Radioisotope Thermoradiative Cell Power Generator
A deep-space power source that generates electricity by radiating heat to the cold of space, the thermodynamic inverse of a solar cell. My primary contribution is the design, construction, and commissioning of the cryogenic measurement system used to characterize prototype thermoradiative cells as well as helping with testing and characterizing device contacts.
This work is funded by NASA's Innovative Advanced Concepts (NIAC) program under a Phase II award, in collaboration with the NASA John Glenn Research Center COMPASS team, which is studying mission applications including a CubeSat in Uranus orbit.
Uranus Cubesat: TITUS
(Technology Innovation Thermoradiative Uranus Smallsat)
A conventional solar cell generates electricity by absorbing high-energy photons from the hot sun (~5800 K) and re-emitting lower-energy photons to the environment. A thermoradiative cell (TRC) reverses this principle, the device sits at an elevated temperature and emits infrared photons toward the cold environment of deep space (~3 K), extracting electrical work from that temperature differential.
Paired with a radioisotope heat source, a General Purpose Heat Source (GPHS) module containing a single 238Pu pellet producing 62.5 W of thermal power, a TRC can generate electricity with no moving parts, no solar panels, and no sunlight. This makes it ideal for missions to the outer solar system, permanently shadowed polar lunar craters, or any environment where solar power is impractical.
The target material, InAs0.91Sb0.09 (bandgap 0.28 eV), can theoretically yield 8 W of electrical power from a single GPHS pellet with a mass-specific power of 12.7 W/kg, over a 4.5× improvement as well as providing 10× mass-specific power and 1000× volume reduction over the heritage Multi-Mission RTG.
Characterizing a thermoradiative cell requires replicating the device's actual operating environment, a hot sample (~575 K) emitting infrared radiation toward an extremely cold "sky." The measurement system must simultaneously hold the device at high temperature while maintaining a cold detector stage at cryogenic temperatures, all under high vacuum. I led the design, construction, thermal analysis, and commissioning of this system.
The measurement enclosure and sample stage were designed in SolidWorks to mount inside a Cryosystems, inc number 22.
The original design was to use a Janis Research ST-500-1 cryostat but unfortunately previous work on the project resulted in severe damage to the system.
SolidWorks CAD of measurement enclosure
A radiative heat transfer analysis was performed to size the measurement enclosure and quantify all thermal loads on the cryogenic detector. The enclosure uses a dual-coating, Acktar Ultra Black (emissivity = 0.98) on surfaces facing the hot TRC device to maximize signal absorption, and gold plating (emissivity = 0.025) on all other surfaces to suppress parasitic radiation from the 300 K chamber walls and prevent oxidation on heated elements.
Config B (Cold Plate) is a flat cold plate faces the TRC device across a small vacuum gap. The view factor between a 575 K source and the detector was computed using the coaxial parallel disk formula (Howell catalog C-11). At a 1 mm gap, F12 = 0.986, delivering 388 mW of device heat to the detector. The signal-to-noise ratio (device heat / parasitic wall heat) peaks at 10.5 for a 1-inch diameter plate. The crossover point where wall heat equals device heat occurs at approximately 1.71 inches diameter, setting a hard upper bound on plate size.
Config A (Cold Cone): A brass cone with Acktar Ultra Black interior coating creates a cavity with effective emissivity of 0.981. The cone conducts heat from its 300 K base to the cold tip through temperature-dependent brass thermal conductivity (C26000). A 1-D conduction model integrated along the cone's tapered geometry predicts a tip temperature of approximately 62 K, yielding a total thermal budget of 830 mW with 17% margin against the cryocooler's 1 W capacity at 10 K.
For a thermoradiative cell to operate at practical power output, the device must run at temperatures up to 650 K. Standard ohmic contact metals for III-V semiconductors degrade rapidly above ~350°C due to interdiffusion, which causes the contact resistance to rise dramatically and ultimately destroys the device.
I helped to fabricated ring Transmission Line Model (TLM) test structures on heavily-doped n-type GaSb using photolithography and e-beam metal evaporation, then measured the specific contact resistivity ρc after stepwise annealing at temperatures up to 400°C. Three contact metallization stacks were evaluated: Ti/Pt/Ag, Pd/Ge/Pd, and Ag, more planned. Contact resistance was extracted by fitting total resistance as a function of ring geometry using a custom Python analysis script.
The ring-TLM geometry was selected to eliminate edge current crowding effects that would introduce systematic errors in the resistivity extraction, particularly important given the small feature sizes patterned on GaSb.
TLM testing
Graphs made by Brian O'Neil graduate researcher
Ti/Pt/Ag maintains a specific contact resistivity of ~1×10−6 Ω·cm² and remains stable throughout extended annealing at 400°C, making it the strongest candidate for a high-temperature TRC ohmic contact on n-GaSb tested so far. Pd/Ge/Pd shows higher and more variable resistivity, while Ag contacts degrade significantly at elevated temperatures.
Creating the TLM test structures used for contact characterization requires patterning metal features at the micrometer scale using photolithography. We developed and optimized the photoresist process for GaSb substrates, including systematic characterization of development time in CD-26 developer to achieve clean, residue-free pattern transfer.
A key challenge was the sensitivity of the development process to ambient humidity. Exposure and development results varied significantly with lab conditions, requiring careful control and systematic testing across a range of development times to identify robust process windows before committing to metallization.
Metal stacks (Ti/Pt/Ag, Pd/Ge/Pd) were deposited by e-beam evaporation in the CHA Industries e-beam system, followed by liftoff in solvent to produce the ring-TLM test structures. The development issues led to further problems down the line with lift off where some of the rings would get stuck and require sonication to remove them which risked damaging the rest of the device and contact.
Development and lift off
Finished contacts
Stuck rings
NPRL research group