface area, on the other hand does not have the advantages of the invention.) Table 4. Summary of characteristics of oxidized samples SampleTll"",Burrloff R R, I, MB Surface R,~ re~ers;~'~Weight Loss R,~ ;b'- Irre~/e.. ~;bl~
No. (C) Amount (~)(counts(~molesarea CapacityCapacity % after Capacity Capaclty (%) per mg) per g) (m2/g) (mAh/g) (mAhlg) reh~ g (mAh/g) (mAh/g) I-l 25 0.0 1.97 5.714.0 2.3 63i5 461 122 1.5 I-2 300 0.7 2.00 5.915.2 2.3 104i 15 459 171 2.8 - -I-3 400 1.2 2.00 6.119.2 1.9 316i8 331 365 7.0 305 240 ,~, I4 452 5.1 2.26 6.016.9 5.2 384i2 316 487 10.6 I-5 484 12.5 2.23 6.322.6 16.6 553i3 370 456 12.5 - -I~ 525 28.2 2.10 6.843.3 27.1 579i2 404 529 13.4 305 484 I-7 550 34.0 2.13 6.735.8 28.1 591i2 397 526 10.7 ~ ~ C,D
I-8 600 54.6 2.07 6.841.6 39.5 797i5 456 546 7.1 - c~

2149~3 Illustrative Example re small angle ~cattering Three precursor materials, i) polyvinylidene fluoride (PVDF), ii) Crowley pitch (tradename), and iii) phenolic resole resin (product no. 29217 from Occidental Chemical Corp.), were pyrolyzed at 1000C and small angle x-ray scattering data was obtained on each as described above.
Figures 14a and b show plots of the intensity versus scattering angle and ln (intensity) versus q2 respectively for each sample. The resole resin sample shows significant scattering (intensity) at small angles and the data in Figure 14b is linear, suggesting that the internal pores are predominantly uniform in size. The data can thus be fit to the Guinier formula giving Rg=5.5A. The resole resin sample is similar to the B1000 sample from CA 2,146,426 which shows all the desirable electrochemical characteris-tics of the instant invention.
The pyrolyzed PVDF sample also shows significant scattering at small angles but the data in Figure 12b is non-linear, suggesting the presence of a variety of pore sizes including pores larger than those of the pyrolyzed phenolic resole resin. The H/C atomic ratio for this sample was 0.053, R was 1.23, and the amount of methylene blue absorbed was greater than 40 micromoles per gram. The reversible and irreversible lithium capacities were 380 mAh/g and 710 mAh/g respectively. This sample has an unacceptably large electrolyte accessible surface area.
The Crowley pitch (tra~en~m~) sample shows minimal small angle scattering indicating that this sample has minimal porosity. The physical and electrochemical charac-teristics of this sample are similar to those of other pyrolyzed cokes. (The H/C atomic ratio for this sample was 0.04 and R was 8.79. The reversible and irreversible lithium capacities were 340 mAh/g and 100 mAh/g respective-ly.) As will be apparent to those skilled in the art in thelight of the foregoing disclosure, many alterations and 21g98S3 modifications are possible in the practice of this inven-tion without departing from the spirit or scope thereof.
For example, mixtures of more than one precursor may be used to prepare compounds. Additionally, precursors might contain significant matter that is not a carbohydrate, as in the case of wood, shells, cotton or straw. Accordingly, the scope of the invention is to be construed in accordance with the substance defined by the following claims.

Claims (33)

WHAT IS CLAIMED IS:

1. A carbonaceous insertion compound comprising:
a pre-graphitic carbonaceous host having a rever-sible capacity for lithium insertion, an irre-versible capacity for lithium insertion, and a surface area accessible to a non-aqueous electro-lyte wherein i) the empirical parameter R, as determined by x-ray diffraction and defined as the height of the centre of the {002} peak divided by the background level, is less than about 2.2;
ii) the H/C atomic ratio is less than about 0.1; and iii) the accessible surface area is suffi-ciently small such that the irreversible capacity is less than about a half that of the reversible capacity;
and lithium atoms inserted into the carbonaceous host .

2. A carbonaceous insertion compound as claimed in claim 1 wherein the accessible surface area is sufficiently small such that the irreversible capacity is less than about a third that of the reversible capacity.

3. A carbonaceous insertion compound as claimed in claim 1 wherein the methylene blue absorption capacity of the carbonaceous host is less than about 4 micromoles per gram of host;

4. A carbonaceous insertion compound as claimed in claim 1 wherein the surface area of the carbonaceous host as determined by BET is less than about 300 m2/gram.

5. A carbonaceous insertion compound as claimed in claim 1 wherein less than about 5% by weight of the carbonaceous host is lost after pyrolyzing at about 1000°C under inert gas.

6. A carbonaceous insertion compound as claimed in claim 1 wherein the non-aqueous electrolyte comprises ethylene carbonate and diethyl carbonate.

