Research

Silicon transistors are approaching their physical scaling limits, and new channel materials are needed for future electronics. We study atomically precise graphene nanoribbons (GNRs) as one of these materials. GNRs are grown on metal surfaces from molecular precursors, which link into polymer chains and then fuse into flat carbon ribbons. The precursor sets the width and edge structure of the ribbon with atomic precision. Since electrons are confined across a width of about one nanometer, quantum effects open a band gap that depends on the exact number of carbon atoms across the ribbon. Most of our work so far has used nine-atom-wide armchair ribbons (9-AGNRs), and we are now also studying other types of GNRs whose different widths and edge structures give different electronic properties. GNR structure can also produce topological electronic states, which we have explored with collaborators in nitrogen core-doped GNRs.

Our device work covers scaling, gating, contacts, stability, and integration. We have fabricated GNR transistors with channel lengths down to 7 nm and shown that double-gate devices with ultrathin high-k dielectrics improve gate control. Charge must pass from metal contacts into a ribbon less than a nanometer wide, so the contact interface often limits performance, and we have studied how contact engineering affects charge injection. We have also shown that atomic-layer-deposited encapsulation improves long-term stability and reliability, and developed an etch-free method for transferring GNRs and other low-dimensional materials from metal films onto 100-mm wafers.

The performance of GNR transistors is still below theoretical predictions. Our current work focuses on improving the interfaces between the ribbons, dielectrics, and contacts, and on moving from single devices toward simple circuits.

The same nanoribbon devices can also be used as sensors. A GNR is a single layer of carbon about one nanometer wide, so nearly every atom is at the surface. Changes in its surroundings, such as trapped charge in the dielectric, defects in the ribbon, or incoming radiation, can change the current through the device in a measurable way.

In a recent study, we exposed 9-AGNR transistors to gamma radiation and tracked both their electrical behavior and the structure of the ribbons. The devices showed a strong electrical response, which we relate to the behavior of charge carriers in such narrow channels. These sensors are small and require little power, which makes them promising for radiation detection in settings such as space systems and other harsh environments. We are extending this work to other types of radiation. In a different direction, we are exploring GNR devices for detecting the very weak light emitted by biological samples, which calls for a different kind of sensor than those used for conventional biological sensing.

Surface acoustic wave (SAW) devices filter radio signals in phones and other wireless systems. Interdigital transducers, metal electrodes patterned on a piezoelectric crystal, convert an electrical signal into a sound wave that travels along the chip surface and then back into an electrical signal. Sound travels much more slowly than electromagnetic waves, which keeps these devices compact. At higher frequencies, the acoustic wavelength becomes shorter, and small fabrication imperfections scatter the waves and increase losses.

We fabricate SAW devices on 128° Y-cut lithium niobate and add periodic structures along the wave path. These phononic structures create stop bands, frequency ranges in which sound cannot propagate, and can be designed to confine and guide acoustic waves. In collaboration with theory and simulation groups, we have worked on composite superlattice RF SAW devices based on this approach. We also use laser direct writing to pattern these structures without masks, which allows us to test new designs quickly and compare them with simulations.

Our longer-term goal is to apply concepts from topological physics to these devices. In a topological design, the lattice structure guides sound along paths that are less affected by defects, which could make high-frequency components more robust. We are also investigating phase-change materials for changing the acoustic properties of a device after fabrication, with the aim of reconfigurable RF components.