Stanford Energy Student Lectures (SESL) 2026
We invite Stanford affiliates to join Stanford students, postdocs, faculty, and staff on Tuesdays from 3:00 - 4:00 pm in Y2E2 299 for a seminar series featuring 17 graduate students and postdocs as they deliver short, accessible presentations on their innovative clean energy research. Discover cutting-edge science and recent breakthroughs in areas such as renewables, energy conversion materials and devices, catalysis, and decarbonization directly from the researchers themselves!
Refreshments will be served starting at 2:45 PM, and you can share your feedback on the speakers for a chance to win a Coupa gift card!
A primary goal of this seminar series is to equip graduate students and postdoctoral researchers with the skills to effectively communicate the significance of their energy research and key findings to a diverse audience interested in energy, but outside their immediate research group. Please note that talks will be in-person only, and details about the speakers are listed below.
The presentations will be judged by a panel of senior staff from the Precourt Institute for Energy and SLAC. Additionally, speakers will receive feedback from a tutor from the School of Engineering's Technical Communication Program. At the end of the summer series, the judges will select three speakers to be honored as Stanford Energy Distinguished Student Lecturers, who will then have the opportunity to present at the Stanford Energy Seminar during the upcoming academic year.
June 23
Amy McKeown-Green
Under Pressure: Watching Photocatalysts Restructure During Ammonia Synthesis
Abstract: Nearly half of the world’s 8 billion people are fed by food grown using artificially synthesized ammonia fertilizer. Currently, artificial ammonia is produced via the Haber-Bosch process, which requires extreme pressures and temperatures and is energy-intensive, contributing 1-2% of global CO2 emissions. Bimetallic photocatalysts can enable green ammonia synthesis at room temperature and pressure by combining a light-active material with a catalytic material. When alloyed together, these two materials synergistically harness light energy to produce ammonia. However, alloy catalysts have a strained internal structure, which makes them susceptible to dynamic restructuring and loss of catalytic activity.
To help translate these catalysts from the laboratory to industrial reactors, we use in situ electron microscopy to watch bimetallic gold-ruthenium ammonia catalysts evolve at the atomic scale during catalysis. By recreating reactor-like gas conditions inside the microscope, we find that gold and ruthenium separate into distinct regions during ammonia synthesis. In some particles, the catalysts even form internal voids and become partially hollow. These atomic-scale transformations can drastically change how the catalysts absorb light and drive chemical reactions. By identifying how these materials restructure and degrade under working conditions, we propose new design principles for more robust and dependable green ammonia photocatalysts.
Jaekwon Lee
Free-space Reconfigurable Optical Interconnects for the Next Generation Datacenters
Abstract: Despite the amazing capabilities the AI revolution has achieved, the hardware, especially the communication between computational nodes, remains a limiting factor in determining the AI performance. As the AI becomes smarter and smarter, the amount of communication between GPUs increases exponentially while moving the data around on-chip or off-chip consumes ~five and ~two-hundred times more energy than the computing itself. As a result, AI performance is now directly correlated with energy security, where datacenters are expected to consume 6.7-12 % of the United States electricity by 2028.
A more energy-efficient, high-speed, and cost-effective optical interconnect is urgently needed to resolve this problem and enable a more “interconnected” system with larger model sizes while suppressing the energy consumption scaling. This talk will review the current trend in optical interconnect research and provide insights into the hardware bottlenecks and future perspectives. Free-space communications with micro-LEDs are suggested to enable energy-efficient and cost-effective optical interconnects while meeting the bandwidth requirements for the short-to-long-distance interconnects.
June 30
Chi Cao
In Situ X-ray Imaging of Aqueous Zinc Ion Battery Anodes
Abstract: Aqueous zinc ion batteries are a promising option for large-scale, long-duration energy storage because of their low cost, improved safety, and reduced resource constraints when compared to conventional lithium-ion batteries. However, they are limited by the lack of reversibility at the zinc metal anode from corrosion and the growth of dendrites. In situ X-ray transmission microscopy is used to visualize the plating of zinc and hydrogen gas formation during the cycling of Zn/Zn cells in 1 M ZnSO4 electrolyte. Zinc initially deposits in the form of hexagonal platelets, but pits that form during stripping lead to hot spots of current density where dendrites form. With the electrode surface becoming increasingly rough during extended cycling, dendrites become dominant. Hydrogen evolution also contributes to the growth of dendrites by causing the local pH of the electrolyte to become more alkaline and promoting the formation of corrosion byproducts.
