Substituted lithium-manganese metal phosphate

Abstract

A substituted lithium-manganese metal phosphate of formula
LiFe.sub.xMn.sub.1-x-yM.sub.yPO.sub.4
in which M is a bivalent metal from the group Sn, Pb, Zn, Ca, Sr, Ba, Co, Ti and Cd and wherein: x<1, y<0.3 and x+y<1, a process for producing it as well as its use as cathode material in a secondary lithium-ion battery.

Claims

1. A lithium-manganese metal phosphate of formula
LiFe.sub.xMn.sub.1-x-yM.sub.yPO.sub.4 wherein M is a bivalent metal selected from the group consisting of Zn, Ca, and Cd, and wherein: x<1, 0<y<0.15 and x+y<1.

2. The lithium-manganese metal phosphate according to claim 1, wherein M is Zn or Ca.

3. The lithium-manganese metal phosphate according to claim 1, further comprising carbon.

4. The lithium-manganese metal phosphate according to claim 3, wherein the carbon is evenly distributed throughout the lithium-manganese metal phosphate.

5. The lithium-manganese metal phosphate according to claim 3, wherein the carbon covers the individual particles of the lithium manganese metal phosphate.

6. The lithium-manganese metal phosphate according to claim 3, wherein the proportion of carbon relative to the lithium manganese metal phosphate is <4 wt %.

7. Cathode for a secondary lithium-ion battery, comprising the lithium-manganese metal phosphate according to claim 1.

8. The cathode according to claim 7, further comprising a lithium-metal-oxygen compound.

9. The cathode according to claim 8, wherein the lithium-metal-oxygen compound is selected from the group consisting of LiCoO.sub.2, LiMn.sub.2O.sub.4, LiNiO.sub.2, LiCoPO.sub.4, LiFePO.sub.4, LiMnPO.sub.4, LiMnFePO.sub.4, and mixtures thereof.

10. The cathode according to claim 7, wherein the cathode is free of added conductive agents.

11. Process for producing the lithium-manganese metal phosphate according to claim 1, the process comprising the following steps: a) producing a mixture in aqueous solution containing at least a Li starting compound, a Mn starting compound, an Fe starting compound, a M.sup.2+ starting compound and a PO.sub.4.sup.3 starting compound, until a suspension forms, b) carrying out a dispersion or grinding treatment of the mixture and/or the suspension, c) obtaining a lithium-manganese metal phosphate of formula LiFe.sub.xMn.sub.1-x-yM.sub.yPO.sub.4 from the suspension by reaction of the suspension under hydrothermal conditions.

12. The process according to claim 11, further comprising, in step a) or step c), adding a carbon-containing component.

13. The process according to claim 12, further comprising mixing the lithium-manganese metal phosphate of formula LiFe.sub.xMn.sub.1-x-yM.sub.yPO.sub.4 obtained in step c) with a carbon-containing component.

14. The process according to claim 11, further comprising carrying out a drying step at a temperature of <100 C. and/or a calcining step at a temperature of >200 C.

15. The process according to claim 11, wherein the process is carried out under hydrothermal conditions at a temperature of 100 C. to 200 C. and at a pressure of 1 bar to 40 bar vapour pressure.

16. Lithium-manganese metal phosphate produced by the process according to claim 11.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 discharge curves at 1 C for a lithium-manganese iron phosphate LiMn.sub.0.66Fe.sub.0.33PO.sub.4 of the state of the art;

(2) FIG. 2 discharge curves at 1 C for the LiMn.sub.0.56Fe.sub.0.33Mg.sub.0.10PO.sub.4 according to the invention;

(3) FIG. 3 discharge curves at 1 C for the LiMn.sub.0.56Fe.sub.0.33Zn.sub.0.10PO.sub.4 according to the invention;

(4) FIG. 4 the charge curves of lithium-manganese iron phosphate (LiMn.sub.0.66Fe.sub.0.33PO.sub.4) of the state of the art;

(5) FIG. 5 the charge curves of LiMn.sub.0.56Fe.sub.0.33Zn.sub.0.10PO.sub.4 according to the invention;

(6) FIG. 6 the discharge curves at different rates for an electrode containing LiMn.sub.0.56Fe.sub.0.33Zn.sub.0.10PO.sub.4 according to the invention;

