Quantifying the Electrode Capping Effect’s Role on Phase Stability in Ferroelectric Hafnium Zirconium Oxide

Ferroelectric hafnia is poised to enable a next generation of microelectronic devices owing to its unprecedented compatibility with mainstream semiconductors and ability to maintain ferroelectricity to single unit-cell dimensions without the necessity of epitaxy. Realizing broad implementation, however, likely requires phase pure materials. It is widely known that ferroelectric hafnium oxide-based films exhibit larger orthorhombic phase fractions and polarization responses when crystallized with a top electrode in place. The top electrode is described as providing a ‘capping effect’ that imparts stress to the hafnia film upon cooling and prevents transformation into the non-ferroelectric monoclinic phase. While compelling evidence of the importance of the top electrode has been presented previously, the quantitative impacts of its role on phase stability are lacking. In this work, the biaxial stresses imparted by the electrode clamping effect are quantified using X-ray diffraction sin²(Ψ) analyses measured on 20 nm thick Hf_0.5Zr_0.5O₂ (HZO) films processed on TaN bottom electrodes with and without top TaN electrodes. Both HZO films are in a state of biaxial stress after crystallization, however the film with a top electrode has a 70% larger (3.5 GPa) stress magnitude than the film without a top electrode (2.1 GPa). Furthermore, the film processed with a top electrode is nearly 100 vol.% the ferroelectric phase whereas the film processed without the top electrode contains only 60 vol.% of the ferroelectric phase. Removal of the top electrode results in a reduction in tensile stress in the hafnia layer by 0.8 GPa and is accompanied by a 10% reduction in ferroelectric phase fraction. On the basis of strain energy reduction, we calculate an energy barrier for the total orthorhombic to monoclinic phase transformation to be 12 meV/f.u. This quantitative assessment of the clamping effect is accompanied by wafer flexure-based stress measurements to quantify stress development for each processing step, as well as temperature-dependent X-ray diffraction and wafer flexure measurements that allow us quantify the temperatures that accompany phase formation and stress evolution. We will show that it is not an in-plane mechanical constraint imposed by the top ‘capping’ electrode that leads to the large biaxial tensile stress in the hafnia layer. Rather, the electrode serves to limit out-of-plane displacements and prevent formation of the monoclinic phase. This work provides a foundation by which phase-pure ferroelectric hafnia can be prepared by careful consideration of top electrode deposition and processing.