Abstract Ground tests of microwave electrothermal thrusters (METs) are conducted under 1 g, where buoyant forces distort the plasma morphology and bias performance data. We investigate these effects in a 2. 45-GHz, TM ᶻ₀₁₁ cavity thruster operated with nitrogen over 200–2000 sccm and at 80–600 W. Three dimensionless surrogates—a modified Bond number Bₒ (buoyancy), a swirl group, and an electromagnetic anchoring group ₄₌ —bring all experimental points onto a single, self-consistent scaling curve. Guided by these surrogates, we then recast the horizontal-nozzle data into a centered, reduced-order surrogate: a log-linear form in ṁ, P₈₍, and the stagnation-pressure ratio P₀, ₇/P₀, ₂ with a single exponential nonlinearity in P₀, ₇/P₀, ₂. The inferred elasticities show displacement grows with mass flow while remaining only weakly sensitive to small, local changes in P₈₍ or P₀, ₇/P₀, ₂ ; the dominant response is a sharp departure once P₀, ₇/P₀, ₂ moves away from its near-optimal neighborhood. Orientation controls performance: with the nozzle upward, buoyancy helps anchor the discharge at the throat and P₀, ₇/P₀, ₂ often approaches 3: 1 (peaking near 3. 3–3. 5), whereas in horizontal/downward tests buoyancy displaces the bubble off-axis, reducing P₀, ₇/P₀, ₂ by 20-40\%. Equal top–bottom injection maximizes swirl confinement and suppresses buoyant drift. The surrogate offers a compact path to correct 1-g measurements toward flight-like conditions and to couple directly with thrust and Iₒ models for MET design.
Biswas et al. (Wed,) studied this question.