Professional Documents
Culture Documents
Atomic Physics at Accelerators - Mass Spectrometry
Atomic Physics at Accelerators - Mass Spectrometry
MASS SPECTROMETRY
Edited by
DAVID LUNNEY
CSNSM-IN2P3-Universite de Paris Sud, Orsay, France
GEORGES AUDI
CSNSM-IN2P3-Unil'ersite de Paris Sud, Orsay, France
and
H.-JURGEN KLUGE
GSI, Darmstadt, Germany
Tutorial on aspects of atomic masses (in honor of the 78th birthday of Aaldert
Wapstra)
D. LUNNEY
About three years ago, Juergen Kluge had the idea of proposing a series of con-
ferences dedicated to what we loosely call "atomic" physics, but performed at
accelerator facilities. Hence the acronym: APAC. Though much of this physics
targets the nucleus, the surrounding atomic system is used to access this recalci-
trant, quantum, many-bodied system that has defied an exact theoretical description
since its discovery, long ago.
Thanks to the inherent precision of lasers, we are able to probe the very sub-
tle effects of the nucleus on atomic energy levels as manifested by the hyperfine
spectrum. Optical spectroscopy, yielding information on nuclear moments and the
charge radius, was the central theme for the first in the series of APAC Euro-
conferences*. Held in Budenheim (Germany), the proceedings of APAC99 were
published in Hyper.tine Interactions (Volume 127, 2000). This second meeting in
the series, held in Cargese (France), was devoted to the field of mass spectrometry
but more specifically: atomic mass measurements. With a profound influence not
only on nuclear but also on atomic physics, such measurements form the heart of
several research programs worldwide, as these proceedings attest. Completing the
Euroconference triumvirate will be APAC2001, to be held in Aarhus (Denmark),
dedicated to studying the interplay between the nucleus and the various atomic
charge states.
The field of mass measurements is as old as nuclear physics itself, even pre-
dating the discovery of the neutron. Since that time, measurement precision has
constantly improved (roughly at a rate of one order of magnitude per decade),
revealing new detail of the mass landscape that guides us in our quest to understand
the nuclear interaction. The application of mass measurements using direct tech-
niques of mass spectrometry at accelerators was pioneered by the atomic masses
group at the Centre de Spectrometrie Nucleaire et de Spectrometrie de Masse
(CSNSM) in Orsay. For this reason, Juergen Kluge asked our group to organize
APAC2000.
Unlike most other properties in physics, the mass is subject to a special treat-
ment known as the Atomic Mass Evaluation. This is necessary due to the very
definition of the mass unit (one twelfth the mass of 12C) and due to the myr-
iad interconnections of masses via reactions and decays. The Grand Inquisitor of
atomic masses, Aaldert Wapstra, has tirelessly devoted himself to this rigorous
The village of Cargese with its Greek (left-handed) and Latin (right-handed) churches. Note the
symmetry breaking.
Aaldert Wapstra, at the conclusion of the tutorial honoring his seminal contribution to the
Atomic Mass Evaluation, recalling some of his memories.
Annette Paul (PTB, Braunschweig) describing the old kilogram (inset) and its proposed re-
placement - a high purity Si ball comprised of a known number of atoms. Annette's talk on
the AVOGADRO project was voted the best of the meeting.
Perhaps the most unique aspect of APAC2000 was the interface between nu-
clear and atomic physics via their respective binding energies. The accuracy of
measurements using trapped, highly charged ions is now such that state-of-the-art
QED calculations must be pushed to higher order.
Nuclear theory made up the balance of the program, notably for its necessary
and important role in stellar nucleosynthesis modeling.
In constructing the APAC2000 program, I relied chiefly on my perception of
how the field of mass measurements ought to be captured for posterity at this
moment of what seems to me, a zenith. Of course, I also relied on the council
of Juergen Kluge for the atomic physics component and general aspects of Eu-
roconference organization. I was in frequent consultation with my colleagues in
Orsay: Georges Audi, Michel de Saint Simon, Catherine Thibault, Hubert Doubre
and Dominique Guillemaud-Mueller. The International Advisory Committee was
solicited for prioritizing the speaker lists and I am grateful to those who responded,
especially to Ernst Roeckl, an ex-officio member who made numerous suggestions
to help shape the tone of the meeting. I would like the acknowledge the role played
by Catherine Thibault in the management of financial matters - an activity that
lay dormant up until the meeting at which point it exploded into frenzied activity
during the week, causing her a significant loss of sleep. (Not enough however to
4 D.LUNNEY
The "official" conference photo, taken by the organizer during his own talk. Note the surprised
look of the audience.
The conference shuttle bus, packed beyond the space-charge limit of passenger density.
The conference site - and perfect occasion for an ISOLTRAP collaboration meeting.
Girolata - midpoint of our excursion and accessible only by sea. (One of us enjoyed it so much he
missed the boat!)
better from the shore than on a boat. Rare in Corsica is rainfall of which we had
the dubious pleasure during our stay on the parched (sometimes even burning!)
isle. I think we all enjoyed the spectacular Calanques de Pinata and the pays des
Quatres Tours.
In summary, my own expectations (not to mention fears) were fulfilled and even,
surpassed. But while personal satisfaction is one thing, the greatest pleasure comes
from the reactions of the participants, a large number of whom have shared with
6 D.LUNNEY
me their appreciation of the meeting. Perhaps one regret: while the conference
barbecue was excellent, we did not resolve the mysterious prophecy, recalled by
Ernst Roeckl, of " ... Charles, nearly rescuing the mechoui from a disaster. .. " But
happily, there are still mysteries in our field.
Hyper(ine Interactions 132: 7-34, 200 l. 7
© 2001 Kluwer Academic Publishers.
GEORGES AUm
Centre de Spectrometrie Nuc/eaire et de Spectrometrie de Masse. CSNSM, IN2P3-CNRS and UPS,
Bdtiment 108, F-91405 Orsay Campus, France
Abstract. The ensemble of experimental data on the 2R30 nuclides which have been observed since
the beginning of Nuclear Physics are bcing evaluated, according to their nature, by different methods
and by different groups. The two 'horizontal" evaluations in which I am involved: the Atomic Mass
Evaluation AME and the NUBASE evaluation bclong to the class of 'static' nuclear data. In this
tutorial lecture I will explain and discuss in detail the philosophy, the strategies and the procedures
used in the evaluation of atomic masses.
Key words: atomic mass evaluation, mass extrapolations. mass spectrometry, nuclear binding en-
ergy, nuclear data.
Nuclear Physics started a little bit more than 100 years ago with the discoveries
of Henri Becquerel and Pierre and Marie Curie. First, it was a science of curiosity
exhibiting phenomena unusual for that time. It is not until the late thirties, weH after
the discovery of artificial radioactivity by Frederic and Irene loliot-Curie, that the
research in that domain tended to accelerate drasticaJly and that Nuclear Physics
became more and more a quantitative science.
Since then, scientists have accumulated a huge amount of data on a large num-
ber of nuclides. Today there are some 2830 variations on the combination of pro-
tons and neutrons that have been observed. Although this number seems large,
specially compared to the 6000 to 7000 that are predicted to exist, one should be
aware that the numbers of protons and neutrons constituting a nuclide are not really
independent. Their special correlation form a relatively narrow band around a line
called the bottom of the vaHey of stability. In Figure 1 this is illustrated for the
known masses (colored ones) across the chart of nuclides. In other words, nuclear
data put almost no constraint in isospin on nuclear models. From there follows the
tendency of nuclear physicists to study exotic nuclides.
Sometimes remeasurement of the same physical quantity improved a previous
result; sometimes it entered in conflict with it. The interest of the physicist has also
evolved with time: the quantities considered varied importantly, scanning all sort
of data from cross sections to masses, from half-lives to magnetic moments, from
radii to superdeformed bands.
* This tutorial lecture is dedicated to A. H. Wapstra on the occasion of his nth anniversary.
8 G.AUDI
l l l NEUTRONS l l l l l :
~&r&&····1··(jfr······y······~······1·······!······1··· .... ~ ...... ~....... ~.~.:, .
90
. :z: . : . . . .
80 ··f······~··o·t·····+·····t·····+·····~·····+····
··[·····+a+····+···+·····L . ..L. :
:t:Ic:=t::I .··"'[·,;. ;.:
70
, 'CJ:::' , .
;;:.;:.:.;;;;;.;;:.;:.:.;;;;;.;;;f.;*'·
30 .. !..... o 1 S u ( 12
o 12 S u ( 60
o 60S u (200
0 200:S u
[J M.sse Extr.polee
o M. sse 1ncorroue
Figure 1. Chart of nuclides for the precision on masses. Only the known masses are colored. exhibit-
ing crudely the narrowness of the valley of our knowledge in this immense landscape. Would these
1970 known masses been scattered around in the (N. Z) plane. our understanding of the nucleus
would have been completely changed.
Certainly, one single experiment, for example a nuclear reaction study, can yield
data for both 'static' and 'dynamic' properties.
It is out of the scope of the present lecture to cover all aspects of nuclear prop-
erties and nuclear data. The fine structure of 'static' nuclear data will be shortly
described and the authors of the various evaluations presented. Then I will center
this lecture on the two 'horizontal' evaluations in which I am involved: the atomic
mass evaluation AME and the NUBASE evaluation, both being strongly related,
particularly when considering isomers.
structures N ~
~<t:<t:<t:
~I~
and
deca s
, .
masse II I
JJ
Figure 2. Schematic representation of all the available 'static' nuclear data (structure, decay,
mass, radius, moments, ... ). Each nuclide is represented as a building with its ground-state at
the ground floor. The mass evaluation is represented on the ground floor. across all buildings.
It includes also data for upper levels if they represent an energy relation to another nuclide.
like a foot-bridge between two buildings that will allow to derive the level difference between
their ground floors.
10 G. AUm
Physics journal for A ~ 44). At the same time the computer files of the evaluation
(the ENSDF: 'Evaluated Nuclear Structure Data Files') are made available at the
NNDC-Brookhaven [1]. In this evaluation network, most ofthe 'static' nuclear data
are being considered.
EExc
Figure 3. Truncated distribution of probability when there is a strong indication about
ordering of ground-state and isomer.
There might be other reasons for 'horizontal' evaluations. The splitting of data
among a large number of evaluators - like in the NSDD network described above
- does not always allow having a completely consistent treatment of a given nu-
clear property through the chart of nuclides. In addition, some quantities may fall
at the border of the main interest of such a network. This is the reason why a
few 'horizontal' compilations or evaluations have been conducted for the benefit
of the whole community. For example, one can quote the work of Otten [7] for
isotope shift and hyperfine structure of spectral lines and the deduced radii, spins
and moments of nuclides in their ground-state and long-lived isomeric states. An
evaluation of isotope shifts has been published also by Aufmuth and coworkers [8],
and Raghavan [9] gave a table of nuclear moments, updated recently by Stone [10].
More recent tables for nuclidic radii were published by Angeli [II] in 1991 and
by Nadjakov et ai. [12J in 1994. Two other 'horizontal' evaluations are worth
mentioning. One is the evaluation of isotopic abundances, by Holden [13]. The
second one is the evaluation of Raman and coworkers [14] for the energy E 2+ and
the reduced electric quadrupole transition probability BCE2) of the first excited 2+
state in even-even nuclides.
The atomic mass evaluation is particular when compared to the other evaluations
of data reviewed above, in that there are almost no absolute determinations of
masses. All mass determinations are relative measurements. Each experimental
datum sets a relation in energy among two (rarely more) nuclides. It can therefore
be represented by one link among these two nuclides. The ensemble of these links
THE EVALUATION OF ATOMIC MASSES 13
generates a highly entangled network. This is the reason why, as I mentioned earlier
(cf. Section 2.2), a 'horizontal' evaluation is essential.
I will not enter in details in the different types of mass experiments, since there
will be another lecture devoted to this subject [15]. Nevertheless, I need to sketch
the various classes of mass measurements to outline how they enter the evaluation
of masses and how they interfere with each other.
Generally a mass measurement can be obtained either by establishing an energy
relation between the mass we want to determine and a well known mass, this energy
relation is then expressed in electron-volts (eV); or obtained as an inertial mass
from its movement characteristics in an electro-magnetic field, the mass is then
expressed in 'unified atomic mass' (u) (or its sub-unit, {w), since it is obtained as
a ratio of masses (cf. Section 3.1.3).
e
The mass unit is defined, since 1960, by 1 u = M 2 C) /12, one twelfth of the
mass of one free atom of carbon-12 in its atomic and nuclear ground-states. Before
1960, as Wapstra once told me, there were two mass units: the physical one 16 0/16,
and the chemical one which considered one sixteenth of the average mass of a stan-
dard mixture of the three stable isotopes of oxygen. Physicists could not convince
the chemists to drop their unit; "The change would mean millions of dollars in the
sale of all chemical substances", said the chemists, which is indeed true! Joseph
H. E. Mattauch, the American chemist Truman P. Kohman and Aaldert H. Wapstra
then calculated that, if 12C /12 was chosen, the change would be ten times smaller
for them, and in opposite direction! That led to unification. 'u' stands therefore,
officially, for 'unified mass unit' !
The choice of the volt in the energy unit (the electronvolt) is not evident. In
the AME, it appeared that not the international volt V should be used, but the volt
V* [16] as maintained in standard laboratories. The latter is defined by adopting
a value for the constant (2e / h) in the relation between frequency and voltage in
the Josephson effect. This choice results from an analysis [17] showing that all
precision measurements of reaction and decay energies are calibrated in such a way
that they can be more accurately expressed in the standard volt. Also, the precision
of the conversion factor between mass units and standard volts V* is more accurate
than that between it and international volts V:
(1)
(2)
where q = Qr - Ma + M b ·
A nuclear reaction usually deals with stable or very-long-lived target 'A' and
projectile 'a', allowing only to determine the mass of a residual nuclide 'B' close to
stability. Nowadays with the availability of radioactive beams, interest in reaction
energy experiments could be revived.
It is worth mentioning in this category the very high accuracies attainable with
(n,y) and (p,y) reactions. They playa key-role in providing many of the most
accurate mass differences, and thus help building the 'backbone' of masses along
the valley of ,B-stability.
Also very accurate are the self-calibrated reaction energy measurements using
spectrometers. When measuring the difference in energy between the spectral lines
corresponding to reactions A(a,b)B and C(a,b)D with the same spectrometer set-
tings [18 J one can reach accuracies better than 100 e V. Here the measurement can
be represented by a linear combination of the masses of four nuclides:
(3)
The most precise reaction energy is the one that determined the mass of the
neutron from the neutron-capture energy of I H at the ILL [J 9]. The I H(n,y )2H
established a relation between the masses of the neutron, of I H and of the deuteron
with the incredible precision of 0.4 eY.
positive, i.e., these particular reactions are exothermic. For the A(B-)B, A(a)B or
A(p)B desintegrations, one can write respectively:
MA - M B , (4)
MA - MB -Ma, (5)
MA - Ms - Mp. (6)
These measurements are very important because they allow deriving masses of
unstable or very unstable nuclides. This is more especially the case for the proton
decay of nuclides at the drip-line, in the medium-A region [20].
a-decays have permitted to determine the masses of the heavy nuclides. More-
over, the time coincidence of a lines in a decaying chain allows very clear identifi-
cation of the heaviest ones.
of other data). If there are no previous results, comparison could be done with
estimates from extrapolations, exploiting the above-mentioned regularity of the
mass surface (cf. Section 3.5), or to estimates from mass models or mass formulae.
Finally, the evaluator might have to establish a dialog with the authors of the
work, asking for complementary information when necessary, or suggesting differ-
ent analyses, or suggesting new measurements.
The new data can now enter the data-file as one line. For example, for the
electron capture of 205Pb, the evaluator enters:
205 890816000cl B 78Pe08 41.4 1.1 205Pb(e)205Tl 0.525 0.008 LM,
where besides a 14 digits ID-number, there is a flag (as described in [2, p. 451]),
here 'B', then the NSR reference-code [291 for the paper '78Pe08' where the data
appeared, the value for the Q of the reaction with its error bar (41.4 ± 1.1 keY),
and the reaction equation, where 'e' stands for electron-capture. The information
in the last columns says that this datum has been derived from the intensity ratio
(0.525 ± 0.008) of the Land M lines in electron capture. The evaluator can add as
many comment lines as necessary, following this data line, for other information
he judges useful for exchange with his fellow evaluator. Some of these comments,
useful for the user of the mass tables, will appear in the AME publication.
In this section, we shall first see how the network of data is built, then how the
system of data can be reduced. In the third and fourth subsections, I shall de-
scribe shortly the least-squares method used in the AME and the computer program
that will decode data and calculate the adjusted masses. A fifth part will develop
the very important concept of 'Flow-of-Information' matrix. Finally, I shall ex-
plain how checking the consistency of data (or of sub-group of data) can help the
evaluator in his judgment.
We have seen in Section 3.1 that all mass measurements are RELATIVE measure-
ments. Each experimental piece of data can be represented by a link between
two, sometimes three, and more seldomly four nuclides. As mentioned earlier,
assembling these links produces an extremely entangled network. A part of this
network can be seen in Figure 4. One notices immediately that there are two types
of symbols, the small and the large ones. The small ones represent the so-called
SECONDARY nuclides; while the nuclides with large symbols are called PRIMARY.
Secondary nuclides are represented by full small circles if their mass is determined
experimentally, and by empty ones if estimated from trends in systematics. Sec-
ondary nuclides are connected by SECONDARY data, represented by dashed lines.
A chain of dashed lines is at one end free, and at the other end connected to
l8 G.AUDI
r
132 136 140 144 148 152 156 160 164 168
40 (e)
36 l .,\
32 r
~
32-
t 28 ~ 28-
N-Z
24 - 24-
20 - 29-
16 -
· . ~
: ~ : ~
~ ~ ~
: ; :
~ 16-
0 :0: 6
'0 ; : :
It . : ' i
L
·' . .
I II 00 " •
·
12 - /2-
II
. .
• II • II
8-
132 /36 140 144 148 152 156 160 164 168
A
Figure 4. Diagram of connections for the experimental data. Each symbol represents one
nuclide and each line represents one piece of data connecting two nuclides. When a nuclide is
connected to carbon-I 2 (often the case for mass spectrometry), it is represented by a square
symbol.
one unique* primary nuclide (large symbol). This representation means that all
secondary nuclides are determined uniquely by the chain of secondary connections
going down to a primary nuclide. The latter are multiply determined and enter thus
the entangled canvas. They are inter-connected by PRIMARY data, represented by
full lines.
We see immediately from Figure 4 that the mass of a primary nuclide cannot be
determined straightforwardly. One may think of making an average of the values
obtained from all links, but such a recipe is erroneous because the other nuclides on
which these links are built are themselves inter-connected, thus not independent. In
other words these PRIMARY data, connecting the primary nuclides, are correlated,
and the correlation coefficients are to be taken into account.
Caveat: the word primary used for these nuclides and for the data connecting
them does not mean that they are more important than the others, but only that
they are subject to the special treatment below. The labels primary and secondary
are not intrinsic properties of data or masses. They may change in any direction if
other information becomes available.
Xn =
fx2
yQ=l (7)
resulting in the calculation of the pre-average is greater than 2.5, the (internal) error
O"j in the average is multiplied by the Birge ratio (O"e = O"j X Xn). The quantity O"e
is often called the 'external' error. However, this treatment is not used in the very
rare cases where the errors in the values to be averaged differ too much from one
another, since the assigned errors lose any significance (three cases in AME'93).
We here adopt an arithmetic average and the dispersion of values as error, which is
equivalent to assigning to each of these conflicting data the same error.
In AME'93, 28% of the 929 cases in the pre-average had values of Xn beyond
unity, 4.5% beyond two, 0.7% beyond 3 and only one case beyond 4, giving a very
satisfactory distribution overall. With the choice above of a threshold of X~ = 2.5
20 G. AUm
for the Birge ratio, only 1.5% of the cases are concerned by the multiplication
by Xn. As a matter of fact, in a complex system like the one here, many values
of Xn beyond 1 or 2 are expected to exist, and if errors were multiplied by Xn in
all these cases, the x 2 -test on the total adjustment would have been invalidated.
This explains the choice made in the AME of a rather high threshold (X~ = 2.5),
compared, e.g., to X~ = 2 recommended by Woods and Munster [30] or, even,
X~ = 1 used in a different context by the Particle Data Group [31], for departing
from the rule of internal error of the weighted average (see also [32]).
Another method to increase the meaning of the final X 2 is to exclude data with
weights at least a factor 10 less than other data, or combinations of other data giving
the same result. They are still kept in the list of input data but labelled accordingly;
comparison with the output values allows to check that this procedure did not have
unwanted consequences.
The system of data is also greatly reduced by replacing data with isomers
by an equivalent datum for the ground-state, if a y-ray energy measurement is
available from the NNDC (cf. Section 2.3). Excitation energies from such y-ray
measurements are normally far more precise than reaction energy measurements.
Typically, we start from a set of 6000-7000 experimental data connecting some
3000 nuclides. After pre-averaging, taking out the data with very poor accuracy
and separating the secondary data, we are left with a system of 1500 primary data
for 800 nuclides.
(e.g., Equation (2) or (3)) or, in matrix notation, K being the matrix of coefficients:
Kim) = Iq). We see immediately that matrix K is essentially filled with zero
values, e.g., for reaction A(a,b)B, Equation (2) shows that the corresponding line
of K has only two non-zero elements. We define the diagonal weight matrix W by
its elements w; = l/(dqjdqi).
The solution of the least-squares method leads to a very simple construction:
(9)
the NORMAL matrix A = tKWK is a square matrix of order M, positive-definite,
symmetric and regular and hence invertible [33]. Thus the vertor 1m) for the ad-
justed masses is:
1m) = A-I tKWlq) or 1m) = Rlq). (10)
THE EVALUATION OF ATOMIC MASSES 21
3.3.5. Flow-of-information
The flow-of-infOllllation matrix is a powerful method that allows to trace back, in
the least-squares method, the contribution of each individual piece of data to each
of the parameters (here the atomic masses). The AME uses this method since 1993.
The flow-of-information matrix F is defined as follows: K, the matrix of coef-
ficients, is a rectangular (Q, M) matrix, the transpose of the response matrix tR is
also a (Q, M) rectangular one. The (i, A) element of F is defined as the product
of the corresponding elements of tR and of K. In reference [34] it is demonstrated
that such an element represents the 'influence' of datum i on parameter (mass) rnA'
A column of F thus represents all the contributions brought by all data to a given
mass rnA' and a line of F represents all the influences given by a single piece of
data. The sum of influences along a line is the 'significance' of that datum. It has
also been proven [34] that the influences and significances have all the expected
properties, namely that the sum of all the influences on a given mass (along a
column) is unity, that the significance of a datum is always less than unity and
that it always decreases when new data are added. The significance defined in this
way is exactly the quantity obtained by squaring the ratio of the uncertainty on the
adjusted value over that on the input one, which is the recipe that was used before
the discovery of the F matrix to calculate the relative importance of data.
A simple interpretation of influences and significances can be obtained in cal-
culating, from the adjusted masses and Equation (8), the adjusted data:
The ith diagonal element of KR represents then the contribution of datum i to the
determination of qi (same datum): this quantity is exactly what is called above the
significance of datum i. This ith diagonal element of KR is the sum of the products
of line i of K and column i of R. The individual terms in this sum are then nothing
else than the influences defined above.
The flow-of-information matrix F, provides thus insight on how the information
from datum i flows into each of the masses rnA'
equal to Q-M, the number of degrees offreedom, with a precision of J2(Q - M).
One can define as above the NORMALIZED CHI, Xn (or 'consistency factor' or
'Birge ratio' ): Xn = J X 2/( Q - M) for which the expected value is 1 ±
I/J2(Q - M).
For our current AME example of 1500 equations with 800 parameters, i.e., 700
degrees of freedom, one gets a theoretical Xn = 1 ± 0.027. The value was 1.062 in
AME'83 for Q - M = 760 degrees of freedom, 1.176 in AME'93 for Q - M =
635, and 1.169 in the AME' 95 update for 622 degrees of freedom. This means that,
on average, the errors in the input values entering the AME'95 were underestimated
by 17%, an acceptable result. The distribution of the Vi'S is also quite acceptable,
with, in AME'93, 17% of the cases beyond unity, 2.6% beyond two, 0.4% beyond 3
and only one case (0.07%) beyond 4.
Another quantity of interest for the evaluator is the PARTIAL CONSISTENCY
FACTOR, XI;', defined for a (homogeneous) group of p data as:
Q I I'
X11I' = - - - - ~V2
Q_Mp~i'
(13)
1=1
Of course, the definition is such that XI;' reduces to Xn if the sum is taken over
all the input data. One can consider, for example. the two main classes of data: in
THE EVALUATION OF ATOMIC MASSES 23
AME'95, for energy measurements xl: = 1.169, and 1.170 for mass spectrometry
data, showing that the two types of input data were equally responsible for the un-
derestimated error of 17% mentioned above. One can also consider groups of data
related to a given laboratory and with a given method of measurement (in AME' 95
there were 164 groups of data) and examine the xl: of each of them. A high value
of xl: might be a warning on the validity of the considered group of data within the
reported errors. In general, in the AME such a situation is extremely rare, because
deviating data are cured before entering the 'machinery' of the adjustment, at the
stage of the evaluation itself (see Section 3.2).
X + 0.64· (a - b) (4)
a2 = (2)
1- If (a - b) 2 + abo (15)
In the appendix of [3] one can find the demonstration and discussion of Equa-
tions (4) and OS).
-----l------ o
.. L\.n p
Nodd Zodd
L\.nn
L\.pp
N even Zodd
Nodd Zeven
N even Zeven
Figure 5. The surface of masses is split into four sheets. This scheme represents the pairing
energies responsible for this splitting. The zero energy surface is a purely hypothetical one for
no pairing at all among the last nucleons.
* Systematic errors are those due to intrumental drifts or intrumental fluctuations, that are beyond
control and are not accounted for in the error budget. They might show up in the calibration process,
or when the measurement is repeated under different experimental conditions. The experimentalist
adds then quadratically a systematic error to the statistical and the calibration ones, in such a way as
to have consistency of his data. If not completely accounted for or not seen in that experiment, they
can still be observed by the mass evaluators when considering the mass adjustment as a whole.
26 G.AUDI
100 105 110 115 no 125
11 --r 22
'::r~
20 2"
S- IMp,
,,.,,
"
~
I" 111(b ~
11I1t~
18
~
~ ...w •
",,'"
"'Tn 16
16
1 1' Yb
m ,,"
U
I"
n,,£, """
JeII Ho
11 "'0, 11
10 /U
8
100 /05 IIQ 1/5 120 125
Neutron number N
Figure 6. Two-neutron separation energies as a function of N (from AME'93, p. 166). Solid
points and error bars represent experimental values, open circles represent masses estimated
from 'trends in systematics'. Replacing some of the experimental data by values estimated
from these trends, changes the mass surface from the dotted to the full lines. The use of a
'derivative' function adds to the confusion of the dotted lines, since two points are changed
if one mass is displaced. Moreover, in this region there are many IX links resulting in large
propagation of errors.
70 75 /JO &5 95 100 lOS 110 115
76
"'w
70
6Ii
62
Nelltroll Nllmber N
Figure 7. Differences. in the rare-earth region. between the masses and the values predicted
by the model of Duflo and Zuker [36]. Open circles represent values estimated from systematic
trends; points are for experimental values.
THE EVALUATION OF ATOMIC MASSES 27
between the mass surface and this function, while displaying reliably the structure
of the former, will vary much less rapidly, improving thus its observation.
A first and simple approach is the semi-empirical liquid drop formula of Bethe
and Weizsacker. The concept of the liquid drop mass formula was defined by
Weizsacker [37] in 1935 and fine-tuned by Bethe and Bacher [38] in 1936. The
binding energy of the nucleus comprises only a volume energy term, a surface
one, an asymmetry term, and the Coulomb energy contribution for the repulsion
amongst protons. The total mass is thus:
(N - Z)2 3 e 2Z 2
M(N, Z) = NMn + ZM H - aA + fJ + yA2/3 + ---1/-3' (16)
A 5 roA
where A = N + Z, is the atomic weight, roA 1/3 the nuclear radius, Mn and MH the
masses of the neutron and of the hydrogen atom. The constants a, fJ, y and ro were
determined empirically by Bethe and Bacher: a = 13.86 MeV, fJ = 19.5 MeV,
y = 13.2 MeV and ro = 1.48.10- 15 m (then ~e2lro = 0.58 MeV). The formula
of Equation (16) is unchanged if M(N, Z), Mn and MH are replaced by their
respective mass excesses (at that time they were called mass defects). When using
the constants given above one should be aware that when Bethe fixed them, he used
for the mass excesses of the neutron and hydrogen atom respectively 7.8 MeV and
7.44 MeV in the 16 0 standard, with a value of930 MeV for the atomic mass unit. In
year 2000, we would have used 8.1 Me V, 7 .3 MeV, and the value of 'u' given in the
header of Section 3. Nevertheless, this should not be a problem for our construction
of a simple function. I refer the reader to the lecture of Mike Pearson [391 for up to
date values for the constants of this semi-empirical formula.
If we subtract Equation (16) from all masses we are left with values that vary
much less rapidly than the masses themselves, while still showing all the structures.
However, the splitting in four sheets will still make the image fuzzy. One can then
add to the right hand side of the formula of Bethe (16) a commonly used pairing
term 6. pp = 6. nn = -121 JA MeV and no 6. np (Figure 5), which is sufficient for
our purpose. (For those interested, there is a more refined study of the variations of
the pairing energies that has been made by Jensen, Hansen and Jonson [40].)
Nowadays it is preferable to use the results of the calculation of one of the
modem models. However, we can use here only those models that provide masses
from the spherical part of the formula (i.e., forcing the nucleus to be undeformed).
The reason is that the models generally describe quite well the shell and subshell
closures, and to some extent the pairing energies, but not locations of deformation.
If the theoretical deformations were included and not located at exactly the same
position as given by the experimental masses, the mass difference surface would
show two artefacts each time. Interpretation of the resulting surface would then be
very difficult.
My two choices are the "New Semi empirical Shell Correction to the Droplet
Model (Gross Theory of Nuclear Magics)" by Groote. Hill' and Takahashi r4l J:
THE EVALUATION OF ATOMIC MASSES 29
and the "Microscopic Mass Formulas" of Duflo and Zuker [36], which has been
illustrated above (Figure 7).
The difference of mass surfaces shown in Figure 7 is instructive:
(1) the lines for the isotopic series cross the N = 82 shell closure with almost no
disruption, showing thus how well shell closures are described by the model;
(2) the well-known onset of deformation in the rare-earth at N = 90 appears
very clearly here as a deep large bowl, since deformation is not used in this
calculation. The contour of this deformation region is neat. The depth, i.e.,
the amount of energy gained due to deformation, compared to ideal spherical
nuclides, can be estimated; and
(3) Figure 7 shows also how the amplitude of deformation decreases with increas-
ing Z and seems to vanish when approaching rhenium (Z = 75).
When exploiting these observations one can make extrapolations for masses
very far from stability. This has been done already [42], but with a further refine-
ment of this method obtained by constructing an idealized surface of masses (or
mass-geoid) [43], which is the best possible function to be subtracted from the
mass surface. In [42], a local mass-geoid was built as a cubic function of Nand Z
in a region limited by magic numbers for both Nand Z, fitted to only the purely
spherical nuclides and keeping only the very reliable experimental masses. Then
the shape of the bowl (for deformation) was reconstructed 'by hand', starting from
the known non-spherical experimental masses. It was found that the maximum
amplitude of deformation amounts to 5 MeV, is located at 168Dy, and that the region
of deformation extends from N = 90 to N = 114 and from Z = 55 to Z = 77,
which is roughly in agreement with what is indicated by Figure 7.
Figure 8. A screen image of DESINT, the interactive graphical display of four cuts in the surface
of masses around 146Gd. The four quadrants display respectively S2n(N), S2p(Z), Q2tl(A) and
(Mcxp - MOuflo-Zukcr)(N) [36) . The lines in black connect nuclides with same Z, N , (Z and N) and
Z , respectively. The boxes at left and bottom serve for various interactive commands. The N = 82
shell closure is clearly seen in quadrant I and in the lower left comer of quadrant 3. The lines in red
illustrate the many consequences of an increase of the mass of 146Gd by 500 keY.
ure 8. Any of the 'derivatives' or of the 'differences' can be displayed in any of the
four quadrants of Figure 8, or alone and enlarged. Any of these functions can be
plotted against any of the parameters N, Z, A, N - Z, and 2Z - N; and connect
iso-lines in any single or double parameters of the same list (e.g., in the third view
of Figure 8, iso-lines are drawn for Z AND for N). Zooming in and out to any
level and moving along the two coordinates are possible independently for each
quadrant. Finally, and more importantly, any change appears, in a different color,
with all its consequences and in all four graphs at the same time. As an example
THE EVALUATION OF ATOMIC MASSES 31
and only for the purpose of illustration, a change of +500 ke V has been applied, in
Figure 8, to 146Gd in quadrant number four; all modifications in all graphs appear
in red.
4. The tables
The most recently published mass table from the "Atomic Mass Evaluation" is of
December 1995 (AME'95) [2]. Urgency in having the first NUBASE evaluation
completed, delayed the planned update of an AME for 1997, since the two evalu-
ators of the AME are also collaborators of NUBASE. The NUBASE evaluation was
thus published for the first time in September 1997 [3]. In order to have consistency
between the two tables, it was decided that the masses in NUBASE'97 should be
exactly those from AME'95. The few cases for which new data required a change
were only mentioned in the table and discussed in the accompanying text. The
electronic ASCII files for the AME' 95 and the NUBASE' 97 tables, for use with
computer programs, are distributed by the Atomic Mass Data Center (AMDC)
through the World Wide Web [45]. The contents of NUBASE can be displayed
by a Java program JVNUBASE [46] through the World Wide Web and also with the
NUCLEUS PC-program [47], all distributed by the AMDC.
In the future, it is planned to have the AME and the NUBASE evaluations, which
have the same 'horizontal' structure and basic interconnections at the level of iso-
mers, to be published together, the first time in a year. Such a publication is urgently
needed, because of the impressive number of new results that have been published
since AME'95 and NUBASE'97.
5. Conclusion
Deriving a mass value for a nuclide from one or several experiments is in most
cases not easy. Some mathematical tools (the least-squares method) and computer
tools (interactive graphical display), and especially the evaluator's judgment are
essential ingredients to reach the best possible recommended values for the masses.
As for the unknown masses, those close to the last known ones can be predicted
from the extension of the mass surface. However, for the ones further out, more
particularly those which are essential in many astrophysical problems, like the
nucleosynthesis r-process, values for the masses can only be derived from some
of the available models. Unfortunately, the latter exhibit very large divergences
among them on leaving the narrow region of known masses, reaching up to tens of
MeV's in the regions of the r-process paths. Therefore, one of the many motivations
for the best possible evaluation of masses is to get the best set of mass values on
which models may adjust and better predict masses further away.
32 G.AUDI
Acknowledgements
I am very proud of having given this lecture in honour of Aaldert H. Wapstra with
whom I have been working now for nearly 20 years. The material used in this
lecture is also his material. He was the one who established in the early fifties the
AME in its modem shape as we know now. Aaldert H. Wapstra has always been
very accurate, very careful and hard working in his analysis in both the AME and
the NUBASE evaluations. During these 20 years I have learned and still learn a
lot from his methods. I wish also to thank my close collaborators Jean Blachot
and Olivier Bersillon with whom I have had much pleasure in having the NUBASE
evaluation become a reality, and Catalin Borcea who built the computer programs
for mass extrapolation, and worked hard at the understanding, the definition and
the construction of a mass-geoid. I am deeply grateful to Dave Lunney for critical
reading and improvements of this text.
References
1. Burrows, T. w., Nucl. lnstrum. Meth. A 286 (1990), 595;
http://www.nndc.bnl.gov/nndc/ensdf!.
2. Audi, G. and Wapstra, A. H., Nucl. Phys. A 595 (1995), 409.
3. Audi, G., Bersillon, 0., Blachot, J. and Wapstra, A. H., Nucl. Phys. A 624 (1997), 1.
4. Irvine, R. J., Davids, C. N., Woods, P. J., Blumenthal, D. J., Brown, L. T., Conticchio, L. F,
Davinson, T., Henderson, D. J., Mackenzie, J. A., Penttilii, H. T., Seweryniak, D. and
Walters, W. B., Phys. Rev. C 55 (1997), 1621.
5. Morek, T., Starosta, K., Droste, Ch., Fossan, D., Lane, G., Sears, J., Smith, J. and Vaska, P.,
Eur. Phys. 1. A 3 (1998), 99.
6. Gallagher, Jr., C. J. and Moszkowski, S. A., Phys. Rev. 111 (1958), 1282.
7. Otten, E. W., Treatise on Heavy-ion Science, D. A. Bromley (ed.) 8 (1989),517.
8. Aufmuth, P., Heilig, K. and Steudel, A., At. Nucl. Data Tables 37 (1987), 455.
9. Raghavan, P., At. Nucl. Data Tables 42 (1989), 189.
10. Stone, N., At. Nucl. Data Tables, to be published.
11. Angeli, I., Acta Phvs. Hungarica 69 (1991), 233.
12. Nadjakov, E. G., Marinova, K. P. and Gangrsky. Yu. P., At. Nllcl. Data Tables 56 (1994), 133.
13. Holden, N. E., Report BNL-61460, 1995.
14. Raman, S., Nestor, Jr., C.W., Kahane, S. and Bhatt, K. H., At. Nucl. Data Tables 42 (1989), 1.
15. Lepine-Szily, A., Experimental overview of mass measurements, tutorial lecture. this issue. 35.
16. Audi, G., Wapstra, A. H. and Dedieu, M., NlIcl. Phys. A 565 (1993), 193.
17. Cohen, E. R. and Wapstra, A. H., Nucl.lnstrul11. Meth. 211 (1983),153
18. Koslowsky, V. T., Hardy, J. C., Hagberg, E .. Azuma, R. E., Ball, G. c.. Clifford. E. T. H.,
Davies, W. G .. Schmeing, H., Schrewe, U. J. and Sharma, K. S .. Nucl. PhI'S. A 472 (1987),419.
19. Kessler, Jr., E. G., Dewey, M. S .. Deslattes, R. D .. Henins, A., Borner, H. G., Jentschel, M.,
Doll, C. and Lehmann, H., Phys. Lett. A 255 (1999), 221.
20. Davids, C. N., Woods, P. J., PenttiHi, H. T., Batchelder, J. c., Bingham, C. R., Blumenthal, D. J.,
Brown, L. T., Busse, B. C .. Conticchio, L. F, Davinson, T., Henderson, D. J., Irvine, R. 1..
Seweryniak, D., Toth, K. S., Walters, W. B. and Zimmerman, B. E., Pin's, Rel', Lelf, 76 (1996),
592:
Davids, C. N, et o/., Masses and proton separation energies obtained from Q" and Qp
measurements, this issue, p, 133,
THE EVALUATION OF ATOMIC MASSES 33
21. Ries, R. R., Damerow, R. A. and Johnson, W. H., In: Proc. AMCO-2 Conf., Vienna, July 1963,
p.357.
22. Barber, R. c., Bishop, R. L., Cambey, L. A., Duckworth, H. E., Macdougall, J. D.,
McLatchie, W., Ormrod, J. H. and Van Rookhuyzen, P., In: Proc. AMCO-2 Conf., Vienna,
July 1963, p. 393.
23. Gillibert, A., Mittig, w., Bianchi, L., Cunsolo, A., Fernandez, B., Foti, A., Gastebois, J., Gre-
goire, c., Schutz, Y. and Stephan, c., Phys. Lett. B 192 (1987),39;
Vieira, D. J., Wouters, J. M., Vaziri, K., Krauss, Jr., R. H., Wollnik, H., Butler, G. W.,
Wohn, E K. and Wapstra, A. H., Phys. Rev. Lett. 57 (1986), 3253;
Savajols, H., The SPEG mass measurement program at GANIL, this issue, 245.
24. Chartier, M., Auger, G., Mittig, w., Lepine-Szilly, A., Fifield, L. K., Casandjian, J. M.,
Chabert, M., Ferme, J., Gillibert, A., Lewitowicz, M., Mac Cormick, M., Moscatello, M. H.,
Odland, O. H., Orr, N. A., Politi, G., Spitaels, C. and Villari, A. C. c., Phys. Rev. Lett. 77
(1996), 2400.
25. Trotscher, J., Balog, K., Eickhoff, H., Franczak, B., Franzke, B., Fujita, Y., Geissel, H.,
Klein, Ch., Knollmann, J., Kraft, A., Lobner, K. E. G., Magel, A., Miinzenberg, G.,
Przewloka, A., Rosenauer, D., Schafer, H., Sendor, M., Vieira, D. J., Vogel, B., Winkelmann,
Th. and Wollnik, H., Nucl. Instrum. Meth. B 70 (1992),455;
Hausmann, M., Isochronous mass measurements of hot exotic nuclei, this issue, 291.
26. Smith, L. G. and Damm, C. c., Rev. Sci. Instr. 27 (1956), 638;
Smith, L. G., Phys. Rev. 111 (1958), 1606;
Lunney, D. et aI., Recent results on Ne and Mg from the MISTRAL mass measurement program
at ISOLDE, this issue, 299.
27. Schwinberg, P. B., Van Dyck, Jr., R. S. and Dehmelt, R. S., Phys. Lett. A 81 (1981),119;
Bollen, G. et al., Mass measurements on short-lived nuclides with ISOLTRAP, this issue, 215.
28. Franzke, B., Beckert, K., Eickhoff, H., Nolden, E, Reich, H., Schaaf, u., Schlitt, B.,
Schwinn, A., Steck, M. and Winkler, Th., Phys. Scr. T59 (1994), 176;
Litvinov, Yu. A. et al., Mass measurements of cooled exotic nuclei, this issue, p. 283.
29. Nuclear Structure Reference (NSR): a computer file of indexed references main-
tained by NNDC, Brookhaven National Laboratory; http://ndcntl.dne.bnl.gov/nsrq/ or
http://www.nndc.bnl.gov/.
30. Woods, M. J. and Munster, A. S., NPL Report RS(EXT) 95 (1988).
31. Particle Data Group, Review of particle properties, Eur. Phys. 1. C 3 (1998), 1.
32. Wapstra, A. H., Inst. Phys. Conf. Series 132 (1993), 129.
33. Linnik, Y. V., Method of Least Squares, Pergamon, New York, 1961; Methode des Moindres
Camis, Dunod, Paris, 1963.
34. Audi, G., Davies, W. G. and Lce-Whiting, G. E., Nucl. Instrum. Meth. A 249 (1986), 443.
35. Audi, G., Epherre, M., Thibault, C., Wapstra, A. H. and Bos, K., Nucl. Phys. A 378 (1982),
443.
36. Duflo, J. and Zuker, A. P., Phys. Rev. C 52 (1995), 23;
private communication February 1996 to the AMDC, http://csnwww.in2p3.fr/AMDC/theory/
du_zu_1O.feb96.
37. von Weizsacker, C. E, Z. Phys. 96 (1935), 431.
38. Bethe, H. A. and Bacher, R. E, Rev. Mod. Phys. 8 (1936),82.
39. Pearson, M., The quest for a microscopic nuclear mass formula, tutorial lecture, this issue, 59.
40. Jensen, A. S., Hansen, P. G. and Jonson, B., Nucl. Phys. A 431 (1984), 393.
41. Groote, H. v., Hilf, E. R. and Takahashi, K.,At. Nucl. Data Tables 17 (1976),418.
42. Borcea, C. and Audi, G., AlP Conf. Proc. 455 (1998), 98.
43. Borcea, C. and Audi, G., Proc. Int. Conf. Exot. Nuc. & At. Masses (ENAM'95), 1995, p. 127.
34 G. AUD!
44. Borcea, C. and Audi, G., Rev. Roum. Phys. 38 (1993), 455;
Report CSNSM 92-38, October 1992, http://csnwww.in2p3.fr/AMDC/extrapolationslbernex.
pdf.
45. The NUBASE and the AME files in the electronic distribution can be retrieved from the Atomic
Mass Data Center through the Web at http://csnwww.in2p3.fr/amdc/.
46. Durand, E., Report CSNSM 97-09, July 1997.
47. Potet, B., Dufio, J. and Audi, G., Proc. ENAM'95 Conj, Aries, June 1995, p. 151.
Hyperjine Interactions 132: 35-57,2001. 35
© 2001 Kluwer Academic Publishers.
ALINKA LEPINE-SZILY
Instituto de Fisica-Universidade de Sao Paulo, c.P. 66318, 05389-970 Sao Paulo, Brasil
Abstract. Nuclear masses and binding energies are an important input for nuclear models, consti-
tuting strong constraints far from the ,B-stability valley. During the last decade, new experimental
techniques for the production of exotic nuclei and the measurement of their mass were developed.
The present paper gives an overview of these techniques.
1. Introduction
The experimental study of atomic masses has been underway almost 90 years.
After the pioneering work of Thompson [1] in 1911, discovering the existence
of isotopcs, it was Aston [2] who performed systematic measurements of atomic
masses with a mass spectrograph. He obtained nuclear binding energies for many
nuclear species. The near constancy of the average binding energy per nucleon
suggested the saturation of the nuclear forces and the independence of the nuclear
density with the atomic weight A. This average behaviour can be described by the
liquid drop model for the nucleus and resulted in the first mass formula [3] by
Weizsacker in 1935.
In the following decades the mass studies were concentrated on stable and long
lived nuclei, close to the ,B-stability valley, and these measurements were per-
formed mainly with high-resolution mass spectrometers (Section 3.2.l.a), which
became more and more accurate due to technical improvements. Other methods
also became available and nowadays the most precise mass spectrometers are the
Penning traps (Section 3.2.l.c), achieving an accuracy of 1 part in 1010 for stable
ions. Penning traps are environments with static magnetic and electric fields, where
single ions are confined and their cyclotron frequency is measured with a very
high precision. The mass differences of the electron-positron [4], of the proton-
antiproton [5], of light and heavy ions [6, 7] were measured by this method with
an accuracy of 10- 10 . The precision attained is such that it can be used to obtain
more accurate values for the fundamental physical constants, as the fine-structure
constant a [6, 7], or the Avogadro number NA [8].
Towards the end of the 1970's, measurements on stable nuclei were essentially
complete. By the mid 1970's, mass measurements began to be extended beyond
the region of stability by the pioneering work of the Orsay group at CERN [9].
36 A. LEPINE-SZILY
Since then the bulk of measurements concerns the masses of nuclei out of the
stability valley, reaching more and more exotic nuclei, produced with accelerators
or nuclear reactors. Nowadays the direct mass measurements, determining either
the cyclotron frequency or the time of flight of the radioactive ions in a magnetic
field constitute the most powerful and precise methods. They will be described in
detail below (Section 3.2).
General trends of nuclear binding energies as a function of the atomic num-
ber Z and the mass number A can be considered as the manifestation of global
properties of nuclear matter, often described in a first approximation by liquid drop
models. Superimposed on this smooth behaviour are rapid changes, reflecting the
microscopic structure of the nucleus. Nuclear models were developed to explain
the trends of nuclear binding energies among other nuclear properties. The nuclear
mass models, empirical (e.g., [10)), macroscopic-microscopic [11, 12] or fully
microscopic [13], have root-mean-square (RMS) deviations between the calculated
and measured masses over the whole chart of nuclides about 0.7-1 MeV. For almost
all masses the experimental precision is much better than the model predictions,
they are measured with precisions of keV or even less.
Going away from stability the predictions of different nuclear models diverge
among them and from experimental values, when available. Differences can be of
several Me V. Thus atomic masses and nuclear binding energies far from stability
constitute a very strong constraint on nuclear models. However, the precision of
measurements at the border of the region of known mass values is only 8m / m =
10-4 -10- 6 . Thus the accurate measurement of short lived nuclear masses is of
great interest for the further progress of nuclear models.
In the following we will concentrate our discussion on mass determinations
of radioactive nuclei, which constitute the bulk of present measurements. The
techniques to be used [14] depend on a large extent on the mass region, on the
production and separation method (energy, intensity), on the half-life, etc. For this
reason the production and separation methods will also be discussed, followed by
the mass measurement methods.
This method [16] is mainly used in association with production modes of poor
kinematical focusing, as target fragmentation or spallation, target fission, deep-
inelastic and fusion-evaporation reactions. The production target is thick enough
to stop the primary beam and the radioactive nuclei produced in the reaction. The
reaction products in neutral charge state diffuse out from the high temperature
target and are ionised (l +) in an ion source and subsequently accelerated (to 10-
100 keY) and electromagnetically separated. The preparation time is mainly due
to the diffusion process ('" 1-1000 ms), which depends on the target matrix but
also on the element in question. The secondary beam produced by this method
has excellent beam quality. The low-energy ISOL beams can be post-accelerated
(Cyclone-LLN, SPIRAL-GANIL, REX-ISOLDE and others); however, for mass
measurements the increase in energy is not desirable.
At the actual stage of developement of the target and ion-source units the pro-
duction of intense low energy beams of many radioactive isotopes of the periodic
table has become possible. At the ISOLDE mass separator facility at CERN, ra-
dioactive nuclides are produced in thick high-temperature targets via spallation,
fission or target fragmentation reactions, induced by an intense (2 j.!A) proton
beam of 1 or 1.4 GeV, delivered by the Proton-Synchrotron Booster (PSB) of
CERN. Until now more than 600 isotopes of more than 60 elements (Z = 2-88)
have been produced with half-lives down to milliseconds and intensities up to
1011 ions/so Facilities with ISOL systems can also be found at GSI, Berkeley,
GANIL, Louvain-Ia Neuve, Jyvaskyla, Oak Ridge, Studsvik and TRIUMF among
others [15, 17].
The main advantage is the good beam quality (emittance ~ 10rr mm mrad,
energy dispersion less than 5 eV) and the low kinetic energy. One limitation is the
slowness of the diffusion process that can cause severe decay losses for very short-
lived nuclides. Also there is a strong dependence of the efficiency on chemical
properties of the element of interest.
38 A. LEPINE-SZILY
nuclei, otherwise difficult to create. This reaction mechanism is mainly used for
mass measurements in association with the ISOL method. The mass-measurement
mode associated is the fi-decay Q-value if the secondary particles are analysed by
a recoil separator; Decay Q-value or mass spectrometry if the ISOL method is used
for separation.
2.2.1.d. Fission induced by light ions on fissile targets. Fissile targets can be
bombarded by thermal neutrons from reactors or by low energy light ions (p, d,
etc.) of an accelerator and undergo fission. The isotopic distribution of the fis-
sion fragments is determined by the excitation energy. For low energy induced
fission, the fragment distribution presents the well-known double-humped form,
forced by strong shell effects. This process is a powerful tool to produce medium
mass neutron rich nuclei, since the A j Z ratio of fissile nuclei is high compared
to stable nuclei. The mass measurement mode associated is mass spectrometry
(Sections 3.1.2 and 3.2.1) following separation by ISOL method.
zone from target and projectile is sliced off or abraded, while the non-interacting
parts of target and projectile act as spectators. The projectile fragments (specta-
tors) are only moderately excited and decay by particle emission and continue
with the same average velocity as the projectile. The abrasion process and the
de-excitation introduces fluctuations in their velocity due to the momentum dis-
tribution of the abraded nucleons. However, the projectile fragments present a very
strong kinematical focalization at 0°, with a small angular and energy dispersion
of the secondary beam. This feature favours the subsequent selection by recoil
spectrometers. The total reaction cross-section of this process is the geometrical
cross section with a correction parameter for the nuclear transparency. This process
produces a large number of fragments with different A and Z. The largest cross-
sections are concentrated along the "spallation corridor", located parallel to the
stability line, shifted to the proton rich side. The Nand Z distribution depends on
the projectile N / Z ratio, representing a memory effect. The A and Z distributions
are broad, the cross-sections are high (mb) and neutron- and proton-rich exotic
nuclei are produced as well. Model predictions realized in the computer codes,
as EPAX [18], LISE [19], INTENSITY [20] show fairly good agreement with
experimental results over the whole periodic table. The separation method is the in-
flight, and the masses can be measured by the direct mass measurement techniques:
Time of flight along a linear flight path (SPEG or TOFI methods; Section 3.2.2.a)
or along a circular or elipsoidal path (Isochronous method in a storage ring; GSI-
ESR; Section 3.2.2.d) Frequency measurements in a storage ring (Schottky Mass
Spectrometry at GSI-ESR; Section 3.2.2.c).
Spallation of heavy targets by high energy (l GeV /nucleon) light particles can
be considered as a target fragmentation. The cross sections are similar to heavy ion
beam target fragmentation. The separation method is mainly the ISOL mode, and
the masses can be measured by the direct mass measurement techniques. In case of
ISOL (low energy): Penning trap (ISOLTRAP, Section 3.2.l.c), Radio Frequency
Mass Spectrometer (MISTRAL, Section 3.2.l.b), conventional mass spectrometers
(Section 3.2.1.a).
11I
,,-....
I
1 1
200 C\J +
,-NO
lD
1'0,-.
'---" 10 1'0
n
150 n to
<:l- n
(J)
..........
100
C
::J
u
0 50
50
Figure 1. Energy spectrum of the IOB( 14N, 13B)11 N reaction obtained after subtraction of the
spectra of the contaminants [25]. The arrow indicates the proton threshold.
and six well defined resonances of the neutron-deficient, unbound liN nucleus,
above the IOC + P threshold (see Figure 1).
The main advantage of this method is the relatively good precision, superior
to those supplied by direct time-of-flight measurements, and the lack of limitation
in the half-life of the nucleus to be studied, allowing even to acess the masses of
particle-unbound nuclei (e.g., 9He [26, 27], IOHe [28], IOLi [23, 30], 13Be [29],
16B [30] and liN [24, 25]). Multinucleon transfer reactions present some experi-
mental difficulties, which arise mostly from the low-reaction cross sections. It is
well known from transfer matching conditions that for strongly negative Q-values
the largest possible l-transfer is dominating the reaction, often resulting in the ex-
cited states being more strongly populated than the ground state. Considerable care
must be given, therefore, to identifying the ground state, especially in the presence
of reactions on contaminations in the target.
Pion double charge exchange (DCX) reactions (JT±, JT'f), with b.Tz = 2, have
been employed with considerable success to measure the masses of a number of
light and medium mass nuclei using the EPICS spectrometer at Los Alamos [26,
31, 32]. The reaction Q-value is related to mass excesses through: Q = b.A - b.D ±
2mec2. Experimentally the method is similar to that used for transfer reactions with
the constraint that the calibration reaction is DCX as well. The mass determinations
EXPERIMENTAL OVERVIEW OF MASS MEASUREMENTS 43
have moderate uncertainties ("-'50-150 keV) the statistics being the main limitation
of the precision. The measurements are challenging owing to the weak reaction
cross sections ("-'100-500 nb/sr) coupled with low pion fluxes. As very thick tar-
gets are used in compensation and the beam spot size is large (typically 10-20 cm)
prohibitively large quantities of target material ("-'20-200 g) are needed.
Another technique which was recently developed to measure masses of light,
unbound nuclei, consists in the simultaneous detection of all decay products (ener-
gies and angles) allowing the construction of the decay energy or so-called invari-
ant mass spectrum, which provides a measure of the mass of the system. Usually
the unbound nuclei are produced in reactions with radioactive beams (for example,
IOHe produced by the one-proton stripping of a II Li beam [33]). In the above
example of IOHe, all decay fragments (SHe and two neutrons) were detected in coin-
cidence in quite large detector arrays. The invariant mass measurements do present
some specific advantages. The method can be extended to much heavier systems
than classical transfer or DCX experiments (although presently available beam in-
tensities limit work to the lightest systems) and its selectivity favours the population
of the ground state contrarily to multi-nucleon transfer reactions. The mass mea-
surement of IOLi and 13Be (performed by fragmentation of a stable 18 0 beam and by
0° measurement of the relative velocities of the neutron and charged fragments, 9Li
and 12Be) suggested [34] that these systems are more bound than was determined
from transfer reactions [29,30,35]. The main limitation is the final precision (300-
400 keY), which is considerably poorer than in multi-nucleon transfer reactions,
due to quite low event rates and relatively low detection resolution.
the Q-value for the decay is derived from the feedings to a number of excited states
as well as to the ground state.
In f3- -decay the Q{3- is determined directly from the endpoint of the electron
energy spectrum. The electron spectra are usually measured using solid state detec-
tors [36, 37], scintillation counters [38] or magnetic spectrometers [39]. Given the
decreasing number of events in the spectra as the endpoint energy is approached,
the data treatment (shape fitting techniques, distortions in the measurements of
the electron energies, etc.) has to be very careful and the response function of the
detector should be well understood [36].
In the direct measurement of the endpoint of the positron energy spectrum
results can be affected due to summing of the positron energy with the annihi-
lation radiation. A summation-free, f3+ -endpoint spectrometer [40] was developed
to bring solution to this problem. The ratio of positron emission to electron capture
is strongly dependent on the endpoint energy, and in principle the QEC may be
deduced from this ratio. However, this method is very sensitive to the feeding of
high lying levels in the daughter. Given the difficulties presented by both methods
care must be taken in interpreting the results of QEC determinations [17,40].
The f3-decay Q-value mass determinations are often limited by the produc-
tion rates, since coincidence measurements and endpoint energy determination
need high statistics. Apart from systematic errors, precisions from ~ 10 keV to
1 MeV may be obtained depending on statistics. Nevertheless in some cases, such
as neutron-rich fission products, f3-decay measurements still represent the most
viable approach.
Measurements of proton or a-decay can also provide mass differences and the
masses of nuclei beyond the drip-line may be accessed in this manner [41]. Given
that the lifetime must be sufficiently long that the decay can be observed, such
measurements are restricted to nuclei above A ::::::: 100. As the decay energies are
discrete quite high precisions (~1 0 ke V) are attainable.
where Bp is the magnetic rigidity, v is the ion velocity, m is the mass, q is the
charge state, y is the usual relativistic factor and We is the cyclotron frequency.
In principle all parameters entering in this equation can be measured in absolute
value. However in order to increase the precision, usually relative values and some
(2 or 3) well-known reference masses are used. This relation can be used at low or
high energy, measuring velocity (time of flight) or cyclotron frequency.
EXPERIMENTAL OVERVIEW OF MASS MEASUREMENTS 45
3.2.1.h. Radio frequency mass spectrometers (RFMS). The radio frequency mass
spectrometer (RFMS) was proposed and realized by Smith in 1960 [47]. He ob-
tained with its last version, a resolving power of 107, allowing the determina-
tion of stable atomic masses with a precision better than 10-9 [47], which was
and remained for a decade the best precision obtained at that time. The impor-
tant advantage of the RFMS over classical mass spectrometers is the fact that
it measures frequency ratios, which are much more precisely determined than
voltage ratios. Using the RFMS one compares masses MA and MB by the com-
parison of their cyclotron frequencies W A and WB in the same homogeneous, sta-
tic, magnetic field B. If the two ions have the same trajectory, then by Equa-
tion (2) we see that the frequencies and masses satisfy the relation wAMA
wBMB.
Recently a new experimental program MISTRAL (Mass Measurements at
ISOLDE using a Transmission and RAdiofrequency spectrometer on Line) was
initiated at ISOLDE/CERN [48]. It is based on a RFMS constructed in Orsay [49]
and installed at ISOLDE/CERN in 1997. In recent data taking periods the masses of
neutron rich isotopes of Ne, Na, Mg, Al K, Ca, Ti and of the very neutron deficient
74Rb were measured with very good precisions (3-7 x 10- 7). The nucleus with
shortest half-life measured was the 28Na (31 ms) [50]. The very short transit time
('"'-'50 1lS), makes MISTRAL ideally suited for high precision mass measurements
of short lived nuclei far from stability [51]. The main parts of the RFMS are a mag-
net with a highly homogeneous magnetic field, a modulator and two symmetrical
beam lines for the injection and ejection of the beam. The beam is injected into the
homogeneous area of the magnet, makes two turns in a slightly helicoidal trajectory
and is again ejected. It crosses the modulator at the first and third half-tum, which
is a radio-frequency electric field that accelerates or decelerates the beam. Ions may
only reach the final detector if this radiofrequency WRF is related to their cyclotron
frequency We through WRF = (n + 1/2)we, which leads to a cancellation of the
two modulations. When the modulator frequency is scanned, narrow transmission
peaks are observed (with resolution R = 105 ) for integer values of WRF/we - 1/2,
where the transmission is maximum.
To perform the mass measurement an unknown mass (from ISOLDE) is injected
alternatively with a reference mass (from a reference ion source), in rapid succes-
sion (15 s) to eliminate short term drift in the magnetic field. However, comparing
masses without changing the magnetic field requires to change the accelerating
voltage of the reference beam and the voltages of the electrostatic and focusing
elements of the beam lines, following the relationship VrefMref = VisoMiso. This
procedure introduces systematic errors and reduces the attainable accuracy. The
EXPERIMENTAL OVERVIEW OF MASS MEASUREMENTS 47
use of reference mass value very close to the mass to be measured (isobars), as in
the PMS method [46] used in Gatchina, could possibly avoid the switch in volt-
age. The unknown mass is obtained by the comparison of the cyclotron frequency
of the unknown mass with the reference mass, since for the same field strength
WrefMref = WisoMiso'
3.2.l.c. lon traps as high precision mass spectrometers. The recent use of a Pen-
ning trap for the high precision mass measurement of unstable nuclei produced
outside of the ion trap [52] represented a breakthrough in the domain of exotic
nuclei. The only working device of such a configuration at the moment is the
ISOLTRAP at CERN [53, 54]. The Penning trap mass spectrometer [55] installed
at the CERNIISOLDE mass separator is divided into three parts: a RFQ (Radio Fre-
quency Quadrupole) ion beam buncher, the "Cooler trap" and the "high-precision
trap". The 60 keY ISOLDE ion beam is injected into the RFQ ion buncher, which is
a buffer gas filled 4-rod structure, with a DC potential applied along the segmented
rods. The buffer gas retards the ions to a few ke V, the oscillating quadrupole
field applied to the 4-rods confines the ions radially and the DC potential drags
the ions near to the end, where they are accumulated as a small cloud. They are
ejected as a well defined bunch and injected into the "Cooler trap", which is a
large cylindrical Penning trap placed in the homogeneous field of a 4.7 T su-
perconducting magnet, where they are mass selectively cooled. The cooler trap
resolves and rejects adjacent isobars and delivers clean ion bunches to the sec-
ond, so-called precision Penning trap, which performs the highly accurate mass
measurements.
