Electronic wiring and the defect chemistry of nano-anatase LixTiO2-δ battery electrodes

Due to the huge number of particles, electronic wiring, i.e. providing an electronically conductive pathway between each of the particles and the current collector, can be extremely demanding in nanoparticle battery electrodes. If average particle size decreases from 10 µm to 10 nm, the number of particles that need to be contacted increases by a factor of 109, for the same amount of active material.

We have found that frozen-in oxygen non-stoichiometry is an ideal tool for tuning the electronic conductivity in anatase TiO2-δ particles [1]. Oxygen non-stoichiometry formed by treatment in hydrogen atmosphere at 450 °C is compensated by formation of additional electronic charge carriers; after cooling down to room temperature, the non-stoichiometry and thus the increased concentration of n-type carriers is frozen in and cannot equilibrate anymore. Even after prolonged use as battery electrode, the increased electronic conductivity prevails.

A fact that may be surprising at a first glance is shown in Figure 1b. Despite much higher electronic conductivity (Fig. 1a), the battery performance of the 7h treated anatase particles is worse than that of 1h treated material. The reason for this non-monotonic behavior can be found in the defect equilibria of the LixTiO2-δ system and is due to the necessity of both ion and electron transport [1,2]. Due to strong interaction between e- and Li+ charge carriers, most of these carriers exist in a neutral associate, leading to an opposite variation of the concentrations of the free e- and Li+. Increased electron concentration from the treatment process therefore leads to decreased Li+ carrier concentration.

Under the assumption, that the initial mobility of Li+ is higher than that of e-, it can be shown that the overall chemical diffusivity of Li in fact exhibits a maximum as a function of reduction treatment time (see also Fig. 1d and Fig. 2), which can be fully understood by the defect chemistry of the LixTiO2-δ system.

Figure 1: (a) electronic conductivity variation during treatment of anatase particles in 5%H2/Ar atmosphere at 450 °C. Up to three orders of magnitude increase in electronic conductivity are possible, until equilibrium is reached after ~7 hours.
(b) Charge/discharge capacities for untreated, 1h and 7h reduction-treated material at 10C (=3.36 A g-1) current. (c) Reversible capacities at the 20th discharge cycle for pristine, 1h and 7h treated anatase material at currents of C/5, 1C and 10C.
(d) Achievable discharge capacities at the 20th cycle for discharge currents of 1C and 10C at different reduction treatment times (A discharge/charge rate of nC denotes an insertion/extraction of 1 Li in 1/n h. In case of Li storage in TiO2, 1C is equivalent to 0.336 A h g-1).  Reprinted with permission from [1]. Copyright 2012 American Chemical Society.

Figure 2: Dependence of (a) defect concentrations, (b) conductivities, and chemical diffusion coefficient of Li in LixTiO2−δ, on oxygen nonstoichiometry δ. Reprinted with permission from [1]. Copyright 2012 American Chemical Society.


  1. J.-Y. Shin, J. H. Joo, D. Samuelis, J. Maier, Chemistry of Materials 24(3), 543–551 (2012). DOI: 10.1021/cm2031009
  2. J.-Y. Shin, D. Samuelis, J. Maier, Solid State Ionics225, 590–593 (2012). DOI: 10.1016/j.ssi.2011.12.003
  3. D. Samuelis, J.-Y. Shin, J. Maier, MPI-FKF annual report 2012, 79.
  4. J.-Y. Shin, D. Samuelis, J. H. Joo, J. Maier, European patent application 2586086, PCT WO/2011/160837
Go to Editor View