Discovery of229Rn and the Structure of the Heaviest Rn and Ra Isotopes
from Penning-Trap Mass Measurements
D. Neidherr,1G. Audi,2D. Beck,3K. Blaum,4Ch. Bo ¨hm,1M. Breitenfeldt,5R.B. Cakirli,6,7R.F. Casten,6,8S. George,4
F. Herfurth,3A. Herlert,9A. Kellerbauer,4M. Kowalska,9D. Lunney,2E. Minaya-Ramirez,2S. Naimi,2E. Noah,10
L. Penescu,10M. Rosenbusch,5S. Schwarz,11L. Schweikhard,5and T. Stora10
1Institut fu ¨r Physik, Johannes Gutenberg-Universita ¨t, 55099 Mainz, Germany
2CSNSM-IN2P3-CNRS, Universite ´ de Paris Sud, Orsay, France
3GSI Helmholtzzentrum fu ¨r Schwerionenforschung GmbH Darmstadt, 64291 Darmstadt, Germany
4Max-Planck-Institut fu ¨r Kernphysik, 69117 Heidelberg, Germany
5Institut fu ¨r Physik, Ernst-Moritz-Arndt-Universita ¨t, 17487 Greifswald, Germany
6Institut fu ¨r Kernphysik der Universita ¨t zu Ko ¨ln, 50937 Ko ¨ln, Germany
7Department of Physics, Istanbul University, Istanbul, Turkey
8Wright Nuclear Structure Laboratory, Yale University, New Haven, Connecticut 06520-8124, USA
9Physics Department, CERN, 1211 Geneva 23, Switzerland
10Accelerators and Beams Department, CERN, 1211 Geneva 23, Switzerland
11NSCL, Michigan State University, East Lansing, Michigan 48824-1321, USA
(Received 4 November 2008; published 19 March 2009)
The masses of the neutron-rich radon isotopes223–229Rn have been determined for the first time, using
the ISOLTRAP setup at CERN ISOLDE. In addition, this experiment marks the first discovery of a new
nuclide,229Rn, by Penning-trap mass measurement. The new, high-accuracy data allow a fine examination
of the mass surface, via the valence-nucleon interaction ?Vpn. The results reveal intriguing behavior,
possibly reflecting either a N ¼ 134 subshell closure or an octupolar deformation in this region.
DOI: 10.1103/PhysRevLett.102.112501PACS numbers: 21.10.Dr, 21.30.Fe, 27.90.+b, 32.10.Bi
The nuclear binding energy reflects the net effect of all
aspects of the underlying fundamental forces . Its evo-
lution as a function of proton or neutron number provides
information of great importance for nuclear structure.
Joining the most basic nuclear characteristics of size and
weight is that of shape, usually quantified by deformation,
when a nucleus is considered not to be spherical.
As the capacious valence orbitals of very heavy nuclides
begin to fill with nucleons, increasing varieties of defor-
mation become possible. It is well known  that one of
these degrees of freedom, of importance in the light acti-
nide nuclei, is that of octupole correlations and octupole
deformations. Mo ¨lleret al.  show octupole contributions
to the binding energy and note that the strongest contribu-
tion is centered at222Rn. The role of octupole deforma-
tions, and therefore the study of their impact, is important
in several respects. The existence of the actinides and the
heaviest nuclei is specifically due to quantum effects in the
underlying single particle levels which also manifest in
associated collective correlations. The measurement of
new masses in this region of octupole correlations there-
fore provides a significant constraint on future microscopic
models, for example, in approaches exploiting density
functional theory which is widely used in many areas of
many-body physics . Further, octupole correlations are
known to greatly enhance the sensitivity to nuclear electric
dipole moments: understanding the relations between
masses and such correlations is therefore of importance
in the wider context of tests of fundamental symmetries.
Finally, another important impact of the study of masses in
this region is to be found in the development of micro-
scopic mass models for nucleosynthesis, which are par-
ticularly vulnerable in heavy nuclei.
In this Letter we report on seven new masses of neutron-
rich radon nuclei including the new isotope229Rn, which
marks the first discovery of a nuclide using a Penning-trap
mass spectrometer. The results allow the extraction of
important new values for the valence proton-neutron inter-
action in the A ? 222 mass region, confirming and signifi-
cantly extending a unique anomaly in these interaction
strengths that may indeed be related to octupole degrees
of freedom in this region.
Penning traps are used extensively  to provide accu-
rate mass data of radionuclides with half-lives as short as
only a few milliseconds and production yields of a few
hundred ions per second . The high resolving power has
allowed the separation of isomeric states  and even the
detection of a new isomer, in the case of65Fe .
The measurements reported in this Letter were per-
formed at the double Penning-trap mass spectrometer
ISOLTRAP  located at the on-line isotope separator
facility ISOLDE at CERN. The radon nuclides were pro-
duced by spallation of a 50-g=cm2thick UCXtarget by
pulses of up to 3 ? 1013protons at an energy of 1.4 GeV
from CERN’s Proton Synchrotron Booster accelerator. The
nuclear reaction products diffused from the hot target
through a water-cooled transfer line into a high-efficiency
arc-discharge ion source, designed via a new approach
PRL 102, 112501 (2009)
20 MARCH 2009
? 2009 The American Physical Society
. Off-line studies showed that the ionization efficien-
cies for noble gases are 5–20 times larger than with the
standard ISOLDE plasma ion source (MK7 FEBIAD ),
and reach 60% for radon (estimated through extrapolation
from the He-Xe noble-gas series).
The singly charged ions were accelerated to 30 keVand
separated with the high-resolution mass separator with a
resolving power of the order of 4000 for a first suppression
of any residual isobaric contaminants. The resulting ion
beam was injected into the recently installed radio fre-
quency quadrupole cooler ISCOOL  (working in trans-
mission mode) to improve the ion-beam emittance. The
ions were stopped and bunched in the ISOLTRAP cooler
and buncher in order to prepare them for capture into the
following two Penning traps. The first, cylindrical trap was
used for cooling and isobaric cleaning. The high-precision
mass measurements were carried out in the second, hyper-
boloidal precision Penning trap. Here, the cyclotron fre-
quency ?c¼ qB=ð2?mÞ of the ion was measured via the
time-of-flight (TOF) ion-cyclotron-resonance detection
technique , where q and m are the charge and the
mass of the ion, respectively, and B the magnetic field
strength. In order to calibrate the unknown magnetic field
at the time of the measurement, the cyclotron frequency of
133Cs, having a well-known mass, was measured immedi-
ately before and after that of the ion of interest. The
resolving power, given by the product of the excitation
time texcand the cyclotron frequency ?c, was ?4 ? 105.
For220;223–228Rn, all resonances were taken with an
excitation time of 1.2 s.For229Rn a totaloffourresonances
with excitation times of 100 ms, 600 ms, and twice 1.2 s
were recorded. The resonances were fitted to the theoreti-
cal curve  in order to extract the cyclotron frequency.
The data analysis followed the procedure described in .
In addition to the systematic uncertainty of 8 ? 10?9, a
relative mass-dependent uncertainty of 1:6 ? 10?10ðm ?
mrefÞ=u was added quadratically to the uncertainty of the
mean of the measured frequency ratios.
Although the cold transfer line allows only gaseous
species to reach the ion source, the TOF ion-cyclotron
resonances were carefully checked for the presence of
any contaminations, that might shift the center frequency,
by a count-rate-class analysis . No anomalies were
The experimental results are given in Table I, which
shows for each investigated radon isotope the frequency
ratio of the reference ion133Csþcompared to the ion of
interest, the relative uncertainties, the derived mass ex-
cesses, and the literature values. The well-known mass of
220Rn  was measured as a cross-check and compares
very well with .
229Rn has never been observed, thus only estimates for
the half-life and mass excess were available. The first
confirmation that the new isotope229Rn was produced at
ISOLDE came from a measurement of the half-life at the
ISOLDE spectroscopy station, shown in Fig. 1. This mea-
surement gives a half-life of 12þ1:2
?1:3s for a nuclide with
and their literature values ?lit. Extrapolated mass excess values are marked with *. The reference ion was133Csþwith mð133CsÞ ¼
132:905451932ð23Þ u .
r ¼ ?c;ref=?c
5 ? 10?8
5 ? 10?8
7 ? 10?8
10 ? 10?8
8 ? 10?8
8 ? 10?8
11 ? 10?8
6 ? 10?8
Frequency ratios r ¼ ?c;ref=?c, relative mass uncertainties ?m=m, and mass excesses of the measured radon isotopes ?
FIG. 1 (color online).
separate acquisitions were added and background was sub-
tracted. The fit includes229Rn with an estimated half-life t1=2
ð229RnÞ of 12 s along with its daughter229Fr (50.2 s) and grand-
daughter229Ra (4.0 m). Inset: Radon half-lives with t1=2(229Rn)
nicely following the trend.
