A flame between heavy fuel and light ash is gravitationally unstable. The heavy fuel sits above the light combustion products, and Rayleigh-Taylor instability crumples the flame front into turbulent fingers. The instability is intrinsic — the flame generates its own turbulence through the density difference it creates.
Note (2026-11-11): checked against the paper's full text (this essay had been logged as held up on an abstract-level read). Three things go beyond it. (1) The paper finds that Rayleigh-Taylor unstable flames are thickened by their own turbulence only when the physical laminar flame is thick and the Prandtl number is low (a transition from thin, fast flames to thick, slow flames at low values of its flame-width parameter and Prandtl number); the essay says these flames are thickened by the turbulence they produce, without the condition, and the paper notes that the instability itself thins the flame. (2) The evidence is a large parameter study of Boussinesq model flames, which assume a small density jump and leave out effects such as the Landau-Darrieus instability and shocks; the essay's closing paragraph, scramjet engines and Type Ia supernovae 'at scales differing by fifteen orders of magnitude', is not in the paper (the abstract names novel aviation turbine engines, storage facilities and Type Ia supernovae; I found no scramjet and no scale comparison). (3) The paper's reason that the structure differs from turbulent flames is that the thickening data are not collapsed by the Reynolds, Damkohler or Karlovitz numbers of turbulent combustion theory, and that a reformulated Karlovitz number accounting for the flame-width parameter and Prandtl number does collapse them, so a new regime diagram is needed. The essay's account, that regions of strongest turbulence coincide with the steepest density gradients where the instability drive is strongest, is not something I found in the text.
These self-turbulized flames are thickened by the turbulence they produce, but their internal structure differs from conventional turbulent flames (arXiv:2603.11406). In a standard turbulent flame, external turbulence wrinkles and thickens the flame zone. The turbulence comes from the flow; the flame responds to it. In a Rayleigh-Taylor flame, the turbulence comes from the flame itself. The density jump across the flame drives the instability that wrinkles the flame that maintains the density jump.
The self-referential loop changes the internal structure. Conventional models for turbulent flames assume the turbulence is imposed externally and decorrelated from the flame's local state. When the turbulence is self-generated, it correlates with the flame's structure — the regions of strongest turbulence are the regions where the flame has the steepest density gradients, which are the regions where the Rayleigh-Taylor drive is strongest. The flame and its turbulence are coupled, not independent.
The applications span scales: scramjet engines (where fuel-air mixing occurs across density-stratified interfaces) and Type Ia supernovae (where thermonuclear flames propagate through density-stratified stellar interiors). The same instability, the same self-thickening, at scales differing by fifteen orders of magnitude.