CERN Accelerating science

ISOLTRAP’s MR-ToF mass separator to plumb a neutron star

by Robert Wolf for the ISOLTRAP collaboration

The mass of an atomic nucleus is one of its most fundamental properties.Via the binding energy, i.e., the difference in mass between its constituents (protons and neutrons) and the entire nucleus, it reflects the sum of all interactions present. Precision mass measurements of short-lived isotopes provide access to the nuclear binding energies which are a rich source of information about nuclear structure effects like pairing, halos, and shell closures, as well as about the deformation of nuclei and the search for super-heavy elements [Lunney2003,Blaum06,Schweikhard2006]. Furthermore, mass measurements are important for the modeling of nucleosynthesis processes that result in the formation of heavier elements and are needed for the explanation of their current abundances [Arnould2007]. In particular the rapid neutron capture, or r-process, is one of the most active fields where precision masses are necessary. Measurements with ISOLTRAP [Mukherjee2008] at ISOLDE-CERN play a key-role in studying the interior of neutron stars and to build an equation of state for this special class of objects [Wolf2013,Kreim2013]. Finally, in the last few years an emerging field of application for high-accuracy mass measurements of radionuclides concerns tests of the Standard Model and specifically of fundamental symmetries like the conserved vector current (CVC) hypothesis, the unitarity of the Cabibbo-Kobayashi-Maskawa (CKM) quark mixing matrix and the search for scalar or  tensor currents in the weak interaction.

Using ISOLTRAP (Fig. 1), the masses m of more than 500 short-lived nuclides have been measured with relative uncertainties of δm/m =1×10-7 and below. ISOLTRAP in effect weighs radioactive nuclides created at ISOLDE by using the elegant technique of exciting the movement of single ions stored in a Penning trap. The basic principle is the determination of the cyclotron frequency ωc=qB/m in the magnetic field B of a Penning trap, from which the charge-over-mass ratio q/m of the unknown ion can be derived.

Until 2010, the experimental apparatus consisted of three ion traps, one linear Paul trap (radio-frequency quadrupole cooler and buncher, in short “RFQ”) and two Penning traps. In the first section, the RFQ serves for retardation, accumulation and buffer-gas cooling of the ions which are delivered from ISOLDE as 60-keV beams. After this, the ions are transferred to a first Penning trap, the preparation Penning trap, where the different species are mass-over-charge selected in some hundreds of milliseconds with mass-resolving powers of about R=m/∆m=105. Furthermore, the ion bunch is centered by buffer-gas collisions and, thus, prepared for the mass measurement in the subsequent precision Penning trap. There, the ion motion is manipulated by a set of radio-frequency (rf) excitations to find the cyclotron frequency. To avoid perturbing Coulomb effects of the charged particles, only a few ions are loaded in the trap at a time. Furthermore, the magnetic field of the 5.9-T solenoid has to be precisely determined. This is achieved by cyclotron-frequency measurements of reference isotopes with well-known masses before and after the measurement of an unknown species. 


Fig.1: Photograph (2012) of the ISOLTRAP precision mass spectrometer at ISOLDE-CERN. The four ion traps are shown as insets.


In order to measure the cyclotron-frequency, the ions are ejected from the Penning trap after each set of rf excitations towards an ion detector through the field gradient of the solenoid. In the decreasing magnetic field, the ions’ radial kinetic energy is converted into longitudinal kinetic energy and the flight time to the detector is shortened when the resonance frequency is hit. Thus, by scanning the excitation frequency and monitoring the time of flight the cyclotron frequency can be determined routinely with a relative mass uncertainty of down to δm/m =1×10-8. The range of accessible nuclei goes down to very low production rates of only 100 ions/s and to nuclei with half-lives below 100ms.

The main limitation for precision Penning-trap mass measurements pursued at ISOLTRAP-ISOLDE and comparable facilities worldwide is the availability of pure samples of the ions of interest. Spallation, fragmentation and fission reactions in the ISOLDE targets produce a large variety of short-lived and stable isotopes. The atoms are ionized and isotope separated via magnetic separators. These separators can reach mass-resolving powers of about m/∆m=5000, which is sufficiently to suppress neighboring isotopes of the same chemical element, but not enough to separate different isobars (nuclides with the same mass number A=Z+N, but different proton number Z and neutron number N). To achieve this prerequisite of precision mass measurements, mass-resolving powers of 104 to 106 are necessary. Until recently, the first of the two ISOLTRAP Penning traps was the device of choice for the separation from contaminating isobars. However, this preparation has to be performed as fast as possible when investigating short-lived species. Furthermore, the species of interest is in most cases the least abundant in the ISOLDE ion beam and its yield can be orders of magnitude below that of the contaminations. To overcome the challenge of “filtering” of the few wanted ions, a fourth ion trap, the multi-reflection time-of-flight mass separator (MR-ToF-MS), has been developed [Wollnik1990,Wolf2012]. As described below, it enhances ISOLTRAP’s mass-purification and measurement capabilities significantly.

