Alternate and steep slope devices

Shriram Ramanathan, Yizhuo Zhou, Omor Shoron, Mohamed Seghir Boucherit, Christopher M. Jackson, Tyler A. Cain, Malinda L. C. Buffon, C. Polchinski, Susanne Stemmer, Siddharth Rajan, Anant D. Bhave, M Li, David Esseni, D. Jena, Hongyun Xing, Anna Szabó, Steven J. Koester, Mathieu Luisier, W Li, Tao Yu, Judy L. Hoyt, Patrick Fay, S. Richter, C. Schulte‐Braucks, Lars Knoll, Gia Vinh Luong, A. Schäfer, Stefan Trellenkamp, Qing‐Tai Zhao, S. Mantl, P. Grünberg, Sonali Agarwal, Eli Yablonovitch, Ziming Hu, Raj K. Jana, Meng Qi, S. Ganguly, Bo Song, E. Kohn2014 · DOI

This document is not a single research paper but rather the official program for Session II.A: Alternate and Steep Slope Devices from a 2014 electronics conference (likely the Device Research Conference). It provides a snapshot of the "Beyond-CMOS" landscape, focusing on devices designed to overcome the fundamental efficiency limits of traditional silicon transistors.

TL;DR

This session explores the frontier of ultra-low-power electronics by showcasing devices that bypass the "Boltzmann Tyranny"—the physical limit that prevents standard transistors from switching faster than 60 mV/decade. By leveraging quantum tunneling, phase-change materials like Mott insulators and $VO_2$, and novel 2D heterostructures, these researchers aim to enable logic circuits that operate at a fraction of the voltage used in today’s chips.

What Problem Does This Paper Address?

For decades, the semiconductor industry has struggled with the "power wall." Traditional Metal-Oxide-Semiconductor Field-Effect Transistors (MOSFETs) are limited by thermionic emission; at room temperature, it takes at least 60 mV of gate voltage to change the source-drain current by one order of magnitude. This is known as the subthreshold swing (SS) limit.

As we try to shrink devices and lower supply voltages ($V_{DD}$) to save energy, this 60 mV/decade limit forces a trade-off: either the device leaks too much power when "off," or it doesn't turn "on" strongly enough to perform quickly. The papers in this session address this by investigating "Steep Slope" devices—architectures that use alternative physics to achieve an SS below 60 mV/decade, potentially allowing for high-performance computing at ultra-low voltages.

Methods

The session covers a diverse array of experimental and theoretical approaches:

  • Tunnel Field-Effect Transistors (TFETs): Instead of pushing electrons over a barrier (thermionic), these devices use quantum mechanical tunneling through a barrier. Papers II.A-4 through II.A-7 explore TFETs using 2D materials, III-V semiconductors (InGaAs/GaAsSb), and lateral heterojunctions.
  • Phase-Change & Mott Transitions: Paper II.A-1 and II.A-10 examine materials that can undergo a rapid transition from insulator to metal (like Mott transistors and $VO_2$ switches), providing a "steep" change in conductivity.
  • Novel Heterostructures: Research into oxide interfaces (SrTiO3/GdTiO3) and GaN-based MOSHEMTs explores how high-density electron gases and specific material stacks can improve switching behavior.
  • Electromechanical Coupling: Paper II.A-8 introduces a Piezoelectric Transformer FET, using mechanical strain to modulate electrical behavior.

Datasets

As this is a collection of experimental device physics papers, "datasets" in the machine-learning sense are not used. Instead, the authors rely on:

  • Experimental Characterization: Electrical measurements (I-V and C-V curves) of fabricated prototype devices.
  • RF Performance Metrics: S-parameter measurements for high-frequency applications (up to 40 GHz for $VO_2$ switches).
  • TCAD Simulations: Used in several papers (e.g., II.A-5) to model the viability of new materials like metal-dichalcogenides before full-scale fabrication.

Key Insights and Ideas

  • The 0.15V Frontier: One of the most significant "aha" moments is the experimental demonstration of NAND and Inverter logic gates operating at a $V_{DD}$ of only 0.15V (Paper II.A-7). This is significantly lower than the ~0.7V-1.0V used in standard commercial processors.
  • 2D Materials for TFETs: There is a strong shift toward "Thin-TFETs" and metal-dichalcogenides. Because these materials are atomically thin, they allow for better electrostatic control, which is crucial for making tunneling efficient.
  • Beyond Logic: These steep-slope materials aren't just for CPUs. $VO_2$ is highlighted as a candidate for "reconfigurable electronics" in the 40 GHz range, suggesting applications in high-speed communications.
  • Breaking the Limit: Paper II.A-9 explicitly reports GaN-based devices achieving sub-threshold slopes steeper than the 60 mV/decade limit, proving that these theoretical concepts are becoming experimental realities.

Key Results

  • High Carrier Density: Modulation of over $10^{14} \text{ cm}^{-2}$ electrons at oxide heterojunctions, showing the potential for high-current "on" states in non-silicon devices.
  • Logic Validation: Successful demonstration of p-TFET logic operations (Inverters/NAND) at ultra-low supply voltages.
  • Frequency Range: $VO_2$ switches demonstrated for wide-band applications ranging from DC to 40 GHz.
  • Steep Switching: Multiple papers confirm experimental sub-60 mV/dec switching, the "holy grail" for low-power digital logic.

Limitations and Future Work

  • On-Current Challenges: While TFETs are great at turning "off" and switching steeply, they often struggle with low "on-current" compared to traditional MOSFETs. Several papers hint at the need to improve the drive current to make these devices competitive for high-speed computing.
  • Material Maturity: Many of these devices use exotic materials (Lithium Niobate, III-V compounds, 2D chalcogenides) that are difficult to integrate into standard high-volume silicon manufacturing (CMOS compatibility).
  • Scalability: While individual gates (NAND/Inverters) were demonstrated, scaling these to millions of transistors on a single chip remains a massive engineering hurdle not fully addressed in these early-stage reports.

Summary updated

BibTeX
@misc{ramanathan2014alternate,
  author = {Shriram Ramanathan and Yizhuo Zhou and Omor Shoron and Mohamed Seghir Boucherit and Christopher M. Jackson and Tyler A. Cain and Malinda L. C. Buffon and C. Polchinski and Susanne Stemmer and Siddharth Rajan and Anant D. Bhave and M Li and David Esseni and D. Jena and Hongyun Xing and Anna Szabó and Steven J. Koester and Mathieu Luisier and W Li and Tao Yu and Judy L. Hoyt and Patrick Fay and S. Richter and C. Schulte‐Braucks and Lars Knoll and Gia Vinh Luong and A. Schäfer and Stefan Trellenkamp and Qing‐Tai Zhao and S. Mantl and P. Grünberg and Sonali Agarwal and Eli Yablonovitch and Ziming Hu and Raj K. Jana and Meng Qi and S. Ganguly and Bo Song and E. Kohn},
  title = {{Alternate and steep slope devices}},
  year = {2014},
  doi = {10.1109/drc.2014.6872274},
}

Loading related content...

Figure