AU2002317657A1

AU2002317657A1 - High performance lithium titanium spinel L14T15012 for electrode material

Info

Publication number: AU2002317657A1
Application number: AU2002317657A
Authority: AU
Inventors: Michael Graetzel; Ladislav Kavan; Jan Prochazka; Timothy Spitler; Francois Sugnaux
Original assignee: XOLIOX SA; Altair Nanomaterials Inc
Current assignee: XOLIOX SA; Altairnano Inc
Priority date: 2001-07-31
Filing date: 2002-07-29
Publication date: 2003-05-29
Anticipated expiration: 2022-07-29

Description

HIGH PERFORMANCE LITHIUM TITANIUM SPINEL L14T15012 FOR ELECTRODE MATERIAL

Field of the invention

The invention relates to a Lithium Titanate Spinel oxide material, more precisely a Li₄Ti_δOi₂ material, which may be used in energy storage devices and more specifically in cathodes and anodes of Lithium Ion batteries or of hybrid capacitance devices.

The invention also concerns a process which may be used for obtaining a Li₄Ti₅Oi₂ spinel material.

State of the art

Spinel oxides Li_1+xTi₂-_xO₄; 0< x <1/3 (space group Fd3 m) were described in 1971 and electrochemically characterized in the early 1990s (K. M: Colbow, J. R. Dahn and R. K Haering, J. Power Sources, 26, 397 (1989) and T. Ohzu u, A. Ueda and N. Yamamoto, J. Electrochem. Soc, 142, 1431 (1995)).The end members of the series, i.e. LiTi₂0₄ and Li_4/3Ti₅/₃O (Li₄Ti₅O₁₂) are metallic (super-conducting below 11 K) and semi-conducting, respectively (M.R. Harrison, P. P. Edwards and J. B. Goodenough, Phil Mag. B, 52, 679 (1985)). Both materials exhibit similar Li-insertion electrochemistry, the formal potential of Li-insertion being (1.36-1.338) V for LiTi₂0₄ and (1.55-1.562) V for Li₄Ti₅Oι₂, respectively (S. I. Pyun, S. W. Kim and H. C. Shin, J. Power Sources, 81-82, 248 (1999)).

In principle, therefore, this latter material can be coupled with a 4V electrode, as LiMn₂O₄ or LiCoO₂, to provide a cell with an operating voltage of approximately 2.5V, which is twice that of nickel-cadmium or nickel-metal hydride cells. Li₄Ti_δOi₂ accommodates Li with a theoretical capacity of 175 mAh/g (.based on the mass of the starting host material) according to the equation:

(Li)^8a(Li_1/3, Ti_5/3)^16dO₄ ^32e + β + Li⁺ (Li₂)^16c(Li_1/3, Ti_5/3)^16dO₄ ^32e (1)

where the superscripts stand for the number of equivalent sites with Wyckoff symbols for the space group Fd3m. Hence, Li⁺ occupies tetrahedral (8a) and octahedral (16c, 16d) sites of the lattice, and the overall insertion capacity is controlled by the number of free octahedral sites. A more detailed analysis points at two-phase equilibrium, which explains the invariance of the electrode potential on the electrode composition. The spinel host structure accommodates Li⁺ without significant changes of lattice constants. Consequently, these materials show excellent cycle life and the Li⁺

-8 diffusion coefficient of about 2.10 cm /s was reported (K. Zaghib, M. Simoneau, M. Armand and M. Gauthier, J. Power Sources, 81-82, 300 (1999)). For an entire battery, both anode and cathode active materials, the charge/discharge cycle may be simply represented by the following equation.

charge Li₄Ti₅O₁₂ + 6LiCoO₂ «--» Li₇Ti₅O₁₂ + 6Li₀.₅CoO₂ E = 2.1V [la] discharge

In previous communications the spinel Li₄Ti₅Oι₂ was prepared by a solid-state reaction of stoichiometric amounts of TiO₂ and Li₂CO₃ or LiOH; the reaction typically occurs within 12-24 hours at 800-1000°C.

A LJ₄Ti₅0i₂ material with smaller particle size was prepared by high-energy ball milling of the conventional microcrystalline spinel . The product exhibited particles around 600 nm in size, but its electrochemical performance was not significantly different from that of the non-milled starting material . The above-mentioned Li₄TisOi2 materials suffer from relatively low Li-insertion capacity at high charging rates. Therefore, there is a need for Li₄Ti₅Oi₂ electrode materials with improved electrochemical performance.

Summary of the invention

The inventors unexpectedly found that Li₄TiδOi₂ material made of nano-sized particles which - according to the nitrogen absorption surface area measurement method of Brunauer-Emmet-Teller (BET method) - have a BET surface area of at least 10 m²/g (i.e. corresponding to particles having a theoretical size of less than 100 nm) exhibit different Li- insertion electrochemistry and show specific electrochemical performances.

In one embodiment of the invention the particles are characterized by a BET surface of between 10and 200 m²/g.

In another embodiment the particles are characterized by a BET surface area of between 20 and 160 m²/g .

In another embodiment the particles have a BET surface between 30 and 140 m²/g.

In a preferred embodiment the particles are characterized by a BET surface area of between 70 and 110 m²/g.

For producing the above cited nano-sized particles the inventors developed new synthetic methods.

The conversion of nanocrystalline Tiθ₂ (anatase) towards Li₄Ti₅Oi₂ was first explored by a reaction of colloidal Tiθ₂ with LiOH. However, this strategy was not successful, neither its variants employing Li₂CO3, LiCH₃COO and LiNO₃ in combination with the stoichiometric amount of colloidal anatase in acidic or alkaline media at temperatures up to 250°C (in autoclave). In all cases, the product contained Liι_+xTi₂-χθ₄ with considerable amounts of unreacted anatase.

The inventors solved the problem by developing a method comprising a step of mixing an organo-lithium compound selected from lithium alcoholates with an organo-titanium compound selected from titanic acid esters in an organic solvent and a step of hydrolyzing said mixture. Particularly preferred alkoxides as starting reagents are Li-ethoxide and Ti(IV)isopropoxide and Ti(IV) n-butoxide. Preferably, the organo-lithium compound and the organo-titanium compound are mixed in a stoichiometric molar ratio substantially equal to 4:5. For some contemplated applications, defined mixtures of anatase and lithium titanate spinel are desirable. These may be obtained by appropriate ratios differing from the above molar ratio.

After hydrolysis, isolation and drying of the precipitate, spinel products were obtained which exhibit BET surface area values of at least 5 m²/g, and generally above 10 m²/g, which correspond to much smaller particle sizes than the particles sizes of state of the art microcrystalline Li₄Ti₅Oι₂ materials. However, after alkoxide hydrolysis, the slurry still contains appreciable mounts of unreacted anatase. Therefore, a preferred embodiment of the process according to the invention further comprises the steps of processing the hydrolyzed mixture with a polymer like polyethyleneglycol (PEG) up to homogeneity, and submitting the homogenized product to a heat treatment effective for removing organic material there from. This polymer is known to form complexes with lithium and oxo-titanium species, while it may also organizes the inorganic structure by supramolecular templating (L. Kavan, J. Rathousky, M. Gratzel, V. Shklover and A. Zukal, J. Phys. Chem. B, 104, 12012 (2000)). After processing the hydrolyzed mixture with PEG and removing the same by annealing, pure spinel materials could be obtained. Materials exhibiting unexpected extremely high BET surface area values of more than 80 m²/g may be obtained. An additional object of the invention is an electrode comprising a nano-structured Li₄Ti₅Oi₂ material exhibiting BET values as above. The PEG processing step and the annealing/sintering step may be conducted to obtain BET values which may range between 50 m²/g and 200 m²/g, corresponding then to values well suitable for electrodes.

A preferred object of the invention is a thin film electrode obtained by coating a conductive support with a hydrolyzed mixture produced by a process as defined above and submitting said coated support to a heat treatment. A particularly preferred thin film electrode is obtained by coating a conductive support with the homogenized product produced by processing the hydrolyzed mixture with PEG, and submitting said coated support to an annealing treatment.

Such annealing treatments may be carried out at 400-500°C, that is to say at much lower temperatures than the solid state spinel preparations of the prior art.

Thus, the invention provides electroactive ion-insertion materials based on nanostructured, tetra-Lithium Titanate spinel allowing extremely high charge and discharge rates, a high number of charge, discharge cycles, Mesoporous electrode materials thereof and processes to produce these materials, including precipitation from a solution, doping with metallic atoms, nano-templating, Microparticles agglomeration, spray-drying, ball-milling and sintering.

This invention also provides electrodes, i.e. anode or cathode, based on nanostructured lithium titanate spinel, and their manufacturing process to build these as rigid films made from nanoparticles of the electroactive material or as flexible layers made from Mesoporous microparticles of the electroactive nanostructured material. Such Microparticles may be obtained from agglomerated and sintered lithium titanate spinel precipitated particles, which are spray dried, processed with an organic binder and coated onto a conductive substrate. It should also be noted that the invention is not limited to a specific process such as the one previously discussed. Several other processes may be used and in particular the Altair processes disclosed in US patent applications 60/306,683 and 60/362,723.

Other features and advantages of the process and products according to the invention will appear to those skilled in the art from the following detailed description and related non limitative examples.

Brief description of the drawings

- Figure 1 shows three X-rays diffractograms of a material according to the invention obtained with different processes.

- Figure 2 shows a cyclic voltammogram of the material (A) shown in figure 1. - Figure 3 shows galvanostatic charging/discharging cycles corresponding to different materials.

- Figure 4 shows 50C rate data corresponding to a material according to the invention.

- Figure 5 shows 100C rate data corresponding to a material according to the invention.

- Figure 6 shows 150C rate data corresponding to a material according to the invention.

- Figure 7 shows 200C rate data corresponding to a material according to the invention. - Figure 8 shows 250C rate data corresponding to a material according to the invention.

Figure 1 shows a powder extract X-ray diffractogram of Li₄TisOi₂ prepared from

Ti(IV) butoxide + Li ethoxide; (A) Material synthesized by the procedure using PEG (BET surface area 183 m²/g); (B) Material prepared as in (A) except the addition of PEG was omitted; (C) Material prepared as in (A) but with the aid of hydrothermal growth of particles (150°C, 10 hours, BET surface area, 119 m²/g). The curves are offset for clarity, but the intensity scale is identical for all three plots.

Figure 2 shows cyclic voltammogram of Li₄TisOi₂ prepared from Ti(IV) butoxide-Li ethoxide. Electrolyte solution: 1 M LiN(CF₃SO2)2 + EC/DME (1 :1 by mass); scan rate 0.2 mV/s. Dashed curve displays the same plot, but with the current-scale expanded by a factor of I0.

Figure 3 shows a chronopotentiometric plot of (A) Li₄Ti₅O₁ prepared from Ti(IV) butoxide-Li ethoxide compared to (B) commercial Li₄Ti₅Oi₂ spinel (LT-2 from Titan Kogyo Japan). Electrolyte solution: I M LiN(CF₃SO₂)2 + EC/DME (1:1 by mass). The current i was adjusted to charging rate of 2C, 50C, 100C, 150C, 200C and 250C for solid curves from top to bottom. Dashed curves display the corresponding galvanostatic discharging at the same rates. For the sake of clarity, the time (t) is multiplied by the absolute value of charging/discharging current i.

