RHIC Beams Change Their Charge
By Spencer Klein
RHIC accelerator physicists inject beams of bare nuclei (with all of their electrons removed) into the accelerator. One might expect that these beams would continue to circulate unchanged. However, every rule has an exception, and, in the 2005 run with copper beams, a team of Brookhaven National Lab, Lawrence Berkeley National Lab and CERN researchers observed that some of these bare nuclei gained electrons, transforming into a beam of single-electron ions.
The single-electron copper beams are produced in a process known as bound-free pair production (BFPP). Bare relativistic nuclei have large charges which induce strong electromagnetic fields which accompany the ions. When these electromagnetic fields collide, the field can excite the vacuum, producing an electron-positron pair. Sometimes, the electron is created bound to one of the copper nuclei, producing a single-electron atom.
For copper, the cross-section for BFPP is about 0.2 barns. At the full copper-copper luminosity, BFPP creates a beam of about 4,000 copper ions per second, carrying about 4 mW of power. The momentum of these ions is unchanged, so they are not immediately deflected from their trajectory. However, as Fig. 1 shows, the ionís reduced charge causes them to be deflected less than the bare nuclei in the accelerator dipoles (bending magnets). At RHIC, copper one-electron ions were predicted to strike the beampipe about 136 meters from each interaction point. As Figure 2 shows, thatís exactly what was seen.
When the scattered ions strike the beampipe, they initiate a large hadronic shower which develops in the RHIC magnets. These showers were detected with a set of PIN diodes which are sensitive to the charged particles in the showers. The main function of these diodes is to monitor the RHIC beam; their utility for physics studies was a nice bonus. For this experiment, they were placed in a special configuration around the area of interest.
Because the diodes are small, about 1 cm square, they only observed a small fraction of the showers; about 10-20 counts/second were observed above background. Still, this was enough for a clear detection. The diode count rates were correlated with the count rates observed in the PHENIX zero degree calorimeters; these calorimeters provide a good monitor of the instantaneous RHIC luminosity. After correction for efficiency, the measured cross-section was found to be in rough agreement with our expectations. Unfortunately, the small size of the diodes and uncertainties in the shower development precluded a precise understanding of the detection efficiency, and, hence, an accurate measurement of the cross-section.
BFPP has implications that go far beyond atomic physics. The cross-section rises rapidly as the beam particles get heavier (as Z7, where Z is the atomic number); for gold at RHIC, the calculated cross-section is about 114 barns, able to produce a beam of 342,000 one-electron gold atoms per second. This slightly depletes the circulating beams, reducing the beam lifetime. For reference, 114 barns is about 15 times the cross-section for hadronic interactions between the two beams. At RHIC, the magnet optics are such that this beam is dispersed before it strikes the beampipe, and so is not concentrated at a single point in the ring.
For lead ions at the LHC, however, this beam is a major concern. At the design luminosity of 1027/cm2/s, the expected beam of 281,000 single-electron lead ions carries about 25 watts of power. These ions strike the LHC beampipe about 380 meters from each interaction point, warming the supercooled magnets. If enough energy is deposited into the wrong part of a magnet, a quench might occur, halting accelerator operations. Calculations of the hadronic energy required to quench a magnet are rather complex, requiring a knowledge of hadronic physics (as e.g. in FLUKA), a detailed model of the magnet materials, and of the magnetic fields, heat and helium flows within the magnet. Figure 3 shows some simulations that were performed for the BFPP studies at RHIC. Our current expectation is that, if the BFPP beam strikes the middle of an LHC magnet, as expected, then the LHC will be able to reach itís design luminosity. Still, the sensitivity and implications of this effect highlight the important of good experimental measurements of the cross-section.
BFPP is just one example of what are known as ultra-peripheral collisions. In these collisions, the nuclei do not physically collide, but interact via the long-ranged electromagnetic force. STAR and PHENIX have also reported results on free electron-positron pair production (a simpler reaction than BFPP) and on photoproduction of vector mesons.
 R. Bruce et al., Phys.a Rev. Lett. 99, 144801 (2007). (arXiv:0706.2292)