Sonoluminescence

Sonoluminescence (SL) is the mysterious and fascinating phenomenon of the ultrashort flashes of light emitted during the catastrophic collapse of a gas bubble caused by an acoustic wave. In some sense, it's the conversion of sound into light. Although the phenomenon was observed as far back as 60 years ago [1], it wasn't until 1990 that it was produced in a single isolated bubble [2], and 1991 that serious studies of it first appeared in the literature [3]. Now Hiller et al. have carried out a precise and careful measurement of the temporal properties of this luminescence [4], further narrowing the range of models to explain the phenomenon.

``Phenomenon'' is a completely appropriate description for SL because there is very little physical understanding, even in the face of an overwhelming body of experimental information. Some simple considerations are instructive: Measurements have shown that typical bubbles have an equilibrium radius of ∼ 5 µm at STP conditions, and that during the acoustic cycle the bubble is expanded to ∼ 40 µm and then rapidly collapsed to a minimum radius of ∼ 0.8 µm. Thus the gas is enormously compressed, suggesting temperatures as high as ∼ 5 × 10⁴ K ∼ 4 eV, enough to produce significant ionization and plasma conditions. However, the ideal gas behavior seems like the least violent scenario for the collapse, and that the far-from-equilibrium conditions that prevail are much more severe.

Experiments are typically done in a 100 mL (2" diameter) approximately spherical flask filled with distilled and carefully degassed water (see figure). The acoustic excitation is provided by one or more piezo-electric transducers fed by a resonant ∼ 25 kHz circuit. Except for an oscilloscope, the equipment costs a few hundred dollars and fits in a coffee can. There are extremely befuddling criteria for the dissolved gas: pure N₂ or O₂ don't work, nor does any mixture of them. The critical ingredient seems to be rare gases, but only in the 1/2 - 2% concentration range. Impurities at few µ-molar concentration levels of many solutes, for example, alcohols, completely quench SL.

The total light output of a flash is ∼ 10^-12 J and the spectral density is consistent with a black-body spectrum near 10⁵ K. Since the flash repetition rate is ∼ 25 kHz, the appearance to a dark adapted observer ∼ 50 cm away is that of a fifth magnitude star, and thus SL is readily observed with the naked eye. The flash duration τ is less than 50 ps. In fact, there is no known photodetection system that is fast enough to measure it, and some detector manufacturers use SL to calibrate the time resolution of their products.

Needless to say, there is a much larger litany of phenomenological information about SL than given above, and no simple theory about the origin of the light, much less its spectral and temporal characteristics, has been given. There are ardent proponents of an incredibly diverse array of speculations about SL, many of them supported by significant arguments and/or detailed calculations. An abbreviated list contains not only the obvious black body, shock waves, bremstrahlung, and subtle collision effects, but also includes more exotic effects that derive from QED. The most recent experiments provide support for the most violent of these mechanisms simply because it seems that everything happens at once. What remains totally mysterious, however, is why there is such strong dependence on the gas mixture and on the choice of liquids, and why such violent events are so strongly dependent on seemingly mild conditions such as the initial water temperature.

[1] H. Frenzel and H. Schultes, Z. Phys. Chem., B27, 421 (1934).
[2] D. Gaitan and L. Crum, J. Acoust. Soc. Am., 87, S141 (1990).
[3] B. Barber and S. Putterman, Nature, 352, 318 (1991).
[4] R. Hiller, S. Putterman, and S. Weninger, submitted to Phys. Rev. Lett.