(7) FIG. 7 the discharge curves at different rates for an electrode containing LiMn.sub.0.66Fe.sub.0.33PO.sub.4 according to the state of the art;

(8) FIG. 8a-b the comparison of the materials according to the invention with LiFePO.sub.4 in full cells versus a lithium titanate anode, at C/10 (FIG. 8a) and at 20 C (FIG. 8b)

(9) FIG. 9 the comparison of the specific discharge capacity at C/10 between a lithium-manganese iron phosphate (LiMn.sub.0.66Fe.sub.0.33PO.sub.4) of the state of the art with substituted lithium-manganese metal phosphates according to the invention;

(10) FIG. 10 voltage profiles at 1 C after aging of materials according to the invention vis--vis lithium-manganese iron phosphate (LiMn.sub.0.66Fe.sub.0.33PO.sub.4) of the state of the art;

(11) FIGS. 11a-c the influence of the electrode density on the discharge rate in the case of material according to the invention;

(12) FIGS. 12a-c the influence of the electrode density in the case of a cathode containing the material according to the invention;

(13) FIG. 13 the discharge curve at C/10 for LiMn.sub.0.80Fe.sub.0.10Zn.sub.0.10PO.sub.4 according to the invention;

(14) FIG. 14 the discharge curve at C/10 for LiMn.sub.0.85Fe.sub.0.10Zn.sub.0.05PO.sub.4 according to the invention.

DETAILED DESCRIPTION

Embodiment Examples

1. Determination of the Particle-Size Distribution

(15) The particle-size distributions for the mixtures or suspensions and of the produced material is determined using the light-scattering method using devices customary in the trade. This method is known per se to a person skilled in the art, wherein reference is also made in particular to the disclosure in JP 2002-151082 and WO 02/083555. In this case, the particle-size distributions were determined with the help of a laser diffraction measurement apparatus (Mastersizer S, Malvern Instruments GmbH, Herrenberg, Del.) and the manufacturer's software (version 2.19) with a Malvern Small Volume Sample Dispersion Unit, DIF 2002 as measuring unit. The following measuring conditions were chosen: compressed range; active beam length 2.4 mm; measuring range: 300 RF; 0.05 to 900 m. The sample preparation and measurement took place according to the manufacturer's instructions.

(16) The D.sub.90 value gives the value at which 90% of the particles in the measured sample have a smaller or the same particle diameter. Accordingly, the D.sub.50 value and the D.sub.10 value give the value at which 50% and 10% respectively of the particles in the measured sample have a smaller or the same particle diameter.

(17) According to a particularly preferred embodiment according to the invention, the values named in the present description are valid for the D.sub.10 values, D.sub.50 values, the D.sub.90 values as well as the difference between the D.sub.90 and D.sub.10 values relative to the volume proportion of the respective particles in the total volume. Accordingly, according to this embodiment according to the invention, the D.sub.10, D.sub.50 and D.sub.90 values named here give the values at which 10 volume-% and 50 volume-% and 90 volume-% respectively of the particles in the measured sample have a smaller or the same particle diameter. If these values are preserved, particularly advantageous materials are provided according to the invention and negative influences of relatively coarse particles (with relatively larger volume proportion) on the processability and the electrochemical product properties are avoided. Particularly preferably, the values named in the present description are valid for the D.sub.10 values, the D.sub.50 values, the D.sub.90 values as well as the difference between the D.sub.90 and the D.sub.10 values relative to both percentage and volume percent of the particles.

(18) For compositions (e.g. electrode materials) which, in addition to the lithium-manganese iron phosphates according to the invention substituted with bivalent metal cations, contain further components, in particular for carbon-containing compositions, the above light scattering method can lead to misleading results as the LiFe.sub.xMn.sub.1-x-yM.sub.yPO.sub.4 particles can be joined together by the additional (e.g. carbon-containing) material to form larger agglomerates. However, the particle-size distribution of the material according to the invention can be determined as follows for such compositions using SEM photographs:

(19) A small quantity of the powder sample is suspended in acetone and dispersed with ultrasound for 10 minutes. Immediately thereafter, a few drops of the suspension are dropped onto a sample plate of a scanning electron microscope (SEM). The solids concentration of the suspension and the number of drops are measured such that a largely single-ply layer of powder particles (the German terms Partikel and Teilchen are used synonymously to mean particle) forms on the support in order to prevent the powder particles from obscuring one another. The drops must be added rapidly before the particles can separate by size as a result of sedimentation. After drying in air, the sample is placed in the measuring chamber of the SEM. In the present example, this is a LEO 1530 apparatus which is operated with a field emission electrode at 1.5 kV excitation voltage and a 4 mm space between samples. At least 20 random sectional magnifications of the sample with a magnification factor of 20,000 are photographed. These are each printed on a DIN A4 sheet together with the inserted magnification scale. On each of the at least 20 sheets, if possible at least 10 free visible particles of the material according to the invention, from which the powder particles are formed together with the carbon-containing material, are randomly selected, wherein the boundaries of the particles of the material according to the invention are defined by the absence of fixed, direct connecting bridges. On the other hand, bridges formed by carbon material are included in the particle boundary. Of each of these selected particles, those with the longest and shortest axis in the projection are measured in each case with a ruler and converted to the actual particle dimensions using the scale ratio. For each measured LiFe.sub.xMn.sub.1-x-yM.sub.yPO.sub.4 particle, the arithmetic mean from the longest and the shortest axis is defined as particle diameter. The measured LiFe.sub.xMn.sub.1-x-yM.sub.yPO.sub.4 particles are then divided analogously to the light-scattering measurement into size classes. The differential particle-size distribution relative to the number of particles is obtained by plotting the number of the associated particles in each case against the size class. The cumulative particle-size distribution from which D.sub.10, D.sub.50 and D.sub.90 can be read directly on the size axis is obtained by continually totaling the particle numbers from the small to the large particle classes.

(20) The described process is also applied to battery electrodes containing the material according to the invention. In this case, however, instead of a powder sample a fresh cut or fracture surface of the electrode is secured to the sample holder and examined under a SEM.

Example 1

Production of LiMn0.56Fe0.33Zn0.10PO4 According to the Process According to the Invention

(21) When producing the material it is to be borne in mind that the material is precipitated from an aqueous Fe.sup.2+ precursor solution. The reaction and drying/sintering is therefore preferably to be carried out under protective gas or vacuum in order to avoid a partial oxidation of Fe.sup.2+ to Fe.sup.3+ with formation of by-products such as Fe.sub.2O.sub.3 or FePO.sub.4.

(22) Production and Precipitation/Suspension of a Precursor Mixture

(23) First, 105.5 g lithium hydroxide LiOH.H.sub.2O was dissolved in 0.9 l distilled water. This solution is called basic solution.

(24) 77.43 g FeSO.sub.47 H.sub.2O, 79.88 g MnSO.sub.4H.sub.2O and 24.27 g ZnSO.sub.47 H.sub.2O were dissolved in approx. 1.5 l distilled water and 103.38 g 80% phosphoric acid added slowly accompanied by stirring. 0.4 l distilled wash water was added. This solution is called acid solution.

(25) The basic solution was introduced into the laboratory autoclave (capacity: 4 liters) at 600 rpm stirrer speed, the autoclave loaded with approx. 6-7 bar nitrogen via the dipping tube and relieved again via the vent valve. The procedure was repeated three times.

(26) A disperser (IKA, ULTRATURRAX UTL 25 Basic Inline with dispersion chamber DK 25.11) was connected to the autoclave between vent valve and bottom outlet valve in order to carry out the dispersion or grinding treatment. The pumping direction of the disperser was bottom outlet valve-disperser-vent valve. The disperser was started on the middle power level (13,500 rpm) according to the manufacturer's instructions.

(27) The prepared acid solution was then pumped with a membrane pump via the dipping tube into the autoclave (stroke 100%, 180 strokes/minute; corresponds to the maximum capacity of the pump) and reflushed with approx. 500 to 600 ml distilled water. The pumping-in lasted for approx. 20 minutes, wherein the temperature of the resultant suspension increased to approx. 40 C. After pumping-in of the acid solution, a deposit precipitated out.

(28) The disperser, which was started before the addition of the acid solution, was used for a total of approx. 1 hour for intensive mixing or grinding of the resultant, viscous suspension (after pumping-in of the acid solution at 50 C.).