Andreas Mühlbauer
Between Hourly Optimization and Sub-Minute Simulation: Implications for Highly Renewable Energy System Planning
Abstract: Standard macro-energy system optimizations typically rely on hourly temporal resolution to determine least-cost infrastructure deployment. However, this temporal averaging can smooth out critical, short-term meteorological volatility and grid fluctuations, potentially underestimating the capacity required to maintain reliability. Our research investigates whether standard hourly energy system optimizations underrepresent the infrastructure required for 100% renewable grids by benchmarking linear capacity expansion frameworks against high-resolution, 30-second operational dispatch simulations. By focusing on the structural effects of temporal resolution on optimal future energy system designs, we identify critical storage and capacity gaps caused by time-step averaging. This work provides a novel framework for testing the resilience of global decarbonization strategies, connecting high-level investment planning with the sub-minute simulation of real-world grid operations.
July 7
Will McNeil
Data Center Load Growth and Power System Costs in California
Abstract: Data centers and other large loads have the potential to reshape U.S. electricity demand growth, generation investment needs, and grid decarbonization pathways. In this study, we develop a techno-economic and capacity expansion modeling framework in PyPSA-USA to quantify the grid impacts of data center load growth in California. We evaluate how proposed California policies and cost allocation frameworks influence electricity prices, infrastructure investment needs, emissions outcomes, and grid reliability. We identify tradeoffs between reliability, affordability, and decarbonization and provide actionable information for policymakers regarding data center policies in California.
Yousif Alkhulaifi
A stirred-tank flow-electrode reactor for direct lithium extraction from brines
Abstract: Global lithium demand has risen rapidly with widespread lithium-ion battery deployment, and this has intensified interest in selective Li recovery from low-grade brines. These resources often contain <200 ppm Li and high concentrations of competing cations such as Na+, K+, and Ca2+. To date, flow-electrode concepts for direct lithium extraction (DLE) have largely relied on cell architectures that pump slurries through long, thin millimeter-scale channels. These geometries increase surface-to-volume ratio and reduce ionic resistance, but they are difficult to seal, prone to clogging, and difficult to scale. Here, we introduce a stirred-tank flow-electrode architecture for DLE, in which iron phosphate (FePO4) and carbon black particles are suspended in brine and vigorously mixed in a cathode tank. Stirring sustains particle motion and causes frequent collisions of particles with a cathode collector and with each other. These collisions facilitate electron transfer to the FePO₄ particle phase and drive selective Li intercalation. In lab-scale experiments, the reactor achieves Li/Na intercalation selectivity up to 280 and current densities up to 1 mA cm⁻², with low impedance and high intercalation rates despite a small cathode area. We will present the reactor design and an experimental study of this system, including measurements of key performance metrics.
July 14
Wylie Kau
A hybrid design framework for ion selective membranes in critical mineral separations
Abstract: Growth of critical mineral (CM) demand driven by rapid deployment of lithium-ion batteries, renewables, and data centers has outpaced conventional CM supply chains. One solution is the separation and generation of high-purity CM products from industrial wastewaters (i.e., industrial wastewater refining), which requires specific ion selectivity. Ion selective membrane separations are distinctly promising due to advantages in scalability and reduced energy consumption compared to alternatives like distillation. Ligand functionalized polymer membranes (LFPMs) achieve ion selectivity beyond size-sieving and electrostatic effects through coordinative interactions between ions and polymer grafted ligand molecules, but no structure-function relationships yet exist that relate ligand identity and polymer structure to selectivity. In this work, we developed a hybrid, data-driven and experimental design framework to quantitatively link LFPM structure to partitioning selectivity in a library of 5 distinct LFPM chemistries. Our framework is built upon a dual mode sorption model that describes ion partitioning as a function of ligand identity and polymer structure, utilizing machine learning to predict ion-ligand coordination affinity. Our model can play a key role in accelerated LFPM design, allowing researchers to bypass trial-and-error experimentation and identify promising chemistries at the speed and scale to match increasing CM demand.
Philip Onffroy
Micro-Architected 3D Printed Carbon Materials for Energy Storage and Inertial Fusion Energy
Abstract: This presentation explores how micro-architected 3D printed carbon materials can enable next-generation energy technologies, from electrochemical energy storage to inertial fusion energy systems. As growing electricity demands from artificial intelligence and electrification place increasing pressure on global energy infrastructure, there is a critical need for advanced materials that combine scalability, precision, and multifunctional performance.
Using light-based additive manufacturing and high-temperature pyrolysis, we develop tunable carbon lattices with microscale structural control and customizable composition. These architected materials overcome limitations of conventional bulk carbon forms by enabling deterministic porosity, improved ion transport, and tailored electrochemical behavior. I will discuss recent advances in polyacrylonitrile-derived carbon lattices, including methods to enhance surface area, conductivity, and electrochemical activity for battery and capacitor applications.
The talk will also highlight emerging work on deterministic fuel capsule architectures for inertial fusion energy, where precisely engineered 3D printed carbon structures may enable scalable, high-throughput target manufacturing with improved symmetry and reproducibility.
By bridging additive manufacturing, carbon materials science, and energy systems engineering, this work demonstrates how architected materials can address key challenges in scalable clean energy technologies and advanced energy infrastructure.
No session on July 21
July 28
Adam Potter
Breaking properties tradeoffs for the next generation of high-temperature hydrogen catalysts
Abstract: Hydrogen is a carbon-free fuel with the potential to replace fossil fuels in industries such as transportation, metallurgy, fertilizer production, and sustainable fuels. Electrolyzers, devices that produce hydrogen from water using electricity, offer a pathway to zero-emission hydrogen production anywhere in the world, but they must improve both efficiency and durability to become economically competitive. High-temperature steam electrolysis is currently the most energy-efficient hydrogen production method known, but operating at temperatures up to 800C requires advanced electrode materials that can maintain high performance while resisting long-term degradation.
Discovering improved materials has been a slow process because researchers must simultaneously optimize several tightly connected properties, where improving one often worsens another. For example, increasing ionic conductivity can improve device performance but often reduces mechanical stability, leading to cracking and failure over time. As the field has advanced, many of these tradeoffs have been linked to fundamental physical mechanisms.
My research focuses on high-entropy materials, which mix five or more elements in a single crystal site, creating unexpected physical effects that can break the expected property tradeoffs. My talk will focus on how key material properties enable affordable hydrogen production, the physics underlying their tradeoffs, and the new mechanisms we have discovered in high-entropy materials that undermine these tradeoffs.
Taeho Kim
Controlling Earthquakes and Fractures for Next-Generation Geothermal Energy
Abstract: At a depth of 6–7 km, the subsurface of the contiguous United States holds enough geothermal heat to power the entire nation. Unlike solar and wind, geothermal energy delivers consistent, near-zero-emissions baseload power that is insensitive to daily and seasonal climate variations, at a fraction of the land footprint. Whether we can safely tap this immense resource depends on our ability to reliably predict and control the growth of subsurface fractures. Conventionally, fractures and pre-existing faults have been avoided due to earthquake risk and the inability to monitor deformation underground. Recent advances in earthquake science and fiber optic seismic monitoring now enable us to seriously pursue next-generation geothermal systems. In this presentation, I argue for a paradigm shift in our perspective on pre-existing faults to unlock geothermal energy at scale: from treating them as a liability to recognizing them as a central design variable. I present the theoretical models, simulations, and field analyses that are needed, including a recent analysis of the largest enhanced geothermal system (EGS) in the world. The analysis suggests that earthquake-enhanced geothermal systems may already exist, and that we can learn to engineer them deliberately, setting up geothermal as a cornerstone of a decarbonized grid.
