Baking better batteries: what happens when a layered titanate loses its water

Anyone who has made milhojas (puff pastry) knows that the water in the dough matters. In the oven the water turns into steam, and the steam pushes the layers apart.

Some battery materials are layered too, and the water between their layers also decides their structure. In our recent paper we studied one of them, a sodium titanate, before and after "baking" it. The result goes in the opposite direction to the pastry: when the water leaves, the layers close.

What is the material?

We prepared sodium titanate nanosheets (we call them NaTiNS) by hydrothermal synthesis: TiO₂ in a 10 M NaOH solution, heated in a sealed autoclave at 190 °C for 48 h. The product has a layered structure close to NaTi₃O₆(OH)·2H₂O. The water molecules sit between the titanate layers and keep them about 8.4 Å apart.

Then we took part of the same batch and calcined it at 850 °C for 4 h. Thermogravimetric analysis (TGA) showed that the nanosheets lose about 11.5% of their weight on heating, and about 7.5% corresponds to structural water. Without that water the layers collapse. They reorganize into a denser tunnel structure, a mixture of 84% Na₂Ti₆O₁₃ and 16% Na₂Ti₃O₇. In the electron microscope, ribbons that were about 1400 nm long became pieces of about 200 nm.

Both samples come from the same precursor. Any difference in the battery therefore comes from the structure and not from the synthesis.

How does each one store lithium?

We tested both materials as anodes in lithium half-cells. A half-cell uses lithium metal as the other electrode, so we see only the reactions of our material.

The hydrated nanosheets behave like a battery material. Lithium ions enter the space between the layers (intercalation), and the voltage curve shows plateaus. At a slow rate (0.2C, a full charge in 5 h) they delivered 208.6 mAh/g.

The calcined material delivered 110.2 mAh/g under the same conditions, about half. Its voltage curve is sloped, without plateaus, and its cyclic voltammogram is more rectangular. Both are signs that more of the charge is stored at or near the surface (pseudocapacitive behavior) and less in the bulk.

What is the trade-off?

The hydrated material stored more charge at every rate we tested, from 0.2C to 5C. At 5C it gave 62.8 mAh/g against 34.0 mAh/g. After the fast cycles we returned to 0.2C, and it recovered 99.1% of its initial capacity. The calcined material recovered 92.9%.

What the calcined material offers is a rigid structure that settles quickly. Over 200 cycles at 2C its capacity stayed flat, at about 44 mAh/g, and its interface with the electrolyte changed less in our impedance measurements. The hydrated material averaged about 130 mAh/g in the same test and kept gaining capacity. We attribute this to a slow activation of the structure.

To separate the processes inside the cell we analyzed the impedance by distribution of relaxation times (DRT). Each process has its own time scale. Contact and interphase (SEI) processes are fast, charge transfer is intermediate, and diffusion in the solid is slow, more than 10 s. We measured the impedance every 30 mV, in a new cell, after the first cycle and after 200 cycles. This let us follow each process separately.

What does it mean?

The water between the layers works as a spacer and is part of the structure. Keeping it gives more capacity. Removing it gives a more rigid material with a more stable interface. The next step we propose in the paper is the middle ground: heating at intermediate temperatures, to see how much water can be removed before the layers close.

Full paper: E. Espinosa-Villatoro, S. Barba, Á. M. Gómez-Coronel, J. Nelson Weker, E. G. Villabona-Leal, E. Ramos-Díaz, H. J. Ojeda-Galván, M. I. Pérez-Valverde, J. Alanis, "Crystalline structure-dependent performance of sodium titanates as anodes for Li-ion batteries", Journal of Electroanalytical Chemistry 1015 (2026) 120266. DOI: 10.1016/j.jelechem.2026.120266