FFUSars Projects

In a fusion power plant, energetic neutrons from the fusion reaction hit the walls around the core and heat them up. That heat is then removed from the vessel by circulating coolants to generate electricity. To keep this process as efficient as possible, we would like to use materials that can retain high thermal diffusivity under the bombardment of fusion neutrons.

Most critically, the thermal resistance at interfaces—for example, between the silicon-carbide (SiC) vessel and the tungsten (W) coating applied to prevent corrosion from the coolant or erosion from the plasma—is notoriously hard to measure and rarely studied under irradiation. The reason is simple: instruments capable of measuring this property are typically incompatible with irradiation facilities.

Transient grating spectroscopy (TGS) is a laser based technique, routinely used at the Plasma Science and Fusion Center to measure thermal and mechanical properties during irradiation. In principle, TGS can measure the interface thermal resistance. Your role will be to prove (or disprove) that this is possible. If successful this project will enable fusion engineers to correctly quantify the efficiency of heat transfer between vacuum vessel and coolant.

In this project you will…

Prepare W-coated SiC samples with W-thickness ranging from few microns to hundreds of microns .
Develop methodology to measure measure interfacial thermal resistance using TGS (same sample, multiple gratings or different samples, same grating)
Find and work with external (or internal) collaborators to compare the interfacial thermal resistance values you measure with TGS to those measured by Time Domain Thermal Reflectance (TDTR): a common and well-established method.
Optional: Capture thermal diffusivity of W-coated SiC at various grating periods at room temperature and at various temperatures up to 800 C.
Optional: Irradiate a W-coated SiC sample to study changes in the interface thermal resistance as a function of dose.

You will develop transferable skills…

Sample preparation, including polishing and physical vapor deposition (PVD).
Safe handling of lasers, optics, vacuum and heating systems.
Hardware design and commissioning.
Establishing connections with researchers both in and outside MIT
Data acquisition, curation, and analysis with aims to write a first author publication and present your results at the ANS, TMS, or MRS conference

Materials in the reactor core face intense neutron irradiation, which leads to complex and poorly understood damage processes. To address this challenge, scientists are developing numerical tools that can efficiently and inexpensively capture radiation damage mechanisms and guide the design of more resilient materials. [1].

The goal of the project is to help the student develop a solid understanding of the science underlying radiation damage and its relevance to scientists and engineers. The student will gain familiarity with key concepts in solid-state physics, statistical mechanics, material properties, and fundamental radiation damage mechanisms. In addition, the project will introduce the Molecular Dy- namics (MD) simulation technique through the use of the open-source package LAMMPS and the Python programming language. The student will charac- terize the main thermodynamic properties of materials at the nanoscale, model ballistic radiation damage, and potentially develop theoretical insights and cor- relations to generalize radiation damage behavior. The project also encourages the development of Python tools and wrappers to automate simulations, data analysis, and visualization. It is worth noting that the outcomes of this research are relevant not only for fusion reactors but also for fission reactor applications.
References

[1] Miaomiao Jin, Penghui Cao, and Michael P Short. Thermodynamic mixing energy and heterogeneous diffusion uncover the mechanisms of radiation damage reduction in single-phase ni-fe alloys. Acta Materialia, 147:16–23, 2018.

In fission and fusion reactors, the structural materials that make up the reactor vessel are exposed to some of the most extreme conditions — high temperatures, pressures, and most importantly, a consistent exposure to high-energy neutrons over extended periods of time. Neutron damage leads to changes in the material properties, which can cause materials to fail catastrophically or unexpectedly. Accordingly, we can’t get regulators to approve new materials without first demonstrating that they can withstand such harsh conditions and remain structurally sound. The problem? For fusion and Gen IV+ fission reactor designs, we do not yet have a reactor on hand to put materials into; existing neutron sources are few and difficult to access. Instead, we use ion accelerators to irradiate our materials; these machines are abundant (we have our own ion accelerator at MIT!) and can damage materials much faster than neutron sources.

However, even with the speedup from using ions, historically all techniques used to evaluate the resulting damage are destructive and time-consuming, and can only get information about the microstructure at a single dose (since the analysis is destructive, we can only look at the final state of the material at whatever dose we irradiate to; we can’t know what happened before that dose unless we do more irradiations to lower doses, then analyze those results to get more data points). Because of this, it takes many irradiations to “see” what happens in our materials as it undergoes irradiation, and this limits how much we know about the underlying physical mechanisms that create problematic microstructural features (bubbles and voids) during irradiation. If we can create techniques to get continuous data on what’s going on inside the material while it’s being irradiated, we could better identify and understand what physics and factors actually control radiation-induced microstructural changes. With such an understanding, we could potentially predict the dose where a material will fail – without needing to irradiate it fully to that dose – by looking at how the material properties are changing during the early stages of irradiation.

You might be wondering, “what kind of technique could we possibly use to ‘see’ what is happening in our samples during irradiation without damaging it?!” Well, our group has an exciting solution! We use transient grating spectroscopy (TGS), a laser technique to nondestructively measure elastic and thermal material properties throughout an ion irradiation experiment. This project will involve exploring possible connections between how thermal and elastic properties are changing during irradiation, what sorts of microstructural changes are happening, and seeing if we could try to identify the physics and mechanisms that actually cause these changes. If this work is successful, we could potentially predict the way a material will change at different doses while we are irradiating, just by looking at trends in the TGS data at the start of an irradiation.

TLDR; you could predict your material’s future fate by using lasers to “see” the physics going on inside during irradiation!

You’ll develop a strong set of highly transferable, marketable skills from this project, including:
- Fabricating your very own customized metal alloys using an arc-melter
- Operating the linear particle accelerator at MIT to run your irradiation experiments
- Training with and operation of laser systems (TGS)
- Independently operating scanning and transmission electron microscopes (SEM, TEM)
o You’ll become a certified tool user at MIT.nano and/or Harvard CNS
- Focused ion beam (FIB) milling to make electron-transparent TEM samples
- Training and experience building custom data analysis tools with MATLAB or Python
- A multitude of sample preparation techniques

This project offers a ton of flexibility to customize the scope, to focus more on experimental or computational skills, and dig deeper into areas you are particularly interested in.

The impact of this work on the field will contribute toward a deeper, more comprehensive understanding of fundamental radiation effects in materials, which remains a critical knowledge gap that has the potential to rapidly accelerate development and deployment of new nuclear energy solutions (both fission and fusion systems) to the grid!

There are at least two slots open for this project - expect to do this research both with Alicia, your mentor, and another FFUSar.

A major hurdle for fusion power is building walls that can survive years of contact with very hot, fast-moving plasma. These walls, called plasma-facing components (PFCs), have to resist heat, wear, and fuel penetration. Silicon carbide (SiC) is a promising candidate because it has good performance at very high temperatures, conducts heat well, and doesn’t let hydrogen move through it easily.

We still don’t know enough about how SiC behaves when it’s directly hit by plasma. To make good decisions about future reactors, we need focused experiments that test SiC under the kinds of conditions a fusion reactor wall would see.

Testing materials inside an actual fusion reactor sounds ideal, but it’s slow, costly, and makes it hard to separate cause and effect. A better first step is to use “linear plasma devices.” These machines create controlled plasma conditions, similar temperatures, densities, and particle bombardment to a reactor’s first wall or divertor, while letting us change one variable at a time and learn faster.

At MIT-PSFC, our linear plasma device is called DIONISOS. If you join our team, you’ll help design and run experiments on DIONISOS to evaluate how SiC stands up to realistic nuclear fusion-like conditions.

*During your time with us, you will be tasked with:
-Sample preparation: polish SiC samples and clean them.
-Design experiments: literature review + choosing the right experimental parameters for your experiments.
-Perform experiments: learn how to use DIONISOS + get your data (assisted).
-Analyze your data: learn how to analyze nuclear reaction analysis (NRA) data (assisted).
-Make sense of your data 🙂

*You will acquire the following transferable skills:
-Manipulate vacuum systems and work in a lab.
-Troubleshoot experiments and problem-solving techniques.
-Communicate your work with colleagues and external researchers.
-Prep samples for any type of plasma experiments or extreme environments.
-Learn how to have fun and enjoy your work 🙂

**You will be directly supervised by Joey Demiane & Keshav Vasudeva. The principal investigator (PI) of the project and lab is Kevin Woller.

The Short Lab here in nuclear science and engineering houses a world-unique facility to evaluate how thermal properties change in conditions simulating a fusion reactor environment. We use transient grating spectroscopy to shock a material with lasers during ion irradiation, plasma exposure, cryogenic conditions and more!

Some of the lab's work has covered several portions of common reactor designs. From structural materials to superconductors, all materials will need to be assessed and qualified under fusion-like conditions to get to a power plant device.

You can learn more about some of our recent work on the poster linked below.

If you have interests and in radiation materials science that line up with our lab, come talk to us! Together, we will come up with a project proposal.

Experience is not required. If the research question is important and well thought out, you will be trained in a variety of engineering techniques, potentially including optical, vacuum, cryogenic, accelerator and plasma systems, and dependent on your line of inquiry.

For questions, or to come and see if this lab is a fit for you, please contact Angus Wylie at awylie@mit.edu.

We will design and build a thermal-insulating casing for a FibHeat-200 molybdenum micro-furnace (to ~1900 °C) to improve temperature stability and protect nearby hardware. The student will prototype a layered casing (ceramic insulation + reflective shield) with small ports for X-rays/visual access and a capillary-fed airline to control the immediate sample atmosphere. Tasks include CAD design, simple thermal analysis using Finite Element Analysis, machining/assembly, PID tuning, and thermal calibration with a thermocouple at the sample position. The student will benchmark the thermal gradients and document safe operating procedures. Designs are inspired by compact, jacket-cooled heaters and capillary-delivered gas environments used at synchrotron beamlines.

(1) Objectives & success criteria
• Achieve stable sample-zone control (±2 °C @ 800–1200 °C; ±5 °C @ ≥1600 °C) with <1 % overshoot after setpoint changes.
• Keep exterior jacket ≤100 °C during 1100 °C operation; ≤150 °C during ≥1600 °C operation.
• Maintain clear line-of-sight ports for imaging/beam as needed.
(2) Design concept (inspired by NSLS-II/FXI compact heater work)
• Layer stack:
1. Inner radiant shield (thin molybdenum sleeve or graphite foil ring pack) around the furnace hot zone to reduce radiative loss.
2. High-temp ceramic insulation (castable alumina or porous alumina fiber board tiles, removable segments).
• Thermometry & control:
o Closed-loop furnace sensor (Type B/R) for power control; secondary K-type at the sample (inside the capillary or touching a dummy pin) to build an offset calibration and hold spec at the sample, following established calibration practice.
o PID retuning with ramp/soak profiles; implement soft-start and dwell steps to minimize drift.
• Maintainability: split jacket; quick-disconnect; modular insulation tiles; removable.

(3) Materials
• Cu jacket (C110), 2-pc clamshell with brazed; Silicone O-rings outside hot face.
• Castable ceramic (e.g., alumina, 1400–1700 °C rated) or alumina fiber boards for the mid-layer.
• Radiant shields: Mo sheet or graphite foils (consider oxidation—use only where not exposed to air jets).
• 2× thermocouples (Type B/R for furnace; K-type for sample).
• Small peristaltic or chiller loop for water (≥1 L/min recommended at ~20 °C).
(4) Analysis
• Heat load to jacket: size channels so jacket copper stays <100 °C at 1100 °C setpoint; finite-element studies for similar designs show this is feasible with close-coupled channels and modest flow.
• Hot-zone uniformity: aim for <10 °C mm⁻¹ over a ~2 mm³ region around the sample; compact designs have demonstrated ~5–14 °C mm⁻¹ gradients—acceptable for micro/nano imaging and most materials tests.
(5) Build & test protocol
1. CAD & FEA (week 1–2): model the clamshell jacket, channel layout, insulation thickness, and port map; steady-state thermal FEA with a radiative inner boundary.
2. Fabrication (week 3–4): machine Cu halves, braze/solder channel tubes or mill internal channels + cap; cast or fit insulation; drill/ream ports; add quick disconnects.
3. Bench tests (week 5): hydro test jacket (2–3 bar), leak check; dry run flow/pressure.
4. Thermal calibration (week 6): insert K-type at sample location; ramp 200→1100 °C in 100 °C steps; record furnace vs. sample offsets; log stability (σT ≤2 °C).
5. Stability/gradient mapping (week 8): small 3D map (±0.5 mm) via stepped thermocouple or IR microprobe, target gradients as above.
6. Application shakedown (week 9): Run a short “interrupted in-situ” trial—raise/lower furnace to image between steps if needed; this approach has precedence when continuous in-situ is impractical.
(6) Risks & mitigations
• Mo oxidation with air exposure: confine air to a small, directed stream at the sample; keep furnace internals shielded; prefer N₂ during warm-up/cool-down.
• Optics/nearby hardware heating: jacket sizing + channel proximity; verify with thermocouples on the jacket skin during 1100 °C holds.
• Sample holder expansion: use low-expansion pins (e.g., Invar) or calibrate for motion; this has been used successfully in Transmission X-ray Microscopy beamline at National Synchrotron Light Source II.
(7) Documentation & deliverables
• CAD pack (STEP + drawings), FEA report, wiring & plumbing diagram, thermal calibration curves, SOP (startup/ramp/atmosphere switch/shutdown), and a short validation note with images/data from a test specimen.
• Optional: a brief note comparing continuous in-situ vs interrupted tomographic approaches and when to use each for stability.

At the ASDEX Upgrade tokamak, a major fusion experiment, microwave reflectometers are used to probe the plasma and measure density profiles. These measurements are essential for understanding plasma behavior, but the data are notoriously noisy and complex. Traditional heuristic-based signal processing methods, such as frequency transforms, filtering, and background subtraction, can only go so far. In many cases, human experts still need to manually distinguish meaningful plasma signals from background noise based on visual pattern recognition.

This is where neural networks come in. The goal of this project is to train a Domain Adversarial Neural Network (DANN) that can automatically learn to identify and extract the true plasma signal from real-world reflectometry data. The project leverages a large dataset of both synthetic “clean” signals and "noisy" experimental data collected directly from the tokamak. By learning to generalize across these two domains, the model will hopefully gain the ability to separate physical signals from artifacts that arise during experiments.

Students joining this project will gain hands-on experience in machine learning, signal processing, and plasma diagnostics, contributing to an effort that directly supports progress toward clean fusion energy.

Participants in this project will:

1. Understand the Problem Domain: Learn the basics of plasma reflectometry and how microwave signals are used to measure plasma density. Explore example datasets of real and simulated signals to understand the structure of the noise and the features of interest.

2. Prepare and Visualize Data: Process and visualize time-series and spectrogram data using Python tools such as NumPy, SciPy, and Matplotlib. Implement preprocessing steps (e.g., normalization, filtering, data augmentation).

3. Design and Train Machine Learning Models: Implement and train a Domain Adversarial Neural Network (DANN) in PyTorch or TensorFlow. Experiment with network architectures and loss functions for optimal noise suppression and domain adaptation. Compare the DANN’s output against existing heuristic-based methods and manual expert analysis.

4. Document and Present Results: Prepare a concise technical report and presentation summarizing the findings.

Nuclear energy is gaining a renewed global momentum as a high-energy-density, reliable, and clean baseload power source essential to decarbonization. However, its continued expansion depends on resolving challenges surrounding spent nuclear fuel (SNF) management. For example, six U.S. states currently restrict the construction of new nuclear reactors until a permanent disposal pathway for SNF is established. This project takes a different perspective: instead of treating SNF only as waste, can we also treat it as an energy resource?

We will investigate the feasibility of recovering energy from SNF through two complementary pathways: (1) using the thermal energy from radioactive decay and (2) converting gamma radiation directly into electricity. We will use non-isothermal subsurface simulators (e.g., TOUGHREACT) to model heat transfer around SNF canisters and quantify how much decay heat can be recovered using low-temperature geothermal technologies such as ground-source heat pumps and organic Rankine cycle systems. These simulations will span a range of realistic conditions—different backfill materials and storage environments—to quantify electricity and heat outputs and compare them with conventional renewable options. In parallel, we will evaluate “radiation-to-energy” concepts that directly harness gamma rays emitted from SNF. Using the recent demonstrations of scintillator-coupled nuclear photovoltaic (“gammavoltaic”) batteries that deliver microwatt-level power, we will estimate the recoverable energy potential as a function of waste mass and facility type.
Undergraduate fellows will contribute to both modeling and conceptual design, gain experience with energy systems, radiation physics, and numerical simulation tools to help quantify how much clean energy could, in principle, be recovered from SNF.

The US National Nuclear Security Administration (NNSA) is funding a large consortium of research designed to root out and reveal the presence of nuclear weapons, for the purposes of non-proliferation treaty (NPT) verification and ultimately seek peace through ensuring no unaccounted-for nuclear weapons exist.

We have developed a highly sensitive environmental sampling using high temperature nanocalorimetry, which we seek to use to detect nuclear weapon reprocessing activities. During Pu-based weapons reprocessing, which is done partially to remove the Americium (Am) from the weapons to restore their effectiveness, microgram-sized crystals of AmO2 will be formed in plentiful numbers. These particles cannot simply be trapped using filtration, as the momentum from a single alpha particle decay is enough to dislodge them from surfaces via counteracting the van der Waals forces keeping them adhered to filters. They therefore transport everywhere throughout the environment – through the air, in wastewater, and via other routes detectable downstream via any environmental path from an enrichment facility. Subsequently, by taking either on-site swipe sampling or monitoring stations positioned at a distance from these facilities to detect stray AmO2 particles, our nanocalorimeter will be utilized to ascertain their age, specifically identifying when the Am was separated from the plutonium during the reprocessing phase.

By developing this new technique, we hope that we can contribute to stopping the spread of nuclear weapons. If you like experimental and/or computational science, have an interest in halting the spread of nuclear weapons, and want to work in a team on a project of national importance, then this is for you!

We can accommodate up to four FFUSars students on this project. We highly hope that at least two will join to work together as a team, so please tell your friends!

We want to develop a generalized self-driven laboratory system to rapidly discover new materials for, and optimize material composition and processing for, a customer-specified fitness function of performance metrics. This process, which we call RAM-3D, is predicated on an iterative feedback loop of ultra-rapid co-fabrication of hundreds to thousands of adjacent material compositions, minimum-information rapid measurements tied to inference models between measurables and the fitness function, and repetition with successively more selective and difficult measurements. Inference models are the key to the success of RAM-3D, as most properties of interest are coupled to those more easily measured, giving us a technological edge to break the speed vs. precision tradeoff.

We will physically embody this general idea via three targeted materials discovery and optimization projects: (1) Improving the Ti-6Al-4V alloy system to optimize for drone usage, (2) Discovering the best magnetic material in the Nd-Fe-B alloy system for drone engine rotors, and (3) Enhancing thermal barrier coating (TBC) performance by optimizing oxygen stoichiometry and cation chemistry in the rare-earth zirconate system to boost both thermal resistance and thermal shock. In each case, we will use combinatorial synthesis via techniques such as off-axis physical vapor deposition (PVD) and liquid-phase multi-particle additive manufacturing, followed by development of inference models between what we want to measure and what we can easily measure. Transient grating spectroscopy (TGS) will be a cornerstone of this effort, as it measures elastic and thermal properties which are sometimes of direct interest, and sometimes related to those of interest. Rapid testing methods, such as thermal shock of a partially diced wafer containing thousands of TBCs, or magnetic force atomic force microscopy (MF-AFM) of thousands of magnetic thick films, is the other key to success of RAM-3D.

If you want to work on ultra-rapid materials science, and understand the expression good enough is good enough when it comes to making key, early decisions about which materials to study in depth, then this project is for you! There will also be lasers, plasmas, high temperatures, atomistic measurements, and opportunities for simulation as well.

This project can accommodate up to four FFUSars, ideally at least two to work in a team together.

Stellarators are a type of magnetic confinement device used in fusion energy research. Their purpose is to confine hot plasma long enough for nuclear fusion reactions to occur. Unlike tokamaks, which use strong plasma currents for confinement, stellarators rely entirely on external magnetic coils arranged in a complex, twisted geometry to create a three-dimensional magnetic field. This design allows for steady-state operation typically without large current-driven disruptions, making stellarators promising for continuous fusion power generation.

Flux surfaces are nested surfaces in a plasma where the magnetic field lines lie. They are desirable because, ideally, they help keep the plasma well-confined. However, in three-dimensional stellarators, flux surfaces are not guaranteed: field lines can break, forming magnetic islands or stochastic regions that can degrade confinement.

In this project you will:
• Learn how to assess whether flux surfaces exist in a given stellarator equilibrium.
• Explore optimization techniques to restore flux surfaces by adjusting coil currents or making minimal changes to the equilibrium.
• Apply these methods to existing designs, creating improved and more desirable stellarator configurations.

You will acquire the following transferable skills:
• Reading, understanding, and working with Python code.
• Applying multi-objective optimization methods.
• Communicating research findings effectively with other researchers.
• Deepening your knowledge of the fascinating world of stellarators.

No existing structural materials can withstand the extreme environments in fusion reactors. To enable fusion energy, new structural materials must be discovered and rigorously tested. Toward this end, this project proposes a new method to screen candidate structural materials for resistance to neutron-like damage.
The proposed method includes (1) triple-ion beam irradiation to find the ion irradiation conditions which match neutron damage, (2) in-situ spectroscopy during ion irradiation to track radiation damage while it’s occurring, and (3) post-irradiation examination of materials to assess radiation damage directly.

This specific project will focus on part (3), the assessment of radiation damage. Specifically, this project will focus on automated methods to determine void swelling from transmission electron microscopy (TEM) and scanning electron microscopy (SEM) images. Void swelling happens when point defects (specifically vacancies) agglomerate into extended 3D defects, and turn an initially fully dense metallic part into porous Swiss cheese. Void swelling can drastically change part geometry and make materials brittle. Automatic characterization of void swelling is a long-standing problem in the field of nuclear materials, and progress toward this end would allow a student to put their mark on this important field of study.
In this project you can:

• Use/build automated image processing techniques to measure void swelling in irradiated materials
• Measure radiation damage (void swelling) using TEM and SEM
• Drive a particle accelerator and use in-situ spectroscopy during irradiation to track radiation damage (optional)

You will develop highly transferable skills including:

• Automated image analysis of radiation damage
• The basics of scanning and transmission electron microscopy
• How to operate a particle accelerator to make radiation damage (optional)