Bimodal lithium transition metal based oxide powder for use in a rechargeable battery

Abstract

A bimodal lithium transition metal oxide based powder for a rechargeable battery, comprising: a first lithium transition metal oxide based powder, either comprising a material having a layered crystal structure consisting of the elements Li, a metal M and oxygen, wherein the Li content is stoichiometrically controlled, wherein the metal M has the formula M=Co.sub.1aM.sub.a, with 0a0.05, and wherein M is either one or more metals of the group consisting of Al, Ga and B; or comprising a core material and a surface layer, the core having a layered crystal structure consisting of the elements Li, a metal M and oxygen, wherein the Li content is stoichiometrically controlled, wherein the metal M has the formula M=Co.sub.1aM.sub.a, with 0a0.05, wherein M is either one or more metals of the group consisting of Al, Ga and B; and the surface layer consisting of a mixture of the elements of the core material and inorganic N-based oxides, wherein N is either one or more metals of the group consisting of Mg, Ti, Fe, Cu, Ca, Ba, Y, Sn, Sb, Na, Zn, Zr, Si, Nb, Mo, Ru, Rh, Pd, Ag, Cd, Sc, Ce, Pr, Nd, Gd, Dy, and Er; the first powder having an average particle size (D50) of at least 15 m; and a second lithium transition metal oxide based powder having the formula Li.sub.1+bN.sub.1bO.sub.2, wherein 0.10b0.25, and NNi.sub.xMn.sub.yCo.sub.zA.sub.d, wherein 0.10x0.60, 0.30y0.80, 0.05z0.20 and 0d0.10, A being a dopant, the second powder having an average particle size (D50) of less than 5 m, and preferably less than 2 m.

Claims

1. A bimodal lithium transition metal oxide based powder for a rechargeable battery, comprising: a first lithium transition metal oxide based powder having an average particle size (D50) of at least 15 m and comprising a material A, having a layered crystal structure comprising the elements Li, a metal M and oxygen, wherein the Li content is stoichiometrically controlled, wherein the metal M has a formula M=Co.sub.1aM.sub.a, with 0a0.05, wherein M is one or more metals selected from the group consisting of Al, Ga and B, and wherein the first lithium transition metal oxide based powder has an electrical conductivity of less than 10.sup.4 S/cm; and a second lithium transition metal oxide based powder having the formula Li.sub.1+bN.sub.1bO.sub.2, wherein 0.10b0.25, and N=Ni.sub.xMn.sub.yCo.sub.zA.sub.d, wherein 0.15x0.30, 0.50y0.75, 0.05<z0.15 and 0d0.10, and wherein A is a dopant; wherein the second powder has an average particle size (D50) of less than 5 m and wherein the ratio of the D50 value of the first powder to the D50 value of the second powder is at least 5:1.

2. A bimodal lithium transition metal oxide based powder for a rechargeable battery, comprising: a first lithium transition metal oxide based powder having an average particle size (D50) of at least 15 m and comprising a material B, having a core and a surface layer, wherein the core comprises a layered crystal structure comprising the elements Li, a metal M and oxygen, wherein the Li content is stoichiometrically controlled, wherein the metal M has a formula M=Co.sub.1aM.sub.a, with 0a0.05, and wherein M is one or more metals selected from the group consisting of Al, Ga and B; wherein the surface layer comprises a mixture of the elements of the core and inorganic N-based oxides, wherein N comprises one or more metals selected from the group consisting of Mg, Ti, Fe, Cu, Ca, Ba, Y, Sn, Sb, Na, Zn, Zr, Si, Nb, Mo, Ru, Rh, Pd, Ag, Cd, Sc, Ce, Pr, Nd, Gd, Dy, and Er, and wherein the first lithium transition metal oxide based powder has an electrical conductivity of less than 10.sup.4 S/cm; and a second lithium transition metal oxide based powder having the formula Li.sub.1+bN.sub.1bO.sub.2, wherein 0.10b0.25, and N=Ni.sub.xMn.sub.yCo.sub.zA.sub.d, wherein 0.15x0.30, 0.50y0.75, 0.05z0.15 and 0d0.10, and wherein A is a dopant; wherein the second powder has an average particle size (D50) of less than 5 m.

3. The bimodal lithium transition metal oxide based powder of claim 1, wherein the first lithium transition metal oxide based powder has a surface area measured by BET of less than 0.20 m.sup.2/g, and the second lithium transition metal oxide based powder has a surface area measured by BET of at least 0.50 m.sup.2/g.

4. The bimodal lithium transition metal oxide based powder of claim 2, wherein the ratio of the weight fraction of the first lithium transition metal oxide based powder to the weight fraction of the second lithium transition metal oxide based powder is at least 2:1.

5. The bimodal lithium transition metal oxide based powder of claim 2, wherein the first lithium transition metal oxide based powder has an electrical conductivity of less than 10'S/cm, measured under a pressure of 63.7 MPa.

6. A positive electrode for a rechargeable battery, comprising the bimodal lithium transition metal oxide based powder of claim 1, and having a porosity 15%.

7. A rechargeable battery comprising the positive electrode of claim 6, wherein less than 15 vol % of organic electrolyte fills pores of the positive electrode.

8. The rechargeable battery of claim 7, wherein the total mass of electrolyte per mass of positive electrode is less than 15%.

9. The bimodal lithium transition metal oxide based powder of claim 2, wherein the first lithium transition metal oxide based powder comprises large dense particles having a D50 of at least 15 m, and wherein the ratio of the D50 value of the first lithium transition metal oxide based powder to the D50 value of the second lithium transition metal oxide based powder is at least 3:1.

10. The bimodal lithium transition metal oxide based powder of claim 2, wherein the first lithium transition metal oxide based powder has a surface area measured by BET of less than 0.20 m.sup.2/g, and the second lithium transition metal oxide based powder has a surface area measured by BET of at least 0.50 m.sup.2/g.

11. A positive electrode for a rechargeable battery, comprising the bimodal lithium transition metal oxide based powder of claim 2, and having a porosity 15%.

12. A rechargeable battery comprising the positive electrode of claim 11, wherein less than 15 vol % of organic electrolyte fills pores of the positive electrode.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1: Voltage profile (capacity of first charge as function of charge voltage) of HLM with (a) 11 mol %, (b) 5 mol % and (c) 0 mol % Co.

(2) FIGS. 2A, 2B, 2C: HLM rate and stability tests after different types of activation

(3) FIGS. 3A, 3B, 3C: electrochemical testing of high voltage stable LiCoO.sub.2

(4) FIGS. 4A, 4B, 4C: Coin cell testing of mixtures of HLM and high voltage stable LiCoO.sub.2

(5) FIG. 5: typical SEM micrograph of dense high voltage LiCoO.sub.2.

(6) FIG. 6: Summary of coin cell testing of dense high voltage LiCoO.sub.2.

(7) FIG. 7: Summary of coin cell testing (cycle stability) of 50 m LCO.

