Light matter interactions form the basis of numerous processes around us (photosynthesis, solar cells, LEDs/lasers). The ability to visualize these interactions at nanometer scales, well below the diffraction limit of traditional optical imaging, is crucial in understanding and engineering these interactions. We have developed a nanometer scale imaging technique called ‘PhotoAbsorption Microscopy using Electron Analysis’ or PAMELA, which uses the sub-nm resolution of secondary electron imaging in electron microscopes coupled with simultaneous laser excitation to obtain nanometer resolution photoabsorption images. We expect to further improve the resolution of PAMELA and explore its various mechanisms in the near future. The details can be found in these previous studies.

• Ze Zhang, Archith Rayabharam, Joel Martis et al. Appl. Phys. Lett. 118, 033104 (2021);

• Ze Zhang, Joel Martis, Xintong Xu et al. Nano Lett. 2021, 21, 5, 1935–1942;

• Joel Martis, Ze Zhang et al.(2021). Microscopy Today, 29(5), 40-44

Hyperspectral imaging has found applications in numerous fields, from remote environmental imaging to medical diagnostics. This project is to develop a broadband hyperspectral imaging system called HyperViper that spans visible to mid-infrared range (0.3-10 m). Most broadband instrumentation is large and expensive, so one of the goals of this project is to create a system that is low-cost and more compact. The canonical example we will use is to image refrigerant leaks which are potent greenhouse gases (~2000x more than CO₂). Using imaging as the method for onsite leak detection would improve the minimum detectable quantity of the sensor.

The distribution of electrons between atoms determines the emergent properties of matter. However, it is non-trivial to image these electrons at the Angstrom scale. Using sub-Angstrom focused electron probes, we imaged the charge density in monolayer MoS2, a 2D semiconductor with numerous applications. We showed that although the charge density image includes contributions from the valence electrons, they are spatially blurred out by the electron probe. Residual aberrations from the microscope also limit the quantitative accuracy of these measurements. We are now exploring monochromated imaging and imaging materials where the valence effects are on much larger length scales than atomic sizes (such as in 2D Moire materials).

Electric fields at the atomic scale give rise to chemical bonding and the emergent properties of matter. Currently there exist very few tools that can directly image these atomic electric field maps. We plan on using differential phase contrast transmission electron microscopy to measure and map out atomic electric fields in various materials with the aim of uncovering new physics. One of the classes of materials we are investigating are 2-dimensional materials that contain atomic defects which are of wide application in photonics and chemistry.

Electrochemical stripping (ECS) of wastewater reduces nitrogen pollution and generates valuable ammonium sulfate fertilizer, which can offset the cost of water treatment. Solar panels can provide sustainable, decentralized electricity to power this process, but face diminishing efficiency as the sun causes them to heat up. On the other hand, ECS becomes more efficient with increased temperature. By transferring heat from the solar panel to ECS, a greater quantity of fertilizer will be produced while simultaneously increasing the electricity output of the solar panel. This tandem approach serves to improve the economic viability of real-world ECS deployment.

Methane is a greenhouse gas responsible for more than 0.5°C of global warming. This project aims to develop processes to chemically decompose (oxidize) methane in order to mitigate its warming potential. These processes could one day be deployed either at methane emitters or to remove methane from atmospheric air. The work to date has been a combination of benchtop catalyst/process development and techno-economic analyses of scaled-up systems. Future work may also include multiphysics modeling of scaled-up methane-oxidizing processes.

Carbon-dioxide free hydrogen production that is economic and scalable would greatly mitigate carbon emissions. Approximately 500 billion m³ is used annually for ammonia production, petroleum refining, and chemical processes, but nearly 95% of that amount is produced from steam-methane reforming, a very carbon-dioxide intensive process (>10-tons CO₂ per 1-ton H₂). We are currently investigating hydrogen production through methane pyrolysis (), a potentially cost-competitive and scalable process. The main hinderance for pyrolysis is catalyst deactivation from the deposited carbon, which eventually stops the reaction. We are currently exploring new methods to prevent catalyst deactivation using oxide supports that act as a “non-stick” layer.

A reduction-oxidation (redox) reaction is a type of chemical reaction that involves a transfer of electrons between two species. Both science and engineering are in critical need to understand and tailor the electron transfer with/without ion transfer in the redox reactions. Photochemical, electrochemical and thermochemical redox reactions lie in the heart of key chemical transformation processes in a variety of important fields. Our approach centers on the discovery-driven investigation in understanding the thermodynamics and kinetic effects in oxide-based phase transition, identification of the active species for chemical bond activation (CH₄, O₂, CO₂), as well as mechanism-informed materials design and process design to bridge the existing gap between the fundamental science and the research frontier in engineering. Our current research interests range from photocatalytic methane (CH₄)-to-methanol (CH₃OH) conversion to the photothermal methane removal, and the direct iron reduction.

There is increasingly high penetration of distributed and decentralized energy resources (PV, wind, storage, etc) around the world, but no detailed information or granular data is available for these systems, which poses challenges for the monitoring, coordination, and control of these decentralized systems. To ultimately revolutionize the global energy system towards a more sustainable, efficient, and intelligent one, we aim to develop automatic tools to map, analyze, and maintain a comprehensive and highly granular energy atlas, which will offer the next-generation platforms for managing heterogeneous and decentralized energy suppliers and integrating them with “active” energy consumption activities on the demand side. The first stage of EnergyAtlas, DeepSolar, constructed a high-fidelity database of solar installation in the contiguous US combining deep learning and satellite imagery.

Over the past decade, the growing impacts of climate change have disproportionately affected certain areas and populations, making environmental justice (EJ) a central focus of our energy transition. It is important to identify the most vulnerable communities where there is a dire need to improve insufficient grid infrastructure and bolster with distributed resources. Existing vulnerability screening tools present data on exposure to environmental hazards, but these tools lack spatial and temporal granularity and direct relevance to energy infrastructure. We introduce DeepEJ, an AI-driven tool that leverages grid maps, Census demographic data, and historical climatic event datasets to develop a dynamic map of climate grid vulnerability, in relation to energy security and resilience. It will serve as an early warning system for the effects of devastating natural events on the grid, and aid in environmental justice efforts to promote equity in our energy system.