Under Ar-atmosphere, 1 ,4 g (0.2 mol) of lithium metal (Aldrich) was dissolved in 110 ml of absolute ethanol and mixed with 71 g (0.25 mol) of titanium (IV) isopropoxide (pract. Fluka) or, alternatively, with 85 g (0.25 mol) of titanium (IV) n-butoxide (pract. Fluka). Still another alternative consisted in using lithium ethoxide or lithium methoxide powders from Aldrich.s 50 mL of the solution of Li + Ti alkoxides was hydrolyzed in 300 mL of water, and the produced slurry was concentrated on rotary evaporator (40°C 20 mbar) to a concentration of 10-20 wt%. Polyethylene glycol (molecular weight 20 000, Merck) was added in the proportion of 50-100 % of the weight of Li₄Ti₅Oι₂, and the mixture was stirred overnight. The resulting viscous liquid was deposited on a sheet of conducting glass (F-doped Snθ2, TEC 8 from Libbey- Owens-Ford, 8 Ω/square) using a doctor-blading technique (L. Kavan, M. Gratzel. J. Rathousky and A. Zukal, J. Electrochem. Soc, 143, 394 (1996)) and finally annealed at 500°C for 30 min. Sometimes, the slurry was homogenized using a titanium ultrasonic horn (Bioblock Scientific; 80 W, 30 x 2s pulses) before deposition. The mass of active electrode material was typically 0.1-0.3 mg/cm²; the projected electrode area 1 cm² and the layer thickness about 2-6 μm. For comparison, an analogous electrode was prepared from commercial Li₄TisOi₂ (LT-2 from Titan Kogyo Japan). The material had a BET surface of 2.9 m²/g (manufacturer's specification; 3.1 m²/g by own measurement). The LT-2 powder was dispersed by mortaring with acetylacetone, and the paste for doctor-blading was prepared by addition of hydroxypropylcelulose and Triton X-100 as described elsewhere (L. Kavan, M. Gratzel. J. Rathousky and A. Zukal, J. Electrochem. Soc, 143, 394 (1996))

The BET surface areas of the prepared materials were determined from nitrogen adsorption isotherms at 77 K (ASAP 2010, Micromeritics). The film thickness was measured with an Alpha-step profilometer (Tencor Instruments). Powder X-ray diffractometry (XRD) was studied on a Siemens D-5000 difractometer using CuKα radiation. The samples for BET and XRD were obtained by mechanical scraping of the film from a glass support.

The BET surface areas of the as-prepared materials were 105 m²/g (synthesis employing Ti(IV) isopropoxide) and 153-196 m²/g(synthesis employing Ti(IV) butoxide), respectively. If the slurry was autoclaved at 150°C for 10 hours, the surface areas decreased to 53 m²/g (isopropoxide-synthesis) or 119 m²/g (butoxide- synthesis), which is due to hydrothermal particle growth by Ostwald ripening.

Fig l/(A) shows the X-ray diffractogram of a material resulting from the butoxide- synthesis (surface area 183 m²/g). All peaks can be indexed as Li₄TisOi2. The crystal size (d_c) can be estimated from the X-ray line width (w) (Scherrer formula):

d_c = 0.9 λ/wcosθ (2) (λ is the X-ray wavelength (0.1540562 nm) and θ is the diffraction angle). Eq. (2) gives d_c about 4-5 nm. This value roughly matches the particle size (d_p) estimated from BET area (S=183 m²/g). Assuming spherical particles, the value of d_p can be approximated as:

d_p = 6/Sp (3)

