A material, in particular for use in electrochemical cells or supercapacitors and a method of making such a material

The present invention relates to a material, in particular for use in electrochemical cells or supercapacitors and to a method of making such a material.

One of the greatest challenges for our society is providing powerful electrochemical energy conversion and storage devices with high energy/power densities and high safety as well as low cost. Such devices could be potentially used in an electric vehicle (EV) or hybrid electric vehicle (HEV) and in other applications. Among all these devices, lithium- based batteries are the most promising candidate in terms of energy density, whilst the achievement of high power density is hindered by kinetic problems of the electrode materials. In the above connection, reference can be made to the papers by J. M. Tarascon and M. Armand in Nature 2001, 414, 359;

A. S. Arico, P. G. Bruce, B. Scrosati, J. -M. Tarascon, W. Van Schalkwijk, in Nature Mater. 2005, 4, 366,

However, the achievement of high power density, which requires relatively large amounts of electrode material, is hindered by kinetic problems of the electrode materials. For example, in order to achieve a high rate capability of lithium batteries, rapid ionic and electronic diffusion is necessary. Such materials often involve relatively poorly conducting compounds which are mixed with conductive materials such as carbon black to improve conductivity. The combination of relatively poorly conducting materials and relatively good conductors is often referred to as mixed conduction. Extensive research work focuses on enhancing mixed conduction by doping the electrode materials with foreign atoms. Examples of this can be found in the

papers named above and in the papers by S. B. Schougaard, J. Breger, M. Jiang, C. P. Grey, J. B. Goodenough, in Adv. Mater. 2006, 18, 905, D. Y. Wang, H. Li, S. Q. Shi, X. J. Huang, L. Q. Chen, in Electrochim. Acta 2005, 50, 2955 and J. Hu, H. Li, X. J. Huang, L. Q. Chen, in Solid State Ionics 2006, 177, 2791.

Mixed conduction can also be achieved by admixing electronically conductive phases (electronic wiring through carbon, Ag, conducting polymers, etc.). Examples of this can be found in some of the papers named above and in the papers by M. Nishizawa, K. Mukai, S. Kuwabata, C. R. Martin, H. Yoneyama, in J. Electrochem. Soc. 1997, 144, 1923; F. Zhang, S. Passerini, B. B. Owens, W. H. Smyrl, in Electrochem. Solid State Lett. 2001, 4, A221; N. Ravet, Y. Chouinard, J. F. Magnan, S. Besner, M. Gauthier, M. Armand, in J. Power Sources 2001, 97-8, 503; H. Huang, S. C. Yin, L. F. Nazar, in Electrochem. Solid State Lett. 2001, 4, A170; F. Croce, A. D. Epifanio, J. Hassoun, A. Deptula, T. Olczac, B. Scrosati, in Electrochem. Solid State Lett. 2002, 5, A47;R. Dominko, M. BeIe, M. Gaberscek, M. Remskar, D. Hanzel, J. M. Goupil, S. Pejovnik, J. Jamnik, in J. Power Sources 2006, 153, 274 The wiring technique is widely applied to microsized or submicrosized particles (typically > 50 nm) and was most systemically studied by Jamnik et al. as can be seen from the above referenced paper. A successful example of this is the well-known carbon coating technique used in the synthesis of LiFePO4 electrode material. However, the rate performance enhance- ment of such electrode materials is still limited, as availability or percolation of the electronically conducting phase and /or the electrolyte become insufficient at very high rates. One of the reasons for this is that the carbon coating has been found to be non-continuous so that continuous conductive patterns are missing. The above referenced papers by Taberna et al and by Reiman et al refer to optimization procedures intended for

high rate performance. P. L. Taberna, S. Mitra, P. Poizot, P. Simon, J. -M. Tarascon, in Nature Mater. 2006, 5, 567 and K. H. Reiman, K. M. Brace, T. J. Gordon-Smith, I. Nandhakumar, G. S. Attard, J. R. Owen, in Electro- chem. Commun. 2006, 8, 517. One is the use of nano-architectured elec- trodes consisting of the electrochemical plating of Fe3θ4 onto Cu nanorods acting as a current collector. The other is the use of porous TiO2 thin films. Both designs lead to enhanced power performance but are naturally not meant for or suitable for achieving high energy demands because the thin films required make it impossible to achieve sufficient electrode mate- rial in a confined space for high energy applications.

The object of the present invention is to provide a material, in particular for use in electrochemical cells or supercapacitors and a method of manufacturing such a material which provide an optimized nanostructure de- sign of materials for both high power and high energy use, in particular in lithium batteries, but also in a variety of other electrochemical devices and applications.

In order to satisfy this object there is provided, in accordance with the present invention, a material in particular for use in electrochemical cells or supercapacitors comprising a poorly conducting active material of relatively low conductivity having regular or irregular passages having average cross-sectional dimensions generally in the size range from 5μm to 200nm and interconnected mesopores having average cross-sectional dimensions in the size range from 2 to 50nm and the active material being covered with a network of an electronically conductive metal oxide of relatively high conductivity extending into said mesopores.

Further there is provided a method of manufacturing such a material comprising the steps of making a material in particular for use in electro-

chemical cells or supercapacitors comprising the steps of preparing a poorly conducting active material of relatively low conductivity having regular or irregular passages having average cross- sectional dimensions generally in the size range from 5μm to 200nm and interconnected mesopores having average cross- sectional dimensions in the size range from 2 to 50nm and the active material being covered with a network of an electronically conductive metal oxide of relatively high conductivity extending into said mesopores.

The material of the invention permits highly Li-permeable materials to be obtained by providing the material with a hierarchical, "self-similar" mixed conducting three-dimensional (3D) networks. The nanoscopic network structure is composed of a dense net of "metalized" mesopores that allow both Li ⁺ and e ^~ to migrate. The term "metalized" is used here because the metal oxides used to form the "metalized mesopores" have electronic conductivity, i.e. a conductivity approaching that of metals. E.g. for RuO2 the conductivity is 5*10 ⁴ S/cm For alternatives such as IrO2, VO2, Moθ2, WO ₂ , Co ₃ O ₄ and Fe ₃ O ₄ it is 2*10 ⁴ S/cm, 2*10 ³ S/cm, 5*10 ³ S/cm, 3*10 ² S/cm, ~10 ² S/cm and 2.5*10 ² S/cm respectively. This network with a mesh size of about 10 nm is superimposed on a similar net on the micro- scale formed by the composite of the mesoporous particles and the conductive admixture. The power of this concept can be demonstrated by reference to Example 1 given below relating to the synthesis of a mesoporous Tiθ2'.Ruθ2 nanocomposite which shows superior high rate capability when used as anode materials for lithium batteries as will be explained further below.

Preferred embodiments of the invention are set forth in the appended subordinate claims and incorporated into the description by reference.

The present invention will be explained in the following the more detail by way of example only and with reference to the accompanying drawings which show:

Fig. 1 (a) a conceptual representation of the desired inventive design comprising a "self- similar" structure concerning the transportation of ions from microscale to nanoscale, with shaded areas representing the efficient mixed conducting parts; (b) a sketch of a realistic composite meeting this concept,

Fig. 2 X-ray diffraction patterns of (a) as-prepared mesoporous Tiθ2 spheres and (b) a mesoporous TiO2:RuO2 nanocomposite,

Fig. 3 Elemental mapping of a mesoporous TiO2:RuO2 nanocomposite in which (a) shows an annular dark-field TEM image of the mesoporous Tiθ2:Ruθ2 nanocomposite and corresponding Ti and Ru EDX maps; (b) shows a HRTEM image taken from the outer edges of a Tiθ2:Ruθ2 sphere and (c) shows a corresponding schematic illustration of the self-wired path of deposited Ruθ2 nanoparticles,

Fig. 4 Rate performance diagrams showing the variation of discharge (square) /charge (round) capacities versus cycle number for different anatase electrodes cycled at different rates between voltage limits of 1 and 3 V, more specifically for (a) 300nm-

TiO2; (b) 5nm-Tiθ2; (c) mesoporous Tiθ2; and (d) a mesoporous

TiO 2: RuO 2 nanocomposite in accordance with the present in- vention,

Fig. 5 Typical SEM (a) and TEM (b) images of mesoporous TiO2 spheres,

Fig. 6 A typical TEM image of mesoporous TiO2:RuO2 spheres,

Fig. 7 Galvanostatic discharge /charge curves of mesoporous TiO2:RuO2 composite electrode cycled at different rates from

C/ 5 to 30 C between voltage limits of 1 and 3 V,

Fig. 8 A typical TEM image of a mesoporous LiFePθ4 grain,

Fig. 9 (a) and (b) typical HRTEM images to different scales of carbon coated LiFePθ4 after coating with RuO2 and (c) a schematic drawing illustrating the effect of the RuO2,

Fig. 10 (a) and (b) further typical HRTEM image of LiFePO ₄ after car- bon coating and coating with Ruθ2,

Fig. 11 (a) and (b) yet further typical HRTEM images of LiFePO ₄ after carbon coating and coating with RuO2,

Fig. 12 Typical charge/ discharge profiles of carbon coated LiFePO ₄ before and after Ruθ2 coating at a current rate of C/ 10, the insert shows the magnified flat region,

Fig. 13 A comparison of the rate performance of carbon coated LiFePO ₄ before and after RuO2 coating,

Fig. 14 Typical X-ray diffraction images of carbon coated LiFePO ₄ and LiFePO ₄ coated with both carbon and Ruθ2, and

Fig. 15 A diagram similar to Fig. 12 showing a surprising increase in specific capacity as the number of charge /discharge cycles increases.

The general scheme of an optimized nanostructure design of electrode materials, which is still simple to fabricate, is shown in Figure 1. The schematic drawing (a) shows an electrode 10 made of a material in accordance with the present teaching with an electrode 10 being in contact at one side with an electrolyte 12 present in an electrochemical cell or supercapacitor, e.g. an electrolyte of a lithium battery permitting transport of Li+ ions to and from the electrode 10. At the opposite side of the electrode 10 there is a current collector 14 consisting, e.g. of a metal foil, e.g. a Ti foil. Suitable electrolytes are e.g. described in WO 2004/034489 (EP 03788901.1) and EP-A- 1505680.