Beta-decay curve of229Rn. Data from 10
PRL 102, 112501 (2009)
20 MARCH 2009
mass number 229 then delivered to ISOLTRAP. The fit
accounted for beta counts from the decays of the daughter
When shown together with the known half-lives of the
lighter radon isotopes (see inset of Fig. 1), the trend is
convincing. When the new S2nvalues are plotted (as a
function of N) alongside neighboring isotopic chains (see
Fig. 2) striking continuity is visible that indicates that the
chain of Rn isotopes has been measured. Though the S2nof
229Fr could fall in the same place, we can exclude this
nuclide since we looked for its well-known mass, over
3.5MeVawayfromthe onewe measured,andeasily within
range of the resolving power of the precision trap. Also the
yield attributed to229Rn follows the normal linear trend
(note the logarithmic scale), as visible in Fig. 3 which
229Fr and granddaughter
229Ra (both known).
shows yields for the different radon isotopes in our preci-
sion trap. For228–229Rn these relative yields are compa-
rable to those determined from the ISOLDE tape station
but differ slightly since the measurements were performed
at different times during the run. Target conditions are
Furthermore, the data were carefully analyzed in order to
exclude anyother atomic
(i) Resonances of possible single-ion contaminants, like
229Fr, are too far away from our measured frequency.
(ii) We have verified all possible molecular contaminants
with up to three different atoms constituting the molecule
in the range of ?2 Hz, which is 4 times the FWHM of the
229Rn resonance and is much larger than the resonance
uncertainty of 0.02 Hz. In addition, none of them is likely
tobe producedatISOLDE. (iii)Thehighamplitude(‘‘TOF
effect’’) of the TOF resonance, comparable with ampli-
tudes for the other radon isotopes, shows that the A ¼ 229
beam was very clean after our purification trap.
Atomic masses give nuclear binding and separation en-
ergies, and various combinations (double differences) of
binding energies provide empirical filters that isolate spe-
cific nucleonic interactions . In particular, the ?Vpn
values givethe averageinteraction between thelast protons
and the last neutrons in even-even and even Z–odd N
asthe target ages.
where BZ;Nis the binding energy. Thus the present mea-
surements of the masses of223–229Rn allows us to inves-
tigate the critical valence p-n interaction which affects
many aspects of nuclear structure such as the single parti-
cle energies, magic numbers, collectivity, the onset of
deformation, and the geometrical shapes in atomic nuclei
Empirical ?Vpnvalues in this region are shown in Fig. 4.
The peak for208Pb is well understood in terms of the high
spatial overlaps of protons and neutrons filling orbits just
below their respective closed shells. This figure, however,
shows a second striking feature: the Ra isotopes around
N ¼ 134 exhibit a sharp peak, similar in magnitude to that
inlead. This was already described in as the Ra puzzle,
and it was speculated that if the mass data are correct, the
p-n interactions might sense a softness to octupole defor-
mation that is known in this region. However, the Ra peak
is so strong that it is important to first confirm it
Although our result for220Rn has somewhat larger un-
certainty (10 keV) than previous results, it clearly confirms
the large value of ?Vpnfor223Ra in Fig. 4 (top), removing
any suspicion that it was due to an incorrectly measured Rn
mass. The peak near N ? 135 is the most extreme excur-
sion of an individual ?Vpnvalue from local trends (except
circles show the yield measured in the ISOLTRAP precision
trap during our measurements. The open circles represent the
yield obtained at the ISOLDE tape station. The comparison for
228Rn gives the ISOLTRAP efficiency of about 0.2%, which is in
the same range as that for229Rn.
Yields of the different radon isotopes. The closed
FIG. 2 (color online).
for Pb-Pa isotopes with N ¼ 123–144 . New radon data are
represented by filled circles. A deviation from the parallel
decrease of S2nin neighboring chains is visible.
Two-neutron separation energies (S2n)
PRL 102, 112501 (2009)
20 MARCH 2009
for doubly magic nuclei and N ¼ Z nuclei whose origins
lie in single particle overlaps and the Wigner force, re-
spectively). Even more importantly, the new results sig-
nificantly extend, and, indeed, complete, the mapping of a
striking anomaly revealing that this point is not an isolated
singularity at N ¼ 135 but rather part of a very well
developed peak in ?Vpn(even-odd) which terminates at
N ¼ 139 and which is unique in the nuclear chart. That it
appears in a region that is also characterized by the best
known example of octupole correlations and of a possible
neutron subshell effect at N ¼ 134  makes it intriguing
and important. It is particularly unexpected that an anom-
aly of this magnitude would occur in a region of collective
structure, where smooth behavior is anticipated due to
many-particle correlations. An understanding of structure
in this region, including octupole deformations, must ac-
count for this anomaly.
In contrast, the ?Vpn(even-even) values for Ra are
consistent with the other nuclei in this region as seen in
Fig. 4 (bottom). One interesting feature is that the ?Vpn
values for neighboring elements have similar behavior but
are systematically shifted to the right for each successive
Z. A similar pattern is visible in the rare-earth-metal region
 except that here it appears quite early in the p-n major
shells while, in the rare-earth-metal region, it is near mid-
shell. Whether this behavior is coincidental, or reflects a
similar microscopic origin, remains to be studied.
