Although porous TiO₂ particles (200 nm) with 30-50 nm internal pores can exhibit a high power response because of shortened Li⁺ ion diffusion lengths, an anode consisting of these porous TiO₂ particles-only exhibited a low overall volumetric capacity due to a low tap density. On the other hand, TiO₂ nanoparticles (20 nm) can exhibit a high volumetric capacity because of a relatively high tap density, but an anode consisting of these TiO₂ nanoparticles-only exhibited a relatively low power response due to restricted Li⁺ ion mobility. 

To exploit optimally the different intrinsic advantages of both types of TiO₂, a rational design of a two-layer anode structure was developed to locate the porous TiO₂ particles in the inner layer of the anode (nearer to the current collector but further away from the separator) where Li⁺ ion diffusion is usually limited, while the TiO₂ nanoparticles were located in the outer layer of the anode (nearer to the separator but further away from the current collector) where Li⁺ ion diffusion is comparatively unrestricted, as shown in Figure 1(a). In both of the layers, an interconnected multi-wall carbon nanotube (MWNT) scaffold decorated with the TiO₂ particles was used to ensure electrical connectivity and prevent TiO₂ agglomeration.

The reversible volumetric capacity of the two-layer anode was 232.3 mAh cm^-3 at a slow (dis)charge rate of 0.1 C, slightly higher than 220.9 mAh cm^-3 for the anode consisting of TiO₂ nanoparticles-only, and much higher than 104.5 mAh cm^-3 for the anode consisting of porous TiO₂ particles-only. Moreover, the volumetric capacity retention of the two-layer anode, as the (dis)charge rate was increased from 0.1 to 4 C, was similar to 38.5% for the porous TiO₂ anode, and much higher than 7.1% for the TiO₂ nanoparticles. Consequently, the two-layer anode structure maintained the intrinsic advantages of both types of the nanostructured TiO₂.

The two-layer electrode structure was made possible by a flexible and scalable layer-by-layer (LbL) atomisation spray processing technique developed in Oxford University, as shown in Figure 1(b). The LbL spray processing method allowed flexible design of nano-scale active materials arrangements at the meso-scale electrode structure. The spray processing method has a fast production rate of ~10 µm cm^-2 min^-1, capable of making 200 nm – 110 µm thick and up to 1 m x 20 cm large area electrodes using a drum coater, and is thus attractive for complementing existing slurry casting methods.

Finally, direct experimental measurements by in-depth X-ray photoelectron spectroscopy (XPS) profiling of post-mortem anodes were performed to investigate Li⁺ ion transfer dynamics through the thickness of various structured anodes. Differences in the spatial distribution of Ti⁴⁺ to Ti³⁺ states were entirely consistent with the different anode designs and performance.

This work shows how it is now possible to better design and tailor layered electrode structure to improve volumetric capacity and rate capability compared with conventional, randomly mixed electrodes.