New research from the lab of mechanical science and engineering professor Nenad Miljkovic identifies the mechanism behind a boiling regime that has been observed since the 1980s but never explained: under strong subcooling, rapid condensation intercepts vapor coalescence and redirects the instability responsible for the boiling crisis into a stable, high-frequency oscillating state. Their work recasts the critical heat flux, long treated as the terminal limit of boiling heat transfer, as a fork in the pathway rather than an absolute thermal ceiling.
A research team led by professor Nenad Miljkovic in The Grainger College of Engineering at the University of Illinois Urbana-Champaign has published a new study in the Proceedings of the National Academy of Sciences. The work delivers the first unified mechanistic account of microbubble emission boiling, a regime that has been observed to carry heat fluxes past the classical boiling limit since it was first reported in 1981, and identifies it as a distinct boiling mode that the team calls condensation-driven oscillatory boiling.
Boiling is the most effective tool engineers have for moving heat off a hot surface, and it underpins the performance of power plants, refrigeration systems, electronics cooling components, and even the divertors that will absorb heat inside fusion reactors. But every boiling system has a performance ceiling. Push the heat flux (heat transfer per unit area) high enough and the vapor bubbles leaving the surface merge into a continuous film that insulates the hot surface, rather than cooling it. Surface temperature spikes, the metal can melt, and the device fails. This is the boiling crisis, and the heat flux at which it occurs — the critical heat flux — has been treated as a hard limit since Novak Zuber’s classical hydrodynamic theory of 1959.
There has long been an exception on the record. In 1981, Japanese researchers reported that once water is subcooled by roughly 30 degrees Celsius or more (i.e. when it’s 30 degrees below its boiling point), boiling can continue past the critical heat flux, accompanied by clouds of tiny bubbles. The phenomenon, named microbubble emission boiling, has been reproduced by many groups over the past four decades and repeatedly proposed for high heat flux applications. What was missing was an explanation that could fully describe the phenomenon. No mechanism accounted for how the system avoids vapor blanketing, and no criterion predicted the conditions under which it would appear.
“The phenomenon itself was never in doubt,” said first author Jiazheng Liu, a PhD student in Miljkovic’s lab. “What was missing was the physics connecting what you see at the surface to how much heat actually moves. Without that link, you can’t predict when the regime will appear, and you can't design for it.”
Watching a surface that never dries out
The team built a high-power pool boiling platform operating in an ice-water bath — roughly 100 degrees Celsius of subcooling, which is the most that water can sustain at atmospheric pressure. The heated surface was deliberately simple: a 1 cm × 1 cm mirror-polished copper sample with nanoscale roughness and no coating, wick, or microstructure of any kind.
They then instrumented it four ways simultaneously: temperature sensors embedded millimeters from the boiling surface, high-speed imaging from the side and from above through an endoscope, a pressure sensor recording the acoustic spectrum, and velocimetry that used the non-condensable gas bubbles released during each collapse — normally treated as a nuisance — as intrinsic tracers.
The picture that emerged was not a vapor film at all. Vapor structures formed on the surface and were destroyed by condensation into the cold bulk before they could coalesce into a blanket, then re-formed in place. The velocity fields and the pressure spectra converged independently on the same cycle frequency, which climbed from roughly 445 to 700 Hz as heat flux increased. The near-wall liquid temperature reversed on entering the regime — a signature not documented in previous studies — as cold liquid was driven back to the surface with every collapse.
The researchers found that feeding that single measured timescale into a transient-conduction model reproduced the measured heat flux to within 20% across the entire regime, up to 1.4 kW per square centimeter, where the researchers stopped increasing power rather than risk the apparatus. Oscillation frequency, not bubble size or departure dynamics, governed the heat transfer.
A criterion, not just a mechanism
If the regime exists because condensation outruns coalescence, then the boundary between the two outcomes should be a ratio of their timescales. The team derived exactly that: a dimensionless criterion, Π, comparing the condensation time of a vapor structure to the time the hydrodynamic instability needs to merge neighboring structures into a film. When Π is less than one, condensation wins and the oscillatory regime is selected; when Π exceeds one, the surface blankets and film boiling follows.
The criterion contains no fitted parameters, and the subcooling threshold it predicts is consistent with the onset conditions reported empirically since the 1980s. Hysteresis tests, in which the heat flux was swept up and back down, confirmed that the regime is a reproducibly stable branch of the post-critical-heat-flux response rather than a transient excursion. The instability Zuber identified still occurs; what changes is what happens next, and phase-change kinetics, not hydrodynamics alone, decides which branch the system follows.
“For sixty years the critical heat flux has been taught as the end of the line,” said Miljkovic. “What this work shows is that it is a branch point, and that we know which branch you get. That reframing is what makes the regime designable rather than merely observable.”
Their finding gives engineers something they didn’t have before: a way to determine in advance whether a given set of conditions will burn out or oscillate. The wider lesson, a point developed in discussions with co-author M. Taher A. Saif, is that phase-change kinetics can redirect a hydrodynamic instability into a stable, high-transport state — a possibility that is not confined to boiling.
_____
“Microbubble emission boiling reinterpreted: Condensation-driven oscillations enable heat transfer beyond the critical heat flux,” Proceedings of the National Academy of Sciences. https://www.pnas.org/doi/10.1073/pnas.2612649123
Authors: Jiazheng Liu, Jiayi Zhang, Duoqi Liu, Kai Luo, Parsa Faghihi, Siyan Yang, Valentin Belosludtsev, Tianxiao Liu, Chi Wang, Vivek S. Garimella, Vishwanath Ganesan, M. Taher A. Saif, Nenad Miljkovic.
Nenad Miljkovic is a Founder Professor and Head of the Department of Mechanical Science and Engineering, with additional appointments in the Department of Electrical and Computer Engineering and the Materials Research Laboratory. He leads the Energy Transport Research Lab at Illinois.
M. Taher A. Saif is the Grainger Distinguished Chair in Engineering and a professor of mechanical science and engineering.
Jiazheng Liu is a mechanical engineering PhD student in the lab of Nenad Miljkovic.