The following presentations by Namlab co-authors were excepted for the upcoming IEEE International Electron Devices Meeting (IEDM) 2019 in San Francisco on December 7-11.
Tutorial, December 7, 4:30 pm – 6:00 pm:
Ferroelectric Memories and Beyond, Johannes Műller, Globalfoundries and Thomas Mikolajick, NaMLab/TuDresden
DOWNLOAD ABSTRACT – Recent advances in scaling and CMOS-compatible implementation of ferroelectric thin films has sparked renewed interest to utilize the unique properties of these materials in advanced CMOS technology nodes. Led by the ferroelectric memory development and further fueled by new applications fields such as steep slope devices and neuromorphic applications, this field has seen a strong growth in R&D activity over the last decade. This tutorial will give an introduction to ferroelectric materials and devices with special emphasis on the utilization of hafnium oxide based thin films. The working principle as well as the challenges of capacitors based ferroelectric random access memory (FRAM), ferroelectric field effect transistor (FeFET) and ferroelectric tunnel junction (FTJ) will be reviewed. In addition, a brief outlook on beyond memory applications of CMOS-compatible ferroelectric thin films will be given.
Session 15 – Memory Technology – Ferroelectrics
Tuesday, December 10, 9:00 a.m.
9:05 AM 15.1 Material Perspectives of HfO2-based Ferroelectric Films for Device Applications
Akira Toriumi, Lun Xu, Yuki Mori, Xuan Tian, Patrick Lomenzo, Halid Mulaosmanovic, Monica
Materano, Thomas Mikolajick, Uwe Schroeder, The University of Tokyo, TU-Dresden
This paper gives material fundamentals and new insights to ferroelectric HfO2 for device applications. The
key role of dopants, effects of the interface on ferroelectric phase, and a detailed discussion of switching
kinetics are of central focus. Based on them, we discuss opportunities of ferroelectric HfO2 for device
11:10 AM 15.5 Next Generation Ferroelectric Memories Enabled by Hafnium Oxide (Invited)
Thomas Mikolajick, Uwe Schroeder, Patrick Lomenzo, Evelyn Breyer, Halid Mulaosmanovic, Michael
Hoffmann, Terence Mittmann, Furqan Mehmood, Benjamin Max, Stefan Slesazeck, NaMLab gGmbH,
Technische Universtität Dresden
Ferroelectrics are an ideal solution for low write power nonvolatile memories. The complexity of
ferroelectric perovskites has hindered the scaling. Ferroelectricity in hafnium oxide solved this issue
making ferroelectric memories in its three variants, ferroelectric RAM, ferroelectric field effect transistors
and ferroelectric tunneling junctions interesting for future memory solutions again.
12:00 PM 15.7 Demonstration of BEOL-Compatible Ferroelectric Scaled Hf0.5Zr0.5O2 FeRAM
Co-Integrated with 130nm CMOS for Embedded NVM Applications
Terry Francois, Laurent Grenouillet, Jean Coignus, Philippe Blaise, Catherine Carabasse, Nicolas
Vaxelaire, Thomas Magis, François Aussenac, Virginie Loup, Catherine Pellissier, Stefan Slesazeck,
Viktor Havel, Claudia Richter, Adam Makosiej, Bastien Giraud, Evelyn Breyer, Monica Materano, Philippe
Chiquet, Marc Bocquet, Etienne Nowak, Uwe Schroeder, Fred Gaillard, CEA-Leti, NaMLab gGmbH, Aix-
We demonstrate scalability of HZO capacitors down to 300nm by co-integrating them for the first time in
Back-End-Of-Line of 130nm CMOS technology. Excellent performance are reported: 2.PR >40μC/cm²,
endurance >1011, switching speeds <100ns, operating voltages <4V, and data retention at 125°C paving the
way towards <10fJ/bit ultra-low power FeRAM for IoT applications.
Session 38 – Memory Technology – Memory for Neural Network
Wednesday, December 11, 1:30 p.m.
3:40 PM 38.6 A 2TnC Ferroelectric Memory Gain Cell Suitable for Compute-in-memory and Neuromorphic Application
Stefan Slesazeck, Taras Ravsher, Viktor Havel, Evelyn Breyer, Halid Mulaosmanovic, Thomas Mikolajick, NaMLab gGmbH, TU Dresden
A 2TnC ferroelectric memory gain cell is proposed, that can be operated either in FeRAM or FTJ mode. The internal gain of the cell mitigates the need for 3D integration of the FeCAPs, making the concept very attractive for embedded memories, compute-in-memory and neuromorphic applications.