The MR-ToF-MS is an ion trap which captures ions with kinetic energies in the keV range and confines them solely with electrostatic fields (fig. 2). This concept differs drastically from both Penning and Paul traps, since it relies on the three-dimensional confinement of “fast” particles, without magnetic or radiofrequency fields. Comparable to an electrostatic storage ring where the ions are periodically focused by static elements, in an MR-ToF-MS this concept is implemented at table-top size. It consists of a coaxial arrangement of two electrostatic ion mirrors between which the charged particles are reflected thousand times forth and back. While the device has a length of only about 0.8m, the folded trajectories can reach a length of kilometers. For medium-mass singly charged ions (A≈100) with a few keV kinetic energy such flights last some tens of milliseconds.

Fig.2: Schematic illustration of the operation principal of the MR-ToF-MS (sectional view) and BNG isobar purification system. Ion bunches from the RFQ, entering the MR-ToF-MS from the left, are captured and separated due to the mass-over-charge ratio of their components. Once the difference in time of flight is sufficient, they are ejected towards a BNG which then selects only the species of interest for further transport to the Penning traps. 


To use the MR-ToF device as a high-resolution mass separator, the electric potential distributions of the ion mirrors have to fulfill certain criteria. The electric fields are optimized such that ions with the same mass have exactly the same revolution period, independent of the spread in kinetic energy, angle and position with respect to the optical axis. Thus, a high mass-resolving power, R=m/∆m=t/2∆t, is achieved, where t is the time of flight and ∆t is the time spread of the single-mass ion bunch at the detector. The mass-resolving powers necessary to separate isobaric ions varies between about R=104 and R=106, depending on the specific isobar combination. For typical flight times of t≈30ms, bunch width of less than ∆t=100ns have to be achieved to reach mass-resolving power of R≈105. In other words, the MR-ToF system needs to be operated as a highly isochronous time-of-flight mass analyzer.

To select the species of interest from the contaminating ions for a subsequent mass measurement, the MR-ToF-MS is followed by a Bradbury-Nielsen gate (BNG) [Bradbury1936,Plass2008]. This is an electrical ion deflector with a very high time resolution, consisting of two sets of multiple, equally spaced thin wires. Neighboring wires belong to different sets and the sets are electrically insulated from each other. In the “transfer” mode, i.e., while the ions of interest are passing the BNG, the two sets and therefore all wires, are on the same electrical potential. In the “deflection” mode, the sets are put on an electric potential of the same value but opposite polarity (typically within ±250V for in 10ns rise/fall time), creating an electric field between the wires which equalizes to nearly 0 in about 1mm before and after the gate. Therefore, ion bunches arriving at the gate with only some tens of nanoseconds delay with respect to each other can be selected or deflected (fig. 2).

Finally, the BNG can suppress about four orders of magnitude of contaminations. Therefore, in combination with the MR-ToF-MS, a highly selective beam purification has been implemented.  Compared to the isobar purification scheme in the Penning trap, this new method is about an order of magnitude faster in time. This enhances the capability of the ISOLTRAP setup to prepare and study contamination-free sample exotic species at the outskirts of the nuclear chart. The expected increase of purified isotopes per time will allow ISOLTRAP to extend the on-going studies even further from stability.


Fig.3: Photograph of the MR-ToF-MS as part of the ISOLTRAP beamline (Image Credit @ CERN)

By using the MR-ToF-MS, the ISOLTRAP collaboration explores new mass regions at and beyond the magic neutron numbers N=50 and N=82. This “terra incognita” is of great importance for nuclear structure in general but also in the context of nuclear astrophysics, more specifically for modeling the composition of neutron stars – among the densest objects known in the cosmos.

Neutron stars are born in a type-II core-collapse supernova, when a massive star of about 3-8 solar masses has fused so much of its material to iron that this “ball” in the centre of the star becomes heavier than the Chandrasekhar limit (about 1.44 solar mass). If the Chandrasekhar limit is overcome, the degeneracy pressure of the electrons cannot compensate the gravitational force anymore and the iron core collapses. In this cataclysmic star death, the in falling material bounces back from the core and is ejected into the interstellar medium, leaving behind a neutron star. Their densities reach from about 104g/cm3 at the outermost regions up to a few times 1014g/cm3 in the core, which is even above the saturation density of nuclear matter.

Modeling the neutron star outer crust, an inhomogeneous layer of some hundred meters in thickness, surprisingly requires only the precise mass of terrestrial stable and radioactive nuclides as relevant input parameters (besides the well-known lattice and electron energy). At this depth, nuclides are supposed to be found in a shell structure with more neutron-excessive nuclides around the magic neutron numbers N=50 and N=82 accumulating in deeper layers (fig. 4).

The first achievement of ISOLTRAP with the newly-integrated MR-ToF separator was the mass of the short-lived N = 52 zinc isotope 82Zn (T1/2=228(10)ms) [Audi2012], predicted to be present in neutron star crusts by several models [Wolf2013]. This measurement was very challenging due to the combination of high contamination yield, mainly of 82Rb, exceeding the yield of the ion of interest by over one order of magnitude, as well as the short half-life and low production rate. The radioactive ion beam from ISOLDE was accumulated in the RFQ after each proton pulse for 100ms and stored for additional 5ms in order to thermalize even the last incoming ions in a helium buffer-gas environment. The ensemble containing the mixture of 82Zn+ and 82Rb+ ions was then injected into the MR-ToF mass separator for a flight time of about 2.5ms, equivalent to 100 revolutions. This was sufficient to separate the two species by multiple signal widths. Subsequently, the 82Rb+ ions were deflected from the beam-line axis by the Bradbury-Nielsen gate and thus removed from the bunch before injection into the next ion trap. This successful purification of an isobaric radioactive beam was decisive for the precision mass measurement.

Before this mass measurement, the mass was only extrapolated by different mass models. These models differed by several 100keV/c2 in mass, but enough to end in different predictions for the composition of the outer crust of neutron stars, the densest objects known in our universe (besides black holes).  

The 82Zn mass was compared with predictions of the three most recent versions of the Brussels-Montreal microscopic mean-field models HFB-19, HFB-20, HFB-21 [Pearson2011]. These models have been developed to derive the binding energy of (unknown) exotic nuclides as well as the equation of state of neutron star matter by using the mass values of the 2149 known masses from the AME2003 and the equations of state of neutron matter coming from realistic calculations with two- and tree-body forces [Goriely2010]. The slight shift to higher mass has initiated a re-evaluation the composition of the outer crust for models HFB-19 and HFB-20 which predicted 82Zn to be found in a depth of about 210m (for a neutron star of 1.4 solar mass and 10 km radius). Finally, with the new mass value, 82Zn is not found in the crust anymore. The reason for this is inaccurate predictions, leading to a different crust composition. These models now predict 79Cu or 78Ni to be the next isotope (fig. 5).

Fig.5: A chart of nuclides showing the neutron-rich regions near the magic numbers N=50 and N=82. Nuclides shown in red are predicted to be present in the outer crust of neutron stars, according to the classic model of Baym et al. [Baym1971] based on experimentally measured masses. The red and black squares correspond to nuclides whose masses are predicted by the mean-field model HFB-19 [Pearson2011]. The red and yellow squares result from the new ISOLTRAP mass measurement of 82Zn. (Image Credit: CERN Courrier).

The composition is of great interest in view of the possible scenario of an r process (rapid neutron capture process) to occur with the decompression of such neutron star matter. A possible site, complementary to the core-collapse supernova r process, is the ejection of neutron-star matter by a merger of two neutron stars (or tidal decompression). This could allow a robust r process to occur as the ejected clump vaporizes into the interstellar medium and undergoes nuclear reactions. This scenario could explain at least part of the total enrichment of the heavy r-process elements in the Galaxy [Goriely2011].




The authors thank J. M. Pearson, Dépt. de Physique, Université de Montreal, Montréal (Québec), Canada, for fruitful discussions in the preparation of the original publication. This work was supported by the German Federal Ministry for Education and Research (BMBF) (Grants No. 06GF9102 and No. 06GF9101I), the Max-Planck Society, the European Union seventh framework through ENSAR (Contract No. 262010), the MASCHE project, the French IN2P3, the Nuclear Astrophysics Virtual Institute (NAVI) of the Helmholtz Association, and the ISOLDE Collaboration.

Note from the editor: Figures 4 and 5 will appear in April’s issue of CERN Courier in the article “Plumbing the depths of Neutron Stars”. The editor of the newsletter would like to thank Christine Sutton and the CERN Courier team for their kind to permission to use these images in this article. 


Lynney2003 D. Lunney, J. M. Pearson, and C. Thibault, Rev. Mod. Phys. 75 (2003) 1021-1082.
Blaum06 K. Blaum, Phys. Rep. 425 (2006) 1-78.
Schweikhard2006 L. Schweikhard and G. Bollen, Int. J. Mass Spectrom. 251(2-3) (2006) 85-312.
Arnould2007 M. Arnould, S. Goriely, and K. Takahashi, Phys. Rep. 450 (2007) 97-213.
Mukherjee2008 M. Mukherjee, D. Beck, K. Blaum, G. Bollen, J. Dilling, S. George, F. Herfurth, A. Herlert, A. Kellerbauer, H. J. Kluge, S. Schwarz, L. Schweikhard, and C. Yazidjian, Eur. Phys. J. A 35 (2008) 1-29.
Wolf2013 R. N. Wolf, D. Beck, K. Blaum, C. Böhm, C. Borgmann, M. Breitenfeldt, N. Chamel, S. Goriely, F. Herfurth, M. Kowalska, S. Kreim, D. Lunney, V. Manea, E. Minaya Ramirez, S. Naimi, D. Neidherr, M. Rosenbusch, L. Schweikhard, J. Stanja, F. Wienholtz, and K. Zuber, Phys. Rev. Lett. 110 (2013) 041101.
Kreim2013 S. Kreim, M. Hempel, D. Lunney, and J. Schaffner-Bielich, Nuclear Masses and Neutron Stars, Int. J. Mass Spectrom. 2013, in print
Baym1971 G. Baym, C. Pethick, and .P. Sutherland, ApJ 170 (1971) 299-317.
Audi2012 G. Audi, F. Kondev, M. Wang, B. Pfeiffer, X. Sun, J. Blachot, and M. MacCormick, Chinese Phys. C 36 (2012) 1157.
Pearson2011 J. M. Pearson, S. Goriely, and N. Chamel, Phys. Rev. C 83 (2011) 065810.
Wollnik1990 H. Wollnik and M. Przewloka, Int. J. Mass Spectrom. Ion Proc. 96 (1990) 267-274
Wolf2012 R. N. Wolf, D. Beck, K. Blaum, C. Böhm, C. Borgmann, M. Breitenfeldt, F. Herfurth, A. Herlert, M. Kowalska, S. Kreim, D. Lunney, S. Naimi, D. Neidherr, M. Rosenbusch, L. Schweikhard, J. Stanja, F. Wienholtz, and K. Zuber, Nucl. Instr. and Meth. in Phys. Res. A 686 (2012) 82-90.
Plass2008 W. R. Plaß, T. Dickel, U. Czok, H. Geissel, M. Petrick, K. Reinheimer, C. Scheidenberger, and M. I.Yavor, Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 4560-4564.
Bradbury1936 N. Bradbury and R. Nielsen, Phys. Rev. 49 (1936) 388-393
Goriely2010 S. Goriely, N. Chamel, and J. M. Pearson, Phys. Rev. C 82 (2010) 035804.
Goriely2011 S. Goriely, N. Chamel, H. T. Janka, and J. M. Pearson, A&A 531 (2011) A78.



Since about a quarter of a century precision mass measurements of short-lived nuclei are performed with the spectrometer ISOLTRAP at CERN. The results are important input parameters for nuclear-structure calculations and tests of fundamental-physics questions. The setup has been improved and extended continuously, repeatedly pioneering techniques for precision mass measurements. The 82Zn measurements were performed by researchers of the University of Greifswald (Germany), the Max-Planck-Institute of Nuclear Physics at Heidelberg (Germany), the CSNSM-IN2P3-CNRS at Orsay (France), the Helmholtz Center for Heavy Ion Research at Darmstadt (Germany), the RIKEN research center (Japan), the Helmholtz-Institute Mainz (Germany), of CERN, as well as from universities at Dresden (Germany) and Leuven (Belgium). The neutron-star calculations were performed at the Université Libre de Bruxelles (Belgium).



Dipl.-Phys. Robert N. Wolf and Prof. Dr. Lutz Schweikhard

Institute of Physics of the Ernst-Moritz-Arndt University Greifswald


Dr. Stephane Goriely and Dr. Nicolas Chamel

Institut d’Astronomie et d’Astrophysique


Dr. David Lunney


Université Paris-Sud


Spokesperson of the ISOLTRAP collaboration

Prof. Dr. Klaus Blaum

Max Planck Institute for Nuclear Physics, Heidelberg, Germany


ISOLTRAP’s local coordinator at CERN

Dr. Susanne Kreim

CERN, bat. 3-1-070, CH-1211 Geneva 23, Switzerland