Future Materials for Beyond Si Integrated Circuits: A Perspective

The integration of novel materials has been pivotal in advancing Si-based devices ever since Si became the preferred material for transistors, and later, integrated circuits. New materials have rapidly been adopted in recent decades to enhance the performance of Si integrated circuits. The imperative to uphold Moore's Law for both More Moore and More than Moore devices has driven the industry to study, and later introduce a plethora of materials and innovative processes into the Si fabrication process, spanning from the front-end-of-line (FEOL) to the back-end-of-line (BEOL). This concerted effort aims to bolster computing power and functionality while curbing costs. Scaling Si-channel transistors down to the nanometer level has presented formidable challenges. The emergence of new materials, such as two-dimensional materials like transition metal dichalcogenides, carbon nanotubes, and metal oxides holds promise for further scaling endeavors. With transistors and interconnects nearing their physical limits, these materials offer potential solutions by enabling the fabrication of high-performance devices without relying solely on Si, while integrated at lower thermal budgets. Moreover, these technologies can be repurposed in the BEOL to add extra functionality while reducing the overall device footprint. Recent breakthroughs, notably the successful demonstration of high-performance devices utilizing ALD metal oxides like In₂O₃, have sparked considerable excitement. Addressing the scaling challenges of interconnects is equally daunting. The quest for materials with lower resistivities than copper interconnects with reduced electromigration at scaled dimensions and efforts to eliminate or minimize barrier layers hold promise in mitigating RC time delay. Non-volatile memories, particularly ferroelectric-based memories, stand to be gained from advancements in materials science. Innovations in such materials as hafnates and enhanced integration techniques for perovskites through electrode stack engineering could facilitate the scaling of current ferroelectric memories. The ongoing introduction of new materials is poised to sustain scaling efforts and unlock novel functionalities in electronic devices for many years.