HfO2-based ferroelectric tunnel junctions (FTJs) represent a unique opportunity as both a next-generation digital non-volatile memory and as synapse devices in braininspired logic systems, owing to their higher reliability compared to filamentary resistive random-access memory (ReRAM) and higher speed and lower power consumption compared to competing devices, including phase-change memory (PCM) and state-of-the-art FTJ. Ferroelectrics are often easier to deposit and have simpler material structure than films for magnetic tunnel junctions (MTJs). Ferroelectric HfO2 also enables complementary metal-oxide-semiconductor (CMOS) compatibility, since lead zirconate titanate (PZT) and BaTiO3-based FTJs often are not.
No other groups have yet demonstrated a HfO2-based FTJ (to best of the author’s knowledge) or applied it to a suitable system. For such devices to be useful, system designers require models based on both theoretical physical analysis and experimental results of fabricated devices in order to confidently design control systems. Both the CMOS circuitry and FTJs must then be designed in layout and fabricated on the same die.
This work includes modeling of proposed device structures using a custom python script, which calculates theoretical potential barrier heights as a function of material properties and corresponding current densities (ranging from 8×10^3 to 3×10^(−2) A/cm2 with RHRS/RLRS ranging from 5×10^5 to 6, depending on ferroelectric thickness). These equations were then combined with polynomial fits of experimental timing data and implemented in a Verilog-A behavioral analog model in Cadence Virtuoso. The author proposes tristate CMOS control systems, and circuits, for implementation of FTJ devices as digital memory and presents simulated performance. Finally, a process flow for fabrication of FTJ devices with CMOS is presented. This work has therefore enabled the fabrication of FTJ devices at RIT and the continued investigation of them as applied to any appropriate systems.
Library of Congress Subject Headings
Nonvolatile random-access memory--Materials; Ferroelectric thin films; Metal oxide semiconductors, Complementary--Design and construction
Microelectronic Engineering (MS)
Department, Program, or Center
Microelectronic Engineering (KGCOE)
Pringle, Spencer Allen, "Modeling and Implementation of HfO2-based Ferroelectric Tunnel Junctions" (2017). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus