Polarized Ion Source Progress:
Past Achievements! Future Aspirations?
Thomas B. Clegg
Department of Physics, University of North Carolina, Chapel Hill, NC, 27599-3255, USA,
and Triangle Universities Nuclear Laboratory,† Durham, NC, 27708-0308 USA
Abstract: Sources of nuclear spin-polarized H and D ions have been developed for over 40 years. These exist today in ~20 laboratories worldwide and are now highly refined to suit the requirements of individual accelerators and physics programs. Differences include capabilities for: particular ion species, positive and/or negative ions, dc and/or pulsed operation, and the variety of vector and (for D) tensor polarizations available. A review of the basic types of polarized ion sources used today is presented, with summaries of performances for each source type. Reference is made to features which make each type of source interesting and to limitations which its users still face. Possible future developments which hold particular promise will be presented.
1 Introduction
Development and use of nuclear-spin-polarized H and D beams over the past 40 years has provided a wealth of technical knowledge and a stable of high-quality polarized ion sources. Today such sources can be found in some 20 laboratories worldwide. At specialized workshops like this, it seems appropriate to step back, review what we have learned in developing these sources, examine the challenges we still face, and then suggest possible productive directions for future research.
Thus, in this review I will attempt to summarize and compare the performances of existing polarized ion source systems. In doing so, I will highlight both the key physics ideas learned and the important technical advances accomplished during their development. Finally, I will discuss known opportunities for, and remaining challenges to, achieving further polarized ion source improvement. Those interested can find much more information about past developments in the proceedings of prior specialized conferences and workshops [1-4].
With lots of help this past week from ‘polarized sourcerers’ worldwide, many of whom are attending this workshop, I was able quickly to gather the latest information about polarized source use and performance. Figure 1 was prepared from replies to my recent survey and provides our first glimpse at the current situation.
Operational polarized sources today are of three basic types: Lamb-shift sources, optically pumped sources (OPPIS), and atomic beam sources (ABPIS). While the oldest of these systems were built 25 years ago, three new systems were just completed and placed in operation in the last 18 months. Individual source designs, even among those of the same type, are often quite different. Features provided are always dictated by the explicit requirements of a laboratory’s accelerator and by needs of the local experimental program. Thus, these factors usually dictate the particular ion beam species provided (e.g. H+, H-, D+, or D-), the beam currents and polarizations used, and the explicit beam time structure (DC or pulsed) and the beam emittance required. With so many technical differences, it is not meaningful in many cases to compare performances of polarized sources head-to-head.
While my recent survey revealed a high general level of local user satisfaction with polarized beams in most laboratories, some reports mentioned remaining challenges if improvements needed by future experiments are to be attained. For example, scientists at COSY/Jülich and at JINR/Dubna want more beam for new accelerators. Our laboratory at TUNL needs more polarized beam for experiments below 600 keV. And, labs in Münich and Groningen want improved deuteron beam polarization and on-line polarimetry.
Collectively, operating polarized sources were scheduled for use in experiments during ~920 days in the past 12 months. However, the distribution of this use varies widely. Surprisingly to me, some very good polarized ion sources which were used extensively in the past have recently fallen into disuse. Among them, at the end of 2001 the source which has long supported the polarization studies program at PSI-Villigen is to be decommissioned with their low-energy cyclotron. The polarized source at Kyushu soon will be moved to a different campus location. Polarized sources at TRIUMF and Wisconsin lie unused awaiting new experimental demand for their beams. The polarized source and tandem accelerator at Kyoto are now instructional tools and are no longer used for basic research. I have no recent information about the source at Uppsala.
In the material which follows, I will consider separately the three ion source types presently in use, the specific design variations employed for each, and their individual performances.
2 Overview of present systems
2.1 Lamb-shift Polarized Ion Sources [5]
The oldest sources in use today are Lamb-shift sources at the Universities of Tsukuba [6], and Köln [7]. These are based on a polarization method [8] which requires the production of H0(2S) [or D0(2S)] metastable atoms by charge exchange of 550eV H+ [or 1100eV D+] beams in cesium vapor. The metastable atoms then enter a region of carefully designed electric and magnetic fields which cause selective destruction of unwanted hyperfine states by quenching them to the ground state. Metastable atoms in the remaining hyperfine state(s) are then selectively ionized by charge-exchange in argon to provide the polarized H- [or D-} beams needed. Ionization of unpolarized ground-state background atoms in the beam is suppressed.
Two variations of this source are used to polarize the beam. The Tsukuba and Kyushu [9] sources employ a Los Alamos-style [10] ‘spin-filter’ to obtain the desired hyperfine states. The Köln source is more versatile. Though it was constructed originally to utilize the Sona field-reversal method [11,12], it is now designed to interchange this system with a ‘spin filter’ when needed. The latter
system has the advantage of producing beams in a single hyperfine state with a wider variety of beam polarizations; the former system provides somewhat higher beam intensity because multiple hyperfine states are used. Detailed performance specifications for these sources are included below in Table 1.
Max. Intensity (mA) / Polarization / Ionizer / Ref.Laboratory / H- / H+ / D- / D+ / % of Max / Type / #
Cologne / 0.0005 / - / 0.0005 / - / 0.7 to 0.8 / Sona, Spin Filter / 7
Kyushu / 0.0003 / 0.0003 / - / 70(p), 65(d) / Spin Filter / 9
Tsukuba / 0.0003 / - / 0.0003 / - / 80(p), 75(d) / Spin Filter / 6
Table 1- Performance of Lamb-Shift Polarized Ion Sources
Lamb-shift polarized sources enjoy the advantages of being relatively simple, reliable, and inexpensive. They also provide DC beams having excellent emittance, a feature which makes them well suited for use with tandem accelerators which have tight phase space acceptance requirements. Users of Lamb-shift sources also enjoy the convenience of intrinsic capabilities for on-line beam polarimetry [13]. However, they often suffer from its one major disadvantage. Lamb-shift sources’ output beam intensity is very small, rarely exceeding 0.5 mA.
An important physics lesson should be learned from this limitation. In Lamb-shift sources, the process used to produce metastable atoms, H+(550 eV) + Cs à H0(2S) + Cs+, also creates intense internal space-charge electric fields from the slow Cs+ ions introduced into the beam. This E-field quickly destroys the desired (2S) metastable atoms by Stark quenching them to the ground state. These ‘lost’ atoms then cannot contribute to the Lamb-shift source’s output polarized beam intensity.
The ‘spin-filter’ mentioned above is an important technical legacy of these sources that has been adopted elsewhere for polarimetry [14], as shown in Figure 2. When it is used in the Lamb-shift source, atoms in individual hyperfine states are selected to provide an output beam with the desired polarization. When it is used as a polarimeter, the relative hyperfine state populations of the entering polarized beam are measured easily [15]. At TUNL, we routinely obtain 1% polarimetry accuracy for H+ and D+ beams, but accuracy with H- or D- beams is diminished by dilution of the beam polarization with unpolarized molecular contributions, for which the spin-filter lacks sensitivity. Accuracy of a spin-filter polarimeter at COSY/Anke is reported [16] to be better than the 1% accuracy reported at HERMES for Briet-Rabi polarimetry [17]. Another spin-filter polarimeter is under construction for the polarized source at Groningen [18].
2.2 Optically Pumped Polarized Ion Sources (OPPIS) [19]
Two optically pumped sources of polarized H- ions exist today, an excellent DC source [20] at TRIUMF which has not operated this year, and an extensively used pulsed source [21, 22] which was first placed into operation at BNL/RHIC in 2000. The OPPIS polarization method is based on pickup by 3 keV H+ ions of a polarized electron in optically pumped, polarized rubidium vapor. To preserve the nuclear polarization, this charge-exchange must occur inside a strong magnetic field. The emerging electron-polarized H0 atomic beam then experiences a rapid magnetic field flip (a Sona reversal, like that used in the Lamb-shift source) which transfers the electronic to nuclear polarization. Atoms in the beam then pick up a second electron in sodium vapor, producing output nuclear-polarized H- ions.
The two existing optically pumped sources have been highly successful. Table 2 summarizes their specifications and performance. Remarkably at TRIUMF, DC H- beams of intensity up to 0.6 mA with Pz = 85% have been attained within a normalized emittance of 2p mm-mrad. For pulsed beams having 500ms pulse length at repetition rates up to 4 Hz, 1.6 mA was attained with Pz = 80% with a similar emittance. Most impressive is the very high stability of these beams [23], exemplified by recent successful use in a series of experimental tests for parity violation.
At BNL/RHIC, polarized H- beam intensities of 0.5 to 1.1 mA with Pz =80% are now available in 400 ms pulses at 1 Hz from the source. Up to 0.58 mA has been measured after acceleration to 200 MeV in the injector linac. Earlier TRIUMF tests mentioned above indicate that further improvement in both current and Pz are possible with a longer solenoid to produce a more uniform B-field over the ECR H+ source and Rb vapor canal.
The principal technical advance that enabled major improvement in OPPIS performance is the development of an improved optical pumping laser. A Cr doped LiSrAl F crystal laser (Cr:LiSAF) developed at Livermore [24] is now commercially available and provides up to 1 kW in 500 ms pulses. Highest simultaneous values of both current and beam polarization Pz become available only when this pulsed laser is used for pumping the Rb vapor.
A smaller, but nevertheless very significant OPPIS performance improvement was obtained when a recirculating sodium jet system was developed for a new ionizer [25]. This has a larger aperture than the previous ionizer cell, providing higher beam current. Biasing the cell at –32 kV improves overall reliability by reducing the loss of Na+ ions into other parts of the source. It also accelerates quickly the desired polarized H- ions, thereby reducing the effects of space charge and improving the output beam’s emittance. The larger aperture also provides for better illumination of the Rb vapor, resulting in higher beam polarization.
What future improvements are still possible? Tests at TRIUMF in 1999 showed that still higher intensity can be achieved by replacing the ECR source with a new intense pulsed H+ plasma source whose design is based on developments at the Budker Institute. In this arrangement [22] shown below in Figure 3, up to 10 A of pulsed H+ beam from the plasma source was neutralized in H2, entered the strong axial magnetic field, and then was re-ionized in He before entering the Rb cell. Output currents I (for H-) of 8 mA and I (for H+ ) of 50 mA with Pz = 42±5% were obtained. Both beam polarization and beam intensity would have been higher with a larger B-field over the He ionizer and Rb neutralizer cells.
2.3 Atomic Beam Polarized Ion Sources (ABPIS)[26]
Atomic-beam sources are the most common type of polarized ion source today, with ~15 operational sources in existence. A summary of their performance specifications for these sources is provided in Table 3. The oldest such sources, at JINR/ Dubna and Bonn, were built in the early 1980’s. The newest, at IUCF and at Munich, were first used for experiments only in the past 18 months. The ABPIS polarization method is based on Stern-Gerlach magnetic focusing to electron-polarize H or D atoms in a slow beam. These atoms then undergo adiabatic-fast-passage RF transitions [27] to convert their electronic polarization to nuclear polarization, and subsequently are ionized and accelerated.
The atomic beam stage for these sources shares most design features with many atomic beam systems developed to inject polarized atoms into polarized H or D targets in storage rings, for example at IUCF [28], or at HERMES [29]. Though older atomic beam sources employ electromagnetic sextupoles for atomic beam focusing, all recently developed atomic beam systems have employed NeFeB permanent magnet sextupoles. The latest have improved pole-tip fields up to 1.6 Tesla [30] and are capable in DC operation of focusing up to ~8 x 1016 nuclear polarized H atoms/s in two hyperfine states into the ionizer downstream. Further gains of a factor of 2 or 3 can be achieved when the power and gas to the dissociator which produces the atomic beam are pulsed. This increase is enabled by improved vacuum and reduced gas scattering of the slow atoms near the dissociator nozzle.