Photo by Emily Theobald
Both of your recent papers turn forces that usually ruin nanoscale devices, like adhesion and surfaces sticking together, into tools. How did you come to see those forces as something to work with rather than against?
A key insight is that when you miniaturize mechanically active devices to the nanoscale, a different regime of forces comes into play. For example: at the macroscale, gravity is important, but because gravity scales with mass, it’s almost totally irrelevant at the nanoscale. This is why insects don’t have any trouble crawling across a ceiling. The things that begin to matter more at small scales are surface forces, like the capillary force which causes fluid to be drawn into small spaces, or van der Waals forces which cause adhesion between surfaces that are very close. Permanent surface adhesion, or stiction, is the classic failure mechanism in micro- and nanomechanical devices, and this has so far really limited how much these devices can be downscaled.
Our concept was, what can we achieve by embracing surface forces instead of working against them? We found that surface forces provide a gentle transduction mechanism to put together devices out of fragile components. For an ultrathin molecular or polymeric layer, directly evaporating a metallic top contact causes metal filaments to penetrate through the layer, giving you shorting and low yield. Using self-assembly principles instead is a way to maintain the excellent conductivity and processability/scalability of metal top contacts while preserving the pristine properties of the active layer.
One device self-assembles around molecules less than a nanometer thick, and the other uses a squishy polymer to “fire” like a neuron. What was the hardest part to get working, and when did you know it would?
Developing the fabrication technique was a very interesting challenge. Because the active device layers are down to just 1 nm thick, surface roughness and planarization of the electrodes were extremely important. The key innovations were finding a way to deposit ultrasmooth gold by massively increasing the evaporation rate; planarization using a thermally cured spin-on glass; using a sacrificial layer just to keep electrode surfaces clean; and accurately timing the undercut step. Once I made the first chip with a 1 nm molecule that had a yield over 95%, I knew we had figured it out.
Photo by Emily Theobald
You co-led both papers with Peter Satterthwaite, and the work pulls in chemistry and materials science as well as electrical engineering. How has that collaboration shaped how you approach problems?
Peter was a crucial collaborator and mentor throughout my PhD, and the main things I admire about him are his breadth of knowledge, curiosity, and willingness to do more work upfront to set up systems that lead to success later down the road. The ability to see problems simultaneously through multiple lenses of chemistry, fabrication, experimental setup design, computing systems, etc. is something I really took away from working with Peter and will continue to try to replicate in my own work. In the Niroui group in general, we have students with backgrounds in multiple areas including chemistry, materials science, mechanical engineering, and electrical engineering, and that interdisciplinary collaboration has enabled our projects and driven our work forward.
The papers point to things like smart prosthetics and low-power wearable health patches. What would it take to get from a test chip to something a person could actually use?
To move towards applications in these devices will require focusing on systems design, controls, and packaging, which I am working on now to achieve a touch-sensitive mechanical neuromorphic computing system. To bring these devices out of the lab and into the real world we’ll need to study and calibrate for the effects of temperature and humidity and protect against things like electrostatic discharge. Incorporating mechanically active Angstrom-scale devices into multifunctional systems for computing, sensing, memory, and actuation will enable compact and efficient environmental and physiological monitoring and beyond; we’re excited to keep working towards these applications goals.
Photo by Emily Theobald