In summary, we have directly determined the masses of
223–229Rn by precision mass spectrometry for the first time.
We have also identified for the first time229Rn and mea-
sured its half-life. These new mass measurements provide
significant extensions of known masses in this region,
which are important for understanding the binding of the
heavy nuclei, which have an impact on nucleosynthesis in
the actinide region, and which provide new constraints on
our study, the Ra anomaly in ?Vpnwas restricted to two
?Vpnvalues, and therefore there was some doubt about its
existence. The present mass measurements prove that it
does exist and that it is part of a systematic deviation from
thegeneral trendof ?Vpnvalues which constitutesthe most
significant anomaly in p-n interaction strengths anywhere
in the nuclear chart (except for N ¼ Z nuclei)—and there-
fore, the most significant anomaly in collective nuclei. Our
results extend the data to show the full behavior of this
anomaly. As this occurs in a region of octupole correla-
tions, these results provide a new signature of such corre-
lations, and a new aspect of their influence on nuclear
This work was supported by the German Federal
(06GF186I, 06MZ215), German DFG under Grant
No. Ko. 142/112-1, Helmholtz Association for National
Research Centers (VH-NG-037), U.S. Department of
Energy (DE-FG02-91ER-40609), the French IN2P3,
Turkish Atomic Energy Authority (TAEK) OUK120100-
4, and the EU FP6 programme (MEIF-CT-2006-042114
and EURISOL DS project/515768 RIDS). We are grateful
to the other members of the ISOLDE technical group for
 D. Lunney et al., Rev. Mod. Phys. 75, 1021 (2003).
 P.A. Butler and W. Nazarewicz, Rev. Mod. Phys. 68, 349
 P. Mo ¨ller, J.R. Nix, and W.J. Swiatecki, At. Data Nucl.
Data Tables 59, 185 (1995).
 M. Bender, G.F. Bertsch, and P.-H. Heenen, Phys. Rev. C
78, 054312 (2008).
 Special issue on Ultra-Accurate Mass Spectrometry and
Related Topics, edited by L. Schweikhard and G. Bollen
[Int. J. Mass Spectrom. 251, 85 (2006)].
 K. Blaum, Phys. Rep. 425, 1 (2006).
 C. Weber et al., Phys. Lett. A 347, 81 (2005).
 M. Block et al., Phys. Rev. Lett. 100, 132501 (2008).
 M. Mukherjee et al., Eur. Phys. J. A 35, 1 (2008).
FIG. 4 (color online).
for the208Pb region (from Ref. ), with the addition of the
new Ra ?Vpnvalues (solid circles). The ?Vpnvalues involving
220Rn, whose mass is confirmed in the present study, are
encircled. Bottom: As top but for even-even nuclei.
Top: Empirical ?Vpn(even-odd) values
PRL 102, 112501 (2009)
20 MARCH 2009
 L. Penescu et al., Nucl. Instrum. Methods Phys. Res., Sect. Download full-text
B 266, 4415 (2008).
 U.C. Bergmann et al., Nucl. Instrum. Methods Phys. Res.,
Sect. B 204, 220 (2003).
 H. Fra ˚nberg et al., Nucl. Instrum. Methods Phys. Res.,
Sect. B 266, 4502 (2008).
 M. Ko ¨nig et al., Int. J. Mass Spectrom. Ion Processes 142,
 A. Kellerbauer et al., Eur. Phys. J. D 22, 53 (2003).
 G. Audi, A.H. Wapstra, and C. Thibault, Nucl. Phys.
A729, 337 (2003).
 B. Grennberg and A. Rytz, Metrologia 7, 65 (1971).
 A.S. Jensen et al., Nucl. Phys. A431, 393 (1984).
 J.-Y. Zhang et al., in Proceedings of the International
Structure, Cocoyoc, Mexico, 1988, Book of Abstracts
(unpublished), p. 109.
 J.-Y. Zhang, R.F. Casten, and D.S. Brenner, Phys. Lett. B
227, 1 (1989).
 I. Talmi, Rev. Mod. Phys. 34, 704 (1962).
 P. Federman and S. Pittel, Phys. Lett. 69B, 385 (1977);
77B, 29 (1978).
 K. Heyde et al., Phys. Lett. B 155, 303 (1985).
 D.S. Brenner et al., Phys. Rev. C 73, 034315 (2006).
 Y. Oktem et al., Phys. Rev. C 74, 027304 (2006).
PRL 102, 112501 (2009)
20 MARCH 2009