Memory is not only vital for the human mind to function but also for technology. Many electronic devices rely on digital memory to compute and store data. The speed and capacity of random access memory (RAM) has increased significantly over the years, but there is one drawback: the digital information contained in these devices needs to be constantly refreshed in order to be maintained, which is not energy-efficient. Memory that can be maintained even after powering down is important for next-generation electronic devices.

In their recent publication in Advanced Functional Materials, Sankaramangalam Ulhas Sharath and Lambert Alff from the Darmstadt University of Technology, Germany, and their co-workers, report how to control the switching modes and conductance quantization of HfO_x-based memristive devices.

Using reactive molecular beam epitaxy, the authors could change the oxygen stoichiometry of the HfO_x-based resistive random-access memory (RRAM) device, enabling identification of the different switching modes and its dependence on oxygen stoichiometry.

A qualitative model shows that in stoichiometric monoclinic HfO₂, the higher electroforming voltage breaks a large number of hafnium–oxygen bonds, creating stronger filaments and accumulating more oxygen ions at the titanium nitride interface. Oxygen-deficient tetragonal HfO_x already possesses a homogeneous oxygen vacancy distribution, allowing the onset of filament formation at much lower fields and with fewer oxygen ions accumulating at the titanium nitride interface.

Considering oxygen ion transport by the electric field and by Joule heating resulted in a model accounting for all occurring switching modes. According to this model, the stabilization of quantum conductance states is caused by reduced thermally driven ionic/vacancy motion in tetragonal, non-stoichiometric HfO_1.5.

To learn more about switching-mode control of HfO_x-based memristive devices, please visit the Advanced Functional Materials homepage.