August 4
Xiaoyu Yang
Bridging Physics, Data, and Learning in Lithium Battery Modeling
Abstract: Battery modeling plays a critical role in the design, control, and safety of next-generation energy storage systems, yet significant challenges remain in achieving both physical fidelity and computational scalability across multiple length scales. This presentation discusses a research framework that integrates physics-based modeling, observation-driven correction, and machine learning to address these challenges. The work spans electrochemical transport within porous electrodes, state estimation under model uncertainty, and scalable surrogate modeling for battery systems operating under thermal gradients. A dual-continuum electrochemical transport formulation is introduced to capture transport-limited behavior beyond conventional electrode assumptions. The presentation also examines how indirect measurements can be combined with imperfect models to improve estimation of internal battery states, particularly core temperature. In addition, a machine-learning-based surrogate framework is presented for efficiently modeling coupled electrochemical–thermal interactions in multilevel battery systems. Together, these studies demonstrate how physics, data, and learning can be combined to develop accurate, efficient, and scalable battery models for applications including fast charging, thermal management, diagnostics, and large-scale battery system optimization.
Ireri Hernandez
Designing Fair Agreements for Solar Energy: What Communities Want from Large-Scale Solar Development
Abstract: Large-scale solar energy projects are central to the clean energy transition and to increasing electricity demands. However given their size and proximity to residents, some projects may face delays at the local siting and permitting stage. One proposed solution has been Community Benefit Agreements, which are meant to address local concerns by specifying what host communities receive in exchange for new development. Little is know about which specific benefit and procedural features are likely to increase local support. This project presents results from a survey experiment on public preferences over Community Benefit Agreements for large-scale solar projects. Respondents in areas with existing or potential for large solar projects evaluated alternative agreement designs that varied in both content and process. Specifically, individuals rated projects based on compensation, local jobs, land-use provisions, public participation, transparency, facilitation, and monitoring. The results suggest that both benefits and process matter to increase public support. On the benefit side, respondents were more supportive of agreements that included compensation, local economic benefits such as supply-chain provisions, and environmental management measures. Additionally, respondents were more supportive of agreements negotiated through open meetings, broad resident participation, and ongoing public oversight. These patterns point to a simple but important lesson for solar deployment: community agreements are not only about what is offered, but also about whether the process feels credible and fair.
Sreya Vangara
From Battery Data to Discovery: AI Agents for Energy Materials Research
Abstract: Batteries are central to a clean-energy future, but understanding how they work and why they fail increasingly requires reasoning across overwhelming experimental evidence. A single battery study can produce electrochemical cycling data, spectroscopy maps, microscopy images, materials characterization, lab notes, and relevant scientific literature. The bottleneck is no longer simply collecting data; it is turning fragmented evidence into insight quickly enough to guide the next experiment.
In this talk, I will present an AI agent tool we are developing for battery materials research and testing on real experimental datasets at SLAC. The system helps researchers ask natural-language questions across multimodal data and literature, extract quantitative trends, connect observations to physical mechanisms, and identify useful follow-up experiments. Rather than replacing scientists or automating the laboratory, the goal is to build AI systems that serve as scientific partners: tools that help researchers reason across data types, length scales, and hypotheses.
Using battery experiments as a case study, I will show how agentic AI can make complex evidence more searchable, interpretable, and actionable. The broader vision is to accelerate next-generation energy storage discovery by helping scientists ask sharper questions and make better decisions from the data they already collect.
August 11
Edem Honu
High temperature defect and microstructural evolution in single crystal Mg
Abstract: Weight reduction in structural applications is among the most direct routes to reducing fuel consumption and CO₂ emissions in transport. Magnesium (Mg), the lightest available structural metal (ρ =1.74 g/cm³), is 30% less than aluminium alloys with a high specific strength and is therefore a critical candidate of next-generation lightweight structures in automotive and aerospace applications. Realizing this potential requires well-controlled thermal processing cycles to restore workability and tailor microstructure; yet the intrinsic, thermally-driven dislocation behaviour of Mg, by grain-boundary effects or applied stress, is poorly understood.
Here, we present the first in-situ three-dimensional dark-field X-ray microscopy (DFXM) study of defect evolution in a single-crystal Mg during a high-temperature annealing cycle under no applied stress, resolving a bulk volume of 255 × 92 × 40 µm³ at sub-micrometre resolution. A pre-existing {11-22} compression twin dissolves by ~202°C, triggering thermally activated dislocation climb across multiple prismatic and pyramidal slip variants. Continued heating drives near-complete static recovery by 318°C, after which the crystal stabilizes as a hierarchical sub-grain boundary network.
Statistical analysis of centre-of-mass rocking-curve maps confirms a significant reduction in lattice orientation spread, confirming the progressive release of stored elastic energy. These critical temperature windows inform optimized annealing regimes directly applicable to the industrial processing of Mg components, with implications for reducing fuel consumption through wider deployment of lightweight Mg structures. The results simultaneously provide the experimental benchmark required to validate emerging thermal field dislocation mechanics (T-FDM) and phase-field dislocation dynamics (PFDD) models in low-symmetry crystal structures.
Joseph Lucero
Seeing Inside Batteries from the Outside: State Estimation for Reliable Energy Storage
Abstract: Lithium-ion batteries are central to electrified transportation and grid energy storage, yet many internal states needed for effective operation cannot be measured directly. Battery management systems must therefore infer quantities such as state of charge from external measurements, typically current and voltage. This talk focuses on how physics-based models can support more trustworthy state estimation from these limited signals. In particular, I will discuss why high-resolution battery models that accurately predict voltage are not necessarily the best models for estimating internal states, and how observability-aware modeling can help identify model structures better suited for inference. Improved state estimation can clarify how battery systems interpret operating conditions, quantify uncertainty, and support more informed control decisions. Ultimately, current and voltage measurements contain valuable information about battery behavior, but extracting that information requires models designed not only to simulate battery dynamics, but also to infer the hidden states that inform battery operation.
August 18
Yun Ni
A Robotic Approach for Real-Time Hybrid Simulation of Floating Offshore Wind Turbines
Abstract: Floating offshore wind turbines (FOWTs) offer a pathway to unlock deep-water renewable energy, yet full-scale controlled testing remains impractical and small-scale experiments suffer from similitude distortions. Real-time hybrid simulation (RTHS) overcomes this limitation by coupling physical hydrodynamic experiments with numerically simulated aerodynamic loads through actuators and sensors. In this study, a lightweight industrial robotic arm driven by electric servomotors was adopted in place of conventional hydraulic actuators, offering a compact footprint and improved force resolution in low-load regimes. A model-free adaptive force control strategy was implemented to emulate 6-DOF aerodynamic loading on a small-scale FOWT specimen. As physical testing is costly and the multi-dynamical system exhibits uncertainty under wave disturbances, we further developed a virtual RTHS platform that integrates the robotic arm, floating specimen, hydrodynamics, mooring dynamics, and a numerical aerodynamic model. This digital environment enables researchers to rapidly pre-tune controllers prior to physical deployment. Experimental results demonstrate that the proposed robotic actuation system achieves satisfactory performance under operational wave conditions. Both the actuation system and virtual RTHS platform advance RTHS testing capabilities for FOWT, establishing a more viable tool for FOWT design validation and reducing the cost required to bring next-generation floating wind technology to the global grid.
Zhiqiao Jiang
Two-Step Photon Absorbers for Obtaining High Solar-Cell Voltages with Low-Energy Photons
Abstract: Expected to make up 50% of global renewable energy generation by 2030, solar cells are one of the most important sustainable pathways to produce clean electricity. Over the past decades, scientists have worked on improving solar-cell efficiencies while lowering production costs. Nevertheless, most of the high-efficiency solar cells require complicated designs and elevated costs. Solar-cell light absorbers that are designed to produce a high voltage in a device by harvesting blue light, waste the low-energy light (e.g., infrared light) that comprises much of the solar spectrum. Likewise, solar-cell absorbers that can use this low-energy light, cannot produce the high voltages we need. This talk presents a type of solar-cell absorber that can simultaneously harvest both ends of the solar spectrum to deliver high voltage and high efficiency. This solar cell architecture, called an intermediate band solar cell (IBSC), is achieved by introducing an intermediate energy band in an expanded 3D halide perovskite analog. Compared to other IBSCs that rely on sophisticated fabrication processes, our solar cells can be fabricated with low-cost solution-based methods that are already implemented in industry. We thus present a potentially lower-cost and higher-efficiency new IBSC design for the solar-cell community to produce useful electricity from otherwise wasted light.