Evidence for Nuclear Emissions During Acoustic Cavitation

R. P. Taleyarkhan,1* C. D. West,1 J. S. Cho,2 R. T. Lahey Jr.,3 R. I. Nigmatulin,4 R. C. Block3

In cavitation experiments with deuterated acetone, tritium decay activity above background levels was detected. In addition, evidence for neutron emission near 2.5 million electron volts was also observed, as would be expected for deuterium-deuterium fusion. Control experiments with normal acetone did not result in tritium activity or neutron emissions. Hydrodynamic shock code simulations supported the observed data and indicated highly compressed, hot (106 to 107 kelvin) bubble implosion conditions, as required for nuclear fusion reactions.

1 Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA.
2 Oak Ridge Associated Universities, Oak Ridge, TN 37830, USA.
3 Rensselaer Polytechnic Institute, Troy, NY 12180, USA.
4 Institute of Mechanics of Ufa-Bashkortostan Branch of the Russian Academy of Sciences, 6 Karl Marx Street, Ufa, 450 000, Russia.
* To whom correspondence should be addressed. E-mail: zrt@ornl.gov


The intense implosive collapse of gas or vapor bubbles, including acoustically forced cavitation

bubbles, can lead to ultrahigh compressions and temperatures and to the generation of light flashes attributed to sonoluminescence (SL) (1-21). Our aim was to study ultrahigh compression and temperatures in bubbles nucleated by means of fast neutrons, whereby bubble nucleation centers with an initial radius R0 of ~10 to 100 nm are created, and the bubbles grow in an acoustic field to a maximum radius (Rm) of ~ 1 mm (19) before an implosive collapse. This approach builds on the observations that (17) increasing Rm modestly (for example, by ~50%), or increasing the rate of collapse (16), can result in very large increases in peak gas temperatures and produce light emission during implosions. In contrast to single-bubble SL experiments, in which the initial bubble radius R0 typically increases to Rm by a factor of only ~10 (for example, from ~10 µm to ~100 µm), our neutron-induced nucleation technique results in Rm/R0 of ~105. For a spherical bubble, the increase of Rm/R0 by a factor of 104 implies a related volumetric ratio increase of 1012 over that produced by conventional techniques. Our expectation was that such an approach, with its vastly increased energy concentration potential during implosions, should give rise to significant increases in the peak temperatures within the imploding bubbles, possibly leading to fusion and detectable levels of nuclear particle emissions in suitable fluids.

To minimize the effect of gas cushioning by promoting rapid condensation during implosive collapse, we elected to work with highly degassed organic liquids. An organic liquid was chosen [normal acetone (C3H6O) as the control fluid and deuterated acetone (C3D6O) as the test fluid] because it permitted the attainment of large tensile states without premature cavitation, and thus a lot of liquid superheat would be present before nucleation. Organic liquids also have relatively large phase change coefficients, which is important, as described later. Unless otherwise noted, the liquid in the chamber was maintained at ~0°C (which was the lowest value obtainable with the equipment we used). The test liquid was degassed and subjected to an acoustic pressure field that oscillated in resonance with the liquid sample and its container. The nucleation of vapor bubbles was initiated with fast neutrons from an isotopic source (Pu-Be) or from a pulsed neutron generator (PNG) that produces 14-MeV neutrons on demand at a predefined phase of the acoustic pressure field.

Experimental system. In the experimental apparatus (Fig. 1), the test liquid was placed in an approximately cylindrical glass flask and driven acoustically with a lead-zirconate-titanate (PZT) piezoelectric driver ring attached to the outer surface. Either a plastic or a liquid scintillation detector was used for detection of neutron and gamma signals (22). The light was detected and amplified in a photomultiplier tube (PMT). A liquid scintillator (LS) detector-based system was set up for pulse-shape discrimination (PSD) (23, 24) [see Web Supplement 1 (25)]. The PSD circuit separates neutrons from gamma rays on the basis of differences in the PSD scintillator signal decay time between neutrons and gamma rays. The system could be operated to permit blocking of gamma rays (hereafter, a mode of operation referred to as "with PSD"). The net efficiency for fast neutron detection was estimated to be ~5 × 103 (26).

Fig. 1.

Schematic of the experimental setup. The distance from the scintillator head to the PNG is ~15 cm; from the scintillator head to the chamber surface, ~0 to 2 cm; from the chamber center to the PNG, ~20 cm; and from the PMT to the chamber surface, ~5 cm. The system (the chamber, PNG, and PMT) is ~1.5 m above the floor. [View Larger Version of this Image (19K GIF file)]

In the experimental sequence of events (Fig. 2), neutrons from the PNG nucleated vapor bubbles
in the tensioned liquid when the cavitation threshold was exceeded at the time of the neutron burst. The nuclear radiation detector typically detected a pulse when the PNG was fired. Thereafter, the vapor bubbles grew until increasing pressure in the liquid during the second half of the acoustic cycle caused them to begin to collapse. If the implosion was robust enough, the bubble emitted a SL light flash, which could be detected by the PMT. In theory, if the liquid is composed of deuterium (D) and/or tritium (T) atoms, and the conditions are appropriate for D-D (or D-T) fusion, nuclear particles (neutrons and gamma rays) would be emitted and seen in the response from either the plastic scintillator (PS) or LS detector. Moreover, in cavitation experiments, when a bubble implodes, a pressure wave that travels at about the speed of sound in the test liquid is also generated and can be detected at the chamber walls by microphones. Significantly, our experiments were characterized by bubble dynamics within clouds of many bubbles.

Fig. 2.

Experimental sequence of events. [View Larger Version of this Image (24K GIF file)]

Timing of key parameters. With our configuration of the PNG and electronic timing systems, we
found, by analyzing the time spectrum of neutrons (25), that neutrons were emitted over a time span of ~12 µs [~4 to 6 µs at full width at half maximum (FWHM)], after which neutron counts were reduced considerably by 15 to 20 µs after the PNG fired (25). We initiated the PNG burst when the fluid tension state was greatest. For multiple-bubble implosions, several bubbles can implode and emit closely spaced SL flashes during any given cycle.

The time between a SL flash and the signals received at two microphones set up on diametrically opposite sides of our chamber was found to be ~27 µs, which is in agreement with the time for a shock wave to travel from the center of the chamber to the glass wall (about 32 mm away). This result indicates that the bubbles generally nucleated and imploded in or around the central axis of the test chamber (27). The efficiency of SL flash detection was dependent on the PMT bias voltage (which determines the gain) and the chosen discriminator settings (29).

Our data were obtained with a PZT drive amplitude much greater than that required for threshold nucleation. Because of this, and because the PNG pulse width was about 4 to 6 µs (FWHM), nucleation could occur a few microseconds before or after the minimum liquid pressure was reached. The timing of the SL flash relative to the PS pulse was analyzed with a multichannel analyzer (MCA). The PZT drive frequency was about 19.3 kHz, which corresponds to a full cycle time of about 52 µs. The time spectrum of events (25) confirmed that the PS flash corresponding to the PNG activation (lasting about 12 µs, with 4 to 6 µs FWHM) was followed by a SL flash (lasting about 4 to 6 µs FWHM) about 27 to 30 µs later.

Experimental observations for C3H6O and C3D6O. We conducted experiments with C3H6O (100% nominally pure) and C3D6O (certified 99.92 atom % D-acetone), filtered before use through 1-µm filters. Degassing was performed by applying a low pressure of about 10 kPa and acoustically cavitating the liquid for about 2 hours. To ensure continued robust nucleation growth and implosive collapse, the drive voltage to the PZT was set to be about double that needed for occasional cavitation (defined here as the occurrence of nucleation and collapse within a 10-s observation period). The negative pressure threshold for bubble nucleation by neutrons and alpha particles in acetone is -7 to -8 bar (20, 21). A pressure map of the chamber was obtained by means of a calibrated hydrophone. Using the scale factor for induced pressures in our chambers versus drive voltage to the PZTs, and gradually increasing the drive amplitude, we determined that the cavitation began at about -7 bar, which is consistent with the known value (20, 21). The pressure amplitude in our chamber was ~ ±15 bar (±220 pounds per square inch).

