Mixed oxide containing a lithium manganese spinel and process for its preparation

Abstract

A mixed oxide containing a) a mixed-substituted lithium manganese spinel in which some of the manganese lattice sites are occupied by lithium ions and b) a boron-oxygen compound. Furthermore, a process for its preparation and the use of the mixed oxide as electrode material for lithium ion batteries.

Claims

1. A process for the preparation of a mixed oxide containing a mixed-doped hyperstoichiometric lithium manganese spinel comprising: (a) providing components comprising a lithium component, a manganese component, and a boron component, and optionally a component comprising N, wherein N is selected from the group consisting of Al, Mg, Co, Ni, Cu and Cr and optionally a component comprising M, wherein M is selected from the group consisting of Zn, Mg and Cu, wherein at least some of the components are in a dry, powdery form; (b) preparing a solid mixture by mixing at least some of the dry, powdery components of (a) to form a mixture in dry, powdery form; (c) preparing a liquid mixture comprising a solvent and a portion of the components of (a) dissolved and/or suspended in the solvent, (d) mixing the solid mixture of (b) and the liquid mixture of (c) with each other; (e) drawing the solvent off the mixture obtained in step (d); (f) calcining the mixture obtained in (e) at a temperature of more than 300 C.; wherein the stoichiometric quantities of the components of the first and second portions are selected in such a way that a mixed oxide of the following formula results:
[(Li.sub.1aM.sub.a)(Mn.sub.2cdLi.sub.cN.sub.d)O.sub.x].b(B.sub.zO.sub.yH.sub.uX.sub.v) wherein: 0a<0.1; d<1.2 3.5<x<4.5; 0.01<c<0.06 z=1, 2 or 4 y=1, 2, 3 or 7 u=0, 1, 2 or 3 v=0, 1, 2 or 3 0.01<b<0.5; M is at least one element from the group of Zn, Mg and Cu; N is at least one element from the group of Al, Mg, Co, Ni, Cu and Cr; and X is at least one element from the group Li, Na, K; whereby a lithium manganese spinel is obtained having a d.sub.50 primary crystallite size of at least 0.5 m.

2. The process according to claim 1, wherein the calcining of (f) is carried out in at least two steps, wherein calcination is at temperatures of 300 C. to 600 C. in a first step and at temperatures of 600 C. to 900 C. in a second step.

3. The process according to claim 1, wherein the D.sub.50 particle size value of the lithium component is less than 30 m.

4. The process according to claim 3, wherein before the mixing according to (d) the manganese component has a D.sub.95 particle size value measured by laser granulometry of less than 30 m.

5. The process according to claim 1, wherein the solvent of the liquid mixture of (c) is water.

6. The process according to claim 5, wherein the liquid mixture of (c) contains a portion of the manganese component and/or the component comprising N.

7. The process according to claim 6, wherein the solid mixture of (b) is stirred into the liquid mixture of (c).

8. The process according to claim 1, wherein the components of (a) are provided in the form of their nitrates, acetates, oxides, hydroxides and/or carbonates.

9. The process according to claim 4, wherein the manganese component is selected from the group consisting of manganese carbonate, manganese oxide and manganese dioxide.

10. The process according to claim 1, wherein the values chosen for b and d are 0.0025<b<0.025 and 0.05<d<0.2.

11. The process according to claim 10, wherein the values chosen for b, d and a are 0.0025<b<0.025, and 0.08<d<0.15, and 0.005<a<0.02.

12. The process according to claim 11, wherein N is selected from Al and Mg, and M is Zn.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The figures show in detail:

(2) FIG. 1: dQ/dE plotted against E for various Co/Al/Li levels of lithium manganese spinels

(3) FIG. 2: comparison of XRD spectra of mixed oxides according to aspects of the invention and pure lithium manganese spinels

(4) FIG. 3: the flowchart of the method according to aspects of the invention

(5) FIG. 4: SEM micrographs of mixed oxides according to aspects of the invention

(6) FIG. 5: position, reflex width and phase-purity of mixed oxides according to aspects of the invention

(7) FIG. 6: the effect of the addition of borate on the electrochemical behaviour

(8) FIG. 7: the manganese dissolution of mixed oxides according to aspects of the invention compared with pure lithium manganese spinels in mg Mn/kg

(9) FIG. 8: the relationship between the BET surface area of mixed oxide crystals according to aspects of the invention and temperature and lithium hyperstoichiometry

(10) FIG. 9: dQ/dE plotted against E for mixed oxides according to aspects of the invention with different lithium hyperstoichiometry

(11) FIG. 10: the discharge capacities of mixed crystals according to aspects of the invention in the first cycle

(12) FIG. 11: the course of the potential in the first complete cycle of mixed oxides according to aspects of the invention

(13) FIGS. 12a-12b: the course of the potential in the first complete cycle of further mixed oxides according to aspects of the invention

(14) FIG. 13: a scanning electron micrograph of a mixed oxide according to aspects of the invention containing lithium manganese spinel.

GENERAL

(15) Analysis methods:

(16) In the following examples, the following analysis methods were used and implemented in accordance with the following instructions: a) Laser granulometry with a Malvern device b) BET surface area according to DIN 66132 c) Cerimetry (based on U. R. Kunze, Grundlagen der quantitativen Analyse, page 207, 2.sup.nd edition, Thieme Verlag, Stuttgart 1986). Initially, the MnO.sub.2, obtained from a sample of the mixed oxide by e.g. acid decomposition, with a defined excess of Mohr's salt was reacted in acidic solution and the Fe.sup.2+ that was not consumed was back titrated with Cer.sup.(IV) sulphate, whereupon the quantity of MnO.sub.2 and therefore the analytical content of the sample, consequently the average degree of oxidation, can be determined via difference calculation. d) XRD measurement

(17) Apertures (1 mm/1 mm/0.2 mm); radiation: CuK range: 10-80; increment: 0.02; measurement period: 5.5 sec/step

EXPERIMENTAL

(18) The starting products used are commercially available from the following suppliers and, unless otherwise indicated, were used as received. Manganese carbonate (MnCO.sub.3) S grade: Comilog/Erachem Lithium hydroxide monohydrate (LiOH*H.sub.2O): Acu Pharma Boric acid (H.sub.3BO.sub.3): Jkle Chemie Aluminium nitrate nonahydrate (Al(NO.sub.3).sub.3*9H.sub.2O): Tropitzsch Zinc nitrate hexahydrate (Zn(NO.sub.3).sub.2*6H.sub.2O): Plato Manganese nitrate solution (50% by weight): Coremax/Taiwan

(19) Electrodes were prepared by mixing 70% active material or mixed oxide according to aspects of the invention, 20% Super PLi conductive carbon from Timcal and 10% PTFE powder in a mortar until a floc formed. The floc was kneaded several times in the mortar before it was rolled out to a thickness of 100 m in a stainless steel roller press. Round electrode films with a diameter of 10 mm were punched out, dried overnight at 200 C. and then arranged in a Swagelok PVDF T-cell for electrochemical measurements. The counter and reference electrodes consisted of a lithium metal film and were separated by a glass wool separator film from Whatman. Merck LP30 was used as electrolyte. The charge/discharge/test cycles were recorded at a current of C/20 and a nominal specific capacity of 120 mA/h in a voltage window between 4.2 volts and 2.8 volts.

Comparison Example 1

Preparation of a Lithium Manganese Spinel of the Formula Composition Li[Mn1.87Al0.10Li0.03]2O4 with No Added Boron-Oxygen Compound

(20) The following components were provided as starting materials:

(21) TABLE-US-00001 MnCO.sub.3 93% 231.13 g Al(NO.sub.3)3*9H.sub.2O 98.50% 38.08 g HNO.sub.3 65% 70.77 g LiOH*H.sub.2O 96.30% 44.88 g H.sub.2O 100% 230 g

(22) The manganese carbonate and the aluminium nitrate nonahydrate are stirred into or dissolved in small stages in the previously introduced water with a laboratory anchor stirrer. The nitric acid is then added slowly and stirring continued for 10 min. The nitric acid is measured such that approx. of the manganese carbonate dissolves as manganese nitrate. The laboratory anchor stirrer is set to 40 rpm in order to keep the suspension floating. The powdery lithium hydroxide monohydrate is then added and the laboratory anchor stirrer set to 100 rpm for 30 min. in order to compensate for the markedly increasing viscosity. The medium-brown paste obtained is dried in the laboratory dish at 100 C. for 18 h (amount weighed in 1431 g/amount weighed out 261 g) and the brown drying product comminuted in the hand mortar. 51.1 g of the drying product was calcined in air in the ceramic crucible at 500 C. for 1 h in the chamber furnace and 32.5 g of a black powder obtained. 9.0 g of the calcination product was sintered in air in the ceramic crucible in the chamber furnace at 750 C. for 12 h, with a heating time of 6 h and a cooling time likewise of 6 h. 8.85 g of sintered product was obtained.

Comparison Example 2

Preparation of an Aluminium- or Cobalt-Doped Lithium Manganese Spinel with Lithium Hyperstoichiometry

(23) The samples are prepared by separately introducing manganese nitrate and dopant solutions (aluminium and/or cobalt nitrate) into a receiver solution of LiOH and NH.sub.3. The samples are transferred into a furnace without a washing process and predried statically at 160 C. A pre-tempering of the precursor at 500 C. follows. After cooling, the intermediate product is pestled and transformed into the finished end-product in a second tempering step at different final temperatures (690 C., 730 C. and 770 C.)

(24) Compared with the cobalt-doped product, the crystallite size of the aluminium-doped spinels is clearly smaller.

(25) A more pronounced markedness of the additional stages and lower cycle stability in the examined cycle range were shown in the case of the electrochemical behaviour of the spinels containing Co/Li. Higher cycle stability was found in the case of the Al/Li-doped sample.

(26) Spinels with different Li/Mn/Co/Al contents were synthesized. Table 1 shows the prepared spinels:

(27) TABLE-US-00002 TABLE 1 Effect of the dopants (doping elements) Al, Co and Li; the quoted values correspond to the portions of replaced Mn ions in % Sample Al Co Li G1007 0 5 1.5 G1008 0 5 1.5 G1010 5 0 1.5 G1011 5 0 1.5 G1019 0 5 0 G1020 5 0 0 G1021 0 5 1 G1022 5 0 1 G1028 2.5 2.5 1

(28) The obtained spinels were firstly characterized electrochemically:

(29) Important information about inter alfa the cycle stability can be found in the course of the potential of the charge/discharge curves. Typically, after approximately half charge/discharge, the charge and discharge curves show a potential step which is attributable to structural order effects in the spinel lattice. The more pronounced and sharper this jump is, the poorer as a rule the cycle stability of the tested material. This effect already gives an advance indication of the cycle stability of the material.

(30) The same applies for an additional potential step which can occur at 3.1 V vs. Li.

(31) The differentials of the potential curves (dQ/dE vs. E) were plotted (FIG. 1) in order to be able to better compare the markedness of the potential steps. The maxima/minima correspond to the plateaus in the charge curves. The sharper the peaks here, the more pronounced the markedness of the steps is.

(32) FIG. 1 shows that the peaks merge more markedly as the aluminium content in the structure increases, which corresponds to a blurring of the potential curves. This effect is a criterion for the structural insertion of the aluminium in the spinel structure. Electrochemically, a higher cycle stability thus results.

(33) Alongside this, an increase in potential was also established by exchanging cobalt with aluminium.

(34) The consequence of the additional insertion of hyperstoichiometric lithium (in manganese lattice sites) in the aluminium-doped sample is an even more pronounced overlapping of the peaks. No shift of the peak position is to be seen.

Example 1

Preparation of a Mixed Oxide According to Aspects of the Invention

(35) The mixed oxide according to aspects of the invention was prepared starting from a stoichiometric doped lithium manganese spinel (precursor value) obtained analogously to comparison example 1. To create the hyperstoichiometry, LiOH and, as boron-oxygen compound, boron oxide in the form of a mixture of LiOH and B.sub.2O.sub.3 were used. The molar ratio of Li to B was 2 to 1 or f=2.5).

(36) The precursor (5% Al (d=0.1) and 1.5% Li (c=0.03) doping) was mixed with different portions of LiOH and B.sub.2O.sub.3 solutions and reacted according to the scheme in FIG. 3.

(37) The addition of the boron-oxygen compound, here the borate, affects the morphology, the half-widths of the X-ray reflexes and the BET surface area of the Al/Li-doped lithium manganese spinels of the mixed oxide according to aspects of the invention.

(38) Even a small addition of borate is already enough to achieve a pronounced jump in important properties.

(39) As can be seen in FIG. 4, the crystallite size increases markedly when borate is added and when the tempering temperature increases. The largest crystallites are accordingly achieved at a temperature of 770 C. and when adding 1% by weight (preferably based on the total mixture of the mixed oxide). However, the BET surface area and, as can be seen in FIG. 5, the half-width of the XRD reflexes clearly decrease.

(40) The examined mixed oxides according to aspects of the invention were single-phase (homophase or homeotype) according to the profile of the XRD reflexes, i.e. a complete regular insertion of aluminium into the crystallite structure had taken place.

Example 2

XRD Spectrum of the Mixed Oxide According to Aspects of the Invention

(41) The XRD spectrum of the mixed oxide according to aspects of the invention (MO1 1% by weight borate and MO2 0.5% by weight borate), obtained according to Example 1, was compared with the XRD spectra of Al-doped hyperstoichiometric lithium manganese spinels obtainable according to comparison example 1 (M29 and M30) (FIG. 2).

(42) All samples were doped with Li/Al in the ratio 1:5. The aim was to test the homogeneous insertion of the dopants into the crystal lattice.

(43) In the XRD spectrum, a homogeneously doped single-phase lithium manganese spinel of the mixed oxide according to aspects of the invention shows individual symmetrical reflexes (M01 and MO2) (after K2 correction).

(44) The curves of M 29 and M 30 show clearly pronounced shoulders (arrows) (after K2 correction). This indicates a multiphase structure or an inhomogeneous distribution of the dopants.

Example 3

Electrochemistry

(45) The presence of a boron-oxygen compound, here the borate, in the mixed oxide according to aspects of the invention obtained according to Example 1 also has a positive electrochemical effect. The cycle stability increases, as can be seen in FIG. 6, as the level of borate in the mixed oxide increases compared with a pure Al-doped lithium manganese spinel.

Example 4

Manganese Solution

(46) A reason for the previous lack of cycle stability of pure doped or non-doped lithium manganese spinels was their decomposition in electrolytes. In order to examine the relative stability of the mixed oxide according to aspects of the invention compared with a decomposition of the spinel component in an electrolyte, in each case 4 g of mixed oxide powder from Example 1 was stored for four weeks at 40 C. with different quantities of borate and an Al-doped lithium manganese spinel from Example 1 in 40 g of LP30 (Merck electrolyte). The electrolyte was then analysed for dissolved manganese by means of ICP. The powders tempered at 770 C. were tested.

(47) As can be seen in FIG. 7, the examined samples showed, in the relative comparison, a decrease in the manganese dissolution as the BET surface area increased. It can be seen that a pure lithium manganese spinel without the added boron-oxygen compound shows the highest dissolution. The addition of the boron-oxygen compound thus stabilizes the lithium manganese spinel in the mixed oxide according to aspects of the invention.

Example 5

Variation of the Lithium Hyperstoichiometry

(48) In further syntheses, the lithium hyperstoichiometry of the spinel component of the mixed oxide was varied progressively with a constant Al content.

(49) Chosen dopings were Al 5% Li 1% (d=0.1; c=0.02), Al 5% Li 2% (d=0.1; c=0.04) and Al 5% Li 2.5% (d=0.1; c=0.05). Each of the samples was first calcined at 500 C. with and without 1% by weight of an LiOH/B.sub.2O.sub.3 mixture (b=0.002; f=2.5) as described and then tempered at 690 C., 730 C. and 770 C.

(50) For the mixed oxides according to aspects of the invention, a clear relationship between reflex half-widths and lithium hyperstoichiometry was able to be recorded. As the lithium hyperstoichiometry increased, the reflexes narrowed.

(51) The BET surface area was then correlated with the lithium hyperstoichiometry:

(52) Due to the influence of the borate, there was a levelling of the scatterings of the reflex half-widths depending on the lithium hyperstoichiometry which occurred in samples from different syntheses (see FIG. 2). The same phenomenon was observed when measuring the BET surface area of the samples. Here too, a levelling was to be recorded in the presence of borate (FIG. 8).

(53) The BET values for the samples from the variation of the lithium hyperstoichiometry are summarized in Table 2.

