MRS Meetings and Events

 

EL20.11.02 2023 MRS Fall Meeting

Wake-Up of Ferroelectric HZO Films -- Towards a BEoL Compatible Wake-Up Free Material

When and Where

Nov 30, 2023
4:15pm - 4:30pm

Hynes, Level 3, Room 301

Presenter

Co-Author(s)

David Lehninger1,Ayse Sünbül1,Anant Rastogi1,Pratik Bagul1,Shouzhuo Yang1,Konrad Seidel1,Maximilian Lederer1

Fraunhofer IPMS1

Abstract

David Lehninger1,Ayse Sünbül1,Anant Rastogi1,Pratik Bagul1,Shouzhuo Yang1,Konrad Seidel1,Maximilian Lederer1

Fraunhofer IPMS1
The discovery of ferroelectricity in thin doped hafnium oxide films has revived interest in the concept of ferroelectric (FE) memory. Zirconium-doped hafnium oxide (HZO) crystallizes at low temperatures (400°C and below), making this material interesting for back-end-of-line (BEoL) implementation. Metal-FE-Metal (MFM) capacitors are critical building blocks for the realization of BEoL-compatible FE memory concepts. Located in the BEoL, they can be connected to either the gate or drain of a standard logic device to realize a one transistor-one capacitor (1T1C) FE field effect transistor (FeFET) or a 1T1C FE random access memory (FeRAM), respectively. [1]<br/>Since its discovery, researchers have made great efforts to improve important properties of hafnium oxide-based ferroelectrics, such as remanent polarization, endurance, retention, imprint, and wake-up. Typical approaches include, but are not limited to, the investigation of different dopant elements, dopant concentrations, film thicknesses, and stacking options such as interface/electrode materials and superlattice structures.<br/>While much progress has been made, certain issues, such as wake-up, remain challenging. Wake-up is the opening of the initially pinched polarization versus electric field hysteresis loop, which leads to a gradually increasing remanent polarization during field cycling until saturation. Since memory applications require stable polarization, an initial field cycling ("wake-up") is necessary. However, the need for this wake-up makes the realization of hafnium oxide based FE memories more complex and reduces the maximum number of endurance cycles. Therefore, a wake-up-free material would be highly desirable.<br/>Several theories have been proposed to explain the wake-up effect. These theories can essentially be summarized as follows [2]: (i) defect redistribution, which reduces internal bias fields and promotes the depinning of domain walls, (ii) defect redistribution and subsequent stabilization of the orthorhombic phase ("field-induced phase transformation"), and (iii) 90° domain wall motion ("ferroelastic switching").<br/>The wake-up effect of FE-HZO films could be mitigated by: (i) using mesa-etched test structures [3]; (ii) reducing the number of defects and carbon contamination by applying longer ozone pulses durations during the atomic layer deposition process [4]; (iii) interface engineering by plasma treatment [5] or epitaxial growth [6].<br/>Herein, we investigate the effect of post-deposition annealing temperature, film stoichiometry, film thickness, and co-doping on the wake-up characteristics of FE HZO films embedded in MFM capacitors. Several structural and electrical characterization techniques, such as X-ray diffraction and polarization versus electric field measurements, are applied to obtain a comprehensive overview. The goal is to present a BEoL-compatible material stack with significantly reduced wake-up without compromising other important parameters such as remnant polarization and reliability.<br/><br/>[1] D. Lehninger et al., <i>2021 IEEE IITC</i>, Kyoto, Japan, 2021, pp. 1-4.<br/>[2] M. Lederer et al., 2021, PSS RRL, 15: 2100086.<br/>[3] J. Bouaziz et al., ACS Appl. Electron. Mater. 2019, 1, 9, 1740–1745<br/>[4] A. Kashir et al., Adv. Eng. Mater., 23: 2000791.<br/>[5] K. -Y. Chen et al., 2017 VLSI Circuits, Kyoto, Japan, 2017, pp. T84-T85<br/>[6] J. Lyu et al., ACS Appl. Electron. Mater. 2019, 1, 2, 220–228

Keywords

atomic layer deposition | oxide

Symposium Organizers

Gina Adam, George Washington University
Sayani Majumdar, Tampere University
Radu Sporea, University of Surrey
Yiyang Li, University of Michigan

Symposium Support

Bronze
APL Machine Learning | AIP Publishing

Publishing Alliance

MRS publishes with Springer Nature