(29) The use of a disperser brings about an intensive mixing and the agglomeration of the precipitated viscous pre-mixture. During the precipitation and crystallization of the suspension, a homogeneous mixture of many small, approximately equally-sized crystal nuclei formed in the disperser as a result of the pre-grinding or intensive mixing. These crystal nuclei crystallized during the subsequent hydrothermal treatment (see below) to very uniformly grown crystals of the end-product with a very narrow particle-size distribution. The power and energy input via the dispersion treatment was respectively more than 7 kW/m.sup.3 and more than 7 kWh/m.sup.3 of the treated precursor mixture/suspension.

(30) Hydrothermal Treatment

(31) Each freshly produced suspension was subjected to hydrothermal treatment in the laboratory autoclave. Before heating the suspension, the autoclave was flushed with nitrogen in order to expel any air present from the autoclave before the hydrothermal process. The product according to the invention formed starting from hydrothermal temperatures of approximately 100 to 120 C. The hydrothermal treatment was preferably carried out for 2 hours at 130 C.

(32) After switching off and disconnecting the disperser the mixture was heated over 1.5 hours to 130 C., for 2 hours. Cooling to 30 C. then took place over 3 hours.

(33) The LiMn.sub.0.56Fe.sub.0.33Zn.sub.0.10PO.sub.4 according to the invention was then able to be dried in air or in the drying oven for example at mild temperatures (40 C.) without visible oxidation.

(34) The thus-obtained material was pumped under nitrogen atmosphere through the bottom outlet valve of the autoclave into a pressure filter (Seitz filter). The membrane pump setting was such that a pressure of 5 bar was not exceeded. The filter cake was subsequently washed with distilled water until the conductivity of the wash water had fallen below 42 S/cm.

Example 2

Production of LiMn0.56Fe0.33Mg0.10PO4

(35) The synthesis was carried out as in Example 1, except that 20.80 g MgSO.sub.4*7H.sub.2O was used as starting material in the corresponding molar weight quantities instead of ZnSO.sub.4.

Example 3

Production of LiMn0.80Fe0.10Zn0.10PO4 According to the Process According to the Invention

(36) The synthesis was carried out as in Example 1, except that 114.12 g MnSO.sub.4*7H.sub.2O, 23.46 g FeSO.sub.4*7H.sub.2O, 24.27 g ZnSO.sub.4*7H.sub.2O, 103.38 g H.sub.3PO.sub.4, (80%) were used as starting materials in the corresponding molar weight quantities.

Example 4

Production of LiMn0.85Fe0.10Zn0.05PO4 According to the Process According to the Invention

(37) The synthesis was carried out as in Example 1, except that 121.26 g MnSO.sub.4*1H.sub.2O, 23.46 g FeSO.sub.4*7H.sub.2O, 12.14 g ZnSO.sub.4*7H.sub.2O, 103.38 g H.sub.3PO.sub.4 (80%) were used as starting materials in the corresponding molar weight quantities.

Example 5

Carbon Coating of the Obtained Material (Variant 1)

(38) The filter cakes obtained in Examples 1 to 4 were impregnated with a solution of 24 g lactose in water and then calcined at 750 C. for 3 hours under nitrogen.

(39) Depending on the quantity of lactose, the proportion of carbon in the product according to the invention was between 0.2 and 4 wt.-%.

(40) Typically 1 kg dry product from Examples 1 and 2 was mixed intimately with 112 g lactose monohydrate and 330 g deionized water and dried overnight in a vacuum drying oven at 105 C. and <100 mbar to a residual moisture of 3%. The brittle drying product was broken by hand and coarse-ground in a disk mill (Fritsch Pulverisette 13) with a 1 mm space between disks and transferred in high-grade steel cups into a protective gas chamber furnace (Linn KS 80-S). The latter was heated to 750 C. within 3 hours at a nitrogen stream of 200 I/h, kept at this temperature for 3 hours and cooled over 3 hours to room temperature. The carbon-containing product was disagglomerated in a jet mill (Hosokawa).

(41) The SEM analysis of the particle-size distribution produced the following values: D.sub.50<0.5 m, difference between D.sub.90 and D.sub.10 value: <1 m.

Example 6

Carbon Coating of the Material According to the Invention (Variant 2)

(42) The synthesis of the materials according to the invention was carried out as in Examples 1 to 4, except that gelatine was also then added (9 g gelatine per 100 g starting product) during the precipitation step a). The end-product contained approx. 2.3 wt.-% carbon.

Example 7

Production of Electrodes

(43) Thin-film electrodes as disclosed for example in Anderson et al., Electrochem. and Solid State Letters 3 (2) 2000, pages 66-68 were produced. The electrode compositions usually consisted of 90 parts by weight active material, 5 parts by weight Super P carbon and 5% polyvinylidene fluoride as binder or 80 parts by weight active material, 15 wt.-% Super P carbon and 5 parts by weight polyvinylidene fluoride, or 95 parts by weight active material and 5 parts by weight polyvinylidene fluoride.

(44) The active material was mixed, together with the binder (or, for the electrodes of the state of the art, with the added conductive agent), in N-methylpyrrolidone, applied to a pretreated (primer) aluminium foil by means of a coating knife and the N-methylpyrrolidone was evaporated at 105 C. under vacuum. The electrodes were then cut out (13 mm diameter) and roll-coated with a roller at room temperature. The starting nip width was e.g. 0.1 mm and the desired thickness progressively built up in steps of 5-10 m. 4 rolled coats were applied at each step and the foil was rotated by 1800. After this treatment, the thickness of the coating was between 20-25 m. The primer on the aluminium foil consisted of a thin carbon coating which improves the adhesion of the active material particularly when the active material content of the electrode is more than 85 wt.-%.

(45) The electrodes were then dried overnight at 120 C. under vacuum and assembled and electrochemically measured against lithium metal in half cells in an argon-filled glovebox.

(46) The electrochemical measurements were carried out against lithium metal using LP30 (Merck, Darmstadt) as electrolyte (EC (ethylene carbonate):DMC (dimethylcarbonate)=1:1, 1 M LiPF.sub.6).

(47) The test procedure was carried out in CCCV mode, i.e. cycles with a constant current at the C/10 rate for the first, and at the C rate for the subsequent, cycles. In some cases, a constant voltage portion followed at the voltage limits (1.0 and 2.0 volt versus Li/Li.sup.+) until the current fell approximately to the C/50 rate, in order to complete the charge/discharge cycle.

(48) Corresponding measurements of the specific capacity and the current carrying capacity were carried out on both LiMn.sub.0.66Fe.sub.0.33PO.sub.4 of the state of the art and materials according to the invention substituted with magnesium and zinc. LiFePO.sub.4 electrodes (available from Sd-Chemie) were likewise also measured.

(49) FIG. 1 shows the discharge curves at 1 C for a LiMn.sub.0.66Fe.sub.0.33PO.sub.4 of the state of the art.

(50) After several cycles, an energy loss is recorded in the range of between 20 and 40 mAh/g. In contrast, the magnesium- and zinc-substituted materials according to the invention (FIGS. 2 and 3) display almost no energy loss in the range of between 20 and 60 mAh/g even after 180 cycles. Nor is any weakening in the capacity at 140 mAh/g to be ascertained even after extended cycles.

(51) FIG. 4 shows charge curves of a LiMn.sub.0.66Fe.sub.0.33PO.sub.4 material of the state of the art with an electrode density of 1.2 g/cm.sup.3 and a thickness of 20 m. By way of comparison, the charge curve for the zinc-containing material according to the invention is shown in FIG. 5. As can be seen from FIGS. 4 and 5, the 1 C value of the material according to the invention is much better during the charge up to 4.3V than in the case of the comparison material of the state of the art.

(52) FIGS. 6 and 7 show the discharge capacity at different rates of the material according to the invention (FIG. 6) as well as of a lithium-manganese iron phosphate of the state of the art (FIG. 7). The electrode density was 1.2 g/cm.sup.3 in the material according to the invention and 1.3 g/cm.sup.3 in the comparison material at a thickness of approx. 20 m.

(53) Here also, there is a significant drop in performance after several discharge cycles of the lithium-manganese iron phosphate not substituted with (electrically inactive) bivalent material compared with the lithium-manganese iron phosphate not substituted with a bivalent metal ion. In addition, the polarization at increased current rates is much greater for the lithium-manganese iron phosphate of the state of the art than for a substituted lithium-manganese iron phosphate according to the invention.