7. A carbohydrate precursor carbonaceous insertion compound comprising:
a pre-graphitic carbonaceous host prepared by pyrolyzing a carbohydrate precursor or a carbohy-drate containing precursor at a temperature above 800°C wherein the empirical parameter R, deter-mined from an x-ray diffraction pattern and defined as the {002} peak height divided by the background level, is less than about 2.2; and alkali metal atoms inserted into the carbonaceous host.

8. A carbonaceous insertion compound as claimed in claim 7 wherein the H/C atomic ratio of the pre-graphitic carbon-aceous host is less than about 0.1.

9. A carbonaceous insertion compound as claimed in claim 7 wherein the methylene blue absorption capacity of the carbonaceous host is less than about 4 micromoles per gram of host;

10. A carbonaceous insertion compound as claimed in claim 7.wherein the surface area of the carbonaceous host as determined by BET is less than about 300 m2/gram.

11. A carbonaceous insertion compound as claimed in claim 7 wherein the tap density of the carbonaceous host is greater than about 0.7 g/ml.

12. A carbonaceous insertion compound as claimed in claim 7 wherein R is less than about 2.

13. A carbonaceous insertion compound as claimed in claim 7 wherein the alkali metal is lithium and the pre-graphitic carbonaceous host has a reversible capacity for lithium insertion, an irreversible capacity for lithium insertion, and a surface area accessible to a non-aqueous electrolyte.

14. A carbonaceous insertion compound as claimed in claim 13 wherein the accessible surface area is sufficiently small such that the irreversible capacity is less than about a half that of the reversible capacity.

15. A carbonaceous insertion compound as claimed in claim 7 wherein the carbohydrate precursor is pyrolyzed at a temperature in the range from about 900°C to about 1100°C.

16. A carbonaceous insertion compound as claimed in claim 15 wherein the pyrolysis temperature is maintained for about an hour.

17. A carbonaceous insertion compound as claimed in claim 15 wherein the pyrolysis temperature is attained by ramping at a rate of about 25°C per minute.

18. A carbonaceous insertion compound as claimed in claim 7 wherein the carbohydrate precursor is a sugar.

19. A carbonaceous insertion compound as claimed in claim 18 wherein the sugar is sucrose.

20. A carbonaceous insertion compound as claimed in claim 7 wherein the carbohydrate precursor is a starch.

21. A carbonaceous insertion compound as claimed in claim 7 wherein the carbohydrate precursor is a cellulose.

22. A carbonaceous insertion compound as claimed in claim 21 wherein the cellulose is selected from the cellulose containing group consisting of red oak, maple, walnut shell, filbert shell, almond shell, cotton and straw.

23. A process for preparing a pre-graphitic carbonaceous host for a carbonaceous insertion compound comprising pyrolyzing a carbohydrate precursor, or a carbohydrate containing precursor, at a temperature above 800°C such that the empirical parameter R, determined from an x-ray diffraction pattern and defined as the {002} peak height divided by the background level, is less than about 2.2.

24. A process as claimed in claim 23 wherein the carbohy-drate precursor is selected from the group consisting of sugar, starch, and cellulose.

25. A process as claimed in claim 23 additionally compris-ing precarbonizing the carbohydrate by washing the carbohy-drate with an acid.

26. A process as claimed in claim 25 wherein the carbohy-drate is sucrose.

27. A process as claimed in claim 25 wherein the acid is concentrated sulfuric acid.

28. An electrochemical device comprising an electrode wherein a portion of the electrode comprises the carbon-aceous insertion compound as claimed in claim 1 or 7.

29. A battery comprising an electrode wherein a portion of the electrode comprises the carbonaceous insertion compound as claimed in claim 1 or 7.

30. A non-aqueous battery comprising:

a cathode comprising a lithium insertion com-pound;
a non-aqueous battery electrolyte comprising a lithium salt dissolved in a mixture of non-aque-ous solvents; and an anode comprising the carbonaceous insertion compound as claimed in claim 1, 13, or 14.

31. The use of a carbonaceous insertion compound in an electrode of an electrochemical device, said carbonaceous insertion compound comprising:
a pre-graphitic carbonaceous host prepared by pyrolyzing a carbohydrate precursor, or a carbo-hydrate containing precursor, at a temperature above 800°C wherein the empirical parameter R, determined from an x-ray diffraction pattern and defined as the {002} peak height divided by the background level, is less than about 2.2; and atoms of an alkali metal inserted into the car-bonaceous host.

32. The use of the carbonaceous insertion compound as claimed in claim 31 wherein the carbohydrate precursor is selected from the group consisting of sugar, starch, and cellulose.

33. The use of the carbonaceous insertion compound as claimed in claim 31 wherein the alkali metal is lithium and the electrochemical device is a non-aqueous battery, the battery comprising a cathode comprising a lithium insertion compound; a non-aqueous battery electrolyte comprising a lithium salt dissolved in a mixture of non-aqueous sol-vents; and an anode comprising said carbonaceous insertion compound.