which gives d_p « 9 nm for S=183 m²/g and p = 3.5 g/cm³. Analogous evaluation routine for the ex-propoxide material produces d_c « 15 nm and d_p « 19 nm. Fig. 1/(B) shows the XRD plot for a material prepared as that in Fig. 1 /(A) except that the addition of polyethylene glycol was omitted. In this case, anatase is clearly distinguished at 2Θ » 25 deg. Fig 1/(C) displays the XRD plot for a material prepared as that in Fig. 1/(A), but the particles were grown hydrothermally (the product's surface area was 119 m²/g, d_c « d_p « 14 nm). The lattice constant of the hydrothermally grown material (Fig. 1/(C)) equals 0.8366 nm, which is in good agreement with the lattice constant of LUTi₅0i₂ made by the conventional high- temperature synthesis: 0.8367, 0.8365 and 0.8358. However, the lattice constant of the nanocrystalline materials (without hydrothermal growth, cf. Fig. I/(A)) is significantly smaller. The actual values fluctuate between 0.8297 nm to 0.8340 nm for various samples, both from the butoxide-and isopropoxide-synthesis.

Whereas the lattice constant, a, of a Liι_+xTi₂-_xO₄ spinel is known to decrease with x (ref.) (M.R. Harrison, P. P. Edwards and J. B. Goodenough, Phil Mag. B, 52, 679 (1985)) according to the relation:

α = 0.8405 - 0.0143* (4)

this reasoning cannot account for the observed decrease of a for the nanocrystalline material according to the invention. The latter conclusion is supported by two arguments: (i) the lattice constant attains its "normal" value after hydrothermal growth and (ii) the nanocrystalline material according to the invention shows the electrochemistry of a Li-rich spinel, vide infra. It may be noticed that a 5 nm-sized particle of Li₄Ti₅Oi2 contains about 200 unit cells only, and such a small particle exhibits marked lattice-shrinking.

Electrochemical measurements were carried out in a one-compartment cell using an Autolab Pgstat-20 controlled by GPES-4 software. The reference and auxiliary electrodes were from Li metal, hence potentials are referred to the Li/Li⁺ (1 M) reference electrode. LiN(CF₃SO₂)2 (Fluorad HQ 115 from 3M) was dried at 130°C/1 mPa. Ethylene carbonate (EC) and 1 ,2-dimethoxyethane (DME) were dried over the 4A molecular sieve (Union Carbide). The electrolyte solution, 1 M LiN(CF₃SO₂)₂ + EC/DME (1/1 by mass) contained 10-15 ppm H₂O as determined by Karl Fischer titration (Metrohm 684 coulometer). All operations were carried out in a glove box.

Cyclic voltammogram of the ex-butoxide material (A) evidences the Li-insertion into Li Ti₅Oi₂ spinel (Fig. 2). The formal potential of insertion equals 1.56 V vs. Li/Li⁺, which matches the potential of ordinary microcrystalline Li₄TisOi₂. The small peaks at 1.75 and 2.0 V can be assigned to anatase . Assuming the insertion ratio Li/Tiθ₂ (anatase) = 0.5, the integral peak area corresponds to the anatase content below 1 %. In most samples, the anatase content was between 0.3 - 0.6 %, and, sometimes, it was even not detectable. Note that the Li-insertion electrochemistry can serve as a very sensitive analytical method for the Li₄Ti₅Oι₂-TiO₂ mixture, which is superior to XRD (cf. Fig. 1 /(A) and Fig 2).

Fig. 3A displays a series of galvanostatic charging/discharging cycles of the same nanocrystalline electrode as in Fig.2 at relatively very high charging rates: 2C, 50C, 100C, 150C, 200C and 250C. The maximum reversible Li-insertion capacity is 160 mC, and about 70% of this charge can still be cycled at 250C with the same cut-off voltage. The commercial microcrystalline Li4TisOi2 shows only ca 19 % of its nominal capacity in the 250C-cycle (Fig. 3B). Fast charging is always reversible, but the nanocrystalline electrode shows considerable irreversibility at 2C. This is apparently due to breakdown processes in the electrolyte solution, such as reduction of trace water, which is more pronounced at high-surface area electrodes. Within the experimental error of weighing of the electroactive film, the microcrystalline (LT-2) electrode gives the theoretical maximum insertion capacity (Eq 1 ), i.e. 630 C/g. The nanocrystalline electrodes usually showed capacities of ca. 550-610 C/g.

Details of the discovery of the particle size requirements for high rate charging of tetra-lithium titanate (spinel):

A unique set of various size samples of tetra-Lithium Titanate was collected for study by the inventors. This unique set of samples ranged in surface area from approximately 1 m²/gm to approximately 200m²/gm. The charging characteristics of tetra-Lithium Titanate (spinel) was determined as a function of Surface Area of the active electrode powder. As a result of this study the charging performance of tetra- Lithium Titanate as a function of the surface area of the active electrode grade powder was discovered.

The unique and comprehensive set of samples was collected from 3-sources:

1. Commercial sources, that is companies who make and sell tetra-Lithium Titanate for use primarily in button cells (i.e. Titan Koygo), were used to procure some of the samples.

2. Altair NanoMaterials provided samples using its new process for producing particulate Li4Ti₅0₁₂. We refer to USA Provisional Patent Application 60/306,683, dated July 20, 2001 and assigned to Altair NanoMaterials Inc. by the inventors Timothy M. Spitler and Jan Prochazka as a source of some of the materials.

3. Samples were produced and included the unique set of samples discussed in the previous chapters.

Preparation of electrodes used in measuring the charging characteristics of particulate tetra-Lithium Titanate samples acquired from commercial sources (item 1. above) and produced using the Altair NanoMaterials process (Item 2. above) proceeded as follows.

The Li₄Ti_δOi₂ powder was dispersed in aqueous medium to form a viscous paste. The powder (1.0 g) was ground for at least 20 min in an agate or porcelain mortar under slow addition of 4 x 0.2 mL of 10 % aqueous solution of acetylacetone. The mixture was diluted with 5 ml H₂O and mixed with 2 mL of 4 % aqueous solution of hydroxypropylcellulose (MW 100,000) and 2 mL of 10 % aqueous solution of Triton- X100 (Fluka). The resulting viscous liquid was stirred overnight before use. When necessary, the mixture was further homogenized using a titanium ultrasonic horn (Bioblock Scientific; 80 W, 30 x 2s pulses). The obtained paste was deposited on a sheet of conducting glass (F-doped Snθ₂, 8 Ω/square) using a doctor-blading technique. The sheet of conducting glass had dimensions: 3 x 5 x 0.3 cm³. A Scotch- tape at both edges of the support (0.5 cm) defined the film's thickness and left part of the support uncovered for electrical contact. The film was finally calcined for 30 min in air at 500°C. After cooling down to room temperature, the sheet was cut into ten electrodes 1.5 x 1 cm² in size; the geometric area of the Tiθ₂ film was 1 x 1 cm². The as-deposited films were controlled by optical microscope, by a simple scratch test (surface scratched by a piece of glass to check the sintering of particles) and by an alpha-step profilometer. The latter method provided information about layer thickness and surface corrugation. The film's mass was determined after scraping the Tiθ₂ layer from the SnO₂(F) support by a piece of glass sheet. The layer thickness was about 1-5 μm. No signs of electrode "aging" (cracking, delaminating) were found, even after many repeated tests of the same electrode. The mass of active electrode material was typically 0.1-0.4 mg/cm²; the projected area was 1 cm².

It is to be noted that when Li₄TisOi₂ was produced as a slurry, this slurry was mixed with hydroxypropylcellulose and Triton-X in the same proportions as with the powder samples. The electrodes were fabricated and tested by the same methods as mentioned above. Electrochemical measurements of electrodes:

Electrochemical measurements were carried out in a one-compartment cell using an Autolab Pgstat-20 (Ecochemie) controlled by GPES-4 software. The reference and auxiliary electrodes were made of Li metal, hence, potentials are referred to the Li/ϋ⁺ (1M) reference electrode. LiN(CF₃SO₂)2 was dried at 130°C/1 mPa. Ethylene carbonate (EC) and 1 ,2-dimethoxyethane (DME) were dried over the 4A molecular sieve. The electrolyte solution, 1 M LiN(CF3SO2)2 + EC/DME (1/1 by mass) contained 10-15 ppm H₂O as determined by Karl Fischer titration. All operations were carried out in a glove box.

A total of 25 electrodes were prepared and tested according to the method of preparation and the experimental setup given above. XRD analysis of these samples showed that they are formed of pure Li₄TisOi₂ with less than 1 % free TiO2 in the rutile or anatase phase.

For each example, galvanostatic chronopotentiometry curves at different charging and discharging rates were measured.

The results obtained with all the samples are summarized in Table 1 , the raw data table. Separate measurements confirm that up to a charging rate of 2C (such that complete charging would be completed in % h; a charging rate of 1 C corresponds to full charge in 1 h), all samples exhibit the same maximum charge quantity. This number is considered as the full capacity of the given sample. Table 1 shows the specific surface area as well as the charging capacity (mC, milli Coulombs) measured for each example. Please note that as the actual mass of tetra-Lithium Titanate (spinel) varied on each electrode, so did the actual charge /discharge currents at the given charging rates.

Table 1 , the raw data table, was processed to produce Table 2, which normalizes the data set to 100% charge capacity at 2C_rate. Table 2 expresses observed capacities as a % of the capacity of sample at 2C. A number smaller than 100% corresponds to a loss of capacity at higher charging rate.

The results of Table 2, the processed data table, are plotted on a logarithmic scale in Figs 4 to 8 each figure presenting the data at a particular rate of charging using the C_rate system. Samples that did not maintain charge capacity of at least 80% of the full charge capacity (the 2C_rate charge) were viewed as materials that failed the testing.

Some general trends are apparent. Clearly, as the charging rate is increased, the charging capacity decreases for the particles with small surface area, while the charging capacity of the particles with large surface area is substantially maintained. As the production of the data set Table 1 is a substantial art, the data is necessarily "noisy". Despite the "noise" of the measurement process some clear performance plateaus are visible. These performance plateaus provide sufficient guidance for a manufacturer to design and produce tetra-Lithium Titanate that conforms to "guaranteed" performance requirements.

Example 1a: The 50C_rate data set as plotted in Figure 4 and contained in Table 1.

These data show that particulate tetra-Lithium Titanate qualifies for use in anode or cathode service in energy storage devices based on Li+ ion electron pair insertion/desertion cycles that function at a 50C_rate. Using a pass/fail test based on equal to or greater that 90% working charge capacity passes and less than 90% fails, the test data may be screened to produce two subsets. This screening process is shown graphically in Figure 4. On an individual sample basis a qualified performance range was determined, that range being a surface area of equal to or greater than 10m²/g to equal to or less than 200m²/g. Example 2a: The 100C_rate data set as plotted in Figure 5 and contained in Table 1.

These data show that particulate tetra-Lithium Titanate qualifies for use in anode or cathode service in energy storage devices based on Li⁺ ion electron pair insertion/desertion cycles that function at a 100C_rate. Using a pass/fail test based on equal to or greater that 90% working charge capacity passes and less than 90% fails, the test data may be screened to produce two subsets. This screening process is shown graphically in Figure 5. On an individual sample basis a qualified performance range was determined, that range being a surface area of equal to or greater than 20m²/g to equal to or less than 160 m²/g.

Example 3a: The 150C_rate data set as plotted in Figure 6 and contained in Table 1.

These data show that particulate tetra-Lithium Titanate qualifies for use in anode or cathode service in energy storage devices based on Li⁺ ion electron pair insertion/desertion cycles that function at a 150C_rate. Using a pass/fail test based on equal to or greater that 80% working charge capacity passes and less than 80% fails, the test data may be screened to produce two subsets. This screening process is shown graphically in Figure 6. On an individual sample basis a qualified performance range was determined, that range being a surface area of equal to or greater than 30m²/g to equal to or less than 140 m²/g.