The material of the electrode is provided with a macropore structure comprising islands or grains of electrode material 16 with passages 18 disposed therebetween. The islands 16 and the passages 18 have macroscale dimensions, i.e. typically in the range from <1 μm to >300 nm and are il- lustrated as being regularly formed and placed. In practice this is unlikely - the islands and the passages would typically be of slightly or even highly irregular shape but have average cross-sectional dimensions in the range quoted.

The islands 16 are also not solid, but are rather porous, more specifically mesoporous with particles of material 20 and passages 22 between them, as can be seen from the magnified schematic view of a portion of the island 16' shown in a circle at the top of Fig. Ia. The particles 18 and the passages 22 typically have dimensions in the range 2 nm to 20 nm and are again shown as being regularly formed and placed. As before, this is

not essential, the particles 20 and passages 22 could also be of slightly or highly irregular shape, with the particles 20 not necessarily being discrete but possibly joined together at various points and permeated by at least partly interconnected voids forming the passages 22. Thus, the micro- structure of the islands is typically (but not necessarily) similar to the macro structure of the material itself but to a smaller scale. This concept is expressed in the present description by the term "similar" or "self similar".

One possible practical realization of this concept is shown in Fig. l(b). Here the light colored regions are particles of a relatively poorly conducting material TiO2 which is a useful electroactive material for an electrode of a lithium battery, but has a relatively low electrical conductivity of below 10" ⁶ S /cm

The particles of Tiθ2 are contained in generally spherical islands or grains 16, which can be better visualized by reference to Fig. 5 and which are mesoporous by virtue of the interconnecting voids and passages between the particles 20. The grains of Tiθ2 are mixed with carbon black which has a relatively high conductivity (0.1-2 S/cm) and the particles 24 of carbon black are represented in Fig. l(b) as large black dots. In addition, the particles 20 and islands 26 are permeated with Ruθ2 an electronic conductor and this forms a conductive network superimposed on the Tiθ2 particles 20 and grains 16. The Ruθ2 is shown by small black dots 26 in Fig. l(b). As can also be seen from Fig. l(b) the electrolyte 12 permeates the pas- sages 18 between the grains 16 and the interconnected pores, voids or passages 22 of the individual grains.

Thus, an efficient mixed conducting 3D nano-network is introduced into the material with a mesh size of a few nanometers only and with channel widths of comparable size. The way this is done is described later with ref-

erence to Fig. 2. As a result of the special mixed conducting three- dimensional network in accordance with the present teaching the insertion kinetics in the electroactive material (here TiO2) becomes indeed negligible and the insertion rate of Li is enhanced to such a degree that the transport within the channels becomes the limiting factor with respect to the insertion rate. Since the diffusion coefficients D of Li in nano-sized anatase is ~ 2 x 1O ¹⁵ cm ² s- ¹ , as reported in the paper by R. Van de Krol, A. Goossens, J. Schoonman, in J. Phys. Chem. B 1999, 103, 7151, the mean diffusion time <teq> = L ² /2D reduces to 120 s for a mean channel distance of about 7 nm. The need to ensure infiltration of the electrolyte into the mesopores sets a lower limit to the mesh size of the conductive Ruθ2 network in the mesopores while the necessity to consider tolerable loss of the electroactive material per volume (meaning that if the passages 18 and 22 are made larger, the amount of electroactive material per unit volume becomes smaller), material stability as well as sufficient connectivity set an upper limit.

This structure design is realized by preparing mesoporous anatase with an average pore size of ca. 7 nm and by subsequently metalizing the pore channels with crystalline Ruθ2. It is noteworthy that in addition to the electronic function described in the papers by J. V. Ryan, A. D. Berry, M. L. Anderson, J. W. Long, R. M. Stroud, V. M.Cepak, V. M. Browning, D. R. Rolison, C. I. Merzbacher, in Nature 2000, 406, 169 and P. Balaya, H. Li, L. Kienle, J. Maier, in Adv. Fund. Mater. 2003, 13, 621, Ruθ2 also allows for quick Li permeation as described also in the paper by M. Armand, F. Dalard, D. Deroo, C. Mouliom, in Solid State Ionics 1985, 15, 205. Furthermore, Ruθ2 is most beneficial as it is - owing to similar bonding properties - expected to spread much better on Tiθ2 than carbon would, and thus can efficiently metallize the tiny channels due to the ionic character- istic of both oxides (Ruθ2 and T1O2). The low wetting angle of electrolyte on

Tiθ2 gives rise to a ready filling of the channels by the liquid electrolyte. As the R11O2 arrangement is highly porous but percolating, a large number of active triple-phase contacts are formed as well. It is a very fortunate circumstance that the contact of Tiθ2 to the electrolyte will result in an in- creased local conductivity owing to dissociation of the ion pairs. Such dissociation of ion pairs is described by A. J. Bhattacharyya, J. Maier, in Adv. Mater. 2004, 16, 811. This effect is particularly pronounced if the channel width is of the order of the screening length. Even though such a mesoporous material is crucial to warrant good electronic and ionic con- tacts on the nanoscale, a feasible electrode design in addition requires a superstructure on a larger scale. One reason is the sheer fact that fabrication of a mesoporous monolith of electrode dimensions is - if possible at all - very intricate and expensive. The second reason is more fundamental and refers to the fact that a large mesoporous monolith can be hardly in- filtrated with electrolyte. The surface interaction provides the necessary driving force (capillary pressure) yet will also lead to a significant pressure loss if the monolith is large.

The inventive concept of superimposing onto the mesoporous structure a similar network on the microscale, with the mesh size being given by the size of the mesoporous particles and with the channel width being selected as a compromise between realization of quick transport and loss of elec- troactive material, i.e. its absence in the volume of the electrode, is actually quite straightforward to achieve. Owing to the less demanding kinetics on the micrometer scale this design is most straightforwardly arrived at by the composite of mesoporous particles and the conducting admixture (e.g., carbon black and Ruθ2 or just RUO2). In the end, as regards qualitative phase distribution and functionality, the hierarchical structure achieved is self-similar on the transition from nanoscale to microscale (see Fig. 1).

The synthesis of the mesoporous TiO2:RuO2 nanocomposite, which meets this design and shows excellent high rate capability when used as anode materials for lithium batteries will be described in more detail in the following. Nanostuctured titania has been of considerable interest as a prom- ising anode material for lithium batteries because of high reversibility of Li insertion /extraction at a low-voltage. This reversibility is e.g. described in the papers by L. Kavan, M. Gratzel, S. E. Gilbert, C. Klemenz, H. J. Scheel, in J. Am. Chem. Soc. 1996, 118, 6716A. R. Armstrong, G. Armstrong, J. Canales, R. Garcia, P. G. Bruce, in Adv. Mater. 2005, 17, 862; Y. G. Guo, Y. S. Hu, J. Maier, in Chem. Commun. 2006, 2783; Y. S. Hu, L. Kienle, Y. G. Guo, J. Maier, in Adv. Mater. 2006, 18, 1421; and E. Baudrin, S. Cassaignon, M. Koelsch, J. P. Jolivet, L. Dupont, J. -M. Ta- rascon, in Electrochem. Commun., 2007, 9, 337.

A first example will now be given for the preparation of an electroactive material based on Tiθ2.

Example I

For the preparation of a material in accordance with the present teaching, mesoporous anatase sub-micron spheres with a uniform grain size (-300 nm) (see Fig. 5) and with surface area of ca. 130 m ² g" ¹ were prepared, according to the method proposed in the paper by Y. G. Guo, Y. S. Hu and Jo. Maier in Chem. Community 2006, 2783. by using a TiO2-CdSO4 composite as intermediate. Peak positions and widths in the X-ray diffraction (XRD) pattern as shown in Fig. 2a of the as-prepared mesoporous T1O2 spheres confirm the fabrication of pure anatase nanocrystalline Tiθ2 with a mean crystallite size of D ₁ Oi along the [101] axis of ca. 7 nm estimated from the widths of the major diffraction peak (2θ = 25.3°) by Scherrer's formula. In- ternal metalization of the mesoporous Tiθ2 spheres was carried out by

wetting the ηO2 powder with 0.1M RuCb solution followed by heat treatment at 450 ⁰ C under O2. The successful introducing of RUO2 is shown in the XRD pattern of Fig. 2b for the resulting sample by the appearance of two new diffraction peaks at 1θ = 28.0° and 35.1°, which can be indexed as the 110 and 101 planes of the tetragonal phase of RuO2 (JCPDS No. 40- 1290). The amount of Ruθ2 is about 5 wt-% corresponding to a complete transformation of RuCb to Ruθ2.

The distribution of crystalline RUO2 within the porous Tiθ2 spheres was characterized using a Zeiss Libra 200FE transmission electron microscope (TEM) operated at 20OkV and equipped with a scanning unit and an energy-dispersive X-ray (EDX) analyzer. Figure 3a shows a typical annular dark-field TEM image taken from a Tiθ2 sphere along with corresponding Ti and Ru maps. All together, this clearly points towards a uniform distri- bution of Ruθ2 down to the 10 nm scale. As visually traced from the high resolution TEM image of Fig. 3b taken from the outer edges of a T1O2 sphere, the as-deposited Ruθ2 nanoparticles form a 3D interconnected network over a portion of the inner surface of the mesopore walls that connects the carbon network on the microscale (Fig. 3c). It is worth noting that the electronic wiring with Ruθ2 does not change the nanostructure of the parent mesoporous Tiθ2 spheres. This can be seen by a comparison of Figs. 5 (especially 5(b)) and 6.

At this stage it is useful to consider the performance of conventionally prepared anatase and carbon composite electrodes. Figure 4, in which rates of up to 3OC (one lithium per formula unit in 1 /30 hour, i.e., 10.08 A g ¹ ) have been employed, compares 300 nm-sized anatase (Fig. 4(a)) with 5 nm-sized anatase (Fig. 4(b)) as well as with mesoporous Tiθ2 but without interior metalization (Fig. 4(c)). Both the nano-sized and the mesopor- ous anatase exhibit much higher capacity and better cycling performance

than the 300 nm-anatase. The extent to which compound formation plays a role in the nanostructured anatase is not yet known. At low current rates, the mesoporous ηO2 has comparable performance with nano-sized anatase (5 nm TIO2). At high current rates above 1OC, its performance be- comes worse. In contrast to the electrolyte which can penetrate into the mesopores, the carbon admixture only contacts the surfaces of the mesoporous grains 16.

For comparison purposes, Figure 4(d) now shows the outstanding rate performance of the obtained mesoporous Tiθ2:Ruθ2 composite after introducing internal metalization via the mixed-conducting network structure. This is confirmed by the curves of Fig. 7 which are galvanostatic discharge/charge curves for a mesoporous TiO2:RuO2 electrode cycled at different rates from C/ 5 to 3OC between voltage limits of IV and 3V. Since Ruθ2 also contributes to the capacity in the voltage range of 1-3 V, the capacity of the Tiθ2:Ruθ2 composite was calculated based on the whole mass (it is noted that no Ru/Li2θ nanocomposite is formed in this voltage range).

The cell was first cycled at C/ 5 and, after 20 cycles, the rate was increased in stages to 3OC. A specific charge capacity of around 214 mA h g ^"1 was obtained at a rate of C/ 5 after 20 cycles; this value reduces to 190 mA h g" ¹ at 1C, to 147 mA h g" ¹ at 5C, and to 125 mA h g" ¹ at 1OC. At the very high rate of 3OC (discharge /charge of all the TiO2 within 2 mini), the specific charge capacity is still 91 mA h g" ¹ , which is about two times larger than that of 5 nm anatase (48 mAh g" ¹ ) and nine times larger than that of mesoporous anatase spheres without interior electronic wiring (10 mAh g" ¹ ). The reversibility is demonstrated by the fact that the capacity of 210 mA h g- ¹ is returned to if the rate is lowered to C/ 5 (not shown here). So far as can be ascertained, such a good performance at these high rates

has never been found before. Especially, the performance at such high rates is much better than those of commercial Tiθ2, nano-sized anatase, Tiθ2-B nanowires, and nano-sized rutile. The best prior result known to date seems to be that achieved in the above referenced paper by Owen et al. However, these authors only used a maximum rate of 8.9 C (i.e., 3 A g- ¹ J and used porous TiO2 thin films with low diffusion distances but relatively low absolute capacities.

Thus, the present invention relates to a new design of electrodes achieved by fabricating a hierarchically nanostructured electrode with highly efficient mixed conducting 3D networks on both nanoscale and microscale levels. A key to its realization is, besides the preparation of mesopores, the use of a suitable electronic conductor - here the oxide R11O2 - that enables favorable surface- surface interactions between the RuO2 and the Tiθ2. The nano-sized network provides negligible diffusion times, enhanced local conductivities and possibly faster phase transfer reactions, and thus plays a key role in achieving the extremely good power performance. The microscopic network guarantees high absolute capacities, ease of fabrication and quick infiltration. The whole procedure is simple, yet very effective and, owing to its versatility, could also be extended to other anode and cathode materials used in lithium batteries and to porous conducting materials in general.

A specific example of the synthesis of the ηO2 - Ruθ2 electrode material in accordance with the present invention will now be given.

In this synthesis, one solution containing 0.92 g cadmium acetate di- hydrate (Merck), 0.38 g thiourea (Merck), and 0.38 g 1-thioglycerol (Sigma) in 40 mL of N,N-dimethylformamide/water (3: 1, in volume) solvent was

added into 80 mL of continuously stirred butanol solution with 3 mL Ti(OBu)4 (Aldrich) and 0.6 mL acetylacetone (Aldrich) at room temperature. Then the mixture was stirred for 20 min and refluxed at 140 ⁰ C for about 3 h. The as-produced white precipitates were collected by using centrifuge and repeatedly washed with ethanol and distilled water. The Ti-Cd precursor was calcined at 500°C under air for 5 h to obtain crystalline TiO2/CdSO4 composites. To form the mesoporous Tiθ2 spheres CdSO4 was completely removed in 10 wt-%HNO3 aqueous solution, followed by thorough rinsing with distilled water. To prepare mesoporous TiO2:RuO2 spheres, 133 mg obtained mesoporous Tiθ2 spheres were wetted by 0.5 mL of 0.1 M RuCb solution. After drying under air, the powders were transferred into a tube furnace and calcined at 450 ⁰ C for 1 h under O2.

XRD measurements were carried out with a PHILIPS PW3710 using fil- tered Cu Ka radiation. A JEOL 6300F scanning electron microscope (SEM) was used to investigate the morphology of the materials. TEM and HRTEM images were collected by using a JEOL 2000EX (operating at 200 kV) and a JEOL 4000EX (operating at 400 kV) transmission electron microscopes, respectively. Ti and Ru maps were collected by using a Zeiss Libra 200FE transmission electron microscope (operating at 200 kV) equipped with a scanning unit and an energy-dispersive X-ray (EDX) analyzer (EDAX, Ametek, USA). The nitrogen adsorption and desorption isotherms at 77.4 K were obtained with an Autosorb-1 system (Quanta Chrome) after the sample was degassed in vacuum at 120 ⁰ C overnight.

Electrochemical experiments were performed using two-electrode Swage- lok-type™ cells. For preparing working electrodes, a mixture of the various samples of Tiθ2, i.e. commercial Tiθ2 (anatase) in 5 nm and 300 nm particle sizes, mesoporous Tiθ2 and the TiO2:RuO2 composite of the present invention were each mixed with carbon black and poly (vinyl difluoride)

(PVDF) at a weight ratio of 60:20:20, were pasted on pure Cu foil (99.6 %, Goodfellow). Glass fiber (GF/ D) from Whatman® was used as a separator. The electrolyte consists of a solution of 1 M LiPFβ in ethylene carbonate (EC) /dimethyl carbonate (DMC) (1: 1, in volume) obtained from Ube Indus- tries Ltd. Pure lithium foil (Aldrich) was used as counter electrode. The discharge and charge measurements were carried out under a similar electrochemical condition on an Arbin MSTAT system. The cells were assembled in an argon-filled glove box.

Example II

This example relates to lithium iron phosphate (LiFePO4) which, as described in the paper by Padhi, A. K.; Nanjundaswamy, K. S.; Goodenough, J. B. J. Electrochem. Soc. 1997, 144, 1188.

Work on this compound by a variety of scientists has shown that it has numerous appealing features such as high theoretical capacity (170 mA h g" ¹ ), acceptable operating voltage (3.4 V vs. Li ⁺ /Li), high safety, environmental benignity, low cost, etc. The scientific work on LiFePO4 is described in the papers by Croce, F.; Epifanio, A. D.; Hassoun, J.; Deptula, A.; Olc- zac, T.; Scrosati, B. Electrochem. Solid State Lett. 2002, 5, A47. Wang, D. Y.; Li, H.; Shi, S. Q.; Huang, X. J.; Chen, L. Q. Electrochim. Acta 2005, 50, 2955, Delacourt, C; Poizot, P.; Levasseur, S.; Masquelier, C. Electrochem. Solid State Lett. 2006, 9, A352, Doeff, M. M.; Hu, Y.; McLarnon, F.; Kostecki, R. Electrochem. Solid State Lett. 2003, 6, A207, Zaghib, K.; Guerfi, A.; Charest, P.; Striebel, K. A. Electrochem. Solid State Lett. 2005, S, A207, Chen, G.; Song, X.; Richardson, T. J. Electrochem. Solid State Lett. 2006, 9, A295, Dominko, R.; BeIe, M.; Gaberscek, M.; Remskar, M.; Hanzel, D.; Pejovnik, S.; Jamnik, J. J. Electrochem. Soc, 2005, 152, A607, Park, K. S.; Schougaard, S. B.; Goodenough, J. B. Adv. Mater. In press,

Wang, Q.; Zakeeruddin, S. M.; Wang, D.; Exnar, I.; Gratzel, M. Angew. Chem. Int. Ed. 2006, 45, 8197, Xie, H. M.; Wang, R. S.; Ying, J. R.; Zhang, L. Y.; Jalbout, A. F.; Yu, H. Y.; Yang, G. L.; Pan, X. M.; Su, Z. M. Adv. Mater. 2006, 18, 2609, Wang, Y.; Wang, J.; Yang, J.; NuIi, Y. Adv. Func. Ma- ter. 2006, 16, 2135, Prosini, P. P.; Lisi, M.; Zane, D.; Pasquali, M. Solid State Ionics 2002, 148, 45,

It has thus attracted extensive interest as a potential candidate for replacing the commercial layered LiCoθ2 material, which shows a relative low capacity and low safety. However, the biggest disadvantage of LiFePO ₄ is its sluggish mass and charge transport, as corroborated by a recent work on the electrical measurement on a single crystal by Amin, R.; Balaya, P.; Maier, J. Electrochem. Solid State Lett. 2007, 10, A13.

Tremendous efforts have been made to overcome the electronic- and ionic- transport limitations by doping with foreign atoms or by decreasing the particle size or by using a reversible redox couple or by coating electronically conductive agents (carbon, Ag, conducting polymers, etc.). The carbon-coating technique is widely applied and was most systemically stud- ied by Jamnik et al. as described in the above referenced papers. However, the rate performance enhancement of such electrode materials is still limited, as availability or percolation of the electronically conducting phase become insufficient at very high rates. It seems that the thickness of carbon layer on the surface of LiFePO4 particles is different on the different crystal planes of LiFePO4. There is even no carbon on some places. That means carbon will not cover all the surface of LiFePO4 particles, leading to a non-continuous electronically conducting network. The reason is that carbon and LiFePO4 have different surface properties. One also has to consider the wetting properties and/or surface- surface interactions when selecting the suitable coating material beyond its electronic function.

