Dopants, Defects and Domains in Wurtzite Ferroelectrics

AlN-based thin films have dominated the market of piezoelectric microelectromechanical system (MEMS) resonators for many years, and they have recently attracted increased interest for their ferroelectric properties and potential as ultrawide bandgap (UWBG) semiconductors. Regardless of the intended application, it is important to understand the effects of defects, including intentional dopants or alloying elements, adventitious oxygen or other contaminants, dislocations or grain boundaries from growth, and a number of other possibilities. Crucially, it is important to remember that an undesirable defect in one application may be entirely inconsequential or even enabling in another.

We focus here on ferroelectric and piezoelectric applications, learning from but not limited to the work in the UWBG community. In particular, we focus on the presence and impacts of structurally-disruptive isovalent (e.g., Sc, La, Gd for Al) substitutions and on nominally donor (e.g., O for N; Si, Hf, Zr for Al) defects. We show that the large bandgap of AlN along with the naturally low carrier mobilities of sputtered films combine to provide a great deal of flexibility for both substitutional and non-stoichiometric alloy/defect engineering AlN-based films for ferroelectric and piezoelectric applications. In fact, we show that films with upwards of 10% donor substitution can not only remain sufficiently insulating for piezoelectric and/or ferroelectric applications but also exhibit reduced coercive field and/or increased piezoelectric response over baseline AlN. Work from our group and others is also revealing effects of such defects and non-stoichiometry on polarity inversion, both static during growth and dynamic during ferroelectric switching, and associated impacts on domain walls and inversion domain boundaries.