Research overview: The TITAN ion trapRelativistic time dilationAtomic parity violation in francium


The TITAN Ion Trap Facility at TRIUMF

In collaboration with the TITAN group at TRIUMF.



Accurate masses of short-lived radioactive nuclei are needed not only for improving nuclear models, but also for tests of the conserved vector current (CVC) hypothesis and the unitarity of the Cabbibo-Kobayashi-Maskawa (CKM) quark mixing matrix. The TITAN facility at TRIUMF in Vancouver aims at mass measurements with a precision down to \approx 10^{-8} for radioactive nuclei with lifetimes as short as T_{1/2} \approx 50 ms [1]. For radioactive isotopes, the lifetime of the nuclei sets a limit for the observation time in ion-cyclotron-resonance based mass spectrometry, which ultimately determines the achievable precision in the mass measurements (see e.g. [2]).

The use of highly charged ions (HCI) is a promising way to push this limit, as the increased charge leads to higher cyclotron frequencies. TITAN will employ an electron-beam ion-trap (EBIT) to breed singly charged, radioactive isotopes delivered by the ISAC radioactive beam facility into high charge states.

The HCI extracted from the EBIT will have energy spreads of tens of eV/charge or more. After direct injection into the mass measurement trap, this spread would lead to large ion clouds susceptible to magnetic field inhomogeneities and decreased sensitivity of the time-of-flight method employed to measure the cyclotron frequency.

To alleviate this problem we are currently building an intermediate cooler trap to pre-cool the HCIs to \approx 1 eV/charge prior to the mass measurement. Electron cooling [3] and sympathetic cooling with initially cold protons [4] have been identified as the most promising cooling methods, since buffer gas cooling, which is well established in work with singly charged ions [5], is ruled out due to the charge transfer between HCIs and the neutral buffer gas. Electron cooling has been demonstrated in the context of anti-hydrogen production [6]. The HITRAP collaboration at GSI plans to use this method as well for their experiments with cold HCI [7]. A great advantage of electron cooling in a strong magnetic field is the self-cooling of the electrons via emission of synchrotron radiation. The biggest potential drawback is the occurrence of electron-ion recombination during the cooling process. As the mass measurement in the precision Penning trap is performed with one particular charge state, a deterioration of the charge state distribution in the cooler trap is unacceptable. The use of initially cold protons instead of electrons would eliminate this problem. However, the power radiated away via synchrotron emission scales inversely with the mass of the particle to the fourth power, rendering this mechanism useless for protons in a Penning trap. We have carried out extensive simulations of electron and proton cooling and find that with attainable densities of 10^7 electrons/cm^3 and 10^8 protons/cm^3, 10^3 HCI can be cooled down to 1 eV/charge within a fraction of a second as shown for the case of electrons in Fig. 1 (a-c) [8].

Fig. 1 Left side: A simple simulation of the electron-HCI cooling process for 10^7 electrons and 10^3 HCI. (a) Within this model, HCI get cooled in a fraction of a second. (b) Synchrotron radiation re-cools the electrons with a delay, which is actually beneficial as it delays the onset of recombination (c). Right side: Schematic model of electron-HCI cooling in a nested Penning trap. Note that unlike in this potential picture, in actual space the HCI penetrate the electron cloud during each passage.

We find that there is a window of opportunity to extract cooled ions before electron-ion recombination sets in. The latter process has significant cross section only for low relative electron-ion collision velocities; the cooling process heats the electrons up, and while synchrotron radiation cools them down again, this happens with a delay on the order of a fraction of a second for a 7 Tesla field. As the relative collision velocity is dominated by the electrons’ lab-frame velocity at this stage, recombination is initially suppressed. Furthermore, the injection and extraction of all particle species into the cooler Penning trap through the strong magnetic field gradient has been simulated in detail [9]. Based on this, we have developed a complete cooler trap design based on a large 7 Tesla superconducting solenoid magnet (funded by a CFI New Opportunity Grant, delivered in March 2008), including machine-shop ready CAD drawings. This design is currently in the TRIUMF design shop for review and will soon enter the TRIUMF machine shop. Figure 2 (a) shows the essential features of the setup. From the left side, HCI from the TITAN EBIT will enter the cooler trap beam line. Two sets of Einzel lenses focus the HCI beam properly through the magnetic field gradient into the trap structure shown in Figure 2 (b). For proton cooling experiments, a proton source is attached sideways on the injection beamline and the proton beam can be merged onto the trap axis with an electrostatic quadrupole deflector.

Fig. 2 (a) Cooler trap design; the magnet dewar has a length of 120 cm. (b) The main electrode structure inside the solenoid bore consists of 28 electrodes. This provides the flexibility to shape time-dependent, complex nested potentials. Some electrodes hidden under the grey macor jackets in the middle of the trap are quad-split for rf-excitation and detection.

In case of electron cooling, a field emission array located off-axis on the extraction side of the beamline produces an electron beam which is steered on-axis with an electrostatic/magnetic Lorentz steerer device. Electrons are injected into the trap and cool down to room temperature via emission of synchrotron radiation within \approx 1/10 of a second. After accumulating at least 10^7 electrons, HCI are entering the trap and cool down via Coulomb collisions with the electrons. As the two species have opposite charge, a special scheme called a nested trap has to be employed (see the right panel in Fig. 1). After cool-down, the HCI are extracted at the other end of the beamline, initially into a diagnostic setup for studying the cooling process, and after installation at TITAN straight into the precision mass measurement trap (Fig. 3).

Fig. 3 The cooler trap in its future location superimposed onto a photo of the current TITAN beamline.

In case of proton cooling, the procedure is similar, but since the protons emit negligible amounts of synchrotron radiation, they cannot dissipate energy and gradually heat up. At least 10^8 protons are therefore needed to cool 10^3 HCI. Charge-changing collisions with background gas are the biggest challenge for this setup as they gradually diminish the charge state of the HCI. Ultra-high vacuum of 10^{-10} torr or better is therefore essential, and great care has to be put into the design of the vacuum vessel and the components going into it. A combination of high-speed turbo-pumps and ion pumps provide high pumping speed and low base pressure, and all-metal gate valves with port apertures not smaller than those of the pumps need to be used. Ultimately, the trap area inside the solenoid bore, which cannot be directly accessed by standard pumps, will be pumped by non-evaporable getter strips lining the walls of the vessel. The electrode structure going into the vacuum vessel are manufactured from oxygen-free high purity copper and macor insulators. The system is designed to be fully bakeable to 200^\circ C.

In the past year, the TITAN mass measurement trap has been brought online using singly charged ions from ISAC and produced measurements of some high-profile short-lived radioactive isotopes such as ^8He [10], ^{11}Li [11], and ^{11}Be . In a second step, the EBIT became operational recently, with the extraction of HCI into the TITAN beamline. The facility is clearly on track towards its goal of performing mass measurements and other experiments with HCI. It is now very important to make progress on preparing these HCI for high-quality mass measurements by readying the cooler trap. In addition to the TITAN collaboration, three other groups are currently working on similar trap designs (at RIKEN [12], HITRAP [7] and MATS at GSI, for a variety of experiments, including mass measurements). We expect the machined trap and vacuum parts to come out of the TRIUMF machine shops in the summer of 2009.


[1] J. Dilling et al., NIM B 204, 492 (2003).
[2] K. Blaum, Phys. Rep. 425, 1 (2006).
[3] D.S. Hall et al., Phys. Rev. Lett. 77, 1962 (1996).
[4] V.L. Ryjkov et al., Eur. Phys. J. A 25, 53 (2005).
[5] G. Bollen et al., NIM A 368, 675 (1996).
[6] S.L. Rolston et al., Hyp. Int. 44, 233 (1988).
[7] J. Bernard et al., NIM A 532, 224 (2004).
[8] Z. Ke et al., Hyp. Int. 173, 103 (2006).
[9] Z. Ke, PhD thesis, Univ. of Manitoba (2008).
[10] V.L. Ryjkov et al., Phys. Rev. Lett. 101, 012501 (2008).
[11] M. Smith et al., Phys. Rev. Lett. 101, 202501 (2008).
[12] N. Oshima et al., NIM B 205,178 (2003).