The work previously conducted has mainly been carried out on solid LiFePO4, however some proposals have also been made for the preparation of porous LiFePO ₄ , notably in the paper by Jamnik et al in The Journal of the Electrochemical Society 152(5) A858 - A863 (2005). This paper describes the preparation of porous LiFePO4 and the carbon coating of the porous material which is formed at the same time as the porous material. The content of that paper with respect to the preparation of the porous LiFePO4 carbon coated material is incorporated herein by reference. How- ever, as mentioned above, there is a significant problem with the fact that the LiFePO ₄ is not coated with a continuous network of carbon.

The present invention is able to overcome this difficulty. More specifically, it has been found that nanosized RuO2 can be used to to 'metalize' tiny pores and even to 'repair* incomplete electronically conducting (carbon) networks in porous carbon-containing LiFePO ₄ , the kinetics and rate capability of the composite are significantly improved.

This is realized by using a low- temperature, solution-infiltration approach that nano-sized RuO2 is deposited on carbon-containing porous LiFePO ₄ by cryogenic decomposition of ruthenium tetroxide (RuO ₄ ), which has previously been used as a precursor to prepare RuO2 thin films at rather low temperature as described in the paper by Yuan, Z.; Puddephatt, R. J.; Sayer, M. Chem. Mater. 1993, 5, 908.

Carbon-containing (~3 wt-%) porous LiFePO ₄ composite materials were prepared by a sol-gel method. 0.01 mol of lithium phosphate (Li3PO ₄ , Aldrich 33,889-3) and 0.02 mol of phosphoric (V) acid (H ₃ PO ₄ , Aldrich

31,027-1) were dissolved in 200 mL water by stirring at 70 ⁰ C for 1 h. Separately, 0.03 mol of iron (III) citrate (Aldrich, 22,897-4) was dissolved in 300 mL of water by stirring at 62 ⁰ C for Ih. The two solutions were mixed together and dried at 6O ⁰ C for 24 h. After thorough grinding with a mortar and pestle, the obtained material was fired in inert (argon) atmosphere at 700 ⁰ C for 10 h. The heating rate was 10°C/min. Aqueous Ruθ4 (~0.5 wt-%) solution was received from Strem Chemicals. A certain amount of porous C-LiFePO4 was placed in a tube-like flask with a sidearm. About 2 mL of pentane was condensed in the sidearm, then warmed to room temperature and allowed to equilibrate with the porous C-LiFePO4 for some time. By slowly cooling the flask, the pentane condensed in the flask and covered and filled the porous C-LiFePθ4. Solution of Ruθ4 in pentane with a very low melting point and low viscosity was used to mimimize capillary forces on the C-LiFePO4 during wetting at such critical conditions. 10 mL of pentane was employed to extract Ruθ4 from aqueous RuO4 (~ 10 mL). A certain amount of solution of RuO4 in pentane was added to the flask containing the porous C-LiFePθ4 pre-cooled to -78 ⁰ C in a dry ice /acetone bath. The flask was allowed to warm slowly to room temperature over a period of several days. All the operations were carried out in a well- vented hood. Until all the pentane was evaporated in the flask, the obtained dry sample was put into a vacuum oven and heated at 200 ⁰ C for 1 h. The amount of Ruθ2 is about -4 wt-% corresponding to a complete extraction and transformation of RuO ₄ .

Structural and Electrochemical Characterizations

XRD measurements were carried out with a PHILIPS PW3710 using filtered Cu KD radiation. TEM and HRTEM measurements were performed a JEOL 4000EX (operating at 400 kV) transmission electron microscopes, respectively. Electrochemical experiments were performed using two- electrode Swagelok-type™ cells. For preparing working electrodes, a mix-

ture of the samples of C-LiFePO4 (C-LiFePθ4-Ruθ2), carbon black, and poly (vinyl difluoride) (PVDF) at a weight ratio of 80:10:10, was pasted on pure Al foil (99.6 %, Goodfellow). Glass fiber (GF/D) from Whatman® was used as a separator. The electrolyte consists of a solution of 1 M LiPFe in ethylene carbonate (EC) /dimethyl carbonate (DMC) (1: 1, in volume) obtained from Ube Industries Ltd. Pure lithium foil (Aldrich) was used as counter electrode. The cells were assembled in an argon-filled glove box. The discharge and charge measurements were carried out under an identical electrochemical condition on an Arbin MSTAT system.

