Ferroelectric domain imaging and characterization of hafnia thin films for integrated ferroelectrics

Abstract

As the market of IoT semiconductors, artificial intelligence hardware grows, non-volatile memory, high performance, and low power memory is required. For this purpose, FRAM has a fast operation speed and has an advantage in cell size, but the existing ferroelectric materials have been difficult to apply due to limitations in CMOS process compatibility, device scaling, and thermal stability. However, with the introduction of ferroelectric hafnia in 2011, hafnia research as a next-generation memory is rapidly emerging. In this paper, first, to understand atomic layer deposition (ALD), a study on 3D TiN nanotubes and ZnO truss structures was conducted using ALD deposition. After improving the understanding of ALD, a HfxZr(1-x)O2(HZO) thin film was deposited by ALD, and changes in the crystal structure and macroscopic electrical properties of ferroelectric hafnia according to heat treatment and changes in electrodes were confirmed. Second, the fabrication of ferroelectric HZO(HfxZr(1-x)O2) thin films using H2O2 as an oxidant was developed using ALD. After deposition of ferroelectric hafnia, the heat treatment temperature was varied, and the GI-XRD, P-V, and C-V analysis confirmed the ferroelectric properties from 500 to 800 degrees. In the case of HZO deposited with H2O2, it was confirmed that the leakage current through P-V and C-V was large, and it was found through the TOF-SIMS that HZO using H2O2 had extensive hydrogen incorporation. To improve this, we checked whether the hydrogen-related bonds were effectively removed from the thin film by changing the electrode and the hafnia interface and O2 process. To confirm the microscopic ferroelectric properties of HZO according to the previous process, domain analysis and switching analysis were performed using a piezoresponse force microscope. Therefore, we confirmed that the RTA 600℃ or less is the optimal condition in terms of domain size by confirming the change of the domain of various process conditions. Also, by analyzing the switching characteristics of the nano-scale, in the case of the HZO thin film, sideways domain growth was confirmed. From Multi-scale visualization of the physical properties according to the process change, it is possible to present a structure-process-physical correlation. As a result, guidance for optimization of hafnia thin film processing for the integrated ferroelectrics process will be presented.