ENG-SEMICON: EPMD: Engineering Perovskite LEDs

The prevalence of artificial intelligence (AI) and of internet of things (IoT) in our life has resulted in an exponential growth in the amount of data being transmitted at any second in a day at an unprecedented rate. The total bit rate for internet traffic already surpassed that of telephone traffic at the beginning of the 21st century and continued to grow at a rate of ~100 times per decade.

While the total internet traffic is high, a decade ago it was already small compared to the data sent over shorter links in data centers which has been growing at an even higher rate. More recently, massively parallel computations that are critical for machine learning (ML) and data analysis have driven bandwidth requirement for ML supercomputers, which have become a major contributor for information data.

The growth rate of global electricity usage by information processing and computing has surpassed that of electricity generation capacity for over a decade, with analysis showing it consumes the majority of global electricity on the horizon. This energy consumption has caused serious alarm to our environment. Without major reduction in energy per bit of information data, the bandwidth of information processing and computing cannot continue to grow. This demands innovations in technology hardware and infrastructure for computing and communication.

Using light to carry information data has not only enabled long-haul high-capacity communication networks and revolutionized our communication technologies, but also offers a route to higher interconnect densities and reduced energy consumption in data centers and ML supercomputers. A key component in optical interconnects is the transmitter, which traditionally utilizes semiconductor lasers.

They have high optical performance but face the challenge of high energy consumption and cost, in addition to difficult integration with their Si CMOS drive circuits which further leads to high system-level energy consumption. LEDs offer a promising alternative as low-cost photonic sources in optical interconnects. However, conventional LEDs cannot be easily integrated with Si CMOS either.

Metal-halide perovskite semiconductors are solution-processable materials with high optoelectronic properties and facile integration compatibility with many platforms including Si CMOS. The objective of the proposed research is to achieve perovskite LEDs with high modulation bandwidth. Three research approaches, each carrying its own intellectual merit, will be conducted in synergy to achieve the goal.

They are: (1) Engineering perovskite semiconductor materials and device design to minimize resistance and increase LED speed. (2) Optimizing modulation format for high-bandwidth operation using data-driven learning. (3) Developing a process for monolithic integration of perovskite micro-LEDs on Si to minimize overall parasitics, improve system-level modulation bandwidth and reduce energy consumption.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.