T detection, monitoring, and estimation. The D-D fusion reaction can have one of two outcomes that occur with almost equal probability. The first leads to the production of helium (He) and 2.5-MeV neutrons; the second to the production of T and protons. Therefore, in addition to the evidence collected for neutron or gamma ray activity, the formation of T would provide compelling evidence of D-D fusion.

To measure T activity, we sampled the experimental fluid directly with a scintillation counter calibrated for detecting T ( 30.
The acoustic chamber was a high-Q system and as such required continuous tuning for optimal performance and bubble implosion occurrence. Time to obtain 100 coincidence data points as shown in Fig. 5B averaged about 30 min. Standard deviation was computed by taking the square root of the sum of the counts in each time bin. From the MCA time spectrum, we calculate an instantaneous rate of ~ 1 to 50 neutrons and gamma rays per second during the time of bubble implosion and SL light emission. For a 20-µs time window, and a rate of about one SL flash/s for a coincidence gathering time of ~1600 s, the number of random coincidences was calculated to be negligible [(20 × 106) × (1 to 50) × 1 × 1600 ~ (0.03 to 1.6)]. In mode 2 operation, data were obtained on a two-channel 500-MHz scope. Simultaneous time spectra data were not possible to obtain for SL and scintillator signals. These were obtained separately with an MCA (25) under identical mode 2 coincidence experimental conditions and revealed insignificant deviation from run to run. These data were then used to estimate random coincidences. It was determined that coincidences occurring during the time of PNG operation would all be random. However, as discussed earlier in this note, the random coincidences during bubble implosion appear to be insignificant.

D. Shapira, M. J. Saltmarsh, "Comments on the Possible Observation of D-D Fusion in Sonoluminescence" (2002) (http://www.ornl.gov/ slsite).

R. P. Taleyarkhan, R. C. Block, C. West, R. T. Lahey Jr., "Comments on the Shapira/ Saltmarsh Report" (31) (http://www.rpi.edu/ ~laheyr/SciencePaper.pdf).

Y. B. Zeldovich, Y. P. Raizer, Physics of Shock Waves and High Temperature Hydrodynamics Phenomena (Academic Press, New York, vols. 1 and 2, 1966).

R. Nigmatulin, Dynamics of Multiphase Media (Hemisphere, vol. 1, 1991), p. 244.

R. F. Trunin, et al., Khimiche Skaya Fizika 11, 424 (1992).

I. S. Akhatov, et al., Phys. Fluids 13, 2805 (2001) [CrossRef].

N. Chodes, J. Warner, A. Gagin, J. Atmos. Sci. 31, 1351 (1974) .

D. E. Hagen, et al., J. Atmos. Sci. 46, 803 (1989).

B. Paul, Raketnaya Technika 9, 3 (1962) .

R. A. Gross, Fusion Energy (Wiley, New York, 1984).

H. S. Bosch and G. M. Hale, Nucl. Fusion 32, 611 (1992).

R. T. Lahey Jr., R. I. Nigmatulin, R. P. Taleyarkhan, in Proceedings of the 3rd International Conference on Transport Phenomena in Multiphase Systems (HEAT-2002), Kielce, Poland, 24 to 27 June 2002, in press.

Y. Tomita and A. Shima, Acoustica 71, 161 (1990).

I. S. Akhatov, V. A. Baikov, R. A. Baikov, Fluid Dyn. 21, July, 161 (1986).

Sponsorship of this research by the U.S. Defense Advanced Research Projects Agency is gratefully acknowledged. We thank D. Steiner of Rensselaer Polytechnic Institute and D. Shapira of ORNL for critical reviews and valuable feedback with constructive comments. We also wish to acknowledge the efforts made by the staff of the Baskortostan Branch of the Russian Academy of Sciences regarding the HYDRO code simulations; in particular, I. Akhatov, R. Bolotnova, N. Vakhitova, A. Topolnikov, and K. Zakirov.

31 October 2001; accepted 30 January 2002
Include this information when citing this paper.

Reprint (PDF) Version of this Article

PubMed Citation

Supplemental Data 1 and 2 are in pdf format

Search Medline for articles by:
Taleyarkhan, R. P. || Block, R. C.

Collections under which this article appears: