What I've been working on

Martian soil contains perchlorate salts. Perchlorate is the oxidizer in many solid rocket propellants, so recovering it on Mars could one day help make fuel there instead of launching it from Earth. Our project tests a new application of membrane filtration: using nanofiltration to pull perchlorate out of a mix of other dissolved ions.

I want to work on the problems a long-term human presence on Mars depends on, and fuel is one of the biggest. This is an approach that hasn't been studied much, so there is a real chance to find out whether it works.

• Ran a sensitivity study on a Python nanofiltration model, varying three membrane parameters: active-layer thickness, pore size, and membrane charge.
• Set a valid range for each parameter and split it into equally spaced values.
• Used parallel processing to run the model on every combination of those values.
• Analyzed the resulting ion fluxes for each combination, focusing on how well the membrane separates perchlorate (ClO₄⁻) from bicarbonate (HCO₃⁻) and chloride (Cl⁻).

I capped the number of iterations in each model run. Testing every combination means the run count multiplies with each value added to a parameter, so total runtime adds up fast. Limiting iterations kept the full sweep practical to run, at the cost of some refinement in any single run.

Lunar regolith (soil) holds small amounts of water that can be released by heating it. The lab's work models that release so that water extraction can be planned for a future Moon base.

Water is one of the first things a lasting base on the Moon needs, both to drink and to split into oxygen and propellant. I wanted to help build the technology that makes staying there possible.

• Owned model validation for the project.
• Found experimental temperature-programmed desorption (TPD) data in published studies to test the model against.
• Calculated desorption enthalpy values from that data.
• Plotted enthalpy of desorption against temperature to compare with the model.

I accounted for hydrated minerals in the calculations. Some of the water in regolith is chemically bound inside minerals rather than sitting on grain surfaces, and it behaves differently when heated. Leaving it out would have made the math simpler, but the comparison would have missed part of the water the experiments actually measured.

A MATLAB simulation of a rocket's flight from liftoff through ascent.

I wanted to take the math from my differential equations and aeromechanics classes and apply it to a real flight problem, start to finish.

• Wrote the equations of motion for a variable-mass rocket whose mass drops as it burns propellant.
• Modeled atmospheric drag along the trajectory.
• Solved the system with MATLAB's ODE45 solver.

I used ODE45, an adaptive-step solver, instead of writing a fixed-step integrator. Mass, thrust, and drag all change quickly early in the flight, and an adaptive solver shrinks its step size there on its own instead of me having to tune it by hand.

A research paper comparing three cleaner propulsion approaches: green monopropellants, electric propulsion, and solar sails.

I wanted to know which lower-impact propulsion methods actually perform well enough to use, and what "sustainable" means once you look past the fuel itself.

• Reviewed the literature on each method.
• Compared them on efficiency and sustainability.
• Wrote the paper presenting the findings.

I judged each method on how it's manufactured, not only on how it performs in flight. That made the comparison harder to pin down, but it surfaced the most important finding.

• Electric propulsion led on specific impulse and fuel efficiency.
• Solar sails came out best for in-space propulsion.
• Manufacturing is where many propulsion systems fall short on sustainability.