A bubble chamber brings new capabilities to the search for WIMPs

The bubble chamber is making a comeback. For a quarter century several decades after its invention by Donald Glaser in 1953, bubble chambers filled with liquid hydrogen or heavy liquids played a leading role in particle physics. But their importance waned in the late 1970s as collision energies increased, events became crowded with tracks, and experimenters concentrated on the search for ever-rarer phenomena.

Now, however, the COUPP (Chicago-land Observatory for Underground Particle Physics) collaboration has reported the successful operation of a small innovative bubble chamber, 100 meters underground at Fermilab, in a search for stable, weakly interacting massive particles. ¹ Called WIMPs, these putative particles are thought to constitute cosmological dark matter (see Physics Today, August 2007, page 16 ). The chamber operates in a novel continuously sensitive mode. The collaboration’s spokesman is Juan Collar (University of Chicago).

The prototype bubble chamber’s year-long run was meant primarily as a proof of principle, with only preliminary efforts at controlling background neutrons and alphas. Nonetheless, this first run has already yielded an improved upper limit on the WIMP cross section for one kind of elastic scattering off protons, for WIMP mass below 30 GeV (see the figure on page 25.)

As the solar system sails through the galaxy’s presumed accumulation of WIMPs, theorists expect that detectors of various kinds should record, at best, a few nuclear recoils from WIMP collisions per kilogram of detector material per year. It’s hard to imagine that a bubble chamber functioning in its traditional pulsed mode would be of much use in a quest for something so elusive.

When pressure on a bubble chamber’s liquid is suddenly lowered, it becomes temporarily superheated and boils only at the site of some disturbance like the ionized trail of a passing charged particle. What’s so promising about bubble chambers in the search for WIMPs is that one can control the threshold for bubbling by adjusting pressures and temperatures so that only nuclear recoils create bubbles and the far more numerous Compton-scattered electrons do not. All the other WIMP detectors now in use are plagued by an enormous background of electrons scattered to high energy by gammas from radioactive contaminants.

At the keV recoil energies in question, electrons cause significantly less heating per unit track length in the bubble chamber liquid than do nuclei. By carefully adjusting the chamber’s operating pressure so that the bubble threshold is just above the local heating the electrons generate, Collar and company have reduced the electron background more than a billionfold.

Without special care, however, the superheated state survives only milliseconds before bubbling becomes general. That was no problem at old-fashioned accelerators with beam-pulse rates not much faster than 10 Hz. One simply synchronized the bubble chamber’s pressure cycle to the beam cycle so that the liquid was in its sensitive superheated state just when the beam came passing through.

Continuous sensitivity

To be useful for WIMP searches, a bubble chamber has to persist in the superheated state for much longer than it takes to regain that state after there’s been bubbling. Glaser had actually envisioned bubble chambers with long-term superheated sensitivity. The major problem has been so-called pool boiling. Irregularities and porosities on vessel surfaces instigate boiling in superheated liquids. Because that’s a safety issue for nuclear reactors, there’s been considerable engineering research on the subject. COUPP has exploited that research to prepare the quartz inner surface of its one-liter prototype chamber so that there’s no pool boiling.

The chamber’s active liquid is trifluoroiodomethane (CF₃I), a fire-extinguisher material twice as dense as water. At ambient pressure it boils at -22 °C. But under pressure it provides COUPP with the convenience of a room-temperature bubble chamber. Pool boiling at the metal surface of the bellows that provides that pressure is avoided simply by interposing a layer of water on top of the immiscible CF₃I.

In the absence of pool boiling, one only has to recompress the chamber when the CCD cameras monitoring the liquid detect bubbling instigated by a nucleus recoiling from radioactive or cosmic-ray background—or from an actual WIMP. Then, regaining the superheated state takes about half a minute in the COUPP bubble chamber. Despite the shallow depth of the Fermilab demonstration experiment and the prototype chamber’s relatively high radioactive contamination, Collar and company were able to maintain a duty cycle of about 80%.

No pretty pictures

In the heyday of bubble chambers, GeV charged particles from high-energy collisions created long, delicate bubble trails that made for some memorable images. But a recoiling keV nucleus in the COUPP chamber travels only a few hundred angstroms before stopping. So it generates only one lone bubble.

How would one distinguish a real WIMP from a background neutron or alpha particle that produces a similar elastic nuclear recoil? Because WIMPs are presumed to have scattering cross sections characteristic of the weak interactions, it’s thoroughly improbable that a WIMP would collide more than once in the chamber. A neutron, by contrast, is unlikely to experience only a single collision, especially in a bubble chamber much bigger than the COUPP prototype. Furthermore, alphas from radioactive decay in the vessel walls or from radon gas creeping around seals give themselves away by the bubble’s location. But radioactive contaminants actually dissolved in the CF₃I, being more problematic, are the object of ongoing purification efforts. Happily one can often distinguish alphas because two or three of them appear in temporal sequences characteristic of particular decay chains.

Spin dependence

Aside from its thermodynamic convenience, CF₃I has important nuclear merits. WIMPs are expected to scatter elastically off nuclei in two different ways: spin-dependent and spin-independent. Even though the intrinsic SD cross section for scattering off a free proton is expected to be much bigger than the corresponding SI cross section, the latter is easier to search for with heavy nuclei. That’s because the SI scattering amplitude, being a coherent sum over all the nucleons in a nucleus, grows like the nuclear mass. But the SD interaction is impervious to spinless nuclei and spin-0 nuclear cores. It sees only free or unpaired valence nucleons.

CF₃I is admirably suited for both searches. With 127 nucleons, the iodine nucleus has more than ten thousand times the SI cross section of a proton. And more important for the COUPP bubble chamber’s first published result, the much lighter spin-1/2 fluorine nucleus—of which there are three in every molecule—is predicted to be especially sensitive to SD interaction with WIMPs.

If WIMPs are found with CF₃I, it would be easy and instructive to swap liquids in a bubble chamber. Perfluorobutane (C₄F₁₀), for example, would be even more sensitive to SD interaction; and there’d be less SI scattering. Comparing observed scattering rates in the two liquids would help distinguish among competing theoretical conjectures as to what WIMPS really are.

The DAMA collaboration, which began operating sodium iodide detectors in Italy’s underground Gran Sasso laboratory in 1996, is the only group so far to claim experimental evidence of WIMPs. ² DAMA reported a small seasonal variation in collision rates, which it attributes to the annual cycle of relative velocity between Earth and the supposed local swarm of WIMPs. Subsequent experimental upper limits have largely excluded the SI scattering predicted by the leading WIMP theories as an explanation for the DAMA result. ³ But, as shown in the figure, the small sliver of SD parameter space that remained consistent with the DAMA claim and later null results has only now been excluded by the COUPP limit.

“If WIMPs do exist,” says Collar, “they are so elusive, and masquerading backgrounds are so insidious, that a lone reported sighting is bound to be received with skepticism. Therefore it’s important that the search go forward with a variety of detector types with different background problems.”