RHIC Low Energy Operations
By Todd Satogata
Part of the challenge of RHIC operations is not only to increase luminosity at high energies, but to explore low energy collisions to search for a possible QCD critical point in the nuclear matter phase diagram. (See previous article by Paul Sorensen.) The search for this QCD critical point has focused on gold-gold collisions in the energy range sqrt(s_NN)=5-50 GeV. The low end of this range is far below the c.m. energy of RHIC injection collisions of sqrt(sNN)=19.6 GeV, and RHIC operation at these low energies encounter several new challenges.
A quick glance at some of the parameters already shows some of the challenges involved. At the lowest energy of sqrt(sNN)=5 GeV, the beam rigidity, power supply currents, and magnet fields are 20% of their values at normal RHIC injection energy. The beam size is larger by over a factor of two, and relativistic gamma=2.68 is lower by a factor of 3.5. The intrabeam scattering (IBS) growth time is reduced by a factor of 20, from hours at normal injection to a few minutes!
With low energy parameters, challenges for RHIC operation include poor magnetic field quality, difficulty controlling chromaticity to keep beam stable, RHIC RF control and harmonic number changes, short luminosity lifetime due to IBS, and simply fitting large low-energy beam into the RHIC aperture.
The RHIC magnets are optimized for high-field performance, and nonlinear fields grow relative to the primary field at lower energies. These nonlinearities, combined with large beam size and strong IBS, conspire to potentially reduce the beam lifetime to hundreds of seconds at sqrt(sNN)=5 GeV. This also affects control of beam chromaticity, which is important for beam stability. In future runs, some sextupole magnet power supplies will need to be changed to correct chromaticity properly at low energies. Main power supply regulation has been tested in RHIC at the lowest energy and shows no problems.
Normally the 28 MHz RF system oscillates h=360 times per turn of the beam around RHIC; each of these cycles is an RF bucket, and h is called the harmonic number. At low energies, Au beam becomes less relativistic and the beam velocity is below the RHIC RF tuning range of 28-28.17 MHz for h=360. The harmonic number must therefore be raised for collision energies less than sqrt(sNN)=17 GeV, up to h=387 at sqrt(sNN)=5 GeV. All beam synchronous clocks, including those that drive experiment triggers, must follow this change. RHIC is three-fold symmetric, so only harmonic numbers divisible by at least 3 can produce simultaneous collisions at both STAR and PHENIX experiments; fill pattern and injection constraints may require harmonic numbers divisible by 9.
Longitudinal and transverse acceptance vs emittance is another challenge at low energies. RHIC Au beam typically has a longitudinal emittance of 0.15 eV-s/u at injection. This beam barely fits into the RHIC 28 MHz RF bucket with 400 kV at sqrt(sNN)=9 GeV. At sqrt(sNN)=5 GeV longitudinal acceptance is only 0.08~eV-s/u, and significant portions of the beam immediately debunch even with perfect longitudinal injection. Transverse acceptance issues in the transfer line provide similar limitations.
To evaluate the severity of these challenges and make projections for RHIC low energy operations, there have been two one-day test runs during Run-6 and Run-7. The lattice for both these test runs was the same, with beta*=10m at all IPs to maximize aperture.
The Run-6 test run was June 5-6 2006. We used protons (the only species available at the time), and explored half-energy field quality at sqrt(s)=22.5 GeV, or a beam rigidity of 37.4 T-m, less than half of nominal proton injection rigidity. Injection efficiency of 80% was achieved, with beam lifetimes of 5-10 hours; power supplies performed well and field quality was very good. Fig. 1 shows good agreement between measured and predicted difference orbits in the yellow ring during this test. The RHIC harmonic number change was not tested because protons are still highly relativistic at this energy.
The Run-7 test run occurred June 6-7 2007. We used gold, and leveraged setup from the previous test run to produce collisions at the same beam rigidity, corresponding to sqrt(sNN)=9.18 GeV. The RF tuning at this energy requires h=366, so one goal of this run was to test this new harmonic number change, including all RHIC instrumentation, injection, and experiment triggers. With circulating beam and operating DAQ, we could also perform vernier scans and measure luminosity and luminosity lifetime. A maximum of 56 bunches were injected into each ring, with injection efficiencies of 60-75%.
Beam lifetime is shown in Fig. 2. Store lengths at this energy were about 15 minutes. Beam intensity fit well to a double exponential with decay component lieftimes of 2 minutes (possibly due to strong nonlinearities introduced to damp fast instabilities) and 20 minutes (consistent with predicted IBS growth rates). These lifetimes will likely improve in future runs as we change sextupole polarities and effect lattice corrections.
Fig. 3 shows vernier scan data and a very clean collision signal in the STAR beam beam counters (BBCs) with peak rates approaching 900 Hz, well above the anticipated rates. Uncogging showed that backgrounds were of the order of 5%. BBC backgrounds did not increase with vernier scans of up to ±9 mm, indicating large transverse aperture in the experiment interaction regions. The estimated peak luminosity was about 1.5x1024 cm-2 s-1 for these conditions.
The RHIC harmonic number change to h=366 worked well during this test for most RHIC instrumentation, but encountered problems with decoders for the RHIC abort and experiment trigger clocks. A workaround was found for both by disabling almost all beam synchronous link events; however, experiment triggers still often lost lock during the study. A small amount of data was taken at STAR, and no useful data was acquired at PHENIX. We believe we understand the root cause of this problem and can test a resolution with experiment DAQs before the next run.
Beam lifetime measurements at sqrt(sNN)=9.18 GeV can be used to optimize the store length, as shown in Fig. 4a. There is a large sensitivity of average luminosity (relative to peak luminosity) to the fast decay rate, and store lengths are optimally between 5-10 minutes for most expected low-energy beam lifetimes. Experience has shown that the detectors can stay on continuously during beam injection, so short store periods are feasible.
Fig. 4b shows the expected event rate vs c.m. collision energy. Lines have been drawn to indicate normal injection, the range of collision rates observed at sqrt(sNN)=9.18 GeV, and projected rates for at the lowest energy of interest. A test of gold collisions at sqrt(sNN)=5 GeV has been proposed for the 2008 RHIC run to determine luminosity and luminosity lifetime, and to evaluate requirements for potential AGS electron cooling. Injection efficiency of 20-50% and IBS lifetimes of a few minutes are expected, so vernier scans and luminosity measurement will be very challenging. Beam synchronous clock issues for harmonic numbers other than 360 will be resolved during the 2007 shutdown, and tested with the experiment triggers.
For more information: