Automated and autonomous scanning probe experiments for manipulating and measuring domain wall properties in ferroelectric thin films

Domain walls in polar materials are a playground for investigating novel physics, due to the symmetry breaking that occurs at the wall, and the corresponding unique functional properties they can exhibit.¹

Traditionally, manipulation of ferroelectric domains at the nanoscale has been achieved largely through piezoresponse force microscopy (PFM), where the electric field is concentrated below the PFM tip and allows domains and domain walls to be both perturbed and measured. However, this manipulation has traditionally been slow and done by human operators, or in automated regimes where e.g., the full hysteresis loop is captured across a grid of points, which does not allow for feedback.²

Here, we introduce a new class of automated experiments which use image-based feedback to find domain wall depinning fields. Initially, a domain is written in (110)-oriented PbTiO₃ thin films, and then scanned to locate domain wall, and then a bias pulse is applied at the center of the wall. If the wall is moved as a result of this pulse, the next section of the wall is scanned; else the bias pulse magnitude is increased until such time that movement is detected. In this way, the local depinning field can be mapped and correlated with the local domain structure. Extensions to finding the bias pulse amplitude-pulse width ‘phase diagram’ for wall depinning, via structured Gaussian processes³, or reinforcement learning, will be discussed.

References

1. 1. Meier, Dennis, and Sverre M. Selbach. Nature Reviews Materials 7.3 (2022): 157-173.
   
   2. Jesse, Stephen, Arthur P. Baddorf, and Sergei V. Kalinin. Applied physics letters 88.6 (2006): 062908.
   
   3. Ziatdinov, Maxim A., Ayana Ghosh, and Sergei V. Kalinin. Machine Learning: Science and Technology 3.1 (2022): 015022.