Example 4a: The 200C_rate data set as plotted in Figure 7 and contained in Table 1.

These data show that particulate tetra-Lithium Titanate qualifies for use in anode or cathode service in energy storage devices based on Li+ ion electron pair insertion/desertion cycles that function at a 200C_rate. Using a pass/fail test based on equal to or greater that 80% working charge capacity passes and less than 80% fails, the test data may be screened to produce two subsets. This screening process is shown graphically in Figure 7. On an individual sample basis a qualified performance range was determined, that range being a surface area of equal to or greater than 30 m²/g to equal to or less than 120 m²/g.

Example 5a: The 250C_rate data set as plotted in Figure 8 and contained in Table 1.

These data show that particulate tetra-Lithium Titanate qualifies for use in anode or cathode service in energy storage devices based on Li+ ion electron pair insertion/desertion cycles that function at a 250C_rate. Using a pass/fail test based on equal to or greater that 80% working charge capacity passes and less than 80% fails, the test data may be screened to produce two subsets. This screening process is shown graphically in Figure 8. On an individual sample basis a qualified performance range was not determined. The 250C_rate data set appears as a "mountain" in figure 8. Graphically, a performance peak was determined at a BET-SA of 90 m²/g.

One of the preferred BET-SA of particulate tetra-Lithium Titanate has been discovered by the inventors to be at least 70m²/g but not more than 110m²/g.

The above examples show that the charging capacity corresponding to a given rate of charge increases with increasing surface area, reaches a plateau, then decreases as the surface area increases further.

Thereby (Examples 1a through 5a), the excellent electrochemical performance of nanocrystalline Li₄Ti₅Oι₂ is clearly demonstrated. Furthermore, it is to be noted that several different processes may be used to manufacture the particles according to invention.

Claims

1. A process for producing a spinel compound of formula U4Ti₅θi₂, comprising a step of preparing a mixture of an organo-lithium compound selected from lithium alcoholates with an organo-titanium compound selected from titanic acid esters, in a liquid medium, and a step of hydrolyzing the mixture of said compounds.

2. A process as claimed in claim 1 , characterized in that said organo-lithium compound is Li-ethoxide or Li-methoxide and said organo-titanium compound is selected from Ti(IV) isopropoxide and Ti(IV) n-butoxide.

3. A process as claimed in claim 1 or 2, characterized in that said organo-lithium compound and said organo-titanium compound are mixed in a molar ratio of substantially 4:5.

4. A process as claimed in any one of claims 1-3, characterized in that it further comprises a step of processing the hydrolyzed mixture with a polymer, in particular, polyethylene-glycol (PEG), up to homogeneity and a step of submitting the homogenized product to a heat treatment effective for removing organic material there from.

5. A process as claimed in claim 4, further comprising a hydrothermal growth step.

6. A process for manufacturing a thin film electrode wherein a conductive support is coated with a hydrolyzed mixture obtained according to the a process as claimed in any one of claims 1-3, the resulting coated support being then heated.

7. A process for manufacturing a thin film electrode wherein a conductive support is coated with a homogenized product obtained according to the process of claim 4, the resulting coated support being then submitted to an annealing treatment.

8. A Li4Ti₅Oi2 spinel material obtainable by the process of any of the previous claims and characterized by a BET surface area of at least 10 m²/g.

9. A Li4TiδOi2 spinel material according to the previous claim characterized by a BET surface area of between 10 and 200 m²/g.

10. A Li4Ti₅O-i2 spinel material according to the previous claim characterized by a BET surface area of between 20 and 160 m²/g.

11. A Li4TiδOi2 spinel material according to the previous claim characterized by a BET surface area of between 30 and 140 m²/g.

12. A Li4TiδOi2 spinel material obtained according to the previous claim characterized by a BET surface area of between 70 and 110 m²/g.

13. A thin film electrode comprising a Li₄Ti₅Oi₂ spinel material as claimed in any of the previous claims 8 to 12.