The Center for E3S combines a broad research and education approach to revolutionize two major fundamental components of digital information processing systems: the communications logic switch and the short-medium range communication of information between logic elements. These goals are pursued by four Research Themes. Three themes focus on logic switching (Nanoelectronics, Nanomechanics, and Nanomagnetics), while an optical approach is used for communication of information (Nanophotonics). The Center for E3S addresses these grand challenges by a collaborative approach, involving engineers, chemists, physicists, and materials scientists from five member institutions (University of California at Berkeley, Massachusetts Institute of Technology, Stanford University, University of Texas El Paso, and Florida International University).
The Fundamental Issue
Binary bits (1's and 0's) are the fundamental units used to process and store information in modern electronic devices. Transistors are electrical switches that mediate the two binary states (current ON and current OFF). Integrated circuits (IC), composed of transistors connected by wires (or “interconnects”), are the cornerstone of the modern computer.
The performance of ICs has improved as they become smaller. Miniaturization also has led to a decrease in power consumption. However, in the past decade, the semiconductor industry has faced severe challenges by the power density of increasingly complex IC chips. In addition, recent advancements such as cloud computing, social networking, mobile internet, wireless sensor swarms, and body-centered networks have further accelerated demand for ever increasing functionality from a fixed amount of battery energy.
Based on the realization that the energy used to manipulate a single bit of information is currently ~105 times greater than the theoretical limit, the Center for E3S has developed an aggressive and disruptive approach to close this gap. The Center for E3S seeks technological breakthroughs for two fundamental components of digital information processing systems: the communications logic switch (transistor) and the short-medium range communication of information between logic elements.
The conventional transistorsuffers from a serious drawback: its conduction is thermally activated. As a consequence, powering voltages of close to 1 Volt are required to provide a good On/Off current ratio, even as transistors become smaller and smaller. The wires connecting the transistor, however, could operate with a very good signal-to-noise ratio, even at voltages below 10 mmili-Volt. Since energy consumption is proportional to the square of operating voltage, the energy currently used to manipulate a single bit of information is several orders of magnitude greater than needed in an ideal system. Therefore, a more sensitive, lower-voltage switch is needed as the successor to the conventional transistor.
Besides energy for switching, an IC chip also consumes energy for communications among the transistors, through the interconnects. The power consumed by interconnects in a microprocessor chip, as high as 50% even a decade ago, has continued to rise steadily to a larger portion of the total consumed power.
Research Goals and Objectives
The research goals have been set from fundamental consideration on the minimum energy needed per digital function. Part of the goal is to explore the lowest limits of each experimental approach compared to the fundamental limits, and to elucidate the factors that determine the minimum energy that each approach can reach.
With respect to a new switch, the Center shares a common objective in the form of the following specifications.
- Sensitivity: ~1 milli-Volt/decade, allowing switches with a swing of only few milli-Volts.
- On/Off current ratio: 106:1
- Current or conductance density (for miniaturization): 1 mS/μm; i.e., a 1μm device should conduct at ~1 kΩ in the on-state.
For optical interconnects to be a low power consumption alternative, the Center’s high level goal is to achieve close to quantum limit detection (20 photons/bit) and atto-Joule/bit communication (~10 aJ/bit).
In addition, the Center recognizes that its research in novel electronics components must be guided by circuit/system perspectives. The detailed technical goals/specifications for the novel electronic components in the context of a circuit has been itself a research endeavor. This System Integration research activity is evolving towards developing new circuit architectures and circuit demonstrations in the Center's second five years.