Where are Chips and High Tech Going?

At SAIC , we track many technologies in Horizon 2 (3-5 years out) and Horizon 3 (more than 5 years out), as some tapping out of Moore’s Law and Dennard Scaling begin to impact computing in general.  As Scotty told us on Star Trek, “[We] can’t change the laws of physics, keptin!”.  Light goes about 18 inches unimpeded in about 1 nanosecond, near the cycle time of today’s chips.  So, it becomes a fancy dance of conduction, induction, and routing to ensure signals get from point A to point B on a chip, which may have hundreds  and even thousands of miles of wiring on board.  So, where do we go if today’s chip technologies can’t ultimately help us to sufficiently increase system performance?  We need to track a number of innovations in chip materials, chip ground rules (how chips are laid out and associated logic and physical design rules), and manufacturing.  Here are the “top 5” innovations and advancements in each of those areas that I’m currently tracking.  Whether it’s high performance #HPC or #quantum computing, or scaling up #AI and data at the #edge, these technologies will be the foundation of Horizons 2 & 3 mission systems integration.

Materials Innovations

1. Hafnium Oxide (HfO2) High-k Dielectrics Hafnium oxide is used as a high-k dielectric material in advanced CMOS technology nodes, replacing silicon dioxide to reduce gate leakage and allow for further scaling of transistors.
2. Gallium Nitride (GaN) GaN is used for high-power and high-frequency applications due to its superior efficiency and thermal properties compared to silicon, enabling advancements in power electronics and RF components.
3. Graphene Graphene's exceptional electrical conductivity and mechanical strength make it a promising material for transistors, interconnects, and sensors, potentially surpassing the limitations of silicon.
4. Phase-Change Materials Phase-change materials, such as Ge2Sb2Te5 (GST), are utilized in non-volatile memory technologies (e.g., PCM), offering faster write speeds and better scalability than traditional flash memory.
5. Two-Dimensional Materials (2D Materials) Materials like molybdenum disulfide (MoS2) and tungsten diselenide (WSe2) are investigated for use in next-generation transistors due to their excellent electronic properties and atomic-scale thickness.

Ground Rule Innovations

1. Gate-All-Around (GAA) FETs GAA FETs provide better control over the channel by surrounding it with gate material, improving performance and scaling beyond FinFET technology.
2. FinFET Technology FinFETs reduce short-channel effects and leakage current, allowing for further scaling and improved performance over traditional planar transistors.
3. Extreme Ultraviolet Lithography (EUV) EUV lithography uses shorter wavelengths (13.5 nm) to create smaller and more precise features on semiconductor wafers, enabling the advancement of technology nodes beyond 7nm.
4. Horizontal Nanowires Horizontal nanowire transistors offer superior electrostatic control and performance, facilitating continued scaling in advanced CMOS technology.
5. Neuromorphic Computing Neuromorphic chips mimic the brain's neural networks to improve computational efficiency and energy consumption for AI applications.

Manufacturing Technologies

1. 3D Stacking and Through-Silicon Via (TSV) 3D stacking integrates multiple layers of semiconductor devices vertically, connected by TSVs, improving performance, reducing latency, and increasing device density.
2. Directed Self-Assembly (DSA) DSA uses the self-organizing properties of block copolymers to create highly regular patterns at the nanoscale, complementing traditional lithography techniques for more accurate and efficient manufacturing.
3. Silicon Photonics Silicon photonics integrates optical components with electronic circuits on a single chip, enhancing data transfer speeds and bandwidth for networking and data center applications.
4. Quantum Dots Quantum dots are nanoscale semiconductor particles with tunable electronic and optical properties, being developed for applications in quantum computing and advanced display technologies.
5. Spintronics Spintronics utilizes the intrinsic spin of electrons in addition to their charge for information processing, promising lower power consumption and faster processing speeds for memory and logic devices.