(8) FIG. 8: Result of DSC measurement of charged electrodes (a) without and (b) with 1.16 mg of electrolyte added

(9) FIG. 9: Micrograph of the LiCoO.sub.2 having a D50 of 100 micrometer

(10) FIG. 10: Micrograph of the HLM cathode having a PSD D50 of 8 micrometer.

(11) FIG. 11: Voltage profile of the sample M05 (left) and sample M20 (right) cycles at C/20 rate (cycle 1) and C/10 rate (cycle 2) between 4.6 and 2.0V at 25 C.

(12) FIG. 12: DSC heat profiles of the (1) electrolyte free (2) 9% electrolyte containing and (3) 46% electrolyte containing electrodes.

DETAILED DESCRIPTION

(13) It was mentioned before that there is a need for Li batteries with very high energy density, that can be achieved by cathodes having preferably both of a high volumetric density and a high specific reversible discharge capacity. Also, as the cathode's charge efficiency should match well with the charge efficiency of typical anode materials, a mixture of LCO and HLM should be an ideal material. In one embodiment, a way of mixing to achieve especially high densities is the bimodal approach, where two powders are mixed. One powder consists of relatively large, dense particles, whilst the other powder has small particles. These small particles may be nano-structured. The small particles can easily be pressed into the voids between the large particles. Obviously this approach will reduce the porosity of the electrode and allows to implement nano structured cathode materials without loosing density. Since the low porosity reduces the transport rate in the liquid electrolyte, it is compensated by a high transport rate in the solid material.

(14) Moreover, the bimodal mixture can be applied practically to lithium batteries if the operating voltages of LCO and HLM are adjusted to each other. HLM requires an activation charge of typically 4.8V. This voltage is generally conceived to be too high, as it was assumed that neither LiNiO.sub.2-based, nor LiCoO.sub.2-based materials can cycle stably between 2V and 4.5-4.6V. In the present invention, a technique is disclosed to lower the voltage of the activation cycle of HLM and/or to increase the operating voltage of the LCO.

(15) In one embodiment, the invention makes use of the advantages of the high voltage stable LiCoO.sub.2 that is disclosed in co-pending application WO2012-171780, describing a lithium metal oxide powder for use as a cathode material in a rechargeable battery, consisting of a core material and a surface layer, the core having a layered crystal structure consisting of the elements Li, a metal M and oxygen, wherein the Li content is stoichiometrically controlled, wherein the metal M has the formula M=Co.sub.1a M.sub.a, with 0a0.05, wherein M is either one or more metals of the group consisting of Al, Ga and B; and the surface layer consisting of a mixture of the elements of the core material and inorganic N-based oxides, wherein N is either one or more metals of the group consisting of Mg, Ti, Fe, Cu, Ca, Ba, Y, Sn, Sb, Na, Zn, Zr and Si. There could also be more elements doped in the surface layer, such as 4d elements and Rare-earth metals, like Er, Nd and Nb. This material has a reversible electrode capacity of at least 200 mAh/g, preferably 210 mAh/g and most preferably 215 mAh/g when used as an active component in a cathode which is cycled between 3.0 and 4.6 V vs. Li.sup.+/Li at a discharge rate of 0.1 C, and preferably at 1 C, at 25 C. It also has a 1 C rate capacity fading value below 60%, preferably below 40% and most preferably below 30%. The capacity fade rateexpressed in %at 1 C per 100 cycles is calculated as follows: (1(DQ32/DQ8))100/23. (DQn=discharge capacity of cycle n).

(16) The cathode material of the present invention has the following properties: (1) Electrodes having a low porosity allowing for (a) the use of less liquid electrolyte providing better safety (b) a high content of active material providing high volumetric and gravimetric capacity and energy density. To achieve low porosity electrodes the cathode may consist of particles which have a bimodal size distribution. In this case the small particles will fit well into the voids between the large particles. (2) Electrodes having a high reversible capacity. A high reversible capacity allows to decrease the electrode thickness at constant capacity, so the total porosity decreases. It is the total porosity (total volume of pores in the electrode) which matters for safety and not the porous fraction (pore volume/electrode volume).

(17) 3) The cathode material allows for very high ionic transport rate in the solid phase. This allows achieving a good rate performance even in the case of low porosity electrodes because solid lithium diffusion can act as a shortcut between the narrow liquid diffusion paths within the electrolyte filled pores. For an efficient solid shortcut, if the cathode powder has a bimodal particle size distribution at least the large particle fraction has a high solid Li diffusion rate. (4) The cathode material has a charge efficiency which matches well the charge efficiency of typical anode materials. (5) The cathode material has a high gravimetric and volumetric capacity and energy density. It performs well (rate performance, cycling stability, . . . ) within the operating window of the lithium battery.

(18) Regarding the safety of the battery, a schematic example for a CCE reaction goes as followsfor LiCoO.sub.2 as cathode material: the battery is charged, meaning that Li is extracted from the cathode by the reaction:

(19) $\begin{array}{cc}{\mathrm{LiCoO}}_2\underset{-x\left({\mathrm{Li}}^{+}-e^{-}\right)}{.fwdarw.}{\mathrm{Li}}_{1-x}{\mathrm{CoO}}_2& \left(1\right)\end{array}$

(20) Charging in this case needs high energy, since a potential of 4V is needed to extract the lithium. Therefore the charged cathode is thermodynamically highly unstable. The equilibrium phases of charged LCO (in air) are LiCoO.sub.2+Co.sub.3O.sub.4. The following reaction is thermodynamically strongly favored:
Li.sub.1xCoO.sub.2.fwdarw.(1x)LiCoO.sub.2+x/3Co.sub.3O.sub.4+x/3O.sub.2(2)

(21) Thermodynamic estimations show that this reaction has a very high G (free enthalpy change) but only a small H (enthalpy change). As a result not much heat evolves.

(22) However, the situation is different in a real battery, where organic electrolyte is present. The evolved oxygen will combust a part of the electrolyte creating large amounts of heat. The electrolyte will also reduce Co.sub.3O.sub.4 to CoO. As an example of organic electrolyte we use ethylene carbonate, and assume, for simplicity, full combustion:
5/2O.sub.2+C.sub.3H.sub.4O.sub.3.fwdarw.3CO.sub.2+2H.sub.2O(3)

(23) Schematically the CCE reaction in a real battery can be written as

(24) $\begin{array}{cc}{\mathrm{Li}}_{1-x}{\mathrm{CoO}}_2+\frac{x}{5}{C}_3{H}_4{O}_3.fwdarw.\frac{3x}{5}{\mathrm{Co}}_2+\frac{2x}{5}{H}_{2}O+\left(1-x\right){\mathrm{LiCoO}}_2+x\mathrm{CoO}& \left(4\right)\end{array}$

(25) This equation tells us about the limitations of the CCE reaction: 1) Decrease x: less de-lithiated cathodes create less heat because less oxygen is evolved to combust the electrolyte. Decreasing x or the amount of LiCoO.sub.2 is not meaningful because this will reduce the capacity of the battery. 2) Decrease the electrolyte: if it would be possible to make a Li ion battery where the electrolyte content is lower than x/5 (x state of charge in mol Li) per mol of LiCoO.sub.2 then safety will improve because not enough electrolyte is present to complete the CCE reaction.

(26) If the lithium transition metal oxide contains Mn or Ni instead of Co, this will not so much change the evolved heat of the combustion of the electrolyte, but can change the kinetics dramatically, because delithiated Li.sub.0.5CoO.sub.2 (with Co in 4 valent state) is much more reactive than delithiated Li.sub.0.5Ni.sub.0.5Mn.sub.0.5O.sub.2 (where the 4 valent Ni is diluted by much more inert Mn.sup.4+). Much less reactive means that it is more difficult to start the CCE reaction (the battery needs a larger external heat or damage). Additionally if Mn is present then eventually less electrolyte is combusted because some 3 valent Mn (e.g. in Mn.sub.3O.sub.4) remains, so less oxygen is delivered (and less electrolyte is combusted). Within this reaction we willin our examplesfocus on LiCoO.sub.2, but the same conclusions are valid for other cathode materials as well.

(27) A very important aspect of cell safety is to lower the amount of exothermic heat generated if a cell gets unsafe. At relatively low temperature (about 200-300 C.) the charged cathode starts to react with the electrolyte. The charged cathode is in an oxidized state, so it is able to release oxygen, which combusts the electrolyte. The combustion reaction contributes a large amount of the heat when a battery gets unsafe. The higher the extracted capacity, the more the cathode is oxidized, and the more oxygen is supplied to combust the electrolyte. Obviously, if less electrolyte is present than needed to consume all oxygen then less combustion will happen and the safety of the battery improves. Reducing the electrolyte below a certain critical amount (where amount of oxygen releasable from the cathode and amount of electrolyte are in balance to allow full combustion) will improve safety. The critical amount increases with increasing extracted capacity in what is called the electrolyte cathode balance.

(28) The pores in the electrodes must be filled with electrolyte otherwise the battery will have a poor performance. From a safety point of view it is desired to achieve a small porosity as much as possible, below the critical porosity which corresponds to the critical electrolyte amount, defined by the electrolyte cathode balance. So it is desired to decrease the porosity of the electrodes (and separator). However, there is a limit because state of the art batteries need electrolyte to facilitate the fast lithium diffusion in the liquid phase. In this sense, the current invention discloses that it is possible to achieve a well functioning battery with very low porosity, much below the critical porosity.

(29) To summarize: one aspect of the present invention is to provide a high density electrode which has a low porosity much below of the critical porosity. Another aspect of the invention is to increase the amount of x (extracted Lithium at full charge per g of cathode). By utilizing these aspects the safety of the Li rechargeable battery is improved because not enough electrolyte is present to allow for a completed CCE reaction.

(30) The bimodal approach of embodiments of the present invention yields high capacity and low porosity electrodes. HLM has a low electrode density, low ionic transport and also very low electronic conductivity. By amending these parameters by mixing HLM with a different large particle powder, a dense cathode powder is obtained. The LiCoO.sub.2 or LiNiO.sub.2 based materials consist of relatively large (>10 m) particles, and the nano-structured HLM occupies the voids, resulting in a high electrode density.

(31) In dynamic cell balancing the irreversible capacity of both anode and cathode are to be compared. The irreversible capacity is the 1.sup.st charge inefficiency Qirr={Q(Charge)Q(Discharge)} divided by Q(Charge). The highest cell capacity is achieved if the irreversible capacities are balanced. If the anode shows a much larger irreversible capacity than the cathode then a lower cell capacity is achieved because not all of the cathode capacity is utilized. On the one hand, the high voltage LiCoO.sub.2 based cathodes of the present invention have very low irreversible capacity. Even at relatively high rate the value is less than 5% which is much less than the value of typical anodes. On the other hand, the HLM type cathode materials have a larger irreversible capacity, which dramatically increases with the discharge rate, and at fast rate the irreversible capacity by far exceeds the value for typical anodes. It has been found that a mixture of high voltage stable LiCoO.sub.2 and HLM can allow improving the cell balancing as well as dynamic cell balancing, thus the capacity of the cell can be increased. The electrode contains large, dense LiCoO.sub.2 based particles. The LiCoO.sub.2 used in the invention has high bulk lithium transport rate and can be cycled in a stable manner at high voltage 4.35V) in full cells. The high voltage stability allows the addition of low cobalt high voltagehigh capacity cathode materialsi.e. the Co doped HLM materialsto the cathode.

(32) The LiCoO.sub.2 forms an ideal framework to add the HLM high voltagehigh capacity but low power cathode materials. As a result electrode density increases (yielding higher capacity) and porosity decreases (yielding a better safety). The LiCoO.sub.2 particles create a fast path for Li diffusion across the electrode, partially replacing the liquid electrolyte. The submicron sized cathode material does itself not allow to obtain mechanically stable relatively thick electrodes without using large amounts of binder. However, the LiCoO.sub.2 based framework supplies stability. By mixing with the large particle dense LiCoO.sub.2 mechanically stable electrodes can be obtained. The high cathode voltage allows extracting more lithium per cathode volume. Thisrelatively to the capacity of the full celldecreases the amount of electrolyte present per unit of capacity. Thus a full combustion of all electrolyte delivers less heat per unit capacity. As a result the safety per unit capacity is improved.

(33) In embodiments of the present invention, cathode materials are characterized by

(34) on the one hand: (1) large dense particles (a) to achieve a high electrode density with low porosity, and (b) to enable a short solid diffusion path across the electrode, (2) excellent high voltage stability (so that it can be cycled in the same voltage window as HLM), and (3) high ionic transport rate.

(35) And on the other hand: agglomerates consisting of primary particles that consist of nano-structured or submicron-sized high capacity cathode material such as HLM, to achieve an electrode with very high electrode density in mixtures that can be used in a high voltage battery.

(36) The invention is further illustrated in the following examples:

Example 1: Importance of Cobalt Doping of HLM

(37) A major problem to cycle a mixture of large particles LCO and smaller particles HLM is the high voltage which is required during first charge of HLM. A typical value is 4.8V, but with each 10 mV that this voltage can be lowered the undesired side effectse.g. the high voltage electrolyte decompositionwill become less severe. The current example shows that the voltage of the HLM can be lowered. Increasing the Co doping level causes a desired decrease of the charge voltage. Otherwise, too much Co doping is not desired because it reduces the reversible capacity.

(38) 3 MOOH precursors are prepared by precipitating a M-NO.sub.3 solution with a NaOH solution in the presence of ammonia. The cobalt content of the solutions is varied from 0 to 8.3 to 16.7 wt %. Table 1.1 lists the ICP analysis of the 3 precursors. The composition is very near to the targeted value. Of each precursor 3 samples with different Li:M ratio are prepared. Each precursor is mixed with Li.sub.2CO.sub.3 and cooked at 800 C. in air for 8 hours. The target Li:M value is calculated as follows: assuming a composition of the final sample Li.sub.1+xM.sub.1xO.sub.2 where all Mn is tetravalent, all Co is 3-valent and Ni can be 3- or 2-valent. The target Li:M ratio is given by Li:M=(1+x)/(1x) where x is calculated to give

(39) (a) all Ni is 2-valent

(40) (b) average Ni valence state is 2.5 and

(41) (c) all Ni is 3-valent.

(42) The first charge voltage profile is obtained by coin cell testing. The charge rate is 80 mA/g upto 4.8V. In the following the best results (highest capacity of each transition metal composition) are listed. FIG. 1 illustrates the voltage profile (capacity in mAh/g of first charge as function of charge voltage V) of HLM with (A) 11 mol %, (B) 5 mol % and (C) 0 mol % Co. With increasing Co doping we observe a lowering of the first cycle charge voltage in the plateau region. The decrease is approx. 60 mV per mol % of Co doping. The example demonstrates that HLM, when mixed with large LCO based material, is preferably doped with cobalt. Preferred doping range is 5-20 mol %. The cobalt doping reduces the necessary charge voltage for FILM, so the LCO and HLM charge voltage profiles matches better.

(43) TABLE-US-00001 TABLE 1.1 Composition of Co containing HLM precursors Ni Co Mn Sample ID wt % wt % wt % Metal composition PHLM-011 10.87 11.04 40.36 Mn.sub.0.664 Ni.sub.0.169Co.sub.0.167 PHLM-010 12.75 4.96 40.45 Mn.sub.0.710Ni.sub.0.209Co.sub.0.081 PHLM-012 15.61 0.092 43.48 Mn.sub.0.747Ni.sub.0.251Co.sub.0.001

Example 2: Activation of Cobalt Doped HLM

(44) A mixed transition metal hydroxide is obtained by conventional precipitation (metal sulfate with NaOH in the presence of ammonia). The metal composition is M=Ni.sub.0.22Mn.sub.0.66Co.sub.0.11. The average particle size (D50 of the PSD) is about 8 m. The precursor is relatively dense (tap density=1.36 g/cm.sup.3). A final lithium metal oxide (HLM type) is prepared by mixing with Li.sub.2CO.sub.3 (Li:M=1.57, corresponding to 50% of the Ni being 2-valent and 50% being 3-valent, assuming all Co is 3-valent and all Mn tetravalent). A high sintering temperature of 1000 C. is applied resulting in relative dense particles with relatively small surface area (0.65 m.sup.2/g). SEM shows that primary crystallites are ranging from about 0.2 to 0.5 m in size. Typical secondary or agglomerated particles have sizes of about 5-10 m.

(45) Such a morphology is not preferred for obtaining a high electrochemical performance for HLM. Good performance is achieved after much softer sintering resulting in significant significantly smaller crystallites. Typically also better performance is achieved in the case of smaller particle sized precursors. However, for experimental purposes, after these strong sintering conditions particular electrochemical performance issues of HLM are clearly detected. After strong sintering the activation typically requires high voltage and several cycles. We speculate that the activation is related to an electrochemical grinding. Large particles have severe rate performance issues, so a large increase of capacity as a result of electrochemical grinding is observed. In the case of cobalt doping however, activation is no problem, as will be shown in this Example.

(46) Coin cells are prepared by known standard methods. Coin cell testing involves 3 different types of activation (different voltages, and 1 or 10 cycles), followed by a test for rate performance and cycle stability. The cycling schedule details are given in Table 2.1 below. Independently of the type of activation, the tests for rate performance and cycling stability give identical results. Very tiny differences even indicate that lower voltage and less activation cycles gives an advantage!

(47) We conclude that in the case of Co doped HLM:

(48) 1) 1 activation cycle is sufficient, and

(49) 2) 4.6V is sufficient as activation voltage.

(50) Details of the coin cell testing are summarized in FIGS. 2A-C, illustrating:

(51) (A): Rate and stability after single cycle activation at 4.6V (N=1, V1=4.6V)

(52) (B): Rate and stability after single cycle activation at 4.8V (N=1, V1=4.8V)

(53) (C): Rate and stability after 10 cycles activation at 4.6V (N=10, V1=4.6V).

(54) For each of FIGS. 2(A), 2(B) and 2(C) the left figure shows cycles 2-7 (from right to left), the right figure cycles 8, 9, 32 & 33 (from right to left).

(55) A similar experiment is performed for a Co doped HLM prepared at lower temperature from a fine particle precursor. The surface area is much larger, about 5 m.sup.2/g. It is confirmed that also for such HLM cathode performance after a single activation at 4.6V gives the best performance.

(56) TABLE-US-00002 TABLE 2.1 Cycle schedule definition Charge Discharge Voltage, Voltage, Cycle current Type current Type Purpose 1 to N V1, 0.1 C CC 2.0 V, 0.1 C CC Activation of HLM 1 + N 4.6 V, CC/CV 2.0 V, CC Rate to 0.25 C/0.05 C 0.2, 0.5, 1, performance 6 + N 2, 3 C 7 + N, 4.6 V, CC/CV 2.0 V, 0.1 C CC Slow reference 31 + N 0.25 C/0.1 C cycle before/after stability test 8 + N, 4.6 V, CC/CV 2.0 V, 1 C CC Fast reference 32 + N 0.25 C/0.1 C cycle before/after stability test 9 + N 4.6 V, CC 2.0 V, 0.5 C CC Stability test to 0.25 C 30 + N (CC = constant current, CC/CV = constant current/constant voltage)

Example 3: Voltage Compatibility of LCO and HLM

(57) HLM type materials are considered to require special formation cycles. A typical formation schedule is an activation either at 4.8V for one or more cycles. 4.8V is a very high voltage; when HLM is mixed with other cathode materials a 4.8V cycle can damage other cathode components. Alternatively it is often recommended to gradually increase the voltage during several formation cycles. Implementing such complex special formation at mass production requires severe investment and makes the application of HLM difficult in real cells. Example 2 showed that for HLM containing cobalt a single activation cycle at 4.6V is sufficient. Cycling HLM at 4.5V gives a high capacity. (At 4.5V the capacity is only about 5% less than 4.6V capacity.) Example 3 will show that these formation conditions of Example 2 are compatible with the high voltage LiCoO.sub.2 discussed before. This is an important requirement if HLM and LCO cathode materials are used together. Example 3 proves that a mixture of HLM with high voltage LiCoO.sub.2 can cycle well in real batteries.

(58) A high voltage stable LiCoO.sub.2 based cathode is obtained from a pilot plant, according to co-pending application WO2012-171780. The Li:Co ratio is 1.0 and the electrical conductivity of the LCO powder is below 10.sup.4 S/cm (1.2*10.sup.5 S/cm). The conductivity is measured under a pressure of 63.7 MPa at room temperature. The LCO contains Mg (1 wt %). The majority of particles have a large size of 20 m. The average particle size (D50 of the PSD) is 18 m. The surface area of the LiCoO.sub.2 based cathode is 0.18 m.sup.2/g.

(59) Coin cell testing is performed according to different schedules named V3.1-V3.2. V3.1 is the charge voltage of cycle 1 whereas V3.2=4.5V, being the charge voltage of cycles 2-32. Table 3.1 summarizes the schedules, and Table 3.2 summarizes the obtained results. The table shows the capacity loss and energy loss (capacityaverage voltage) per 100 cycles (in percent) extrapolated from cycle 7 and 31 (for 1 C rate) and cycle 8 and 32 (for 0.1 C rate) The data in Table 3.2 show that V3.1=4.3V gives an excellent cycle stability. A similar stability is obtained with V3.1=4.6V. V3.1=4.7V still shows acceptable cycle stability (but less than V3.1=4.6V) whereas V3.1=4.8V shows some deterioration.

(60) From Table 3.2 we can conclude that high voltage LiCoO.sub.2 used in the present invention is compatible with HLM. High voltage stable LiCoO.sub.2 tolerates an activation cycle at 4.6V, and allows for cycling without significant capacity fade at 4.5V. Example 2 demonstrated similar electrochemical testing for HLM. Example 3 demonstrates that the electrochemical properties of high voltage stable LiCoO.sub.2 is compatible with HLM type voltage range, particularly if the HLM contains Co allowing to reduce the voltage V1 of a single activation cycle to 4.6V.

(61) TABLE-US-00003 TABLE 3.1 Cycling schedule V 3.1-V 3.2 (1 C = 160 mA/g) Charge Discharge Voltage, Voltage, Cycle current Type current Type Purpose 1 V3.1, 0.1 C CC 2.0 V, 0.1 C CC Activation of HLM 2-6 V3.2, CC/ 2.0 V, CC Rate performance 0.25 C/0.05 C CV 0.2, 0.5, 1, 2, 3 C 7, 31 V3.2, CC/ 2.0 V, 0.1 C CC Slow reference cycle 0.25 C/0.1 C CV before/after stability test 8, 32 V3.2, CC/ 2.0 V, 1 C CC Fast reference cycle 0.25 C/0.1 C CV before/after stability test 10- V3.2, 0.25 C CC 2.0 V, 0.5 C CC Stability test 31

(62) TABLE-US-00004 TABLE 3.2 Coin cell testing results for high voltage stable LiCoO.sub.2 Q Q.sub.4.5V Q.sub.4.5V Voltage DQ Qirr (0.1 C) Q (1 C) E (0.1 C) E (1 C) (0.1 C) (1 C) V3.2 = 4.5 V mAh/g % %/100 %/100 %/100 %/100 mAh/g mAh/g V3.1 = 4.3 V 161.0 1.1 3.1 6.9 3.5 9.0 194.5 189.6 V3.1 = 4.6 V 228.8 1.3 3.4 7.6 3.9 10.0 194.1 188.4 V3.1 = 4.7 V 250.9 1.6 4.7 11.6 5.8 14.7 193.2 187.2 V3.1 = 4.8 V 255.9 1.7 5.8 15.6 7.2 19.5 193.0 187.0

(63) The following definitions are used for data analysis: (Q: capacity, DQ: Discharge Capacity, CQ: Charge Capacity)

(64) The discharge capacity QD is measured during the first cycle in the 4.3-3.0 V range at 0.1 C (in mAh/g). Irreversible capacity Qirr is (QC1QD1)/QC1 (in %).

(65) Fade rate (0.1 C) per 100 cycles, for capacity Q (0.1 C): (1QD31/QD7)*100/23.

(66) Fade rate (1 C) per 100 cycles, for capacity Q (1 C): (1QD32/QD8)*100/23.

(67) Energy fade E (0.1 C)& E (1 C): instead of discharge capacity QD the discharge energy (capacityaverage discharge voltage) is used.

(68) Q.sub.4.5V (0.1 C) and Q.sub.4.5V (1 C): discharge capacity of cycle 7 (at 0.1 C) and of cycle 8 (at 1 C)

(69) FIGS. 3A-C illustrates the electrochemical testing of high voltage stable LiCoO.sub.2 for:

(70) (A) first cycle 4.3V

(71) (B) first cycle 4.6V

(72) (C) first cycle 4.8V.

(73) For each of FIGS. 3(A), 3(B) and 3(C) the left figure shows cycles 1-6 (from right to left), the middle cycles 7, 8, 31 & 32 (from right to left). The right figures show the fade rate: capacity in mAh/g against cycle number (charge: small circles, discharge: bigger circles).

Example 4: Voltage Compatibility of LCO and HLM, and Compatibility of HLM-LCO Mixtures with Anodes

(74) A mixed transition metal hydroxide is obtained by conventional precipitation (metal sulfate with NaOH in the presence of ammonia). The metal composition is M=Ni.sub.0.22Mn.sub.0.66Co.sub.0.11. The average particle size (D50 of the PSD) is about 3-4 m, the precursor is consisting of relatively loose agglomerated sub-micrometer particles. The precursor has a low tap density of about 0.6 g/cm.sup.3. Final HLM type cathode material is prepared by mixing the precursor with Li.sub.2CO.sub.3 (Li: transition metal blend ratio=1.442, at this ratio it is assumed that Ni is divalent) and firing at 800 C. for 10 hours. The chemical formula of the final product is estimated as Li.sub.1.181Ni.sub.0.182Mn.sub.0.546Co.sub.0.091O.sub.2. The surface area of the final sample is 4.5 m.sup.2/g. The morphology is fluffy, meaning that the powder consists of loose agglomerates of sub-micrometer particles. These submicrometer particles have a size about 100 nm.

(75) A high voltage stable LiCoO.sub.2 based cathode is obtained from our pilot plant (made according to the process in WO2012-171780). The Li:Co ratio is 1.00 and the electrical conductivity of the LCO powder is below 10.sup.4 S/cm.sup.2. The LCO contains Mg (1 wt %). The majority of particles have a large size of 25 m. The average particle size (D50 of the PSD) is 22 m. The surface area of the LiCoO.sub.2 based cathode is below 0.15 m.sup.2/g.

(76) The LiCoO.sub.2 and the HLM powder are mixed using different weight ratios:

(77) Sample LCO:HLM

(78) S4A: 90:10;

(79) S4B: 75:25;

(80) S4C: 50:50.

(81) Coin cells are prepared and tested by the schedule used in Example 2(A) for testing HLM, here with is 1 activation cycle (N=1) at V1=4.6V followed by rate test and stability test at V2=4.5V (instead of 4.6V in Table 2.1). Table 4.1 summarizes the obtained results.

(82) TABLE-US-00005 TABLE 4.1 Coin cell testing of mixtures of HLM and high voltage stable LCO. Coin Cell Schedule: V1 = 4.6V, V2 = 4.5V Fading rate (per 100 cyc) DC Q Qirr DC Q 1 C Rate 3 C Rate Capacity Energy Sample mAh/g % 0.1 C %/0.1 C %/0.1 C 0.1 C 1 C 0.1 C 1 C S4A 234.6 3.8 196.8 95.6 90.6 6.5 12.3 7.5 14.1 S4B 232.5 5.9 194.8 93.8 88.3 10.1 16.1 12.1 19.7 S4C 239.3 7.6 201.2 90.5 84.2 14.5 19.0 17.4 25.0

(83) FIGS. 4A, 4B, and 4C summarize the coin cell testing details. (A=S4A, . . . ). For each of FIGS. 4(A), 4(B) and 4(C) the left figure shows cycles 2-7 (from right to left), the middle cycles 8, 9, 32 & 33 (from right to left). The right figures show the fade rate: capacity in mAh/g against cycle number (charge: small circles, discharge: bigger circles).

