Every Gateway test diode leaked, but the dose was 7.1 times flight

What the test found

NASA tested three Z4J solar-cell coupons and a diode board representing current management at the base of a Gateway photovoltaic wing. The campaign accumulated ultraviolet exposure, electron and proton radiation, xenon-ion erosion, thermal-vacuum cycles from +100 C to -183 C, and electrostatic-discharge arcs rather than testing each stress in isolation.¹

After the lunar-transit radiation step, all 15 diodes showed reverse-bias leakage. The result initially looked like a fleet-wide defect. Review then found that the coupon exposure was 7.1 times the modeled PPE dose and 4.5 times the maximum geostationary-orbit case.

Dose, temperature, and the current budget

Leakage matters only through its circuit consequence. The presentation calculates that 1.0 mA costs 0.1 W per string, while 10 mA costs 1 W per string. It also notes that the leakage path becomes active only when the affected string is shorted to ground or mostly shadowed, and that changes this small would sit below PPE telemetry resolution.

The team irradiated 29 additional flight diodes, bringing the sample to 44, at 1.35 x 10^15 one-MeV electrons per square centimetre. A 24-hour anneal at +90 C reduced leakage, relevant because the diodes run warm in flight. At 99.9 percent confidence using average plus six standard deviations, the reported maxima were 0.94 mA at -100 V and 12.5 mA at -160 V. The investigators judged the leakage acceptable over the operating range, while continuing the broader cumulative test program.

Where a skeptic should push

The evidence closes one identified diode-leakage risk; it does not certify the complete wing. Coupon tests simplify geometry, connectors, manufacturing spread, shadow patterns, and system-level fault combinations. The published presentation also predates completion of the full campaign, which it expected in spring 2024.

The strongest part of the work is not the word acceptable. It is the chain from exposure ratio, to larger sample, to annealing condition, to circuit-level watt loss. Any later change in mission duration, shielding, thermal state, or operating voltage must reopen that chain.

The bottom line

High confidence: the first 15-device result overstated flight leakage because its radiation fluence overshot the PPE case. Moderate confidence: the follow-up distribution supports acceptable diode loss within the tested operating range. The claim would fail if combined stresses create leakage outside the single-stressor statistical envelope or if flight temperatures do not supply the assumed annealing.

Space-power qualification needs consequence, not merely damage

The non-obvious implication is that detecting degradation is only the first half of qualification. The opportunity is to translate accelerated-test damage into a bounded system penalty. The threat is symmetric: dismissing a test because it is conservative can hide a coupled failure, while accepting raw leakage counts can reject good hardware. Mission exposure and circuit behavior have to meet in the same model.