In order to coat the porous carbon coated LiFePO4 a solution of Ruθ4 in pentane is infiltrated into the porous carbon coated LiFePO4 and its temperature is increased slowly from -78°C after equilibration with the carbon-containing porous LiFePθ4. The Ruθ2 preferentially forms on the bare surface of LiFePU4 rather than on the surface of carbon due to the similar ionic characteristic of the two compounds (LiFePO4 and RUO2) . This point is clearly supported by the HRTEM measurements whose results are shown in Figure 9 and in Figures 10 and 11. It can be seen from these features that nano-sized Ruθ2 with a particle size of about 5 nm is directly deposited on the bare surface of LiFePU4 instead of on the carbon surface which has been confirmed by extensive HRTEM studies. The non- continuous electronically conducting network of carbon is "repaired" because of the RUO2 preferential deposition on the bare surface of LiFe PO4. As the Ruθ2 arrangement is also porous and as percolation occurs, a large number of active triple-phase (RuO2, LiFePO4, liquid electrolyte) contacts are formed, greatly facilitating Li insertion. The boundary between Ruθ2 and LiFePO4 can be readily observed, indicating the good wetting properties as seen in Fig. 9(a). It is worth noting that after further coating with Ruθ2, the resulting C-LiFePθ4.Ruθ2 composite still retains the morphol-

ogy, porous and crystal structure of the porous C-LiFePO ₄ substrate, as corroborated by TEM and XRD measurements shown in Figures 8 and 14.

In order to check the potential application in high-power Li batteries, the electrochemical performance for Li insertion/ extraction has been investigated. Before discussing the results, it is useful to briefly consider the performance of the carbon-containing porous LiFePO4, which was successfully prepared by Jaminik et al. After several initial cycles, the reversible capacity stablizes at about 140 mA h g- ¹ . All the electrochemical results discussed below are the stable performance obtained after several initial cycles. Figure 12 shows the typical charge /discharge profiles of carbon- containing porous LiFePO4 before and after nano-sized RuO2 coating at a current rate of C/ 10 (one lithium per formula unit in 10 hours). After Ruθ2 coating, the polarization between the charge and discharge plateaus is reduced to 36 mV from the 51 mV for the sample without Ruθ2 coating, indicating that the kinetics of the LiFePO4 is indeed improved after Ruθ2 coating. Figure 13 shows the comparison of rate performance of carbon- containing porous LiFePO4 before and after Ruθ2 coating, where rates of up to 3OC have been employed. At low rates, they exhibit comparable per- formance. However, at higher rates, the difference is very clear, e.g. specific reversible capacities of 124 and 93 mA h g" ¹ were obtained at rates of 2C and 1OC respectively for the sample after RuO2 coating, which is much higher than those of the sample before RuO2 coating. These results reveal that carbon coating on porous LiFePO ₄ works fine for lithium insertion at low current rates but does not at high current rates due to the insufficient electronically conducting network. After "repairing" the electronically conducting network by nano-sized Ruθ2, the kinetics and rate capability of the composite are significantly improved. The results here give clear evidence of the utility of Ruθ2 in generating very effective mixed conducting heterogeneous electrodes.

We proposed a new design by fabricating a hierarchically nanostructured electrode with highly efficient mixed conducting 3D networks on both nanoscale and microscale levels. A key to its realization is, besides the preparation of mesopores, the use of a suitabe electronic conductor - here the oxide Ruθ2 - that enables favorable surface- surface interactions. While the nano-sized network provides negligible diffusion times, enhanced local conductivities and possibly faster phase transfer reactions, and hence is the key to the extremely good power performance, the micro- scopic network guarantees high absolute capacities, easy of fabrication and quick infiltration. The whole procedure is simple, yet very effective, and owing to its versatility, could also be extended to other anode and cathode materials used in lithium batteries, such as Li _-1 TiSO ₁₂ , V2O5, Li- CoO2, LiMn ₂ θ4, LiCθχNi _y Mn ₁ -χ-yθ2 (0<x< l, 0<y< l, 0<x+y< l), LiMnPO ₄ and so on.

Although the above examples have all utilized RuO ₂ as the conductive metal oxide other electronically conducting metal oxides can also be used, for example Irθ2, VO2, MoO ₂ , WO ₂ , Cθ3θ ₄ and Fe3θ4.

As mentioned above the material of the invention can also comprise particles of a conductive material dispersed in the active material and present in the passages 18. The particles of conductive material preferably comprise carbon black.

The active material preferably comprises generally spherical mesoporous grains of one Of TiO ₂ and LiFePO ₄ of a diameter in the range from 400 to 2000nm with mesopores having cross-sectional dimensions in the size range from 2 to 50nm with a conducting network of crystalline Ruθ2 coating the grains and extending inside the mesopores, with the proportion of

Ruθ2 to Tiθ2 being in the range from 4% to 20% by weight, with particles of carbon black having diameters in the range from generally 30nm to 50nm being interspersed with the mesoporous grains and located in the passages between the grains and optionally in the mesopores and with the proportion of carbon black lying in the range from the range from 10 to 30% by weight of the combined weight of TiO ₂ and RuO ₂ or of LiFePO4 and RuO ₂ .

The RuO ₂ generally fills any non-continuities between adjacent grains of carbon black, i.e. gaps between them. The application of this invention is not restricted to the lithium batteries and can also be extended to other electrochemical devices such as supercapacitors and photoelectrochemical devices such as Dye-sensitized solar cells (DSSC) where TiO ₂ is used as a photoelectrode .