(84) Usually, the properties of 100% LCO do not match real anodes well. 100% LiCoO.sub.2 has a too high charge efficiency (near to 98.5%) and a high rate performance at 3 C rate (91.5%).

(85) At the same time 100% HLM, compared with typical anodes, has a low charge efficiency (<90%) and a low rate performance. Therefore also HLM does not match well with a typical anode. In the invention, mixtures match well with real anodes, and a sufficient high cycling stability is achieved. The region between Sample S4A and S4B corresponding to 10-25% HLM shows particular promising properties since an optimum highest density coincides with higher capacity, slightly lower rate and increased irreversible capacity when comparing with 100% LiCoO.sub.2. The Example illustrates that the voltage window of (cobalt doped) HLM and LiCoO.sub.2 are compatible and that a mixture of LiCoO.sub.2 and HLM matches much better to real anodes in terms of charge efficiency and rate performance than either HLM or LiCoO.sub.2. At 4.5V about or more than 195 mAh/g capacity can be achieved.

Example 5: LiCoO2 Based Electrodes with Little or no Liquid Electrolyte

(86) A high voltage stable LiCoO.sub.2 is prepared by a double firing, as in WO2012-171780. The D50 of the particle size distribution is 50 m which is consistent with the result of a PSD measurement. The particles are dense, with a pressed density of at least 4 g/cm.sup.3 and a BET value 0.2 m.sup.2/g. FIG. 5 shows a typical SEM micrograph of the LiCoO.sub.2. Left: 1000 magnification. Right: relatively small particles located on surface of a larger 40 m particle in larger magnification (5000). Besides of a very small amount of fine particles the large particles are 100% dense with no open porosity.

(87) Coin cells are prepared, wherein the electrode consists of 96 wt % of active material. The electrode loading is about 6 mg/cm.sup.2. Table 5.1 lists the coin cell test results. FIG. 6 shows the electrochemical testing results: rate performance in thin electrode configuration: left Figure: from right to left Cycles 1 to 6, with corresponding rate: C/10, 1C, 5C, 10C, 15C, 20C; right figure: rate % versus 0.1C against C rate (in hour). A typical discharge rate for lithium batteries is a 1C rate. The high density, large particle LiCoO.sub.2 demonstrate 92% at 1C rate, which is sufficient for practical applications. Alternatively, cells with 12 mg/cm.sup.2 load were cycled with V1=4.3, V2=4.5V schedule described in Example 3 and showed very high cycle stability. Results are summarized in Table 5.2 and FIG. 7.

(88) TABLE-US-00006 TABLE 5.1 Results of coin cell testing (4.4-3.0 V, 1 C = 160 mA/g) Cy 1(Charge) Cy 1 (Disch) Cy 2 (DisCh) Cy 3 (DisCh) Cy 4 (DisCh) 0.1 C, mAh/g 0.1 C, mAh/g 1.0 C, % 5.0 C, % 10 C, % 180.7 175.7 91.8 82.4 76.2 The % value of the rate performance is obtained by dividing the discharge capacity of the cycle by the discharge capacity at a rate of 0.1 C.

(89) TABLE-US-00007 TABLE 5.2 Results of coin cell testing (see schedule of Ex. 3, with V3.1 = 4.3 V, V3.2 = 4.5 V, 1 C = 160 mA/g) rate Fade Fade E Fade E Fade Q(C/10) Qirr (1 C) (0.1 C) (1 C) (0.1 C) (1 C) mAh/g % % %/100 %/100 %/100 %/100 158.6 4.17 90.5 8.2 8.4 8.2 11.4

(90) FIG. 7 gives a summary of the coin cell testing results (cycle stability) of 50 m LCO particles. The left figure shows cycles 1-6 (from right to left), the middle cycles 7, 8, 31 & 32 (from right to left), following the schedule used in Example3, with V3.1=4.3V, V3.2=4.5V. The right figures show the fade rate: capacity in mAh/g against cycle number (charge: small circles, discharge: bigger circles).

(91) After the test the cell is disassembled and the particles in the electrode are analyzed by FESEM and XRD. FESEM shows that the particles remain dense, they do not shatter into pieces. XRD shows similar narrow peaks as before testing, proving that the crystal structure does not deteriorate. In general, as the particle size of dense compact particles increases, the rate performance of these materials decreases. This is because larger particles have a longer Li diffusion path length. Comparing Li diffusion path lengths is a useful tool to study the expected rate performance for different shaped particles or objects. A possible definition of the path length (diffusion length) is the average shortest distance of each Li atom in the particle to the surface. We can estimate the average solid diffusion path length of the large PSD LCO. If we assume that all particles are spheres of 50 m diameter then the path length is the average of the distance from the surface R-r multiplied with the volume fraction. At position r the volume traction is 4r.sup.2/Vol where Vol=4/3R.sup.3. Integrating over the sphere all distances multiplied by the volume fraction gives

(92) $\frac{3}{R^3}{}_0^{R=25\mathrm{.Math.m}}r\left(R-r\right)r^2=\left(\frac{R}{4}\right)=6.25\mathrm{.Math.m}$

(93) A flat, dense (plate type) electrode of 12.5 m thickness has the same average diffusion length, so it should exhibit roughly a similar rate performance. With a density of 5.05 g/cm.sup.3 for LiCoO.sub.2 the cathode loading of the plate type electrode is 6.3 mg. Hence we estimate the rate performance of a liquid free electrode plate as >90% at 1C rate for 6.3 mg/cm.sup.2 loading. 6.3 mg/cm.sup.2 loading is less than the 15 mg/cm.sup.2 loading for typical state of art lithium batteries. Otherwise, the powder requirement for a solid cathode can be less. We can use the diffusion length L defined as L.sup.2=2d Dt by using the relation

(94) $t\frac{L^2}{D}$
where L is the diffusion length and D the diffusion constant to estimate the change of rate performance for different thick electrodes if we know the rate for a given thickness. For a 20 mg/cm.sup.2 dense electrode plate using the 10C rate performance of Table 5.1 we obtain a 1C rate performance of about 75%. In theory even thicker plate electrodes could be utilized because the solid diffusion does not show the rate limiting effects (electrolyte shut-down) known for binary electrolytes.

(95) To summarize: The solid diffusion constant of high voltage LiCoO.sub.2 is sufficiently high to allow for liquid electrolyte free electrodes. In mixed electrodes (for example HLM+LiCoO.sub.2) the large LiCoO.sub.2 particles can act as solid diffusion short-cuts between 2 regions with HLM particles, thus dramatically reducing the required amount of liquid electrolyte. Such electrodes will have extremely good safety properties. The example clearly demonstrates that electrodes with very low or even zero content of liquid electrolyte can have sufficient power.

Example 6: Low Porosity Electrode

(96) The example will demonstrate low porosity electrodes. The large particle LiCoO.sub.2 (50 m size) of Example 5 and the HLM of Example 4 are used. Slurries for coating are prepared from mixtures of LCO and HLM. As conductive additive super P is used, as binder and solvent a 5% PVDF in NMP solution is used. Table 6.1 summarizes the slurry composition. A part of the slurries are coated on Al foil as a thick film, dried, pealed off, and grinded to obtain electrode powder. The electrode powder density is measured on pressed pellets. Generally the electrode powder density is a very good approximation for a real electrode density which can be obtained with the powder. The diameter of the pressed pellets is 1.311 cm and the mass is 3.0 g, the applied force is 207 MPa. The obtained pellet density values are 4.153 g/cm.sup.3 (5% FILM); 4.036 g/cm.sup.3 (10% HLM) and 4.123 g/cm.sup.3 (15% HLM). Use of the true density of LiCoO.sub.2 (5.05 g/cm.sup.3), HLM (4.251 g/cm.sup.3), PVDF (1.7 g/cm.sup.3) and carbon (2 g/cm.sup.3) allows to calculate the porosity. The result is 11.3%; 11.6% and 7.6%. These data show that electrodes with very low porosities can be achieved. The porosities are much less than the critical porosities calculated below in Example 8. Typical electrode porosities in commercial cells are >12%, often 15-20%.

(97) The remaining slurry is used to coat electrodes of high loading followed by drying and compacting. The active cathode load is 58, 50 and 45 mg/cm.sup.2. These loads are very high, 3-4 times larger than normal electrodes in commercial cells. Coin cell testing shows full capacity within 1-2% of expected value when tested at C/20 rate. The reference value is obtained for 3.0-4.3V with 12 mg/cm.sup.2 electrodes at C/10 rate, 1C=160 mA/g and is similar to the theoretically expected value (average of LCO and HLM capacity). During further cycling of the cells with thick electrodes the electrodes pealed peeled off, known current coating technology is not optimized to obtain stable cycling with very high load. However the Example demonstrates that very thick electrodes with low porosity, promising excellent safety can be reversible cycled.

(98) TABLE-US-00008 TABLE 6.1 Electrode making using mixtures of fluffy HLM and 50 m LCO HLM PVDF Carbon HLM LCO Solid active % sol g g g g % % 5 12.377 0.619 1.5 28.5 72.7 96.04 10 15.231 0.762 3 27 68.5 95.17 15 15.115 0.756 3.75 21.25 64.9 94.30

Example 7: High Safety if Less Electrolyte Present

(99) This example demonstrates the absence of a major exothermic reaction of delithiated cathodes when no electrolyte is present. For this experiment commercial LiMO.sub.2 with M=Ni.sub.0.5Mn.sub.0.3Co.sub.0.2 is used as model material. Coin cells (with small positive electrodes (cathode) and Li metal anode) are prepared and charged to 4.3V. Cathode active material weight is about 3 mg. The coin cells are opened, thereby paying attention not to short circuit the cell. The electrode is washed in DMC, then dried at 120 C. for 10 min in air.

(100) The dried electrode is inserted in a stainless steel DSC cell and a defined amount of electrolyte is added. In this Example, 1.1 mg electrolyte is added in one experiment, in the other experiment no electrolyte is added. Then the cells are hermetically sealed. The exothermic heat is measured during heating to 350 C. at a rate of 5 K/min. The mass of the DSC cell is checked before and after measurement to ensure that the DSC cell does not leak.

(101) FIG. 8 compares the result of DSC measurements of charged electrodes (A) without and (B) with 1.16 mg of electrolyte added. The measurement shows the typical strong exothermic event at about 290 C. where charged cathode and electrolyte start a violent reaction. However, in the absence of electrolyte almost no heat evolution is observed. At first glance this resultthe absence of an exothermic eventis a surprise since delithiated Li.sub.0.4MO.sub.2 is thermodynamically highly unstable. This is because during the charging process (reaction 1 below) the extraction of Li requires a large amount of energy (at 4.3V about 3.7V170 mAh/g). Then, during heating the cathode collapses and releases oxygen (reaction 2 below). At first glance it can be expect that the reaction 2) from unstable compounds to stable compounds is strongly exothermic. So the result (small exothermic heat for 2)) appears surprising. But if we look at the whole thermodynamic cycle by adding reaction 3) we can understand that reaction 2) does not evolve much exothermic heat.
Charge: stable.fwdarw.unstable: LiMO.sub.2.fwdarw.Li.sub.0.4MO.sub.2+0.6Li1)
Cathode collapse: unstable.fwdarw.stable: Li.sub.0.4MO.sub.2.fwdarw.0.4LiMO.sub.2+0.3M.sub.2O.sub.3+0.15O.sub.22)
Complete the cycle: 0.4LiMO.sub.2+0.3M.sub.2O.sub.3+0.15O.sub.2+0.6Li.fwdarw.LiMO.sub.23)

(102) We can split reaction 3) into 2 reactions:
0.6Li+0.15O.sub.2.fwdarw.0.3Li.sub.2O (strong exotherm)3a)
0.3Li.sub.2O+0.3M.sub.2O.sub.3.fwdarw.0.6LiMO.sub.23b)

(103) In these reactions only 3a) is strongly exotherm (Li burns with the evolution of lots of heat). None of the other reactions 1), 3b) is strongly endotherm: In reaction 1) H.sub.1 is near zero because a battery does not change much in temperature when charged or discharged. In reaction 3b) H.sub.3b is near zero because generally the creation of double oxides from simple oxides does not create or consume much heat. Therefore reaction 2) cannot be strongly exotherm, because in a thermodynamic cycle the sum of the formation enthalpies adds up to zero: 0=H.sub.1+H.sub.2+H.sub.3aH.sub.3b. If however, reaction 2) is not strong exotherm, then no heat evolves during cathode collapse and the battery does not self-heat and does not go into thermal runaway. In the presence of electrolyte however, the situation is different, because the cathode collapse occurs simultaneously with the very exothermic combustion of electrolyte. It can be concluded that reducing the electrolyte: active mass ratio allows to improve the safety of real batteries.

Example 8: Improved Safety by Lowering the Electrolyte: Cathode Ratio

(104) Charged Li batteries are potentially unsafe. In the worst case scenario the delithiated cathode and the lithiated anode react with the electrolyte, causing a thermal runaway. The major contribution to the exothermic heat is the electrolyte oxidation: the charged cathode can release oxygen which combusts the electrolyte. If we assume that x in equation (4) (of the CCE reaction described before) is 0.5, then 1 mol charged Li.sub.0.5CoO.sub.2 can combust 0.1 mol C.sub.3H.sub.4O.sub.3 (ethylene carbonate (EC) based electrolyte). We can estimate how much electrolyte can be combusted by the oxygen released from the cathode. We assume x=0.5, and for simplicity we use EC to represent the electrolyte. 1 mol LiCoO.sub.2=97.9 g 0.1 mol C.sub.3H.sub.4O.sub.3=8.8 g

(105) The electrolyte: cathode mass ratio which allows for 100% combustion is about 9 percent (8.8/97.90.09, theoretically for x=1 this would be 18%). The result is that the safety of a battery can be improved if much less than about 9% by weight of electrolyte relative to the active cathode is added to a battery. If less electrolyte is present, then less electrolyte combusts and the thermal safety will improve.

