Intel Foundry recently announced major advancements in semiconductor technology at the IEEE International Electron Devices Meeting 2024. These innovations aim to push the industry towards achieving 1 trillion transistors on a chip by 2030, a necessity for developing more efficient, high-performing computing technologies like AI.
Key breakthroughs presented include:
- Subtractive Ruthenium: Intel has introduced subtractive ruthenium as a new material for chip interconnects. This approach can reduce capacitance by up to 25%, potentially enhancing performance and paving the way for denser chip designs.
- Selective Layer Transfer (SLT): This method promises a 100x increase in the speed of chip-to-chip assembly. It allows for the creation of ultra-thin chips that are more flexible and efficient, potentially benefiting AI applications.
- Silicon RibbonFET CMOS: Intel showcased further advancements in transistor scaling with silicon RibbonFET CMOS technology. This development is crucial for continuing the trend of miniaturising components, a key aspect of Moore’s Law.
- Gate Oxide for 2D FETs: The company is exploring new materials for transistors, focusing on improving the gate oxide in scaled 2D FETs. This research could lead to a future where silicon is replaced by more advanced materials.
Additionally, Intel is advancing its research in gallium nitride (GaN) technology, which holds promise for higher performance in power and RF electronics.
Intel also highlighted its vision for future semiconductor development, focusing on innovations in memory integration, interconnect bandwidth, and modular expansion. The company emphasised the need for creating transistors that operate at ultra-low voltages to overcome thermal challenges and enhance energy efficiency.
These advancements underscore Intel’s commitment to leading the semiconductor industry into a new era, driven by innovative technology and the support of initiatives like the U.S. CHIPS Act.