(54) FIG. 8a-b shows the discharge curves in full-cell configuration versus a lithium titanate (Li.sub.4Ti.sub.5O.sub.12) anode at D/10 and at 20D for a carbon-coated lithium iron phosphate of the state of the art (available from Sd-Chemie) and for the LiMn.sub.0.56Fe.sub.0.33Zn.sub.0.10PO.sub.4 material according to the invention. The electrode composition was 90 wt.-% active material, 5% Super P graphite and 5 wt.-% polyvinylidene fluoride. The cell was balanced such that the mass of the cathode and the mass of the anode were as similar as possible.

(55) Here it is shown that the LiMn.sub.0.56Fe.sub.0.33Zn.sub.0.1PO.sub.4 material according to the invention has a long 4-volt plateau up to 80 mAh/g and at the same time a specific discharge capacity at D/10 comparable to the (carbon-coated) lithium iron phosphate of the state of the art, which means a clear increase in the energy density vis--vis lithium iron phosphate.

(56) FIG. 9 shows the C/10 or discharge capacity of LiMn.sub.0.56Fe.sub.0.33Zn.sub.0.10PO.sub.4 and LiMn.sub.0.56Fe.sub.0.33Mg.sub.0.10PO.sub.4 materials according to the invention compared with two LiMn.sub.0.66Fe.sub.0.33PO.sub.4 materials of the state of the art. Here also it can surprisingly be seen that an increase in the manganese plateau potential is achieved by zinc and magnesium substitutions, although in these cases Mg and zinc are electrically inactive ions and the manganese content is smaller than in the case of the materials of the state of the art.

(57) After 20 cycles, the charge and discharge cycles at C/10 and 1D for the LiMn.sub.0.56Fe.sub.0.33Zn.sub.0.10PO.sub.4 and LiMn.sub.0.56Fe.sub.0.33Mg.sub.0.10PO.sub.4 materials according to the invention and for lithium-manganese iron phosphates of the state of the art (FIG. 10) were measured. Here also a stabilization of the length of the 4-volt manganese plateau by the materials according to the invention compared with lithium-manganese iron phosphate of the state of the art is shown.

(58) FIGS. 11a to c show the variation in electrode density relative to the discharge capacities at different rates. In FIG. 11a, the density of the material is 1.6 g/cm.sup.3, in FIG. 11b 1.7 g/cm.sup.3 and in FIG. 11c 2.0 g/cm.sup.3.

(59) It is shown that it is possible with the materials according to the invention to increase the electrode density while preserving the discharge capacities.

(60) FIGS. 12a to c show the influence of the electrode thickness on the discharge capacities. The thickness of the electrodes in FIG. 12a was 25 m, in FIG. 12b 33 m and in FIG. 12c 51 m. It is shown here also that the plateau can be kept at 4 volt and the discharge capacity at least up to 5 C, and the active matter load can simultaneously be increased.

(61) Even higher energy densities can be achieved by increasing the manganese content of these substituted materials according to the invention.

(62) FIGS. 13 and 14 show the discharge capacity at D/10 of the LiMn.sub.0.80Fe.sub.0.10Zn.sub.0.10PO.sub.4 and LiMn.sub.0.85Fe.sub.0.10Zn.sub.0.05PO.sub.4 materials according to the invention (produced according to Examples 2 and 4). It is shown that the 4V manganese plateau is even longer than in the case of LiMn.sub.0.56Fe.sub.0.33Zn.sub.0.10PO.sub.4 according to the invention.

(63) In summary, the present invention makes available mixed lithium-manganese iron phosphate materials substituted with bivalent metal ions, which can be produced by means of a hydrothermal process. The specific discharge capacity for room temperature exceeds 140 mAh/g despite the substitution with sometimes 10% electrochemically inactive bivalent metal ions. Very good discharge rates were measured for all the substituted materials.

(64) Compared with non-substituted LiMn.sub.0.66Fe.sub.0.33PO.sub.4 it was shown that the discharge voltage profile at 1 D for the bivalently substituted novel materials according to the invention remains unchanged even after several charge and discharge cycles (the length of the Mn plateau at 4 volt remained unchanged).

(65) It was found with respect to the energy density that the substitution with zinc or with magnesium gave the best results compared with copper, titanium and nickel.