(106) The electrolyte fills the pores in the electrodes. The pores must be filled with electrolyte otherwise the battery will have a poor performance. For simplicitylets assume that of the electrolyte is in the pores of the cathode (this is achieved for example if cathode and anode have the same thickness and same porosity). Lets furthermore assume that the separator is thin, so we can neglect its porosity, and lets assume that the electrolyte exactly fills the porosity, meaning no leftover electrolyte is present in the battery after assembly. The critical porosity is the porosity whichwhen filled with electrolytecorresponds to a ratio of electrolyte to cathode active material of 9 wt %. Using these assumptions together with the densities of cathode and electrolyte we can estimate a critical porosity as follows: LiCoO.sub.2: Density 5.05 g/cm.sup.3 Typical electrode composition: LiCoO.sub.2: 96 wt % (2% Binder (1.77 g/cm.sup.3), 2% carbon (2 g/cm.sup.3)).fwdarw.electrode theoretical density (0% porosity)=4.92 g/cm.sup.3 C.sub.3H.sub.4O.sub.3: Density 1.32 g/cm.sup.3 Electrolyte: 9% (by weight, per mass of cathode) present in the battery, of those 9% (=4.5%) present in the cathode.

(107) Using these data allows to calculate the critical porosity as 13.7% (in volume), knowing that vol1=vol electrode=1/(0.96*4.92)=0.214; vol2=vol electrolyte=0.045/1.32=0.034; and porosity=vol2/(vol1+vol2). The critical porosity increases with charge voltage. The high voltage stable cathodes allow for reversible cycling at >4.5V. At 4.5V we achieve a reversible capacity of >185 mAh/g (the total charge being 280mAh per gram when all Li is extracted). This corresponds to x>0.675 in the 4.5V charged Li.sub.1xCoO.sub.2, corresponding to a critical porosity of 20.5%. If the real cathode porosity is further decreased or if x is increased (more lithium extracted at full charge by charging at a higher voltage) then we can assume that the safety of a battery having a certain capacity improves. The example shows that electrodes with small porosity allow to reduce the electrolyte sufficiently to achieve improved safety.

Example 9

(108) A LiCoO.sub.2 with an average particle size (D50) in excess of 100 m is prepared by using a large excess of lithium and sintering at high temperature. After sintering the excess Li is removed, resulting in a stoichiometric LiCoO.sub.2, by performing the following steps:

(109) firing a blend of Li.sub.2CO.sub.3 and Co.sub.3O.sub.4 (mol ratio Li:Co=1.2) at 990 C. for 12 h

(110) removing excess Li.sub.2CO.sub.3 by washing,

(111) followed by adding of more Co.sub.3O.sub.4 (about 6% Co per mol LiCoO2), and

(112) re-firing at 950 C. The particles of the final sample are dense rock shaped.

(113) FIG. 9 shows a micrograph of the particles of the obtained powder.

(114) A Li and manganese rich cathode materialreferred to as HLMwith composition Li.sub.1+xM.sub.1xO.sub.2 is prepared from a suitable MCO.sub.3 precursor by mixing with Li.sub.2CO.sub.3 and firing in air at 800 C. The final composition has a Li:M ratio of about 1.42 and a transition metal composition M=Mn.sub.0.67Ni.sub.0.22Co.sub.0.11. FIG. 10 shows a micrograph of the particles of the obtained powder. Note that the magnification in FIG. 10 is 10 that of FIG. 9.

(115) The LiCoO.sub.2 and HLM cathode powders are mixed. 3 mixtures are prepared containing 5, 10 and 20% by mass of HLM powder, labeled M05, M10, M15. The powder density is measured by compacting the pellets to a Density 1, then, after relaxing the pressure a Density 2 is measured. A very high density of 4.17 g/cm.sup.3 is measured for M10. This high density indicates that a very low porosity of 5 vol % or so can be achieved in real electrodes.

(116) The mixtures are tested in Li coin cells. Coin cell are tested at a rate of C/20 (corresponding to a rate of 8 mA/g) between 2.0 and 4.6V. Cycle 2 is at C/10 rate (16 mA/g) between the same voltage limits. A very high reversible capacity is achieved, proving that even the very large and dense LiCoO.sub.2 particles can cycle well. Table 9 summarizes the results. FIG. 11 shows the obtained voltage profiles.

(117) TABLE-US-00009 TABLE 9 LiCoO.sub.2:HLM Q discharge Density1 Density2 Sample mass ratio mAh/g g/cm.sup.3 g/cm.sup.3 M05 95:5 227.4 mAh/g 4.30 4.17 M10 90:10 230.7 mAh/g 4.17 4.07 M20 80:20 237.0 mAh/g 3.97 3.89

Example 10

(118) This example will demonstrate the improved safety of a cathode when less that than the critical amount of electrolyte is present. The example predicts that a cathode with a very low porositywhich only allows for less than the critical amount of electrolyte to be presentwill provide improved safety.

(119) The safety of sample M10 of Example 9 is estimated by a DSC measurement: 5 coin cells are prepared and charged at 25 C. to 4.5V at C/10 rate (16 mA/g). The obtained capacities are 197.2-197.8 mAh/g. The cells are disassembled directly after reaching the target voltage. The electrodes are washed in DMC to remove the electrolyte. After drying, small electrode discs with 3 mg of active material are inserted into DSC cells. 3 different types of DSC cells are prepared:

(120) Cell type 1) No electrolyte is added, the cell is just crimped,

(121) Cell type 2) about 2.6 mg of an electrolyte (ED/EMC) diluted 1:10 by DMC is added. After a few moments most of the DMC is evaporated, and the cell is crimped,

(122) Cell type 3) about 2.6 mg of electrolyte is added and the cell is crimped.

(123) In this way DSC cells with a electrolyte:cathode ratio of (1) zero (being much lower than the critical ratio), (2) 0.08being less than the critical ratioand (3) 0.46by far exceeding the critical ratioare obtained.

(124) The heat evolution is measured during heating at a rate of 5 K/min to 350 C.

(125) Table 10 summarizes the obtained results. FIG. 12 shows the obtained DSC heat profiles. Clearly, as the amount of electrolyte decreases below the critical ratio the evolved heat decreases, thus the safety of a battery with a small amount of electrolyte will be high.

(126) TABLE-US-00010 TABLE 10 0% electrolyte 8% electrolyte 46% electrolyte Integrated heat (kJ/g) 0.136 0.725 1.551 (1.601) cathode

(127) It should be clear that the conclusions of the different examples embodying the invention are valid for undoped and doped LCO, the dopants being for example Mg, Ti, Fe, Cu, Ca, Ba, Y, Sn, Sb, Na, Zn, Zr, Si, Er, Nd and Nb.