Multiscale Investigations on Si-integrable Ferroelectric Hafnia-Zirconia Systems: From Fundamental Understanding to Everyday Electronics

Project Information

FERHAZ

Grant agreement ID: 794954

DOI

10.3030/794954

Project closed

EC signature date 21 February 2018

Start date 1 May 2018

End date 30 April 2020

Funded under

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Total cost

€ 165 598,80

EU contribution

€ 165 598,80

Coordinated by

RIJKSUNIVERSITEIT GRONINGEN
Netherlands

Periodic Reporting for period 1 - FERHAZ (Multiscale Investigations on Si-integrable Ferroelectric Hafnia-Zirconia Systems: From Fundamental Understanding to Everyday Electronics)

Reporting period: 2018-05-01 to 2020-04-30

Summary of the context and overall objectives of the project

"Home electronic devices contribute to an increasing part of residential electricity usage (50% by 2030 as estimated by international energy agency). The “endless” shrinkage of transistors to pack and process more information enhances these static losses. Ferroelectric materials and corresponding low-power electronic devices provide neat solutions to minimize these losses. However conventional ferroelectric materials lose their properties when miniaturized, limiting the microelectronic applications of these materials.

The material systems explored in FERHAZ are Hafnia based ferroelectrics. These are Si compatible simple oxides. Ferroelectricity in these materials becomes robust at nanaoscale, quite opposite to conventional ferroelectrics. FERHAZ included thorough investigations that lead us to gain key insights into this ""new kind"" of ferroelectricity. The project began with systematic studies on the effect of strain, and size in stabilizing a novel ferroelectric phase in Hafnia-based compounds. Guidelines to stabilize this ""new kind"" of ferroelectricity were arrived at. These films were grown then epitaxially on Si, and ferroelectric properties of such capacitors were investigated. Ferroelectric tunnel junctions were explored as potential low-power memories, and the structure-property studies were carried out through state-of-the art electron microscopy. The important role of oxygen vacancies and their migration is crucial for this new kind of ferroelectricity.

The conclusions from FERHAZ has the potential to inspire the design of other simple oxide ferroelectric systems."

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

Hafnia-Zirconate (HZO) system was chosen as the thin-film of choice. In the first part of FERHAZ the growth of this system was performed on various substrates: perovskites, hexagonal substrates wuth an aim to understand the effect of size and strain in the stabilization of ferroelectric (FE) phases. On SrTiO3 (STO) substrate, by using a correlated metal (La0.67Sr0.33MnO3) as the back electrode (image 1), we were able to demonstrate unequivocally a new FE rhombohedral phase in HZO [1], with the largest FE polarization in HZO. By systematically optmizing the growth on various substrates, susbequently we were able to understand and provide suitable crystallographic guidelines for the stabilization of this phase, understanding that the stability of this new phase relies on carefully engineering biaxial compressive strain, and a favorable crystallographic orientation at nanoscopic domain sizes (<10 nm). State-of-the-art aberration corrected microscopy techniques such as differential phase contrast STEM, provided unprecedented insights about the crystallographic polarization in these systems [3, 5]. (Image 2)
Taking lessons from these studies, next we were able to carefully engineer this new FE phase directly on Si, and study the FE properties of metal-insulator-semiconductor capacitors. Although it was not possible to engineer just a pure FE phase owing to the low strain that Si substrate provides to the films, the direction to optimize the growth in this way has been set [2].

From a device perspective, FE (and multiferroic) tunnel junctions have been fabricated with rhombohedral HZO tunnel layer sandwiched between Co and LSMO, both ferromagnetic metals [5]. Structure-property analysis of these devices and their failure mechanisms has been strudied through a combination of electrical characterization including memristive hysteresis, endurance, fatigue measurements, and atomic resolution STEM techniques and spectroscopy techniques such as iDPC-STEM, HAADF-STEM and energy dispersive analysis using the Themis Z microscope purchased at the University of Groningen. The influence of oxygen vacancy migration facilitated by LSMO is very crucial to explain the tunnel resistance properties observed in these devices [manuscrript under preparation].

In-situ biasing and heating microscopy protocols (sample preparation, and measurement set up)- which do not exist previously- were being designed to carry out operando studies of device operation inside the electron microscope. However, the COVID pandemic stopped the progress on this towards the end of FERHAZ, only to recently resume (after FERHAZ).

FERHAZ has resulted in 5 direct publications (thus far) in good journals, with ref [2, 5]. These articles have gained significant public interest. For e.g. the discovery of the rhombohedral phase for the first time is reported here: https://www.eurekalert.org/pub_releases/2018-10/uog-nfb101918.php and inside the university media channels here: https://www.rug.nl/sciencelinx/nieuws/2018/10/20181022_noheda?lang=en
Our results on direct epitaxy of HZO on Si is explained in the ACS live slides platform here: https://pubs.acs.org/doi/suppl/10.1021/acsaelm.9b00585/suppl_file/el9b00585_liveslides.mp4

Our paper on multiferroic tunnel junctions was the editors suggestion in physical review applied in Sept 2019
In addition information about these works have been shared on social media platforms such as facebook and linkedin.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

The following are the important discoveries arising directly out of FERHAZ:
1. Discovery of a novel polar rhombohedral phase in HZO thin-films, which shows huge polarization. The stability and guiding principles of that give rise to this metastable phase have been thoroughly investigated. Furthermore, this phase has been replicated and investigated by several other labs too (University of Twente, Barcelona etc).
2. The understanding about nanoscopic and new kind of ferroelectricity as the effect of interplay between metastability, size, strain and oxygen vacancies is much clearer owing to the the investigations on pure and single-phase FE samples synthesized over the course of FERHAZ.
3. The influence of oxygen vacancy migration facilitated by oxygen conductor LSMO in Co/HZO/LSMO tunnel junction devices. These insights are very crucial to help design newer nanoscopic ferroelectric materials systems, and corresponding devices.
4. Direct epitaxial growth of HZO directly on Silicon. This is quite crucial to take these materials towards microelectronic applications and in creating a resurgence in ferroelectric materials.

FERHAZ has resulted in creating two off-shoot full scale masters projects at the University of Groningen.
1. Hafnia based ferroelectric tunnel junctions (E Stylanidis) and
2. Depolarization effects on the stability of the rhombohedral phase in HfZrO4 ferroelectric thin-films (V de Haas).

FERHAZ is a part of the wider scheme of EU projects- at a scientific level- to create a resurgence of ferroelectrics in microelectronics. The insights gained in this project opens avenues not only towards efficient current day computational hardware and materials, but also towards newer schemes of computation such as neuromorphic computation. Memristors such as ferroelectric tunnel junctions- studied here- are important components of neuromorphic computation. FERHAZ has been carried out with close collaboration with Cognigron: Groningen center for cognitive materials, a novel center dedicated towards engineering materials that have the capability to self-learning materials and networks that can compute analogous to human brain [REFER]. These efforts are directed towards designing systems that perform low-power computation.

Differential scanning TEM image of HZO on GaN. Polarization can be estimated as ~2 uC/cm2

High resolution scanning TEM image of HZO grown on STO with LSMO as a back electrode using PLD