(54) TABLE-US-00003 TABLE 2 BET surface area: Li-hyperstoichiometry variation Li Sample BET m.sup.2/g T C. (mol % on Mn) G1051B1 0.62 770 1.0 G1051B1 1.07 730 1.0 G1051B1 2.5 690 1.0 G1031B2 0.6 770 1.5 G1031B2 2.2 690 1.5 G1052 4.91 770 2.0 G1052B1 0.31 770 2.0 G1052B1 0.92 730 2.0 G1052B1 1.66 690 2.0 G1053 4.85 770 2.5 G1053B1 0.4 770 2.5 G1053B1 1.04 730 2.5 G1053B1 1.7 690 2.5

(55) Overall, as well as the addition of borate and the temperature, the lithium hyperstoichiometry also affects the BET surface area. BET values <1 m.sup.2/g were found.

Example 6

Cerimetry

(56) The fact that hyperstoichiometric lithium is inserted into the spinel lattice was demonstrated by means of cerimetry using a mixed oxide according to aspects of the invention with aluminium-doped manganese spinel. The insertion into the spinel lattice changes the average degree of oxidation of the manganese ions. Were the lithium present only as fluxing agent (in the form of Li.sub.2O), the average degree of manganese ion oxidation would be +3.56, in the case of a structural insertion, +3.62.

(57) The two structural borderline cases can be represented as follows:

(58) TABLE-US-00004 Pure fluxing agent Li.sub.1[Li.sup.+1.sub.0.029Al.sup.+3.sub.0.098Mn.sup.+x.sub.1.874]O.sub.4Mn.sup.+3.56 0.071Li.sub.2O*wB.sub.2O.sub.3 Structural lithium insertion Li.sub.1[Li.sup.+1.sub.0.067Al.sup.+3.sub.0.094Mn.sup.+x.sub.1.839]O.sub.4Mn.sup.+3.62 wB.sub.2O.sub.3

(59) The obtained values which are listed in Table 3 lay in the range from +3.59 to +3.69, which proves the structural insertion.

(60) TABLE-US-00005 TABLE 3 Average degree of oxidation, measured by means of cerimetry, of the manganese ions in the mixed oxide Li[LiAlMn]O.sub.4B.sub.2O.sub.3. Temp. Stoichiometry M.sup.X+ Sample C. Li Li Al Mn cerimetric STDDEV+/ G1031 B2 500 1 0.081 0.095 1.824 3.69 0.04 G1031 B2 T1 700 1 0.067 0.094 1.839 3.59 0.01 G1031 B2 T2 730 1 0.070 0.095 1.835 3.63 0.01 G1031 B2 T3 690 1 0.071 0.094 1.835 3.59 0.01 G1051 B1 T2 730 1 0.076 0.092 1.832 3.60 0.01 G1053 T1 770 1 0.055 0.094 1.851 3.63 0.01 G1053 B1 T1 770 1 0.100 0.093 1.807 3.62 0.02 G1053 T2 730 1 0.053 0.093 1.854 3.59 0.00 G1053 B1 T2 730 1 0.097 0.092 1.811 3.60 0.01

(61) The degree of oxidation achieved depends on the synthesis temperature. At higher temperatures, the degree of oxidation was somewhat lower. This finding correlates well with the above-described structural temperature-related shift of the reflex layers in the XRD spectrum.

(62) In FIG. 9, dQ/dE vs. E of materials of different mixed oxides according to aspects of the invention with different lithium hyperstoichiometries are plotted.

Example 7

Electrochemical Examination

(63) FIG. 10 compares the discharge capacities of samples of mixed oxides according to aspects of the invention in the first cycle. The theoretical value was calculated from the analytically determined spinel stoichiometry. The measured values lie between 117 mAh/g for a purely lithium-doped (hyperstoichiometric) sample (EXM 1666) and 96 mAh/g for one doped with Li/Al/Zn.

(64) The theoretical maximum capacity decreases from 129 mAh/g for EXM 1663 to 109 mAh/g for EXM1666. The measured values show the same tendency at a lower level.

(65) The reference samples showed comparatively small capacities, which can be attributed to the increased lithium insertion.

(66) FIG. 11 shows in the overview the courses of the potential of the reference samples in the first complete cycle. In FIG. 12, the emphasis is on potential ranges which gave initial indications of the cycle behaviour of the samples. Both the occurrence of an additional potential step at 3.2 V and a well-pronounced step at half charge/discharge are considered a feature of poor cycle stability: The potential plateau at 3.2 V (FIG. 12a, left-hand side) is clearly pronounced for the Al-free sample, still difficult to spot for the sample doped with 3% Al and no longer present in the models doped with 5% Al. The markedness of the stage of potential at half charge/discharge decreases in the order EXM1663 after EXM1666.

Example 8

Industrial-Scale Preparation of a Mixed Oxide Containing Lithium Manganese Spinel According to Aspects of the Invention of the Composition Formula (Li0.99Zn0.01)[Mn1.87Al0.1Li0.03]O4.(0.9 B2O3, 2*0.9Li2O

(67) In each of several stirring mixtures, the following quantities of raw materials were used:

(68) TABLE-US-00006 MnCO.sub.3 4.360 kg 93.50% Mn(NO.sub.3).sub.2 sol. 2.843 kg 50% Al(NO.sub.3).sub.3 * 9H.sub.2O 0.867 kg 101% Zn(NO.sub.3).sub.2 * 6H.sub.2O 70.2 g 99% H.sub.2O (dist.) 3.530 kg LiOH * H.sub.2O 1.063 kg 55% Boric acid 12.80 g 100.3%

(69) The water, the manganese nitrate solution, the aluminium nitrate nonahydrate and the zinc nitrate hexahydrate were each placed in a 121 plastic bucket and mixed with a Pendraulik stirrer with stirring disk and dissolved completely. The finely-powdered boric acid and the manganese carbonate pre-ground to a D95 value of 27 m were then added. The Pendraulik stirrer was first set to level 3 in order to keep the still highly liquid suspension homogeneous. The powdery lithium hydroxide was then added within 1 min. and stirred at level 4 for a further 15-20 min. The viscosity of the suspension increased rapidly.

(70) The stirring mixtures were combined in a stirred receiver container and the still pumpable suspension was dried in a Storck Bowen spray drier with two-fluid nozzle and spraying from above in the cocurrent process at an air-entry temperature of 550 C. and an air-exit temperature between 140 C. and 145 C. A brown powder with a bulk density of 1011 g/l was obtained.

(71) The drying product obtained in this way still contains nitrates and was calcined on stainless steel tray sheets in a continuous belt furnace with an attached exhaust-gas cleaning unit for nitrous vitriol gases at 450 C. and an average residence time of 1 h in the heating zone. Surprisingly, small quantities of nitrous vitriol gases formed and a fine black powder with a bulk density of 823 g/l was obtained. The calcination product was then tempered in air in an Alsint-ceramic crucible in the chamber furnace at 770 C. for 12 h with a heating time of 6 h and a cooling time of 12 h.

(72) The bluish black tempering product was ground on an Alpine AFG100 air separator mill with ceramic separator wheel and ceramic milling air nozzles 3.0 mm in diameter at a separator speed of 5500 rpm.

(73) The milled product collected in the centrifugal cyclone separator had a bulk density of 1045 g/l and an apparent density of 1713 g/l. FIG. 13 shows a scanning electron micrograph of the product. The chemical composition determined by means of ICP was:

(74) TABLE-US-00007 Li 3.9% Mn 59.7% Al 1.6% Zn 3800 mg/kg B 600 mg/kg S 0.23% Na 920 mg/kg

(75) The particle-size distribution was monomodal and characterized by the following parameters: D.sub.10=2.5 m D.sub.50=10.8 m D.sub.90=20.7 m D.sub.100=35.6 m

(76) Further product properties were: pH value: 9.5 residual moisture content: 0.26% by weight (Karl-Fischer method) BET surface area: <1 m.sup.2/g micropore volume: <0.001 cm.sup.3/g (ASAP 2010, Micromeritics) cubic lattice constant a=8.210 (according to X-ray diffraction)

(77) In the electrochemical cycle test, the mixed oxide containing lithium manganese spinel according to aspects of the invention shows a discharge capacity of 105 mAh/g and a cycle loss of less than 0.1% per cycle measured in a half-cell of the LMS//LiPF6-EC-DMC//Li type.