New semiconductor material developed
Researchers at Georgia Institute of Technology collaborated with Tianjin University to develop a new semiconductor material. Using this approach, it should be possible to fabricate chips that switch and operate several orders of magnitude faster than current devices, which could lead to significant changes in electronics.
Moore's law facing limits
According to Moore's law, the number of transistors on a chip roughly doubles every two years. This observation was first described by Gordon Moore in 1965, and Moore later acknowledged that the pace would eventually slow. Increasing the transistor count on a single silicon chip is becoming progressively more difficult and expensive, and physical limits are being approached.
Continued transistor scaling
Companies such as TSMC, IBM, and Intel continue to invest heavily to shrink transistor dimensions further so that Moore's law remains useful. Current transistor densities are on the order of hundreds of millions per square millimeter, an impressive figure. Over the past 50 years, transistor density per square millimeter has increased by several hundred thousand times.
3D transistors raise thermal challenges
To increase device density, manufacturers have started building transistors in three dimensions and stacking them vertically. Vertical stacking packs more transistors into the same footprint, but it introduces severe thermal management challenges because there is no straightforward heat dissipation path, which can lead to overheating. IBM and others are exploring new architectures, such as vertical transistors, but these devices remain largely at the research stage.
Graphene offers strong thermal conductivity
Researchers are also searching for alternative semiconductor materials that can offer better thermal performance than silicon. Graphene has attracted attention due to its crystalline structure and exceptional thermal conductivity. Its thermal conductivity is much higher than silicon, so graphene-based devices could offer better heat management.
Zero bandgap problem in graphene
Graphene's advantages include potential for smaller, faster devices with reduced heat, and electronic properties not available in silicon. However, graphene has a fundamental drawback: it is not a semiconductor in the conventional sense because it has a zero-electron-volt bandgap. Modern chips such as CPUs and GPUs are built from transistors, which are essentially switches controlled by voltage. The presence of a bandgap is crucial because it determines the minimum energy required to switch between on and off states. Ideally, a transistor bandgap should be small enough to allow conduction when desired but large enough to provide an off state. Graphene's zero bandgap means it behaves like a semimetal with high conductivity, which prevents conventional transistor operation.
SEG: semiconducting epitaxial graphene
Researchers found a way to produce a graphene-based material that behaves like a semiconductor. In production, silicon carbide is heated above 1,000°C and then doped. Just as with modern silicon, doping changes the electronic properties of the material. By modifying graphene's properties in this way, the researchers introduced a bandgap. They call this new material semiconducting epitaxial graphene, or SEG.
Faster switching rates
SEG transistors are built on a silicon carbide substrate. When the gate-source voltage exceeds a threshold, the transistor conducts and current flows. Graphene's flatness and band structure enable ballistic transport, meaning electrons can move faster through the semiconductor. This allows higher switching speeds and reduced heat generation. These transistors can switch very rapidly, potentially supporting frequencies in the terahertz range.
Compatibility with conventional processes
Another advantage of this work is that the fabrication process used is compatible with conventional manufacturing techniques. Economically, the approach appears feasible. The researchers report not only creating graphene-based chips in the lab but also successfully producing them at scale on large silicon carbide wafers without defects.