In this Penning trap a homogeneous static magnetic field is used for radial con-
finement and an axially symmetric static electric quadrupole field produces an axial
confinement. The superposition of the magnetic dipole and the electric quadrupole
fields results in a circular cyclotron motion with a modified frequency W+ and a
slow precession of this circular motion around the trap axis called magnetron mo-
tion with frequency W_. This circular motion is combined with a simple harmonic
axial oscillation with w z . The sum of W+ and w_ is exactly the cyclotron frequency
Wc and they are given by the following relation:
Wc
w±=-± ( W c)2 (3)
2
2 2
For stable isotopes, mass determinations of ultimate precision (10- 11 ) are achieved
by measuring all three eigenfrequencies to get We (see [56, 57], this issue). Since
this process involves too much time compared to the half-lives of radioactive nu-
clides, a different technique is used by ISOLTRAP. In order to measure the mass of
the nucleus in the trap an azimuthal quadrupole RF field is applied. The externally
applied RF is in resonance with the sum of w_ + W+ = We with a maximum
energy gain detected by a time-of-flight technique. After excitation, the ions are
ejected from the trap and allowed to drift through the inhomogeneous fringe field
48 A. LEPINE-SZILY
of the magnet, which gives rise to an axial force resulting in an increase in the
axial momentum. Measuring the time-of-flight from the trap to the detector as a
function of the applied RF frequency, a resonance curve centered at We is obtained
and the mass of the ion can be determined. The width of the cyclotron resonance
curve ~we (FWHM) = 0.9/ TRF is determined by the Fourier limit of the driving
RF field, with duration TRF • The magnetic field calibration is made by measuring
the cyclotron resonances of the stable reference isotopes with well known mass
values. A singly charged ion with mass number of A = 50 in a magnetic field
B = 6 T has We = 2 MHz, and if TRF = 0.9 s, ~wc = 1 Hz, corresponding to a
frequency resolution of 5 x 10-7.
Initially the technique was applied to alkali and alkali-earth metal ions, mea-
suring the mass excesses of many unstable Cs, Rb, Sr, Ba, Fr and Ra isotopes [52,
54, 58] with a mass resolving power of 106 , improving considerably many existing
mass values. The ground and isomeric states of 78Rbm,g and 84Rbm,g were resolved
for the first time by direct mass spectrometry [58] using ISOLTRAP, with a mass
resolution of 70 keV, allowing the clear separation of ground and isomeric states.
More than 40 isotopes of neutron deficient rare earth elements ranging from Pr,
Nd, Pm, Sm, Eu, Dy and Ho [59] have been measured with a typical accuracy of
14 keY. Since the ISOLDE beam for these elements has a large amount of isobaric
contamination, these measurements only became possible after the installation of
the new cooler trap. Recently the masses of neutron deficient isotopes of Hg,
Pb, Bi and Po were measured by this technique. Increasing TRF from 0.9 to 8 s
the mass precision could be reduced to 50 ke V for these heavy nuclei, allowing
the resolution of ground and isomeric states for several Hg isotopes. Very recent
mass measurements of 33-35,42,43 Ar isotopes (TI/2 (33Ar) = 174 ms) and of 74Rb
(T1/2 = 65 ms) have shown that the half-life limits were dramatically reduced. On
the other hand, measurements on 74Kr (T1/2 = 11.5 m) attained an unprecedented
accuracy of a few times 10- 8 . These measurements benefit from the improvements
of sensitivity (of two orders of magnitude) due to the RFQ beam buncher.
A Penning trap mass spectrometer (Canadian Penning Trap, CPT) was con-
tructed at the TASCC facility of the Chalk River Laboratories, After the shut-down
ofTASCC in 1997, CPT is being installed at Argonne National Laboratories [60].
The problem of deceleration of an in-flight separated, energetic radioactive beam
to very low energies and its injection and capture into a Penning trap is addressed
for the first time. The problem of deceleration and capture of very long-lived highly
energetic antiproton beams has already been addressed by Gabrielse et al. [5].
Penning trap mass spectrometers have as main advantage their high precision.
When coupled to a highly versatile and efficient source of radioactive beams (as
ISOLDE), they constitute an outstanding facility for the high precision mass mea-
surements of radioactive ions with half-lives TI/2 ~ tens of ms. It is important to
observe, that the half-life limitation, connected both to the ISOL and trap confine-
ment method - where the time duration is inversely proportional to the precision -
has been dramatically reduced recently.
EXPERIMENTAL OVERVIEW OF MASS MEASUREMENTS 49
3.2.2. Direct mass measurements oj radioactive ions with intermediate and high
energies
3.2.2.a. Time-oj-flight (TOF) method on linear path. One of the most important
advances in the last decade in the measurement of masses far from stability has
been the development of direct time-of-flight (TOF) techniques [61], coupling the
projectile fragmentation and target spallation reactions to high resolution magnetic
spectrometers - SPEG at GANIL [62] and TOFI at Los Alamos [63]. A system-
atic survey of the nuclear mass surface for the neutron-rich nuclei from 11 Li to
72Ni [64-69] was obtained from experiments undertaken by both the TOFI and
SPEG groups.
The direct TOF technique, as described by Equation (2), requires only the knowl-
edge of the magnetic rigidity (Bp) and the determination of the velocity v of the
ion on a long linear path, by the direct measurement of its TOP. In the case of
the mass measurement program based on the TOFI (Time-of-Flight Isochronous)
recoil spectrometer [65, 66, 68, 69], light and medium mass neutron-rich nuclei
were produced via the spallation and fission of a thin natTh target bombarded
by the high current (1 mA) 800 Me V proton beam from LAMPP. Low energy
('"'-'2-3 A Me V) products traversing am j q filter are transmitted to TOF!. The TOFI
spectrometer consists of four identical dipole elements arranged such that the TOF
of ions of a particular m j q traversing the system is independent of the velocity
(larger v implies longer path), thus the TOF is a direct measure of m j q. The mass
resolution is determined by the timing resolution: typical numbers are ot = 200 ps,
TOF'"'-' 600 ns over the 14 m flight-path of TOFI, which results in omjm '"'-' 3 x
10-4 . Care must be taken to avoid isobaric contamination, which can be present
due to the limited resolution of the mass filter preceding the TOFI spectrometer
and the low energy of the secondary beam. The masses of neutron rich nuclei in
the fp shell (50 < A < 80) were determined recently by the TOFI group with a
resolution of 2 x 10-4 [70].
In the case of the SPEG (Spectrometre aPerte d' Energie du Ganil) program [64,
67, 71, 72] the nuclei of interest are produced by projectile fragmentation of high
intensity, intermediate energy heavy-ion beams from the GANIL coupled cyclo-
trons. The projectile-like fragments emitted from the production target, located at
the exit of the second cyclotron, are selected by the beam analysis spectrometer
and transported along a doubly achromatically tuned beam line to the focal plane
of SPEG, a flight path of 82 m. The TOF measurement is made, as in the TOFI
experiments, using a pair of microchannel plate (MCP) detectors - TOF is '"'-' 1 Il-s.
The magnetic rigidity of each ion is derived from a position measurement made
using a thin position sensitive detector system (PPAC or MCP) located at the dis-
persive image plane of SPEG. Resolutions (FWHM) of '"'-' 10-4 in the rigidity and
'"'-'2.5 x 10-4 in TOF are typically achieved, corresponding to a mass resolution of
'"'-'3 x 10-4 . The SPEG mass measurements yielded masses for 39 n-rich nuclides
around N = 20 [64, 67]. More recently the masses of 31 n-rich nuclides near to
N = 28 were measured by the SPEG collaboration [72]. The direct mass measure-
50 A. L13PINE-SZILY
ment of neutron deficient isotopes of Ga, Ge, As, Se and Br was performed with
this technique [71] and the data are being reanalysed [73].
In both the TOFI and SPEG measurements a large number of different nuclides
("-' 100) are simultaneously transmitted through the beam line and spectrometer.
With a large number of reference masses providing calibration over a wide range
of Z and A, mass determinations with precisions of the order of 10-6 are possible
for the mass region A ~ 80. The final uncertainties range from "-' 100 keV for
nuclei relatively close to stability (many thousands of events) to "-'I MeV for those
approaching the ends of isotopic chains (some tens of events). The popUlation of
isomeric states with TI/2 "-' TOF "-' 1 J.lS presents a potential problem, as the resolu-
tion is not sufficient to resolve the typical mass difference between the ground and
excited states. Clearly, well known nuclei with isomeric states must be excluded
from the calibration.
The main advantages of this mass measurement method are related to the ad-
vantages of the projectile fragmentation production method, as the simultaneous
production and measurement of a large number of nuclear masses and short transit
times allowing the measurement of very short lived nuclides (T1/2 "-' 1 IJ-s). The
main limitation is in the precision attained, which is intrinsically limited by the
path length and by timing resolution (detection-electronics). In order to extend
the TOF measurement to heavier nuclei (A ~ 80) and to increase the preci-
sion for the lighter ones, a substantially longer flight path is required. Due to
the practical limitation on a rectilinear flight path, this requirement can be best
achieved with spiral or circular path. New mass measuring methods have been
developed, which use TOF or frequency measurements associated with circular or
spiral trajectories.
B m Bp
--=y-=- (4)
WRF/ h q v'
~ ~~ [ L _ _ _ _ _ _ _- - L _
--.. 4.4
~
'U
~t:
4.35
rbi, §3D
Sn
~ 4.3
'U
detector
~ 4.25
'"
to 4.2
t5
__ 4.15 J
S -25. - IS. -5. 5. 15. 25. 35. 45.
~ Plio e degree)
(
415
~ 410
"-
fRF = 13 MHz and NT = 250 one finds that, (omjq)j(mjq) = 1.7 x 10- 5 x ot
(ns), A time resolution of "'300 ps is readily attainable between start and stop
detectors, resulting in a mass resolution of 4 x 10-6.
This method was used at GANIL for mass measurements of nuclei far from
stability [74-77]. The GANIL accelerator consists of a system of two identical
separated-sector cyclotrons (CSSl and CSS2), where the beam is delivered from
an ECR ion-source and preaccelerated in a small compact cyclotron CO. In the
GANIL system all three cyclotrons (CO, CSSl, CSS2) use the same RF frequency
with the same initial phase. In standard use a stripper is located between the two
cyclotrons to ionize and inject the beam for further acceleration in CSS2. In the
mass measurement technique the stripper was substituted by a production target,
where the secondary nuclei are produced by fusion-evaporation reaction in inverse
kinematics to be then injected and accelerated in CSS2. With a suitable detection
system and a calibration based on ions of known masses it is thus possible to
determine the mass of the secondary ions to high precision. This method was tested
with light heavy ions of Ajq = 3.0 (6He, 9Li, 12Be [74] and the precision attained,
when the masses were accelerated simultaneously, was of 8 x 10-7.
The abundantly produced 100 Ag was the reference mass and the 1OoCd, 100 In
and 100Sn ions had their masses measured [75, 76], the Cd in very good agreement
with previous results, and the In and Sn for the first time. The precision attained,
when not limited by statistics, was of 2 x 10- 6. Recently the A = 80 isobars, 80Zr,
sOY, 80Sr (q = 20+) were simultaneously accelerated in CSS2 with 76Sr, 76Rb,
76Kr (q = 19+) and 68Se and 68As (q = 18+) and their mass differences were
determined [78].
The main advantage of this mass measurement method, which benefits from
characteristics of the fusion-evaporation reaction in inverse kinematics (low en-
ergy, foreward focusing, not many recoiling species) is the high precision (2 x 10-6)
and the short transit time (40-70 I1-s). Very short-lived nuclei can be studied by this
method. The difficulty of identification of the transmitted nuclides, the potential
existence of isomeric states and the low m j q acceptance of CSS2 ('" 10-4 ), accept-
ing and accelerating simultaneously only ions with very close m j q ratios, are some
of the limitations of this method.
A somewhat similar method, which also makes use of two cyclotrons was used
at the SARA cyclotrons ofISN/Grenoble [79]. The acceleration mode in the second
cyclotron at SARA is asynchronous, what means that the radiofrequency of the
second cyclotron is different from the RF of the first one (RF2 #- RF 1) and particles
arrive with random phase relative to RF2. A change in mjq value can be obtained
by changing the radio-frequency RF2 of the second cyclotron and thus a larger
range of m j q values can be measured. A precision around 10-6 was obtained
in the mass measurements of A = 80 isobars [79]. The advantage of decoupled
frequencies has to be payed by a lower transmission and the necessity to measure
the TOF by a start detector between the target and the injection, which usually
changes the charge state of the nuclides of interest before injection.
EXPERIMENTAL OVERVIEW OF MASS MEASUREMENTS 53
_ SCHOTTKY MASS SPECTROMETRY _ _ _ _ _ _ ISOCHRONOUS MASS SPECTROMETRY,
Seplum
C<x>led Fragments
L.......:......=::::::!~_---:t. 1 Yt ~Y I Hoi Fragments
In the other method [88, 89] the velocity dependent term in the right-hand side
of Equation (6) disappears by choosing the beam energy and tuning the ring such
that Y = Ytf' i.e., fullfilling the isochronous condition. Here, no electron cooling is
needed and the mass measurement is achieved immediately after injection.
Isochronicity has been achieved for ion energies corresponding to Y = Ytf =
1.37. In this case the ESR can be used as a Time-of-Flight (TOF) mass spectrometer
for the mass determination of very short lived exotic nuclei [90]. The first measure-
ments [91] on hot fragments of 58Ni and 22Ne primary beams, with known mass
values, obtained a mass resolution of 8mjm '" 2 x 10- 6 , or Mj /'),M = 150,000
and masses differing from the litterature by a few ke V. A dedicated detector system
based on MCP stacks, with an overall time resolution of 85 ps, was also developed.
Thus, this very promising method, which has as main advantages the short time
and the good precision, using about 50 turns, can measure masses of very short
lived hot nuclei with a precision of about 10-6 .
Acknowledgements
This talk was based on a review article written in collaboration with W. Mittig
(Ganil) and N. A. Orr (LPC Caen), their contribution is gratefully acknowledged.
My participation at the conference was supported by FAPESP.
References
1. Thompson, J. J., Philos. Mag. VI 24 (1999), 209, 668.
2. Aston, F. W, Philos. Mag. 45 (1923), 934; Proc. Roy. Soc. London Ser. A 115 (1927), 487.
3. Von Weizsacker, C. F., Z. Phys. 86 (1935), 431.
56 A. LEPINE-SZILY
J. M. PEARSON
Departement de Physique, Universite de Montreal, Montreal (Quebec), H3C 3J7 Canada
Abstract. After stressing how well the purely macroscopic 1935 mass formula of von Weizsacker
works, we discuss the general problem of deriving nuclear masses from basic nucleonic interactions.
We then describe the very recent Skyrme-Hartree-Fock-BCS mass formula of Goriely et al., the
first and only to be entirely microscopic. We conclude by stressing how much more work has to be
done before reliable extrapolations can be made from the mass data out to the highly neutron-rich
part of the nuclear chart where the r-process of nucleosynthesis takes place.
1. Introduction
The interest of nuclear masses lies in the fact that the mass M nue (N, Z) of a nucleus
with N neutrons (n) and Z protons (p) is measurably different from the sum of the
masses of the free nucleons, whence a direct determination of the internal energy
Enuc (the negative of the binding energy) of the nucleus is possible,
(1)
worked very well, but there are serious ambiguities arising from the decision,
dictated by computational limitations, to ignore the common origin that the two
aspects must have in the basic nucleonic interactions.
Actually, the macro-micro approach is most easily understood in terms of the
purely microscopic approach of having both OM and SM features emerge on an
equal footing from the common starting point of nucleonic interactions, either
realistic or effective. We shall thus violate the historical order of development and
first outline, in Section 3, the latter approach; in the same section we also present
the only mass formula developed so far that can be said to be purely microscopic,
being based essentially on the Hartree-Fock (HF) method. Section 4 then describes
the older macro-micro approach, including variants in which the macroscopic part
is evaluated semi-classically rather than by the OM. Finally, in Section 5, we look
to the future, showing how much work remains to be done.
It is appropriate at this point to recall just why this sort of work continues to be
important. In the first place, measuring the masses of more and more nuclei further
and further away from the stability line with ever increasing precision is surely
one way of testing our theories of nuclear structure. But much of the interest in
constructing theoretically sound mass formulas that give precision fits to the mass
(and other) data lies in the possibility thereby offered of being able to make reliable
extrapolations away from the data out to the highly n-rich nuclei that playa crucial
role in the r-process of nucleosynthesis but which are so unstable that there is no
prospect of being able to measure them in the laboratory [3]. Mass formulas also
permit an even greater leap from the mass data: the extrapolation out to infinite
nuclear matter (INM), which also is of astrophysical interest.
3e 2 2 -4/3
Enuc _
- - - avol +asfA
-1/3
+ -Z A + ( asym +assA -1/3) I 2 , (2)
A 5~
where A = N + Z. This is just the simplified form given in 1936 by Bethe and
Bacher [4], with the addition of the surface-symmetry term (ass) introduced in 1966
by Myers and Swiatecki [5]. The two leading terms correspond to the OM, which
was inspired in large part by the observation that the radius R of any nucleus
(N, Z) is given by R ::::: roA 1/3, where ro is a constant. The roughly constant
density thereby implied is what one would expect in view of the finite range of
nuclear forces, along with their strong short-range repulsion, and is referred to as
the saturation property. With this picture of nuclei consisting simply of differently
sized pieces of nuclear matter we expect the energy per nucleon to be constant,
whence the term avol in Equation (2); the term in asf is then simply a correc-
tion for surface tension. However, the Coulomb force, being of infinite range, is
not saturated, and must be taken into account explicitly: this is the third term of
THE QUEST FOR A MICROSCOPIC NUCLEAR MASS FORMULA 61
Equation (2). One must also allow the specifically nuclear terms, both volume and
surface, to depend not only on the total number of nucleons but also on the n-p
composition. To a good approximation this dependence can be expected to be an
even function of I = (N - Z)/ A, since Mn =:::: Mp and nuclear forces are more or
less charge-symmetric.
In the limit A ---+ 00 the energy per nucleon becomes infinite and positive,
because of the Coulomb term, unless we are dealing with a pure neutron system,
I = 1. However, if we imagine the Coulomb force to be switched off we find
a finite energy per nucleon, ayol + a sym l 2 . This corresponds to the equilibrium
energy of INM, which is, of course, a hypothetical medium, since the Coulomb
force cannot be switched off. However, a similar situation is realized in neutron
stars by the presence of electrons, which neutralize the small fraction of protons
present.
We have fitted the mass formula (2) to the 1995 compilation of Audi and Wap-
stra [6], finding aYol = -15.65 MeV, a sym = 27.72 MeV, asf = 17.63 MeV,
ass = -25.60 MeV, and ro = 1.233 fm, with an rms error of 3.02 MeV, which
means that about 98% of the variation in the binding energy is being reproduced.
This is remarkable, given that nearly 2000 data points are being fitted by 5 para-
meters. Nevertheless, the graphs of the residual errors (Figures la,b) show clearly
the importance of the shell effects omitted in Equation (2); less apparent is the
neglected even-odd pairing effect. Figure 2 shows how closely the drip lines pre-
dicted by the completely microscopic mass formula HFBCS-l (Section 3D) are
reproduced by the mass formula (2). In Figure 3 we see how well the limits of the
zones of instability with respect to ,B-delayed nucleon emission lying just inside
the drip lines are described by this same mass formula, while Figure 4 makes a
similar comparison for the region of a-instability.
As a final example of what the semi-empirical formula can do, let us generalize
it to include gravity. Since this has a non-saturating character formally identical to
that of the Coulomb force, though with opposite sign, Equation (2) becomes
a
yol
+ a,sf A- I / 3 + ~{e2(1
5ro 4
_1)2 _ GM2}A2/3
A
+ (a sym + ass A -l!3)12 . (3)
For normal nuclei the gravitational correction will be utterly negligible, so let us
consider very large values of A, and limit ourselves to systems consisting only of
neutrons, I = 1, since otherwise the Coulomb repulsion would diverge. Using the
above values of the parameters, we find that such a system will become gravita-
tionally bound for A > 3.4 X 1056 , i.e., for a mass in excess of 5.6 x 1029 kg,
which is within an order of magnitude of the mass of a typical neutron star. Since
furthermore the above value of ro implies that the density of a nucleus is within an
order of magnitude that of a neutron star, we see that the semi-empirical formula
can actually encompass the picture of a neutron star as consisting of an enormous
gravitationally bound nucleus.
62 J. M. PEARSON
15
S. E. Mass Formula
:>
v 10
::0:
5
v,
Ul
~
0
::0:
0.
-5
><
v ~~.
::0:
-10
... ......
(a)
15
S. E.Mass Formula
:>Q.I 10
::0:
5
v,
~
Ul
::0: 0
...
0. -5
><
v
::0:
-10
20 40 60 80 100 120
Z
(b)
Figure 1. (a): Errors in mass fit to semi-empirical mass formula of Equation (2), as function
of N. (b): Errors in mass fit to semi-empirical mass formula of Equation (2), as function of Z.
3. Microscopic approaches
The ideal mass fonnula would be one in which the binding energies of all nuclei
were derived from the basic nucleonic interactions. Work on these lines was initi-
ated in the fifties by Brueckner and co-workers, and since then a tremendous effort
THE QUEST FOR A MICROSCOPIC NUCLEAR MASS FORMULA 63
120
100
80
N 60
O ~ ______~______~________~______~______~__~
o 50 100 150 200 250
N
Figure 2. Drip lines according to semi-empirical mass formula of Equation (2) and mass
formula HFBCS-J .
120 ,-'
1 00
80
:
N 60
40
. .'
Limits 01 delayed N emission (s.-e.)
20
Drip lines (HFBCS-l)
o~ ______ ~~ ______ ~ ______ ~ ________ ~ ______ ~ __ ~
has been expended in pursuing different possible approaches. Two main types of
calculation can be discerned, as follows.
(a) Non-relativistic methods. These calculations assume that the nucleus is de-
scribed by a non-relativistic Schr6dinger equation
(4)
64 1. M. PEARSON
120
100
80
N 60
.'
40
- - Limits of alpha-Instability (s.-9.)
20
,.,
Drip lines (HFBC5-1)
0
0 50 100 150 200 250
N
Figure 4. Dots represent nuclei that are unstable with respect to a-emission according to
mass formula HFBCS-l; the solid line represents inner limit of zone of such nuclei as given
by semi-empirical mass formula of Equation (2).
where
H =-
/i 2
2M L \1; + L Vij + L
. i>j i>j>k
"ijk. (5)
Here Vij and Vijk are potentials representing the two-nucleon (N-N) and three-
nucleon interactions, respectively, as determined by the scattering and bound-
state properties of such systems, and by meson-exchange theories. To calculate
the properties of complex nuclear systems in this non-relativistic framework
both variational methods and Brueckner-Bethe-Goldstone theory have been
used: guides to the literature can be found in [7] for the former, and [8, 9] for
the latter.
(b) Brueckner-Dirac methods. Here the nucleons are treated fully relativistically,
being represented by Dirac spinors. The degrees of freedom associated with
the exchange of the mesons responsible for the nucleonic interactions are taken
into account explicitly; moreover, the meson parameters are fitted to the N-N
scattering data and measured meson properties. When fully developed this
approach should be at least as reliable as the non-relativistic methods. (For
a review see [l0].)
All of these different methods based on realistic nucleonic interactions are hor-
rendously complicated, and the only system for which quantitatively satisfactory
results have been obtained is the relatively simple case of INM. Even if this success
shows that these theories are fundamentally correct, the fact remains that the few
calculations that have been performed on finite nuclei give results whose accuracy
is inacceptably poor. Nevertheless, many-body theories based on realistic nucle-
onic forces will be able to serve as a qualitative guide in tying down some of the
THE QUEST FOR A MICROSCOPIC NUCLEAR MASS FORMULA 65
1'l2
H eff = -2M~
- "" ,;2 + ""
~ I
eff
vIJ' (6)
. i>j
in which V~jff is some effective N-N potential that does not have to fit the N-N data.
One way in which this force could be determined, particularly appropriate to the
present context of nuclear masses, would be to optimize the fit of the expectation
values EHF = (<I> IHeffl <1» to all the measured values of Enuc.
The method proceeds by minimizing EHF with respect to arbitrary variations in
the unknown s.p. functions tPi (Xi), which are then given as eigensolutions to a s.p.
Schrodinger equation
( _~V2
2M + U)A-.. = '1'1
E'A-..
I'l'l ,
(7)
and on an equal footing in this picture, so that a much more unified approach to
the mass formula is offered than is possible with the hybrid macro-micro methods
(Section 4).
A particularly suitable form of effective force is the Skyrme form [12]:
i
+ li 2 Wo«1; + (1 j)'P;j x 8(rij)pij. (8)
All terms here are fonnally of zero range, although the momentum dependence of
the tl and t2 terms simulates a finite range.
With Skyrme forces the s.p. equation (7) takes the form
in which i labels all quantum numbers, and q denotes n or p. All the field tenns
are now local, essentially because one can introduce a position-dependent effective
mass M;. There are two such quantities, corresponding to the two types of charge,
which can be expressed at the INM equilibrium density in terms of an isoscalar
and an isovector effective mass, M; and M:, respectively; these are unique com-
binations of the Skyrme-force parameters. Expressions for all quantities appearing
in this equation, and for E HF , can be found in [13].
Most nuclear HF calculations that have been perfonned use Skyrme forces, al-
though the Gogny group uses forces that are explicitly finite-range [14]. While the
latter may be regarded as more realistic, the essential non-locality of the s.p. fields
complicates the calculations considerably. Until very recently no HF effective force
had been fitted to more than ten or so nuclei, all spherical, presumably because of
the computer-time limitations that arose in the past with defonned nuclei. However,
it is now possible to fit to the masses of all nuclei (Section 3.4).
3.3. CORRELATIONS
Even when the Slater determinant <P satisfies the HF equations (7), it can never be
identical to the exact nuclear wave function \II. Thus we must expect nuclear prop-
erties to show features that cannot be accounted for within the HF framework; such
irreducible deviations from the mean-field picture are referred to as correlations.
T =0 pairing between neutrons and protons is responsible (see [19] and references
quoted therein), but no systematic study has been made so far.
4. Macroscopic-microscopic approaches
The way in which shell corrections are grafted on to the DM can be understood as
an approximation to the HF method. If 15 is any smooth diagonal approximation to
the HF density matrix PHP then an expansion of E [PHF] in powers of op == PHP - 15
leads to the Strutinsky theorem (see [17, p. 93]),
where we have neglected pairing and terms 0 (op)2. Also h == h [15] is the smoothed
s.p. Hamiltonian approximating the exact s.p. Hamiltonian h[PHP]' and the Ei are
the corresponding eigenvalues, with the sum going over all occupied states: all the
shell-model fluctuations arise in this sum.
The choice of 15 is arbitrary, and this approach originally took it to correspond
to the DM. As for the s.p. Hamiltonian h, this must have the same form as in
Equation (7), whence a choice must be made for the field U. There is no un-
ambiguous way of doing this for the DM choice of 15, and in practice one has
been guided by considerations of plausibility and convenience. Once this choice
has been made the determination of the Ei is straightforward, and there remains
only the last term to calculate. Formally, this corresponds to a smoothed version
of the sum in the preceding term, tr hp == LiEi' and it is evaluated by the various
'Strutinsky smoothing procedures' that have been devised (see, for example, [16,
Section 12.4]). We have here a second source of uncertainty, especially for nuclei
close to the n-drip line [27].
As for pairing corrections, BCS calculations are straightforward in this ap-
proach, once the s.p. states have been calculated.
The most recent and elaborate mass formula of this type is the 'finite-range
droplet model' (FRDM) of [2]. Here the simple DM of Equation (2) is replaced
by the more sophisticated 'droplet' model, but otherwise the model conforms to
the above scheme. The data fit has an rms error of 0.689 Me V for the above set
of 1888 nuclei, which is somewhat better than for HFBCS-l. However, in addi-
tion to the usual uncertainty associated with Strutinsky smoothing, the inevitable
uncertainty associated with the choice of s.p. field is exacerbated by the fact that
the symmetry coefficient a sym (J in their notation) corresponding to this field has
the value of 35 MeV, while this same coefficient takes the value of 32.73 MeV in
the macroscopic part. We do not know whether the quality of the fit would have
deteriorated if it had been reiterated until a sym took the same values in both parts.
Moreover, it is difficult to assess the impact on the reliability of the extrapola-
tions.
• •
3.5 FROM
:;-
Ql 3 HFBCS-1
6
c.
c3 2.5
Qi
.c
C/) 2
1.5
of these mass fonnulas is to be believed? The ultimate mass fonnula, i.e., the one
that gets everything right, must be microscopic, but one cannot pretend that with
HFBCS-l we have reached the end of history, as far as mass fonnulas go, with
nothing more to do than refit the parameters every time Audi and Wapstra crank
out another evaluation. On the contrary, I believe that the remaining uncertainties
in the extrapolations of HFBCS-l out to the n-drip line are at least as great as the
differences between HFBCS-l and the two modem macro-micro fonnulas. Here
are some of the outstanding theoretical problems that must be addressed before we
can really start to extrapolate out to the n-drip line with any confidence.
(1) Effective mass. Our data fits have confinned the Bernard-Giai renonnaliza-
tion of M* to take account of surface modes [24]. But their calculations were made
only in the region of the stability line, so it is really only the isoscalar component
M; / M whose (V-mass value is well tied down to close to 1, and it is essential that
they be extended towards the n-drip line in order to detennine the (V-mass value of
M~ / M. It should be stressed that there is no justification at the present time for our
assumption that M; = M~ as far as (V-mass values are concerned; all that can be
said is that there are no data against it.
(2) Pairing. Firstly, if we are to have any confidence in the extrapolations out
to the n-drip line, it is essential that the HF-BCS method used here be replaced
by HFB. Calculations performed with the Skynne force SkP [38, 39] on spherical
nuclei indicate that the effect of replacing HF-BCS by HFB will be to quench
shell effects at the n-drip line, but there have been no extensive HFB calculations
for arbitrary defonnations. Actually, a somewhat improvised attempt to modify the
original ETFSI calculations [18] for Bogolyubov quenching has already been made
[40], but a complete HFB calculation would be preferable. However, the extent of
the quenching depends on the s.p. spectra, and thus on the value of M* / M assumed
for highly n-rich nuclei, and it must be stressed that there is as yet no finn support
for the value of 1 taken for the force SkP. Thus before undertaking any extensive
HFB calculations it will be necessary to resolve the problem of M~ / M.
Beyond that, attention has to be paid to the choice of the pairing interaction,
noting that it does not have to be identical to the HF force. Should it be density-
dependent? Should it depend on the gradient of the density in order to confine it
to the surface? These are all open questions, and one welcomes the recent efforts
to relate the pairing force, hitherto treated on a purely phenomenological basis, to
more basic nuclear processes, as in [41-43].
Experiment. There will always be a need for more and more data further and
further from the stability line, not only to tie down the parameters of the various
theories, but also to weed out the weakest of them. Of particular importance here is
the need to shed light on the M~ / M problem, and to find firm evidence for the ex-
istence of shell quenching. In both these connections I would stress the importance
not only of mass data, but particularly of s.p. spectra. And again, while the n-rich
side of the stability line is the one of most direct interest for the r-process, data on
the p-rich side can be just as invaluable in tying down theories. Finally, to conclude
THE QUEST FOR A MICROSCOPIC NUCLEAR MASS FORMULA 73
on a completely different note, let me point out that it would be very useful indeed
if one could measure neutron-skin thicknesses, since these provide an independent
determination of the symmetry coefficient a sym .
Acknowledgements
I wish to express my gratitude to F. Tondeur and the many others in Montreal and
Brussels with whom it has been my pleasure to collaborate over the years.
References
1. v. Weizsacker. C. F., Zeit.! Phys. 96 (1935), 431.
2. Moller, P., Nix, J. R., Myers, W. D. and Swiatecki, W. J., At. Data NucZ. Data Tables 59 (1995),
185.
3. Goriely, S., this issue, 105.
4. Bethe, H. A. and Bacher, R. F., Rev. Mod. Phys. 8 (1936),82.
5. Myers, W. D. and Swiatecki, W. D., NucZ. Phys. 81 (1966), 1.
6. Audi, G. and Wapstra, A. H., Nucl. Phys. A 595 (1995), 409.
7. Akmal, A., Pandharipande, V. R. and Ravenhall, D. G., Phys. Rev. C 58 (1998), 1804.
8. Baldo, M., Giansiracusa, G., Lombardo, U. and Song, H. Q., Phys. Lett. B 473 (2000), 1.
9. Lejeune, A., Lombardo, U. and Zuo, w., Phys. Lett. B 477 (2000), 45.
10. Machleidt, R., Adv. Nucl. Phys. 19 (1989), 189.
11. Gomes, L. c., Walecka, J. D. and Weisskopf, V. F., Ann. Phys. 3 (1958), 241.
12. Vautherin, D. and Brink, D. M., Phys. Rev. C 5 (1972) 626.
13. Tondeur, F., Goriely, S., Pearson, J. M. and Onsi, M., Phys. Rev. C 62 (2000), 024308.
14. Decharge, J. and Gogny, D., Phys. Rev. C 21 (1980), 1568.
15. Lalazissis, G. A., Raman, S. and Ring, P., At. Data Nucl. Data Tables 71 (1999), 1.
16. Preston, M. A. and Bhaduri, R. K., Structure of the Nucleus, Addison-Wesley, Reading, MA,
1975.
17. Ring, P. and Schuck, P., The Nuclear Many-Body Problem, Springer, New York, 1980.
18. Aboussir, Y., Pearson, J. M., Dutta, A. K. and Tondeur, F., At. Data Nucl. Data Tables 61
(1995), 127.
19. Satula, w., Dean, D. J., Gary, J., Mizutori, S. and Nazarewicz, w., Phys. Lett. B 407 (1997),
103.
20. Goriely, S., Tondeur, F. and Pearson, J. M., At. Data Nucl. Data Tables 77 (2001), 311.
21. Myers, W. D. and Swiatecki, W. 1., NucZ. Phys. A 612 (1997), 249.
22. Nadjakov, E. G., Marinova, K P. and Gangrsky, Yu. P., At. Data Nucl. Data Tables 56 (1994),
134.
23. Barranco, M. and Treiner, J., NucZ. Phys. A 351 (1981), 269.
24. Bernard, V. and Van Giai, Nguyen: Nucl. Phys. A 348 (1980), 75.
25. Chabanat, E., Bonche, P., Haensel, P., Meyer, J. and Schaeffer, R., Nucl. Phys. A 627 (1997),
710.
26. Chabanat, E., Bonche, P., Haensel, P., Meyer, 1. and Schaeffer, R., Nucl. Phys. A 635 (1998),
231.
27. Nazarewicz, w., Werner, T. R. and Dobaczewski, J., Phys. Rev. C 50 (1994),2860.
28. Dutta, A. K, Arcoragi, J.-P., Pearson, 1. M., Behrman, R. and Tondeur, F., NucZ. Phys. A 458
(1986), 77.
29. Tondeur, F., Dutta, A. K., Pearson, J. M. and Behrman, R., NucZ. Phys. A 470 (1987), 93.
74 J. M. PEARSON
30. Pearson, J. M., Aboussir, Y., Dutta, A. K., Nayak, R. C., Farine, M. and Tondeur, F., Nucl. Phys.
A 528 (1991), 1.
31. Aboussir, Y., Pearson, J. M., Dutta, A. K. and Tondeur, F., Nucl. Phys. A 549 (1992),155.
32. Goriely, S., In: S. Wender (ed.), Proceedings of the 10th International Symposium on Capture
Gamma-Ray Spectroscopy and Related Topics, Santa Fe, New Mexico, USA, September 1999;
A.I.P. Proceedings 529 (2000),287.
33. Mamdouh, A., Pearson, J. M., Rayet, M. and Tondeur, F., Nucl. Phys. A 644 (1998), 389.
34. Mamdouh, A., Pearson, J. M., Rayet, M. and Tondeur, F., Nucl. Phys. A 679 (2001), 337.
35. Myers, W. D. and Swiatecki, W. J., Nucl. Phys. A 601 (1996), 141.
36. Onsi, M., Nayak, R. c., Pearson, J. M., Freyer, H. and Stocker, w., Phys. Rev. C 55 (1997),
3166.
37. Nayak, R. C. and Pearson, J. M., Phys. Rev. C 58 (1998), 878.
38. Dobaczewski, J., Hamamoto, I., Nazarewicz, W. and Sheikh, J. A., Phys. Rev. Lett. 72 (1994),
981.
39. Dobaczewski, J., Nazarewicz, W. and Werner, T. R., Phys. Scr. T 56 (1995), 15.
40. Pearson, J. M., Nayak, R. C. and Goriely, S., Phys. Lett. B 387 (1996), 455.
41. Barranco, F., Broglia, R. A., Gori, G., Vigezzi, E., Bortignon, P. F. and Terasaki, J., Phys. Rev.
Lett. 83 (1999),2147.
42. Garrido, E., Sarriguren, P., Moya de Guerra, E. and Schuck, P., Phys. Rev. C 60 (1999), 064312.
43. Bertsch, G. F. and Esbensen, H., Ann. Phys. 209 (1991), 327.
Hyperjine Interactions 132: 75-103,2001. 75
© 2001 Kluwer Academic Publishers.
Abstract. Atomic binding energies are calculated at utmost precision. A report on the current status
of Lamb-shift predictions for hydrogenlike ions, including all quantum electrodynamical corrections
to first and second order in the fine structure constant Ci is presented. All relevant nuclear effects are
taken into account. High-precision calculations for the Lamb shift in hydrogen are presented. The
hyperfine structure splitting and the g factor of a bound electron in the strong electromagnetic field
of a heavy nucleus is considered. Special emphasis is also put on parity violation effects in atomic
systems. For all systems possible investigations beyond precision tests of quantum electrodynamics
are considered.
Key words: atomic binding energy, electron g-factor, highly-charged ions, Lamb shift, parity viol-
ation, quantum electrodynamics.
1. Introduction
Quantum Electrodynamics (QED) can claim to be one of the most precisely tested
theories of physics. One impressive example is the g factor of the free electron
which is known today as
gfree = 2+2 x 1159652188.4(4.3) x 10- 12 (Experiment [l])
gfree = 2+2 x 1159652216.0(1.2)(67.8) X 10- 12 (Theory [2]).
(The second error indicated in the second line is due to the value of ex employed
in this calculation [3].) Similar precisions are nowadays obtained in systems like
positronium or the Lamb shift in hydrogen where insufficiently known nuclear
parameters limit the accuracy of theoretical predictions [4, 5]. The electric field
involved in all these systems is rather low, however, compared to the strongest elec-
tromagnetic fields accesible to experimental investigation today. In atomic systems,
these strong fields are obtained by stripping all but one electron from a heavy atom,
* Dedicated to Prof. A. Wapstra on the occasion of his 78th birthday.
76 G. SOFF ET AL.
~ 10 17. -_ _ __ _ _ _ __ _ _ _ _-,
E
o
~ 10 16
iii
.::: 10 15
.. '
............ . ....-
10 13
/ ../. /... f
....<".:. / . '
:' .j
:' . /
1/
:!
10 10 ff
109~:~~~~~~~~~~~~~~
10 20 30 40 50 60 70 80 90
nuclear charge Z
Figure 1. Expectation value of the electric field strength for the lowest-lying states of a
hydrogenlike atom in the range Z = 1-92.
e.g., lead or uranium. The single electron is bound in a system similar to hydrogen,
and the whole system is therefore called hydrogen-like system. The expectation
value of the electric field strength in these systems is depicted in Figure 1. The
field strength at the nuclear surface is even higher. For example, at the surface of a
uranium nucleus, lEI ~ 2 . 10 19 V/cm. This is only a factor of 2 less than the field
strength in superheavy systems with Z :? 170 where spontaneous pair production
is predicted to take place if the total charge can be confined in a sufficiently small
volume for a sufficiently long time [6--8]. It seems evident that in such strong fields
'normal' atomic physics - valid for a hydrogen atom where the field probed by
the electron is six orders of magnitudes smaller - may be questioned. A precise
knowledge about the validity of QED in strong external fields is also very promis-
ing for the detection of new physics beyond QED [9]. Thus it is a primary goal to
explore the behaviour of electrons in some of the strongest electromagnetic fields
accessible to experimental investigation. Additional evidence can be obtained from
muonic atoms because the muon is localized much closer to the nucleus and probes
therefore even stronger fields. On the other hand, it is also more sensitive to nuclear
effects that are theoretically less well known.
In atoms, the coupling of the electron with the binding field of the nucleus is
determined by the coupling constant Za where Z is the nuclear charge number and
a ~ 1/137 is the fine structure constant. In highly charged ions like uranium, the
coupling constant is no longer a small parameter but amounts to Za ~ 0.6 and
therefore a perturbation expansion in Za becomes meaningless. In contrast, most
approaches for the calculation of QED effects in low-Z atoms like hydrogen still
EFFECTS OF QED 77
employ such an expansion. In heavy ions, the interaction with the Coulomb field
of the nucleus has to be considered to all orders in Za. Bound-state QED provides
the relativistic description of an electron in highly charged atomic systems. To test
the predictions of this theory with utmost precision is a major challange in todays
atomic physics. This review is focused on the recent developments in this exciting
area. The examples given here will indicate the predictive power of QED in 'simple
systems' where many-body effects do not have to be considered except for nuclear
structure calculations. The paper is organized as follows:
In Section 2 we will consider the present status in the evaluation of all QED
corrections to the Lamb shift in hydrogen-like uranium and lead and compare theo-
retical and experimental results. The general theoretical and experimental precision
of Lamb shift measurements will be discussed, and an interesting future direction
of research will be mentioned. In Section 3 the hyperfine splitting of atomic levels
is considered. We will report on the present status of the hyperfine splitting effect
in hydrogen- and lithiumlike ions and we will compare the theoretical results with
the available experimental values. Another quantity accessible for high-precision
experiments is the g factor of a bound electron which is also in the scope of
Section 3.
In Section 4 parity violating effets in atomic systems will be considered. The
advantages of investigating these effects in highly charged ions will be underlined
and a new experiment for detecting a parity violation in heliumlike europium and
gadolinium is proposed.
2. Lamb-shift calculations
The term Lamb shift refers to the difference between the Dirac energy eigenvalue
of a single atomic level and its actual value which is shifted due to nuclear and
QED effects. The binding energy of an atomic level is one of the quantities best
to measure and therefore QED effects are well detectable. Precise measurements
in hydrogenlike heavy ions nowadays concentrate on uranium [10] where the best
measurement recently was achieved by Stohlker et al. [11, 12]. Also in gold [13],
lead [14], and bismuth [15] the ground state Lamb shifts were investigated.
The major contributions to the IS 1/ 2 Lamb shift are the radiative corrections of
order a (Figure 2), and also the effect of the nuclear size which accounts for the
SE) VP)
Figure 2. The QED corrections of order Ct, self energy (SE) and vacuum polarization (VP).
The double solid line denotes the bound electron and the wavy line indicates the photon.
78 G. SOFF ET AL.
finite extension of the nuclear charge distribution. Therefore the potential that is
present in the Dirac equation changes and the wave functions and energy eigen-
values are slightly altered. Examples can be found, e.g., in [16]. For uranium,
even the different models for the shape of the nucleus at a fixed size «(r2) 1/2 =
5.860 ± 0.002 fm) cause differences in the K-shell binding energy of nearly one eY.
The two major QED corrections are self energy and vacuum polarization. The
self energy (SE) is the result of the emission and reabsorption of a photon by
an electron. For high-Z ions, it has been calculated employing many different
methods, beginning from the pioneering elaborations of Brown et al. [17] and
Desiderio and Johnson [18] to the more accurate approach developed by Mohr
[19, 20]. Blundell and Snyderman presented an alternative approach [21, 22] of
calculating the first-order self energy also in a non-Coulomb potential.
The first-order vacuum polarization (VP) correction accounts for the interaction
of the electron with a virtual electron-positron pair in the field of the nucleus. It is
commonly divided into two major parts, the charge divergent but nowadays well-
known Uehling part [23-26] and the finite Wichmann-Kroll part [27] which was
evaluated by Gyulassi [28] and later with high precision by Soff and Mohr [29],
and by another approach also by Fainshtein et ai. [30]. Persson et ai. could even
improve the numerical accuracy [31].
The finite nuclear size has also an effect on the radiative corrections. For ura-
nium, the difference between the self-energy correction with and without consid-
ering the nuclear size amounts to more than 1% [32] and therefore the influence of
the nuclear size on the radiative corrections should be taken into consideration at
least for heavy systems with Z > 50. An elaborated investigation on this topic for
both self energy and vacuum polarization is given in [33]. The results of all these
contributions for hydrogen-like uranium and lead are depicted in Table I, together
with the contributions of order a 2 and the recoil contributions.
All present experimental results can be theoretically well explained by con-
sidering the effect of the finite nuclear size and the QED corrections of order a
as discussed above. However, at the GSI in Darmstadt, a precision of about 1
eV in measurements of the ground-state Lamb shift seems likely in the near fu-
ture [12]. At this level, QED corrections of order a 2 have to be taken into account.
The complete set of radiative corrections is displayed in Figures 3(a)-(k). These
diagrams are naturally divided into separately gauge invariant subsets: SESE (a),
(b), (c); VPVP (d); VPVP (e), (f); SEVP (g), (h), (i); and S(VP)E (k). In addition
to a complicated renormalization scheme, the calculation of higher-order QED
corrections also requires the application of special numerical methods to perform
multiple summations over the complete eigenfunction spectrum of the Dirac equa-
tion for a bound electron. Only recently this task was completed for the last missing
contributions, the so-called reducible part of the SESE (a) diagram together with
the SESE (b) and (c) diagrams. Now all of these corrections have been numeri-
cally calculated for the ground state of hydrogenlike uranium and lead [34]. The
so-called irreducible contribution SESE (a) can be separately renormalized and
EFFECTS OF QED 79
Table I. Lamb-shift contribution for the ground state of 238U91+ and 208Pb 81 +
(in eV), including the full recoil contribution. The finite nuclear-size correction
for uranium is calculated for a Fermi distribution with (r2) 1/2 = 5.860± 0.002
fm. The finite nuclear-size correction for lead is calculated for a Fermi distri-
bution with (r2) 1/2 = 5.505 ± 0.001 fm. Known error-margins are linearly
added
evaluated. It denotes that part of the SESE (a) diagram where the energy of the
intermediate electron states between the two self-energy loops is different from the
energy of the state under consideration. This contribution has been first calculated
by Mitrushenkov et al. [35] for the nuclear charge numbers Z = 70, 80, 90, and 92,
and recently for arbitrary values of Z in [36, 37]. Although the obtained results are
in fair agreement for high Z, a discrepancy between these in [36] and [37] has been
observed for the case of low and intermediate Z values. Only the results of [37]
agree with an analytical expansion based on powers and logarithms of Za [38].
The reason for this discrepancy is under current investigation and two recent works
devoted to this subject again end up with different conclusions: the calculation
of Yerokhin [39] supports [36] whereas Manohar and Stewart [40] find the same
logarithmic term as Karshenboim in [38]. It would be not the first time however,
that modem high- Z methods prove to be more successful also for low Z than the
conventional Za expansion [41]. For the high-Z region, there is no doubt about the
results presented in Table I.
The VPVP contributions (e) and (f) (also known as Kallen-Sabry corrections)
have been investigated in Uehling approximation [42, 43]. Calculating the domi-
nant Uehling part of the lowest order VP correction results in a precision of about
80 G. SOFF ET AL.
~ ~ cD
SESE
a) b) c)
t:2~~
VPVP
d) e) f)
~1-0 J-<'
g) h) i)
SEVP
t:o
k)
S(VP)E
5% for the ground state of hydrogenlike uranium and lead ions. In the Uehling
approximation one restricts to the first term in a Za expansion of the bound electron
propagator in the electron loop. This also corresponds to the expansion of the bound
propagator in terms of the nuclear potential. Recently the VPVP (e) contribution
has been determined to all orders in Za [44]. In this case the inaccuracy of the
Uehling approximation turned out to be about 25%. The VPVP (d) contribution
was considered in [31] and tabulated in [45]. The SEVP (g), (h), and (i) contribu-
tions were evaluated in [46] employing the Uehling approximation and in the exact
form in 147]. The inaccuracy of Uehling approximation for hydrogenlike uranium
amounts to only 2% in this case. The S(VP)E (k) contribution has been calculated
only in Uehling approximation up to now [47, 48]. A recent overview about all
first- and second-order QED corrections is presented in [49].
In addition to the QED corrections of order a 2 , the internal structure ('polar-
izability') of the nucleus and the nuclear recoil effect cause additional binding-
energy corrections of the same order of magnitude as these QED corrections. The
nuclear-polarization contribution was derived and evaluated in the framework of
an effective photon propagator in [50-52]. The recoil effect accounts for a non-
infinitely heavy nucleus and takes into account its movement. Nonrelativistically
this can be considered by the reduced mass of the electron, an approximation which
is more than 50% wrong for heavy hydrogenlike systems like uranium. The com-
EFFECTS OF QED 81
p1ete effect to all orders in Za was calculated in [53, 54] for point-like nuclei and
in [55] for extended nuclei. The extension of the nuclear charge distribution reduces
the effect by about 10% in the case of U 91 +. In Table I the results are compiled
for uranium and lead together with all QED corrections of order a and a 2 and
also with the nuclear effects. We point out that the compilation in this table does
not follow the convention of [56] who do not include the nonrelativistic reduced-
mass correction in the Lamb shift because it does not contribute to the 'classical'
2S 1j2 -2plj2 splitting in low-Z atoms. For high-Z systems, however, a complete
recoil correction inherently includes the nonrelativistic reduced-mass correction
and there is no sense to consider this contribution separately [57]. All our present
results allow for high-precision tests of QED in the strong field of the nucleus that
are expected to be experimentally available in the near future.
The high-precision calculations for high-Z systems also influence the predictive
power in the low-Z region where higher precision is gained by including more
and more terms of an expansion in the interaction with the binding field, i.e., in
powers of Za and some logarithmic terms. For an overview about these 'analytical'
techniques, we refer to [58]. The precise determination of radiative corrections
by numerical calculations even for low nuclear charge numbers Z is currently an
emerging field [59]. Here, we will focus on the one-photon self energy, which
constitutes the dominant contribution to the Lamb shift in hydrogen by two orders
of magnitude.
One of the calculational challenges in the regime of low nuclear charge are
numerical cancellations. In order to understand the origin of the numerical can-
cellations it is necessary to consider the renormalization of the self energy. The
renormalization procedure postulates that the self energy is essentially the effect
on the bound electron due to the self interaction with its own radiation field, minus
the same effect on a free electron. There it is absorbed in the mass of the electron
and therefore not observable. The self energy of the bound electron is the difference
of two large quantities. Terms associated with renormalization counterterms are of
order unity in the Za-expansion, whereas the residual effect is of order (Za)4.
This corresponds to a loss of only one significant figure at Z = 92, but roughly 9
significant digits at Z = 1. Accurate numerical methods have to be employed, and
the convergence of certain angular momentum expansions has to be accelerated by
powerful numerical algorithms (see, e.g., [60]) which reduce the computation time
by roughly three orders of magnitUde.
We start our discussion here from the well-known (regularized and renormal-
ized) expression for the one-photon self energy ~ESE,
.
hm
A--+oo
{-1. e2Re 1. f
-dw
e 2JT
3
-d- k
3 DfJ-V
(2JT)
(2)
k ,A
/).m = -a
rr
(3-
4
InA 2 3)
+ - (.B).
8
(3)
D (k2 A) - _( (4)
+i E +i E
g{tv _ g{tv )
{tV , - k2 k2 - A 2 •
The contour e used in the numerical calculation is not the Feynman contour but
represents the mathematically equivalent contour depicted in Figure 4. The ana-
lytic structure of the propagators plays an important role in the evaluation, and
the choice of the contour has to reflect this structure which is indicated by the
branch cuts in Figure 4, while at the same time providing a convenient basis for the
numerical evaluations (as discussed below). The energy variable z in Equation (2)
assumes the value
z = En - tv, (5)
where En is the Dirac energy ofthe atomic state, and tv denotes the complex-valued
energy of the virtual photon. It is understood that the limit A ---+ 00 is taken after
all integrals in Equation (1) are evaluated.
The contour e naturally leads to a separation of the two scales in the self-
energy problem: the atomic energy scale (Za)2 m and the relativistic electron mass
Im(w)
Gil
GL
En 2m Re(w)
GL
Gil
Figure 4. Integration contour C? for the integration over the energy w = En - z of the virtual
photon. The contour C? consists of the low-energy contour CL and the high-energy contour
CH. Lines shown displaced directly below and above the real axis denote branch cuts from the
photon and electron propagator. Crosses denote poles originating from the discrete spectrum
of the electron propagator.
EFFECTS OF QED 83
scale. It is perhaps interesting to note that similar contours are also employed in
analytic evaluations of the self energy [41, 61, 62], which are based on the Za-
expansion (there, too, a separation of the two energy scales is necessary). The
decisive observation is that the separation of the scales by an appropriate choice of
the contour facilitates the numerical evaluations considerably. Different techniques
are employed for the high- and the low-energy part.
In the low-energy part, the most challenging problem is the accurate numerical
evaluation of the bound electron propagator to the required relative precision of
10- 24 , whereas the convergence of the partial wave expansion represents a less
involved problem. Resummation techniques [63] which associate a finite value
to an otherwise divergent series are used in this part of the calculation. Observe
that the crosses in Figure 4 are shifted infinitesimally above the positive real axis
for excited states; the corresponding imaginary contributions to the integral (1)
yield the autoionization decay width. For the energy shift, we are only interested
in the real part of the radiative correction. For excited states, the subtraction of the
pole contributions of low-lying states to the required accuracy represents another
challenge in numerical calculations (this subtraction necessarily has to be done
before the final photon-energy integrations are carried out). Difficulties associated
with the regime of ultra-soft photons and the subtraction of the pole contributions
necessitate a further separation of the low-energy part into an infrared part and a
middle-energy contribution (see Figure 5).
We now discuss one of the fundamental differences between numerical and
analytic evaluations. The free electron propagator
1
F=----- (6)
a·p+,B-z
and the full electron propagator G defined in Equation (2) fulfill the following exact
identity,
G =F - F VF +F V G V F.
Im(w)
->O<XX X Re(w)
eIR
Figure 5. Separation of the low-energy contour CL into the infrared part CrR and the
middle-energy part CM. As in Figure 4, the lines directly above and below the real axis
denote branch cuts from the photon and electron propagator. At the separation, we have
Rew = 0.1 En.
84 G. SOFF ET AL.
I
+ I
I
I
I I I
>'< >'<>'<
Figure 6. The exact expansion of the bound electron propagator in powers of the binding field
leads to a zero-potential, a one-potential, and a many-potential term. The dashed lines denote
Coulomb photons, the crosses denote the interaction with the (external) binding field.
G=F-FVF+FVFVF- .. ·.
in hydrogen, it is possible to derive both the IS I / 2 Lamb shift and also the Ryd-
berg constant Roo = ot 2 C mel (2h). Its value is given at present to be Roo =
10 973731.568549(83) m -I [66]. Aiming for an even higher accuracy, measure-
ments would be able to detect a time variation in the fine structure constant ot.
Absolute frequency measurements can be reproduced years later whereas mea-
surements that monitor only relative shifts in ot need to be in continuous oper-
ation. The current limit on variation of ot with time is given by comparing
a hydrogen maser with a Hg microwave atomic clock over a range of 140
days [67],
This number refers to a time variation in the recent epoch. Looking on longer time
scales or on astrophysical data, the boundary becomes even more stringent but for
earlier epochs of the universe [68]. Indications for an existing variation of ot with
time are taken from observations of spectra of very far red-shifted quasars [69]
where the data might indicate a small negative a (i.e., ot seemed to have been
larger in former times) and no other explanation has been found yet. The data are
too poor to give clear evidence, however.
It is beyond the scope of the present article to discuss the underlying theory for
the time variation of fundamental constants in detail. We are just going to point
to some major articles in this field. The possible non-constancy of fundamental
constants was already considered by Dirac [70]. Nowadays, the time variation
of ot is predicted by a number of theories such a string theory (e.g., [71-73 D.
In superstring theories, Einstein's equations of General Relativity are obtained in
a straightforward manner but with a scalar extension. This scalar particle links
the time variation of ot to the Hubble constant, a ~ Ho [73]. Alternatively, theo-
ries have been considered which introduce new scalar fields [74]. Their coupling
to the Maxwell scalar FJ-LvFJ-LV allows for a time variation of ot. The fine struc-
ture constant is not the only 'constant' which might show a possible variation in
time. In Kaluza-Klein theories, the unification of forces takes place in an enlarged
space-time of 4 + N dimensions where N is the number of additional spatial
dimensions and N ;? 7. These extra dimensions are supposed to form a very
small compact manifold with a mean radius RKK where RKK is thought to be of
the order of the Planck length lp = l(n G N )/C 3 ) ~ 1.6 x 10- 35 m. Here G N
denotes the gravitational constant. In Kaluza-Klein theories, an expanding universe
leads quite naturally to RKK -I 0 [71], and if one allows for a varying ot one
finds a ~ GN. In addition, other observables might also change with time such
as the masses of non-fundamental particles, e.g., d/dt(me/mp) -I o. Godone et
al. [75] put a limit on the product of the proton's g factor and that mass ratio,
Id/dt In[gp (me/mp)]1 ~ 5.4 x 10- 13 yc l .
86 G. SOFF ET AL.
. .-. .....
~ 106~-----------------------------,
t:.
-::. 10 5
Ie
.-
.
..... -........... "
..... ..
.-...
.... . .
-......... .
.
. "
.. ."
• .f e •
10
..
"
-1 ~
10 0~~1~0~~2wO~3wO~~4LO~w50~~6~0~w7~0~8~0~~90
nuclear charge Z
Figure 7. Expectation value of the magnetic field strength for the lSI /2 state of a hydrogenlike
atom in the range Z = 1-92. For each Z, the odd isotope with the highest natural abundance
or longest lifetime was chosen. The nuclear magnetic moments are different for each system
and do not follow a simple functional law. Therefore no continuous curve is obtained. The
values were calculated employing wave functions for extended nuclear charge distributions.
EFFECTS OF QED 87
one of the quantities in nature that is most precisely known but it also demonstrates
the theoretical difficulties connected to it. Measurement and theoretical calculation
are conventionally not even presented in a comparable way because of effects re-
sulting from the proton structure [56, 76]. They lead to a small deviation from the
idealized point-dipole magnetic field, and this deviation still cannot be described
in a proper theoretical manner because its exact form is unknown.
The present numbers are
where the nuclear strucure effects include all contributions from the proton, from
finite size and mass to form factors and internal structure. Most of the discrepancy
between both numbers is already removed if the finite size of the proton is taken
into account [79] but the theoretical precision does not increase. In the theoretical
value, quantum elctrodynamical effects are included up to the reasonable accuracy.
The theoretical handling of quantum electrodynamical corrections to the hyper-
fine structure splitting is the same as for the g factor since in both cases radiative
corrections to a magnetic perturbation are considered. Up to now, highly charged
ions were experimentally investigated aiming for the hyperfine strucutre splitting,
at GSI [80-82] as well as at the Lawrence Livermore National Laboratory (USA)
[83, 84]. Corresponding to the rather high experimental precision, also theoretical
investigations were carried out in particular on the QED contributions of order a to
the hyperfine structure splitting in heavy highly charged ions. Again, a perturbation
expansion in Za is not feasible, and the employed computational methods are very
similar to those for the Lamb shift.
The ground-state hyperfine structure splitting of hydrogen-like ions can be writ-
ten in the form
6.E = 4 3 fJv m 2I + 1 2[
-a(aZ) - - - - m e A(aZ)(1 - 8)(1 - E) + XQED ], (8)
3 fJvN mp 2I
where m is the electron mass, mp is the proton mass, fJv is the nuclear magnetic
moment and fJvN = (eli)/(2m p ) ~ 3.152 x lO-8 eVrr is the nuclear magneton.
The terms in the square brackets represent the four corrections to the classical
nonrelativistic hyperfine splitting, i.e., the relativistic factor which for the IS 1/ 2
state is given by [85]
1
A(aZ) = with y = Jl - (aZ)2, (9)
y(2y - 1)
the Dirac equation where the Coulomb potential is slightly modified around the
origin. These wave functions are then employed for the hyperfine structure splitting
calculations. The uncertainty of this effect is governed by the insufficiently known
nuclear charge distribution and does not play any role at the current level of preci-
sion for the theoretical predictions of the hyperfine structure splitting. A reasonable
nuclear charge distribution, e.g., a two-parameter Fermi distribution, allows to
evaluate the effect very easily (e.g, [86]).
The effect due to a deviation from the point-dipole model for the nuclear mag-
netization distribution is often termed Bohr-Weisskopf effect, after A. Bohr and
V. Weisskopf who performed the first numerical investigations [87,88]. Theyem-
ployed a so-called single-particle model where the extended magnetization dis-
tribution is due to a single valence nucleon moving around a core formed by
all other nucleons. One elaborated form of this model was employed to obtain
numerical values for that effect in the range Z = 49-83 for hydrogen- and lithi-
umlike ions [86, 89, 90]. For 209Bi 82+ a related model with the valence proton
moving relativistically was also considered [91, 92]. It gives similar results. A num-
ber of other approaches also try to model the magnetization distribution of the
nucleus due to some outer valence nucleon [93] or treat the whole magnetiza-
tion distribution on a purely phenomenological base by introducing parameters
that allow to model nearly any distribution of the magnetization within the nu-
cleus [94]. The total interaction of all nucleons, however, is not taken into ac-
count in any of these models. An approach pointing more into that direction is the
so-called 'dynamical-correlation model' first evaluated by Arima and Horie [95,
96] and applied to highly charged ions by Tomaselli et al. [97, 98]. This model,
however, depends even more than those mentioned before on input parameters
that have to be obtained from independent experiments. Therefore it is highly
sensitive to uncertainties from nuclear physics. An elaborated discussion on the
current difficulties estimating the Bohr-Weisskopf effect is given in a recent re-
view [99].
The QED corrections of order ex to a bound electron interacting with a perturb-
ing magnetic field are given in Figure 8. They have all been calculated during
the last decade by several independent groups for the lSI/2 and the 2S 1/ 2 state
[86, 90, 100-105] and their values are well established. Unfortunately, their mag-
nitude is of similar size as the uncertainty of the Bohr-Weisskopf effect caused by
the model-like structure which has to be employed for the nuclear magnetization
distribution. Therefore any precision test of bound-state QED by measuring the
hyperfine structure splitting in any highly charged ion is prevented until a reliable
model for the nuclear structure becomes available.
It is possible, however, to combine measurements of the hyperfine structure
splitting in hydrogenlike and lithiumlike ions of the same species, as was pointed
out by Shabaev [106, 107]. By extracting a value for the Bohr-Weisskopf effect
from one experiment, its magnitude for the other charge state of the ion can be
adjusted and its uncertainty is much less than that due to employing different
EFFECTS OF QED 89
models. The Bohr-Weisskopf effect for a hydrogenlike ion can be obtained from a
measurement by
b.E(Is) + b.E(ls) _ b.E(ls)
E(ls) = Duac QED Exp
(10)
b.E(ls)
Dirac
where b.Egi;~C is the value of the lsl/2-hyperfine structure splitting including the
nuclear charge distribution, b.Eg~b is the QED correction to the hyperfine structure
splitting for the lsl/2 state, and b.E~~~ is the experimental value ofthe lsl/2 hyper-
fine structure splitting. The first and the last of these quantities are well known, and
the QED value is also known but put under test when a similar experiment is carried
out on the 2s 1/2 state of a lithiumlike ion, where a similar formula has to be applied.
The expected ratio for the Bohr-Weisskopf effect is E(2s) IE(ls) = 1.078 for the case
of 209Bi [106]. This quantity has to be tested by experiments. The QED contribu-
tions have to be known very precisely, and in addition to the diagrams shown in
Figure 8, those corresponding to electron-electron interactions have to be taken
into account at least to order a 2 . A few of them are displayed in Figure 9. Their
value was estimated by Shabaev and co-workers [90, 105, 106, 108]. The above
proposal was taken up already. A precision search was started at GSI. However, it
did not yet yield any positive result [110] although the region under consideration
was investigated very carefully. Due to technical problems, the search is not yet
completed and therefore any high-precision test of the QED contributions to the
hyperfine structure splitting in heavy highly charged ions is still not performed.
For 209Bi, theoretical and experimental values are displayed in Table II.
In addition to the hyperfine structure splitting, the Feynman diagrams of Fig-
ure 8 also represent the QED corrections of order a to the g factor of bound
electrons. Whereas the magnetic field is generated by the spinning nucleus for the
hyperfine structure splitting case, it is externally applied for g-factor measurements
90 G. SOFF ET AL.
Table II. Recent detailed theoretical predictions for the hyperfine structure splitting
of the ground state in hydrogenlike and lithiumlike bismuth. For 209Bi 82 +, the QED
corrections and the relativistic one-electron value (including the finite nuclear-size ef-
fect) were taken from [104]. The separate finite nuclear-size effect was obtained by
subtracting the corresponding value of a point-like nucleus calculated by Formula (8)
(without 8, E, and XQED). The Bohr-Weisskopf effect is given in [105]. For 209Bi80+,
all values were taken from [105]
where the Zeeman effect is investigated. These experiments therefore employ much
weaker but homogeneous magnetic fields of only several T which, on the other
hand, are much better known and therefore allow much more precise theoretical
predictions. The magnetic moment of a charge q of mass mq is connected with its
angular momentum J by
(11)
where g j is the g factor of this particular charge. For an electron, the constant in
Equation (11) is expressed as the Bohr magneton JLB = (e It) / (2 me) ~ 0.579 X
10-4 e Vrr. The magnetic moment of an electron is
(12)
where J denotes now the total angular momentum of the electron. For a free elec-
tron, the total angular momentum is equal to its spin S, and the Dirac theory
yields gfree = 2. The deviations from this value due to QED were already given in
the introduction. For bound electrons, a number of additional corrections appear.
The most important is the transition from S to J, because only the total angular
momentum operator commutes with the Hamiltonian for a bound electron. This
modification of the g factor is sometimes called 'relativistic correction' and some-
times 'due to spin-orbit coupling'. However, it is not a 'correction' at all but is
entirely contained in the Dirac equation. The g factor of a bound electron due to
this was already obtained by Breit in 1928 [111] (cf. also [112] and [113] for a
detailed derivation). It is given by
-2
3
(
1 +2 J +J
1 I -
2
(z,,)2) (14)
for the lSI/2 and 2S 1/ 2 states in hydrogenlike ions. This value again is modified
due to quantum electrodynamical effects and also due to nuclear properties. Equa-
tions (13) and (14) were obtained employing wave functions for a point-like nu-
clear charge distribution. Extended nuclei lead to the finite-size correction which
amounts to up to 10- 3 for uranium. This correction is easy to handle and limits
are put by the insufficiently known nuclear charge distribution which leads to an
uncertainty of the order 10- 7 for uranium but much less for medium-range and
low Z. The finite nuclear mass leads to a recoil correction similar to the Lamb shift.
However, for the g factor no complete correction to all orders in Za is known yet,
and the existing values from a perturbation series in Za [114, 115] are justified
only for small Z with an uncertainty of at least 1% in the region of carbon [116].
92 G. SOFF ET AL.
~CP~~€J-4-+
~t~tf~~
lr<F~¢cFcr
~t9~ta>~~q
~f~ftq:t~
~~~p;~~~
~~~~~~~
Figure 10. Graphs of order (a/lr)2 to the g factor of a bound electron and to the hyper-
fine structure in hydrogenlike atoms. The interaction with the magnetic field is denoted by a
triangle. The figure is taken from [118).
(15)
EFFECTS OF QED 93
Table III. Known theoretical contributions to the g factor of an
electron bound in the ground state of 12C5+. All values are
given in units of 10- 9 . The uncertainty for the recoil results
from the Za expansion that is employed [116]. The values for
the QED of order (al:rr:) are taken from [118]. The uncertainty
for the order-(al:rr:)2 value is estimated to be 100% (cf. the
discussion in the text). The error margins for the 'total' value
are due to the (Za) expansion for the recoil contribution, the
numerical uncertainties for the QED effects of order (al:rr:), and
the estimated uncertainty for the bound-state QED contribution
of order (al:rr:)2
total: 2001041589.8(9)(12)(11)
where A (2n) is the correspondinr, term in the expansion in powers of (a / rr) for g /2
of the free electron (cf. [2]). A]4) = 197/144 + (1/2 - 3ln2)S(2) + 3/4s(3) =
-0.328478965 ... , where sen) is Riemann's s function [120]. Investigations for
the hyperfine structure splitting have shown, however, that the Za expansion is
not feasible at all to approximate radiative corrections for even medium-range
Z [10 I] and any result of such an approximation should therefore be considered
with enormous care. In Table III we present all known theoretical contributions
to the g factor of hydrogenlike l2c. The bound-state QED contribution of order
(a/rr) can be obtained by subtracting this contribution for the free electron from
the total value for that order. This gives 8443 (12) x 10- 10 which has to be compared
with 7422 x 10- 10 from the the corresponding Za expansion [121]. It is clear
that the available terms of the Za expansion considerably underestimate the value
even for the low Z = 6. For hydrogenlike carbon, an experiment was carried out
which yielded a precision similar to the theoretical one [122, 123]. At this level, it
becomes possible even to evaluate a new high-precision value for the electron mass
that is given elsewhere in this volume [124].
Direct measurements of g factors in elements heavier than carbon have not
been carried out up to now. They are under way for oxygen and the results are
rather promising [125]. However, if the lifetime of the higher of the two hyperfine-
structure splitting levels is measured it is also possible to derive a value for the g
factor of the electron. This was pointed out by Shabaev [126]. The transition proba-
94 G. SOFF ET AL.
Here 1:l.E is the transition energy, g(e) is the bound-electron g factor defined above
and gjn) is the nuclear g factor. For the lowest-order QED and for a pointlike
nucleus a connetion between wand g(e) has been derived first in [127].
The lifetime of the upper hyperfine structure level in 209Bi82+ was measured to
be i exp = 397.5(1.5) f1S by Winter et aZ. [82]. Using Equation (16) and employing
the experimental value of the hyperfine splitting 1:l.E together with the transition
amplitude for 209Bi82+, the experimental value of the bound-electron g factor is
found to be 1.7343(33). This result is in remarkable good agreement with the theo-
retical value of 1.7310 for 209Bi 82 +. Also the value for 207Pb 81 +, measured by Seelig
et al. [81], was found to be in good agreement with the theoretical prediction. Only
the older measurement of Klaft et aZ. [80] yielded a shorter lifetime than predicted
by theory. This triggered many investigations in the past. Only recently, the puzzle
was solved by reinvestigating that experiment [1lO].
The high precision in g-factor measurements could also be used to reinvestigate
the nuclear magnetic properties, in particular the nuclear magnetic moments them-
selves that enter the above formulae (8) and (16) via f1 and g [, respectively. It was
pointed out [128] that due to the necessary corrections used in the old experiments,
they might be less precisely known than stated or the tabulated values are simply
wrong. We therefore conclude this section by pointing to the urgent need of rein-
vestigation also for the nuclear magnetic moments which became obvious by very
precise experiments and calculations of QED effects.
beyond the standard model, due to the high precision which is possible for theoret-
ical calculations in atomic physics. Possible discoveries might be, e.g., a second Z
boson or supersymmetric counterparts of existing particles.
The most precise recent experiment was carried out by measuring the 7Sl/2 -+
6S 1/ 2 transition probability in atomic (i.e., neutral) WCs where the weak charge
Qw of the nucleus is probed. To lowest order ('tree-level'), Qw is given by
where sin2 8 w is the weak mixing angle ('Weinberg angle', sin2 8 w = 0.2224(19)
[66]) and N is the number of the neutrons in the nucleus under consideration.
Adding radiative corrections, one obtains within the standard model Qw
= -73.20(13) for this system [131]. The experimental value was found to be Qw
= -72.06(28)exp. (34)theo. [132] which differs slightly from the standard-model
value. This measurement is thought to be the first indication of the long-looked-for
'anapole moment' of the atomic nucleus [133] which is a parity-violation effect
within the nucleus due to the weak interaction and manifests itself by a toroidal
dipole moment. It was first proposed by Zel'dovitch just after the discovery of
parity violation [134]. A detailed recent discussion can be found in [135] which
also contains all important references on the subject. The discrepancy between
Qw standard model and Qw Cs 7s-+6s, however, may be simply due to the underestima-
tion of theoretical uncertainties in the calculations for the Cs system [136] and does
not yet clearly prove new physics beyond the standard model. Here, we face the
general problem of atomic many-electron systems that although very precise cal-
culations are possible in principle, they are difficult to handle due to the enormous
amount of computer power that is required for these tasks.
In highly charged ions, on the other hand, the binding to the nucleus dominates
by far the electron-electron interaction which may be considered as a perturbation.
The spin-independent part of an effective Hamiltonian for a zero-momentum trans-
fer interaction between a bound electron and a nucleus, mediated by a Zo boson, is
given by
~ GF
Hw(r) =- M Qw PN(r)ys, (18)
2",2
where G F denotes the Fermi constant, PN(r) is the nuclear density and Y5 is the
Dirac matrix. Due to the parity-violating exchange of Z bosons, every electron
state has a small admixture of a wave function with opposite parity, W -+ W +
iryw', where the coefficient iry is pure imaginary because of the T -invariance of the
Hamiltonian (18). Accordingly, the amplitudes of the different processes in atoms
look as follows [130]:
A = Ao + iryAJ, (19)
where Ao is the amplitude of the basic process and A J is the amplitude of the
process caused by parity violation. The 'degree of the parity violation' P in an
96 G. SOFF ET AL.
l81s IS;
14521.93
0.13(8) 14522.10
0.67(5) EI
MI 182 3p-
P 114540.67
1818 3 S; 14636.06
0.24(13) 0.67(16)
MI giO.25 (13) EI
hfqEI
gJ 0.19 (8) 0.93(12)
0.8(9) 2EI
hfqMI
EIMI
Is2I S+
_-'----'_...L...._ _ _--'-_ _'-----''-- 0 57784.06
Figure 11. Energy-level scheme of the first excited states of heliumlike gadolinium. Num-
bers on the right-hand side indicate the ionization energies in eY. The partial probabilities
of radiative transitions are given in s-l. Numbers in parentheses indicate powers of 10. The
large radiative width for the Is2 p 3 PI state is indicated as a bold line. The double lines denote
two-photon transitions.
A 1 = 2 17 (W)I/2
P = 217- _1 , (20)
Ao Wo
where Wo, WI are the probabilities corresponding to the amplitudes Ao and AI,
respectively.
Electroweak radiative corrections were considered both for neutral atoms, e.g.,
in [137] and in the newer literature dealing with the situation in Cs ([136] and
references therein) and also for highly charged hydrogen-like ions [138, 139]. The
consideration of these radiative corrections becomes necessary because they can
contribute up to 10% to the magnitude of PNC effects.
For experiments, a promising situation occurs in heavy helium like ions due to
the near-degeneracy oftwo levels with opposite parities, 21 So and 23 Po. These lev-
els cross near the nuclear charge numbers Z = 64 (gadolinium) as well as close to
Z = 92 (uranium). The case of uranium was elsewhere considered in detail [140].
Here we restrict our consideration the heliumlike highly charged ions of gadolin-
ium (Z = 64) and europium (Z = 63). Especially we consider a quenching-type
experiment with interference of hyperfine- and weak-quenched transitions [141].
Such an experiment would require the use of a polarized highly-charged ion beam
together with a beam-foil time-of-flight technique, a rather challenging task for
experimentalists.
The standard parity violation situation in atoms concerns the M 1 transition with
an admixture of the E I transition. For heliumlike gadolinium and europium, the
EFFECTS OF QED 97
,.
Is28 IS;
v
Is2p 'p~ 14030.82
0.1~8) 14031.22
J 0.12(6) 0.52(5)
0.2 MI "- Is2 'p-
El p '14048.20
M1 0.11(10)
182s'S,
+
/ EI
14141.70
0.20L) 0.63(16)
Ml g; 0.59 (13) EI
hfqEl
gl20.35 ( 8) 0.84(12
0.7(9) hfqMI 2EI
EIMI
Is2 'S+
_-'---L._'--_ _ _- L - _ - ' - - - _ - L - 0 55866.01
Figure 12. Energy-level scheme of the first excited states of heliumlike europium. Notations
are the same as in Figure 11.
energy level scheme is shown in Figures 11 and 12, respectively. The one-photon
hyperfine-quenched transition 21 So -+ 11 So, via the magnetic photon emission
(MI), is due to the hyperfine mixing of the 21 So and 23 SI levels. The weak inter-
action of electrons with the nucleus also opens another one-photon decay channel
21 So -+ 11 So, via the electric photon (E 1) emission through the mixing of the 21 So
and 23 PI levels by the operator Hw in Equation (18). As a result, the total amplitude
A in Equation (19) is represented by the mixture of the basic M 1 (magnetic photon)
amplitude Ao == As and the additional EI (electric photon) amplitude Al == Ap.
The corresponding transitions rates are Ws and W p , respectively. The weak mixing
coefficient 17 in Equation (19) is determined by
(21)
where ~o = E2ISo - E23Po. The theoretical predictions for europium and gadolin-
ium are listed in Table IV. Due to the admixture of states of opposite parity, in a
polarized-beam experiment a small asymmetry in the number of emitted photons
per direction should become visible, expressed by
two-photon transition is about one order of magnitude smaller than the hyperfine-
quenched 23 Po lifetime. This implies a strong background in experiments with
Gd62 + ions.
In europium the weak asymmetry effect reduces up to 28 ::::::: 0.23).0 x 10- 3 .
However, the 21 So level lives significantly longer then the hyperfine-quenched 23 Po
level. The 21 So lifetime equals to about 1.19 ps and corresponds to a decay length
of about 0.1 mm in the laboratory. The peculiarity of the situation is that unlike in
standard hyperfine-quenching experiments we are not aiming at the measurement
of the lifetime defined in our case by the two-photon transition. The experiment
should result in a measurement of the ratio !'J..n/no, where no ± !'J..n/2 are the
numbers of counts for two directions ~ of the beam polarization. This ratio is
directly proportional to the weak interaction matrix element: !'J..n / no = 28.
Since photons being observed in this experiment originate from the single-
photon decays of the hyperfine- and weak-mixed F = I state, the success of the
experiment will depend on the production of a significant degree of polarization
for this state of the heliumlike ion. This is a task to address to the experimentalists
in order to employ the advantages of heavy highly charged ions for parity-violation
investigations and the search for possible new physics beyond the standard model.
5. Summary
During the last two decades there has been increasing interest in quantum electro-
dynamics of strong fields and parity-violating effects in atoms. For one-electron
systems, the fundamental QED contributions can be tested most directly in strong
fields. Therefore, in this review we considered the present status of the Lamb-
shift predictions for hydrogenlike uranium and lead. We emphasized the necessity
to include even the second-order QED corrections in the Lamb-shift calculations.
The recent evaluation of the two-photon self-energy contribution reduces the main
uncertainty in the theoretical Lamb-shift predictions. For hydrogen, the current
experimental and theoretical precision not only allows high-precision tests of QED
but also to determine fundamental constants like Roo with high accuracy. Quan-
tum electrodynamics might therefore be helpful even for looking for new physics
beyond the standard model such as a variation of the fundamental constants.
EFFECTS OF QED 99
We demonstrate that hyperfine splitting effects are also suitable for tests of QED
in strong electromagnetic fields. For any test of QED in strong magnetic fields
by hyperfine-structure splitting .. the Bohr-Weisskopf effect is the main source of
uncertainty in the present predictions, together with possibly erroneously known
magnetic moments of the nuclei. In contrast, experiments on the g factor yield an
up-to-now unmatched precision and agreement with theory on the same level for
any bound state system heavier than hydrogen.
Finally we discussed parity violation effects in few-electron systems. Up to now,
experimental data of atomic parity violation effects are available only for neutral
atoms. For highly charged ions with a Z close to Z = 64, however, the effect
of PNC is ten times larger as in neutral atoms because of the near-degeneracy
of two levels for Z = 64. In particular, to measure the parity-violation effect for
heliumlike gadolinium (Z = 64) and europium (Z = 63), we propose a quenching-
type experiment with interference of hyperfine- and weak-quenched transitions.
Acknowledgements
References
1. Van Dyck, Jr., R. S., Schwinberg, P. B. and Dehmelt, H. G., Phys. Rev. Lett. 59 (1987), 26.
2. Hughes, V. W. and Kinoshita, T., Rev. Mod. Phys. 71 (1999), 133.
3. Jeffrey, A.-M., Elmquist, R. E., Lee, L. H. and Dziuba, R. E, IEEE Trans. Inst. Meas. 46
(1997),264.
4. Pachucki, K., Phys. Rev. A 60 (1999), 3593.
5. Karshenboim, S. G., Can. J. Phys. 77 (1999), 241.
6. Pieper, W. and Greiner, w., Z. Phys. 218 (1969), 327.
7. Zel'dovich, Y. B. and Popov, V. S., Usp. Fiz. Nauk 105 (1971), 403 [Sov. Phys. - Usp. 14
(1972),673--694].
8. Miiller-Nehler, U. and Soff, G., Physics Reports 246 (1994), 101.
9. Ionescu, D. c., Reinhardt, J., Miiller, B. and Greiner, w., Phys. Rev. A 38 (1988), 616.
10. Beyer, H. E, Menzel, G., Liesen, D., Gallus, A., Bosch, E, Deslattes, R., Indelicato, P., Stdh-
lker, T., Klepper, 0., Moshammer, R., Nolden, E, Eickhoff, H., Franzke, B. and Steck, M., Z.
Phys. D 35 (1995), 169.
100 G. SOFF ET AL.
11. Beyer, H. F. and StOhlker, T., Test of QED in high-Z hydrogen-like systems, In: E. Zavattini,
D. Baka10v and C. Rizzo (eds), Frontier Tests of QED and Physics of the Vacuum, Heron
Press, Sofia, 1998, pp. 356-370.
12. StOhlker, T., Mokler, P. H., Bosch, F., Dunford, R. w., Franzke, F., Klepper, 0., Kozhuharov,
c., Ludziejewski, T., Nolden, F., Reich, H., Rymuza, P., Stachura, Z., Steck, M., Swiat, P. and
Warczak, A., Phys. Rev. Lett. 85 (2000), 3109.
13. Beyer, H. F., Liesen, D., Bosch, F., Finlayson, K. D., Jung, M., Klepper, 0., Mosharnmer, R.,
Beckert, K., Eickhoff, H., Franzke, B., Nolden, F., Spadtke, P. and Steck, M., Phys. Lett. A
184 (1994), 435.
14. Mokler, P. H., StOhlker, T., Kozhuharov, c., Moshammer, R., Rymuza, P., Stachura, Z. and
Warczak, A., J. Phys. B 28 (1995), 617.
15. StOhlker, T., Mokler, P. H., Geissel, H., Mosharnmer, R., Rymuza, P., Bernstein, E. M.,
Cocke, C. L., Kozhuharov, c., Miinzenberg, G., Nickel, F., Scheidenberger, c., Stachura, Z.,
Ullrich, J. and Warczak, A., Phys. Lett. A 168 (1992), 285.
16. Franosch, T. and Soff, G., Z. Phys. D 18 (1991), 219.
17. Brown, G. E., Langer, J. S. and Schaefer, G. w., Proc. R. Soc. London A 251 (1959), 92.
18. Desiderio, A. M. and Johnson, W. R., Phys. Rev. A 3 (1971), 1267.
19. Mohr, P. J., Ann. Phys. (New York) 88 (1974), 26.
20. Mohr, P. J., Phys. Rev. Lett. 34 (1975),1050.
21. Snyderman, N. J., Ann. Phys. (New York) 211 (1991), 43.
22. Blundell, S. A. and Snyderman, N. J., Phys. Rev. A 44 (1991), R1427.
23. Uehling, E. A., Phys. Rev. 48 (1935), 55.
24. Wayne Fullerton, L. and Rinker, Jr., G. A., Phys. Rev. A 13 (1976), 1283.
25. Borie, E. and Rinker, G. A., Rev. Mod. Phys. 54 (1982), 67.
26. Hylton, D. J., Phys. Rev. A 32 (1985), 1303.
27. Wichmann, E. H. and Kroll, N. M., Phys. Rev. A 101 (1956), 843.
28. Gyulassy, M., Nucl. Phys. A 244 (1975), 497.
29. Soff, G. and Mohr, P., Phys. Rev. A 38 (1988),5066.
30. Fainshtein, A. G., Manakov, N. L. and Nekipelov, A. A., J. Phys. B 24 (1991),559 (Misprints
of volume number and year in the header of the printed paper!)
31. Persson, H., Lindgren, I., Salomonson, S. and Sunnergren, P., Phys. Rev. A 48 (1993), 2772.
32. Mohr, P. J. and Soff, G., Phys. Rev. Lett. 70 (1993), 158.
33. Beier, T., Mohr, P. J., Persson, H. and Soff, G., Phys. Rev. A 58 (1998),954.
34. Goidenko, I. A., Labzowsky, L. N., Nefiodov, A. v., Plunien, G., Soff, G., Zschocke, S.,
Second-order self energy calculations for tightly bound electrons in hydrogenlike ions,
In: S. G. Karshenboim and F. S. Pavone (eds), Hydrogen Atom II, Proceedings of the Satellite
Meeting "Precision Physics of Simple Atomic Systems" of the 17. ICAP Conference, Springer,
in press.
35. Mitrushenkov, A., Labzowsky, L., Lindgren, I., Persson, H. and Salomonson, S., Phys. Lett. A
200 (1995), 51.
36. Mallampalli, S. and Sapirstein, J., Phys. Rev. Lett. 80 (1998),5297.
37. Goidenko, I., Labzowsky, L., Nefioodv, A., Plunien, G. and Soff, G., Phys. Rev. Lett. 83
(1999),2312.
38. Karshenboim, S. G., Zh. Eksp. Teor. Fiz. (1993).
39. Yerokhin, V. A., Phys. Rev. A 62 (2000), 012508.
40. Manohar, A. V. and Stewart, I. w., Phys. Rev. Lett. 85 (2000), 2248.
41. Pachucki, K., Ann. Phys. (New York) 226 (1993), I.
42. Beier, T. and Soff, G., Z. Phys. D 8 (1988), 129.
43. Schneider, S. M., Greiner, W. and Soff, G., J. Phys. B 26 (1993), L529.
44. Plunien, G., Beier, T., Soff, G. and Persson, H., EPJD 1 (1998),177.
45. Beier, T., Plunien, G., Greiner, M. and Soff, G., J. Phys. B 30 (1997), 2761.
EFFECTS OF QED 101
46. Lindgren, 1., Persson, H., Salomonson, S., Karasiev, v., Labzowsky, L., Mitrushenkov, A. and
Tokman, M., 1. Phys. B 26 (1993), L503.
47. Persson, H., Lindgren, I., Labzowsky, L. N., Plunien, G., Beier, T. and Soff, G., Phys. Rev. A
S4 (1996), 2805.
48. Mallampalli, S. and Sapirstein, J., Phys. Rev. A S4 (1996),2714.
49. Mohr, P. J., Plunien, G. and Soff, G., Physics Reports 293 (1998), 227.
50. Plunien, G., Muller, B., Greiner, W. and Soff, G., Phys. Rev. A 43 (1991), 5853.
51. Plunien, G. and Soff, G., Phys. Rev. A SI (1995), 1119; ibid. S3 (1996), 4614(E).
52. Labzowsky, L. N., Nefiodov, A. v., Plunien, G., Beier, T. and Soff, G., J. Phys. B 29 (1996),
3841.
53. Artemyev, A. N., Shabaev, V. M. and Yerokhin, V. A. Phys. Rev. A 52 (1995), 1884.
54. Artemyev, A. N., Shabaev, V. M. and Yerokhin, V. A., J. Phys. B 28 (1995), 5201.
55. Shabaev, V. M., Artemyev, A. N., Beier, T., Plunien, G., Yerokhin, V. A. and Soff, G., Phys.
Rev. A 57 (1998), 4235.
56. Sapirstein, J. Rand Yennie, D. R, Theory of hydrogenic bound states, In: T. Kinoshita (ed.),
Quantum Electrodynamics, Advanced Series on Directions in High Energy Physics 7, World
Scientific, Singapore, 1990, chapter 12, pp. 560--672.
57. Beier, T., Mohr, P. 1., Persson, H., Plunien, G., Greiner, M. and Soff, G., Phys. Lett. A 236
(1997),329.
58. Pachucki, K., Leibfried, D., Weitz, M. Huber, A., Konig, W. and Hansch, T. w., J. Phys. B 29
(1996), 177.
59. Ientschura, U. D., Mohr, P. J. and Soff, G., Phys. Rev. Lett. 82 (1999),53.
60. Jentschura, U. D., Mohr, P. J., Soff, G. and Weniger, E. 1., Comput. Phys. Commun. 116
(1999),28.
61. Jentschura, U. D. and Pachucki, K., Phys. Rev. A 54 (1996),1853.
62. Jentschura, U. D., Soff, G. and Mohr, P. J., Phys. Rev. A 56 (1997),1739.
63. Weniger, E.J., Comput. Phys. Rep. 10 (1989), 189.
64. Jentschura, U. D., Mohr, P. J., Soff, G., Calculation of QED effects in hydrogen, Hyp. Interact.
(this issue).
65. Niering, M., Holzwarth, R., Reichert, J., Pokasov, P., Udem, Th., Weitz, M., Hansch, T. w.,
Lemonde, P., Santarelli, G., Abgrall, M., Laurent, P., Salomon, C. and Clairon, A., Phys. Rev.
Lett. 84 (2000), 5496.
66. Mohr, P. J. and Taylor, B. N., Rev. Mod. Phys. 72 (2000),351.
67. Prestage, J. D., Tjoelker, R L. and Maleki, L., Phys. Rev. Lett. 74 (1995),3511.
68. Dzuba, V. A. and Flambaum, V. V., Phys. Rev. A 61 (2000), 034502.
69. Webb, J. K., Flambaum, V. V., Churchill, C. w., Drinkwater, M. J. and Barrow, J. D., Phys.
Rev. Lett. 82 (1999), 884.
70. Dirac, P. A. M., Nature (London) 139 (1937), 323.
71. Marciano, W. J., Phys. Rev. Lett. S2 (1984), 489.
72. Barrow, J. D., Phys. Rev. D 35 (1987), 1805.
73. Damour, T. and Polyakov, A. M., NucZ. Phys. B 423 (1994), 532.
74. Carroll, S. M., Phys. Rev. Lett. 81 (1998), 3067.
75. Godone, A., Novero, C., Tavella, P. and Rahimullah, K., Phys. Rev. Lett. 71 (1993), 2364.
76. Bodwin, G. T. and Yennie, D. R, Phys. Rev. D 37 (1988), 498.
77. Hellwig, H., Vessot, R F. c., Levine, M. w., Zitzewitz, P. w., Allan, D. W. and Glaze, D. I.,
IEEE Trans. Insf. Meas. IM-19 (1970),200.
78. Essen, L., Donaldson, R w., Bangham, M. J. and Hope, E. G., Nature (London) 229 (1971),
110.
79. Zemach, A. c., Phys. Rev. 104 (1956), 1771.
80. Klaft, 1., Bomeis, S., Engel, T., Fricke, B., Grieser, R., Huber, G., Kuhl, T., Marx, D.,
Neumann, R., Schroder, S., Seelig, P. and Volker, L., Phys. Rev. Lett. 73 (1994),2425.
102 G. SOFF ET AL.
81. Seelig, P. Bomeis, S., Dax, A, Engel, T., Faber, S., Gerlach, M., Holbrow, c., Huber, G.,
Kiihl, T., Marx, D., Meier, K., Merz, P., Quint, W, Schmitt, E, Tomaselli, M., Volker, L.,
Winter, H., Wiirtz, M., Beckert, K., Franzke, B., Nolden, E, Reich, H., Steck, M. and Winkler,
T., Phys. Rev. Lett. 81 (1998),4824.
82. Winter, H., Bomeis, S., Dax, A, Faber, S., Kiihl, T., Marx, D., Schmitt, E, Seelig, P., Seelig,
W., Shabaev, V. M., Tomaselli, M. and Wiirtz, M.: Bound electron g-factor in hydrogen-like
bismuth, In: GSI Scientific Report 1998, GSI, DE-64291 Darmstadt, Germany, 1999, p. 87.
83. Crespo L6pez-Urrutia, J. R., Beiersdorfer, P., Savin, D. W. and Widmann, K., Phys. Rev. Lett.
77 (1996), 826.
84. Crespo L6pez-Urrutia, J. R., Beiersdorfer, P., Widmann, K., Birkett, B. B., Martensson-
Pendrill, A.-M. and Gustavsson, M. G. H., Phys. Rev. A 57 (1998),879.
85. Pyykko, P., Pajanne, E. and Inokuti, M., Int. J. Quantum Chern. 7 (1973), 785.
86. Shabaev, V. M., Tomaselli, M., Kiihl, T., Artemyev, A. N. and Yerokhin, V. A, Phys. Rev. A
56 (1997), 252.
87. Bohr, A. and Weisskopf, V. E, Phys. Rev. 77 (1950), 94.
88. Bohr, A., Phys. Rev. 81 (1951), 331.
89. Shabaev, V. M., J. Phys. B 27 (1994),5825.
90. Shabaev, V. M., Shabaeva, M. B., Tupitsyn, I. I., Yerokhin, V. A., Artemyev, A. N., Kiihl, T.,
Tomaselli, M. and Zherebtsov, O. M., Phys. Rev. A 57149 (1998); ibid. 58 (1998), 161O(E).
91. Labzowsky, L. N., Johnson, W. R., Soff, G. and Schneider, S. M., Phys. Rev. A 51 (1995),
4597.
92. Labzowsky, L., Nefiodov, A., Plunien, G., Soff, G. and Pyykko, P., Phys. Rev. A 56 (1997),
4508.
93. Schneider, S. M., Schaffner, 1., Greiner, Wand Soff, G., J. Phys. B 26 (1993), L581.
94. Finkbeiner, M., Fricke, B. and Kiihl, T., Phys. Lett. A 176 (1993), 1l3.
95. Arima, A. and Horie, H., Prog. Theor: Phys. 11 (1955), 509.
96. Noya, H., Arima, A. and Horie, H., Prog. Theor: Phys. 8 (1958), 33.
97. Tomaselli, M., Schneider, S. M., Kankeleit, E. and Kiihl, T., Phys. Rev. C 51 (1995), 2989.
98. Tomaselli, M., Kiihl, T., Seelig, P., Holbrow, C. and Kankeleit, E., Phys. Rev. C 58 (1998),
1524.
99. Beier, T., Physics Reports 339 (2000), 79.
100. Schneider, S. M., Greiner, Wand Soff, G., Phys. Rev. A 50 (1994), 118.
101. Persson, H., Schneider, S. M., Greiner, W., Soff, G. and Lindgren, I., Phys. Rev. Lett. 76
(1996), 1433.
102. Yerokhin, V. A., Shabaev, V. M. and Artemyev, A. N., Pis'ma Zh. Eksp. Tear: Fiz. 66
(1997), 19 [JETP Lett. 66(1) (10 July 1997), 18-21]: E-print archive, physics/9905029 (1997)
(http://xxx.lanl.gov ).
103. Blundell, S. A., Cheng, K. T. and Sapirstein, J., Phys. Rev. A 55 (1997), 1857.
104. Sunnergren, P., Persson, H., Salomonson, S., Schneider, S. M., Lindgren, I. and Soff, G., Phys.
Rev. A 58 (1998), 1055.
105. Shabaev, V. M., Artemyev, A. N., Zherebtsov, O. M., Yerokhin, V. A., Plunien, G. and Soff,
G., Hyp. Interact. 127 (2000), 279.
106. Shabaev, V. M., Shabaeva, M. B., Tupitsyn, I. I. and Yerokhin, V. A., Hyp. lnteract. 114
(1998),129.
107. Shabaev, V. M., Hyperfine structure of highly charged ions, In: H. F. Beyer and V. P. Shevelko
(eds), Atomic Physics with Heavy Ions, Springer, Berlin, Heidelberg, 1998, chapter VI,
pp. 138-158.
108. Shabaeva, M. B. and Shabaev, V. M., Phys. Rev. A 52 (1995), 2811.
109. Beiersdorfer, P., Osterheld, A. L., Scofield, J. H., Crespo L6pez-Urrutia, J. R. and Widmann,
K., Phys. Rev. Lett. 80 (1998), 3022.
EFFECTS OF QED 103
110. Bomeis, S., Dax, A., Engel, T., Holbrow, C., Huber, G., KUhl, T., Marx, D., Merz, P., Quint,
W., Schmitt, E, Seelig, P., Tomaselli, M., Winter, H., Beckert, K., Franzke, B., Nolden, E,
Reich, H. and Steck, M., Hyp. Interact. 127 (2000), 305.
111. Breit, G., Nature (London) 122 (1928), 649.
112. Margenau, H., Phys. Rev. 57 (1940),383.
113. Rose, M. E., Relativistic Electron Theory, Wiley, New York, 1961.
114. Faustov, R., Phys. Lett. 33B (1970),422.
115. Grotch, H., Phys. Rev. A 2 (1970),1605.
116. Beier, T., Lindgren, I., Persson, H., Salomonson, S. and Sunnergren, P., Hyp. Interact. 127
(2000), 339.
117. Persson, H., Salomonson, S., Sunnergren, P. and Lindgren, 1., Phys. Rev. A 56 (1997), R2499.
118. Beier, T., Lindgren, 1., Persson, H., Salomonson, S., Sunnergren, P., Haffner, H. and
Hermanspahn, N., Phys. Rev. A 62 (2000), 032510.
119. Eides, M. 1. and Grotch, H.,Ann. Phys. (New York) 260 (1997),191.
120. Abramowitz, M. and Stegun, 1. A., Handbook of Mathematical Functions, 8th edn, Dover,
New York, 1972.
121. Grotch, H. and Hegstrom, R. A., Phys. Rev. A 4 (1971), 59.
122. Hermanspahn, N., Haffner, II., Kluge, H.-J., Quint, w.. Stahl, S., Verdli, J. and Werth, G.,
Phys. Rev. Lett. 84 (2000), 427.
123. Haffner, H., Beier, T., Hermanspahn, N., Kluge, H.-J., Quint, W, Stahl, S., Verdli, J. and
Werth, G., Phys. Rev. Lett. 85 (2000), 5308.
124. Werth, G., Haffner, H., Kluge, H.-J., Quint, w., Valenzuela, T. and Verdu, J., A possible new
value for the electron mass from g-factor measurements on Hydrogen-like ions, this issue, 209.
125. Werth, G., private communication.
126. Shabaev, V. M., Can. J. Phys. 76 (1998), 907.
127. Schneider, S. M., Greiner, Wand Soff, G., Z. Phys. D 31 (1994), 143.
128. Gustavsson, M. G. H. and Martensson-Pendrill, A.-M .. Phys. Rev. A 58 (1998), 3611.
129. Kriplovich, 1., Parity Nonconservation in Atomic Phenomena, Gordon and Breach, New York,
1991.
130. Labzowsky, L., Klimchitskaya, G. and Dmitriev, Y. Y., Relativistic Effects in the Spectra of
Atomic Systems, Institute of Physics Publishing, Bristol, 1993.
131. Marciano, W J. and Rosner, J. L., Phys. Rev. Lett. 652963 (1990); ibid. 68 (1992), 898(E).
132. Bennett, S. C. and Wieman, C. E., Phys. Rev. Lett. 822484 (1999); ibid. 82 (1999), 4153(E);
83 (1999), 889(E).
133. Wood, C. S., Bennett, S. c., Cho, D., Masterson, B. P., Roberts, J. L., Tanner, C. E. and
Wieman, C. E., Science 275 (1997), 1759.
134. Zel'dovich, Y. B., Zh. Eksp. Teor. Fiz. 33 (1957), 1531; [JETP 6 (1958), 1184].
135. Flambaum, V. V. and Murray, D. W, Phys. Rev. C 56 (1997), 1641.
136. Dzuba, V. A. and Flambaum, V. v., Phys. Rev. A 62 (2000), 052101.
137. Lynn, B. Wand Sandars, P. G. H., 1. Phys. B 27 (1994), 1469.
138. Bednyakov, 1., Labzowsky, L., Plunien, G., Soff, G. and Karasiev, v., Phys. Rev. A 61 (1999),
012103.
139. Bednyakov, 1., Labzowsky, L., Soff, G., Plunien, G. and Karasiev, v., Hyp. Interact. 127
(2000), 301.
140. Schafer, A., Soff, G., Indelicato, P., MUller, B. and Greiner, W, Phys. Rev. A 40 (1989), 7362.
141. Labzowsky, L. N., Nefiodov, A. v., Plunien, G., Soff, G., Marrus, R. and Liesen, D., Phys.
Rev. A 63 (2001), 054501.
~ Hyperfine Interactions 132: 105-114,2001. 105
" © 2001 Kluwer Academic Publishers.
s. GORIELY*
Institut d'Astronomie et d'Astrophysique, Universite Libre de Bruxelles, Campus de la Plaine,
CP 226, 1050 Brussels, Belgium; e-mail: sgoriely@astro.ulb.ac.be
Abstract. Although important efforts have been devoted in the last decades to measure atomic
masses, the modelling of the r- and p-processes of nucleosynthesis still requires the use of theoretical
predictions to estimate experimentally unknown masses in the neutron-rich and neutron-deficient
regions. Different mass models are available to extrapolate nuclear masses far away from the ex-
perimentally known region. These models are compared and used to estimate the reaction rates of
relevance in the r- and p-processes. The impact of the different mass models on the astrophysics
predictions are discussed.
1. Introduction
Among the ground state properties, the atomic mass is obviously the most fun-
damental quantity and enter all chapters of nuclear astrophysics. Its knowledge
is indispensable to estimate the rate and energetics of any nuclear transforma-
tion. Although important effort has been devoted in the last decades to measure
atomic masses, important nuclear astrophysics applications involve exotic neutron-
rich and neutron-deficient nuclei for which no experimental data exist. On the
neutron-rich side of the valley of ,B-stability, this concerns principally the so-called
rapid neutron-capture process (or r-process) responsible for the origin of approx-
imately half of the A > 60 stable nuclei observed in nature. On the proton-rich
side, we are dealing mainly with the p-process of nucleosynthesis called for to
explain the production of the stable neutron-deficient isotopes with Z > 34. In
order to model the r-(p-)process, a reaction network that includes essentially all
nuclei located between the line of stability and the neutron (proton) drip line is re-
quired. For each nuclide of the network, several nuclear properties must be known,
including neutron, proton, a-capture rates, photodisintegration rates, ,B-decay half-
lives, .... In both applications, few (if any, as in the r-process) binding energies are
known along the nuclear flow, so that nucleosynthesis calculations have to make
use of theoretical extrapolated nuclear masses. These specific applications lead us
to favour microscopic or semi-microscopic predictions based on sound and reliable
* S. G. is FNRS Research Associate.
106 S.GORIELY
nuclear models which, in tum, can compete with more phenomenological highly-
parametrized models in the reproduction of experimental data. The selection crite-
rion of the adopted model is fundamental, since most of the nuclear ingredients in
reactions rate calculations need to be extrapolated in an energy and mass domain
out of reach of laboratory measurements, where parametrized systematics based on
experimental data can fail drastically. In addition, the r- and p-processes involve a
large number (thousands) of unstable nuclei, so that only global approaches, i.e.,
a unique description in the whole (N, Z)-plane, are relevant. In Section 2, the
different mass models used for astrophysics applications are described and their
predictions away from the experimentally known regions compared. The predicted
abundance distributions may be affected by the mass model uncertainties. Section 3
illustrates the impact of the different mass predictions on the r-abundance distribu-
tion predicted by the simple canonical and multi-event models. Section 4 presents
p-process calculations performed with different mass models and the impact of the
uncertainties remaining in the neutron-deficient region.
masses well in excess of 2 Me V. This is far from reaching the level of precision
found by droplet-like models.
The result is that the most microscopically founded mass formulas of practical
use were till recently those based on the so-called ETFSI (extended Thomas-Fermi
plus Strutinsky integral) method [4]. The ETFSI method is nothing else than a
high-speed macroscopic-microscopic approximation to the HF method based on
Skyrme forces, with pairing correlations generated by a 8-function force that is
treated in the usual BCS approach (with blocking). The macroscopic part consists
of a purely semi-classical approximation to the HF method, the full fourth-order
extended Thomas-Fermi method, while the microscopic part (based on what is
called the Strutinsky-integral form of the Strutinsky theorem) constitutes an at-
tempt to improve this approximation perturbatively, and in particular to restore the
shell corrections that are missing from the ETF part. In the latest version of the
ETFSI mass model (ETFSI2), eleven parameters are found to reproduce the 1719
experimental masses of the A ~ 36 nuclei (the N = Z, Z ± 1 nuclei subject to
Wigner-term anomalies are not included) with an rms deviation of 709 keV [5].
The precision of the ETFSI mass table is therefore comparable with the one ob-
tained by the droplet-like formula. The ETFSI model remains an approach of the
macroscopic-microscopic type, although it provides a high degree of coherence be-
tween the macroscopic and microscopic terms through the unifying Skyrme force
underlying both parts. A logical step towards improvements obviously consists in
considering now the HF method as such. It was demonstrated very recently [6, 7]
that HF calculations in which a Skyrme force is fitted to essentially all the mass
data are not only feasible, but can also compete with the most accurate droplet-like
formulas available nowadays.
The force used in the latest HF-BCS mass calculation of [7] is a conventional
lO-parameter Skyrme force, along with a 4-parameter 8-function pairing force
(pairing correlations are introduced in the framework of the BCS method). The
Skyrme and pairing parameters are determined by fitting to the full data set of
1719 A ~ 36 masses, both spherical and deformed. The best fit is obtained for an
effective nucleon mass M* = 1.05M and a symmetry coefficient J = 28 MeV. In
order to describe the IN - Z I ~ 1 nuclei, a phenomenological Wigner correction
term is added to the total HFBCS binding energy. The resulting rms deviation from
the 1772 masses with A ~ 36 amounts to 0.683 MeV. The MSk7 force has been
used to estimate the complete mass table HFBCS 1 made of 9200 nuclei, including
all those lying between the drip lines over the range Z, N ~ 8 and Z ~ 120. The
rms error to the 1888 nuclei in this range for which measured masses are given in
the 1995 Audi-Wapstra compilation [8] is 0.738 MeV.
Another microscopically rooted approach worth mentioning is the development
by Dufio and Zuker [9] of a mass formula based on the shell model (SM). The
nuclear Hamiltonian is separated into a monopole term responsible for the satura-
tion properties, and a residual multipole term. The monopole term is responsible
for saturation and single-particle properties, and fitted phenomenologically. The
108 S. GORIELY
multi pole part is derived from realistic interactions. The latest version of the mass
formula made of 10 free parameters reproduces the 1888 Z, N ~ 8 experimental
masses with an rms error of 553 ke V.
The quality of the mass models available is traditionally estimated by the rms
error obtained in the fit to experimental data and the number of free parameters.
However, this overall accuracy does not imply a reliable extrapolation far away
from the experimentally known region in view of the possible shortcomings de-
scribed above and linked to the physics theory underlying the model. The reliability
of the mass extrapolation is a criterion of first importance when dealing with astro-
physics applications. The quality of a mass model can also be tested by its ability to
estimate other quantities than masses, such as deformations, charge radii or fission
barriers. It should be mentioned that the calculation of thermonuclear reaction rates
requires the knowledge of various ground-state properties in addition to the nuclear
mass, namely the deformation, density distribution, single-particle level scheme,
pairing force and shell correction energy. Coherent nucleosynthesis calculations
demand the different nuclear quantities to be taken from one unique model. Mass
models which do not provide all the quantities of relevance are consequently of
reduced applicability for astrophysics. The importance of estimating these proper-
ties reliably should not be underestimated. For example, the nuclear level density
of a deformed nucleus at low energies (typically at the neutron separation energy)
is predicted to be significantly (about 30-50 times) larger than of a spherical one
due principally to the rotational enhancement. An erroneous determination of the
'"""'
(;:) - 10
U
a:)
~ - 15
~
20 40 60 80 100 120 20 15 \0 5 o
Z S n [MeV]
Figure 1. Comparison of nuclear masses predicted by HFBCSI and FRDM models. Left
panel is plotted as a function of the charge number Z and right panel as a function of the
HFBCS I neutron separation energy Sn. The nuclei involved in the r-process are characterized
by Sn ,:S 3 MeV and in the p-process by Sn ;?: 9 MeV.
NUCLEAR MASSES AND THE r- AND p-PROCESSES OF NUCLEOSYNTHESIS 109
deformation can therefore lead to large errors in the estimate of radiative capture
rates.
In summary, the most popular mass models used nowadays for astrophysics
applications are FRDM, ETFSI2, HFBCSI and SM (although the latter does not
provide ground-state properties other than masses). Despite the rough similarity
in the quality of the data fits of all four mass formulae, the overall quality of the
models are not identical. In particular, FRDM suffers from all the shortcomings
described above, while ETFSI and HFBCS approaches are still subject to unsatis-
factory pairing treatment [3]. At this stage, these four models should be regarded as
providing a lower limit to the remaining uncertainties in the mass extrapolations.
Unknown physical effects affecting exotic nuclei might not be described by neither
of these models, and higher uncertainties can consequently be foreseen. For exam-
ple, the existence of the shell quenching at the neutron drip line remains an open
question that will be resolved by future developments of microscopic mass models
constrained by new measurements. At the moment, striking differences emerge on
extrapolating far from the experimental data, as illustrated in Figure 1, especially
between HFBCS 1 and FRDM or SM predictions (HFBCS 1 and ETFSI2 masses
are relatively similar).
Solar System
-e- HFBCSI
-B- ETFSI2
10-1 FRDM
10-2 .
10-3
104~~~~~~~~~~~~~~~~~~~
70 80 90 100 IlO 120 130 140
A
Figure 2. Comparison of the r-abundances calculated with the HFBCSl, ETFSI2 and FRDM
mass models in the framework of the canonical model. The r-process is characterized by
N n = 1021 cm- 3, T = 1.2 x 109 K and r = 2. I s. The curve with full dots corresponds to
the solar system r-abundance distribution arbitrarily normalized.
pressed in 109 K) and an irradiation time r = 2.1 s. This type of event can be seen
as extreme conditions found during the passage of a supernova shock front through
the He-burning shell of a massive supernova progenitor star [14]. As seen in Fig-
ure 2, significantly different abundance distributions could be expected according
to the mass model used. In this canonical model at fixed irradiation time, the lo-
cation of the r-abundance peaks principally depends on the nuclear masses. The
r-process flow follows during the neutron irradiation a time-independent r-process
path defined by the equation
3
1
"lS2n(Z, A) = Sa0 [MeV] = ( 34.075 -log N n [cm- 3] + "llog T9 )T9
5.04'
where S2n is the 2-neutron separation energy and S~ the so-called astrophysical
parameter. The r-process path and the predicted r-abundance peaks depend on two
major characteristics of the mass model, namely the symmetry energy coefficients
and the neutron shell correction energy. A small symmetry energy coefficient mod-
ifies the slope of the mass parabola and leads to an r-process developing in a more
neutron-rich region involving faster ,8-decaying isotopes (the ,8- -decay rates of
exotic neutron-rich nuclei are relatively insensitive to the mass uncertainties [10]).
The shell correction energy modifies the nuclear deformability and affects the pro-
duction of nuclei in the vicinity of the r-peaks [15]. For all these reasons, in the
example illustrated in Figure 1, FRDM masses lead to a weaker r-process.
A second r-process model, called the multi-event r-process model, is now con-
sidered to estimate the impact of the masses. In such a model, a given number of
different non-equilibrium canonical events are superposed in order to reproduce
NUCLEAR MASSES AND THE r- AND p-PROCESSES OF NUCLEOSYNTHESIS 111
Solar r-abundances
Th in CS 22892-
--HFBCS1
- - - ETFSI2
-----FRDM
.-....... S'v1
-o- HFBC81
-o- FRDM 21.6
8M 1\
22.4 0"
(JQ
23.2 Z
/\ :::J
o «:I 24 n
CZl
V 2.5 s,
24.8 ,--,""
V
25.6
2 26.4
80 100 120 140 160 180 200 220 240
A
Figure 3. (a) Fit to the solar system r-abundances derived from a superposition of
non-equilibrium canonical events (see text) with the HFBCS I, ETFSI2, FRDM and SM
masses. The square corresponds to the Th abundance observed in the CS 22892-052 star [16].
(b) Average astrophysical parameters S~ and neutron density Nn responsible for the fits shown
in (a). The curves obtained with ETFSI2 masses are very similar to HFBCS I.
density (at fixed T9 = 1.3) responsible for the production of each r-nuclide in solar
abundance are displayed in Figure 3(b). In this r-process model, the changes in
the nuclear physics input do not affect the abundance distributions, but rather the
statistical distributions of the events involved. Figure 3(b) clearly shows that the
thermodynamic conditions required to reproduce the solar r-abundance distribution
are significantly different with HFBCS 1 than with FRDM or SM.
A direct application of the multi-event r-process model is found in the Th cos-
mochronology. Recent observations of Th in ultra-metal poor stars revived the old
idea that the Th cosmochronometry could provide an age estimate of the oldest
stars in the Galaxy, and therefore a lower limit to the age of the Galaxy [16].
Assuming that a solar-like mix of the r-elements ingested in these halo stars orig-
inates from a small number of nucleosynthetic events that took place just before
the formation of the stars, the age of the star can be estimated without calling for a
complex model of the chemical evolution of the Galaxy. The only difficulty of the
methodology is related to the theoretical estimate of the original production of Th
by the r-process. The r-process production of Th is obviously model dependent and
numerous uncertainties in the astrophysics conditions, nuclear physics input and
solar r-abundances still keep the Th cosmochronometry on unreliable grounds [18].
In particular, the nuclear masses play a crucial role in the reliability of the Th
cosmochronometry. As illustrated in Figure 3(a), the Th abundance predicted with
different mass models can differ significantly. The fit to the stable nuclei are of
the same quality, and there is obviously no reason to favour one or another mass
formula on grounds of this parametric fit. The extrapolation to the Th abundance
obtained with the different mass models lead to a star age of T* = 12.1, 11.3, 18.8
and -10.7 Gyr for the HFBCS 1, ETFSI2, FRDM and SM models, respectively.
2.3 and 2.9 billions degrees. The error bars correspond to the deviation found when
using different mass models, namely HFBCS 1, ETFSI2 and FRDM in the estimate
of all the n-, p- and a-capture rates, as well as the respective photodisintegration
rates. Experimental masses [8] are used in the rate calculation when available. This
significantly reduces the differences in the predicted rates, but not the discrepan-
cies in the overproduction factors obtained for the heaviest p-nuclei (Figure 4).
Measuring the remaining unknown masses of the neutron-deficient isotopes with
65 :s;; Z :s;; 81 close to the valley of ,B-stability would resolve this disagreement.
5. Conclusions
A reliable prediction of the abundances produced by the r- and p-processes requires
an accurate knowledge of the nuclear masses. When dealing with astrophysics
applications, the reliability of a mass model should be estimated not only on its
ability to reproduce experimental data, but also on the quality of the nuclear physics
theory it is based on. Only global microscopic approaches can provide the accuracy
and reliability requested for nucleosynthesis applications. This continued effort to
improve the microscopic predictions of nuclear masses is concomitant with new
mass measurements far away from stability.
References
I. Moller, P., Nix. J. R. , Myers, W. D. and Swiatecki, W. J., At. Data Nuc!. Data Tables 59 (1995),
185.
2. Goriely, S. and Arnould, M., Astron. Astrophys. 262 (1992), 73.
3. Pearson, 1. M. (2001) this issue, 59.
114 S. GORIELY
4. Aboussir, Y., Pearson, J. M., Dutta, A. K. and Tondeur, F., At. Data NucZ. Data Tables 61
(1995), 127.
5. Goriely, S., In: S. Wender (ed.), Proc. of the 10th Int. Symposium on Capture Gamma-Ray
Spectroscopy and Related Topics, AlP, 2000, 287 p.
6. Tondeur, F., Goriely, S., Pearson, J. M. and Onsi, M., Phys. Rev. C 62,024308.
7. Goriely, S., Tondeur, F. and Pearson, 1. M., At. Data NucZ. Data Tables 77 (2000), 311.
8. Audi, G. and Wapstra, A. H., NucZ. Phys. A 595 (1995), 409.
9. Dufio, J. and Zuker, A., Phys. Rev. C 52 (1995), R23.
to. Goriely, S. and Arnould, M., Astron. Astrophys. 312 (1996), 327.
II. Goriely, S., In: N. Prantzos and S. Harissopoulos (eds), Proc. of the 5th International Corif. on
Nuclei in the Cosmos, Editions Frontieres, 1998, 314 p.
12. Demetriou, P. and Goriely, S., In: Proc. of the 6th International Con! on Nuclei in the Cosmos;
NucZ. Phys. A (2000), to be published.
13. Tachibana T., Yamada M. and Yoshida N., Prog. Theor. Phys. 84 (1990),641.
14. Thielemann, F.-K., Arnould, M. and Hillebrandt, w., Astron. Astroph. 74 (1979),175.
15. Pearson, J. M., Nayak, R. C. and Goriely, S., Phys. Lett. B 387 (1996), 455.
16. Sneden c., Cowan J. J., Burris D. L. and Truran J. W., Astrophys. 1. 496 (1998), 235.
17. Arnould, M., Astron. Astrophys. 46 (1976), 117.
18. Goriely, S. and Clerbaux, B., Astron. Astrophys. 346 (1999), 798.
19. Rayct, M., Arnould, M., Hashimoto, M., Prantzos, N. and Nomoto, K., Astron. Astrophys. 298
(1995),517.
20. Baleisis, A., Arnett, D., In: Proc. of the 6th International Con! on Nuclei in the Cosmos; Nucl.
Phys. A (2000), to be published.
Hyperfine Interactions 132: 115-126,200l. 115
© 2001 Kluwer Academic Publishers.
Abstract. Superallowed ,'l-decay provides a sensitive means for probing the limitations of the
Electroweak Standard Model. To date, the strengths ([t-values) of superallowed 0+ -+ 0+ ,'l-decay
transitions have been determined with high precision from nine different short-lived nuclei, ranging
from lOC to 54Co, Each result leads to an independent measure for the vector coupling constant
Gv and collectively the nine values can be used to test the conservation of the weak vector current
(CVC). Within current uncertainties, the results support CVC to better than a few parts in 10,000
- a clear success for the Standard Model! However, when the average value of Gv, as determined
in this way, is combined with data from decays of the muon and kaon to test another prediction of
the Standard Model, the result is much more provocative. A test of the unitarity of the Cabibbo-
Kobayashi-Maskawa matrix fails by more than two standard deviations. This result can be made
more definitive by experiments that require extremely precise mass measurements, in some cases
on very short-lived (( 100 ms) nuclei. This talk presents the current status and future prospects for
these Standard-Model tests, emphasizing the role of precise mass, or mass-difference measurements.
There remains a real challenge to mass-measurement technique with the opportunity for significant
new results.
Key words: atomic mass, CKM unitarity, CVC hypothesis, standard model, superallowed beta-
decay.
1. Introduction
Superallowed 0+ --+ 0+ nuclear ,B-decays provide both the best test of the Con-
served Vector Current (CVC) hypothesis and, together with the muon lifetime,
the most accurate value for the up-down quark-mixing matrix element Vud of the
Cabibbo-Kobayashi-Maskawa (CKM) matrix. This matrix should be unitary, and
experimental verification of that expectation constitutes an important test of the
Standard Model. With current world data for 0+ --+ 0+ ,B-decays [1] used to obtain
a value for Vud, and the standard values [2] taken for the other required elements
of the CKM matrix, the unitarity test from the sum of the squares of the elements
116 J. C. HARDY AND I. S. TOWNER
in the first row fails to meet unity by more than twice the estimated uncertainty.
This result is tantalizingly close to establishing a definitive disagreement with
the Standard Model, and prompts renewed efforts to improve the precision with
which the test can be made. The challenges in doing so are both experimental and
theoretical.
The nine 0+ -+ 0+ transitions that currently comprise the nuclear input data for
Vud are all between T = I analog states. Thus, nuclear-structure effects only enter
at the level of differences between the parent and daughter wave functions. These
differences, which are caused by Coulomb and charge-dependent nuclear forces,
tum out to be very small, and introduce a correction, denoted Dc, of 0.2--0.6%
when the experimental ft-values are used to extract a value for the effective weak
vector coupling constant, G~. Even a conservative estimate of the uncertainties
in this correction indicates that structure-dependent uncertainties should not afflict
the experimental determination of G~ above the level of approximately ±0.1 %. To
date, ft-value measurements, which have focused on nuclei with 10 (; A (; 54,
have aimed at achieving this level of precision or slightly better.
This requirement has placed stringent demands on experiment. To determine the
ft-values of these transitions to the required precision, the half-lives and branching
ratios have had to be measured to ±0.05%, and the total Ii-transition energy to plus
or minus a few hundred eV! If the unitarity test is to be made more definitive, not
only will future experiments have to achieve this level of precision or better, but
there must be improvements - or increased confidence - in the calculated values
of Dc.
To calculate Dc for a particular superallowed transition, it is important to have
a reliable nuclear model that demonstrably fits nuclear properties in the same
mass region. To refine the model's effectiveness in calculating charge-dependent
effects, it is also valuable to have accurate experimental data on the band c co-
efficients of the isobaric-multiplet mass equation (lMME) for the same 0+ mul-
tiplet and, if possible, to have data on other, non-analog 0+ -+ 0+ decays from
the same parent state. The nine superallowed 0+ -+ 0+ decays currently known
with high precision occur among nuclei where the IMME coefficients are known,
structure information is relatively abundant and reliable models exist. Any future
improvements in Dc are likely to come from measurements of additional superal-
lowed transitions, especially those for which Dc is anticipated to be particularly
large. Such measurements will then constitute a test of the Dc calculations and,
if successful, will give a better indication of the uncertainty that should be ap-
plied to Dc in the cases currently known, where the correction itself is smaller.
All these future measurements will, of necessity, require very precise mass mea-
surements, both to determine the Ii-transition energies and to establish the IMME
coefficients.
This paper will outline the current status of the nuclear measurements bearing
on Vud , the significance of a demonstrated disagreement with CKM unitarity, and
future directions for study, with special emphasis on mass measurements.
STANDARD-MODEL TESTS WITH SUPERALLOWED tl-DECAY 117
(1)
with
G~ = G v {1 + ~~f/2, (Mv)2 = 2(1 - oc),
K 2;r3/i In 2 -10 -4 (2)
( flhc)6 ~ 'i = (8120.271 ± 0.012) x 10 GeY s,
(mec-)-
where f is the statistical rate function, t is the partial half-life for the transition,
(Mv) is the Fermi matrix element and Gv is the primitive vector coupling con-
stant. The physical constants used to evaluate K were taken from the most recent
Particle Data Group publication [2]. These equations also include three calculated
correction terms - all of order 1%. We write OR as the nucleus-dependent part of
the radiative correction, ~~ as the nucleus-independent part of the radiative correc-
tion, and Oc as the isospin symmetry-breaking correction. A general description of
these three correction terms and the methods used in their calculation has appeared
elsewhere [1, 3]. In the present context, it is sufficient to note that nuclear structure
plays a small role in the determination of OR, but it is predominant for that of oc.
Equations (1) and (2) can now be combined into a form that is convenient for
the analysis of experimental results:
K
:Ft =ft(1 + oR)(1 - oc) = 2 V . (3)
2G v O + ~R)
Here we have defined :Ft as the "corrected" ft-value. From this equation, It IS
evident that the :Ft-values obtained from 0+ ~ 0+ transitions in different nuclei
can constitute a stringent test of CYC, which requires them all to be equal.
To date, superallowed 0+ ~ 0+ transitions have been measured to ±O. 1%
precision or better in the decays of nine nuclei, 1OC, 140, 26m AI, 34Cl, 38mK, 42SC,
46y, 50Mn and 54Co. In each case, three quantities are required from experiment: the
transition energy QEC which is used in calculating f; the half-life tl/2 ofthe parent
nuclide and the branching ratio for the superallowed transition, which together
yield the partial half-life t. World data on Q-values, lifetimes and branching ratios
- the results of over 100 independent measurements - were thoroughly surveyed
[4] in 1989 and then updated several times since, most recently for the WEIN98
conference [1]. By applying the OR and Oc corrections listed in [1], we converted
the results into :Ft-values. These values are plotted in Figure 1.
It is important to appreciate that the values of OR and Oc each result from more
than one independent calculation. In the case of OR, results come from a variety
118 1. C. HARDY AND I. S. TOWNER
I
6000
..C "CI "SC "Mn
3080
5000 "0 -.. .... "V "Co
X300/ ::::
mt14t-
-
.!!.
41
4000
/ 3074
1000
10 15 20 25 30
10 20 30 40
Z of daughter
of primary sources, all in complete accord with one another; for oc, we have used
an average of two independent calculations [5, 6] with assigned uncertainties that
reflect the (small) scatter between them. Thus, in a real sense, both experimentally
and theoretically, the :Ft-values illustrated in Figure 1 represent the totality of cur-
rent world knowledge. The uncertainties reflect the experimental uncertainties and
an estimate of the relative theoretical uncertainties in oc. There is no statistically
significant evidence of inconsistencies in the data (X 2/ v = 1.1), thus verifying the
expectation of CVC at the level of 3 x 10-4 , the fractional uncertainty quoted on
the average :Ft-value: :Ft = 3072.3 ± 0.9 s.
Now, in using this average :Ft-value to determine Vud and test CKM unitarity,
we must add another ± 1.1 s to account for the systematic difference between the
two calculations of Oc that we have combined in reaching this result. (For a more
complete discussion of how we treat these theoretical uncertainties, see [4].) Since
this systematic difference arising from theory is not in any sense a Gaussian prob-
ability distribution, we take a conservative approach here and add the two errors
linearly to obtain the value we use in subsequent analysis:
The value of Vud is obtained by relating the vector constant G v determined from
this Tt value, to the weak coupling constant from muon decay, G F/(lic)3 =
(1.16639 ± 0.00001) x 10- 5 GeV- 2 , according to:
V2 _ G~ _ K
(5)
ud - G~ - 2G~(l + ll~)Tt .
With the nucleus-independent radiative correction adopted from [7], ll~ = (2.40±
0.08)%, we obtain the result
We are now in a position to test the unitarity of the CKM matrix by evaluating
the sum of squares of the elements in its first row. With the value just obtained
for V ud , combined with the values of Vus and Vub quoted by the Particle Data
Group [2], the unitarity sum becomes
cannot be at fault. Not only do they originate from a large number of independent
measurements, but also the error bar associated with IVud I is not predominantly
experimental in origin. In fact, if experiment were the sole contributor, the uncer-
tainty would be only ±O.OOOI. The largest contributions to the IVudl error bar come
from 6.~ (±0.0004) and Oc (±0.0003).
Thus, if we are to determine whether the minimal Standard Model has failed,
we must eliminate all possible "trivial" explanations for the apparent non-unitarity.
To do so, nuclear physicists focus on the reliability of the calculated corrections
in V ud . (Others are re-evaluating Vus - see [1].) But, if there is a fault in the
corrections, what size effect are we seeking? To restore unitarity, the calculated
radiative corrections (OR or 6.~) for all nine superallowed transitions would all
have to be shifted downwards by 0.3%, or the calculated Coulomb corrections Oc
all shifted upwards by 0.3%, or some combination of the two. Such changes would
constitute a substantial fraction of the total values of these small quantities. We
have recently re-examined [1] the calculation of the various correction terms and
conclude that such large changes are very implausible, particularly in the case of OR,
the calculation of which involves standard QED and is well verified.
Even Oc is well substantiated by several independent, yet concordant calcula-
tions [5,6,9,10] and by measurements [11,12] of the weak non-analog 0+ --+ 0+
,B-transitions from 38mK, 42Sc, 46V, 50Mn and S4CO. Such transitions can only occur
via admixed components from the analog-state wavefunction, and are a sensitive
test of charge-dependent mixing in five of the very nuclei whose superallowed
branches contribute to the determination of Vud ' In all cases, the calculations used to
obtain Oc also yielded values for the non-analog transitions that agree very closely
with experiment [13]. Thus, we must conclude that oc, like OR, is firmly based, and
unlikely to conceal any fault that could shift all of its values substantially outside
their quoted error bars. Therefore, it is the reduction of those error bars that should
preferably be the goal of future experiments.
Given the importance of their contribution to the uncertainty now quoted for Vud,
it is the precision of 6.~ and Oc that must be improved if the unitarity test is to be
sharpened substantially. Indeed, 6.~ is inherent to the determination of Vud whether
the latter is obtained from neutron decay or from the superallowed transitions and,
until the calculation of 6~ is improved, its relative imprecision will continue to
limit the unitarity test at about the present level regardless of any other improve-
ments, experimental or theoretical. Improved calculations of 6~ must therefore
take priority as the most critical task for the future. This appears to lie entirely
within the realm of theory: there have been no suggestions for experiments that
could help to refine existing calculations or to confirm new ones.
The charge-dependent correction, oc, is a different story. Theoretical improve-
ments are certainly possible, but experiment can provide independent controls on
the calculations which, if successful, will reduce the uncertainty of the results.
On the theoretical side, a large-basis (no core) shell model has already been used
[101 to calculate Oc for the JOc superallowed transition. The extension of such
STANDARD-MODEL TESTS WITH SUPERALLOWED ,B-DECAY 121
detailed calculations to heavier nuclei will make important additions to the results
of the more phenomenological methods used to date. Experimentally, there are
three different approaches currently being followed:
(1) increasing experimental precision on the nine knownft-values;
(2) measuring new 0+ ---+ 0+ decays from odd-Z, Tz = 0 nuclei with A ;) 62; and
(3) measuring new 0+ ---+ 0+ decays from even-Z, Tz = -1 nuclei with
18:::;: A :::;: 42.
4234
-> 4233
! ---
--
Q)
.:.::
Q)
:l
4232
CO
>u
d 2832
2831
2830
cases that uncertainty is comparable to, or greater than, the uncertainty arising from
the theoretical correction terms. However, as Figure 2 illustrates, if these values are
to be improved, any new measurement must be of substantially higher precision
than any of the previous ones. This should be possible with an on-line Penning-
trap mass spectrometer but it will require special efforts. Are those efforts justified?
Certainly, if they are accompanied by other improvements, such as a more precise
branching ratio for 10C and better half-lives for tOc, 34CI and 38mK, then the test of
CVC will be made more demanding and, to the extent that the Tt-values continue
to agree with one another, this would demonstrate at the same time the reliability
of the Oc calculations, which compensate for transition-to-transition variations that
appear in the uncorrected It-values. Of course, it is only the relative values of
Oc that can be tested by this method. However, Oc itself represents a difference
- the difference between the parent and daughter-state wave functions caused by
charge-dependent mixing. Thus, the experimentally determined variations in Oc
STANDARD-MODEL TESTS WITH SUPERALLOWED tl-DECAY 123
Contributions to Ft uncertainty
16
14 -
~
,!
! I
i
o
o 12 - I
6 10
~
t t
c 8 f- -
en
~
- ,- l·
t: 6 r-
!
1----- f-
C\1
a... 4 f- - ,- r- r- I - - r-
2
H5l 7 12 16 18 20 22 24 26
Z of daughter
Figure 3. Contributions from experiment and theory to the overall Ft-value uncertainty for
the nine well-known superallowed transi tions.
are actually second differences. It would be a pathological fault indeed that could
calculate in detail these variations (i.e., second differences) in DC while failing to
obtain their absolute values (i.e., first differences) to comparable precision.
The alternative experimental approach to testing DC is offered by the possibility
of increasing the number of superallowed emitters accessible to precision studies.
Two series of 0+ nuclei present themselves: the even-Z, Tz = -1 nuclei with
18 :s; A :s; 42, and the odd-Z, Tz = 0 nuclei with A ~ 62. These are illustrated in
Figure 4. The main attraction of these new regions is that the calculated DC values
for the superallowed transitions are larger, or show larger variations from nuclide
to nuclide, than the values applied to the nine currently well-known transitions. In
principle, then, they afford a valuable test of the accuracy of the DC calculations. It
is argued that if the calculations reproduce the experimentally observed variations
where they are large, then that must surely verify their reliability for the original
nine transitions whose DC values are considerably smaller.
Beyond the large values of Dc anticipated for both, these two series of new
superallowed emitters have very different characteristics from one another in most
other respects. Those nuclides with 18 :s; A :s; 42 share the same nuclear shell-
model space as the nine currently contributing to the unitarity test and, particularly
in the sd-shell, that model is extremely successful in calculating a wide range of
nuclear properties. Thus, any discrepancies observed between theory and exper-
iment for this series would directly reflect on the DC values now being used in
the extraction of Vud . In contrast, the nuclides with A ~ 62 are in a completely
124 1. C. HARDY AND I. S. TOWNER
50~------------------------------------------~~~---'
40
N
..
'Ii
c:
0
-
0
"-
0..
30
0
"-
CD
.0
E
:::J 20
Z
10
10 20 30 40 50 60
Number of Neutrons, N
Figure 4. Portion of the nuclear chart showing the nine currently well-known superallowed
emitters (textured grey) and the two additional series of reasonably accessible emitters (light
grey). The stable nuclei are shown in black. Five cases currently under study are specifically
identified.
different nuclear region, one for which reliable nuclear model calculations do not
yet exist. Indeed, much experimental information essential to the development of
a good model and to the calculation of reliable charge-dependent mixing is sparse
to non-existent in this region: the T = 1 isobaric-multiplet states are not all known
and, in fact, relatively few excited states of any kind are well specified. If any
discrepancies were to be observed between theory and experiment in this region
it would be difficult to disentangle the effects of deficiencies in the local model
chosen to describe the nuclear structure from any more general failings of the DC
calculations.
The experimental challenges to precise it-value measurements in the two re-
gions are quite different too. For the emitters with A ? 62, the superallowed
transition should be their predominant branch, which in principle simplifies the
branching-ratio measurement; however, they also have half-lives ranging from
120 ms down to about 20 ms, which makes all measurements, but particularly the
mass measurements, very difficult to accomplish at the required level of precision.
The half-lives of the nuclides with 18 :::;; A :::;; 42 make them more tractable, but in
most cases their ,B-branching is also much more complex and requires branching-
STANDARD-MODEL TESTS WITH SUPERALLOWED tJ-DECAY 125
5. Conclusions
I have, I hope, convinced you that the apparent non-unitarity of the CKM matrix,
as derived from superallowed nuclear ,B-decay, is an important problem that affects
a broad range of physics. Clearly, new measurements - including demanding new
mass measurements - are required to improve the statistical definition of the unitar-
ity test. From the perspective of mass measurements, the best short-term prospects
are afforded by the series of even- Z, Tz = -1 nuclei with 18 :<S; A :<S; 42, whose
masses can probably be measured to the required precision with current technol-
ogy. In the longer term, though, it would be highly desirable to develop techniques
to determine the precise masses of the odd-Z, Tz = 0 nuclei with A ~ 62, which
have half-lives of less than 100 ms.
126 J. C. HARDY AND 1. S. TOWNER
Acknowledgements
The work of J. C. H. was supported by the U.S. Department of Energy under Grant
number DE-FG03-93ER40773 and by the Robert A. Welch Foundation.
References
1. Towner, 1. S. and Hardy, J. C., In: P. Herczeg et ai. (eds), Proc. of the 5th Internat. WEIN
Symposium: Physics Beyond the Standard Model, World Scientific, Singapore, 1999, pp. 338-
359.
2. Groom, D. E. et ai., European Phys. J. C 15 (2000), 1.
3. Towner, 1. S. and Hardy, 1. c., In: E. M. Henley and W. C. Haxton (eds), Symmetries and
Fundamental Interactions in Nuclei, World Scientific, Singapore, 1995, pp. 183-249.
4. Hardy, J. C. et al., Nuclear Phys. A 509 (1990),429.
5. Towner,1. S., Hardy, J. C. and Harvey, M., Nuclear Phys. A 284 (1977), 269.
6. Ormand, W. E. and Brown, B. A., Phys. Rev. C 52 (1995), 2455.
7. Sirlin, A., In: P. Langacker (ed.), Precision Tests of the Standard Electroweak Model, World
Scientific, Singapore, 1994.
8. Reich, J. et ai., Nuc!. Instrum. Methods A 440 (2000), 535.
9. Sagawa, H. etal., Phys. Rev. C 53 (1996), 2163.
10. Navratil, P. et al., Phys. Rev. C 56 (1997),2542.
11. Daehnick, W. W. and Rosa, R. D., Phys. Rev. C 5 (1985), 1499.
12. Hagberg, E. et ai., Phys. Rev. Lett. 73 (1994), 396.
13. Hardy, J. C. and Towner, 1. S., In: C. Baktash (ed.), Nuclear Structure 98, AlP Conf. Proc. 481,
1999, p. 129.
14. Harty, P. D. et ai., Phys. Rev. C 58 (1998),821.
15. Kikstra, S. W. et ai., Nuclear Phys. A 529 (1991),39.
16. Koslowsky, v.T. et ai., Nuclear Phys. A 472 (1987), 419.
17. Brindhaban, S. A. and Barker, P. H., Phys. Rev. C 49 (1994),2401.
18. Audi, G. and Wapstra, A. H., Nuclear Phys. A 595 (1995), 409.
Hyperfine Interactions 132: 127-131,2001. 127
© 2001 Kluwer Academic Publishers.
A. H. WAPSTRA
I will make a few observations that I hope will be of some general use.
The first physicist who did a good job in measuring masses was Aston [1]. He
used a combination of electric and magnetic fields and took good care to build a
reasonably focusing mass spectrometer. My experience, in the course of 55 years
in evaluating data, has been that precision measurements with non focusing instru-
ments should be considered with a healthy distrust. This is true both for measure-
ments of atomic masses and for measurements, in electric or magnetic spectrome-
ters, of energies of particles involved in nuclear reactions or decays.
Aston's spectrometer focused in one direction. Later improvements, around
1935, by groups around Mattauch [2], Bainbridge [3], Nier [4] and some oth-
ers introduced focusing not only in two space directions, but also in momen-
tum space, taking care that variations in velocity did not result in objectionable
shifts in spectral line positions. This resulted in so-called double-focusing mass
spectrometers.
Their measurements led to accurate ratios of atomic masses. In order to convert
them to absolute values, it was decided to define an atomic mass unit; the mass of
oxygen was taken to be 16 units.
I use here an element name, not a specific isotope, this because a confusion
arose. Chemists took as standard the average atomic mass of the three stable oxy-
gen isotopes in the 'naturally occurring mixture', physicists that of the isotope 16 0.
Only when it was discovered that the ratio in which the three isotopes occur is not
unique, chemists agreed to change their definition. But they furiously objected to
using 16 0. That would have meant a decrease of 278 ppm in all atomic weights.
This would imply a change of many millions of dollars in the price of all chemicals
sold yearly. Mattauch, the chemist Kohman and I [5] then found out that the change
would be almost precisely lO times less (in the other direction) if 12C was taken to
be 12 units. And this proposal, as you will know, was accepted.
A discussion arose about a name for that unit. I proposed the name 'aston', but
the English especially did not want this. I have the feeling that, for one reason or
another, the person Aston was not popular among them. So the name 'unified unit'
was adopted, symbol u. I remember that some time later a new discussion was
started. I then defended the symbol u; in fact I said "let us be firm in retaining the
u, let us even make it a double u!" And I was rather amazed that nobody seemed to
understand my little joke! (Or, may be, they were trying to be polite ... )
128 A. H. WAPSTRA
The new techniques of measuring masses in Penning traps and storage rings are
so new that I suppose I cannot say anything about it that you do not yet know. But
I must say that I am very impressed by the results and very pleased with them.
The other source of data on atomic masses are measurements of energies of
particles and/or gamma rays involved in nuclear reactions and decays. And it has
been quite interesting to follow the relative importance of the two. The first evalua-
tion of data obtained by such means was the famous collection of papers by Bethe,
Bacher and Livingston [6]. They found a disagreement between the data from the
two sources of information. They decided that the reaction data were then more
dependable than the mass spectroscopic ones.
This was the beginning of a curious competition between the groups involved
in these measurements. I think it happened three times, in the course of time, that
it appeared that the results from the other field were better. And only after some 3
decades some harmony was established.
I have been involved in one of the changes. Around 1955, spectrometers used
in reaction energy measurements were calibrated with the alpha particle energy
in the decay of 21OPO. This energy was accurately measured by Rytz, a person
who did much in the field of precision measurements of particle energies. But
the spectrometer he used was not a focusing one. I [7] decided to investigate the
situation, together with Mattauch and his collaborators in Mainz. We did this by
making a least-squares evaluation in which those reaction energies were treated
as ratios with this calibration energy. The outcome for 21Opo was one part in a
thousand smaller than Rytzes value, some four times larger than his reported error.
I am quite pleased to remember that Rytz then repeated his measurements with
steadily improving accuracy; it is now about 10 ppm. Also, he even recently took
care [16] to give reca1ibrated values for many alpha-particle measurements with
precisions of 5 ke V or better.
Now I want to say something about an important matter for general physics:
the error assigned to a measured value. It is really disturbing how often papers
do not state estimated precisions (errors). Audi and I have some experience in
doing experiments in the field, and we can then make reasonable guesses; still, it
must be considered a serious defect. And then, if they do give errors, they do not
always say what is meant: standard errors (as defined in least-squares estimates),
so-called limits of error or even something different. A special case is the way
mass spectroscopists used to give errors. They repeated their measurements many
times, and calculated errors from the spread of the data. This is a good method, but
it neglects the possibility of systematic errors. Now, the several mass ratios they
published in one paper mostly form an over-determined set. Thus, one may make
a least-squares evaluation of this set. We [8], and sometimes the experimenters
themselves, did so, and the result was that the errors mostly had to be multiplied
by about 2.5 in order to get consistency.
Of course, it is not certain in itself that making this correction to the errors takes
full care of the influence of systematic errors. But in combining mass spectro-
MEMORIES OF MASS DETERMINATIONS 129
scopic data, thus corrected, with data from reactions and decays in a least squares
calculation, the total collection always came out sufficiently consistent.
I was, of course, curious to see, how the Penning trap precision measurements
of the last decades behave in such tests. I was quite pleased to note, that these
analyses did not indicate a necessity for changing the errors assigned to them.
Another quantity entering these least-squares evaluations we also had to treat
with some care. Mass spectroscopic measurements are expressed in the mass units
I mentioned, but reaction energy measurements in electronvolts. Thus, we have to
include the ratio of the two units. The precision in this quantity, at the beginning of
the evaluation work, was not much better than that in the other data. So we also used
it as a variable in our calculation. The results were not of much use for evaluators
of physical constants. Reversely, their work has been of much use to us. I will
therefore mention with pleasure those evaluators: DuMond [9], Cohen [10] and at
present Mohr and Taylor [11]. And in this respect, the following is of interest. The
ratio electronvolt to mass unit can be expressed using the internationally accepted
volt unit, but for a long time it could be done more precisely in standard volt defined
by accepting a standard value for the constant in the Josephson relation between
voltage and frequency. In 1987 [10], the mass-energy ratio expressed in the first
had a precision of 300 ppb, in the latter of only 89 ppb. Thus, we expressed the
results of our mass evaluation in standard electronvolts. In the recent evaluation
of Mohr and Taylor [11], the situation is improved by not less than an order of
magnitude: the precisions are now 40 ppb and 8 ppb respectively. Only for very
few of our output data, using standard electronvolts has still an advantage over
using international electronvolts. But since the ratio of the two volts now differs
from 1 by only as little as 4 ppb, I think we will still use 'standard' electronvolts.
As said, we put the data obtained in a least-squares calculation, which auto-
matically gives best values. So, you may ask: what is your real contribution? Well,
I will give two examples.
In the 1983 atomic mass adjustment [12] it was shown that direct measurements
of masses of mercury isotopes [13] did not agree with directly measured masses
of isotopes of several other elements combined with results of nuclear reaction
measurements. The difference is more than 15 keY compared with errors, on both
sides, of 3 keY or less (see [13, Fig. 1]). Whom do we trust? Well, it appears that
the mass spectroscopic values for odd mass Hg isotopes differ about 5 ke V more
than those for even mass ones. This is explained by the fact that they are compared
with molecules containing the rare l3C isotopes, the other ones not. There seems to
be a dependance of the result on the intensity of the ion beams. Since we feared that
such an effect may have influenced all their data, we decided not to use any of them.
Of course, we then tried to convince people to make new Hg mass measurements
- but this did not lead to new results in the 20 years that we did so.
The other one concerns the alpha decay of 176 Au. Measurements were made by
two groups. The first one [14] reported branches with energies 6260 and 6290 ke V
with the intensities 80% and 20%. The second one [15] found 6228 and 6282 keY,
130 A. H. WAPSTRA
no intensities given. All estimated errors are 10 keV. The difference between the
energies is a bit large, but not disturbingly so. The second group reported the low
energy line to be in coincidence with a 138 keY gamma ray. It was assumed that
this combination feeds the ground-state in the daughter 172Ir, and the other a known
isomer; the energy difference then agrees with the known excitation energy of the
latter.
This seems satisfactory! But my very recent analysis of the new GSI mass
measurements indicated that the resulting energy difference between the 196 Au and
i92Ir groundstates is probably too high. Thus, I felt forced to look at the published
alpha particle spectra. This showed that the intensity ratio of the reported alpha
branches were found drastically different!
This can be explained in two ways. It may be that the strong branch in [14]
is due to a contamination. This, though, is made improbable by the fact that its
excitation curve agrees well with that of the other branch. Or, there may be two
176 Au isomers produced in quite different ratio's in the two experiments. In [14]
the activity is made in a bombardment of 141Pr with 40Ca, in [15] with the rather
different reaction between 92Mo and 88Zr. It is therefore not impossible that [15]
mainly observes the upper 176 Au isomer, whereas the 6260 keV of [14] one occurs
between the ground-states. This assumption would solve the mentioned problem!
Finally, I want to add a remark about the application of the least-squares method.
It is often said to be valid only for Gaussian probability distributions, but this is not
true. The only condition is, that the probability distribution gives an average for
the measured value and for its square: the standard error is the square root of the
difference between the square of the average and the average of the square.
A case in point is the following. In modem mass measurements on radio-active
substances, both in Penning traps and in storage rings, one often cannot separate
isomers. And often the ratio, in which a ground-state and its upper isomer occur, is
not known. The only fair way to derive a value for the mass of the ground-state is
then to assume that the probability distribution of that ratio is flat between 0 and 1.
This allows calculating both its average and its standard error. Thus, one can derive
a value for the mass of the ground-state - that is, if the excitation energy of the
upper isomer is known.
And I want to add a warning. The least-squares distribution gives the 'best'
value for average and error if, of the distribution, only the two conditions outlined
above are known. And for Gaussian distributions, they are indeed the best ones.
But this is not necessarily true for other distributions. For Poisson distributions,
one can derive a best value different from the least-squares one, and with a smaller
error!
References
I. Aston. F. w., Phil. Mag. 39 (1920),611.
2. Mattauch, 1. and Herzog, R., Z. Physik 89 (1934), 786.
3. Nicr, A. 0., Ph),s. Rev. 50 (\936), \041.
MEMORIES OF MASS DETERMINATIONS 131
Abstract. For many nuclei beyond the proton drip line in the Z > 72, N > 82 region, both proton
and a emission are energetically allowed. In the case of some proton emitters, there are a-decay
chains emanating from both parent and daughter nuclei. This means that if the mass excess of one
member of an a-decay chain is known, then the mass excesses for all members of both chains can be
obtained. In addition, proton separation energies may be derived for nuclei in the a-decay chain of the
proton emitter. The method of time- and space-correlations also allows the identification of isomeric
states in these nuclei. As an example, a large number of mass excesses and proton separation energies
for ground and metastable states have been derived from Qa and Qp values obtained from the proton
emitters 165,166, 167 Ir, 171 Au, InTI, and their daughters.
Key words: alpha-decay, atomic mass, drip-line nuclides, nuclear binding energy, proton radioactiv-
ity, Q-values.
* Work supported by the U.S. Department of Energy, Nuclear Physics Division, under Contract
W-31-109-ENG-38.
134 c. N. DAVIDS ET AL.
1. Introduction
The search for examples of proton radioactivity has resulted in the discovery of a
large number of proton emitters in the region 50 < Z < 84 [1]. Many of these
proton emitters and their daughters are also a-emitters, and in some cases the
a-decay chain from the daughter terminates on a nuclide closer to stability whose
mass excess is known. This opens up the possibility of using a- and proton-decay
Q-values to determine the mass excesses of a large group of nuclei connected by
particle decay. The Q-values are derived from the measured kinetic energies of the
emitted protons or a-particles. Where the decay chains are not connected to nuclei
with known mass excesses, proton separation energies can be measured in some
cases and derived in others.
For the a-decay ofthe parent nucleus (Z, A) to the daughter (Z - 2, A - 4), the
energy and momentum relations used to convert between Q-value, mass (M) and
mass excess (ME) are:
2. Experiments at Argonne
A number of proton emitters have been studied at Argonne National Laboratory,
using a double-sided silicon strip detector (DSSD) coupled to the Fragment Mass
Analyzer (FMA). The FMA [2] is a recoil mass spectrometer, separating reaction
products from the primary beam, and dispersing them according to their mass-to-
charge (M / q) at the focal plane. The DSSD allows time- and space-correlation of
implanted ions with their charged particle decays. In cases where the daughter nu-
cleus decays by a emission, extremely clean spectra can be obtained by correlating
MASSES AND PROTON SEPARATION ENERGIES 135
ANL·P·22,OlS
I I I I 1 ,I 1
M (a) A= 167, T<100 ms
-
<:)
..--
8 -
1670S 167
><1> F
6-
~
-
--
<:)
C\J
en
4- -
!z
~
0
() 2- -
.Y--
1 1 1 1 I 1 I
40 .....
167m
-
UP
~ II I I I I I
A
I
1234567
DECAY PARTICLE ENERGY (MeV)
Figure 1. (a) Energy spectrum of decay events in the DSSD from 357 MeV 78 Kr + 92Mo,
after requiring that the decay occurred in the same pixel within 100 ms following a mass 167
implant. (b) Same as (a) with the additional requirements that a second decay event occurred
in the same pixel within 100 ms, having an energy of 6000 ke V, the known a-decay energy of
1660s.
the proton and a decays. Figure 1 shows such a situation, where M / q and time
interval gating are used to select the very weak proton decay channels [3].
A long a-decay chain leads from 1660S, the proton decay daughter of 167Ir, to
150Er, which decays by f3+ IEC emission to 150Ho. The QEC for the 150Er decay
to the 2- state in 150Ho has been measured to be 4108(15) keY [4]. Combining
this with the mass excess of -61950(27) keY determined for the 2- state in 150Ho
at ISOLTRAP [5], we obtain a mass excess for 150Er of -57842(31) keY. Based
on this number and the known Qa values for its a-decay parents [4,6-8], we can
derive mass excesses for the nuclides 154Yb, 158Hf, 162W, 1660s, 170pt, and 174Hg
from Equation (4).
136 C. N. DAVIDS ET AL.
AI{..p.z I,8 11
167
Decay of Ir
Figure 2 shows the decay scheme 167Ir [3]. The proton decay Q-value for 167Ir
[3] enables us to obtain its mass excess, using Equation (5) and the mass excess of
1660S. The ground-state mass excesses of the 167Ir a-decay daughters 163Re, 159Ta,
155Lu, and 151Tm, and that ofthe 112+ 17l Au ground state may be obtained from the
1671r mass excess and the measured a-decay Q-values [3,9]. Note that each ofthese
odd-A isotopes has two a-emitting states, with spin 112+ and 1112-. Table I gives
the experimental mass excesses tied to 167rr by proton and a-decay, along with the
corresponding values from the 1995 Atomic Mass Evaluation (AME95) [4].
As in the case of 1671r, the proton decay of I77TI results in a long a-decay chain
extending from the daughter 176Hg down to 148Dy, whose mass excess has been
measured at 1SOLTRAP [5]. This allows the mass excesses of all the a- and proton-
decay parents of 148Dy to be obtained. The a-decay chain from 177Tl has been
traced down to 165Re [9], permitting mass excess values for it, 169rr, and 173 Au to be
derived. Knowing the Q a for 180Pb [10] also yields a mass excess for 18oPb. Table II
shows these mass excess values, compared with the values from AME95 [4].
also can derive this quantity for the a daughters of proton emitters, if the proton
daughter nucleus is also an a-emitter. This is illustrated in Figure 3 for the case of
171Aum(lI/2-) decay.
Using measured values of Qp and QOI> proton separation energies can be derived
for a large number of a-daughter nuclei of proton emitters. Figure 4 shows these
Sp values, plotted along with predictions from the Liran-Zeldes mass formula [11].
From Figure 4 one can see the beginnings of systematic trends in the proton sepa-
ration energies. There are several sets of pairs of diagonal lines running from upper
left to lower right. Extrapolating down the lower lines of the two central pairs,
one sees that the yet-unobserved nuclides 164Ir and 170 Au are possible candidates
for proton emission. In addition, the N = 82 nuclide 155Ta has a larger proton
138 C. N. DAVIDS ET AL.
1C
" 2 r - - - -.....
171Au
0+ --".--"'--
17°Pt
W _ _.L.--~_
"2
Figure 3. Q-value loop involving the 11/2- states in 171 Au and 167Ir.
0.5 Tm Re
>
CI)
~
>-
...
Cl
CI)
0
c
W
c
.
0 -0.5
1ii
10
Q.
CI)
en ·1
c
0
...
"0
Q.
-1.5
·2+---~---+--~~--+---~--~--~--~r---~
140 150 160 170 180
Mass Number
Figure 4. Ground state proton separation energies for Tm. Lu, Ta, Re, Ir, Au, and Tl. Filled
circles denote measured energies, open circles represent derived values, and solid Jines are the
Liran-Zeldes [II) prediction.
separation energy than might be concluded from its neighboring Ta isotopes. The
Liran-Zeldes prediction [11] seems to systematically over-bind the last proton by
~200 keVin this region of the drip line, except for the Z = 81 proton emitter 177Tl.
MASSES AND PROTON SEPARATION ENERGIES l39
3. Conclusions
Measurements of both proton- and a-decay Q-values for individual proton emitters
have enabled the extraction of the mass excesses for nuclides well beyond the
proton drip line. Knowledge of these mass excesses, along with proton separa-
tion energies, helps to precisely delineate the proton drip line in this region. Both
quantities also allow the testing of mass models at and beyond the proton drip line,
far from the mass region where the models were developed.
References
I. Woods, P. J. and Davids, C. N.,Annu. Rev. NucZ. Part. Sci. 47 (1997), 541.
2. Davids, C. N. et ai., Nuc!. Instrum. Methods Phys. Res. B 70 (1992),358.
3. Davids, C. N. et ai., Phys. Rev. C 55 (1997), 2255.
4. Audi, G. and Wapstra, A. H., Nuc!. Phys. A 595 (1995), 409.
5. Beck, D. et ai., NucZ. Phys. A 626 (1997), 343c.
6. Page, R. D. et al .• Phys. Rev. C 53 (1996), 660.
7. Bingham, C. R. et ai., Phys. Rev. C 54 (J 996), R20.
8. Seweryniak, D. et al., Phys. Rev. C 60 (1999), 031304; Phys. Rev. C 61 (2000), 039902(E).
9. Poli, G. L. et al., Phys. Rev. C 59 (1999), R2979.
10. Toth, K. S. et ai., Phys. Rev. C 60 (1999), 011302.
11. Liran, S. and Zeldes, N., At. Data Nuc!. Data Tables 17 (1976), 431.
,.....
, Hyperjine Interactions 132: 141-146,200l. 141
© 2001 Kluwer Academic Publishers.
Abstract. The transition rates of I':!.L = 0 a transitions in the lightest Po isotopes are consid-
erably slower than the transition rates of the heavier ones, even leading to a considerable devi-
ation in the classical Geiger-Nuttall plot. Evidence will be given that due to particle-hole exci-
tations through the Z = 82 proton shell, the groundstate of the lighter Po isotopes becomes de-
formed and direct a decay to the spherical groundstate of the cOITesponding Pb isotopes is highly
retarded.
1. Introduction
The transition strength of ex decay, one of the oldest observables in nuclear physics,
can be easily interpreted as a tunnelling process through the Coulomb barrier and
the gross features of f:).L = 0 transitions (e.g., between even-even nuclei) leads
to the so-called Geiger-Nuttall law where a linear relation is observed between
the loglO TI/2 and E1/2. The polonium isotopes have often been used as text book
examples and Figure 1 gives the Geiger-Nuttall relation of all even-even Po iso-
topes with neutron number less than 126, the shell closure. From mass 192 on the
loglO TI/2 values start to deviate from the linear behavior. This effect can more
clearly be seen in the reduced ex widths (8 2 ) where the energy dependence from
tunnelling through the Coulomb barrier is divided away. The systematic of the
Z = 82 region is given in Figure 2: the Po nuclei (full circles) definitely deviate
from the neighbouring nuclei.
l ,E+04
l ,E+OO
18 Po
• •
l ,E-04 •
2,2 2 ,3 2,4 2,5 2 ,6
2'~Ea
Figure 1. Geiger-Nuttall plot of the even--even Po isotopes below N = 126. The half-life is
in seconds and the energy in MeV.
1000
Ra
>Q) Rn
.::tt.
A
oS
~
;: 100
~
"Q)
(,)
::J
-0
Q)
II:
10
mation is summarised in Figure 3 for the odd-mass isotopes and in Figure 4 for the
even-mass Po isotopes; here also results from older studies have been included (for
references see [1]).
494
595rr:J
471 374 6963,9.5%, 41 keY
En In IF ,e
7264, 76% ,132 ke V
7316,15%, 15 keY
7378,43%,1.8 keY 7336,90.5'1.21 keY
7540, 11%, 2.5 keY
187mpb 1879Pb
631:"
0[.
6378,48 keY 549
a: 6423,15 keY 6 } :
6949, 22 keY 0
Y,e'
6699,28 keY 0 '
3ye'
9 keY I
757
Y
7003, 26 keY y,e
6606 23 ke V 0 6385, 24 ke V
189mpb 1899Pb 191mpb 1919Pb 193mpb
Figure 3. A survey of fine structure ex decay observed in odd-mass Po isotopes. The energies
and reduced widths (8 2 ) are given in keY. For 189. 191 po also the relative ex feeding (in %) is
given,
2.45(5) ms
-725,767
577 650 0+ 6896 21 (8) keY
L
(0·)
7350 460(250) keY
35 (20)%
532
E+ 70123,3 % 90(18) keY 5 9 1 j
o· 7915 16 keY 7535 96,4 % 51(4) keY
184Pb 65 (20)% 186Pb 188Pb
194PO o· 196PO o.
410 ms 2pOh 5.5s
L
0' 619444 keY
O· 684254 keY
o·
190Pb 192Pb 194Pb
8000
_ Odd-mass Po (h.s.)
Odd-mass Po (I.s.)
>Q) - . - Even-even Po
..l<: 7500
>.
~
Q)
c 7000
Q)
>.
CI:I
u 6500
Q)
'?
d Mid-shell N= 104
6000
eXIstmg. This unique behavior has recently been reported and discussed in [8].
The gradual change in the grounds tate of the even-even Po mother from a pure
spherical state over a mixed configuration towards a pure oblate ground state in
190pO then explains the deviation in the Geiger-Nuttall plot (see Figure 1) and
in the reduced a-decay widths (see Figure 2). By adding the reduced a-decay
widths of the feeding to the excited states to the reduced a-decay widths of the
groundstates (black circles in Figure 2), the total reduced a-decay widths (open
circles) do exhibit a normal behavior as compared to the neighbouring nuclei.
5. Conclusions
Near the Z = 82 shell closure, the properties of the nuclei in their ground state and
in their low-energy excitation are mainly governed by the valence neutrons and by
particle-hole excitations through the proton shell closure. The main decay mode
in the neutron-deficient region is a decay and the decay rate is governed by the
available energy (mass) and by the formation probability of the a particle in the
mother nucleus. This formation probability strongly depends on the overlap of the
wave functions of the initial and final states. Most of the a-decay data presented
in this article has mainly a II L = 0 character and therefore one does not have
to take into account the influence of the centrifugal barriers. As can be seen in
the Figures 2, 3 and 4 enormous differences are present in the a-reduced widths
with the most prominent example the case of 188pO where the excited 0+ state is
more than 10 times strongly fed than the 0+ groundstate implying an important
retardation in the groundstate feeding. This retardation has a severe influence on
the half-life of the a-decaying state with a deviation of the Geiger-Nuttall law as
result. The retardation can be understood by considering the shape evolution in
146 M. HUYSE ET AL.
the Po isotopes when moving towards N = 104, mid-shell for the neutrons. The
nucleus becomes deformed in the grounds tate under the proton-neutron interac-
tion by breaking up the Z = 82 shell closure and creating a number of proton
2-particle-2-hole pairs. As discussed in [5] and [8] the specific behavior under
deformation of the proton orbitals in the neighbourhood of Z = 82 implies that an
oblate shape involves the promotion of one proton pair from below to above thus
creating one proton 2-particle-2-hole pair while a prolate shape involves two or
more proton 2-particle-2-hole pairs. The groundstate of the Pb daughter nucleus
remains spherical and thus has little 2-particle-2-hole character. As consequently
there is only a limited overlap in the wave functions of mother and daughter, the
a-decay to the groundstate will be retarded and the highest a-decay strength will
go to excited states in Pb with considerable 2-particle-2-hole character. In this
region a-decay provides a strong spectroscopic tool for the identification of shape
coexistence.
The onset of deformation around N = 104 could also influence the masses of
the involved nuclei and could eventually be seen in the systematics of the a-decay
energies. It is tempting to interprete the deviation in the a-decay energy of I89pO
(see Figure 5) as due to a shape change from spherical/oblate in the heavier Po
isotopes to prolate as predicted by theory. Extra binding due to the deformation
in I89pO could then lead to a reduced Qa transition energy. But the experimental
evidence for this interpretation is still weak as the deviation could also be due to a
missing a branch to the ground state. Such a transition could be strongly retarded
and thus remained unobserved in the experiment. Further a-decay experiments
combined with in-beam spectroscopy are required.
Acknowledgements
This work was supported by the FWO-Vlaanderen, the GOA programme (Vlaan-
deren, Belgium) and by the Swedish National Research Council (NFR).
References
1. Van Duppen, P. and Huyse, M., Hyp. Interact. 129 (2000), 149.
2. Davids, C. et at., Nucl. Instr. Meth. B 70 (1992), 358.
3. Leino, M. et at., Nucl. Instr. Meth. B 99 (1995), 653.
4. Hofmann, S. Rep. Prog. Phys. 69 (1998),639.
5. Andreyev, A. et at., Phys. Rev. Lett. 82 (1999), 1819.
6. Andreyev, A. et at., Eur. Phys. 1. A 6 (1999), 381.
7. Andreyev, A. et at., In: Proc. of ENPE 99, Sevilla, Spain, AlP Conference Proceedings 495,
1999,p.121.
8. Andreyev, A. N. et at., Nature 405 (2000), 430.
9. Andreyev. A. N. et at., Phys. Rev. C, submitted.
Hyperfine Interactions 132: 147-152,2001. 147
© 2001 Kluwer Academic Publishers.
Abstract. A mass measurement experiment by a time of flight method with the SPEG spectrometer
at GANIL has been performed to investigate the N = 20 and N = 28 shell closures far from stability.
The masses of 31 neutron-rich nuclei in the range A = 29--47 have been measured. The precision
of 19 masses has been significantly improved and 12 masses were measured for the first time. The
neutron-rich CI, S and P isotopes are seen to exhibit a change in shell structure around N = 28.
Comparison with shell model and relativistic mean field calculations demonstrate that the observed
effects arise from deformed prolate ground state configurations associated with shape coexistence.
The evidence of an isomeric state in the 43 S and its interpretation by a shell model calculation confirm
the analysis of the masses and constitutes the first evidence of the predicted shape coexistence around
N=28.
1. Introduction
One of the fundamental questions, which emerge from the study of nuclei far from
stability, concerns the persistence of the magic character of certain configurations
of protons and neutrons. Such magicity breaking has already been observed at the
N = 20 shell closure for neutron-rich nuclei, where an island of inversion has been
pointed out [1-3]. More recently, the determination of the lifetime [4] and of the
148 F. SARAZIN ET AL.
deformation of 44S [5] has indicated that the strength of the N = 28 shell closure
is weakened below igca.
Experimentally, nuclear binding energies are very sensitive to the existence of
shells and may provide clear signatures of shell closures [6]. It was in this spirit
that a mass measurement experiment using a direct time of flight technique was
undertaken to investigate the N = 20 and N = 28 shell closures for nuclei from
Ne (Z = 10) to Ar (Z = 18) produced by the fragmentation of a 60 A MeV
48Ca beam on a Ta target located in the SISSI device. A general description of
the experimental technique, which has already been used at Gani1 to measure the
masses of a large number of neutron-rich nuclei [1, 7], can be found in [8].
From this experiment and its subsequent analysis, the masses of 31 neutron-rich
nuclei in the range A = 29~7 have been measured. The precision of 19 masses has
been significantly improved and 12 masses were measured for the first time. Details
of the analysis and experimental mass excesses deduced from this experiment can
be found either in the already published paper [9] or in [10].
:;- 25
Q)
~ 20
C
N
en
15
Ca
18 20 22 26 28 30
Number of neutrons
Figure I, Experimental 5211 values in the region of the N = 20 and 28 shell closures, The
circles con'espond to values from [II], the bold circles to values for which the precision was
improved and the filled circles to masses measured for the first time [9].
SHAPE COEXISTENCE AND THE N = 28 SHELL CLOSURE 149
s-
III
~24
I:
o
+::
u
... 20
~
8 CI
Gi
~ 16
+15MeV
S
12
+10MeV
4
Si 0
0
0
0 ' CD ' 00 '
0 ~'
u'
~'
u' u'
~'
z: z: z:
16 20 24 28 32
Number of neutrons
Figure 2. Shell corrections as defined in the text of the mass of Si, P, S, CI and Ca isotopes.
The points correspond to experimental values (see caption of Figure I for details), the lines to
shell model calculation [3, 13]
A more direct way to see shell effects on nuclear masses is to subtract from
the mass excesses the contribution of the macroscopic properties of the nuclei.
Here we have used the finite range liquid drop model of [12]. The difference -
the microscopic or shell correction energy - is plotted in Figure 2, together with
predictions of the shell model [3, 13].
As in Figure 1, the qualitatively different behaviour of the Sand P isotopes as
compared to the Ca isotopes is clearly evident. The Ca isotopes show pronounced
shell correction minima around N = 20 and N = 28. The Sand P isotopes do
not exhibit such effects at N = 28, but instead show a discontinuity in the slope at
N = 26. The observed trends are well reproduced by large scale shell model (SM)
calculations undertaken within the sd-fp model space [3]. A similar agreement was
obtained in relativistic mean field (RMF) calculations [10, 14].
""
70
60
50
10
0
0
Energy (keV)
Figure 3. Experimental spectra for the isomeric state in 43S. The lower inset compares the SM
predictions and the experimental values (see [20] for the 940 keY state); the numbers beside
the transitions are B(E2) values in units of e 2 fm4.
Shape coexistence is often associated with isomerism (see, for example, [17]).
In the present experiment, we observed such an isomeric state in 43S, thanks to a
47T NaI array surrounding the detector telescope providing the identification of the
nuclei. In a subsequent experiment employing a much shorter flight path ("-'43 m)
[18], both the energy of the y-rays and the half-life were measured by delayed
coincidences between two Ge-detectors and a Si-telescope. A single transition with
Ey = 319 keY and a lifetime of 478(48) ns was observed (Figure 3).
Energy and lifetime appear to be consistent with an E2 transition and a B(E2)
of 0.357(0.036) e2 fm4, or 0.04 Weisskopf units. Such an isomeric state is pre-
dicted by SM calculations similar to those of [3] as a shape isomer, whereby the
weak strength of the transition arises from the two very different wave functions
involved. A recent measurement of the Coulomb excitation of 43S revealed a new
level at 940 keY with a B(E2) of 175(69) e2 fm4 [20). The SM predictions are
compared to experiment in Figure 3. As can be seen both the transition strengths
and the energies of the levels are well reproduced.
4. Interpretation
If we consider the ground state of 43 S in the case of a closed f7/2 shell, it would
correspond to a 7/2- neutron hole. The SM, however, predicts an inversion of
SHAPE COEXISTENCE AND THE N = 28 SHELL CLOSURE 151
the 7/2- and the 3/2- levels, with the former being the isomeric state. Analysis
of the SM wave functions reveals that 48% of the 3/2- ground state corresponds
to coupling of a neutron to a deformed core of protons in a 2+ state. A similar
conclusion was reached by [19]. Moreover, RMF calculations [9] predict prolate
deformed ground states in this region. In addition, deformation is suggested by the
measured B(E2) values [5, 20]. The discontinuity observed at N = 26 (Figure 1)
can now be understood in a simple Nilsson picture. For a prolate deformation of
fh ~ 0.2, a large gap appears between the lowest three orbits and the fourth orbital
arising from the 1f7/2 and higher orbitals. Consequently, a pseudo shell-closure due
to deformation can be considered to appear at N = 26. Oblate deformations would
not be compatible with these observations.
One may see in Figure 4 that the average occupation number of the f7/2 for
N = 28 nuclei in the shell model calculation reflects the configuration mixing of
the different shapes. Interestingly, in the RMF calculations, it is possible to extract
it as a function of the constraint applied on the shape of the nuclei. Then, one can
see in the figure that this average occupation number is very comparable when the
nuclei are constraint to be of an oblate shape or when they are predicted to be
spherical. A "closed shell" configuration is then believed to persist (as an excited
state) and is still consistent with the observation of the isomeric state in 43S.
References
l. Orr, N. A. et ai., Phys. Lett. B 258 (1991 ), 29.
2. Warburton., E. K. et ai., Phys. Rev. C 41 (1990), 1147.
152 F. SARAZIN ET AL.
E.ROECKL
Gesellschaft fur Schwe rionenforschung, D-6429 J Darmstadt, Germany; e-mail: e.meekl@gsi.de
Abstract. By using heavy-ion induced fusion-evaporation reactions at the on-line mass separator
of GSI, decay properties of neutron-deficient isotopes between 56Ni and lOOSn were investigated.
Recent experimental results will be presented and discussed, with particular emphasis on the role of
masses, Q values and separation energies.
Key words: alpha-decay, atomic mass, beta-decay, N-Z nuclides, nuclear spectroscopy.
1. Introduction
In 1976, i.e., almost a quarter of a century ago, the Third International Conference
on Nuclei Far From Stability was held here at the Institut d'Etudes Scientifique
in Cargese. Meeting at the same scene of action again, with a similar topic of
action, is an appropriate occasion to review the topic of masses, decay Q values
and separation energies by taking N '" Z nuclei as examples.
The study of decay properties at and near N = Z in general, and of nuclei
between the double shell closure at 56Ni and lOOSn in particular, is of great current
interest to nuclear physics and astrophysics. The astrophysical 'links' involve, e.g.,
the detection of solar neutrinos [I, 2], the rp process [3], and the electron-capture
cooling of supernovae [4]. As far as nuclear physics is concerned, the main points
of interest are, e.g., the proton dripline, the island of ot emission beyond lOOSn,
special microscopic effects due to protons and neutrons occupying identical orbits,
and the high energy release in f3 decay. The detailed nuclear-physics aspects, as
well as their astrophysical relevance, cannot be discussed within the scope of this
report, but can be found in the references cited throughout the text.
There has been a recent upsurge of both theoretical and experimental work on
decay properties of N '" Z nuclei, measurements being in particular carried out
at ISOLDE/CERN, GANIL and GSI. The present report deals mainly with work
performed in the latter laboratory by using the isotope separator on-line (ISOL)
to the heavy-ion accelerator UNILAC. The paper is structured as follows: After a
presentation of the techniques used at ISOL, recent experimental results will be
discussed for direct charged-particle decay and for f3 decay, with particular empha-
sis on the determination of decay Q-values; finally, a summary and an outlook will
be given.
154 E.ROECKL
2. Experimental techniques
In this way, one can, e.g., distinguish between fJ+ and EC transitions preceding
proton emission, determine the (QEC - Sp) value for a selected level of the final
nucleus populated by proton transitions, deduce information on the lifetime of the
proton-emitting levels by means of the proton X-ray coincidence technique [11],
and use proton-y coincidence data to identify excited states in the final nucleus.
10
114
Sa 110X
8 ~ /9
>Q)
~
0
6
'<t
CIl 4
C
:::J
0
() 2
0
2000 3000 4000 5000 6000
Energy (keV)
Figure 1. Triple a chain involving the decays of 114Ba, 110Xe and 106Te.
156 E.ROECKL
u 9000
IJJ
D, Cano-Ott. PhD thesis (2000)
°8500
¢
8000
o
o
7500
: : {(tP-"'
7000 ............ ,..... .................. , ...... " ........ , ............. ,...... ,... ,qQ..J;rf}J', ..........- - - - - - - ,
: region of Interest : QEC (exp. val.)
.;.
.
'~~
8500 7372t27 keY
6000 ....... .- ............. ......... " .....!.. .
5500
3000 4000 5000 6000
energy (keV)
Figure 2. Determination of QEC values from TAS data: The case of fs-oHo (2-) [21).
To reliably determine the endpoint energies (QEC - Sp) of the continuous ,8-delayed
particle spectra, which are typical of heavy nuclei, is as difficult if not dubious a
task as to find the endpoint of a ,8+ spectrum. As has been recognized since the
early days of ,8-delayed particle spectroscopy, it is preferable to take advantage
of the energy dependence of the statistical rate function for ,8+ and EC transition.
This method is similar to ECI,8+ -ratio procedure discussed in Section 3.2.1. The
(QEC - Sp) mass-links offer a powerful method to measure masses of very neutron-
deficient isotopes, provided the relevant levels in the intermediate and final nuclei
can either be unambiguously identified (or safely approximated as being closely
spaced and continuously populated). The latest ISOL works published on this topic
deal with I13Xe and 114Cs [25, 26]. A comparison of ,8+ -coincident and singles
proton spectra yielded (QEC - Sp) values of 7920(150) and 8730(150) keY, respec-
tively. While the evaluators [27] accepted the former result to deduce the mass of
113Xe from that of I 12Te, the 114CS result has been rejected. This is probably due to
the fact that the authors of [26] discussed the problem of the ambiguity in assigning
158 E.ROECKL
>Q)
830
C 820
Off, 810
60
z:.
'iii 50
t:
Q) 40
E
ro 30
Q)
CD 20
10
0
0 2 3 4 5 6 7 8
Excitation Energy (MeV)
Figure 3. Lower panel: Beta intensity for the decay of 98 Ag as a function of 98Pd excitation
energy, obtained from the de-convolution of the experimental TAS spectra (solid line) and
from the Cluster Cube data (shaded area). The f3 intensity is defined as the (f3+ + EC) feeding
in percent per decay, integrated over an excitation-energy interval of I MeV. Upper panel:
QEC value derived from a joint least -squares analysis of the TAS data (horizontal dashed line),
and QEC values determined on the basis of IEC/ 1f3+ ratios for the different excitation-energy
intervals (squares). The error bars stem from the uncertainties of the individual QEC values,
which are mainly determined by the uncertainty of the integrated EC/total ratio. See text and
[24] for details.
the observed proton-coincident y rays to 113 1 levels, and added the following state-
ment: "Because of the dubiety of this assignment, we have included a contribution
from this source into the final uncertainty". It is hoped that such ambiguities will
be remedied as soon as TAS data will become available that were recently taken
for 96 Ag, lOoln [28], 113Xe, 114CS, 117Ba and ll9Ba [17]. The advantage of these
measurements is that, in contrast to earlier charged particle-y experiments, the TAS
allows one to identify the levels in the final nucleus that are populated by tl-delayed
particle emission. In this way, the occurrence of an excited 686 keV level of 95Rh
(N = 50) has preliminarily been interpreted, with reference to a shell-model
calculation [29], to be the hitherto unobserved single-proton rr g7/2 configuration.
assignment for the parent states, the Gamow-Teller resonance and its quenching
all the way to a waiting-point nucleus of the astrophysical rp process. These in-
vestigations are part of an ongoing research programme which includes a study
of the 12+ 'yrast-trap' in 52mPe [30], an accurate determination of the branching
ratio for the transition from the 58CU ground-state to the 58Ni ground-state [31],
decay measurements of the Tz = -1 odd-odd nuclei 56CU [32] and 60Ga [33] and
of the N = Z odd-odd nuclei 70Br [34] and 94Ag [35, 36], a re-investigation of
the ,B-delayed proton decay of 57Zn [39], and a first study of ,B-delayed y-rays of
61Ga [37] and lOoln [36], and the identification of ,B-delayed protons and y-rays
emitted from the the rp-process waiting-point nucleus 93Pd [38]. In all these cases,
the 'experimental' determination of ,B-transition rates involves QEC values derived
from systematics and thus represents a semi-empirical approach only. Therefore, it
is highly desirable to get experimental QEC and Sp values for these N ~ Z nuclei.
All in all, a wealth of ,B-decay data has become and will continue to become
available for N ~ Z nuclei between the double closed-shell nuclei 56Ni and IOOSn.
Among the ISOL experiments planned for the immediate future, there are a TAS
,B-strength measurement of 52mpe, an attempt to identify the ,B or isomeric decay
of 991n, a proton-hole nucleus with respect to IOOSn, and a search for direct proton
decay of 105Sb and ll3Ba. It is indeed amazing to see the high data quality that can
be obtained for nuclei close to the proton drip line, and it is also encouraging to
observe what could be called, at least in the authors's judgement, a renaissance in
the 'interface' between in-beam and decay spectroscopy.
Last not least, I would like to thank the APAC2000 organizers, and to wish
them (and us) good luck also with the conference barbecue ('mechoui') which
evidently deserves special attention: The Preface of the Proceedings of the 1976
Cargese Conference Cargese 1976, acknowledges the support "of many people
from the Cargese village", and names a few "who nearly rescued the mechoui from
a disaster. .. ".
References
1. Trinder, W. et aI., Study of the fJ decays of 37 Ca and 36Ca, Nucl. Phys. A 620 (1997),191-213.
2. Liu, W. et aI., Beta decay of 40Ti and 41Ti and implication for solar-neutrino detection, Phys.
Rev. C S8 (1998), 2677-2688.
3. Schatz, H. et at., rp-process nucleosynthesis at extreme temperature and density conditions,
Phys. Rep. 294 (1998), 167.
4. Caurier, E. et at., Shell-model calculations of stellar weak interaction rates: 1. Gamow-Teller
distributions and spectra of nuclei in the mass range A = 45-65, Nucl. Phys. A 653 (1999),
439.
5. Kirchner, R. et at., Intense beams of mass-separated neutron-deficient indium, tin, thallium and
lead isotopes, Nucl. [nstr. Meth. Phys. Res. A 234 (1985),224-229.
6. Kirchner, R. et at., An ion source with bunched beam release, Nucl. [nstr. Meth. Phys. Res. B
26 (1987),204-212.
7. Kirchner, R., On the thermoionization in a hot cavity, Nucl. [nstr. Meth. Phys. Res. A 292
(1990),203-208.
160 E. ROECKL
8. Hu, Z. et aI., Energy and efficiency calibration of an array of six Euroball Cluster detectors
used for beta-decay studies, Nucl. Instr. Meth. Phys. Res. A 419 (1998), 121-131.
9. Keller, H. et ai., A summation-free f3+ -endpoint spectrometer, Nucl. Instr. Meth. Phys. Res. A
300 (1991), 67-76.
10. Karny, M. et ai., Coupling a total-absorption spectrometer to the GSI on-line separator, Nucl.
Instr. Meth. Phys. Res. B 126 (1997), 411-415.
11. Hardy, J. C. et ai., Nuclear lifetimes in the region of 10- 16 sec measured by a new technique,
Phys. Rev. Lett. 37 (1976), 133.
12. Klepper, O. et aI., Direct and beta-delayed proton decay of very neutron-deficient isotopes
produced in the reaction 58Ni + 92Mo, Z. Phys. A 305 (1982), 125-130.
13. Shibata, M. et ai., Beta decay studies of 107Sb and other neutron-deficient antimony isotopes,
Phys. Rev. C 55 (1997),1715-1723.
14. Schardt, D. et aI., Alpha decays of neutron-deficient isotopes with 52:(Z:(55, including the
new isotopes lO6Te (T1/2 = 60 ~s) and llOXe, Nucl. Phys. A 368 (1981), 153-163.
15. Guglielmetti, A. et ai., Nonobservation of 12C cluster decay of 114Ba, Phys. Rev. C 56 (1997),
R2912-R2916.
16. La Commara, M. et aI., Production of very neutron-deficient isotopes near lOOSn via reactions
involving light-particle and cluster emission, Nucl. Phys. A 669 (2000), 43-50.
17. GSI Experiment U180 (Z. Janas et al.), unpublished result; Janas, Z. and Mazzocchi, c., private
communication.
18. Roeckl, E., Alpha decay, In: D. N. Poenaru (ed.), Nuclear Decay Modes, Institute of Physics
Publishing, 1996, pp. 237-274.
19. Keller, H. et ai., f3+ -endpoint measurements near lOOSn and 146Gd, Z. Phys. A 340 (1991),
363-370.
20. Keller, H. et ai., Beta-decay energy and atomic mass of 148Tb, Z. Phys. A 352 (1995), 1-2.
21. Cano-Ott, D., Ph.D. Thesis, University of Valencia, 2000, unpublished.
22. Beck, D. et ai., Direct mass measurements of unstable rare-earth isotopes with the ISOLTRAP
mass spectrometer, Nucl. Phys. A 626 (1997), 343c-352c.
23. Hu, Z. et ai., Beta decay of 97 Ag: Evidence for the Gamow-Teller resonance near 100 Sn, Phys.
Rev. C 60 (1999), 024315-1-024315-17.
24. Hu, Z. et ai., Beta decay of 97 Ag: Evidence for the Gamow-Teller resonance near lOOSn, Phys.
Rev. C 62 (2000), 064315-1-064315-9.
25. Plochocki, A. et ai., Measurement of proton separation energies close to the proton drip line in
the antimony-cesium region, Phys. Lett. B 106 (1981), 285-288.
26. Tidemand-Petersson, P. et aI., Beta-delayed particle emission from neutron-deficient isotopes,
Nucl. Phys. A 437 (1985), 342-366.
27. Audi, G. et ai., The Nubase evaluation of nuclear and decay properties, Nucl. Phys. A 624
(1997), 1-124.
28. GSI Experiment U176 (Batist, L. et ai.), unpublished result; L. Batist, private communication.
29. Johnstone, 1. P. and Skouras, L. D., Particle-hole excitations in N = 50 nuclei, Phys. Rev. C 55
(1997), 1227-1230.
30. Gadea, A. et ai., Gamma and f3 decay of the 12+ isomer in 52Fe, In: D. Rudolph and
M. Hellsrrom (eds), Pmc. Int. Workshop on N = Z Nuclei (PINGST 2000), lnt. Rep. of Lund
University, 2000, in print.
31. Janas, Z. et ai., GSI Sci. Rep. 1999, GSI-2000-1, 2000, p. 13.
32. Ramdhane, M. et ai., Beta decay of 56Cu, Phys. Lett. B 432 (1998), 22-28.
33. GSI Experiment U 173 (Gadea, A. et ai.), unpublished result; Z. Janas and C. Mazzocchi, private
communication.
34. Doring, J. et ai., Beta-decay study of the N = Z odd-odd nuclei 62Ga and 70Br, In: D. Rudolph
and M. Hellstrom (cds), Pmc. Int. Workshop on N = Z Nuclei (PINGST 2000), Int. Rep. of
Lund University, Lund, 2000, in print.
N ~ Z NUCLEI 161
35. Schmidt, K. et al., Decay properties of the new isotopes 94 Ag and 95 Ag, Z. Phys. A 350 (1994),
99-100.
36. Mazzocchi, C. et aI., Decay properties of N ~ Z nuclei below lOOSn, In: D. Rudolph and
M. Hellstrom (eds), Proc. Int. Workshop on N = Z Nuclei (PINGST 2000), Int. Rep. of Lund
University, Lund, 2000, in print.
37. Oinonen, M. et al., Beta decay of 61Ga, Eur. Phys. J. A 5 (1999), 51-156.
38. Schmidt, K. et al., Beta decay of 93 Pd, Eur. Phys. J. A 8 (2000),303-306.
39. GSI Experiment U173 (Janas, Z. et al.), unpublished result; A. Jokinen et aI., private
communication.
Hyperfine Interactions 132: 163-175,2001. 163
© 2001 Kluwer Academic Publishers.
Abstract. Based on the use of a single ion, isolated at the center of a cryogenically cooled Penning
trap, an environment is produced which makes this mass spectrometer remarkably free of systematic
errors. The most notable developments in our quest for an ultra-high accuracy instrument were (a) the
compensation of the trapping potential, (b) the discovery that motional sidebands could manipulate
radial energies, (c) the use of multiply-charged ions that could improve signal-to-noise, and (d) the
use of an ultra-stable superconducting magnet/cryostat system with drift <0.010 ppblh. The domi-
nant systematic errors are associated with radial electric fields caused by image charges in the trap
electrodes and with the rf-electrical drive field used to determine the harmonic axial resonance. To
illustrate the potential of this improved spectrometer, the four-fold improved measurement of the
proton's mass and the eight-fold improved measurement of oxygen's atomic mass will be described.
Key words: mass spectrometer, Penning trap, proton mass, oxygen atomic mass.
endcap
insulator
(macor)
split guard
ring
3.71 em
split guard
insulator
endcap
skimmer
reOector
Figure 1. Cross-sectional view of our current Penning trap. Critical alignment of the gold
plated phosphor bronze electrodes is maintained by the seating of the macor insulator
rings within the electrode registration grooves. The entire assembly is held together with
spring-loaded alumina rods (not shown).
The basic operation of this instrument is described in some detail in the lit-
erature [2-4]. However, to appreciate the improvements that will be described,
it is necessary to outline the origin of the detected signal. The motion of the
ion along the trap's axis of symmetry (at liz) can be excited by a frequency syn-
thesizer, causing current to be induced in the endcaps of the trap. One of these
electrodes is used to observe this motion by means of an attached LC circuit that
is tuned to liz = J q Vol md 2 12rr, where Vo is the ring-endcap potential difference,
d = 0.211 cm is the characteristic trap dimension, and qlm is the ion's charge-
to-mass ratio. The resulting signal voltage is amplified by a cryogenically cooled
preamp and mixed with the original phase-shifted drive voltage to generate an error
signal. If something happens to shift liz, the error signal becomes non-zero and is
integrated to produce an offset that moves the ring voltage supply to a new value
that again forces the error signal to zero. In this way, the ion's axial motion is
kept 'frequency locked' to the stable drive synthesizer and the resulting correction
ULTRA-PRECISE MASS MEASUREMENTS 165
-.-=~'.-
Figure 2. The bracketed cyclotron resonance for a single C4 + ion using our extrapolated trig-
ger anharmonic detection method. The superimposed straight line segments are least-squares
fitted lines and the typical 'linewidth' (between comer frequencies) is 0.1-0.2 ppb. Also, full
sweep range is one ppb and typical sweep time for each direction is 200-300 sec.
(2)
The stability of the magnet/cryostat system can strongly influence the ultimate
resolution that can be achieved by any mass spectrometer. Yet for years, it was a
mystery why superconducting magnets were only stable to few ppb/h. Experimen-
tally, we found the culprit to be any material in the high-field region of the magnet
that had a large inverse temperature dependence in its magnetic susceptibility [11].
If such a material was not fixed in temperature by the liquid helium, then changes
in boil-off (due to the wandering atmospheric pressure) changed its temperature. To
fix this problem, we had a new custom-designed magnet/cryostat constructed that
has been observed to have drifts less than 0.010 ppb/h for up to three consecutive
weeks on occasion. To illustrate the exceptional stability of this magnet/cryostat
system, refer to the data in Figure 3 for which a very small quadratic of 0.010(1)
ppt/h2 was remove for the last nine days of a 33-day run (for which the overall
!
~
-4
c -6
g -8
0
U 600 650 700 750 800
>.
0
time (hours)
Figure 3. Cyclotron resonance data taken using a single 0 6 + ion in the new magnet system.
The data are fitted to a very small quadratic drift near the end of a long run.
168 ROBERT S. VAN DYCK ET AL.
drift was less than 10 ppt/h). For this new system, liquid helium now provides the
temperature control for all critical surfaces (instead of the inferior control produced
from the boil-off vapor employed in earlier designs) and an active stabilization
of the pressure over the liquid helium keeps the liquid's temperature reasonably
constant. Furthermore, by using an internal superconducting flux-stabilization coil
[12] and an external active magnetic compensation system (i.e., Helmholtz coils
and custom-fabricated flux gate sensor), this system obtains an overall shield factor
that exceeds 104 for sufficiently distant field disturbances. See [13, 14] for detailed
description.
0.03
• carbon 4+ o
,..., o DD
N o proton
0.02
b
'"
~
;::l
"0
0.01
.;;;
...
0)
>.
<)
0
c
0)
::>
-0.01
]"
c
0
!:I
-0.02 cPo
o
.9 Jon Mass Ratio - 2.977,783,718,335(74)
<) o 0
>.
<) -0.03
0 50 100 150 200 250 300
time (hours)
Figure 4. The cyclotron frequency residuals versus time for a single C4+ ion and a single
proton. By fitting data, obtained from the bracketed comer frequencies, to a common drift, the
fitted cyclotron frequency ratio (as shown) allows all data to be displayed on the same record.
ULTRA-PRECISE MASS MEASUREMENTS 169
N
::x:: 30
-5 single carbon 4+
"<t 25
r-:
~
trl 20
trl
v)
r-- 15
0
0
«l
10
&
~ 5
g
~
0
0
u» 0 0.1 0.2 0.3 0.4 0.5 0.6
u
axial drive in micro-watts (arb. ref.)
Figure 5. The cyclotron frequency of a single C4 + ion is measured as a function of the axial
drive power. The strong dependence reflects the use of drive powers much larger than normally
used as well as the use for detection of a larger than normal value of the anharmonic C4
coefficient (~6.7 x 10- 5 ). The uncertainty of the linear fit extrapolated to zero power is
~2 x 10- 11 .
ing resonance for the cyclotron resonance in the excitation/detection sweeps and
measuring Vm at the extremes of C4 = ±5.6 X 10-5 , one could then determine
8vm/v m versus axial drive power. By extrapolating to zero axial drive, one can use
the residual 8v m /v m to estimate an upper limit of 0.6 meV for the C4+ magnetron
energy (see [13]). On the other hand, the C4 + theoretical cooling limit is given by
(vm/vz)Ez(thermal) < 0.06 meV and the proton has a limit three times smaller
still. Thus, for these two ions, it is safe to assume IEm I < 1 meV for each ion. This
implies a relative error due to magnetron energy to be < 0.04 ppb in the final mass
ratio.
There is also an important effect due to axial drive which can be observed by
simply varying this drive power and observing its effect on v~, see Figure 5. This
procedure allows us to correct the cyclotron data for the actual axial drive used in
the detection, which at that time yielded a 0.02 ppb uncertainty for C4+. However,
this method could not be used for protons since significantly higher axial drives
caused too much loss of resolution. To keep the magnitude of the possible effects
the same for both ions, we reduced axial drive by 7-11 dB (to compensate for
change in axiallinewidth). In addition, we estimated its possible relative magnitude
by using the first-order perturbation relations, from which we conclude a possible
0.07 ppb uncertainty in the final comparison.
There is also a concern that available cyclotron energy could produce a signif-
icant shift in v~. This concern is based on the magnitude of Ec which is needed
to see an axial shift above the detection noise, typically given by Ec(det) :( 1 eY.
However, the frequency shift detection method relies on the fact that the expected
relative width of the cyclotron resonances is quite small, typically < 10- 11 due to
residual external magnetic noise and thermal axial noise coupled into the cyclotron
resonance. Thus, the cyclotron drive frequency must be essentially right on the
170 ROBERT S. VAN DYCK ET AL.
Mp = 1,007,276,466.89(14) nu,
where 1 nu = 10-9 unified atomic mass units. This result contains a final 0.14 ppb
possible uncertainty and it represents at least a factor of four improvement over any
previous determinations [15, 16].
8v e (0 6 +) :;:::j
[3.2]Ez - [0.042]Ec + [2.1]Em' (3)
Vc
8v c (C 6+) :;:::j
[3.0]Ez - [0 .080]Ec + [2.7]Em' (4)
Vc
where C4 ~ -5 X 10- 5 and B2 ~ 0.5 G/cm 2 . One can obtain an estimate for
Ec by using our experimental observation that the comparison of C 6 + /C4+ has
yielded the expected value of 12.0 on the average to within 0.007(7) ppb. One
can then show that the above perturbation relations for C6+ and C4+ would predict
ULTRA-PRECISE MASS MEASUREMENTS 171
,-..,
N
::r: 1
.,.,E- 0.8 single carbon 6+
.,.,
~
0.6
v5
'"'"
,..)
0.4
~ 0.2
""'"
tri 0
""'"
>. -0.2
u
eCI) -0.4
;:I
cr
CI) -0.6
~
e -0.8
g
..Q -1
u 0 0.2 0.4 0.6 0.8
>. 1.2 1.4 1.6
u
cyclotron drive in nano-watts (arb. ref.)
Figure 6. The cyclotron frequency of a single C 6+ ion is measured as a function of the
cyclotron drive power. The dependence on cyclotron power is extremely weak, even with the
nominal choice of C4 = 5 x 10- 5 . There is negligible shift «O.003-ppb) for typical cyclotron
drives that have zero effective 'linewidth' (i.e., between respective corner frequencies).
a systematic shift at least ten times larger if the fractional error in the comparison
were determined by Ec(det). Therefore, it follows that the critical cyclotron energy
used in the above perturbation relations should be less than Ec(det)/IO '"'-' 0.1 eY.
We have checked experimentally on the shift due to the strength of our cyclotron
drive by increasing it by 6-9 dB above the weakest drives for C6+ (see Figure 6).
We obtain a shift of -0.0029(7) ppb when extrapolated to zero cyclotron drive
[corresponding to Ec = 0.036(9) eV]. As a result, with IEm I :s; 1 meV and
Ec :s; 100 meV for these ions, we see that the sum of the last two terms in each
expression in Equations (3)-(5) yields a comparable relative shift that is each less
than 0.01 ppb.
As in the case of H+ /C4+ comparison, the dominant relative shift is associated
with the energy of the driven axial motion and is approximately 0.03-0.05 ppb.
However, unlike H+, we can determine these shifts directly (as described in Sec-
tion 3 for C4 +) for every ion in every run, yielding an uncertainty <0.010 ppb. (The
typical uncertainty is generally an order of magnitude below the actual corrections
applied to the data.)
In the course of still further inspection, we verified that errors often existed
in setting the phase of the detection circuit. The corresponding effect has been
determined to be comparable to that produced by the axial drive. The problem
is that such an error will cause a shift in what we believe to be the 'true' axial
frequency. Basically, the detection system locks the drive frequency to the point in
the axial resonance where the error signal is zero. However, the observed cyclotron
frequency depends on the frequency of the free motion through the shift due to the
magnetron rotation. Unfortunately, we do not know the latter, only the former and
172 ROBERT S. VAN DYCK ET AL.
ml) =(ml)
(m2 m2
true calc
'{I+fh~</J(m2-1)},
ml (7)
(8)
where q] is the charge on ion #1. This relation yields a shift of 0.12 ppb for the
oxygen/carbon comparison with an uncertainty in the applied correction of 0.013
ppb.
Figure 7 gives an example of the runs that compare cyclotron frequencies of
oxygen and carbon ions. Because hundreds of hours are generally used in such
runs and there is often a slow wander in the data that does not correlate with any
of the parameters that we monitor, we often group or re-bin the data in different
ways (using linear and quadratic fits) to obtain a realistic mean and error for each
run. The example shown utilizes all the data in a quadratic fit. Table I summarizes
the error budget for all the sources described above except the common error due
to the image charge shift for all of the recent oxygen runs. The combined error
ULTRA-PRECISE MASS MEASUREMENTS 173
Table I. Summary of atomic mass calculations for all oxygen runs in 1999 and 2000 with the three
major error contributions listed in ppt (10- 12 )
for each run is obtained from the quadrature sum of its statistical error, its axial-
power extrapolation error, and its phase-uncertainty error. Table I also summarizes
the value of the atomic mass computed for each run as well as the final weighted
average of all the runs. When the common error associated with the uncertainty of
the image charge shift is taken into account, we obtain:
M C60) = 15,994,914,619.18(25) nu.
This result agrees quite well with the previous 1995-MIT measurement of this
atomic mass [15], but is more than eight times as accurate. It also agrees well
with the value obtained from the 1993 and 1995 Atomic Mass Evaluations by Audi
and Wapstra [18].
As a future possibility, a trap 1.6 times larger could be used that would see the
image charge shift reduce by four times; and since the linewidth will be only half
as large in such a trap, the shift due to the axial phase error would be reduced as
well. Alternatively, the phase error could be reduced still further by getting better
statistics on the fit to the measured line shape and more precise measurements of the
image charge shift could be made. In addition, one could consider alternating axial
drive detection and cyclotron excitation to eliminate the extrapolation to zero axial
drive. However in that case, the ultimate limitation will undoubtedly come down to
statistical uncertainty, with the increased demand on the magnet stabilization loops
to keep the drift in the field constant. Short of even-more Draconian methods, we
suspect that 0.01 ppb may be near the practical limit in accuracy for the present
version of the UW-PTMS.
References
I. Brown, L. and Gabrielse, G., Rapid Communications of Phys. Rev. A 25 (1982), 2423-2425.
2. Moore, F. L., Brown, L. S., Farnham, D. L., Jeon, S., Schwinberg, P. B. and Van Dyck, Jr.,
R. S., Phys. Rev. A 46 (1992), 2653-2667.
3. Van Dyck, Jr., R. S., Farnham, D. L. and Schwinberg, P. B., In: Physica Scripta T59, Pro-
ceedings of Nobel Symposium on 'Particle Traps and Related Fundamental Physics', Lysekil,
Sweden (1995),134-143.
4. Van Dyck, Jr., R. S., In: R. Hulet and B. Dunning (eds), Atomic, Molecular, and Optical
Physics: Charged Particles, Experimental Methods in the Physical Sciences 29A, Academic
Press, New York, 1995, pp. 363-389.
5. Gabrielse, G., Phys. Rev. A 27 (1983), 2277-2290.
6. Brown, L. S. and Gabrielse, G., Rev. Mod. Phys. 58 (1986),233-311.
7. Farnham, D. L., A determination of the proton/electron mass ratio and the electron's atomic
mass via Penning trap mass spectroscopy, Ph.D. Thesis, University of Washington, Seattle,
1995.
ULTRA-PRECISE MASS MEASUREMENTS 175
8. Van Dyck, Jr., R. S., Wineland, D. J., Ekstrom, P. A. and Dehmelt, H. G., Appl. Phys. Letters
28 (1976),446-448.
9. Van Dyck, Jr., R. S., Schwinberg, P. B. and Dehmelt, H. G., In: New Frontiers in High Energy
Physics, Plenum, NY, 1978, pp. 159-181.
10. Cornell, E. A., Weisskoff, R. M., Boyce, K. R. and Pritchard, D. E., Phys. Rev. A 41 (1990),
312-315.
11. Salinger, G. L. and Wheatley, J. c., Rev. Sci. Instrum. 32 (1961),872-874;
Lockart, J. M., Fagaly, R. L., Lombardo, L. W. and Muhlfelder, B., Physica B 165 & 166
(1990), 147-148.
12. Gabrielse, G. and Tan, J., 1. Appl. Phys. 63 (1988),5143-5148.
13. Van Dyck, Jr., R. S., Farnham, D. L., Zafonte, S. L. and Schwinberg, P. B., In: Trapped Charged
Particles and Fundamental Physics, AlP Conference Proceedings 457, 1999, pp. 101-110.
14. Van Dyck, Jr., R. S., Farnham, D. L., Zafonte, S. L. and Schwinberg, P. B., Rev. Sci. Instrum.
70 (1999), 1665-1671.
15. DiFilippo, E, Natarajan, v., Bradley, M., Palmer, E and Pritchard, D. E., Physica Scripta T59
(1995),144-154.
16. Carlberg, c., Hyp. Interact. 114 (1998),177-195.
17. Van Dyck, Jr., R. S., Moore, E L., Farnham, D. L. and Schwinberg, P. B., Phys. Rev. A 40
(1989),6308-6313.
18. Audi, G. and Wapstra, A. H., Nuc. Phys. A 565 (1993),1-65; 595 (1995), 409-522.
Hyperfine Interactions 132: 177-187,2001. 177
© 2001 Kluwer Academic Publishers.
Abstract. We report new values for the atomic masses of the alkali 133Cs, 87Rb, 85Rb, and 23Na
with uncertainties ( 0.2 ppb. These results, obtained using Penning trap single ion mass spec-
trometry, are typically two orders of magnitude more accurate than previously measured values.
Combined with values of h/matom from atom interferometry measurements and accurate wavelength
measurements for different atoms, these values will lead to new ppb-level determinations of the molar
Planck constant N A h and the fine structure constant ex. This route to ex is based on simple physics.
It can potentially achieve the several ppb level of accuracy needed to test the QED determination
of ex extracted from measurements of the electron g factor. We also demonstrate an electronic cool-
ing technique that cools our detector and ion below the 4 K ambient temperature. This technique
improves by about a factor of three our ability to measure the ion's axial motion.
Key words: single ion mass spectrometry, ppb, cesium, rubidium, sodium, electronic cooling, feed-
back.
If QED correctly predicts the g-2 anomaly of the electron, the fine structure con-
stant, ex, is known to 3.8 ppb [1, 2].t In order to test QED at this accuracy, and
to remove theoretical uncertainties highlighted by the recent readjustment of the
theoretical value of g-2 [3], other measurements of ex at this level of accuracy are
needed. Unfortunately, the next most precise measurements are far from this accu-
racy: ex has been measured to 56 ppb using the AC Josephson effect [4, 5], to 37 ppb
using a beam of neutrons [6], and to 24 ppb from the Quantum Hall effect [7] (see
Figure 1). There is obviously a great need for a QED-independent determination
of ex at the ppb level: not only could it probe QED an order of magnitude more
accurately, but it could possibly suggest the source of the discrepancies in these
other measurements.
* Corresponding author.
** Currently at NIST, Gaithersburg, MD 20899-8424, USA.
t Groups at the Universities of Washington and Harvard plan to reduce this error to about 1 ppb.
178 S. RAINVILLE ET AL.
a Josephso effect
I
Q tanturn Hat effect
t-+-jI
ij
!
~~ g-2 of ~leclron + ~ED
I
1
I "his wot
I
Figure 1. Current measurements of the fine structure constant ex with uncertainty below 100
ppb are plotted with respect to the 1998 CODATA recommended value [12]. The measurement
of the g-2 factor of the electron combined with QED calculations yields a value of ex with an
uncertainty of (3.8 ppb) [1,2]. The relative uncertainties of the other measurements are: 24 ppb
from the Quantum Hall effect [7], 37 ppb from neutron beam interferometry [6], and 56 ppb
from the AC Josephson effect and the magnetic moment of the proton [4, 5]. The method
discussed here (this work) currently yields a preliminary value at 29 ppb.
A route to ex that appears likely to yield a value at the ppb level is opened by the
relationship (in SI units)
2 2Roo 103 mp
ex = -----(NAh), (1)
c Mp me
where mx and Mx represent the mass of particle x in SI and atomic units respec-
tively, and NAh is known as the molar Planck constant. The Rydberg constant, Roo
has been measured to an accuracy of about than 0.008 ppb [8], the ratio of the mass
of the proton to the electron mp/ me is known to 2 ppb [9], and we have determined
the mass of the proton Mp (in atomic units) to 0.5 ppb [10] (Van Dyck et ai. have
recently reported a value of Mp accurate to 0.14 ppb which agrees with ours [11]).
The speed of light c is a defined constant. Thus a measurement of NAh, the "molar
Planck constant", at the ppb level can determine ex to about 1 ppb.
The molar Planck constant can be precisely determined from the quantum rela-
tionship between de Broglie wavelength and momentum:
h h 103 N A h
A=- AV= - = - - - (2)
mv m M
Note that to determine NAh we need to measure the particle's atomic mass M as
welI as the wavelength of the light A and the particle's velocity v. This simple
PRECISE MEASUREMENTS OF CS, RB AND NA 179
2. Experimental setup
Our experimental apparatus and procedure for measuring ion mass ratios have been
described earlier in the literature [17]. We will briefly outline here the general
features of this experimental setup.
We trap a single ion in an orthogonally compensated Penning trap (of charac-
teristic size d = 0.549 cm) [18, 19] placed in a highly uniform magnetic field
of 8.5 T. The harmonic axial oscillation of the ion in the trap is detected with a
superconducting resonant circuit (niobium coil) having a stable Q of about 50000.
The heart of our recently improved detector is a DC SQUID which has a technical
noise floor 10 times smaller than the RF SQUID it replaced. The trapping voltage is
adjusted to match the ion's axial frequency wz , with the resonance frequency ofthe
detection circuit ('" 160 kHz). Typical frequencies for the other two normal modes
of motion of an Ar+ ion in our trap are 3 MHz for the trapped cyclotron frequency
w~, and 4 kHz for the magnetron motion Wm' The cyclotron and magnetron modes
are observed and cooled indirectly by coupling them to the axial mode with a tilted
180 S. RAINVILLE ET AL.
quadrupole RF field applied on the split guard rings of our trap. With the proper
amplitude-time product, such a coupling pulse ("rr-pulse") can phase-coherently
transfer all the energy of the cyclotron motion into the axial mode, in a manner
qualitatively and formally equivalent to the Rabi two-state problem [20].
The absence of direct damping of the radial modes avoids the possibility of
frequency shifts due to such coupling and means that the cyclotron resonance has
practically zero linewidth. In addition, this procedure allows us to work with nearly
ideal fields (pure quadrupole electric field and uniform magnetic field) which helps
to keep the calculated systematic errors below a few parts in 1011.
The trapped cyclotron frequency w~ is obtained by directly exciting the ion's
cyclotron mode, allowing it to evolve "in the dark" for a delay time T, and applying
a "rr -pulse". The axial ring down is then observed to extract the accumulated phase.
Since phases are only defined between 0 and 2rr, a series of measurements is made
with different delay times T to determine the proper phase unwrapping. With a
phase error of about ± 12 degrees, a measurement of one minute gives a precision
of'" 1-2 x 10- 10 . From this, we can extract the free space cyclotron frequency
We = q B / m (where q is the ion charge, B is the magnetic field strength, and m
is the mass) using the invariance relationship (valid even if the trap is tilted with
respect to the B field or not perfectly cylindrical):
We don't directly measure the magnetron frequency Wrn when taking data, but we
use the relation
W; (
Wrn ;:::;; - ' 1 + -9.sm(ern) ) , (4)
2we 4
where ern (typically, < 0.003 rad) is the measured angle between the magnetic
field and the trap axis obtained by actually measuring the magnetron frequency
once. The precision of a single ion cyclotron frequency measurement is limited by
the'" 2.5 x 10- 10 short term fluctuations of the magnetic field.
Finally, a mass ratio is determined by alternately measuring the cyclotron fre-
quencies of two ions during one night (when the magnetic field fluctuations are
reduced). A typical set of data is shown in Figure 2. The final precision on the mass
ratio (typically, ~ 1 x 10- 1°) is currently limited mainly by the drift (few ppb per
hour) of our magnetic field during the time it takes to make and isolate a new single
ion. In this setup, the ions are produced in the trap by ionizing neutral gas with an
electron beam. Obviously, this method is not selective and unwanted ions are also
produced and trapped in the process. A crucial point for us is to be able to isolate
the ion of interest as quickly as possible. The case of Cs3+ was particularly difficult
since the making of a single Cs3+ would produce about 100 and 10 of the singly
and doubly charged states respectively. Furthermore, the hydrocarbons csHt and
C3H~ used as comparison species break apart into many unwanted fragments under
the electron-beam bombardment. The overall effect of these factors was to increase
PRECISE MEASUREMENTS OF CS, RB AND NA 181
0.290 0.570
N 0.285 0.565
::r: oc
'"
:;.:J
r- 0-
<:')
0.280 0.560 :1:
N
~
0 0""
<:')
0.275 0.555
360· VJ
.....u 0
0 00
0 0
+.... 0.270 0.550 VI
0
0:: 00
u'" 0.265
0'\
0.545 ~
0.260
1 169
0.540
02:00 03:00 04:00 05:00 06:00
Time of Measurement (on 11126/98)
Figure 2. Typical night of data. The solid line is a second order polynomial fit to the field drift.
The 360 0 bar shows the magnitude in Hz of a 360 0 error in phase unwrapping. Also shown is
a bar representing the magnitude of a ppb change in frequency.
the time to make and isolate a single ion to 20-30 minutes from the 5-10 minutes
required for previous measurements.
•
•
•
I»CS .... /
• / C,H:
'''Cs~'/
o / co,'
Figure 3. Example of Variation of Mass Ratio from Night to Night. A measurement of the
neutral mass of 133Cs is extracted from each night's run of cyclotron frequency ratio measure-
ments and plotted in ppb relative to our final published value of the neutral mass of 13 3Cs.
The open and closed circles are from frequency ratio measurements of Cs+++ /cot and
Cs++ /CsHt, respectively. The error bars on each night's measurement are extracted from the
low order polynomial fit to both ion's cyclotron frequencies and reflects the distribution of the
cyclotron frequency measurements during that night. The shaded region represents the one
sigma confidence interval arrived at in the final analysis.
Species MIT mass (u) ppb 1995 mass (u) [22] ppb delta
0-1995
this excess scatter, the uncertainties in the weighted average of the ion mass ratios
involving Cs++ fCsHt and Cs+++ fcoi were increased by factors of 2.6 and 2.2,
respectively, so that X~ = 1. Since the Rb measurements all had similar mjq, we
assumed that the night-to-night fluctuations involving the Rb ratios were drawn
from a common statistical distribution. Therefore, we increased the uncertainties
for the Rb ion ratios by a factor of 2.2 so that the overall Rb X~ was reduced to l.
For Na, X~ ~ 0.8 so the uncertainties were not adjusted.
After correcting for molecular binding and electron ionization energies [25], we
obtained a set of neutral mass difference equations. We added to this the set of mass
difference equations used to determine the atomic masses in [10]. A global least
square fit to this overdetermined set of linear equations gave the neutral masses
of the alkali metals (see Table II) with uncertainties aod as well as the previously
published neutral masses with X~ = 0.83. The previously published masses were
essentially unchanged [10]. Uncertainties in M[160] and M[H] (the only atoms
other than 12C in the ratios of Table I) contributed less than 0.1 ppb uncertainty to
the alkali masses.
The use of two distinct reference ions gave a check on systematics by pro-
viding two independent values for each neutral mass. For Rb and Cs X~ is less
than 1. However, because of the larger uncertainty on M[Cs] from Cs++ jC 5 Ht
we quote a final uncertainty of 0.20 ppb (cf. aod(Cs) = 0.16 ppb). For 87, 85 Rb we
quote a od(87, 85 Rb) as the final uncertainties. For the neutral masses from Na++ jC+
and Na+ jCi, the statistical uncertainties are 0.09 and 0.07 ppb, respectively. The
0.2 ppb disagreement of the two values may be evidence for a systematic at the
0.1 ppb level. To reflect this we assigned Me 3Na] a 0.12 ppb uncertainty (cf.
aod(Na) = 0.06 ppb) which spans both independent measurements.
Table II quotes the final values for M[133CS], M[ 87 Rb], M[ 85 Rb] and Me 3Na]
obtained from the global least square fit to the overdetermined set of mass dif-
ference equations with uncertainties from the above discussion. Also included in
Table II are the alkali masses from the 1995 mass evaluation [22]. Our values
differ from the 1995 values by typically 1.5a1995, which suggests that the uncer-
tainties on the masses from the 1995 evaluation were slightly underestimated.
Our value for M[133Cs] lies within the uncertainty of the recent measurement
184 S. RAINVILLE ET AL.
4. Electronic cooling
We have recently developed techniques to cool the detector and ion below the 4 K
ambient temperature of the coupling coil and trap environment. This is done with
electronic cooling [29]. Not only does this technique greatly improve our signal-
to-noise, but it also reduces the amplitude of the thermal motion of the ion, an
important source of error in our measurements. These significant improvements
will be crucial in our efforts to improve our accuracy beyond 10- 10 . In particular,
we plan to make simultaneous measurements of the cyclotron frequencies of two
different ions to alleviate the scatter due to magnetic field fluctuations between the
measurements which are now the limit on our precision. Our first approach to do
this will be to put two ions in the same trap at the same time.
The essence of electronic cooling is to measure the thermal noise, phase shift the
signal and then apply it to the coupling coil in such a way that it cancels the thermal
excitation in the coil. The key is that our DC SQUID has technical noise much
lower than 4 K and can measure the current in the coil well in a time shorter than
the thermalization time of the coil (Qo/w '" 30 ms). This feedback also decreases
the apparent quality factor Q of the coil. Figure 4 shows the thermal noise of the
coil at different gain settings. Analyzing these data, we find that the thermal energy
in the coil, corresponding to the area under the peak, is reduced below 4 K by
the factor Q/ Qo, as expected from the detailed solution of the circuit (assuming a
parallel LRC coupling coil where the resistor R = QowoL has the usual Johnson
noise current).
Another effect of this feedback is to effectively reduce the impedance presented
to the ion by the detector, thereby increasing the damping time of the ion. This
reduces the bandwidth of our signal, increasing our signal-to-noise ratio (the John-
son noise is a constant current/ $z). This translates directly into a better ability
to estimate the parameters of the axial oscillation of the ion. With this technique,
we can now measure the phase of the cyclotron motion with an uncertainty as
low as 5 degrees, which is more than a factor of 2 improvement. Our ability to
determine the amplitude of the ion signal has also improved, again by more than
a factor of 2, and we can measure the frequency of the axial motion with 4 times
better precision. The better phase noise allows us to obtain the same precision on a
cyclotron measurement in a shorter time. This will be very important in the future
since we would have to acquire data for 10 minutes to reach a precision of 10-11
with the previous phase noise ('" 12 degrees), or 100 minutes for 10- 12 ! We can
PRECISE MEASUREMENTS OF CS, RB AND NA 185
0.5 -.,--------~-r-------___,
Negalive feedback
0.1
Figure 4. Thermal profile of the detector coil as a function of the quality factor Q adjusted
with the gain of the feedback. The thermal energy in the coil (area under the peak) is propor-
tional to Q/ Qo, where Qo is the Q of the detector coil without feedback. This shows that the
feedback does indeed reduce the thermal fluctuations in the coil.
also use the improved signal-to-noise to reduce the cyclotron amplitude we use,
which in tum reduces the frequency shifts due to relativity and field imperfections.
Finally, this technique gives us the ability to arbitrarily select the damping time of
the ion by changing the gain of the feedback.
The improved detector opens some exciting possibilities for our experiment.
For example, we can now probe and characterize the electrostatic anharmonicities
of our trapping potential much more effectively. It also opens the door for us to
very high precision at small mass-to-charge ratio, (e.g., 6.7Li, 3He, 3H) where we
used to suffer from excessively short ion damping times.
Before every measurement, we cool the ion's motion by coupling it to our de-
tection circuit until it comes into equilibrium with the detector. (Only the axial
motion is coupled to the detector, but we cool the two radial modes using the
mode coupling field mentionned in Section 2.) This remaining "4K" motion of
the ion adds vectorially to the displacement from our cyclotron drive pulses and
hence prevents us from establishing an exactly reproducible amplitude and phase of
motion with each excitation pulse. This effectively adds random noise to the phase
we measure. Since it is the same Johnson noise that drives the ion's thermal motion
and is added to the ion image current to form our detected signal, these two sources
of noise both contribute to our measurement error (phase noise). Moreover, the
thermal cyclotron amplitude fluctuations cause relativistic mass variations and also
combine with field imperfections to introduce random fluctuations of the cyclotron
frequency which can be of the order of few parts in 10- 11 . After magnetic field
fluctuations, this is the dominant source of noise in our measurements, and will
be the main obstacle to higher precision in our simultaneous measurements of two
186 S. RAINVILLE ET AL.
different ions. With the electronic cooling technique described above, the ion's
motion should now come into equilibrium with the colder detector thereby greatly
reducing these problems, and allowing a precision better that 10- 11 in a reasonable
amount of time.
The first practical result of this improved detector was a new calibration of our
cyclotron amplitude. Knowing the absolute ion cyclotron radius is important to be
able to calculate and correct (if necessary) the cyclotron frequency we measure for
frequency shifts due to relativity and trap imperfections. Also knowing the ion-
ian separation will be crucial when we make measurements on two ions in the
same trap. Before this work, there was a factor of two uncertainty in our amplitude
calibration, despite our having worked hard on three different methods to determine
it. During the month of July 2000, we accurately measured the relativistic cyclotron
frequency shift (a natural law) versus cyclotron radius for Ne++ and Ne+++ ions
using the feedback technique described above to maintain adequate SIN over a
broad range of cyclotron radii. The measurements on each ion separately led to
two independent amplitude calibrations with an uncertainty of 1.6% and 1.8%,
respectively. The agreement between the two calibrations was about 3%. These
two measurements were done at very different cyclotron frequencies so that any
systematic error would affect them differently. Thus we feel confident that we now
know the absolute amplitude of motion of our ions to about 3%, a tremendous
improvement.
Acknowledgements
This work is supported by the National Science Foundation and a NIST Precision
Measurements Grant.
References
1. Kinoshita, T., IEEE Trans. Instrum. Meas. 46 (1997), 108.
2. Van Dyck Jr., R S., Schwinberg, P. B. and Dehmelt, H. G., Phys. Rev. Lett. 59 (1987), 26.
3. Kinoshita, T., Phys. Rev. Lett. 75 (1995), 4728.
4. Williams, E. R et ai., IEEE Trans. Instrum. Meas. 38 (1989), 233.
5. Cohen, E. R. and Taylor, B. N., Rev. Mod. Phys. 59 (1987), 1121.
6. Kruger, E., Nistler, W. and Weirauch, w., Metrologia 35 (1998), 203.
7. Jeffery, A. M. et ai., IEEE Trans. Instrum. Meas. 46 (1997), 264.
8. Udem, T. et ai., Phys. Rev. Lett. 79 (1997), 2646.
9. Farnham, D. L., Vandyck, R S. and Schwinberg, P. B., Phys. Rev. Lett. 75 (1995), 3598.
10. Difilippo, F., Natarajan, v., Boyce, K. R. and Pritchard, D. E., Phys. Rev. Lett. 73 (1994), 1481.
11. Van Dyck Jr., R, Farnham, D., Zafonte, S. and Schwinberg, P., In: Trapped Charged Particles
and Fundamental Physics, AlP 457, Asilomar, CA, 1998, pp. 101-110.
12. Mohr, P. and Taylor, B., Rev. Mod. Phys. 72 (2000),351.
13. Kruger, E., Nistler, W. and Weirauch, w., Metrologia 32 (1995), 117.
14. Weiss, D. S., Young, B. C. and Chu, S., Phys. Rev. Lett. 70 (1993),2706.
15. Udem, T., Reichert, J., Holzwarth, R. and Hansch, T. w., Phys. Rev. Lett. 82 (1999), 3568.
PRECISE MEASUREMENTS OF CS, RB AND NA 187
16, Haffner, H. et al., In: F. Fuso and F. Ccrvelli (eds), Abstracts of 17th International Conference
on Atomic Physics, AlP, Firenze, Italy, 2000, pp. 27-28.
17. Weisskoff, R. M. et aI., 1. Appl. Phys. 63 (1988), 4599.
18. Van Dyck Jr., R, Wineland, D., Ekstrom, P. and Dehmelt, H. G., Appl. Phys. Lett. 28 (1976),
446.
19. Gabrielse, G., Phys. Rev. A 27 (1983), 2277.
20. Cornell, E. A., Weisskoff, R M., Boyce, K. R and Pritchard, D. E., Phys. Rev. A 41 (1990),
312.
21. Bevington, P. and Robinson, D., Data Reduction and Error Analysis for the Physical Sciences,
2nd edn, McGraw-Hill, Boston, 1992.
22. Audi, G. and Wapstra, A. H., Nucl. Phys. A 595 (1995), 409.
23. Huber, P., Robust Statistics, Wiley, New York, 1981.
24. Bradley, M. P. et ai., Phys. Rev. Lett. 83 (1999), 4510.
25. Mallard, W. G. and Linstrom, P. J. (eds), NIST Chemistry WebBook, NIST Standard Reference
Database No. 69, NIST, Gaithersburg, MD, 1999 (http://webbook.nist.gov).
26. Carlberg, c., Fritioff, T. and Bergstrom, I., Phys. Rev. Lett. 83 (1999), 4506.
27. Young, B., Ph.D. thesis, Stanford University, 1997.
28. Udem, T., private communication, 1999.
29. Forward, R, 1. Appl. Phys. 50 (1979), 1.
Hyperfine Interactions 132: 189-194,200l. 189
© 2001 Kluwer Academic Publishers.
Abstract. The aim of the AVOGADRO project is to replace the kilogram artefact by a high-purity,
perfect single crystal of natural or isotope-enriched silicon. The isotopic composition and the impu-
rities of the silicon crystal must, therefore, be known with highest possible accuracy and precision.
The only method to obtain all this information without destruction of the massive samples is prompt
(n,y)-spectrometry. The measurements are performed at a thermal neutron guide of the ILL (Institut
Max von Laue Paul Langevin) in Grenoble, France. The spectrometry of y-radiation emitted by
a nucleus promptly after thermal neutron capture allows a highly precise determination of atomic
mass differences, as well as the determination of isotope abundances leading to the molar mass.
The uncertainties assigned to the results for the respective atomic masses determined by the mass
differences amount to up to 10- 10 , while the molar mass of an isotope-enriched Si single crystal has
so far been determined with an uncertainty of 1 . 10- 4 . A direct comparison (for example, relative
value of isotope abundances determined by (n,y)-spectrometry omitting the thermal neutron cross
section) furnishes a value of 7 . 10-5 . The final aim of the AVOGADRO project is to provide a well
specified crystal, which allows a more accurate value of the Avogadro constant to be determined.
This constant is the key input parameter for tabulated values of fundamental constants and for a new
definition of the unit of mass - the kilogram itself.
1. Introduction
125 years ago, on May 20th, 1875, by the meter convention signed in Paris, the
'Kilogramme des Archives' and the 'Metre des Archives' were acknowledged
by 17 national governments which concluded the first international agreement on
units. To this day, the kilogram is the only base unit in the International System of
Units (SI) (which followed the Meter Convention in 1960), which is still defined
in terms of a material artefact and not in terms of physical constants. Great efforts
are therefore made to erase this last unknown spot on the metrological map. The
approach of the AVOGADRO project [1] for the new realization of the kilogram is
to determine a more accurate Avogadro constant NA with an assigned uncertainty
of the order of 10- 8 . The relation 1 kg = {NA kmol}u would thus serve as a defining
equation, with the atomic unit of mass being u = 1/12· m(12C). The determination
of a more accurate Avogadro constant can be based on a silicon single-crystal of
190 A. PAUL ET AL.
known volume v, mass m, molar mass M(Si) and volume Vo of the unit cell (with
n atoms): NA = (n . M(Si) . v)/(Vo . m).
The spectrometry of y-rays promptly emitted after thermal neutron capture has
been realized within the scope of this project to analyse the silicon single-crystals
involved. This on-line measurement is carried out to determine the characteristics
of nuclei (atomic mass differences, cross-sections, level schemes, etc.), as well
as non-destructive investigations into the isotopic composition (and, therefore, the
element composition) of any samples. The thermal neutrons are captured in the
sample according to their isotope-specific cross sections and cause high internal
excitation of the nuclei concerned (Tmax ~ 1011 K). This leads to an immediate
rearrangement of the nucleons in a new state most favourable with respect to en-
ergy, and the excess energy is dissipated by prompt emission of electromagnetic
radiation in the form of characteristic y -quanta. The process takes place for all
isotopes (in contrast to neutron activation analysis); it is non-destructive and does
not require sample preparation (in contrast to mass spectrometry). Moreover, it is
independent of the sample dimensions and its chemical state.
At the end of a curved neutron guide tube and at great distance from the reactor
core, the sample to be investigated is exposed to the parallel thermal neutron beam
(E ~ 25 meV) of known flux density. The energy of these neutrons is lower
than the energy of the fission neutrons near the reactor core by a factor of up to
109 and, therefore, far from any resonance capture. All isotopes are simultane-
ously detected by multi-channel analysis (in form of a single y-spectrum). This
involves the analysis of the promptly emitted y-rays according to their energy (up
to 12 MeV) and intensity after thermal neutron capture. The y-energy spectrum is
used to identify all nuclides and to determine atomic mass differences in the case of
direct transition to the ground state (e.g., pure 2-body decay), while the intensity of
the promptly emitted y-rays is a measure of the abundances of the isotopes inside
the samples.
The (n, y )-experiment is installed at the high-flux reactor of the ILL in Grenoble.
It is placed at the end of the large neutron guide hall at position S51 of neutron
guide tube H22, providing a distance of about 100 m to the reactor core, thus
reducing its y -background radiation by a factor of 106 . Moreover, S51 is shielded
by heavy concrete, and the beam shutter is placed outside. The shielding of the
large Ge(HP)-detectors is optimized to reduce the y -background. In addition, a
neutron collimation system with a horizontal aperture of (Xh = 0,587(3)° and a
vertical one of (Xv = 0,850(4)° has been developed. A capture shield against diffuse
neutrons has also been installed. The experimental set-up includes two Ge(HP)-
detectors with relative efficiencies of 150% and 115% (compared to a cylindrical
PROMPT (n,y) MASS MEASUREMENTS FOR THE AVOGADRO PROJECT 191
I l000 mm I
~. 2200 - - - 1 --11
3/1 . 3/1 NaI-detector measuring a point source of 60Co at the y-energy of 1332.5
keY at 25 cm distance). yy-coincidence measurements for the determination of
accurate level schemes of all isotopes involved are thus possible. Ultra-high-rate
electronics with sampling ADCs is used in order to achieve count rates above
105 S-l, nearly without pile-up. Both, the energy resolution and the temperature
stability are improved.
Energy and efficiency calibration of the detector is ensured by calibration with
different target materials, for example the 14N(n,y)15N reaction (moreover, the re-
actions lH(n,yfH, 13C(n,y)14C, 27AI(n,y)28AI, 35CI(n,y)36Cl provide transitions
well known as regards y-energy and/or the transition probability) and standard
y-calibration sources. Uncertainties for the y-energy Ey can be reduced by taking
into account the energy intervals of the full-energy, single-escape and double-
escape peaks. These intervals are very precisely known: m e c 2 = 510999.06(15) eV,
and they are, therefore, well suited for highly accurate energy calibration. The
analysis of the prompt (n,y)-spectra for the determination of the isotopic com-
position involves literature data (checked or improved by the (n,y)-experiment)
for the y-energy Ey and the corresponding transition probability Pyas well as
the thermal neutron capture cross section. Since the data published for transitions
in excited nuclei often differ by more than the given uncertainty, or as the stated
uncertainty is too large, check and improvement of published data is a primary task
for the (n, y )-experiment.
192 A. PAUL ET AL.
Table I. Results for the masses of silicon isotopes expressed in the conventional form of the molar
mass M in flU compared with literature data
In keV In /LU
A X (n,y)A+1X
Ey,gs ER (I) EB A M(AX) Refs. [2-4]
28 27976926.532 (22) 27976926.5327 (20)
28Si(n, y) 29 Si 8472.222(5) 1.329 8473.551 (5)
29 28976494.720 (21) 28976494.719 (30)
29 28976494.720 (21) 28976494.719 (30)
29Si(n, y) 3O Si 10607.164(5) 2.013 10 609.178(5)
30 29973770.217 (35) 29973770.218 (45)
30 29973770.217 (35) 29973770.218 (45)
30Si(n, y )31 Si 6586.646(5) 0.751 6587.397(5)
31 30975363.275 (70) 30975363.275 (70)
31 30975363.275 (59) 30975363.275 (70)
31 Si(n, y) 32 Si 920 l. 798( 5) 1.420 9203.218(5)
32 31974148.131(822) 31974148.129 (2324)
m(A+~ X) A ) EB
m ( zX +m n - -2'
C
m _ Ey,gs .
n c2
(1 + ~ . 2
Ey,gs
m(A+i X) . c 2
)
'
by stepping up and down the mass scale, starting with a single nuclide. Table I
shows some of the results. The calculation can be based on either the conversion
(u B keY) or (g B keY).
4. Isotope abundances
The determination of relative isotope abundances involves the identification of the
y-ray transitions by their y-ray energy and the analysis of the probability of each
transition, that is to say, the number of events in a peak. All peaks of a single isotope
are weighted by their y-ray transition probabilities P(~X, Ey) [5] and relative
detection efficiencies 17rel (E y ).
PROMPT (n ,y) MASS MEASUREMENTS FOR THE AVOGADRO PROJECT 193
10'
....
;;;}
c~
....
0
" Si [n,y)fSi
2235. 1+ keV
110'
Z
10'
. <11
....
10'
~A
~
Si2~~
Si O~
S1 ++
2000 2200 2400 2600 2800 keY 3000
Ey
Figure 2. Comparison of (n,y)-spectra of three silicon samples (Si 2: natural silicon; Si OR:
28Si-enriched material from Oak Ridge; and Si++: highly enriched 28Si material from Rus-
sia), in the energy range from 2000 keY to 3000 keY. The vertical scale of the spectra have
been shifted for comparison.
Table II. Isotope intensity T* of the enriched samples Si OR, Si+, and Si++, including the
statistical uncertainties, uncertainties from systematic effects, and the isotope ratios T (11 Si) =
T(11Si)/(T(i~Si) + T(i~Si) + T(f~Si» calculated by (3) on this basis together with the thermal
neutron capture cross sections of Raman et al. [5]
SiOR 0.004459 (81)(30) 0.002570 (65)(21) 0.98971 (37) 0.00627 (27) 0.00402 (25)
Si+ 0.004438 (79)(34) 0.002600 (65) (21) 0.98970 (37) 0.00624 (27) 0.00407 (25)
Si++ 0.000 570 (21) (5) 0.000301 (27) (5) 0.99872 (8) 0.000 81 (5) 0.00048 (6)
This way, a mean (n: number of peaks of a single isotope) isotopic intensity
T* (~X) is established, for each detected isotope:
I n A
T *(A)_
X --.~" N(t,Eyi,ZX) and
Z n i=1 f)(Ey) . P(~X, Ey)
(2)
T(~X) = a(zX,
A I . T*(~X).
En)
194 A. PAUL ET AL.
Typically, for prompt (n, y )-spectrometry, either this isotopic intensity or the iso-
topic intensity divided by the capture cross-section a(1X, En) (En is the energy of
the thermal neutron), that is to say, T(1X), is given as a ratio to the major isotope
of the sample. In this normalization the isotopic intensity divided by the capture
cross section gives the isotope abundances.
The isotope intensities of natural silicon crystals have been determined from
these measurements with an uncertainty of 10-6 . Their accurate isotope abun-
dances can be calculated as soon as accurate thermal neutron capture cross sections
are known. In a first test for the sensitivity of the (n, y )-analysis, isotope-enriched
materials are measured. Three samples of 28Si of unknown enrichment, two from
Russia and one from Oak Ridge, are compared, cf. Figure 2 and Table II. The
material from Oak Ridge and the less enriched material from Russia proved to be
identical within 3 x 10-4 by (n,y)-analysis. Its enrichment is determined to be
98.97(4) % taking all uncertainties into account.
The highly enriched material from Russia is found to have 99.872(8) % of 28Si;
the uncertainty is 8 x 10-5 . Thus, a molar mass of 27.9790(30) g.mol- 1 with a
relative uncertainty of 1 x 10-4 is measured by prompt (n, y )-analysis for the first
time.
5. Conclusions
Within the scope of a research project, the method of isotopic analysis merely
exploiting nuclear physics phenomena is developed into a measuring arrangement
set up at the ILL, for the non-destructive qualitative and quantitative determination
of the composition of any samples. The determination of the isotopic composition
of the silicon spheres for the Avogadro project [1] is an example of its metrological
application.
Acknowledgements
We are indebted to the staff members of PTB department Q.2 and to the staff of the
ILL as well, for their valuable assistance.
References
I. Bettin, H., Koenders, L., Martin, J., Nicolaus, A., Becker, P. and Rottger, S., PTB-Mitteilungen
106 (1996),321-329.
2. DiFilippo, E, Natarajan, v., Bradley, M., Palmer, E and Pritchard, D. E., Physica Scripta T 59
(1995), 144-154.
3. Bradley, M., Palmer, E, Garrison, D., I1ich, L., Rusinkiewicz, S. and Pritchard, D. E., Hyperjine
Interactions 108 (1997), 227-238.
4. Audi, G .. Wapstra, A. H., Nuclear Physics A 595 (1995), 409.
5. Raman, S., Jurney, E. T., Starner, J. W. and Lynn, 1. E., Phys. Rev. C 46 (1992),972.
Hyperfine Interactions 132: 195-207,2001. 195
© 2001 Kluwer Academic Publishers.
Abstract. A new experiment for a high-precision measurement of the pion mass at a 1 ppm level
is presented. It combines an improved cyclotron trap that produces pionic and muonic atoms in
a small volume with a doubly focusing crystal spectrometer to measure the corresponding exotic
X-ray transitions with high accuracy and a novel type of CCO detector. The muonic X-rays lines serve
as highly accurate calibration lines. The measurement has been accomplished recently. A detailed
analysis of the data is on the way.
Key words: charged pion mass, exotic atom, X-ray spectroscopy, Bragg spectrometer.
1. Introduction
The precise knowledge of the mass of charged pions is of outstanding importance
in several fields of modem physics research.
In nuclear physics the simple rr-p and rr-d systems serve as a testing ground
to study the hadronic interaction [1-3]. For instance the strong interaction shift can
provide a sensitive check of the methods of chiral perturbation theory. It is note-
worthy, that the most recent rr-p experiment [3] suffers from the fact that 20% of
its systematic error is due to the actual uncertainty of the pion mass (2.5 ppm) [4].
For a forthcoming precision experiment to measure the complex scattering lengths
[5] the availability of a precise pion mass value is mandatory. For the determination
of the mass of the fL-neutrino MVII the pion mass Mn plays a key role. They both
are connected through the decay of the free rr+ as [6]
rr + ~ fL ++ vJL"
* Corresponding author.
196 G. L. BORCHERT ET AL.
Taking into account the CPT theorem and the conservation laws of energy and
momentum the relation for pion decay at rest reads as
with M/1+ and P/1+ being the mass and the momentum of the muon, respectively.
Keeping the muon mass M/1+ as constant, as it has the smallest uncertainty
(0.05 ppm) [4], a significant improvement for !)'MvI" can only be expected from
a more precise pion mass Mn-.
Combining theoretical considerations with cosmological bounds, the range for
allowed values of MVI" can be drastically reduced: for unstable neutrinos, MVI"
should be larger than 70 keY jc 2 while for stable neutrinos, MVI" should be less
than 65 keY jc 2 [7-9]. An improvement of the accuracy of Mn- by a factor of 5 is
necessary to reach the critical boundary of 70 keY jc 2 for MVI"'
0.05
0.00
-+-
I
-0.15
-0.20
M'lt (MeV/f?)
Figure I. This graph shows the evolution of the upper limit of M VI" as a function of the value
of Mn-' We can see that only the B value of [13], which has an error of 3 ppm, is compatible
with a non-negative value of M~ [6], and that our previous determination of the pion mass
I"
[14] is not accurate enough (4 ppm): a more accurate value of the pion mass is necessary
to reduce the possibilities of interpretation (represented by the hatched area in the figure) to
determinate a precise neutrino mass.
PRECISION MEASUREMENT OF THE CHARGED PION MASS 197
2. General considerations
As the pion is an unstable particle (r = 26 ns) the classic method of mass spec-
troscopy cannot be applied. Yet it still lives long enough to form a bound atomic
system. Its deexcitation radiation of the pionic atoms can be measured with high
precision by the method of X-ray spectroscopy.
The binding energy of a boson in a hydrogen-like atomic system is given to first
order by the Klein-Gordon equation:
During the formation process the pion is captured in a highly excited state of the
host atom. When it cascades down, the electrons of the atom are quickly ejected by
Auger effect. For nitrogen gas, cascade calculations [19] show that when the pion
reaches states with main quantum number n = 5, the probability for finding one
remaining K electron is less than 2% with a gas pressure around 1 bar (see last row
of Table I) - the data of our previous experiments of 1995 and 1997 confirmed that
no satellite transition with a yield >2% could be indentified [14].
198 G. L. BORCHERT ET AL.
Table I. Properties of the 5g --+ 4f transitions in pionic nitrogen and muonic oxygen
]TN /.LO
Transition energies (eV) 4055.381 (5g --+ 40 4023.752 (5g9/2 --+ 4f7/ 2)
Parallel transitions (eV) 4057.693 (5f --+ 4d) 4023.509 (5g7/2 --+ 4f7/ 2)
4061.940 (5d --+ 4p) 4025.806 (5g7/2 --+ 4f5/ 2)
4053.076 (5d --+ 40
99.6% 99.8%
Abundance
0.4% 0.2%
Finite size effect (!-l-e V) 3 3
The situation is different when the pion is captured in solid materials. During
the deexcitation a strong refilling of the emptied electron shells occurs from neigh-
boring atoms and satellite transitions due to the presence of K electron become
significant. In this case they cannot be resolved experimentally, the interpretation
must be based on cascade calculation arguments. This situation was the reason for
the problems with the recent M rr - results [13, 15], as they were performed with
solid magnesium.
Hence for a precision measurement it is mandatory to use gaseous atomic sys-
tems. Slight difficulties arise if the gas is composed of molecules. In this case,
when the pion cascades down after capture, the ejection of the binding electrons
will cause the breakup of the molecule. As the fragments are accelerated by the
Coulomb force, the pionic X-ray transitions exhibit a (symmetrical) Doppler broad-
ening, which yields information about the charge states at the moment of the
molecule separation [20].
An essential contribution to the final uncertainty of the pion mass is introduced
by the absolute energy calibration of the X-ray transitions. Generally fluorescence
X-rays are used for this purpose, but they show a severe disadvantage: their large
naturallinewidths [211 together with a complex satellite structure [22], which actu-
ally depends on the mode of excitation. Even though some standard X-ray lines are
PRECISION MEASUREMENT OF THE CHARGED PION MASS 199
known with ppm accuracy, their use in spectroscopy gives only an accuracy to 4 or
5 ppm. In contrast, exotic atom X-rays are extremely narrow and can be calculated
with high precision by the methods of QED for hydrogen-like systems.
It is obvious that the energy and the line shape of the calibration line should be
as similar to the pionic X-ray transition as possible to avoid systematic errors. In
this project the 5g --+ 4f transition in pionic nitrogen is compared to muonic oxy-
gen. Interestingly enough, as pions decay to muons, both show up simultaneously,
so that by suitable experimental tuning their intensity ratio can be optimized. In
Table I some essential parameters are shown. The energies of the most intense
transitions are given [17], and the Bragg angles for a Si (220) reflection. The
difference between pionic and muonic line is 32 e V or 36'. The energies of the
parallel transitions allow a clean separation in a high resolution spectrometer. In
the following row the natural line widths of pionic and muonic transitions are given
- these widths are more than two orders of magnitude smaller than the response
function of the crystal spectrometer (~300 meV, Monte Carlo simulation [23],
using the rocking curve values from [24]).
In the next 5 rows some effects due to the interaction with the nucleus are listed.
While the isotope lines are well separated [25], the following effects are obviously
negligible [25, 26].
3. Experiment
The aimed-for precision of 1 ppm for Mn- is only achievable with a crystal dif-
fraction device. In the envisaged energy region, a Bragg reflection spectrometer
has to be used. To meet the condition of high resolution and high efficiency, spher-
ically shaped large Si (2 20) crystals (10 cm diameter) have been developed*. Their
resolution proved to be close to the theoretical expectation. From geometrical con-
siderations it is obvious that the X-ray image is curved. To correct for this effect,
the detector has to be a 2-dimensional position sensitive device. Furthermore, it
has to be large enough to cover the spectral range between the pionic line and the
muonic calibration line completely. In this way the spectrometer motions, which
may introduce systematic uncertainties during the measurement, are avoided. To
fulfil these conditions a CCD detector array consisting of 2 x 3 adjacent chips,
each 24 x 24 mm 2 , has been developed. The pixel size of 40 !-Lm is well adapted
to yield the necessary spatial resolution at the focal length resulting from a crystal
radius of curvature of 3 m.
A typical feature of high-resolution crystal diffraction is the small solid angle.
Thus the total efficiency of the spectrometer incorporated in this experiment is
around 10- 7 . With typical accelerator currents of about 1.6 rnA count rates of
up to 500 per hour are achievable for pionic X-rays, but only about 15 per hour
for muonic ones. Therefore an efficient suppression of the background is rather
* The technique has been developed in collaboration with the optical company Carl Zeiss,
Oberkochen.
200 G. L. BORCHERT ET AL.
essential. In this respect the CCD detector offers two excellent features: firstly, the
energy resolution is about 160 eV at 4 ke V, which permits to set narrow energy cuts.
Secondly, the discrete pixel structure allows for a geometrical characterization of
the recorded events: 4 keY X-rays deposit their energy in one single pixel for about
70% of the hits and in two adjacent pixels for about 30%. Larger clusters are due
to particle or high-energy gamma radiation and can thus be discarded. In this way
the background can be reduced by almost three orders of magnitude.
Figure 2. Technical sketch of the experimental setup. The apparatus is composed of three
main components: the cyclotron trap. the crystal spectrometer, and the CCO detector. The
X-ray beam, which passes through vacuum tubes, is reflected by a Bragg crystal and de-
tected by a cooled CCO detector. A lever arm moved by a linear drive and stabilized by
a piezo-electric unit rotates the crystal. The rotation of the crystal relative to the detector
is measured by a high-precision angular encoder, which is rigidly connected to the crystal
rotation axis. The detector arm and the CCO position are adjusted by linear drives.
PRECISION MEASUREMENT OF THE CHARGED PION MASS 201
The new cyclotron trap is the ideal device to create such a intense source of
pionic and muonic X-rays [27]. It consists of a superconducting split coil magnet
with weakly focusing characteristics [27]. The target gas cell (54 mm diameter,
230 mm length) is placed in its center. The pions injected from the rr E5 beam
line of the PSI accelerator facility are forced into orbit around the target by the
magnetic field. They lose their kinetic energy going through several degraders. In
the central region the pions are slow enough to be stopped in the gas target and to
be captured in bound atomic states. As the decay muons are present as well, the
simultaneous measurement of rr Nand fJ.,O becomes possible, when the gas cell is
filled with a mixture of nitrogen and oxygen. By an optimization of the degraders in
the trap and of the gas mixture and pressure it was possible to measure the 5g --+ 4f
transitions in rr Nand fJ.,O simultaneously with comparable intensities. The center
of the trap can also be equipped with various solid state targets that are excited to
emit X-rays by means of an external X-ray tube shining through a special aperture.
In this way stability and calibration can easily be cross-checked by measuring these
high intensity X-ray transitions (mainly Cu Ka, in second order).
The experimental setup is shown in Figure 2.
4. Measurement
Based on the experience from our previous experiment that already yielded a pre-
cision of 3.7 ppm for the pion mass [14], the following schedule for the actual
measurement was adopted: During the first 2 weeks, the beam injection, the cy-
clotron trap, and the gas targets were optimized. Consequently the fine tuning of
the crystal spectrometer was performed with Cu Ka, fluorescence radiation. The
following 6 weeks were used for data taking. With the accelerator current of 1.5-
1.7 mA and a target gas mixture of 90% oxygen and 10% nitrogen at 1.4 bar a
count rate of 15-18 events per hour in each of the prominent lines was obtained.
At the end about 9000 events in the rr N line and about 10000 events in the fJ.,O line
were accumulated. A typical spectrum is shown in Figure 3.
The remaining 2 weeks were dedicated to 2 series of special calibration mea-
surements:
l. The relation between the distance of 2 lines in the detector plane and the cor-
responding interval of the Bragg angle has to be known with high accuracy.
This measurement was performed by rotating the crystal in small steps, and
determining the corresponding displacement of the line on the CCD detector.
For this purpose, both the Cu Ka, and rrN line were used. In this case the
trap has been optimized for pion production, yielding up to 300 pionic events
per hour. Furthermore, the rr N measurements have the same statistics as the
measurement with the gas mixture: we can check the result for the pion mass
by this independent measurement.
2. For a careful analysis of the data it is essential to determine the intrinsic re-
sponse function of the spectrometer. As nitrogen and oxygen are diatomic
202 G. L. BORCHERT ET AL.
' 100
800 4060 4050 4040 4030 4020 4010 .. E[eV I
100
600
500
~ ,00
.~ 300
E'" 200
100
j ~ .. I \/\
tr.6 ~8 ~r.O 5J.2 -;J:4 ----s~ ..."" (- )
0 50 100 150 200 250 300 ..xIchns I
Detec tor position
Figure 3. X-position spectrum (the solid line represents the fit of the data); the cali-
If N-fLO
bration lines - fLO (5g -+ 4f) - are on the left of the spectrum. On the top, the residuals are
plotted.
molecules, their X-ray transitions exhibit the Doppler broadening due to Cou-
lomb explosion [20] and are not suitable for this purpose. Therefore the 6 -+ 5
transitions in nNe at 4.5 keY have been measured.
s. Data analysis
The CCD images are acquired at the rate of one frame every minute. The noise
peak is removed from each frame before the image is stored. Typically 60 frames
where stored in a single file, containing up to 300 useful events. A complex data
processing is required to extract the meaningful information from a very large
number of pixel hits. First, we eliminate the CCO "bad pixels", i.e., pixels that
always give a signal due to CCO chips defects or synchronization signals.
Afterwards we must also estimate the displacement between the CCD chips:
the evaluation of the horizontal shift between both columns of three vertical CCD
chips represents the main correction on the final value of the pion mass, which is
extracted from the relative difference between the positions of the pionic line and
the muonic line, while the knowledge of the vertical shifts is important to determine
the parameters of the curvature correction. We have two main methods to evaluate
these displacements:
1. We use cosmic rays which deposit an intense, narrow and linear track on the
CCO chips, since they have a high energy. If one cosmic ray hits two adjacent
PRECISION MEASUREMENT OF THE CHARGED PION MASS 203
Scatter plot
o
o
.n
OL-:--OR
.I · ~~I -
.,
4i 0
.~
.2;~
0
I I
c
~
'iii
o
0..
ML-~MR
I
>-
o
o
.n
I
· UR
U-L ---
500 1000
X-Position (pixels)
Figure 4. Scatter plot obtained with the mask in front of the CCD detector and an 55Fe source.
It allows us to deduce the shifts between the CCD chips.
CCO chips, we align thus both parts of this track during the data analysis, to
find the relative position of these CCO chips.
2. We put a mask with several narrow slits and apertures in front of the CCO
detector and use a radioactive source to project the image of this mask on the
CCO chips (see Figure 4). We deduce the shifts between the CCO chips by
changing their positions to reproduce the dimensions and the proportions of
these openings.
The combination of both independent methods gives a very accurate estimation
of relative shifts between the CCO chips.
Because of the large pion induced background, the spectrometer is completely
enclosed in a concrete house (see Figure 5). The background events recorded by
the CCO are removed by using the large discrepancy with the topology of energy
deposition created by low-energy X-rays. These X-ray events deposit their energy
in one or two adjacent pixels (see Figure 6), while background events deposit their
energy in large clusters of pixels.
204 G. L. BORCHERT ET AL.
'IT'E5
Figure 5. Setup of the experiment in the :rr E5 area at PSI, with the concrete shielding used to
suppress the high neutron induced background.
Energy spectrum
....
'" 0
cO
::J
o
u
..,o
3000
Energy (Pixels)
Figure 6. :rr N-JhO energy spectrum after the application of cluster analysis; we can separate
the pionic and muonic lines - the peak between the two vertical dashed lines - energetically
from the parasitic lines that come from background-induced scattering of X-rays from mate-
rials inside the CCD cryostat. (The small picture shows the energy spectrum without cluster
analysis.)
PRECISION MEASUREMENT OF THE CHARGED PION MASS 205
Scatter plot
o
o
Lfl
c
·2
'iii
a
0..
I
>-
500 1000
X-Position (pixels)
Figure 7. nN-j.LO scatter plot. We apply energy cuts to keep only the clusters in the correct
energy range and reduce the background. (The small picture shows the scatter plot without
energy cuts.)
A cluster analysis [14] eliminates these parasitic events and keeps only single
and double pixel X-ray events. Further this elimination of the parasitic events is
achieved by taking advantage of the good energy resolution of the CCD and doing
energy cuts around the energy where the X-rays of interest are expected (see Fig-
ure 7). We typically obtain a signal to noise ratio in excess of 300. This method thus
yields very low background scatter plots (X/Y-position plots), while the original
ones look almost uniformly illuminated.
The next step consists of a projection of the 2-dimensional images provided
by the CCD onto the axis corresponding to the dispersion plane of the spectrom-
206 G. L. BORCHERT ET AL.
eter. However, we must correct for the curvature of the X-ray lines, due to the
shape of the crystal and the geometrical constraints imposed by the Bragg law for
X-rays reflected outside of the dispersion plane. The parameters of this curvature
are obtained by the fit of a measured line, and a analysis allow to compensate,
numerically, for the conical shape of this line: We analyze the CDD chips by slices,
and we change the relative position of these slices to 'redress' the studied line.
Finally, we must take into account the different sources of systematic errors
to analyze the X-position spectra correctly. The main systematic error is due to the
positioning of the CCD detector. The geometrical Bragg reflection imposes that the
CCD detector is exactly on the Rowland circle, because the focus of the reflected
Bragg X-rays is located on this circle. But, since we have a large detector and
we do a simultaneous measurement of both exotic X-ray lines, it is impossible to
match this strict geometrical condition for each line. We have chosen to optimize
the positioning of the CCD detector to have the focus in the intermediate position
between both exotic X-ray lines. Thus, the defocusing of exotic lines leads to a
small asymmetrical broadening of their shape that can be estimated by a Monte
Carlo simulation. Another source of systematic error is due to the uncertainty of the
angular encoder, used to measure the rotation of the crystal, that is less than 0.2".*
This correction is only needed for the non-simultaneous measurements, i.e., the
measurement with pure nitrogen gas where we use the Cu Ka X-ray as calibration
line.
6. Conclusion
The pion and muon stop rate, the peak-to-background ratio, and the stability and
sensibility of the detector system are sufficient to reach an accuracy of 1 ppm for the
pion mass determination. The utilization of the gas target eliminate all ambiguous
interpretations of the electronic screening during the pionic and muonic transitions.
As the measurement has been accomplished only a couple of weeks ago, the
analysis of the data has just started. At the envisaged level of precision it is a
rather complicated and time-consuming process including numerous checks and
corrections of higher-order effects. We have just estimated the statistical error from
the X-position spectra. We can expect to obtain a accuracy of about three times
better than the accuracy of our previous pion mass determination [14].
References
1. Sigg, D. et at., Nuclear Phys. A 609 (1996), 269.
2. Chatellard, D. et at., Nuclear Phys. A 625 (1997), 855;
Hauser, P. et at., Phys. Rev. C 58 (1998), 1869.
3. Schroder, H.-Ch., Phys. Lett. B 469 (1999), 25.
4. Groom, D. et at., Eur. Phys. 1. C 15 (2000), 1.
5. Anagnostopou1os, D. etat., PSlproposa1 R-98-01.1, 1998.
Abstract. The mass of the electron in atomic units (me) represents the largest error contribution
in an experiment to determine the g-factor of the electron bound in hydrogen-like carbon. Recent
progress in the calculation reduces the uncertainty of the theoretical value to such a low value that
me can be determined from a comparison of experimental and theoretical g-factors. The present
preliminary value of the electron mass agrees with the accepted value but reduces the uncertainty by
about a factor 2.
Key words: electron g-factor, electron mass, highly charged ions, Penning trap.
To date the most precise value for the electron mass in atomic units is provided by
a comparison of the cyclotron frequencies of electrons and carbon ions in the same
magnetic field. These experiments, carried out by Van Dyck and coworkers [1], are
performed by storing single particles or small clouds of each kind in a Penning ion
trap. A carbon ion is used as a reference since its mass can be easily converted to
the mass of the carbon atom which represents the atomic definition of the mass
unit, M (12C) = 12. The present value from Van Dyck [1], as included in the 1999
CODATA references of fundamental constants, is quoted as
Attempts to reduce the uncertainty of this value are required in view of ongoing ex-
periments to determine the finestructure constant a with improved precision from
photon recoil on cesium atoms and a number of other measurable quantities [2]:
2 2Roo 1 Mp
a = ------(NAh).
c Mp m
(2)
The Rydberg constant Roo has been measured to an accuracy of 0.008 ppb [3], the
proton atomic mass to 0.14 ppb [4], and NAh can possibly be determined by atomic
interferometry to the sub-ppb level [2]. Then the present value for the electron mass
would represent the term with the largest error bar and any improvement would
reduce the overall uncertainty for a.
210 G. WERTH ET AL.
ms = ±1/2 is the spin orientation quantum number and JLB the Bohr magneton.
The g-factor can be derived from the ratio of VL and Vc = (q / M)B of the ion of
charge state q:
VL me q
g=2---. (4)
Vc M e
Here the first number in brackets is the statistical and the second one the systemat-
ical uncertainty, while the third number represents the error of the electron/carbon
mass ratio [1].
The g-factor for an electron bound to a nucleus can be calculated accurately.
While for a free electron the Dirac equation predicts the value g = 2, relativistic
corrections due to the binding to a nucleus of charge Z lead to small deviations.
Breit [6] has solved the Dirac equation for the Sl/2 ground state of hydrogenic
systems and obtained
(6)
In addition radiative corrections have to be taken into account. For the free elec-
tron the different orders of Feynman diagrams, representing an increasing number
of virtual exchange photons, are calculated as a series expansion in a/n. Ki-
noshita and others have performed such calculations up to the order (a/n)4 [7]
and obtained a result in excellent agreement with the experimental value [8]
(8)
The first number in brackets is the numerical uncertainty of the one-loop correc-
tions while the second one is an estimate on the contribution of the two-loop part.
This value is in good agreement with the experimental one.
While in general a comparison of the experimental and theoretical value is con-
sidered as a test of the theoretical calculations, we can here change our point of
view: When we consider the theoretical value as granted and take the value of g
from Equation (5) we can use it to determine a value for the mass ratio Mime from
Equation (4). We then obtain:
This value agrees within the limits of the error with the value obtained by Van Dyck
(Equation (1» with a slightly larger error margin. This error is almost completely
determined by the uncertainty of the theoretical calculation, predominantly the
estimate of the higher order contributions.
It is evident that even a modest improvement of the theoretical calculation will
lead to a new and more precise value for the electron mass. Such calculations,
although extremely difficult and tedious, are presently under way [12]. Two authors
have recently presented new theoretical values for the g-factor of 12C5+: Beier et al.
[13] and Karshenboim [14] have included the known a 2(Za)2 term [15] into the
calculation of [11] and obtained [13]:
and [14]:
The three uncertainties in Equation (10) refer to the nuclear recoil part, the numer-
ical uncertainty of the order (aln), and the estimated error of the order (aln)2,
respectively.
Adding the errors in Equation (10) quadratically and taking conservately the
larger uncertainty of the two theoretical numbers we calculate a new value for the
electron mass
It agrees with the direct measured value of Van Dyck [1] but the uncertainty is
reduced by a factor of about 2. We consider this value as preliminary in the sense
that a complete calculation for the two-loop correction, preferably independent by
different authors, is required before claiming a new value for the electron mass.
As Karshenboim [14] pointed out, it may even not be necessary to perform a
complete calculation of the higher order part, as long as the error in the (Za)-
expansion compared to the nonperturbative calculation is small: If its general struc-
ture is known, containing a certain number of higher order terms, one may exper-
imentally determine the coefficients of these terms by measurements of g-factors
of hydrogenic ions with different Z and fits of the coefficients to the experimental
results. Such experiments are under way in our laboratory. A first result on 16 07+
can be expected soon with similar accuracy as in the case of 12C5+.
Of course, it remains desirable to improve the directly determined value of the
electron mass by more refined measurements of the electron and carbon cyclotron
frequencies [16]. Of particular interest in this respect is the method of cooling the
electron motion into the lowest quantum state of the cyclotron harmonic oscillator
as recently successfully demonstrated by Gabrielse et al. [17]. This would reduce
substantially systematic uncertainties arising from relativistic mass shifts. The al-
ternative way, as outlined in this contribution, includes quantumelectrodynamic
calculations. There are good prospects, however, that these calculations - although
tedious and very difficult - can be performed sufficiently accurate in the near future.
A value for the electron mass from our g-factor measurements can at least serve as
confirmation of the directly measured mass which has not yet been performed at a
similar level of precision and which is desirable for any fundamental quantity.
Acknowledgements
We acknowledge stimulating discussions with Thomas Beier, Savely Karshenboim,
Viktor Shabaev and Gerhard Soff. Our experiments are part of the TMR-network
'Eurotraps' of the European Union.
References
1. Farnham, D. L., Van Dyck, R. S. and Schwinberg, P. B., Phys. Rev. Lett. 75 (1995),3598.
2. Weiss, D. S., Young, B. C. and Chu, S., Phys. Rev. Lett. 70 (1993), 2706.
3. Udem, Th. et al., Phys. Rev. Lett. 82 (1999),3568.
4. Van Dyck, R. S. et al., In: D. Dubin and D. Schneider (eds), Trapped Charged Particles and
Fundamental Physics, AlP Conf. Proc. 457 (1998), p. 101.
5. Haffner, H. et al., Phys. Rev. Lett. 85 (2000),5380.
6. Breit, G., Nature 122 (128), 649.
7. Hughes, V. W. and Kinoshita, T., Rev. Mod. Ph)"s. 71 (1999), 133.
8. Van Dyck, Jr., R. S., Schwinberg, P. B. and Dehmelt, H. G., Phys. Rev. Lett. 59 (1987), 26.
9. Blundell, S. A., Cheng, K. T. and Sapirstein, J., Ph),s. ReI'. A 55 (1997), 1867.
10. Persson. H. et al., Ph)":;. Rev. A 56 (1997). R2499.
II. Beier, Th. et al., Phys. Re\,. A 62 (2000), 032510.
A POSSIBLE NEW VALUE FOR THE ELECTRON MASS 213
Abstract. Penning trap mass spectrometry has reached a state that allows its application to very
short-lived nuclides available from various sources of radioactive beams. Mass values with out-
standing accuracy are achieved even far from stability. This paper illustrates the state of the art
by summarizing the status of the ISOLTRAP experiment at ISOLDE/CERN. Furthermore, results of
mass measurements on unstable rare earth isotopes will be given.
Key words: atomic mass, nuclear binding energy, Penning trap mass spectrometer.
1. Introduction
Penning traps have since long proven to be very accurate mass spectrometers.
A large variety of mass measurements with an accuracy down to 10- 10 have been
performed on stable, mostly light particles [1-5]. With ISOLTRAP [6] at ISOLDE/
CERN it has been demonstrated that Penning traps can be applied with advantage
also to short-lived radioactive ions [7-17]. The success of this experiment has trig-
gered a number of similar ion trap mass measurement projects [18-20] installed or
to be installed at radioactive beam facilities.
The intention of this paper is to briefly summarize the status of ISOLTRAP.
Some emphasize will be put on a discussion of the accuracy potential of ISOLTRAP
also with respect to the study of very short-lived species. Furthermore, results of
mass measurements on rare earth isotopes will be presented for the first time.
A discussion of other more recent ISOLTRAP results and technical improvements
can be found in further contributions to these proceedings [21-24].
216 G. BOLLEN ET AL.
2. ISOLTRAP at ISOLDE
2.1. EXPERIMENTAL SET-UP
The tandem Penning trap spectrometer ISOLTRAP has been operated at ISOLDE
for many years. More than 200 masses have now been measured, those determined
since 1995 are listed in Table I. The essential technique of trapped-ion cyclotron
motion excitation coupled with time-of-flight detection has not changed, but the
spectrometer has considerably evolved in the method [6, 24-26] of stopping the
radioactive ion beam and of introducing the ions into the trap systems.
Figure 1 shows the present layout of the ISOLTRAP spectrometer [6]. The first
main component of the spectrometer has the task to stop the 60 keY ISOLDE beam
and to prepare it for efficient transfer into the cooler trap. This is achieved by an
RFQ trap ion beam buncher system [24, 26], which allows to capture the continu-
ous ISOLDE beam in flight. The second component is a Penning trap [25] which
has the task to accumulate, cool, and mass separate with a resolving power up to
105 the ions delivered from the RFQ trap and to bunch them again for an efficient
delivery to the second Penning trap. This precision trap [27] is the actual mass
spectrometer where the cyclotron frequency Vc = (l/2JT)(q Im)B of an ion with a
charge-to-mass ratio q I m stored in a magnetic field of strength B is determined.
Table I. List of nuclides investigated with ISOLTRAP from 1995 until end
of 2000
Ar 33-37,42,43 Sm 136-143
Kr 73-78, 80, 82 Eu 139,141-149,151,153
Rb 74 Dy 148, 149, 154
Sr 76, 77 Ho 150
Sn 124.128-132 Tm 165
Xe 114-121,124,130,132 Yb 158-164
Ba 123,125,127,131 Hg 179-197
Ce 132.133, 134 Pb 196, 198
Pr 133-137 Bi 197
Nd 130.132.134-138 Po 198
Pm 136-141, 143 At 203
P
MASS MEASUREMENTS ON SHORT-LIVED NUCLIDES 217
TOF detectlon
cyclotron fnoquency [
determlnaUon magnet
~ ~ ~ Trap3
man •• Iactlve
cooling
eccumul.uon
and bunching
structure
-AI
RFQ
ISOLDE Ion
beam (de) Trapl Ion bunches
30 .. 60 keY
> :J.-. -2.5keV
-.&'9d'bender
HV platfoml
This means that the resolving power can be increased by enlarging the observation
time of the ions, but also that the half-life of a nuclide poses a limit on the maximum
achievable resolving power.
The statistical uncertainty (8vc/vc)stat with which the cyclotron frequency can
be determined is inversely proportional to both the resolving power R and to the
square root of the number N ian of the detected ions. The fore factor depends on
the resonance detection scheme used in the experiment. An investigation of a large
number of data obtained with ISOLTRAP shows that this factor is close to unity
resulting in (8m/m)stat ~ 1 . R- 1 . Ni~~/2. Using these relations the capability of
ISOLTRAP is evaluated in Table II for different scenarios. It is assumed that the
maximum storage time is twice the half-life of the investigated nuclide. The first
case (A = 19) illustrates a measurement of a very light short-lived nuclide. Only
400 detected ions are sufficient to achieve a statistical uncertainty of a few ke V.
Case 2 describes the situation of a high accuracy mass measurement on a short-
lived nuclide like 74Rb. In order to obtain a statistical accuracy as required for
example for a meaningful test of the Constant Vector Current hypothesis, a few
hundred thousand ions have to be detected. This can be achieved in a measurement
time of about one day if the beam intensity is sufficient to reload the trap every 100
ms with one ion. Case 3 (A = 131) describes the situation of a mid mass nucleus
218 G. BOLLEN ET AL.
Table II. Cyclotron frequency vc, minimum line width Llvc (FWHM), resolving power R, statis-
tical accuracy (8m/mhtat for ions with mass number A and half-life TI/2, investigated with an
observation time Tobs and with Nion ions detected
A TI/2 Tobs Vc (MHz) Llvc (Hz) R (10 6 ) Nion (8m/m)stat 8m stat (keV)
with a rather long half-life. Here a high resolving power can be achieved, enabling
for example the resolution of low-lying isomeric states [29].
Of course, the total accuracy of the mass values has to include possible system-
atic errors. The design of the present ion trap system for nuclear mass measure-
ments is such that systematic errors due to field imperfections are typically below
10 ppb. ISOLTRAP has shown in a series of measurements on stable nuclides that
an accuracy level of 10 ppb can be achieved. For example, in a recent measure-
ment a cyclotron frequency ratio r = vc(133CS+)jvc(85Rb+) = 0.639039326(6)
was determined with a statistical uncertainty of 8r j r = 1 . 10- 8 . This can be
compared with the ultra-precise O.l-ppb mass measurements [3] obtained by the
MIT group for these two ions. With their mass value for 133Cs as reference and
with the above frequency ratio a value for the mass excess of 85Rb of ME ( 85 Rb) =
- 82166.94(19) (keV) is obtained. The relative deviation from the MIT value is
only 2 . 10- 8 . The most common trap imperfections (higher order electric field
components, field eccentricity, tilt of electric field with respect to magnetic field
axis) [28, 30] result in a systematic error in the mass determination which scales
linearly with the mass difference between reference and investigated ion. From the
numbers given above it can be concluded that in the case of ISOLTRAP such a
systematic error is smaller than 10- 9 per mass unit.
The major uncertainty during on-line runs is still the stability of the magnetic
field. In the evaluation of most ISOLTRAP measurements with calibration mea-
surements performed only every few hours a value of 8mjm = 1 . 10- 7 has been
used as a conservative estimate for any arising systematic error. This value can be
reduced significantly if the magnetic field calibration is performed more frequently.
Additional care has to be taken to minimize systematic effects due to Coulomb
interaction between ions of different masses stored simultaneously [29]. Such ef-
fects can be completely avoided if the measurements are performed with only a
single trapped ion at any time.
for practically all of them. The most recent measurements concentrated on neutron-
rich nuclides around Z = 82, neutron-rich tin isotopes, neutron-deficient xenon
isotopes, and in the lighter mass region on argon isotopes.
In order to extract nuclear structure information from binding energies it is nec-
essary to study the fine structure of the mass surface. This requires measurements
in long isotopic or isotonic chains. For example, mass measurements have been
carried out in a long chain of mercury isotopes, 179- 197 Hg [15]. These measure-
ments have given very precise information on binding energies in this region. As
discussed in detail in [15] a deviation from expected trends is observed in the two-
neutron separation energies which may have its origin in the mixing of weakly and
strongly deformed states [31]. Other recent highlights are the measurements on
33 Ar [17, 22] and even more recently on 74Rb. With half-lives of TI/2 = 173 ms
and TI/2 = 64 ms these two nuclides are the shortest-lived ever studied in an
ion trap. These measurement became possible due to an improved efficiency of
the spectrometer and the employment of a fast measurement cycle [22]. As dis-
cussed in [17] the result on 33 Ar turns out to be an important new test for the
Isobaric Multiplet Mass Equation (IMME) [321. The mass measurement of 74Rb
is, in connection with present efforts in precision ,B-decay studies at ISOLDE and
at TRIUMFNancouver, an important step towards a precision test of the CVC
hypothesis at high Z.
Table III. Nuclides or molecules investigated in a beam time in September 1997. Oxide
ions were studied in the case of cerium. Column 2 gives the ratios of the cyclotron
frequency of the studied molecular or atomic ion to that of 142Sm. The first error
corresponds to the statistical uncertainty, the second to the total uncertainty of the fre-
quency ratios. Column 3 lists the mass excesses as tabulated in the 1995 mass tables
(AME95) [34] (except for 148Eu, see text). The last column shows the result of a new
mass adjustment in which all ISOLTRAP data are included
The resulting mass excess values are given in the last column of Table III. For
comparison, the values of the 1995 mass adjustment (AME95) [34] are included
in column 3. In the case of cerium, the mass excess values have been corrected for
the known mass excess of 16 0. For 148Eu the value listed in column 3 is not from
AME95, but one obtained in a subsequent preliminary mass adjustment in which
earlier ISOLTRAP data [14] were taken into account. This resulted in a mass value
change of 100 keV. This change renders possible a yet non-observed very weak EC-
decay of QEC = +25(10) keY whereas earlier such a decay was forbidden with
QEC = -41 (17) ke V. In the case of 133Ce the presence of the isomer was taken
into account by assuming an abundance ratio of 1.0(0.5) and making a correction
of 19(5) keY in order to derive the ground state mass. In the cases 132Ce and 134Ce
the original errors were increased from 22 to 32 keVin order to account for possible
BaF contaminations in the trap. In summary, this set of measurements demonstrates
once more the large impact ISOLTRAP has on the knowledge of binding energies
of unstable nuclides. For more than one third of the studied nuclides masses were
MASS MEASUREMENTS ON SHORT-LIVED NUCLIDES 221
not known before and in most cases a mass accuracy of better than 25 ke V was
achieved.
3. Outlook
The high performance of ISOLTRAP is the guarantee for many more import<mt
mass measurements to come. In the very near future measurements are planned
on 32 Ar, which will provide an important contribution to a fully experimental de-
termination of the limits of scalar currents in weak interaction [35]. In a similar
line it is foreseen to continue the very recent measurement on 74Rb, important in
the context of a test of the Constant Vector Current hypothesis. After the first suc-
cessful experiment on this nuclide it is now planned to push the accuracy towards
the I keV level. Further plans exist to continue studies on neutron-rich nuclides in
the vicinity of doubly magic nuclei like 132Sn and 208pb. This will contribute to a
better understanding of the single particle structure in these regions and provide
important experimental input data for the modeling of the astrophysical r-process.
References
I. Proc. of Nobel Symposium 91 on Trapped Charged Particles and Related Fundamental Physics,
Lysekil, Sweden, 1994, Phys. Scripta T59 (1995).
2. Proc. of Internat. Can! on Trapped Charged Particles and Fundamental Physics, Asilomar,
CA, USA, 1998, AlP Conf. Proc. 457, Amer. Phys. Soc., 1999, p. 111.
3. Bradley, M. P. el at., Phys. Rev. Lett. 83 (1999), 4510.
4. Carlberg, C., Phys. Rev. Lett. 83 (1999), 4506.
5. This issue.
6. Bollen, G. et al., Nucl. Instrum. Methods A 368 (1996), 675.
7. Bollen, G. et at., Hyp. Interact. 38 (1987), 793.
8. Kluge, H.-J., Phys. Scripta T22 (1988), 85.
9. Stolzenberg, H. et aI., Phys. Rev. Lett. 65 (1990), 3104.
10. Bollen, G. et at., 1. Mod. Opt. 39 (1992), 257.
II. Otto, T. et aI., Nuclear Phys. A 567 (1994), 281.
12. Beck, D. et at., Nuclear Phys. A 626 (1997), 343c.
13. Ames, F. et at., Nuclear Phys. A 651 (1999), 3.
14. Beck, D. et aI., European Phys. 1. A 8 (2000),307.
15. Schwarz, S. et aI., in preparation for Phys. Lett. B (2000).
16. Dilling, J. et aI., in preparation for European Phys. 1. A (2000).
17. Herfurth, F. et aI., submitted to Phys. Rev. Lett. (2000).
18. Savard, G. et at., this issue, p. 223.
19. Dilling, J. et at., Hyp. Interact. 127 (2000), 491.
20. Szerypo, J. et aI., In: Proc. of Radiocative Nuclear Beams 5, Divonne, France, 1999, in press.
21. Dilling, J. eta/., this issue, p. 331, p. 495.
22. Herfurth, F. et al., this issue, p. 309.
23. Schwarz, S. et al., this issue, p. 337.
24. Kellerbauer, A. et aI., this issue, p. 511.
25. Raimbault-Hartmann, H. et aI., Nucl. Instrum. Methods B 126 (1997), 374.
26. Herfurth, F. et at., submitted to Nucl. Instrum. Methods 2000, Preprint CERN-EP/2000-142.
27. Becker, St. et aI., Internat. 1. Mass Spectrom. Ion Proc. 99 (1990),53.
222 G. BOLLEN ET AL.
Abstract. The Canadian Penning Trap (CPT) mass spectrometer is a device used for high-precision
mass measurements on short-lived isotopes. It is located at the ATLAS superconducting heavy-ion
linac facility where a novel injection system, the RF gas cooler, allows fast reaction products to
be decelerated, thermalized and bunched for rapid and efficient injection into the CPT. The CPT
spectrometer and its injection system will be described in detail and its unique capabilities with
respect to its initial physics program, concentrating on isotopes around the N = Z line with particular
emphasis on isotopes of interest to low-eneri'v tests of the electroweak interaction and the rp-process,
will be highlighted.
1. Introduction
The Canadian Penning Trap (CPT) Mass Spectrometer is a device aiming at the
precise determination of the atomic masses of short-lived isotopes. It was designed
for specific measurements of interest to weak interaction tests at low energy which
require the ability to obtain mass measurement accuracy in the 1-10 ppb range
on isotopes with half-live as short as 50 ms. The required accuracy with only
a very small sample of short-lived ions available dictated the use of an ion trap
system. This then required the development of a new injection technique for short-
lived isotopes in ion traps, both more efficient and more universally applicable than
existing systems, combined to a versatile production mechanism. Such a system is
now operating at Argonne National Laboratory and its characteristics are presented
in the following.
* Corresponding author; e-mail: SAVARD@anlphy.phy.anl.gov.
224 G. SAVARD ET AL.
2. Physics motivation
The physics program at the CPT mass spectrometer involves mass measurements
ranging from those on heavy stable nuclides, to broad mass survey in regions far
from the valley of beta stability, to high-precision measurement on specific iso-
topes. Mass measurements are of general interest to test nuclear model predictions
in regions where little data is available but more importantly provide key input
to our understanding of many physical phenomena involving isotopes far from
stability. Of particular interest are explosive astrophysical processes such as the
rp- and r-process but also specific decays along the N = Z line which are key
laboratories to test fundamental interactions at low energy. The N = Z region
is of particular interest because of the symmetry between the neutron and proton
wave functions which often lead to simplifications that allow precise interpretation
of physical processes occurring in complex systems. A particular example of this
is the study of superallowed 0+ to 0+ ,B-decays where, because the axial-vector
decay strength is zero for such decays, the measured ft values are directly related
to the weak vector coupling constant through [1]:
(1)
The unstable isotopes are produced at the target location of the Enge split-pole
magnet by fusion-evaporation reaction of the heavy ion beams from the ATLAS
superconducting linac on thin targets. The ATLAS linac is capable of producing
very high intensity beams of stable ions at Coulomb barrier energy allowing sig-
nificant production of unstable isotopes. The targets and thin windows used in the
production of unstable isotopes are all cooled by gas to better sustain the intense
beam heating.
T Rf gos coclpr"
bunchpd opc.., spec tf'cgroph
II !lAO:; "POM
RFQ trop
Lo ser 10"" source
Figure 1. Schematic layout of the CPT mass spectrometer and its injection system at ATLAS.
226 G. SAVARD ET AL.
The kinematics of the reaction carries the reaction products forward and into the
spectrograph. The Enge split-pole spectrograph, operating in the gas-filled mode,
is used to separate the reaction products from the primary beam. The spectrograph
is used at zero degrees to maximize the collection efficiency and a mini Faraday
cup (made of tantalum) also located at zero degrees removes the bulk of the pri-
mary beam. A transmission type parallel-plate avalanche counter is used to locate
the reaction products in the spectrograph and tune the magnetic field. A tunable
degrader made of a series of thin aluminum foils on a rotating stage is then used
to decelerate the fast reaction products to an energy of roughly 15-30 MeV before
they reach the focal plane of the spectrograph.
This initial production and preseparation is performed with an efficiency vary-
ing between 10% and 100% depending on the kinematics of the reaction and
limited mostly by the acceptance of the spectrograph. The time involved in this
part of the process is of the order of 0.5 f.,l,S.
The reaction products must then be transformed from a fast continuous beam of
enormous longitudinal and transverse emittance to a bunched cooled beam suit-
able for efficient injection into ion traps. A new device, the RF gas cooler, was
developed for this task (see Figure 2). The reaction products are stopped in a large
volume gas cell filled with high-purity helium at a pressure of about 150 Torr.
In the slowing down process the reaction products recapture charge with the vast
majority of them ending up in the singly-charged state where they remain since
further neutralization on the helium is forbidden by its high ionization potential.
The singly-charged reaction products are then extracted from the gas cell by a
combination of DC and RF fields (in the high pressure gas cell) which bring the
Figure 2. Extraction end of the gas cell system leading to the first two sections of the RF gas
cooler. The gas cell and the RF gas cooler are segmented to apply guiding electrical fields
while applied RF signals provide radial confinement of the ions.
THE CANADIAN PENNING TRAP SPECTROMETER AT ARGONNE 227
ions towards an exit nozzle where the flow carries them out. The ions are pushed
by the flow into a segmented linear RF structure which pulls the ions through
two differential pumping apertures while the gas is pumped away by large root
blowers. The final section of the linear RF structure is a linear RF trap where the
ions are accumulated and cooled. The bunched ions are then ejected every 50 ms
and transported through a shielding wall by a low energy electrostatic transport
line to the CPT mass spectrometer located in an adjacent room. The transfer line is
equipped with movable microchannelplate and Si detectors to provide diagnostics
for both stable and low-intensity pulsed radioactive ions.
The RF gas cooler compresses the longitudinal phase-space of the reaction
products by over lO orders of magnitude with a total efficiency of about 20%. This
technological breakthrough now forms the basis of a new approach for the pro-
duction of radioactive beams which has been adopted for the next generation large
scale radioactive beam facility, RIA [5], in preparation in the US. The delay time
introduced in the extraction out of the gas cell is about 7 ms with the maximum ex-
traction fields usable at this time and a new cell currently under construction should
allow this delay to be reduced to a ms or less. The total delay time introduced in
this part of the cycle can be as low as 10--20 ms.
The mass measurements take place in the CPT mass spectrometer which consists of
a radio-frequency quadrupole (RFQ) trap, used for accumulation and preparation
of the ion sample, and a precision machined Penning trap located inside the highly
homogeneous 5.9 T field of a superconducting solenoid.
The RFQ trap can receive ions from two sources: (1) the laser ion source pro-
vides the CPT with ions of stable isotopes either for calibration of the device
or actual precision measurements, and (2) the RF gas cooler provides unstable
isotopes gathered in the Enge split-pole spectrograph. Both sources are synchro-
nized to inject ions into the RFQ trap at a specific RF phase to maximize capture
efficiency. The RFQ trap is operated with a helium buffer gas pressure of 10-5 -
10- 6 Torr to cool the ions before the capture of the next ion bunch. This part is
also essential to remove any 'memory' effect due to the use of different sources
for calibration and measurement. After accumulation of a sufficient number of
ions, the ions are extracted from the RFQ trap and transfered to the Penning trap
via a ramped cavity which matches the phase-space of the extracted beam to the
acceptance of the Penning trap.
Capture efficiency in the RFQ trap is about 50% from the laser ion source and
5-10 times less from the RF gas cooler. The minimum cycle time is of the order of
] 0 ms, limited by the cooling time in the RFQ trap. A second ramp cavity is being
added to the transfer line to better match the phase-space of the ions extracted from
the RF gas cooler to the acceptance of the RFQ trap. This should raise the capture
efficiency for ions from the RF gas cooler by a factor of 5.
228 G. SAVARD ET AL.
330 . - - - - - - - - - - - - - - ,
310
190
170 -I--~----~-~-----'
75 80 85 90 95 100
AppIJod Ftoq ... ncy ·459000 Hz
Figure 3. Cyclotron resonance obtained in lO min of data taking on stable 197 Au with a 0.5 s
interaction time in the 5.9 T of the superconducting magnet.
The Penning trap is fed by the RFQ trap and is used to determine accurately the
mass of the stored ions by measuring their cyclotron frequency
(2)
4. Performance
The CPT mass spectrometer and its injection system are operating on-line at
ATLAS with a total efficiency, from isotope production to measurement in the
Penning trap, of the order of 10-3 . The demonstrated capabilities of the machine
with radioactive isotopes are 1 ppm mass precision for isotopes produced with
cross-section at the fraction of a mb level and 100 ppb mass precision for isotopes
produced at the 1-10 mb level. Both are demonstrated in Figure 4. The left panel of
Figure 4 shows resonances at the modified cyclotron frequency where two species
e 20Cs and 120Xe) produced on-line are simultaneously loaded into the trap and
selectively removed via a mass selective resonant excitation. This approach yields
mass accuracy at the ppm level. The right panel of Figure 4 shows a cyclotron
resonance obtained on l20Cs after the l20Xe is selectively removed from the trap.
The cyclotron resonance yields more accurate mass values, below the 100 ppb level
THE CANADIAN PENNING TRAP SPECTROMETER AT ARGONNE 229
100 ,eo
90
80
I ,eo
70 1 '10
~
60
50
f-f - ..X
) 160
8 40 150
:JO ~ 140
20
,30
10
0 120
752350 752400 752450 752500 752550 752600 2.
Frequency 1Hz) Applied Frequency · 754000 HiE
Figure 4. Left panel: Modified cyclotron resonance for l20Cs and 120Xe loaded simultane-
ously in the Penning trap. Right panel: Cyclotron resonance on l20Cs after the 120Xe ions are
removed from the trap.
5. Conclusion
The CPT mass spectrometer and its novel injection system are now operating on-
line at the ATLAS accelerator in Argonne. They now provide the capabilities for
mass measurement accuracy at the 100 ppb level for short-lived isotopes produced
by fusion-evaporation reaction, independently of their chemical properties. This
allows precision measurements on short-lived isotopes not normally available at
ISOL facilities because of the limitations of the extraction out of thick targets.
These capabilities will soon be extended to the 10 ppb level of accuracy.
Acknowledgements
This work was supported in part by the US Department of Energy, Nuclear Science
Division, under contract W-31-109-ENG-38, and by NSERC of Canada.
References
I. Towner, I. S., Hagberg, E., Hardy, J. c., Koslowsky, V. T. and Savard, G., In: M. de Saint Simon
and O. Sorlin (eds), Proc. Int. Can! on Exotic Nuclei and Atomic Masses, Editions Frontieres,
Saclay, France, 1995, p. 711.
230 G. SAVARD ET AL.
2. Towner, I. S. and Hardy, J. C., In: E. M. Henley and W. C. Haxton (eds), Symmetries and
Fundamental Interactions, World Scientific, Singapore, 1995, p. 183.
3. Particle Data Group, Phys. Rev. D 54 (1996), 1.
4. Ball, G. C. et al., submitted for publication to Phys. Rev. Lett.
5. Savard, G. et al., accepted for publication in NucZ. Phys. A; see also http://www.phy.anl.govlRIA.
6. Bolien, G., Moore, R. B., Savard, G. and Stolzenberg, H., 1. Appl. Phys. 68 (1990), 4355.
Hyperjine Interactions 132: 231-244,200l. 231
© 2001 Kluwer Academic Publishers.
Abstract. The Penning trap mass spectrometer SMILETRAP has been considerably improved
during the last two years. The helium pressure has been carefully stabilized and is now indepen-
dent of irregular air pressure. The temperature of the hyperboloidal precision trap is stabilized to
±0.03°C. Remaining temperature instabilities are compensated by changes in the current of a warm
coil surrounding the precision trap. The frequency synthesizer is now locked to GPS. This means
that it is much easier to accurately mcasure resonances during several days. The improvements have
demonstrated that in mass doublet measurements with an excitation time of I s it is possible to
determine the mass of ions with q / A = 1/2 at an uncertainty to a few times of 0.1 ppb, using
selected rather than cooled ions. In routine measurements lasting for one day it is possible to reach
a mass uncertainty of 1 ppb. The masses of the following particles and atoms have been measured
with uncertainties in the region 0.3-2 ppb: p, 3H, 3He, 4He, 22 Ne, 28Si, 36 Ar, 76Ge, 76Se, 86Kr and
133Cs. It has also been shown that though we are using a warm bore the trap pressure is sufficiently
low to prevent electron capture from the rest gas for excitation times of 3 s and for ion charges as
high as 50+.
Key words: Penning trap, EBIS produced ions, precision mass measurements.
1. Introduction
The SMILETRAP facility is a Penning trap mass spectrometer connected to a an
electron beam ion source named CRYSIS at MSL, Stockholm University, able to
deliver pulses of highly charged ions with some 10+9 charges per pulse. The ob-
jective of the project is to exploit the fact that the precision in mass measurements
using a Penning trap increases linearly with the charge of the ion. The main features
of SMILETRAP have been reported at two earlier Conferences [1, 2], in Lysekil
1994 and in Ferrara 1997. Therefore, in this paper we concentrate on improvements
and results achieved during the last two years.
The main merit of using highly charged ions is clear from a very simple consid-
eration. As is well known the cyclotron frequency Vc for an ion with mass m and
* Corresponding author.
232 TOMAS FRITIOFF ET AL.
(2)
As mass reference ions we have used suitable charge states of 12C ions and some-
times highly charged ions of light ions whose masses are known to about 0.1 ppb.
It is obvious that the atomic mass is obtained by adding to the ionic mass m2
the mass of the missing electrons and their binding energies. Up to about 40 Ar
they are known from ionization energies with high accuracy. As will be seen at
the end of this paper we have determined the masses of several atoms where the
electron binding energies have had to rely on calculated values. Some of these
values have been reported at this conference by Eva Lindroth, Paul Indelicato and
their collaborators. Their calculations are claimed to be accurate to about 20 eV
which means that the resulting uncertainty in our atomic masses for heavy atoms
is only about 0.1-0.2 ppb.
It should be strongly pointed out that a serious possible systematic uncertainty
in precision mass measurements using Penning traps are associated with the va-
li(lity of Equation (2). This means that the magnetic field has to be extremely
stable as well as the frequency synthesizer. Therefore we have concentrated our
improvements on the following measures:
- Stabilization of the helium pressure in the dewar.
SMILETRAP PENNING MASS SPECTROMETER 233
Variations in air pressure changes the boiling point of the liquid helium that cools
the magnet. This temperature may change the field directly (distance variations) but
it can also affect it indirectly by changing the boil rate and the gas flow out from the
dewar. When the boil rate is changed the gas flow is changed inside the structure
that is holding the magnet. The structure will therefore be cooled or warmed by
the flowing gas. Differential contraction of the material might move the solenoid,
which will change the field inside the trap. Regulating the pressure inside the dewar
with a needle valve solves this problem. Not only does the pressure become con-
stant, but also since the heat load is quite constant, the flow also stays constant (we
have no automatic monitoring of the flow but manually it has been observed with a
flow meter). The pressure transducer that we use is a temperature compensated
absolute meter that is installed inside the box. The locking point is set with a
resistor, which is also inside the temperature controlled box. Changing the current
through the solenoid of a needle valve then regulates the pressure (Figure 1).
In order to monitor the air pressure a barometer was also installed. In Figure 2
one can see the regulated He pressure and the air pressure changes during the
hurricane that passed over Sweden at the beginning of December 1999. There are
no signs of any remaining air pressure effects.
To He
F====?:~==:::::::=--Recovery
system
Liquid Control
Helium Valve
Dewar
1085
~ -\\
I~ J
1
=
1075
Wv.
I-
1065
I~
1!U ~~ t-
I-
1055 ::4
I ~.
.. '''". . .' .
.... . .
';; '''';. , oJ' " . ' .:•• ~~
~ 1045 -
1093.4
1093:1
': , ... '\ . :-~ .... '~'" " '
E
! 1035 I ~ ';
I~
. -~
~
Q: 1015
1005
\ /'.
995 .-. /
985
\ ~ \ /
975
\ / \ /
965
" V
30-Nov l-Dec 2-Dec 3-Dec 4-Dec 5-Dec 6-Dec 7-Dec
Figure 2. Air pressure and helium pressure during the passage of two strong low-pressure
centers. As seen from the figure the violent changes in air pressure are not noticed by the
helium gas in the dewar.
air flow
fan with
heating coil
[ control
unit
I
temperature ~
probe ~
L$~~~
Figure 3. Temperature regulated air was blown into the closed system shown in the figure. In
this way a temperature stability of ±O.03°C was achieved.
SMILETRAP PENNING MASS SPECTROMETER 235
Trap Temperature
23.00 . , - - - - - - - - - - -- - - - - - - - ,
22.98 +-----------------~--l
22.96 +---:;o;;;j&_- - - - - - - -- - --...-jF----I
t=:==:__
U 22.94 +--JlL~,.-------------III_~---I
-;;: 22.92
~ 22.90 +- jli~=====~~t~=j
f- 22.88 + - - - - - --""'_ ,.-- - -......~-------i
22.861-------'4~1A~--r-----___j
22.84 + - - - - - - - - ----"....""'dIIII~-------I
22.82 +--~--__r_-_____.----.-::::JIIL-.__--.----l
1119/00 11\9/00 1/20100 1120100 1120100 1120100 1/20100 1120100
19:12 21:36 0:0 0 2:24 4:48 7:12 9:36 12:00
Figure 4. Temperature close to the trap.
1.95
,. 1.8
:t: 1.65
... 1.5
" 1.35
=
=
~
.
g' 1.2
"- 1.05
0.9
0.75
1/19/00 1/19/00 1/20100 1/20/00 1/20100 1/20100 1/20/00 1/20100
19:12 21 :36 0:00 2:24 4:48 7:12 9:36 12:00
Changes in the temperature of the trap and the steel of the vacuum chamber around
the trap changes the magnetic susceptibility of the materials. The room temper-
ature is regulated within 0.5 degrees for long periods but there are sometimes
uncontrolled sudden larger changes in the laboratory temperature. Therefore, a
temperature-regulated system (Figure 3) that forces air to circulate in the space
between the magnet bore and the vacuum chamber was installed some years ago,
which made the trap temperature stable within a few tenths of a degree. That is not
good enough and to investigate if it could be improved the old system was removed
and the space around the magnet was sealed from the outside. A temperature probe
was installed close to the trap. Figures 4 and 5 show the temperature change close
to the trap and the cyclotron frequency during the same period. There is a strong
correlation between the temperature and the measured resonance. A change in
temperature by 1°C corresponds to a cyclotron frequency change by as much as
9Hz.
236 TOMAS FRITIOFF ET AL.
40002
0>
U)
~ 39998 ..... , ..... _.. ........ ",.. 1, .... "....... ..t ....,,, ......... j.._. ...
..···'·'· .. ·1······,_······1·············-
c: ,
~
.2 39995
o
>-
o 39994~. .~nn~~,,~nn~~~~rMMn~TT~~~~
a 2 4 6 8 10 12 14 16 18
Hours
Figure 6. Measured cyclotron frequency of Hi with and without the active regulation of the
magnetic field.
(3)
SMILETRAP PENNING MASS SPECTROMETER 237
60
50
40 61 =
:hoo ms'
'"~
'E 30 . - -. ----- ---f-- -- t---- -
o
(,)
20
rI ._ ..i.
I
_. +I .....
10
O~~'-~~~-r~~~~~~~~~~
o . 25 ~ ~ 100 1~ 1~ 1~ ~
Time-ol-flight hIS)
40
I
i
. -+-.....
30 • .• ..j
61 - 3700 rris
70G 2~'
e i i
'"
'E
~
20 r-i ... --.-. -t-_·_·-· 1
o
(,) ,
i
i
10
O~~,-~~~~~~~~~~~~~~
o 25 50 75 100 125 150 175 200
Time-ol-Ilight (lIS)
Here R (J/mol K) is the molar gas constant, T (K) is the gas temperature, v (crnls)
is the velocity of the highly charged ions, a (cm2) is the charge exchange cross
section, and N is the Avogadro constant. The cross section for the charge exchange
reaction has been taken from [2] and ale VIcharge ion energy has been assumed.
For the experiment that we carried out on 76Ge22+ [5] the obtained ratio was
0.088(41)% which corresponds to a pressure <6.4(3.6) x 10- 12 mbar (less because
238 TOMAS FRITIOFF ET AL.
oo,o ~------------------------------,.
SO,O •• : I
70,0 !
Ui 00,0 i
i 50,0 .:
~ 40,0 ·1
:l: 30,0 .i
20,0 ·1
10,0 .i
0,0 +---r------r---,----.----r---r------r---,---~
the ions created during 3.6 s is the sum of 76Ge22+ and Hi ions). This value is quite
good considering the fact that we us a warm system and since the precision trap is
not easily pumped down due to several small apertures (holes for entrance and exit
of ions).
Assuming that this pressure value is correct we estimated the corresponding
half-life for various ions and the electron capture rate for 1 and 3 second excitation
times (Figure 8 and Table I).
The pressure in the trap is thus not a limiting factor up to 5 seconds excitation
times even for very high charges. However, as a safety it is always possible to apply
a 'cleaning' pulse during the excitation in order to get rid of the unwanted charge
states. The pressure may be reduced by a factor 2-3 by inserting more NEG-strips
and new turbopumps. To reduce the pressure still more the only way is probably
use to a cold bore. There are, however, several projects that can be investigated
with the present or modestly improved vacuum.
SMILETRAP PENNING MASS SPECTROMETER 239
Table II. Examples of some relevant mass determinations in SMILETRAP
7 ~--------------~'---~------~~------~--~---3
6
15 25
4+ 5+
T
5
3
All others
(C,Ne,Ar)
2 L -__~__- L__~____L-__~__- L_ _~_ _ _ _L-__~__- L__~
Figure 9. In a series of experiments we used various charge states of the highly charged ions
indicated in the figure as mass standards in a determination of the proton mass.Hi ions were
used as a carrier of the proton. Indicated with a solid line is the presently accepted proton
mass. As seen our value of the proton mass is close to the accepted value except for the alpha
particle. This observation could either be due to a wrong accepted helium mass or an unknown
systematic error in our determination of light masses.
Table III. Comparison between the atomic masses of 4He and 3He determined in
SMILETRAP and the accepted value
mass values deviate from the values determined by the Seattle group [5] with as
much as as about 2 ppb compared to the given uncertainty of 0.25 ppb. We have
been informed by R. S. van Dyck Jr. (Head of the Seattle group) that the large
deviation might be due to a night to day change of the magnetic field that at the
time of their measurements was not known by the group.
The difference between the masses of 3He and the 3H gives the Q-value of the
tritium beta-decay. Most Q-values of the decay of 3H have been obtained from
the end point of the beta-spectrum. In the search for the limit of the antielectron
neutrino it could be useful with a very accurate Q-value obtained from precise
values of the masses of 3H and 3He. Therefore, we have also measured the tritium
mass using a dedicated source, in which the tritium is trapped in a uranium bed.
The tritium is released by warming up the metal. CRYSIS was used for two reasons.
Firstly, the ionization efficiency is extremely high in an electron beam ion source.
During five days of running only 0.14 ml tritium was consumed. Secondly, no
molecules survive and therefore all ions are 3Hl+. We agree with the Q-value of
the Seattle-group. Thus very likely there is the same systematic error in both their
3He and 3H mass determinations [8]. In our case we have so far relatively high
uncertainties when comparing the frequencies of Hi and 3He2 + , respectively 3H I +,
due to the fact that their q / A-values differ. Therefore we will also measure the mass
of 3He 1+, that will give an even more reliable Q-value
2041.5...---,-------,-------,-..,
2041.0
2040.5
2040.0
>
~
~ 2039.5
Q) 2039.0 ....•.......•..•••...••.•.••.•....•••••.••••...•.•.••.•:11; ..
::J
tU 2038.5 ·····································1················ ....
>
62038.0
2037.5
2037.0
2036.5
2036.0 -'--,-----....----..------r-...J
2 3 4
Measurement
Figure 10. 76Ge double beta decay Q-value detenninations. Deduced from: 1 - [14], 2 - [12],
3 - [13], 4 - our recent measurement (2000).
to neon-like ions (or almost neon-like) for which the electron binding energies can
be very accurately calculated. In this way we reached a total uncertainty of 1 ppb
for the masses of the two atoms, which in both cases represents a 17-fold mass
improvement. However, in the mass difference the systematic uncertainties practi-
cally cancel since the measurements of the two ions species were made in the same
way and thus are associated with practically the same systematic uncertainties. The
Q-value obtained is 2039.005(50) keY, a 7-fold improvement (Figure 10). Thus the
uncertainty of 50 eV is only a small fraction of the resolution of the Ge-detectors
used. Therefore, the new Q-value can be used with great confidence for locking
the position of the electron peak in the analysis of the electron spectrum of the
sophisticated Ov2f3 - double beta decay of 76Ge. It should be pointed out that if the
last measurement reported by the Manitoba group would be correct, the reported
uncertainty would be sufficient for giving a reliable position of the expected elec-
tron peak. However, in precision physics at least two groups should agree since it is
sometimes very hard to control systematic errors. We have thus not only confirmed
the last reported Q-value of the Manitoba group but also considerably improved
the value.
Acknowledgements
We gratefully acknowledge the support from the Knut and Alice Wallenberg Foun-
dation, the Carl Trygger Foundation, the Bank of Sweden Tercentenary Founda-
tion, the Swedish Natural Research Council, the European Community Science
and TRM-Programs and the Manne Siegbahn Laboratory.
The upper part of SMILETRAP is more or less a copy of ISOLTRAP at CERN,
however, aiming at a much higher precision. We are very much indebted to Prof.
H.-J. Kluge and Prof. G. Bollen for their continued interest and support, in partic-
ular during the building up period at the University of Mainz
References
1. Carlberg, c., Hyp. Interact. 114 (1998), 177.
2. Carlberg, C. et aI., Physica Scripta T 59 (1995), 196.
3. Graff, G., Kalinowsky, H. and Traut, J., Z. Phys. A 297 (1980),35.
4. Bollen, G. et aI., Nucl. Instr. Meth. Phys. Res. A 368 (1996), 675.
5. Beck, B. R., Steiger, J., Weinberg, G., Church, D. A., McDonald, J. and Schneider, D., Phys.
Rev. Lett. 77 (1996),1735.
6. Douysset, G. etal., Phys. Rev. Lett. 86 (2001), 4259.
244 TOMAS FRITIOFF ET AL.
H. SAVAJOLS
GANIL. BP 5027, 14076 Caen cedex 05. France
Abstract. The measurement of masses (or binding energy) of nuclei far from stability is of fun-
damental interest for our understanding of nuclear structure. Their knowledge over a broad range
of the nuclear chart is an excellent and severe test of nuclear models. This is why considerable
experimental and theoretical efforts have been and are invested in this domain. Here we want to give
a description of direct mass measurements combining high-resolution time-of-f1ight determinations
and accurate momentum measurements in a spectrometer. This very fast and direct method was
developed at GANIL with the high-resolution spectrometer SPEG. The mass measurements realised
and scheduled provide important first indications of new regions of deformation or shell closures
very far from stability.
Key words: atomic mass, nuclear binding energy, mass spectrometer, time-of-f1ight.
1. Introduction
The precision required in a mass measurement is highly variable. In some applica-
tions the binding energy must be known with a precision of a keY or better (QED,
QeD). Very far from stability, the mass predictions diverge to the level of MeV.
Hence measurements at this level of accuracy or somewhat better, of the order of
100 keY, will provide significant constraints. Nonetheless, direct mass measure-
ments corresponding to a determination of the binding energy with an error of
100 ke V need a high resolution and a high precision. Indeed, for a nucleus of mass
A = 100, a precision ofthe order of 1.10- 6 is necessary. By a time-of-flight method
this level of accuracy requires a relatively long flight path ("'" 100 m) to a high
resolution spectrometer. The SPEG technique has no significant lifetime limitation
and can be employed with very low secondary beam intensities (0.01 particle/s).
Therefore mass measurements of very exotic nuclei, located close to the drip-line
in the mass region of A "'" 10-40, are accessible.
The following discussion is divided into four main parts: (a) the production
of exotic nuclei far from stability for mass measurements, (b) the SPEG method,
(c) some results, and (d) the future of the SPEG mass measurement program.
production target. The dominant mechanism in this energy domain is the fragmen-
tation of the projectile. The projectile like fragments are then selected in flight by
the beam line a-shaped spectrometer and transported to the high-resolution spec-
trometer SPEG [1]. The energy of the fragments are of the order of a few Ge V. The
power of this process for the production of exotic nuclei stems essentially from the
concentration of the cross section within a very small angular and velocity range as
the incident beam energy increases. This production mechanism is well suited for
achieve a high transmission in a spectrometer, that may at the same time provide
the selection of nuclei of interest. Predictions of cross-section are incorporated
in codes, such as LISE [2] or as INTENSITY [3], which include selection by a
magnetic spectrometer. Such codes are very useful for preparing an experiment, to
define the most suited reaction and the magnetic and eventually electric parameters
of the spectrometer.
Typically, projectile fragmentation of 40-48Ca and 64Ni has been used to measure
masses of very neutron-rich nuclei. The fragmentation of a 78Kr beam was used to
measure masses of proton-rich nuclei near the N = Z line.
To increase the transmitted yield of exotic nuclei, a dedicated device called
SISSI [4] has been constructed. It consists of a set oftwo superconducting solenoid
lenses of very short focal length (0.6 m). The first solenoid is used to focus the
incoming beam in a spot of ±0.2 mm on a fast rotating target. The second provides
an increase in the transport line angular acceptance. A large acceptance angle of up
to ±80 mrad is achieved with a small emittance of 16rr mm mrad due to the small
beam spot. This allows to transport the secondary beam with a standard beam line.
In order to have a broad range of reference masses which is a crucial point in a
direct mass measurement, the production target rotating at around 2000 turns/min
in the SISSI device has steps of different thickness. In the last mass measure-
ment the tantalum target was divided in three parts, i.e., 89% has a thickness of
550 mg/ cm2 for the nuclei of interest and 10 and 1% has a thickness of 450 mg/ cm2
and 250 mg/ cm 2 to produce heavy reference masses with suitable count rates.
3. Experimental procedure
The method used is a direct time-of-flight technique (Figure 1). The power of this
approach comes from the coupling of the projectile fragmentation as a produc-
tion method to the high-resolution spectrometer SPEG. The very broad elemental
and isotopic distributions resulting from such reactions combined with the fast in-
flight electromagnetic selection can provide the mapping of an entire region of the
nuclear mass surface in a single measurement.
The beam line from the exit of the a-spectrometer and the focal plane of the
SPEG spectrometer is doubly achromatic. In such a device, the mass is deduced
from the relation
ymov
Bp= - - ,
q
THE SPEG MASS MEASUREMENT PROGRAM AT GANIL 247
Op (Ga loLLe)
/
Powl obJel !-'PEG
/
Tslar'l (Gallemps)
where Bp is the magnetic rigidity of a particle of a rest mass rna, charge q and
velocity v and y the Lorentz factor. This technique requires only a precise deter-
mination of the magnetic rigidity and the velocity of the ion.
The time-of-flight (ToP) is measured using a pair of microchannel plate detector
systems located near the production target (start signal) and at the final focal plane
of SPEG (stop signal). The flight times are typically of the order of 1 f.!s for a path
82 m long. The intrinsic resolution of the start and stop detectors are of the order of
100-200 ps (PWHM) leading to a time-of-flight resolution of /'+,.t j t "" 2.10- 4 . One
crucial problem in direct mass measurements is to eliminate systematic errors in
the time-of-flight determination due to differential nonlinearities of the electronic
chains, i.e., time-to-amplitude and analog-to-digital converters. Por that purpose,
we use time signals from a precise high-frequency clock uncorrelated with the
beam. The period used is TC = 80 ns. Time differences from such signals and both
start and stop signals are measured for each nucleus. These time differences are
random, and differential nonlinearities are then averaged over the whole range of
TACs and ADCs. According to this technique, determination of the time-of-flight
follows the relation
where T,qart and T"top are the time differences, N is the number of TC periods and
To is a constant.
The magnetic rigidity 8 of each ion is derived from two horizontal position
measurements. The first measurement is performed by a thin position-sensitive
microchannel plate system located at the dispersive image planes of the analysing
magnet, i.e., at the conventional target chamber where the dispersion in momen-
tum is big (10 cmj%). The second is made by two drift chambers used after the
spectrometer. Thus reconstruction trajectories of each ion is possible and we accu-
rately determine the value ofthe magnetic rigidity independantly of the object size.
A momentum resolution of 10- 4 is commonly achieved.
248 H. SAVAJOLS
The identification of each ion arriving at the focal plane of SPEG is achieved
by the measured flight time and the energy loss and total energy signals from a
detector telescope. The telescope consists of four cooled silicon detectors of typ-
ical ticknesses of 50, 300, 6000 and 6000 !--lm. The energy loss signal is obtained
by summing the signals from the first two elements of the telescope, while the
total energy signal is obtained by summing all first three detectors. The last silicon
detector is used in anticoincidence.
A number of points should be considered carefully. The population of isomeric
states with lifetimes of the order or greater than the flight time through the system
(~ i!--ls) is a potential problem, as the resolution of the system is not sufficient to
resolve the typical mass difference between the ground and isomeric excited states.
The contribution of such states will lead to a systematic shift in the measurement
towards less bound masses. For 32 Al an isomeric state in a ratio of about 2%
has been found [8]. As a check that the deduced masses are not affected by the
existence of isomers, the device includes the detection of delayed y-rays by a 4rr
NaI array surrounding the telescope.
A mass resolution of 2-4 . 10-4 is obtained from the combination of the time-
of-flight and the magnetic rigidity measurement. This corresponds to ±3 MeV of
the mass excess, for a nucleus A = 40 the final uncertainties range from 100 ke V
for thousands of events (nuclei relatively close to stability) to 1 MeV for tens of
events (nuclei approaching the ends of isotopic chains).
As noted above, a large number of nuclei (~ 100) are transmitted in a single set-
ting of the beam line and the spectrometer (Figure 2). The nuclei with well-known
masses and with adequate yield provide a calibration from which the unknown
masses are derived. Such a large number of reference masses over a wide range
of Z and A is particularly important for providing final mass determinations with
precisions of the order of 10-6 . In the last mass measurement experiment [24],
only masses where three or more independant and compatible mass determinations
were available were used as references.
From these references we estimate in each case the systematic uncertainty in-
troduced by the method. A mass is then deduced from an interpolation done by
a Taylor series between the references. The terms of this series reflect the small
inaccuracies (or adjustments) in the method. A typical development for the mass
determination is
where the first four terms could be considered as a first order adjustment. The last
two terms are higher order corrections. The fifth term corresponds to the difference
in the time collection of ions in the silicon telescope. The last term corresponds to
effects related to energy loss of particles in the emisive foils of microchannel plate
detectors or entrance foils of both small drift chambers. The constants (Yl-a6 are
adjusted for known nuclei. Mass uncertainties include statistical, calibration and
THE SPEG MASS MEASUREMENT PROGRAM AT GANIL 249
...!!, t.400
Os
M
"
1200
a f ~. ~ • J
s I~~~{f Ti,
p
1t)OO
.lI
AI
II,
aoo
/'h.
III ~
soo F
400 C
JJ
200 ..........~.........~...........~.J......-.~..........~..........~........~~~...L..~"-'-'~......
700 750 800 1160 goo 5160 'DOO '060 " 00 1160 1200
extrapolation errors. Of course the farther away the extrapolation, the bigger is the
extrapolation error.
The first mass measurement with SPEG took place in 1986. Since then many
experiments have been performed in the region of light and medium nuclei. In-
deed, the limit for obtaining reasonable mass determinations with present detector
technology is A "-' 80.
On the proton-rich side, the fragmentation of a 78Kr beam on a Ni target has
been used to reach nuclei near the N = Z line in the mass region A "-' 60-80
SPEG98 [5]. These masses are important for providing input for astrophysical
modelling for the rp process and for information on the nuclear structure in a region
of high deformation. As more than 200 nuclides are produced in this fragmentation
reaction, with high yields for nuclei close to the stability, it was necessary to purify
the secondary beams. The method is based on the stripping of the ions in a thin foil
located the two dipole stages of the a-spectrometer. It allows the selection in terms
of atomic number Z, and does not increase beam emittance. The mass excesses of
70Se and 71 Se have been determined with a precision of "-'5.10- 6 [5]. These results
agree well with the estimates of Audi and Wapstra based on systematic trends.
250 H. SAVAJOLS
20000n-,"--,,--r--.--,---r--.--,---, - - .
15000
:>-
Q)
~
'--"' 10000
.::
ru
(f}
AI
5000
15 16 17 18 19 20 21 22 23 24 25
Neutron Number
Figure 3. Two neutron separation energy S2n as a function of the neutron number N for
neutron-rich nuclei in the vicinity of the N = 20 shell closure.
On the neutron-rich side, efforts were made to measure masses near the N = 20
and N = 28 shell. Indeed, the evolution of shell closures far from stability is a
subject of much actual debate. Deformations, shape coexistence or variations in
the spin orbit strength as a function of the neutron to proton ratio could provoke
the modification of magic numbers. The more natural observable that may probe
this change in the standard shell ordering is the separation energy of the last two
neutrons, S2n'
Around N = 20, two groups using different time-of-flight recoil spectrome-
ters, the TOFI [6] and the SPEG systems, dedicated mass measurements in that
region. Figure 3 shows the S2n-values as a function of the neutron number in the
vicinity of the N = 20 shell closure collected in the SPEG86 [7], SPEG87 [8] and
SPEG91 [9] campains. The results of Ne, Na and Mg isotopes exhibit an anomaly
around N = 20. An overbinding, corresponding to an increase in the two neutron
separation energies, at the neutron sd-shell closure is observed, where a decrease
would be expected. Additionally, a much lower than expected excitation energy
was determined for the first 2+ state in 32Mg [10, 11]. This overbinding is related by
the breaking of the N = 20 magi city where an island of deformation is observed.
These results have been confirmed by Coulomb excitation [20] and theoretical
THE SPEG MASS MEASUREMENT PROGRAM AT GANIL 251
-
>GJ
~
25
20
0::
en
('oj
15
Ca
10 K
0
18 20 22 26 28 30
Number of neutrons
Figure 4. Experimental 521l values in the region of the N = 20 and N = 28 shell closures.
The circles correspond to values from [25] . the bold circles to values for which the precision
was improved and the full circles to masses measured for the first time.
5000~--------------------------~------.
;; 4000
!~3000
Q
;] 2000
;( 1000
'=.
>(
41 o
/
~
-= -1000
-200015
20 25 30
Number of neutrons
Figure 5. Shell correction as defined in the text of the mass Ca, Sand P.
where d Mexp is the experimental mass excess and d MpRLDM is the macroscopic part
of the FRLDM. The shell correction energy is plotted in Figure 5. As in Figure 4,
the qualitatively different behavior of the Sand P isotopes as compared to the
Ca isotopes is clearly evident. The Ca isotopes show pronounced shell correction
minima around N = 20 and N = 28. The Sand P isotopes do not exhibit such
effects at N = 28, and show a discontinuity in the slope at N = 26.
From experimental masses, two different independant observables, the S2n-va-
lues and shell correction, show the same behavior. These results are reproduced by
shell model [19] and relativistic mean field calculations [28]. The models predict
that deformed prolate ground state configurations associated with shape coexis-
tence are necessary to explain the experimental data. The existence of an isomeric
state in the 43S in the same experiment and its interpretation by shell model cal-
culations confirm the analysis of the masses and constitutes the first evidence of
shape coexistence in this region [29].
6. Summary
The direct time-of-flight method with SPEG is one of the important advances in
the last decade for the measurement of masses far from stability. The technique
has no lifetime limitation and can be used with very low event rates. Therefore,
this is at present the only method for measuring masses up to the neutron drip-
line in the mass region A '"'-' 10-50. The masses thus obtained provide a means of
identifying new nuclear structure effects that are well illustrated by our work in the
N = 20 and N = 28 region. Further studies of a more detailed and experimentally
complex nature, such as Coulomb excitation and decay spectroscopy, may then be
performed on selected nuclei.
Improvements in the technique, which are still possible, will give us a unique
opportunity to measure masses for nuclei at the limit of stability. The measurement
of nuclear binding energies far from stability provides an early stimulus for the
development of nuclear models.
Acknowledgements
I am grateful to Wolfgang Mittig, who has initiated the mass measurement program
with SPEG, and to all people who collaborated in SPEG experiments (see [5, 7-9,
24]).
254 H. SAVAJOLS
References
1. Bianchi, L. et aI., Nuclear Instrum. Methods A 276 (1989),509.
2. Tarasov, O. et ai., http://www.ganil.fr/lise/proglise . html.
3. Winger, J. et aI., Nuclear Instrum. Methods B 70 (1992), 380.
4. Joubert, A., IEEE 1 (1991).
5. Chartier, M. et ai., Nuclear Phys. A 637 (1998),3.
6. Wouters, J. M. et aI., Nuclear Instrum. Methods A 240 (1985), 77.
7. Gillibert, A. et aI., Phys. Lett. B 176 (1986), 3l7.
8. Gillibert, A. et ai., Phys. Lett. B 192 (1987),39.
9. Orr, N. A. et aI., Phys. Lett. B 258 (1991), 29.
10. Detraz, C. et ai., Phys. Rev. C 19 (1979), 164.
11. Guillermaud-Mueller, D. et ai., Nuclear Phys. A 426 (1984), 3.
12. Motobayashi, T. et ai., Phys. Lett. B 346 (1995), 9.
13. Campi, X. et ai., Nuclear Phys. A 251 (1975), 193.
14. Wildenthal, B. H. et aI., Phys. Rev. C 28 (1983), 1343.
15. Poves, A. et ai., Nuclear Phys. A 571 (1994), 221.
16. Warburton, E. K. et ai., Phys. Rev. C 41 (1990), 1147.
17. Werner,T. R. et aI., Phys. Lett. B 335 (1994), 259; Nuclear Phys. A 597 (1996), 327.
18. Retamosa, J. et ai., Phys. Rev. C 55 (1997), 1266.
19. Caurrier, E. et ai., Phys. Rev. C 58 (1998),2033.
20. Ren, Z. et ai., Phys. Lett. B 380 (1996), 241.
21. Pfeiffer, B. et ai., Z. Phys. A 357 (1997), 235.
22. Sorlin, O. et ai., Phys. Rev. C 47 (1993), 2941.
23. Glasmacher, T. et ai., Phys. Lett. B 395 (1997), 163.
24. Sarazin, F. et ai., Phys. Rev. Lett. 84 (2000), 5062.
25. Audi, G. et ai., Nuclear Phys. A 624 (1997), 1.
26. Orr, N. A. and Mittig, w., unpublished.
27. Moller, P. and Nix, J. R., At. Data Nucl. Data Tabies 59 (1995), 185.
28. Ren, Z., RMF code, informations on request.
29. Sarazin, F. et aI., In: APAC'2000 Proceedings.
30. Savajols, H., E364 GANIL proposal.
31. Haustein, P. E. (ed.), At. Data Nucl. Data Tabies 39 (1988),185.
32. Utsuno, Y. et at., to be published.
33. Tapper, R. J., Rep. Progr. Phys. 63 (2000), 1273.
34. Berdermann, E. et ai., preprint 2000-09 GS!.
Hyperfine Interactions 132: 255-263, 2001. 255
© 2001 Kluwer Academic Publishers.
DAEG S. BRENNER
Clark University, Worcester, MA 01610, USA
Abstract. A program to measure masses along the astrophysical rp-process path in the
A ~ 60-80 region is underway at the Wright Nuclear Structure Laboratory, Yale University. The
classic technique of end-point determinations for f3+ spectra measured in coincidence with daughter
y-rays is used to determine QEC which, in tum, is used to calculate the mass. Several innovations
have been incorporated to increase the sensitivity and selectivity of the method. Results of recent
experiments are reported.
Key words: atomic mass, nuclear binding energy, nuclear spectroscopy, Q-beta.
1. Introduction
Network calculations, which are used to simulate the astrophysical rapid proton
capture (rp) process and to model energy production in stellar scenarios, require
considerable input data including masses, half-lives, and rates of particle-induced
and photodisintegration reactions [1]. These calculations, which map reaction path-
ways, are hampered by the limited amuunts of experimental data available beyond
Z = 32. Atomic mass measurements are especially important since these enter the
calculations in several ways: as QEC values and as proton and a-particle binding
energies. Experimental mass measurements in the rp-process region with uncer-
tainties ~ 100 ke V, somewhat better than can be estimated using mass formulas,
could help refine these calculations.
2. Experimental methods
Mass measurements have been made for reaction products produced using the
ESTU tandem Van de Graaff accelerator at the Wright Nuclear Structure Labora-
tory, Yale University. The layout of the nuclear spectroscopy experimental station
that was used in early experiments is shown in Figure 1. The target is located
immediately behind a set of two circular collimators that define a 3-mm diameter
beam spot. Recoiling reaction products are collected on an aluminized Kapton tape
positioned approximately 6 cm behind the target. In order to prevent the primary
beam from melting the tape a 3-mm diameter Au plug is centered on the beam axis
approximately 5 cm behind the target and in front of the tape. The placement of
the plug is adjusted for individual experiments depending on the calculated angu-
256 D. S. BRENNER
Tar et box ,- - - --
~-------- ,
,-
~______ _ ____ ;1
Holding box
,,
1'---------Jf-------Jr{j ~-----'r
, - - - I
Detector area
Figure 1. Schematic diagram of the Yale nuclear spectroscopy measurement station used for
the 71 Se experiment.
59"
_ Ge Clover Detector
LEPS Detector
D ~-Detector
Figure 2. Multi-detector configuration used for recent experiments.
grounds. In order to simulate a Fermi-Kurie analysis the positive square root of the
counts is plotted as a function of the energy channel up to the first channel where
a negative number is encountered. Data for that channel and all higher channels
are eliminated from the fit region to avoid distorting the linearity of the square-root
plot. We find this method of analysis to be more reliable than a de-convolution
approach for spectra with low statistics, as is often the case for nuclei far from
stability. The known decay scheme is then used to guide the selection of the high-
energy region of the f3+ spectrum where only a single f3+ -decay branch is thought
to be present.
Difficulties arise in determining QEC values from f3+ -spectrum end points due
to the effects of summing in the f3+ detector from random and coincident radia-
258 D. S. BRENNER
tions. The former is a function of counting rate and is not a serious problem when
counting rates are low, as is the case in our experiments. Summing of coincident
radiation, on the other hand, is a significant problem which can be quite complex
and important, depending on the decay scheme and type of detector used for the
f3+ measurements. A detailed discussion of summing effects can be found in an
earlier publication [2] that was used to guide our analysis.
Two examples of mass measurements will be briefly discussed, 71Se and 72Br.
71Se was produced by the 58Nie60,2pn) reaction. A 1 mg/cm2 metal foil target of
58Ni, enriched to an isotopic abundance of >99%, was bombarded with a ""3~0
particle nA beam of 160 ions with incident energy of 65 MeV. At lO-min intervals
the deposit spot was moved to a shielded location where decay of the sample was
measured using the apparatus illustrated in Figure 1. The plastic scintillator was
calibrated from f3+ -end-point measurements for 67Ge and 70 As produced under
nearly identical experimental conditions by bombardment of 58Ni with 160 ions at
75 MeV. The calibration curve derived from the 67Ge and 70 As ,8+ -spectrum end
points is shown in Figure 3(a) together with those for 71Se decay.
An example of the end-point spectra obtained for 71Se is shown in Figure 3(b)
for the f3+ spectrum in coincidence with a 147 keV y-ray which de-excites a level
of the same energy in 71 As. The f3+ end-point energy and QEC value calculated for
this gate and four others are listed in Table I. The weighted mean QEC value for all
gates is 4.762(35) MeV which when combined with the adopted mass excess [3]
of its daughter, 71 As, gives a mass excess value of -63.130(35) MeV for 71 Se.
In some respects measuring QEC for 71 Se was an easy experiment since the
production cross section was large, "" 150 mb for the reaction chosen, and there
were several strong f3+ branches to levels in the daughter nucleus 71 As. As one
moves further from stability experiments become increasingly more difficult; pro-
duction cross sections drop, QEC increases on average, half-lives shorten, and the
f3+ -decay strength is fragmented and spread over more levels at higher excitation
energies in the daughter. 72Br, our next candidate for study, presented some of
3500
67G e, 7°As
2500
(a)
2000LL~~~~~~~~~~~~~~
(b)
15
t! 10
Z
5
o
280 320 360 400 440
Channel
Figure 3. (a) Calibration curve for the plastic scintillator fJ+ detector determined by end
points for 67 Ge and 70 As. Also shown are the 71 Se fJ+ end points that were used to calculate
the QEC. The equation shown is the fJ+ detector energy calibration in keY. (b) Square root
plot of the end-point region of the fJ+ spectrum in coincidence with the 147 keY 71 As y-ray
which depopulates a level of the same energy. The linear fit to the data was made over the
channel range 300-410 which corresponds to a fJ+ energy range of 2486-3370 keY. This fit
range excludes any significant contributions from fJ+ feeding to levels higher in the decay
scheme.
260 D. S. BRENNER
Table /I. {3+ end-point energies and QEC values for 72Br
a From [4]. Separate systematic and statistical errors were reported by the authors: 70Se, ±0.35
(syst.) ±0.30 (stat.); 71Se, ±0.25 (syst.) ±0.20 (stat.). These have been added in quadrature for
comparison to our results.
b From [8].
these challenges since QEC is estimated to be "-'8.7 MeV and the production cross
section is lower. In order to address these we replaced the single Ge detector with
three clover detectors as shown in Figure 2. For the 72Br experiment the LEPS
detector was not present. Figure 4(b) shows the fJ+ spectrum in coincidence with
the 753 keV y-ray which de-excites a level at 3124 keV in 72Se. The fJ+ end-point
energy and QEC value calculated for this gate and one set on the 1125 keV y-ray
which de-excites the same level are listed in Table II. The weighted mean QEC
value is 8.869(95) MeV, which when combined with the adopted mass excess [3]
of 72Se, gives a mass excess of -59.025(96) MeV for 72Br as shown in Table III.
This result is in good agreement with the adopted value of -59.150(260) MeV [3].
The close correspondence shows that fJ-y coincidence spectroscopy can be used
to measure masses reliably and with good precision, even when QEC is large, as
long as the production cross section for the nuclide of interest is not too small, in
this case "-'75 mb. The 72Br experiment also illustrates how each new measurement
can be used to demonstrate the linear response of the fJ+ detector. Note that we
now use 71 Se as a "standard" to provide several additional calibration points and
these nicely lie on a linear fit that includes the other well-known calibrators, 65Ga,
67 Ge, and 68 As.
MASS MEASUREMENTS 261
6000
,-..
>-Col 5000
.:r::
'-'
..
~
E
en
+ 4000
~
3000
(a)
2000
300 500 700 900 1100
Endpoint Channel
10
(b)
8
72Br
6
N
"'
.-
Z
4
o ~~~~~~~~~~~~~~~~~
Kr(36)
Br(35) 48 50
Sa (34)
As (33)
Ge(32)
Ga(31)
zn (30)
Cu(29)
Ni (28)
Co (27)
Fa (26)
22
3. Results
A motivation for our work was to measure masses to a higher precision than
currently available in the region along the rp-process path. Figure 5 shows the
status of mass data in the rp-process region above the waiting point nucleus 56Ni.
The stars show the nuclides where we supply new, improved results. In Table III
our mass excess values are compared to those from direct mass measurements,
to adopted values and estimates, and to predictions of several recently published
theoretical mass models. In general, our results are in quite good agreement, within
stated errors, with those of Audi et al. [3]. Masses for 70Se and 71Se, produced by
fragmentation of a 78Kr beam and measured by a time-of-flight method at GANIL,
were recently reported [4]. The higher precision that we obtain for these nuclides
is evidence that the classic method of f3-y spectroscopy, enhanced by the use of
clover Ge y-ray detectors and a highly efficient recoil collection and transport sys-
tem, is a competitive method for mass measurements in this region of the nuclidic
chart.
Our method is, of course, limited by the availability of target-projectile com-
binations that can produce the nuclide of interest with sufficient yield and by the
half-life and branching in the decay scheme. We estimate the practical limitations
to be a production cross section of ;:: 10 mb and a half-life of ;:: 1 s for the present
apparatus. Extension to less favorable cases may be possible in the future by using
a more efficient y-ray detector configuration; we expect to add a fourth clover
detector to our system. Eventually, the availability of intense beams of radioactive
nuclides produced directly at a Rare Isotope Accelerator could be used to extend
mass measurements to nuclides lying on or beyond the rp-process path.
MASS MEASUREMENTS 263
Acknowledgements
The authors who have benefited from the hospitality and assistance of the staff
of the WNSL while visiting Yale University wish to express their gratitude. This
work is supported by the U.S. Department of Energy under Grant Nos. DE-FG02-
88ER40417 and DE-FG02-91ER40609, and Contract No. DE-AC02-98CH10886.
References
1. Schatz, H., Aprahamian, A., Gorres, J., Wiescher, M., Rauscher, T., Rembges, J. F., Thiele-
mann, F.-K, Pfeiffer, B., Moller, P., Kratz, K-L., Herndl, H., Brown, B. A and Rebel, H.,
Phys. Rep. 294 (1998), 167.
2. Foy, B. D., Brenner, D. S., Davids, C. N., Seweryniak, D., Blumenthal, D. Gill, R. L.,
Zamfir, N. v., Warner, D. D. and Barton, C. J., Phys. Rev. C 58 (1998), 749.
3. Audi, G., Bersillor, 0., Blachot, J. and Wapstra, A, Nuclear Phys. A 624 (1997), 1.
4. Chartier, M., Mittig, W, Orr, N. A, Angelique, J.-C., Audi, G., Casandijian, J.-M., Cunsolo, A.,
Donzaud, c., Foti, A, Lepine-Szily, A., Lewitowicz, M., LUkyanov, S., MacCormick, M.,
Morrissey, D. J., Ostrowski, A. N., Sherrill, B. M., Stephan, c., Suomijarvi, T., Tassan-Got, L.,
Vieira, D. J., Villari, A C. C. and Wouters, J. M., Nuclear Phys. A 637 (1998), 3.
5. Moller, P., Nix, J. R., Myers, W D. and Swiatecki, W J., At. Data Nucl. Data Tables 59 (1995),
185.
6. Aboussir, Y., Pearson, J. M., Dutta, A. K and Tondeur, F., At. Data Nucl. Data Tables 61
(1995), 127.
7. Nayak, R. C. and Satpathy, L., At. Data Nucl. Data Tables 73 (1999), 213.
8. Sharma, K. S., Hagberg, E., Dyck, G. R., Hardy, J. C., Koslowsky, V. T., Schmeing, H.,
Barber, R. c., Yuan, S., Perry, Wand Watson, M., Phys. Rev. C 44 (1991), 2439.
Hyperfine Interactions 132: 265-273, 2001. 265
© 2001 Kluwer Academic Publishers.
YD. E. PENIONZHKEVICH
loint Institute for Nuclear Research, Flerov Laboratory of Nuclear Reactions, 141980 Dubna,
Russia
Abstract. The present status of mass measurements from reactions producing nuclei at the driplines,
including those unstable to nucleon or cluster emission, is discussed. The results of recent heavy ion
and Jr-meson induced experiments on the study of the superheavy hydrogen isotopes (4H, 5H, 6 H),
helium (9 He, IOHe), lithium (IOLi, 11 Li) and beryllium (l3Be) are given. The possibilities of mass
measurements in radioactive ion beam experiments are also considered.
Key words: nuclear reactions, mass excess, missing mass, invariant mass, resonance scattering.
1. Introduction
The mass of a nucleus is one of its main characteristics for nuclear physics as
well as for other sciences. In metrology the mass of the nucleus is used in the
determination of the fundamental constants, in atomic physics - for calculating the
electron binding energy and in astrophysics - in the planning of different projects,
including projects on nucleosynthesis.
In nuclear physics, the mass of the nucleus is one of the basic quantities used
in the development and testing of different nuclear models, also for predicting the
properties of new nuclei, including superheavy ones. Detailed discussions on the
measurement and use of nuclear masses in nuclear physics take place at traditional
conferences like AMCO and ENAM. For this reason, this paper is dedicated only
to some experimental results connected with the mass measurement and structure
of several nuclei at the driplines, obtained in nuclear reactions.
In the present work, results are presented, obtained in the Flerov Laboratory of
Nuclear Reactions (JINR, Dubna), as well as in collaboration with other scientific
centers.
nuclear systems having N / Z > 2.5 and lying at the limit of particle stability is
to investigate the isotopes of lightest elements and determine their mass in the
ground and excited states. The structure of such nuclei may tum out to be quite
different from what is observed close to the ,B-stability line. The most direct way
to measure their mass and to study their structure is the missing mass method [1],
from reactions, which have two products in the exit channel, allowing to determine
the properties of one partner by means of the energy spectrum of the other. This
circumstance becomes of great importance when the studied nucleus has no bound
states. The neutron binding energy in such nuclei (negative or positive) can reach
several ke V, and for this reason the mass measurements should be performed with
similar precision. The missing mass method was for the first time successfully
applied in experiments on mass measurements of the super neutron-rich nuclei
such as 5,6H, 7,8,9He in heavy ion reactions.
Information on the properties of the heavy isotopes of hydrogen, helium, lithium
and beryllium is given in details in [2]. In the present work presented are the latest
results obtained by the collaboration FLNR (JINR, Dubna) - the Hahn-Meitner
Institute (Berlin) - RSC "Kurchatov Institute" (Moscow) by the missing mass
method in heavy ion binary reactions.
From the laws of energy and momentum conservation in a binary reaction:
where M and P are the rest masses and momenta of the incoming particles (1, 2)
and the final products of the binary reaction (3,4), Q is the energy release in the
reaction, it follows that:
P1 M 3 cose
M1+M2
+{ 2M3M4[E1M2I(M] M2) + +Q - E*] _ [P]M3 Sine]2}1/2,
M3+ M4 Ml +M2
• I
•
g.$.
- bound slates
ot nucleusB
E(b). MeV
Figure 1. Schematic representation of the energy spectra obtained in two-body reactions
A(a, b)B.
98e( 14 C. 140)9He
30
Qo--34.58(10)MeV 337 MeV
(7.9MeV)
3.98MeV
13.04MeV
1.21MeV
15
17 0°->
g.s.
10 140+3n
5
160°_> 140+2n
150,-----------------------------------------------------,
i "C("B••2 N) "Be
190 MeV
? 4.6 deg
"5
u 100
50 en
c.:i
o~~~~~~~~~~~~~~~~
-2.0 -0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0
EeJcc:(MeV)
Figure 3. Spectrum of the 14C(1l B, 12N)\3Be reaction after subtraction of the background
component from target impurities. The experimental spectrum is presented by the histogram.
The fitted spectral components as well as the sum of the three-body continuum for 12N + 12Be
+ n and breakup \3 N* -+ 12N + n are shown by full lines. The black points denote the sum
of fitted components shown by full lines. The dashed line is a four-particle phase space [6].
In this way, as follows from the above, the mlssmg mass method makes it
possible to obtain unique information on the masses and the structure of stable
as well as of unstable nuclei. However, the further advent into the region of even
more exotic nuclei using this method is connected with great difficulties due to the
necessity to have a stable complementary partner.
This problem can successfully be solved using the missing mass method for
three-body reactions. For instance, the nucleus 12Li can be studied in the reaction
MASS MEASUREMENTS IN NUCLEAR REACTIONS 269
i~Mn + I~O ---+ I~Li + iHe + i~Ni, where two magic nuclei are formed (helium
and nickel), which in turn gives a gain in the reaction Q-value and correspondingly
in the cross sections. Such a method can be used to measure the masses of nuclei
such as 12He, 12Li, 16Be and heavier ones like 260, etc.
...
'"
°c
::I
~ 1
t:I
:E
;;
-d' SO
"0
>:
O~~~~LWLC~~~
-2 -1 0 1 2 3 4 5 6
Es ' MeV
He-n-n
Figure 4. Invariant mass spectrum from the reaction CD2(IILi, nn 8He) at E = 61 A MeV.
Curves 1-4 show possible contributions from the II Li fragmentation. Curve I corresponds to
the IOHe resonance at E8He-n-n = l.2 MeV. Curve II illustrates sensitivity to energy and
corresponds to E8Hc-n-n = 0.5 MeV.
found with the results given in [5], where this nucleus was investigated in reactions
using the missing mass method with primary beams. It should be noted that in [5],
in addition to the ground state a few excited states with good energy resolution
were observed for the nucleus IOHe. The mass excess of IOHe was found to be
48.94 MeV in [7, 8] and 48.81(7) MeV in [5].
If in the Garvey-Kelson relation [10]
the masses given in [11] are used, then for the mass excess of IOHe one obtains the
value M.E. = 49.09(24) MeV, which is very close to the experimental value [5, 7].
Of great interest are the radioactive ion beams, what concerns the mass mea-
surements of light nuclei close to the driplines. For example, 25, 26 0 can be studied
in the dCZ40,p) and tCZ 40,p) reactions. It is planned to carry out such experiments
at FLNR with the use of a gaseous deuterium and liquid tritium targets.
Quite recently in Dubna at the U-400M cyclotron, using a secondary 6He beam
in the reaction p(6He,pp)5H the mass excess and respectively the binding energy
of 5H was measured [12]. There have been attempts to produce this nucleus during
the last 20 years in different reactions [1]. However, only in radioactive ion beam
experiments was it possible to get information on its mass excess, to observe qua-
sistationary states ( • ~ 0.5 Me V) and determine the binding energy with respect
to the decay 5H ---* 3H + 2n, which was found to be ~ 2.0 MeV. Thus, the nucleus
5H was found to be more stable than 4H, B(4H) ~ 2.5 MeV.
MASS MEASUREMENTS IN NUCLEAR REACTIONS 271
1200
10C+p
1000
'"'
......
II) 800
-e 600
c::
"tl
......
.g 400
200
00 2 3 . 5
~.m. (MeV)
Figure 5. Part of the excitation function of 10C + P elastic scattering. The intensity of the
10C beam was 7 x 103 pps and the measuring time 16 hours. The fully drawn curve shows a
potential model fit to the data in the low energy region and the dashed curve is a result of a fit
with a coherent sum of Breit-Wigner resonances.
272 YU. E. PENIONZHKEVICH
References
1. Ogloblin, A. A. and Penionzhkevich, Yu. E., In: Treatise on Heavy Ion Sciences, Vol. 1, Plenum
Press, New York, 1989, p. 261.
2. Kalpakchieva, R., Penionzhkevich, Yu. E. and Bohlen, H. G., Phys. of Particles and Nuclei
30(6) (1999), 627.
3. Bohlen, H. G. et al., Z. Phys. A 320 (1985),237.
4. Bohlen, H. G. et at., In: Yu. Ts. Oganessian, Yu. E. Penionzhkevich and R. Kalpakchieva (eds),
Proc. School-Seminar on Heavy Ion Physics, JINR Publishing Department, Dubna, 1993, p. 17.
5. Ostrowski, A. N. et at., Phys. Lett. B 338 (1994), 13;
Bohlen, H. G. et al., Nucl. Phys. A 583 (1994), 775.
6. Belozyorov, A.v., Kalpakchieva, R., Penionzhkevich, Yu. E. et aI., Nucl. Phys. A 636 (1998),
419.
7. Korsheninnikov, A. A. et aI., Phys. Lett. B 326 (1994),31.
8. Penionzhkevich, Yu. E., Phys. of Particles and Nuclei 25 (1994), 930.
9. Korsheninnikov, A. A. et at., Nucl. Phys. A 588 (1995), 23.
MASS MEASUREMENTS IN NUCLEAR REACTIONS 273
10. Garvey, G. T. and Kelson, I., Phys. Rev. Lett. B 157 (1966), 120.
II. Audi, G. and Wapstra, A., NucZ. Phys. A 565 (1993), I.
12. Bogdanov, D. D. et al., In: Proc. Int. Conf. Clustering Phenomena in NucZ. Phys., St.
Petersburg, 2000, p. 280.
13. Artemov, K. P. et al., Journal of Nucl. Phys. (Rus.) 52(6) (1990), 634.
14. Axelsson, L. et ai., Phys. Rev. 54 (1996), R1511.
15. Seth, K. K., Phys. Rev. Lett. 58 (1987), 1930.
16. Seth, K. K., In: Proc. Tnt. Con! ENAM'95, ArIes, Frontiers, 1995, p. 109.
17. Gomov, M. G. et al., NucZ. Phys. A 531 (1991),613.
Hyperjine Interactions 132: 275-281,2001. 275
© 2001 Kluwer Academic Publishers.
Abstract. The mass-measurement programme at GANIL aims to measure the masses of heavy nuclei
close to the N = Z line which is the ideal region to study neutron-proton pairing. An original direct
time-of-flight mass-measurement method was developed at GANIL which uses the CSS2 cyclotron
as a high-resolution spectrometer. The masses of ions of A = 68, 76, 80 and 100 have been measured
with a precision of a few 10- 6 . Mass measurements will be performed with the new CIME cyclotron
of SPIRAL using a similar method based on the measurement of the phase of the accelerated ions
for different radio-frequencies. A recently approved experiment will help develop this new technique
and aims to measure the mass of 31 Ar radioactive nuclei with a precision of 10- 6 .
Key words: atomic mass, mass spectrometer, cyclotron, N = Z nuclides, nuclear binding energy.
1. Physics motivation
Among nuclear structure studies far from the valley of stability, mass measure-
ments are of primary importance as they allow the determination of the limits
of existence of nuclei. The mass of a nucleus, and thus its binding energy, is
one of the most fundamental nuclear properties since all the information about
forces acting in such a system of nucleons is contained in this quantity. Various
models have been developed to reproduce and predict masses, particularily far
from the valley of stability: collective semi-empirical microscopic-macroscopic
models, phenomenological microscopic models such as shell-model calculations,
276 M. CHARTIER ET AL.
N
c o
~ 7 • N.. Z
Q)
~Q)
6 o
>
Q) S
>'
'"I 4
o
-IOU-~I~
O ~~W
~~~
~~~
~~~~~~@~~7~
O ~~W
~~~
~~IOO
atomic number I
Figure 1. Empirical n-p strength for even--even nuclei obtained from double binding energy
differences.
on the measurement of the cyclotron frequency of cooled trapped ions, the time
necessary for this operation represents a fundamental limitation for nuclei with
very short life times, although recent developments at ISOLTRAP have allowed
measurements with half-lives of about 100 ms (see Bollen et ai. and Herfurth et ai.
in these proceedings).
For these reasons, an original method was developed at GANIL, which is well
suited to the configuration of its coupled cyclotrons, based on the use of the second
separated-sector cyclotron, CSS2, as a high-resolution spectrometer [6, 7]. This
method required very low-intensity beam-tuning techniques and enabled the ac-
celeration of several exotic beams simultaneously for the first time at GANIL.
The CSS2 method consists of simultaneously accelerating ions of similar m/q
ratios and measuring, at the end of the acceleration process inside the cyclotron,
their time-of-flight difference ot, which corresponds to their phase difference with
respect to the radio-frequency o</>. When considering two ions of slightly different
masses, the heavier spirals longer inside the cyclotron and so the time-of-flight or
phase difference is proportional to their m / q difference:
o(m/q) ot
m/q t
If the two ion species are simultaneously accelerated, the unknown mass is de-
termined with respect to the other one taken as a reference, knowing the number
of turns inside the cyclotron or the total time-of-flight, or from several reference
masses. The mass resolution, 3 x 10- 5 , is better than the one obtained when using
a magnetic spectrometer because the time-of-flight of the ions is much longer:
around 50 fJ.-S over a flight-path of the order of a km. Masses of nuclei with half-
lives of a few tens of fJ.-S can therefore be measured. The CSS2 cyclotron has
a relatively large acceptance [6] which is also an advantage compared to other
devices. The exotic nuclei are produced by fusion-evaporation from a primary
beam delivered by the first GANIL cyclotron, CSSI, impinging on a target located
between the two cyclotrons. The ions are intercepted in a silicon-detector telescope
(~E 30 fJ.-m-E 300 fJ.-m) mounted on one of the probes movable radially along the
axis of a magnetic sector of the cyclotron (see Figure 2).
CSSI CSS2
co
production
target
Figure 2. The experimental set-up of the CSS2 mass-measurement method.
THE MASS PROGRAMME AT GANIL USING THE CSS2 AND CIME CYCLOTRONS 279
than the magnetic field changes, consists of varying the radio-frequency of the
cyclotron proportionally to the relative m j q difference of the ions:
8(mjq) 28/
-m-j-q- = -Y -/ .
References
1. Zeldes, N., Phys. Lett. B 429 (1998), 20.
2. Zhang, J. Y. et aI., Phys. Lett. B 227 (1989), 1.
3. Brenner, D. S. et aI., Phys. Lett. B 243 (1990), 1.
4. Van Isacker, P. etaZ., Phys. Rev. Lett. 74 (1995), 4607.
5. Van Isacker, P. et al., Phys. Rev. Lett. 82 (1999), 2060.
6. Auger, G. et al., NIM A 350 (1994), 235.
7. Chartier, M. et al., NIM B 126 (1997),334.
8. Chartier, M. et aI., Phys. Rev. Lett. 77 (1996), 2400.
9. Audi, G. et aZ., Nucl. Phys. A 624 (1997), 1.
10. Bazin, D. et al., Phys. Rev. C 45 (1992), 69.
11. Antony, M. S. et ai., At. Data NucZ. Data Tables 34 (1986), 279.
12. E. K. Warburton et ai., Phys. Rev. C 41 (1990), 1147.
Hyperjine Interactions 132: 283-289, 2001. 283
© 2001 Kluwer Academic Publishers.
Abstract. Projectile fragments of a 209Bi beam were separated in flight with the fragment separator
FRS and injected into the experimental storage ring ESR. In the ESR a beam containing up to about
100 different isotopes was cooled to a relative velocity spread of 8 v! v = 10-6 by means of the
electron cooler. The image currents of the ions induced in a Schottky pick-up probe at each tum were
recorded. A subsequent Fast Fourier Transformation of these signals yields the revolution frequencies
of the different isotopes stored in the ESR. Unknown masses of more than 150 neutron-deficient
nuclides in the element range of 52 ~ Z ~ 85 have been measured directly by Schottky Mass
Spectrometry and in addition more than 60 new masses have been obtained from a-decay chains.
These new mass data allow the location of the one-proton drip line and the prediction of the two-
proton dripline for heavy nuclides. The experimental masses are compared with different theoretical
predictions.
1. Introduction
The combination of the heavy-ion synchrotron SIS [1], the fragment separator
FRS [2], and the experimental storage ring ESR [3] at GSI is a unique facility to
study exotic nuclei. The FRS spatially separates the fragments in flight and injects
them into the ESR for precise mass determination performed by measuring the
revolution frequency of the stored ions. A large number of different nuclei can be
stored simultaneously which allows systematic investigations of large areas of the
chart of nuclides. In our experiments at the ESR, new masses of a wide area of
neutron-deficient heavy nuclides have been measured directly by Schottky Mass
284 YU. A. LITVINOV ET AL.
Production Target
8glCrrf 9Be
Projectiles /
2Q9 B~
~
from SIS Schottky
Projectile F ragments .... - Probe
Figure I. Layout of the high-energy facility at GSI used for mass measurements of exotic
nuclei.
Spectrometry (SMS) [5]. By using Qa-values it has been possible to extend the
knowledge of masses up to the proton drip-line and beyond [6]. In this paper we
describe the experimental progress and results obtained in our experiments.
2. Experimental setup
The layout of the experimental facility used for mass measurements is presented
in Figure 1. Projectiles of (800-900) MeVlu 209Bi were extracted from the SIS
and focused on a 8 g/cm 2 beryllium target located at the entrance of the FRS.
Projectile fragments were produced and emerged from the target with energies
of about 350 MeVlu. At this energy the fragments were completely stripped or
carried up to three electrons. In the last experiment the FRS and the ESR were
operated at a constant magnetic rigidity of Bp = 6.5 Tm. Different energies of
the primary beam were used to select different sets of nuclei to match this rigidity.
Usually less than 500 particles were injected at one injection. After injection the
stored fragments were forced to an identical mean velocity by the electron cooler.
A relative velocity spread of less than 8 v / v ~ 10-6 has been achieved for this low
number of stored ions [4]. The revolution frequencies in the ESR (~1.9 MHz) are
an explicit measure for the mass-to-charge ratio. Every circulating ion induced an
image current in the Schottky pick-up probe at each tum. The 34th harmonic of
these time-correlated signals was analyzed in order to match the signal bandwidth
with the input bandwidth of the signal analyzer (320 kHz). This bandwidth is just
large enough to cover the full acceptance of the ESR. The signals were mixed-
down, digitized with a sampling rate of 640 kHz and stored. The relative mass-to-
charge acceptance of the ESR of ± 1.5% allows to store more than 100 different
species of up to 25 elements at the same time.
.
~
.~15 ~lA
:I:
iii
::> o' ~., It.,. lL, alA..
..... 1M.. 0
i'"
i'" ~
;;; :0
0>
:I:
.ri OJ ,. , ... ,.,...... ~ • V Do>
.0 e
~ 10 150
~ g ID '""..,.
150.5
~ ! ~
151 151 S !::....
i: ;. '.0,:
.J:
0..
~ f M
~_
a:::
~&
:!U')
?_ "'~
.. " i .; • F
:e :3
i?:' O::_ ~i"'<i. ::: ~¥O::iii"'l- ;$.!=D:.o
~ m <c :....
.r' ~ a:
0
.~ 0:: 0.. _. :;j ci 0. t. 3: '" '\ /; ~~ ~
I l'!
rj, 0
:t. it
,,0 0.. :
~
~ ~
5
O~~~~~~~~~~~~~~~~~~~~~~
/ / 'I
50 70 90 110 130 150 170 190 210 230 250
Frequency I kHz
Figure 2. Part of a typical frequency spectrum from 50 to 250 kHz. The insert shows a typical
isobaric triplet. The peaks of Ir and Re correspond to a single stored particle.
~0.45
c
Q)
E 0.40
0.35
Figure 3. Ground and isomeric states of hydrogen-like 145Gd63+ ions. The frequency lines
here contain one particle each. The determined excitation energy is in good agreement with
the published value of (748.7 ± 0.1) keY [7] .
1.4
1 30 sec I (190:t50j keY
'~m Ho 67< 0+
1.0 3.7m
.l!l 154 Er 67"
'c 0.6
::J
>99%
..0
i5 0.2 2032
W ,':'.,-
Q EC= 8+ 260 31
;::.
'iii 1360 sec I ~
c
Q) 1.0 (2 ~ 11 .8m
C 1""9 Ho ST' 154 Ho67+
0.6
0.2
0.0 0.05 0.01 0.15
Frequency difference I kHz
Figure 4. Mass determination of ground and isomeric states of bare 154g ,mHo67 + ions. Our present
interpretation is the following: 30 s after injection only the isomeric state of 154Ho ions is observed,
see upper panel. 360 s after injection the 154mHo ions decayed and the ground state of 154Ho was
populated from the EC-decay of 154Er, see lower panel. The shift of the 154Ho peak matches the
known excitation energy of (260 ± 50) keY [7] . The corresponding decay scheme is shown in the
right part.
masses of ground and isomeric states with excitation energies of E* < 500 ke V re-
quires cautious interpretation of the data. Nuclides with very close mass-to-charge
ratios consequently have very close orbits in the ESR. The average transverse
distance of the trajectories can be as small as a few micrometers. These close-
lying frequency peaks can be distorted due to Coulomb repulsion which is seen
SCHOTTKY MASS MEASUREMENTS OF COOLED EXOTIC NUCLEI 287
5. Results
Masses of more than 150 neutron-deficient heavy nuclides were directly measured
for the first time [5]. These nuclides cover a large part of the chart of nuclides in
the element range from 52 to 85. This area contains a large amount of a-emitters,
which are linked by decay chains. Nuclides at the end of these decay chains usually
have half-lives longer than 10 s, which is presently the minimum time required for
employing SMS. The masses of these nuclides were directly measured by SMS .
The combination of these measurements with a-spectroscopy data allowed in ad-
dition the determination of masses of more than 60 nuclides [6]. This extends the
measured mass-surface to very exotic regions of the chart of nuclides beyond the
proton drip-line.
The amount of new mass values represents a crucial test of the predictive power
of theoretical mass models. We define the uncertaincy of a model as
2 l~ -
arms = -n ~(Mexp - Mmodel)i'
i=l
Mexp is the experimental mass value, Mmodel denotes the model-predicted mass
value and n is the number of nuclides. Figure 5 shows the arms values for some
-
1000
I
<:
r- 0
- '"
<:
.><
(I)
-
.
>.
iii (I)
.
-
.c
I 1
(I)
Z
q; Qj
.><
0
-, x
<: :l >. :;;:
0 N Qj ~ .c
'~ ~~ Gl c '
f!! 6 'c<0 .!! <0 ~lil i;'~
(II
=:l a. t!?(6
~~
~~I ~~
Gl --
Gl ;0 <0
o a.. 0 c% ~ C/) ~~ ~~ Ll
Microscopic- extrapolations
Microscopic Macroscopic Mass relations
macroscopic Audi et al.(1997)
Mass Model
Figure 5. Uncertaincy arms values for different theoretical mass models obtained from the
comparison with our experimental data [5, 6].
288 YO. A. LITVINOV ET AL.
f ooo·plOlon dnplina,predicted by
Umn and Zatdes. AI. Dala Nud Dala
Tablas 17.431(1976)
o kn<7o'<lll nuclides
Nd • stable nuclides
Pr
Ce new masses mea.sured alihe ESR
La
Sa
Figure 6. Part of the chart of nuclides showing the new masses measured at the ESR. The one-proton
dripline is shown for odd-Z nuclides and the two-proton dripline is shown for the even-Z nuclides.
mass models [7, 11-14]. The calculations differ by several hundred keY from the
experimental data and only the extrapolations of [7] are in a good agreement with
our results (see also [6]).
The explored mass surface is well suited for investigations of the one and the
two-proton dripline. The location of the experimental one-proton dripline is done
for heavy elements from bismuth to protactinium. The compilation of our data with
previously known information about masses and proton-separation energies allows
to reliably predict the one-proton dripline for elements below bismuth. With our
results we can also make linear short-range extrapolations and predict the location
of the two-proton drip line [6]. Figure 6 shows the new masses measured at GSI and
summarizes the present situation on the one and the two-proton dripline of heavy
nuclides.
and thus shorter times needed for the measurements. With these improvements
more than 200 nuclides with unknown masses in the half-life regime of 1-10 s
can be accessed by SMS in the future. In order to resolve isomers with lower
excitation energies without Coulomb distortion, new mass measurements with a
larger momentum compaction factor ex p are planned.
Acknowledgements
This work has been financially supported by BMBF under Contract No. 06GI4751
and No. 06LM363, and the Beschleunigerlaboratorium Munchen. Yu. A. L. and
Yu. N. N. would like to acknowledge the support from the WTZ grant No. RUS-
654-96. Z. P. would like to thank the Polish Commitee for Scientific Research
(KBN) grant No. 2P03B11715 for partial support.
References
1. Blasche, K. and Franczak, B., In: H. Henke, H. Homeyer and Ch. Petit-Jean-Genaz (eds), Proc.
of the Third European Part. Ace. Con!, Berlin, Vol. 9, Editions Frontiere, Gif-sur-Yvette, 1992.
2. Geissel, H. et aI., Nucl. Instr. Meth. B 70 (1992), 286.
3. Franzke, B. et aI., Nucl. Instr. Meth. B 24/25 (1987), 18.
4. Steck, M. et aI., Phys. Rev. Lett. 77 (1996),3803.
5. Radon, T. et aI., Nucl. Phys. A 677 (2000), 75.
6. Novikov, Yu. N. et aI., submitted to Nucl. Phys. A.
7. Audi, G. et aI., Nucl. Phys. A 624 (1997), 1.
8. Bollen, G. et aI., Phys. Rev. C 46 (1992), R2140.
9. Geissel, H. et aI., In: 7th Int. Con! on Nucleus-Nucleus Collisions, Strasbourg, France, July
2000; to be published in Nucl. Phys. A.
10. Firestone, R. B. and Shirley, V. S., Table of Isotopes, 8th edn, J. Wiley & Sons, 1996, p. 1660.
11. Haustein, P. E., Atom. Data Nucl. Data Tables 39 (1988), 185.
12. Aboussir, Y. et al., Nucl. Phys. A 549 (1992), 155; Atom. Data Nucl. Data Tables 61 (1995),
127.
13. Dufio, J. and Zuker, A., Phys. Rev. C 52 (1995), R23.
14. Myers, W. D. and Swiatecki, W. J., Nucl. Phys. A 601 (1996), 141.
15. Nolden, F. et aI., Nucl. Instr. Meth. A 441 (2000), 219.
Hyperjine Interactions 132: 291-297,2001. 291
© 2001 Kluwer Academic Publishers.
Abstract. A novel method for mass measurements of short-lived exotic nuclides is presented. Exotic
nuclides were produced and separated in flight at relativistic energies with the fragment separator
(FRS) and were injected into the experimental storage ring (ESR). Operating the ESR in the isochro-
nous mode we performed mass measurements of neutron deficient fragments of 84Kr with half-lives
larger than 50 ms. However, this experimental technique is applicable in a half-life range down to a
few JLS. A mass resolving power of 110000 (FWHM) has been achieved. Results are presented for
the masses of 68 As, 70,71 Se and 73Br.
Key words: mass measurements, 68 As, 70Se, 71Se, 73Br, exotic nuclei, storage rings.
1. Introduction
The atomic masses are known for most of the stable and long-lived isotopes, but
new experimental data are needed for nuclides far off stability. Small production
cross sections and short half-lives of these exotic nuclides are the reasons why
conventional techniques for mass measurements cannot access them. In this paper
we present a method that has been developed for mass measurements of rare and
short-lived exotic nuclear species. This technique is complementary to Schottky
Mass Spectrometry of cooled exotic nuclei [1, 2], which is used at the Experimental
Storage Ring (ESR) [3] for mass measurements of longer-lived exotic nuclides.
* Present address: NSCL, Michigan State University, South Shaw Lane, East Lansing, MI48824-
1321, USA.
292 M. HAUSMANN ET AL.
storage of
digitized
signal traces
MCP1
Digital Sampling
Oscilloscope
secondary electron
Figure I. Schematic view of the detector used for mass measurements of hot exotic nuclei [7].
Secondary electrons are released from the foil and are guided to MCP detectors by perpen-
dicular electric (E) and magnetic (B) fields. The magnetic field is generated by an external
magnet. The MCP2 branch of the detector is shown only in a very schematic way. The signals
of both branches are sampled by a digital sampling oscilloscope and the digital information
are stored.
1
> 0.5
-- 0
(ij -0.5
~ -1
·iii -1.5
-2
o 50 laO 150 200
time I J..LS
(a)
. .
0.5
.
0.5
'I I
.. *"... .. . . ... .. --
0 .,It > 0
--
>
(ij -0.5 (ij -0.5
c
*
*
c
Ol
..
Ol
·iii ·iii
*.
-1 -1
-1.5 -1.5
-2 -2
100 105 110 115 120 125 109.53 109.54 109.55 109.56
time I J..LS time / J..LS
(b) (c)
Figure 2. Trace of the signals from one detector branch. (a) Complete trace; (b) part of this
trace with signals of two particles (marked as .. and *). (c) Two individual signals of these
two particles expanded.
294 M. HAUSMANN ET AL.
1000
100
Ul
Q)
U
t
cO
a. 10
510 515 520 525 530 535 540 545 550 555
Revolution time / ns
Figure 3. Revolution time spectrum of neutron deficient krypton fragments. The ESR was set
to the magnetic rigidity of 53Fe at an energy of 344.66 A . Me V. The nuclides in this spectrum
cover a range of ± 7.4% in mass-to-charge ratio.
* Obtained from [8] and corrected for the missing electrons using [9].
296 M. HAUSMANN ET AL.
Table I. Atomic masses measured in this work and compared to other experi-
mental works and the compiled tables of [8]. The mass excess is given in units
of ke V. Extrapolated values from [8] are marked by #
masses of 68As and 70Se we chose 31p, 62Zn, 64Ga, 33S, 66Ge, 35CI, 39K, 41Ca, 43Sc,
45Ti and 47y in the first measurement, while in the second one we used 56CO, 58Ni,
60Cu, 62Zn, 66Ge, 35Cl, 39K, 41Ca, 43Sc, 45Ti and 47v. The obtained calibration was
used to calculate the mass-to-charge ratio for the nuclides under investigation. The
result (mass excess) for the individual measurement was derived from this.
In order to investigate the systematic error, for each reference nuclide except the
outermost ones a mass value was calculated by calibration to the other reference
nuclides. From the comparison of these values to the literature [8] (by a X2 test) a
systematic error (::::"12-55 keY) was deduced for each measurement and added in
quadrature to the uncertainties from the calibration. The final results were derived
by averaging the results of the two measurements with different settings of the
FRS, these are given in Table I. In addition, we compare our results to the compiled
experimental values and extrapolations of [8] as well as to the recent experimental
results from time-of-flight measurements [10]* and from f3 end-point determina-
tions [11]. We achieved a resolving power of m/l1m = 110000 (FWHM). This
result is slightly improved compared to the value reported for this method ear-
lier [12]. The improvement is a result of the increased number of turns observed
for a single particle.
5. Outlook
Further mass measurements of shorter-lived krypton fragments have been per-
formed. Here, the cross sections strongly favour the production of undesired con-
taminants, therefore the Bp-I1E-Bp method was used at the FRS [4]. Several
nuclides with half-lives far below one second have been identified in the revolution
time spectra of these measurements. Six of these short-lived nuclides (49 Fe, 48Mn,
45Cr, 43,44y and 41Ti) have unknown masses [8]. The nuclide 45Cr with a half-life
of 50 ms is at present the shortest-lived species identified. The nuclide 43y was
* The authors of [10] report separate statistical and systematic uncertainties; these have been
added in quadrature in Table I.
ISOCHRONOUS MASS MEASUREMENTS OF HOT EXOTIC NUCLEI 297
observed in the spectra, its mass is of relevance to determine the endpoint of the
astrophysical ap process. The data from these measurements are under analysis.
The duration of the individual records (200 JLs) shows that the presented method
is suitable for mass measurements of nuclides with half-lives down to several
JLS. In the future we intend to measure the masses of short-lived neutron-rich
fission fragments, which are important for the understanding of the astrophysical
r-process.
Acknowledgements
This work was financially supported by BMBF under Contract No. 06 GI 849 I and
No. 06 LM 363 and by the Beschleunigerlaboratorium Miinchen.
References
l. Radon, T. et at., Nucl. Phys. A 677 (2000), 75.
2. Litvinov, Yu. A. et aI., this issue, 283.
3. Franzke, B., Nucl. Instr. Meth. B 24/25 (1987), 18.
4. Geissel, H. et at., Nucl. 1nstr. Meth. B 70 (1992), 286.
5. Wollnik, H., Nucl. 1nstr. Meth. A 258 (1987),289.
6. Hausmann, M. et aI., Nucl. Instr. Meth. A 446 (2000), 569.
7. Trotscher, J. et ai., Nuc!. 1nstr. Meth. B 70 (1992), 455.
8. Audi, G., Bersillon, 0., Blachot, J. and Wapstra, A. H., Nucl. Phys. A 624 (1997), 1.
9. Huang, K. N. et at., At. Data Nucl. Data Tables 18 (1976), 243.
10. Chartier, M. et aI., Nucl. Phys. A 637 (1998), 3.
11. Brenner, D., this issue, p. 255.
12. Stadlmann, J. et aI., In: H. O. Meyer and P. Schwandt (eds), Nucl. Phys. at Storage Rings, Prac.
of the 4th lnternat. Can!: STORI99, Bloomington, 1999, AlP Conf. Proc. 512, 2000, p. 305.
Hyperfine Interactions 132: 299-307, 2001. 299
© 200 I Kluwer Academic Publishers.
Abstract. The MISTRAL experiment (Mass measurements at ISOLDE with a Transmission and
Radiofrequency spectrometer on-Line), conceived for very short-lived nuclides, has reached the end
of its commissioning phase. Installed in 1997, results have been obtained consistent with all aspects
of the projected spectrometer performance: nuclides with half-lives as short as 30 ms have been
measured and accuracies of 0.4 ppm have been achieved, despite the presence of a systematic shift
and difficulties with isobaric contamination. Masses of several nuclides, including 25-26 Ne and 32Mg
that forms the famous island of inversion around N = 20, have been significantly improved.
Key words: mass spectrometry, nuclear binding energy, exotic nuclei, neon, magnesium.
1. Introduction
The determination of atomic masses suffers from two particular difficulties: the fact
that masses are often determined from a series of reaction and/or decay values and
the requirement of very high measurement precision. The former is the problem of
the evaluator, such as explained by Audi in this issue, but can be helped by so-called
"direct" methods of mass spectrometry. The second is the problem of the experi-
menter who must develop a reliable technique. This difficulty is compounded for
studies of exotic nuclei where the half-lives become very short and the production,
very weak.
Among several complementary direct techniques (see review by Lepine-Szily
in this issue) one that is particularly attractive and has now delivered on its promise
for precision measurements of short-lived nuclides is that exploited by MISTRAL,
installed at the ISOLDE isotope separator facililty at CERN in 1997.
At once a magnetic mass spectrometer and a time-of-flight spectrometer, MIS-
TRAL is a unique instrument. It was inspired by the original spectrometer con-
* Now at NSCL, Michigan State University.
300 D. LUNNEY ET AL.
RF modulator detail
magnet:
beam
1 m
O.aT
field
secondary 60 kV
electron ISOLDE
multiplier beam
(single ion
detection)
ion source
(0-80 kV)
Figure 1. Overview of the MISTRAL spectrometer at ISOLDE. Insets show (lower left) the
trajectory envelope and (upper right) front and side views of the modulator.
ceived and built by Smith [1] in the 1960's. The main difference is that for use with
weak (e.g., radioactive) beams, the very small slit that Smith used to define the ion
trajectories must be enlarged.
MISTRAL determines the mass m of an ion of charge q in a homogeneous
magnetic field B by measuring its cyclotron frequency Ie = (l /2rr)(q / m) B from
transmission peaks produced by a radiofrequency voltage that modulates the ki-
netic energy of the injected (60 keY) ISOLDE beam. The technique is rapid, re-
quiring only the time-of-flight of the ions through the apparatus at full transport
energy. It is also very precise since it is capable of very high resolving power
(> 105 ), the magnetic field fluctuations are small and comparisons with a reference
mass are performed very frequently to eliminate field drift effects. A layout of
the spectrometer, together with the envelope of the modulated, isochronous ion
trajectories, is shown in Figure 1. For detailed descriptions see [2, 3].
c'"
~
o
~ ISV
'\
y~,,-
SOV' ~ 1500
gpo
rf
"1
Trsoluin6000r
r.
'
o
c:
.2
40000
250 V 90000
493825 493875 493925 493975 494025 494075 494 125 494 175
frequency (kHz)
ow
rt vo~age a"1llilude
modulallon eHiclency
(a) (b)
Figure 3. (a) The frequency response of the (tuned) impedance-matching circuit of the mod-
ulator. Though a fiat response over the 250-500 MHz frequency band is sought, only two
resonances are achieved. The response is measured by an antenna mounted near the modulator
having its own frequency response which is roughly deconvoluted. (bottom) The modulation
efficiency for 85 Rb ions at 60 keY, computed by the Monte Carlo simulation as, in fact, it is not
accessible experimentally. The absolute amplitude for a given frequency is the product of these
two quantities. (b) The 254 MHz modulation amplitude of a 60 keY 23Na beam measured at
the spectrometer midplane.
frequency response of the circuit used to match the modulator impedance to the
signal source, shown in Figure 3(a) (top). The realization of a matching circuit
over such a large frequency range is far from trivial [5]. The second factor, shown
in Figure 3(a) (bottom), is kinematic, varying with frequency and depending on
ion velocity and modulator geometry. The ion sees two electric field gaps on each
passage through the modulator (see Figure 1 inset). For maximum modulation
efficiency, the ion velocity must be such that the field changes polarity during
its flight through the central drift electrode to receive two tandem accelerations.
If the velocity is such that the ion feels equal and opposite electric fields, the
net acceleration (and resulting modulation efficiency) is zero. As seen from the
figure, this effect is periodic as the modulation frequency can go through many
oscillations for slow ions. The efficiency diminishes with frequency since the net
accelerating field is lowered for wider phase windows [5]. The modulation ampli-
tude, the experimentally observable quantity, is the convolution of the above two
factors. A radial profile of a modulated beam as a function of voltage is shown in
Figure 3(b) with the modulation amplitude given by the position of the maximum
intensity. (For mechanical reasons, it is not possible to measure the entire profile
which is normally symmetric.)
In the case of isobaric contaminations it is not only resolving power that is
important but also the relative position of the different peaks. A weak isobar can
be very nicely resolved but if it falls inside the flat, integer peak of another isobar
RECENT RESULTS ON Ne AND Mg FROM MISTRAL 303
produced with greater abondance (e.g., a stable species several orders of magnitude
stronger), we must change the frequency range in order to move the relative spacing
of the two peaks. It can also happen that the mass difference of two isobars is
large enough to be resolved and the two mass peaks still overlap but with different
harmonic numbers. Again, a change in frequency is required to change the relative
spacings. In cases where several isobars are present, this manoeuver - because
of the limited range of modulation efficiency coupled with frequency response
(Figure 3) - can easily become impossible.
-
>Q) 200 reference: 23Na 39/41K I ISOLDE
I
.
I
~
.........
100 29Mg
\2JNa
.L
I
ffi 2JNa
... "Na
"Ne "Mg
I .- !
I "Na
("Mg)
~ 0
I
~ T
1 1
=
,: i,,~, I !
~
I(
"AI "Mg
0) 2JNe 26Mg 29AI ("AI)
0) -100
-J "Ne
~
.....,: -200
CJ) MISTRAL (1999) : 26Mg
~ -300
ensemble of results
1 32Mg
Figure 4. Absolute difference of MISTRAL masses compared to the 1995 mass evalua-
tion [8]. The results are divided into three groups corresponding to the reference isotope
used: (left) 23Na, (center) 39 K, (right) isobar in the ISOLDE beam, indicated in parentheses.
The resulting dispersions of the calibrant masses (full symbols) are respectively, 1.1 . 10- 6,
7.7.10- 6 and 3.5.10- 7 .
was necessary to add a systematic error of 1.2 . 10- 6. The reader is refered to [7]
for a discussion of the calibration procedure and detailed presentation of the results
which are summarized here.
The results of the ensemble of nuclides measured is shown in Figure 4 as the
absolute difference with respect to the 1995 atomic mass evaluation [8]. On the
left side of Figure 4 are several nuclides whose masses were measured using 23Na
as the reference isotope. The calibrant masses are reproduced fairly well (with an
aVI~rage accuracy of 8.10- 7 ) and new values with reduced errors were obtained for
two short-lived neon isotopes 25- 26 Ne.
In the center of Figure 4 are shown those nuclides measured with respect to
39K and/or 41 K. The larger error bars and dispersion in this case reflects the results
of the less than satisfactory calibration. Despite the accuracy of only 3.5 . 10- 6
achieved for 32Mg (also compounded by statistical error) it was still possible to
improve the mass value for this exotic nuclide.
But why not turn the isobaric problem around and use it to our advantage?
Provided we can identify them, isobars practically represent the closest thing to a
narrow doublet, free from the deleterious effects of a large voltage jump. Indeed,
the groups that are breaking the records of mass measurement precision (see the
papers of Rainville et al., and Van Dyck, Jr. et al., in this issue) all rely on such
doublets. There were three cases where we were able to identify isobaric pairs
for measurement, shown in Figure 4 (right), and obtain an excellent measurement
accuracy: 2 . 10- 7 .
RECENT RESULTS ON Ne AND Mg FROM MISTRAL 305
Table I. New. MrSTRAL mass measurements. In column 2: the mass (and error),
in column 3: the mass excess (and error) and in column 3: the absolute deviations
from the mass table [8]
NannSO
(reaction)
0rr91
(SPEG) MISTRAL99
69%
Figure 5. Absolute mass difference of MISTRAL results compared to the other results in [8]
used to derive the recommended mass for 26Ne (left) and 25Ne (right). The influence of the
MISTRAL values are also given.
The final derived mass values for the three cases we were able to improve are
listed in Table I along with their uncertainty and deviation from the mass table
value [7].
An evaluation was performed for the results in Table 1. In the case of the two
neon isotopes, two values constituted the recommended mass in the table. For 26Ne,
the reaction 26Mg(n-, n+) 26 Ne was studied at LAMPF [9] and a direct measure-
ment was made using SPEG at GANIL [10]. These values are compared to the
MISTRAL value in Figure 5 (left) and can be seen to be perfectly compatible. The
MISTRAL value now accounts for 75% of the weight.
The two data comprising the 25Ne mass, shown in Figure 5 (right), come from
reactions: 26MgCLi,8B)25Ne [11] and 26Mg(\3C, 14 0)25Ne [12]. Though the latter
result has a smaller error bar due to the higher reaction cross-section and 26Mg tar-
get enrichment, the MISTRAL value is closer to the former. However the new ad-
justed mass from the evaluation is still compatible with the old. Note the MISTRAL
result carries almost 70% of the weight.
The result for 32Mg is shown in Figure 6, again compared with the individual
data that make up the recommended mass. All these data, from direct measure-
306 D. LUNNEY ET AL.
~ 400
-
TOFI86
.Y
1.0 300 SPEG87 TOFI91
<J') a
IIIIIIII
~ 200 SPEG91
~~
~ 0
100
-1 00
f-,:
(f) -200
~ -300
Figure 6. Absolute mass difference of MISTRAL result compared to the various results in [8]
used to derive the recommended mass for 32Mg. The later TOFI value is in fact an average of
two measurements as indicated in the figure. Inset: recorded mass peak with some 30 counts
and a resolution of 80,000.
ments with either TOFI [13, 14] or SPEG [10, 15], are compatible with each other
and with the MISTRAL value with the exception of the later TOFI value [14] which
is in fact an average of two runs, designated a and b in the figure. Figure 6 (inset)
shows the summed mass peak. The high resolving power of 80,000 was necessary
to combat the high density of isobaric contamination and is reflected by the poor
statistics, contributing an uncertainty of about 3 . 10-6.
Despite an uncharacteristically large error bar, the MISTRAL mass, represent-
ing a 268 ke V overbinding with respect to the AME95 value, indicates the defor-
mation effects thought to weaken the normally stabilizing shell closure at N = 20
are more pronounced. The paper of Sarazin et al. in this issue treats this physics
question.
4. Conclusion
The recent results show that in spite of high resolving power, the technique em-
ployed by MISTRAL of using harmonics is highly prone to the overwhelming
isobaric contamination brought by the plasma ion source. Another run is planned
using a laser ion source to avoid this.
On the other hand, there is absolutely no complication related to measuring the
masses of nuclides having very short half-lives and provided that the calibration is
' ''ne correctly, results of sufficient accuracy to improve mass data are obtained.
This is demonstrated in the cases of 25-26 Ne and 32Mg. The latter nuclide, for
which we find an overbinding compared to previous measurements, indicates a
RECENT RESULTS ON Ne AND Mg FROM MISTRAL 307
Acknowledgements
The results presented in this paper constitute part of the doctoral thesis of Mon-
sanglant [7]. The authors would like to acknowledge the expert technical assistance
of G. Le Scomet, M. Jacotin, J.-F. Kepinski, G. Conreur from CSNSM, Orsay,
M. Duma from lAP, Bucharest and G. Lebee from CERN. We thank Christoph
Scheidenberger from GSI for his assistance during one of the experiments. This
research project is funded by the IN2P3, France and some of the work at ISOLDE
supported by the European RTD Programme "Access to Large-Scale Facilities".
References
1. Smith, L. G., Phys. Rev. 111 (1958), 1606.
2. de Saint Simon, M. et aI., Phys. Scripta T59 (1995), 406.
3. Lunney, D. et ai., Hyp. Interact. 99 (1996), 105.
4. Coc, A. et ai., Nucl. Instrum. Methods A 271 (1988), 512.
5. de Saint Simon, M., Jacotin, M. and Lebee, G., Internal CSNSM Report 94-22 (1994) (available
from [16]).
6. Toader, C. F., Doctoral thesis, Universite de Paris Sud, Orsay, 1999 (available from [16]).
7. Monsanglant, c., Doctoral thesis, Universite de Paris Sud, Orsay, 2000 (available from [16]).
8. Audi, G. and Wapstra, A. H., Nuclear Phys. A 595 (1995), 409.
9. Nann, A. et aI., Phys. Lett. B 96 (1980),261.
10. Orr, N. A. etal., Phys. Lett. B 258 (1991), 29.
11. Wilcox, K. H. et aI., Phys. Rev. Lett. 30 (1973), 866.
12. Woods, C. L. et ai., Nuclear Phys. A 437 (1985), 454.
13. Vieira, D. J. et ai., Phys. Rev. Lett. 57 (1986), 3253.
14. Zhou, X. G. et ai., Phys. Lett. B 260 (1991), 285.
15. Gillibert, A. et ai., Phys. Lett. B 192 (1987), 39.
16. Atomic Mass Data Center (AMDC), http://csnwww . in2p3. fr /amdc/.
Hyperjine Interactions 132: 309-3l4,200l. 309
© 2001 Kluwer Academic Publishers.
Abstract. Recently, the applicability of Penning trap mass spectrometry has been extended to nu-
clides with a half-life of less than one second. The mass of 33 Ar (Tl /2 = 174 ms) was measured using
the ISOLTRAP spectrometer with an accuracy of 4.2 ke V. This measurement provided a stringent test
of the Isobaric Multiplet Mass Equation (IMME) at mass number A = 33 and isospin T = 3/2. The
fast measurement cycle that shows the way to other measurements of very-short-lived nuclides is
presented. Furthermore, the results of the IMME test are displayed.
Key words: atomic masses, ion trap, on-line mass spectrometry, radioactive ion beams, 33 Ar.
1. Introduction
Mass measurements of short-lived nuclides have been of great interest for many
years. This is because the atomic mass is the total Hamiltonian of the nucleus.
Generally, a large scale survey of the mass surface is needed in order to detect
trends in the nuclear binding energy or to fit the free parameters of nuclear models.
However, there are nuclides of special interest, spread all over the nuclear chart,
the mass of which is needed to an extraordinarily high accuracy for studies such as
fundamental tests. Some of these nuclides have very short half-lives.
One prominent example is the mass of 32 Ar. This mass plays an important role
for the result of the recently performed ,B-neutrino correlation experiment using
the ,B-delayed proton emission of 32 Ar [1]. The significance of the limit set by this
* Corresponding author.
310 F. HERFURTH ET AL.
2. Experimental setup
The ISOLTRAP mass spectrometer is installed at the online mass separator ISOL-
DE/CERN [2]. It consists of a linear radio-frequency ion trap and two Penning
traps. The continuous beam delivered from ISOLDE is accumulated, cooled and
bunched in the radio-frequency quadrupole (RFQ) ion beam buncher [3]. The main
task of this device is to transform the 60 keV continuous ISOLDE beam into
low-energy (2-3 keV) low-emittance (~107T mmmrad) ion bunches that can be
efficiently delivered to the first Penning trap. There the ion cloud is purified and
again formed into a cold bunch [4]. A mass selective buffer gas cooling technique is
employed that allows this trap to be operated as an isobar separator with a resolving
power of up to R R::! 105 for ions with mass number A = 100 [4]. The prepared ion
bunch is then delivered to the second trap, which is the high-precision trap used for
the mass measurements of the ions [5]. The mass measurement is carried out via a
determination of the cyclotron frequency Vc = (q / (27T m)) B of an ion with mass m
and charge q in a magnetic field of known strength B (determined by a reference
mass). For a detailed description see [5, 6].
Important from the point of efficiency is, that the decay losses during one mea-
surement cycle are kept as low as possible. Therefore, the total time for the mea-
surement of one frequency step must not exceed the half-life of the nuclide by
much. During this time the ion preparation (accumulation, cooling, bunching, iso-
bar separation) and the actual cyclotron frequency determination have to take
place.
ISOLTRAP MASS MEASUREMENTS 311
3. Mass measurements of 33 Ar
i axial cooling
i magnetron excitation
h ~===~----------------
I cyclotron excitation ~===::::;-I ------------
i radial cooling
!
n
. ~.- - - - - - - - - - -
I ejection ,
r!m_a~g~ne_tr_o_n_ex_c_irn_ti_on_ _ _ _ _ _ _ _ _ _ _~~~========~
! cyclotron excirntion ,-- L
i ejection, srnrt TOF acquisition I
~/rl"-'--'-'--'-'--'-'--'I-'-~-'-~-'I-'-~-'--'--'
o 70 100 150 200 250
t time [ms]
proton impact on target
Figure 1. A measurement sequence for very-short-lived nuclides as. for instance, 33 Ar.
312 F. HERFURTH ET AL.
Table [. Equations for input into a least square adjustment as [7] and mass excess values (ME) as
determined in this work. The last column gives literature values from [7]
The frequency ratio vcC 6 Ar)/vc C3 Ar) can be determined with a relative accuracy
of 9 . 10- 8 , governed by statistics and resolving power.
In total, each cycle of the measurement of the cyclotron frequency took 175 ms.
The overall efficiency of ISOLTRAP was about 10-4 given by the ratio of the num-
ber of 33 Ar ions detected after the second trap using a multi channel plate (MCP)
detector, and those in the ISOLDE beam. This efficiency includes decay losses,
detection efficiency and transfer losses. The main losses occur while injecting into
the first Penning trap where the ion optics has still to be improved. On the average,
less than one 33 Ar ion was stored in the precision trap in each cycle.
Adding quadratically a conservative estimate for the systematic uncertainty of
om/m = 1 . 10- 7 , that covers unobserved magnetic field changes [10], the mass
of 33 Ar is determined with a precision of om/m = 1.3 . 10- 7 . This corresponds
to a mass uncertainty of om = 4.2 ke V and represents a seven fold improvement
of the error as compared to the previous measurement [11]. In addition, the longer-
lived isotopes 34.42.43 Ar have been measured with a mass uncertainty om = 3.5, 5.8
and 5.3 keY, respectively. The results are summarised in Table I. Also given are the
equations that are the input for the least square adjustment of all atomic masses [7].
Including the ISOLTRAP data all members of the A = 33, T = 3/2 isospin
quartet are very well known now. This allows a more stringent test of the quadratic
IMME than before (for a detailed discussions on all input data and the IMME test
see [12]). Until now the test of the quadratic IMME showed consistency for the
"ground state quartet" but was limited by the 30 ke V [11] error on the 33 Ar ground
state mass value. Including the new mass value the data is not very well described
by the quadratic IMME relation M(Tz ) = a + bTz + cTi- M is the mass of the
level with isospin T and projection Tz . The fit of a quadratic function to the data
results X2 = 10.6 that has a probability of only 0.1 % to occur if the data was
described by a quadratic function. Allowing for an additional cubic term in dTi
IMME the data yields dC 3 Ar)= -2.95 ± 0.90 keY, which is not consistent with
zero.
In Figure 2 the c/-coefficients for all completely measured ground state quartets
are plotted together with the significance of their deviations from zero. All together
there are now four ground-state quartets with a d-coefficient farther than two stan-
ISOLTRAP MASS MEASUREMENTS 313
40
20
;; 11:
0
6 11:
"" -20
·40
o 0
3 ............................................................................. .
o
{ 2 ........... ~ ....... o ...................................................... .
°0 °
I ··0······························0···················· ...................... .
o 00
o
to 20 30 40 50
A
Figure 2. The d-coefficients of all 18 completely measured ground-state quartets (upper
graph) and the significance of their deviation from zero S = Idl/Ud (lower graph). The data
for the A = 33 quartet include the ISOLTRAP data, the others are from [13].
dard deviations from zero and thus, in significant disagreement with the generally
accepted quadratic IMME.
There has been great effort to explain a nonzero d-coefficient, triggered by the
A = 9 ground state quartet. The significant result for the A = 9 quartet has partly
been explained by isospin mixing effects in the Tz = -1/2 and + I /2 members
[14, 15], by the expansion of the least bound proton orbit in 9C, as well as by
charge-dependent nuclear forces [16]. However, the situation rests unresolved.
In conclusion, even though it is not clear which effect causes the breakdown
of IMME for the described A = 33, T = 3/2 quartets, it is necessary to be very
careful in deriving high-accuracy masses of proton-rich configurations from the
IMME.
4. Summary
Due to the improved efficiency and a fast measurement cycle, it is now possible to
measure nuclides with half-lives well below one second using the ISOLTRAP mass
spectrometer. In addition to the measurement of 33 Ar, which is with
TI/2 = 173 ms the shortest-lived nuclide ever measured in a Penning trap, 34,42,43 Ar
314 F. HERFURTH ET AL.
have been measured. In the case of argon further experiments are planned to im-
prove the accuracy, especially for 34 Ar, and to extend the measurements to 32 Ar.
Even more challenging is the planned mass measurement on 74Rb, due to the
even shorter half-life of only 65 ms. This measurement will be accomplished by
a high-accuracy (om/m < 1 . 10-7) measurement of 74Kr, which will allow to
determine the Q-value of the 74Rb ,B-decay to an accuracy of about 10 keY.
Acknowledgements
This work was supported by the European Commision within the EUROTRAPS
network under contract number ERBFMRXCT97-0144, within the RTD project
EXOTRAPS under contract number ERBFMGCET980099 and by NSERC of Ca-
nada.
References
l. Adelberger, E. G. et ai., Phys. Rev. Lett. 83 (1999), 1299 and 31Ol.
2. Kugler, E. et ai., Nuci. Instrum. Methods B 70 (1992), 41-49.
3. Herfurth, F. et ai., A linear radiofrequency ion trap for accumulation, bunching, and emittance
improvement of radioactive ion beams, Nucl. Instrum. Methods A (2000), accepted; also CERN
preprint: CERN-EP/2000-062.
4. Raimbault-Hartmann, H. et ai., Nucl. Instrum. Methods B 126 (1997),378.
5. Bollen, G. et ai., Nucl. Instrum. Methods A 368 (1996), 675-697.
6. Bollen, G. et ai., this issue, 215.
7. Audi, G. and Wapstra, A. H., Nuclear Phys. A 595 (1995), l.
8. Carlberg, C. et ai., 1999.
9. Mohr, P. J. and Taylor, B. N., Rev. Modern Phys. 72 (2000), 351; 1998 CODATA values.
10. Beck, D. et ai., Nucl. Instrum. Methods B 126 (1997),374.
II. Nann, H. et ai., Phys. Rev. C 9 (1974), 1848.
12. Herfurth, F. et ai., submitted to Phys. Rev. Lett.; also CERN preprint: CERN-EP/2000-142.
13. Britz, J., Pape, A. and Antony, M. S., At. Data Nucl. Data Tables 69 (1998), 125-159.
14. Henley, E. M. and Lacy, C. E., Phys. Rev. 184 (1969), 1228.
15. Janecke, J., Nuclear Phys. A 128 (1969), 632.
16. Bertsch, G. and Kahana, S., Phys. Lett. B 33 (1970),193.
Hyperjine Interactions 132: 315-322,2001. 315
© 2001 Kluwer Academic Publishers,
Abstract. Mass measurements of the N = Z nuclei 80Zr, 76Sr, 68Se were performed for the first
time and a new measurement was obtained for 80y, using the second cyclotron CSS2 of GANIL
as a high-resolution spectrometer. Ions around N = Z were produced by fusion-evaporation in the
inverse 58Ni (4.32 MeV A) + 24Mg and I2C reactions. New masses were measured by a time-of-
flight method, with a precision of 2 . 10-6 , by using well-known masses as references. Study of
the double binding energy difference 8Vnp is then performed leading to a strong N = Z Wigner
effect around N = Z = 40. Knowledge of new masses in this region also plays a crucial role in the
modelling of the astrophysical rp process.
1. Introduction
Properties of nuclear structure can be investigated by measuring fundamental quan-
tities for nuclei far from stability, in particular masses. For exotic nuclei close to the
N = Z line, masses are mostly known until A -::::: 62 and partially until A -::::: 100.
Going towards the proton drip line and measuring binding energies of A = [68,80]
nuclei provide a good test of neutron-proton pairing role in N = Z nuclei. Even-
even N = Z nuclei studies (in particular, mass measurements) are also motivated
by their contribution in the astrophysical rapid proton capture process, rp process,
as they constitute waiting points of this process.
2. Experimental technique
Mass measurements have been performed with a time-of-flight method using the
second cyclotron of GANIL, CSS2, as a mass spectrometer of high resolution [1].
316 A. S. LALLEMAN ET AL.
This method has already been used successfully for the mass measurement of
doubly magic IOOSn [2]. Residues are produced by fusion-evaporation reactions
between a primary beam delivered by the first cyclotron (CSS I) of GANIL, and a
target placed between the two cyclotrons.
Secondary ions produced in the target are injected in CSS2 tuned with the mag-
netic rigidity: BpCSS2 = (ymO/q)V2 for ions with a relativistic mass ymo, a charge
state q and an injection velocity V2. Because the two cyclotrons are coupled in
frequency, V2 is related to the primary beam velocity VI by: V2 = (2/5)(h l /h 2)vI
where hi and h2 (integers) are the harmonics of the two cyclotrons respectively.
Because the mass acceptance of the cyclotron is (8m/q)/(m/q) :S; a few 10-4, dif-
ferent nucleides can be simultaneously accelerated in CSS2, making it possible to
accelerate known and unknown masses at the same time. At the final radius, where
the time-of-flight is about 70 I1S, ions are detected in a ilE(30 I1m)-E(300 11m)
Silicon Telescope located on a radially moving rod inside a magnetic cavity of
CSS2. In the isochronous mode of acceleration, the fundamental equation implies
directly that (8m/q)/(m/q) = 8t/t if, and only if, the number of turns inside
the cyclotron are the same. The unknown mass can be detennined from a well-
known reference mass after the same number of turns inside CSS2 by measuring
their time-of-flight (or phase) relative to the high frequency (HF) signal of CSS2
(7.6 MHz in this experiment).
3. Experimental results
:;-.,
80s r
• .' . ' t.;. •• "
."
c • ' Jo. • f~.·.> .~ :--~
400
' .. ... •.• )
'F •• • '
r ..,:
.~i!J
. '.
L 'I
$' , '
~. "
:.~ -t;.', ~..... :"~ ~~... ..,.:,
\
·
.... ,~,: ...... ':>C. -"k: t,
.-;-.'~)- ~~ .,~ I: .':.~ ,:-. ,.
7~ ' ~~..:.t~. :.(;!~ : '. 'J
380 ;;~:~$~-,-:~6:Rb ~
· St.. ,."I.· :;.•
· 72 .
~~-S:tr-7.X ·:
Kr ": ." --:~ lr-:-F : t;:r•.·,
360 -
h
.. .. , .. ;...}
. . . ....... •
.... . "~")-.
..:-.-.; ,.
(oI~
.~ ,fO.
., .,
- .... .'~~':. . -
320 '
'
'."
300
-
280 -
65 70 75
.
~1..L11,1'11111111
80 85 90
I
95
.'
Phase (ns)
(a)
90 r 3 4
80 t- 80
70 ~ Sr
60 t
c'" 2
:::J
0
u
50!
40
30
20 " 1
1O =-
0
-3000 -2000 - 1000 0 1000 2000 ns
(b)
Figure I. (a) Energy-phase spectra obtained with target of 12C + 24Mg at the final radius of
CSS2. (b) Intercepted turns for 80Sr obtained with a larger base time (obtained with start given
by the ion in telescope and stop signal given by the cut of beam every of 5.395 J-Ls). Time scale
(x-axis) corresponds to an arbitrary reference. For these nuclei 8 turns are intercepted.
The energy-phase spectra obtained with the 12C + 24Mg is presented in Fig-
ure lea) where well-separated ions can be seen. In the mass region A = 80 the
relative mass differences between isobars are less than 2 . 10-4, so with the set up
of CSS2 we obtain many exotic residues produced by the different reactions.
The separation in A / q of the ions results from their different total kinetic energy
(E + f}.E) at the end of acceleration and the difference in phase is (for the same
number of turns in the cyclotron) directly related to their mass differences. The
318 A. S. LALLEMAN ET AL.
width in phase is not similar for each of the detected nuclei. This is due to the
fact that the size of the detection system is larger than the separation between the
turns. The detector intercepted part of several turns, depending on the particular
trajectories of the differents species accelerated, then the distribution in phase for
each nuclei is more or less different. On the other hand, some tail at higher energy
is observed in phase with each residue due to the pulse height defect in the E
detector.
A use of two cuts of primary beam (at the beam source) has been done in order
to have beam every 5.395 !-lS and 6.710 !-lS instead of the HF period (131.58 ns).
This larger base time (with a "stop" every 5.395 !-lS or 6.710 !-ls) allows to observe
the last turns intercepted for each species (example given Figure 1(b) for 80Sr), with
the time revolution period equal to 4 . 131.58 ns = 526.32 ns. The poor statistics
and the insufficient resolution in time make it impossible to determine tum by tum
the phase difference between the observed nuclei. However, Figure 1(b) shows that
the distribution is Gaussian centered on the more intercepted tum.
Consequently in Figure l(a), with a base time of 1 HF period, each of the nuclei
are fitted with a Gaussian, with the centroid corresponding to the average phase
equivalent to an average number of turns. After analysis of each nucleus, in the
large base time, a correction of 11 laps for 76Kr, 80Sr and 68 As has been performed
to compare their phases with the other ones with the same average number of
laps [3]. The resulting calibration is presented in Figure 2(a) where 4 reference
nuclei are used: 76Kr, 80Sr, 68 As and 76Rb. For the known ones, mass excesses were
taken from Audi and Wapstra Table [4]. The masses of 80Zr, 76Sr, 68Se and 80y are
determined by extrapolation from their phase difference measurement (8t related
to 76Rb), represented in Figure 2(b). This calibration is then checked with two other
nuclei: 80Rb and 72Kr. A second calibration has been made with exactly the same
method, after a small change of the magnetic field.
Final results from the 2 calibrations are given in Table I. The error bars take into
account both the statistic error (j / -IN (low because of the thousands of counts N
for A = 80,76 and 68 ions), where (j is the width of the Gaussian, and a systematic
error (dominant) of 0.1 ns (corresponding to the maximum 8t for one tum) due to
the error on the average number of intercepted turns between the different nuclei
(±1 tum).
The measured mass excess of 80Zr and 76Sr is in agreement with both macro-
scopic-microscopic predictions (from FRDM [5]) and estimations from systematic
trends (from [4]). Nucleus 68Se is less bound than predicted by both approaches
(given in Table I). This could be due to the existence of an isomeric state with
a long half-life (70 !-lS minimum due to the total time-of-flight) which would be
populated in the 58Ni + 12C reaction. However, such a state has still not been
observed experimentally. Moreover, we note that 69Br, which is one proton more
that 68Se, is presumably unbound [6, 7]. Three previous mass measurements of 80 y
(all in agreement with each other) were made from beta end-point energy [8-10].
A more recent experiment, performed with a time-of-flight method using a cy-
MASS MEASUREMENTS OF EXOTIC NUCLEI 319
~
Q,l 2 Reference ions
~ u~ed forthe cal ibration
~ 0
-2
-4 ~~~-L~__~~~~~~__~~
-1.S 10-4 -110- 4 -S 10-5
Om/q/(m/q)(Ref
10 " ~ ' 1 '
- ----------------------- ~ ---
1
/ b)
1 /
=5 I
en
fi g a)
t/-r
;1
-.
.Q
I
~
-e ref ~
....
t-- ••
Q,l
0 ,
:
""
80 y _ ~Oy (prev.)
~ , 6 Sc - "
....
'--'
GIO
-s
80Zr
-10 L........~-'-~~~--'-~~L.....~...L....~,
-110- 4 -S 10- 5 0 510- 5 110- 4
Om/q/(m/q)(Ref : 76Rb )
Figure 2. Calibration 8t vs. (8m/q)/(m/q) (a) for nuclei of reference and checking points,
and (b) for all nuclei (with the new masses determined by extrapolation).
Table I. Experimental mass excesses measured in this work (exp) in comparison with FROM
predictions [5J and AME95 estimations [4]. Are also given the previous measurements for Say
clotron at SARA (Grenoble, France) led to another result [11], different by 2 MeV
from the others. The new mass measurement presented in this work is in between
the previous results.
0
.- -1
;>
Q,I
~ -2
"-'
Q.. -3
=
;> -4
r,Q
-5
-6
-7
-8
0 20 40 60 80 100
Z
Figure 3. Residual interaction between the last proton and the last neutron 8Vnp (Z, N) as a
function of Z for even-even Z and N (with a line for N = Z) obtained with known masses
and our new experimental ones (denote as "Exp.").
MASS MEASUREMENTS OF EXOTIC NUCLEI 321
5. Rp process modelling
The astrophysical rp process is characterized by a sequence of p and a captures
competiting with f3+ IEC decays in X-ray bursts scenarios [14]. In particular, even-
even N = Z = 34 to 42 nuclei are waiting points of this process because of
their long half-life (a few seconds) and because proton capture is forbidden (from
predictions), which inhibits its flow. Modelling of this process depends crucially
on nuclear physics data (masses, f3 decay) and particularly on the mass differences
(Qp, Q2p, QjJ, QQ,) around these waiting-point nuclei.
The resulting abundances of the N = Z nuclei ~~Se, ~~Kr, ~~Sr and ~gZr have
been calculated in collaboration with M. Wiescher and A. Aprahamian from Notre-
Dame University, Indiana (USA) and are presented in [3]. Locally there is a one
order of magnitude difference between the abundances calculated for 68Se using
either the measured or the predicted mass of 68Se.
In fact, predictions from Correlation Scheme [15] for both masses of 68Se and
69Br for the proton capture Qp(68Se(p,y)69Br) were used in the first case, and
the experimental mass measurement for 68Se and prediction for 69Br [15] for the
second. As a result the Qp was negative (proton capture forbidden) in the first case
and positive (proton capture authorized) in the second case. This example shows
the importance of measuring masses around waiting points in rp process modelling.
6. Conclusion
Mass measurements of the N = Z nuclei 80Zr, 76Sr and 68Se and 80y have been
performed with a precision of 2 . 10-6. A new experiment is programmed to dispel
the disagreement between our mass measurement for 80y and previous ones and to
check the difference of 2 MeV between the measured and predicted mass of 68Se.
The systematics on 8 Vnp , extended to higher masses by our measurement, shows
peculiar effects as N = Z = 40 is approached. This systematics should be checked
and extended further. Our measurements have yielded the mass of several waiting
point nuclei and turned out to be crucial in rp process abundance calculation.
Acknowledgements
We would like to thank M. Wiescher and A. Aprahamian from Notre-Dame Univer-
sity, Indiana, USA, for performing calculations of nuclei abundances concerning rp
process modelling.
References
1. Auger, G. et ai., Nucl. Instrum. Methods A 350 (1994), 235.
2. Chartier, M. et ai., Phys. Rev. Lett. 77 (1996), 2400.
3. Lalleman, A. S., PhD report GANIL TOO 02.
4. Audi, G. and Wapstra, A. H. et ai., Nuclear Phys. A 624 (1997), 1.
322 A. S. LALLEMAN ET AL.
Abstract. Information on atomic masses and nuclear binding energies can be extracted from (n, p)
and (n, ex) reactions with thermal and resonance neutrons. This is illustrated by means of several
selected examples.
Key words: atomic mass, nuclear binding energy, neutron reactions, Q-values.
1. Introduction
Valuable information on atomic masses and nuclear binding energies can be ex-
tracted from (n, p) and (n, a) reactions with thermal as well as with resonance
neutrons.
Thermal neutron-induced (n, p) and (n, a) spectroscopy delivers values for the
grounds tate transition reaction energies Q(n, p) and Q(n, a), with an accuracy of
a few ke V. These quantities are directly linked to the atomic masses of the initial
and final nuclei M(A, Z) and M(A, Z - 1) and M(A - 3, Z - 2), respectively, via
the relations:
Since the neutron mass M(n) and the masses of the iH and 4He atoms are very
well known [1], an accurate determination of Q provides an accurate value for the
mass difference between the initial and the final nucleus, which is especially inter-
esting when one of them is short-living. This will be illustrated by a few selected
examples.
The second part of the paper is devoted to resonance spectroscopy. Indeed, by
combining resonances observed in (n, p) reactions with the corresponding reso-
nances in the inverse (p, n) reactions, the difference Sn - Sp between the neutron
and the proton binding energy can be determined. The principle is illustrated in
Figure 1 by means of the 4oK(n, p) 40 Ar and 40 Ar(p, n) 40K reactions. A few cases
will be discussed where a comparable or slightly better accuracy is obtained as
compared to the values reported by Audi and Wapstra [1]. Similarly, a combination
324 C. WAGEMANS ET AL.
40K + n~41K*~4°Ar + p
Sn - Sp = (Ep - En) x !~
E p=3 06 keY
(1ab. energy!)
E =2346.9 keY
Gab. energyl)
3+ /7///$/7#//7//7
41K
Figure 1. Detennination of Sn - Sp from a 41 K level excited via the 40K(n, p) 40 Ar and the
40 Ar(p, n) 40K reactions. The resonance energies are taken from experimental data and have
to be transfonned in center of mass units.
of resonances in (n, a) and (a, n) reactions yields accurate Q(n, a) values. Also
here, examples will be given.
beams. A detection chamber was installed at the end of the 87 m curved neutron
guide H22, delivering a flux of 3.5 x 108 neutrons/cm2 s at the sample position.
Moreover, the ratio of slow neutrons to epithermal and fast neutrons is 106 and the
direct y -ray flux from the reactor is reduced by a factor of about 106.
The charged particles emitted were detected with suited high-resolution Si sur-
face barrier detectors. Great care was given to the energy calibration of the de-
tection chain, for which a variety of well-known reactions such as 6Li(n, a)t,
IOB(n, a) 7Li and 143Nd(n, a) 140Ce were used.
A good example of the value of the method is the (nth, a) reaction on 153Gd
(Tl/2 = 241.6 d). From the energy of the 153Gd(nth, ao) transition, Q(n, a) =
(9.79 ± 0.03) MeV was determined [2], which is 0.21 MeV higher than the value
calculated from the 1977 atomic mass tables of Wapstra and Audi. In the 1983
edition [3], however, they concluded (based on neutron capture data including
[2]) that the 1977 data were wrong, since they were based on erroneous electron
capture decay measurements. The 1993 edition [1] reports Q(n, a) = (9.8150 ±
0.0012) MeV, in agreement with the above given experimental result.
This example illustrates that a-spectroscopy permits a simple verification of
other methods.
The experimental value Q(n, a) = (9.79 ± 0.03) MeV combined with rela-
tion (2) results in (3.004448 ± 0.00003) u for the 153Gd - 150Sm mass difference.
So with a value of 149.917272 u for the 150Sm mass [1] we obtain a 153Gd mass
350
a1 99
300 t- Ru(n,a)
,I
250 t- i I
I
......
(/) 200 - II i
c
::::I ) \
0 150 ~
I
()
100 t-
.\ I
50 I- ao
E nerg y [Me V]
Figure 2. Pulse-height spectrum of the 99Ru(n, ao) 96Mo and 99Ru(n, a]) 96Mo reaction
with thermal neutrons.
326 C. WAGEMANS ET AL.
l.O 2286.19
4
1
1
1
1
1
1
1
3+ / / //);r////)r/T/T/7
4 1K
Figure 3. Combination of resonances in the 40 K(n, p) 40 Ar and 40 Ar(p, n) 40K reactions.
En [keY] Ep [keY]
8.37 ± 0.04
>-....---~--1986.0 ± 0.3
7.90 ± 0.02
3.53 ± 0.]0 --+-,--............-lI,.---1982.1 ± 0.3
1.340 ± 0.003 --!-I-~~'t---1979 .9 ± OJ
3/2+/////////)W///////////
37Cl
Figure 4. Schematic representation of some common resonances fed in 36Cl +n and 36S +p.
Another example are the 178.3 and 253.0 ke V resonances in 170(n, a) 14C [12]
combined with the 2553 and 2642 keY resonances in 14C(a, n) 17 0 [14], yielding
Q(n, a) = (1817.2 ± 3.5) keY in perfect agreement with the evaluated value of
(1817.52 ± 0.21) keY [I]. A last example is 33S(n, a) 30Si [15] combined with
30Si(a, n) 33S [16] yielding Q(n, a) = (3496.9 ± 5.0) keY in agreement with the
DETERMINATION OF ATOMIC MASSES AND NUCLEAR BINDING ENERGIES 329
evaluated value of (3493.13 ± 0.23) keY [1]. Details ofthe resonances used in the
latter case are given in Table I, which illustrates the stability of Q(n, a) over the
resonances.
In both cases, the uncertainty is completely due to the uncertainty on the
a-energy in the inverse reaction.
4. Conclusion
In the present paper we have shown that valuable information on atomic masses
and nuclear binding energies can be extracted from neutron-induced reactions,
occasionally in combination with proton- or a-induced reactions leading to the
same compound nucleus.
References
1. Audi, G. and Wapstra, H., Nuclear Phys. A 565 (1993), 1.
2. Wagemans, c., Allaert, E., Barreau, G., Emsallem, A. and D'hondt, P., Nucl. Instrum. Methods
190 (1981), 167.
3. Wapstra, A., Audi, G. and Hoekstra, R., Nuclear Phys. A 432 (1985), 185.
4. Wagemans, C., Druyts, S. and Geltenbort, P., Phys. Rev. C 50 (1994), 487.
5. Giese, C. and Benson, J., Phys. Rev. 110 (1958), 712.
6. Goeminne, G., Wagemans, c., Wagemans, J., Serot, 0., Loiselet, M. and Gaelens, M., Nuclear
Phy.\'. A 678 (2000), 11.
7. Alderliesten, c., Aerts, P., Van Bijlest, H. and Van der Leun, C., Nuclear Phys. A 220 (1974),
284.
8. Weigmann, H., Wagemans, c., Emsallem, A. and Asghar, M., Nuclear Phys. A 368 (1981),
117.
9. Parks, P., Beard, P., Bilpuch, E. and Newson, H., Nuclear Phys. 86 (1966),504.
10. Bieber, R., Wagemans, c., Heyse, J., Balcaen, N., Barthelemy, R. and Van Gils, J., In: Proc. (~f
Internat. Symp. on Capture y-ray Spectroscopy and Related Topics, Budapest, 1996, p. 443.
11. Nooren, G. and Van der Leun, C., Nuclear Phys. A 423 (1984), 197.
12. Wagemans, J., Wagemans, c., Goeminne, G. and Geltenbort, P., Nuclear Phys. A 688 (2001),
490.
13. Skelton, R., Kavanagh, R. and Sargood, D., Phys. Rev. C 35 (1987), 45.
14. Sanders, R., Phys. Rev. 104 (1956), 1434.
15. Wagemans, c., Weigmann, H. and Barthelemy, R., Nuclear Phys. A 469 (1987), 497.
16. Wiechers, G., McMurray, W. and Van Heerden, I., Nuclear Phys. A 92 (1967),175.
Hyperjine Interactions 132: 331-335,2001. 331
© 2001 Kluwer Academic Publishers.
Abstract. The masses of the xenon isotopes with 114 ,:;: A ,:;: 123 were directly measured for the
first time. The experiments were carried out at the ISOLTRAP triple trap spectrometer at the on-
line mass separator ISOLDE/CERN. A mass resolving power of the Penning trap spectrometer of
m / /).m "'" 500000 was chosen and an accuracy of 8m "'" 12 ke V for all investigated Xe isotopes was
achieved. An atomic mass evaluation was performed and the results of this adjustment are compared
with theoretical predictions. The new results for the xenon isotopes and their effects on neighboring
nuclides are discussed within the two-neutron separation energy picture.
Key words: Penning trap, mass spectroscopy, xenon, nuclear binding energy, radioactive isotopes,
atomic masses.
1. Introduction
Penning traps are devices especially suited for high-precision experiments such as
g-factor determination [1, 2], test of CPT-invariance [3] or atomic mass measure-
ments. Recently the mass of stable 133Cs was determined on a sub-ppb level [4]. At
ISOLTRAP such a device is employed for the measurements of masses of unstable
nuclei. So far, high-precision measurements on nearly 200 radioactive nuclides
have been performed with an accuracy of typically 10-1 .
* Corresponding author.
** Present address: GSI, Darmstadt.
t Present address: NSCLIMSU, East Lansing.
332 J. DILLING ET AL.
2. Setup
The ISOLTRAP mass spectrometer is installed at the on-line facility
ISOLDE/CERN in Geneva. It consists of a linear radio frequency quadrupole
(RFQ) trap [5] and two Penning traps [6, 7]. The quasi-continuous ion beam de-
livered by ISOLDE with typically 30 or 60 keV is injected into the linear RFQ
trap filled with He buffer gas. Here, the beam is electrostatically retarded, cooled
by buffer gas collisions, bunched and extracted at low energy. The ions at typical
2.5 keV transport energy are transferred to the first Penning trap where mass se-
lective buffer gas cooling is applied to further cool and isobarically clean the ion
sample. Subsequently these ions with a charge-to-mass ratio q / m are delivered to
the precision Penning trap where their cyclotron frequency Vc = (q / (m . 2rr» . B is
measured. The cyclotron frequency is determined by exciting the ion motion with a
radiofrequency field for a period TRF and by employing a time-of-flight technique.
The magnetic field B is calibrated via a cyclotron frequency measurement of stable
133Cs, delivered from a test ion source.
3. Measurements
In the case of the here presented Xe mass measurements a radio frequency exci-
tation time of the ions of TRF = 0.9 s was chosen resulting in a resolving power
of R = 5 . 105 . With typically 6000 detected ions for each investigated nuclide, a
statistical accuracy in the mass determination of 8m/m = 3.10- 8 was achieved. As
a conservative estimate of possible systematic errors an additional error of 1 . 10- 7
is added quadratically yielding a total uncertainty for ISOLTRAP mass values of
all xenon isotopes of 8m ~ 12 keY.
4. Results
Mass measurements were performed for the Xe isotopes with mass number
114 ~ A ~ 124 and A = 130. Figure l(a) shows the difference between mass
values from the Atomic Mass Evaluation 1995 (AME 95) and an evaluation in-
cluding the new ISOLTRAP data. The masses of 114Xe, 1l5Xe, and 116Xe were
previously unknown. In the case of 118Xe, the experimental mass uncertainty could
be reduced by a factor of 100 whereas the uncertainties of the masses of 117Xe
and 119Xe are reduced by one order of magnitude. The masses of 124Xe and 130Xe
are known from literature with high accuracy. The deviation of those values from
the ISOLTRAP data is 8me 24 Xe) = 1 ± 12 keY and 8m(130Xe) = 3 ± 13 keY
demonstrating the reliability of the ISOLTRAP data.
In the case of 120Xe, a drastic discrepancy by 8 standard deviations between
the literature value and the ISOLTRAP result is observed. Other notable deviations
for l2lXe (3a), 122Xe (2a), 118Xe (la) and 119Xe (la) are reported. The discrep-
ancies were mainly due to incorrect ,B-endpoint measurements or underestimation
of errors for those. The high-accuracy ISOLTRAP data influence via the manifold
MASS MEASUREMENTS OF 114-124, 130XE AT ISOLTRAP 333
200 - - Incl, ISOLlRAP
# ¢ 1WE95
100 #
# I
'0
I
-400
I
112 114 116 118 120 122 124 126 128 130 132
MlsstbTber
(a)
10000
Ba (Z=56)
9000
Cs(Z=65)
8000
Xe (Z=54)
7000
I (Z=53) '.
6000
Te (Z=52)
~ 5000
4000
,J
3000
2000
1000
0
110 112 114 116 118 120 122 124 126
MillstUmer
(b)
Figure J, (a) Difference between mass values from the Atomic Mass Evaluation 1995 (AME
95) [8] (data points with error bars) and an evaluation including the ISOLTRAP data (zero line
with error band), For isotopes marked with # masses are estimated from the extrapolation of
systematic trends [8], (b) Two-neutron separation energy as a function of mass number, Filled
circles show the new values, open circles data from AME 95.
correlation the mass values of 23 other nuclides. This impact can be seen most
directly by a plot of the two-neutron separation energies in the xenon region as a
function of mass number (Figure I (b)). Open circles indicate literature values taken
from the previously published atomic mass evaluation AME 95 [8]. The values
determined or significantly changed by the present work are shown as full dots.
Generally, a very smooth behaviour of the two-neutron separation energies (es-
pecially for the nuclides with even proton number) is found in this region of the
chart of nuclides, indicating the absence of any drastic nuclear structure effects
in these neutron mid-shell nuclides. The only stronger irregularities observed at
A = 116 for cesium and at A = 118 for iodine might be due to the rather large
uncertainties for the masses of 114CS and 116CS, respectively, or a nuclear structure
effect for the case of 1181.
334 J. DILLING ET AL.
A comparison with three selected nuclear mass models is carried out within this
work. More details about this comparison can be found in [9], and an overview
on those formulas is given in [10]. The graphs in Figure 2 show the difference
between experimental and theoretical values. The model of Pearson et al. [11],
based upon an extended Thomas-Fermi-Strutinski ansatz includes a Skyrme term
to describe the interaction between the nucleons. The contributions accounting for
the deformation effects seem to be overestimated in the region around A = 121,
where no changes in the experimental deformation values are observed. A rather
smooth development would be expected. The root-mean-squares (RMS) deviation
for the xenon isotopes in the mass region 114 :S; A :S; 136 is ~m (RMS) = 301 ke V.
The mass values by Dufio and Zuker [12] are derived by a microscopic mass
model. The odd-even staggering is clearly over-estimated in this mass formula.
The trend towards the closed neutron shell N = 82 (A = 136) is well covered.
The deviation found is ~m(RMS) = 413 keY.
For the macroscopic-microscopic model of Moller and Nix [13] the odd--even
staggering is also overestimated. The overall trend is best described by this model
where a smooth development towards the N = 82 shell is found. The RMS-
difference for this model is ~m(RMS) = 253 keY.
5' BOO
600
I -.-Pecnon I BOO
BOO
- . - Du!Ioaud<er
Q) 400
400
........
~
200 200
0 0
-200 -200
~ .<10) .<10)
fI\/
.aJI .aJI
:
.aJO .aJO
~. ·ICDJ
·ICDJ
·1200 ~--.-~~~--r-~----r~--'-'--' ·1200
110 115 120 125 130 135 140 110 115 120 125 130 135 140
MassNuTber MassNuTber
5': I- -Milier I
Q) 400
~
........
O+-________________
200
~L------
: ·200
~-400
.aJI
: .aJO
~ ·ICDJ
·1200 +-~-.-__r_~~~__,_ _~~~
110 115 120 125 130 135
MassNuTber
Figure 2. Comparison of the experimental mass values with the prediction of selected mass models
for xenon isotopes with 114 :( A :( 136. The experimental masses are given as a zero line with a
practically invisible error band.
MASS MEASUREMENTS OF 114--124, 130XE AT ISOLTRAP 335
References
L Van Dyck, R S, et aI" Phys. Rev, Lett, 59 (1987), 26,
2, Hennanspahn, N, et aI" Phys, Rev. Lett, 84 (2000), 427,
3, Gabrielse, G. et aI" Phys. Rev, Lett. 74 (1995),3544,
4, Bradley, M, P, et al., Phys, Rev, Lett, 83 (1999), 4510.
5, Herfurth, F. et aI" Nucl. Instrum. Methods (in press),
6, Bollen, G, et aI" Nuc!. Instrum, Methods A 368 (1996), 675,
7. Raimbault-Hartmann, H. et aI" Nuc!. Instrum, Methods B 126 (1997), 378,
8, Audi, G. and Wapstra, A. H., Nuclear Phys, A 595 (1995), 409,
9, Dilling, J, et aI" (in preparation for European J, Phys,),
10, Patyk, Z, et aI" Phys, Rev, C 59 (1999), 704,
1L Aboussir, Y. et ai" Nuclear Phys, A 549 (1992), 155,
12, Dufio, J, and Zuker, A. P, Phys, Rev, C 52 (1995), R23,
13, Moller, P et aI" At, Data Nuc!. Data Tables 59 (1995), 185,
Hyperjine Interactions 132: 337-340, 2001. 337
© 2001 Kluwer Academic Publishers.
Abstract. The recent implementation of gas-filled radiofrequency traps for efficient ion beam bunch-
ing extended the applicability of the Penning trap mass spectrometer ISOLTRAP/CERN to non-
surface ionizable species. In a first series of successful runs the masses of 182- 197 Hg, 196, 198Pb,
197Bi, 198po and 203 At have been determined with an accuracy of 1 . 10-7. In order to unambigu-
ously determine the ground state mass the ground and isomeric states of 185,187,191,193, 197Hg were
separated applying a resolving power of up to 3.7 . 106. First experimental values for the isomeric
excitation energy of 187, 191 Hg were obtained.
Key words: radioactive nuclides, atomic masses, Penning trap, mass spectrometry.
1. Introduction
* Corresponding author.
338 S. SCHWARZ ET AL.
2. Experimental setup
Figure 1 shows a schematic layout of the ISOLTRAP mass spectrometer. ISOL-
TRAP consists of three main components, a gas-filled radiofrequency quadrupole
(RFQ) ion trap and two Penning traps. Ions continuously delivered from the mass
separator ISOLDE are accumulated, cooled and bunched in the RFQ trap where-
upon they are efficiently transported to the first Penning trap. In 1996 and 1997 a
large Paul trap was used for the accumulation and bunching stage of the experiment
first allowing for about one half of the results reported here. Mainly in order to
improve the emittance of the ejected pulses and to increase the bunching efficiency
the Paul trap was replaced by a linear Paul trap in 1999 [7]. The first Penning trap
is used as an isobar separator with a resolving power of up to R ~ 105 by applying
a mass-selective buffer gas cooling technique [6]. A cold and isobarically clean
Penning trap #2 :
cyclotron frequency
determination ,
0'0
isomer separation
Penning trap #1 :
cooling, isobar
separation ~! !~
Accumulation and bunching
I I
J ~~der
Paul trap ("96 /'97) I
I --i~.-~--'~"" I
Figure 1. Schematic layout of the experimental setup of the ISOLTRAP mass spectrometer.
MASS DETERMINATION OF NEUTRON-DEFICIENT NUCLIDES 339
ion bunch is then sent to the second Penning trap, a high-precision trap. There,
the mass m of the ions with charge q is determined by measuring their cyclotron
frequency Vc = q 1m· B 12l( with a resonant time-of-flight technique. The strength
of the magnetic field B is determined by a reference mass.
21
20
,,- Po a) AME 95
19
18
;;-
Q)
17
~
~ 16
(J)
15
14
13
12
11 L L. l --L... ...
96 100 104 108 112 124 128 132
N
(a)
21
b) AME includ ing
20 ISOL TRAP data
19
18
;;-
'"
:::;: 17
~ 16
(f)
15
14
13
12
11
96 100 104 108 112 11 6 120 124 128 132
(b)
Figure 2. Experimental two-neutron separation energies S2n in the region of Z = 80. (a) Data
as of [8], (b) including ISOLTRAP data. Full circles indicate S2n-values that are either
obtained for the first time or whose errors are decreased by at least a factor of two (40 cases).
340 S. SCHWARZ ET AL.
The results from the four ISOLTRAP runs reported here were integrated into the
network of known mass relations in a mass adjustment (called Atomic Mass Eval-
uation, AME [8]). Due to numerous mass links (mainly a-decays) the ISOLTRAP
data have a considerable influence on the mass region surrounding the investigated
nuclides. This impact can clearly be seen in Figure 2 which depicts experimental
two-neutron separation energies S2n for the elements Ho (2 = 67) to Ac (2 = 89)
as a function of neutron number. Besides the discontinuity at neutron number
N = 126, caused by the neutron shell closure, the new S2n-values exhibit a smooth
behaviour along the individual isotopic chains. A nearly linear trend can be seen
around 2 ;?: 78. Deviations from this linear trend are observed for the Pt, Hg and
Po chains around mid-shell, N = 104, with a magnitude of up to 200 ke V. This
"fine structure" may be caused by nuclear structure effects due to the presence of
shape coexistence. A more thorough discussion on these effects will appear in [9].
References
I. Heyde, K. et aI., Phys. Rep. 102 (1983), 291.
2. Wood, J. L. et ai., Phys. Rep. 215 (1992), 1Ol.
3. Bonn, J. et ai., Phys. Lett. B 38 (1972), 308.
4. Kiihl, T. et ai., Phys. Rev. Lett. 39 (1977), 180.
5. Bollen, G. et aI., Nucl. Instrum. Methods A 368 (1996), 675.
6. RaimbauIt-Hartmann, H. et aI., Nucl. Instrum. Methods B 126 (1997),378.
7. Herfurth, F. et ai., submitted to Nucl. Instrum. Methods; also preprint CERN-EPI2000-062.
8. Audi, G. and Wapstra, A. H., Nuclear Phys. A 595 (1995), 409.
9. S. Schwarz et ai., submitted to Nuclear Phys. A.
Hyperjine Interactions 132: 341-348,200!' 341
© 2001 Kluwer Academic Publishers.
Abstract. The present status of calculations of quantum electrodynamical, nuclear, and interelec-
tronic-interaction corrections to the binding energies in heavy few-electron ions is reviewed. The
currently available theoretical results for the Lamb shift in H-, He-, and Li-like ions are compared
with recent experimental data. A special attention is focused on testing quantum electrodynamics in
a strong electric field.
Key words: bound-state QED, Lamb shift, binding energy, heavy ions.
1. Introduction
R = 5
{ -(r) 2[ 3 2(3-(r4)
1 - -(aZ) - - -1)]}1/2 (4)
3 4 25 (r2)2 7
The formula (3) determines the finite nuclear size correction with the precision
"'-'0.2% in the range Z = 1-100.
Next we should take into account the QED corrections of first order in a. To
this order, the QED correction is defined by the self-energy (SE) and vacuum-
i:
polarization (VP) diagrams (Figure 1). The energy shift from the self-energy dia-
gram (Figure l(a» combined with the related mass counterterm is defined as
/j.,E = 2ia dw f f
dXl dX21j;,; (xdaflG(E" - w, Xl, X2)
(5)
QED EFFECTS IN HEAVY FEW-ELECTRON IONS 343
(a)
2 (b)
Figure I. First-order self-energy and vacuum-polarization diagrams.
where 1/1a (x) is the Dirac-Coulomb wave function of the state under considera-
tion, 1/1 = 1/I t yo, G(W, Xl, X2) is the Coulomb-Green function, DILv(w, Xl - X2)
is the photon propagator, aIL = (1, a), and a is a vector incorporating the Dirac
matrices. The most accurate calculations of the SE correction to all orders in a Z
were performed by Mohr [9] and by Indelicato and Mohr [10] for the point nucleus
case and by Mohr and Soff [11] for the extended nucleus case. The VP correction
(Figure l(b)) is given by
!)'EH = _i_
2n M
roo dW(al(D(W) _
J
-00
v:])[p,
W + 10
xG(w + Ea)(D(W) + v:])
[p, la}. (8)
W+10
Here p is the momentum operator, G (w) is the Coulomb-Green function, Dm (w) =
-4naZa,Dl m(w), and Dik(W, r) is the transverse part of the photon propagator in
the Coulomb gauge. In equation (8), the scalar product is implicit. The term 6.E L
contains all the recoil corrections within the (aZ)4m2 / M approximation and for
the point nucleus case one can obtain [20]
m2 _ E2
!)'E L = a (9)
2M
The term !)'EH contains the contribution of order (aZ)5 m 2 / M and all contributions
of higher order in a Z which are not included in !). E L. The calculations of the recoil
correction to all orders in aZ were performed in [22, 23] for point and extended
nuclei, respectively.
Finally, one must take into account the nuclear polarization correction. This
correction arises from diagrams describing the interaction of the electron with the
nucleus where the intermediate states of the nucleus are excited. It was evaluated
by Plunien and Soff [24] and by Nefiodov et at. [25].
The individual contributions to the ground-state Lamb shift in 238U91+ are pre-
sented in Table I. The uncertainty of the Dirac binding energy results from the
uncertainty of the Rydberg constant [26]. As one can see from the table, the present
status of experiment on the ground state Lamb shift in hydrogenlike uranium pro-
vides testing the QED effects of first order in a on the level of 5%.
I
Figure 2. One-photon exchange diagram.
"QED" part, is the residue. As one can see from the table, to test the second-order
QED effects in these experiments one should improve the experimental precision
by an order of magnitude.
In heavy Li-like ions, in addition to the one- and two-electron contributions dis-
cussed above, one has to calculate the three-electron corrections. In second order in
ex the three-electron contribution is defined by the two-photon exchange involving
all three electrons. The accurate QED calculations of all the two- and three-electron
corrections of second order in ex to the 2pI/2-2s transition energy were performed
recently in [36-38]. To get the theoretical precision on the level of the experiman-
tal one, in addition to an evaluation of all one-electron QED corrections of the
second order in ex, one should estimate the interelectronic-interaction corrections
of third- and higher-orders in l/Z. Such an estimate based on the Breit equation
approximation was done in the present work.
In Table III we present the individual contributions to the 2pI/2-2s transition
energy in Li-like uranium. The total theoretical value of the transition energy,
280.46(9) ± 0.20 e V, is in agreement with the related experimental value,
280.59(10) eV [4]. As one can see from the table, the first-order QED contribu-
tion is -42.93 eV while the total second-order QED contribution beyond the Breit
approximation amounts to 1.33 ± 0.20 eY. Comparing these values with the total
theoretical and experimental uncertainties indicates that the present status of the
theory for Li-like uranium provides a test of the QED effects of first order in ex on
QED EFFECTS IN HEAVY FEW-ELECTRON IONS 347
the level of about 0.5% and of the QED effects of second order in ex on the level of
about 15%.
5. Conclusion
In this paper we reviewed calculations of the Lamb shifts in heavy few-electron
ions. To date, the best opportunity for probing QED in a strong electric field appears
in heavy Li-like ions. The present status of the theory and experiment in Li-like
uranium provides a test of the QED effects of first order in ex on the level of about
0.5% and of the QED effects of second order in ex on the level of about 15%. To
increase the accuracy of the theoretical predictions the complete calculation of the
one-electron second-order QED corrections is needed.
Acknowledgements
The work ofY. M. S., Y. A. Y., O. M. Z., and M. M. S. was supported by the Russian
Foundation for Basic Research (Grant No. 98-02-18350) and by the program "Rus-
sian Universities. Basic Research" (project No. 3930). The support by the BMBF,
by GSI and by DFG is also acknowledged.
References
1. Beyer, H. E, IEEE Trans. Instrum. Measm. 44 (1995), 510;
Beyer, H. E, Menzel, G., Liesen, D., Gallus, A., Bosch, E, Deslattes, R., Indelicato, P., SWhl-
ker, Th., Klepper, 0., Moshammer, R., Nolden, E, Eickhoff, H., Franzke, B. and Steck, M.,
Z. Phys. D 35 (1995), 169.
2. SWhlker, T., Mok1er, P. H., Bosch, E, Dunford, R. w., Klepper, 0., Kozhuharov, C., Ludziejew-
ski, T., Franzke, E, Nolden, E, Reich, H., Rymuza, P., Stachura, Z., Steck, M., Swiat, P. and
Warczak, A., Phys. Rev. Lett. 85 (2000),3109.
3. Marrs, R. E., Elliot, S. R. and SWhlker, T., Phys. Re lJ. A 52 (1995), 3577.
4. Schweppe, J., Belkacem, A., Blumenfeld, L., Claytor, N., Feinberg, B., Gould, H.,
Kostroun, V. E., Levy, L., Misawa, S., Mowat, J. R. and Prior, M. H., Phys. Rev. Lett. 66 (1991),
1434.
5. Beiersdorfer, P., Osterheld, A., Scofield, J., Crespo Lopez-Urrutia, J. and Widmann, K., Phys.
Rev. Lett. 80 (1998), 3022.
6. Bosselmann, Ph., Staude, U., Hom, D., Schartner, K.-H., Folkmann, E, Livingston, A. E. and
Mokler, P. H., Phys. Rev. A 59 (1999), 1874.
7. Franosch, T. and Soff, G., Z. Phys. D 18 (1991),219.
8. Shabaev, V. M., J. Phys. B 26 (1993), 1103.
9. Mohr, P. J., Ann. Phys. (N.Y.) 88 (1974), 26, 52; Phys. Rev. A 46 (1992), 4421.
10. Indelicato, P. and Mohr, P. J., Phys. Rev. A 58 (1998), 165.
11. Mohr, P. J. and Soff, G., Phys. Rev. Lett. 70 (1993), 158.
12. Soff, G. and Mohr, P., Phys. Rev. A 38 (1988), 5066.
13. Manakov, N. L., Nekipelov, A. A. and Fainstein, A. G., Soviet. Phys. JETP 68 (1989), 673
(Zh. Eksper. Teoret. Fiz. 95 (1989), 1167).
14. Persson, H., Lindgren, I., Salomonson, S. and Sunnergren, P., Phys. Rev. A 48 (1993),2772.
15. Mohr, P. J., Plunien, G. and Soff, G., Phys. Rep. 293 (1998), 227.
348 V. M. SHABAEV ET AL.
16. Beier, T., Mohr, P. J., Persson, H., Plunien, G., Greiner, M. and Soff, G., Phys. Lett. A 236
(1997), 329.
17. Mitrushenkov, A., Labzowsky, L. N., Lindgren, I., Persson, H. and Salomonson, S., Phys. Lett.
A 200 (1995), 51.
18. Mallampalli, S. and Sapirstein, J., Phys. Rev. A 57 (1998), 1548.
19. Goidenko, I., Labzowsky, L., Nefiodov, A., Plunien, G., Soff, G. and Zschocke, S., Hyp.
Interact. 127 (2000), 293.
20. Shabaev, V. M., Theor. Math. Phys. 63 (1985), 588.
21. Shabaev, V. M., Phys. Rev. A 57 (1998), 59.
22. Artemyev, V. M., Shabaev, V. M. and Yerokhin, V. A., Phys. Rev. A 52 (1995),1884; 1. Phys. B
28 (1995), 5201.
23. Shabaev, V. M., Artemyev, A. N., Beier, T., Plunien, G., Yerokhin, V. A. and Soff, G., Phys.
Rev. A 57 (1998), 4235.
24. Plunien, G. and Soff, G., Phys. Rev. A 51 (1995),1119; 53 (1996), 4614.
25. Nefiodov, A. V., Labzowsky, L. N., Plunien, G. and Soff, G., Phys. Lett. A 222 (1996), 227.
26. Mohr, P. J. and Taylor, B. N., Rev. Modem Phys. 72 (2000), 351.
27. Blundell, S. A., Mohr, P. J., Johnson, W. R. and Sapirstein, J., Phys. Rev. A 48 (1993), 2615.
28. Lindgren, I., Persson, H., Salomonson, S. and Labzowsky, L. N., Phys. Rev. A 51 (1995), 1167.
29. Yerokhin, V. A., Artemyev, A. N. and Shabaev, V. M., Phys. Lett. A 234 (1997),361.
30. Artemyev, A. N., Shabaev, V. M. and Yerokhin, V. A., Phys. Rev. A 56 (1997), 3529.
31. Persson, H., Salomonson, S., Sunnergren, P. and Lindgren, I., Phys. Rev. Lett. 76 (1996), 204;
Persson, H., Salomonson, S., Sunnergren, P., Lindgren, I. and Gustavsson, M. G. H., Hyp.
Interact. 108 (1997), 3.
32. Sunnergren, P., PhD Thesis, Goteborg University and Chalmers University of Technology,
Goteborg, 1998.
33. Artemyev, A. N., Beier, T., Plunien, G., Shabaev, V. M., Soff, G. and Yerokhin, V. A., Phys.
Rev. A 62 (2000), 022116.
34. Mohr, P. J. and Sapirstein, J., Phys. Rev. A 62 (2000), 052501.
35. Shabaev, V. M., Artemyev, A. N. and Yerokhin, V. A., Phys. Scripta T86 (2000), 7.
36. Artemyev, A. N., Beier, T., Plunien, G., Shabaev, V. M., Soff, G. and Yerokhin, V. A., Phys.
Rev. A 60 (1999), 45.
37. Yerokhin, V. A., Artemyev, A. N., Beier, T., Plunien, G., Shabaev, V. M. and Soff, G., Phys.
Rev. A 60 (1999), 3522.
38. Yerokhin, V. A., Artemyev, A. N., Shabaev, V. M., Sysak, M. M., Zherebtsov, O. M. and
Soff, G., Phys. Rev. Lett. 85 (2000),4699.
Hyperfine Interactions 132: 349-363,200l. 349
© 2001 Kluwer Academic Publishers.
Abstract. We present recent results in the field of total binding energy calculations, Lande factors,
quantum electrodynamics corrections and lifetime that are of interest for ion traps and ion sources.
We describe in detail MCDF and RMBPT calculation of ionic binding energies, which are needed
for the determination of atomic masses from highly charged ion measurements. We also show new
results concerning Lande factor in 3-electron ions. Finally we describe how relativistic calculations
can help understand the physics of heavy ion production ion sources.
Key words: MCDF, RMBPT, correlation, Lande factor, highly charged ions.
1. Introduction
The development of slow, highly charged ion sources has opened many new oppor-
tunities for accurate measurements of atomic properties of few-electron ions. Si-
multaneously, Penning ion traps are used to measure ionic masses accurate enough
to be sensitive to electron binding energies [1-3], and highly accurate Lande gj fac-
tors [4]. It thus becomes necessary to improve our ability to evaluate theoretically
those quantities with the most advanced relativistic calculation schemes.
The SMILETRAP group at Stockholm has undertaken a series of accurate mea-
surements of binding energy of highly charged ions, starting with Cs, Ge and
* Present address: GSI, Planckstrasse 1, DE-64 291 Darmstadt, Germany.
350 P. INDELICATO ET AL.
Se, and aiming also at very heavy elements (Pb, Hg). Their initial accuracy on
atomic masses was around 10-9 [2] and is improving. Singly or doubly charged
ion masses have also been measured by the MIT group with an uncertainty of a few
parts in 10- 10 [3]. These masses, together with values for the binding energy and
mass of the missing electrons, allow for the determination of the atomic masses.
An accurate knowledge of the latter is needed in current experiments aiming for
a determination of the fine structure constant independent of the free electron
g factor [5, 6].
The Penning traps, which enabled to measure the well-known electron of anom-
alous magnetic moment as (g - 2)/2 = 1159652188.4(4.3) x 10- 12 [7], are now
being used for measurement of one-electron ions Lande factors with accuracy bet-
ter than 10-9 [4]. This experiment may also provide a new way to measure the
electron mass.
The understanding of the mechanism of heavy-ion production in ion sources
is also important for any future improvement. Recently it was demonstrated that
X-ray spectroscopy inside an Electron-Cyclotron Resonance Ion Source (ECRIS)
could lead to a very deep insight into the highly-charged ions production mech-
anism [8]. Similarly an experiment looking at X-rays from very heavy elements
provided accurate data for comparison with relativistic many-body and QED cal-
culations [9-11]. In both cases it is important that reliable information on radiative
and Auger transition rates, ionization and excitation cross sections is available to
provide detailed analysis of the ion production mechanisms [12, 13].
=L + L V(lr; - rj I),
N
where JeD is the one-electron Dirac operator and V is an operator representing the
electron-electron interaction of order a, which couple only positive energy states:
(2)
1
where At+ = At A is an operator projecting onto the positive energy eigenstates
of the single-particle Dirac Hamiltonian JeD (ri). This operator is the one-photon
exchange kernel from QED and, if Equation (1) is solved exactly, provides the
atom energy including all contributions from diagrams in the ladder approximation
(i.e., Feynman diagrams with series of parallel photon lines exchanged between
the electrons - each new photon line adding a 1I Z factor to the size of the correc-
tion), excluding all contributions from the negative-energy continuum, which can
be calculated only in the framework of QED.
The expression of Vij in Coulomb gauge and in atomic units is
1
Vij = (3)
rij
C'ii . C'i}
(4)
rij
_ C'ii_·
_ C'i_} [cos (_WI_ir_i}) - IJ
rij C
where ri} = Iri - r} I is the inter-electronic distance, Wij is the energy of the photon
exchanged between the two electr" .. .), C'ii are the Dirac matrices and c = l/a is
the speed of light. The term (3) represents the Coulomb interaction, the second
one (4) is the Gaunt (magnetic) interaction, and the last two terms (5) stand for
the retardation operator. In this expression the V operators acts only on rij and
not on the following wave functions. By a series expansion of Equations (4)-(5)
in powers of Wi} rij Ie « I one obtains the Breit interaction, which includes the
leading retardation contribution of order 1I c 2 . The Breit interaction is the sum of
the Gaunt interaction (4) and ofthe Breit retardation
The usual starting point for all many-body calculations is the Dirac-Fock (DF)
approximation. In this model, the electrons are treated in the independent-particle
approximation, and their wave functions are evaluated in the Coulomb field of
the nucleus and in the spherically-averaged field from the electrons. One can use
a variety of methods to take into account correlation effects. In the no-pair ap-
proximation, correlation is equal to the effect of the sum of all ladder diagrams
contribution from which the Dirac-Fock energy has been subtracted. Correlation
can be calculated by perturbation or by variational methods. Here we have used
simultaneously two methods. We used relativistic many-body perturbation theory
(RMBPT), in a variant called coupled clusters approximation (see, e.g., [21]) in
which large classes of contributions are included to all order. We also used the Mul-
ticonfiguration Dirac-Fock method (MCDF), a very efficient variational method. In
the MCDF, the total wave function of the system is made of a linear combination of
configuration state functions, which are Slater determinants with well defined total
angular momentum and parity. In MCDF both the mixing coefficients between
configurations (as in the relativistic interaction of configurations - RCI) and radial
wave functions are optimized by a variational procedure.
Among the progress done in the last few years, one can note that it is now
possible to treat all the parts of the electron--electron interaction on the same foot-
ing. In the RMBPT method contributions mixing operators in Equations (3), (4)
and (6) are included, while in the MCDF the self-consistent field equations contain
contributions from all three terms for both the mixing coefficients and the wave
function determination. Only the energy dependent part of the electron--electron
interaction, i.e., the difference between Equations (5) and (6) is treated only as a
perturbation, as the one-electron orbital energies cannot be defined except in the
Dirac-Fock case.
Finally, it should be noted that there is more to relativistic calculations than
a straightforward generalization of non-relativistic methods. For example, relax-
ation or correlation can lead to wrong non-relativistic limit for fine-structure level
splitting [22] or transition rates [23]. These problems arise because in a relativistic
calculation for each l there are two orbitals with j = l ± 1/2, which have different
radial parts (for example, P 1/2 and P3/2). If one calculate the energy of the ground
configuration of F, for example, the Is 2 2s 2 2p 5 J = 1/2 and J = 3/2 levels have
a different number of PI/2 and P3/2 electrons. In this case the relaxation act on the
atom in a different way for the two levels and the non-relativistic limit is not zero.
grams with at least one photon line crossing one or more rungs of the ladder. At the
present time both contributions are known only for the ground state of two-electron
systems [19, 20].
The second group is composed of radiative corrections, and include all diagrams
with either electron-positron closed loops (corresponding to virtual pair creations
in the field of the nucleus - the so-called vacuum polarization terms), or a self-
energy loop (i.e., a photon emitted and reabsorbed by the same electron). This
second group is the one which gives rise to the Lamb shift. In the next order a full
QED treatment, one obtains the radiative corrections (important for the innermost
shells) to the electron-nucleus interaction (self-energy and vacuum polarization).
The one-electron self-energy is evaluated using the one-electron values of Mohr
and coworkers [24-28]. Vacuum polarization and self-energy screening are treated
with the approximate method developed by Indelicato and coworkers [29-32].
These methods give results in close agreement with more advanced methods based
on QED (see for example [33-38]). Loop-after-loop contributions of the vacuum
polarization can be easily included by adding the Uelhing potential to the nuclear
potential in the self-consistent process.
The wave functions calculated with the methods described in Section 2 can be used
for the evaluation of other atomic properties. Usually one treats the different op-
erators in first order of the perturbation theory between correlated wave functions.
The situation of these calculation is much more complex than for energies as it has
been shown that effect of negative-energy states cannot be neglected, even at low -Z
[39-41]. Moreover, it is not even always possible to recover correct non-relativistic
limits for non-diagonal operators like transition probabilities [23].
Our implementation of the MCDP method, developed by Desclaux [42, 43]
and one of us (P. I.) can be used to evaluate any transition probability (En or
Mn) [44], hyperfine matrix elements [44, 45], parity violation matrix elements
[46], Auger cross-sections [47, 48], two-photon transition rates [49], and Lande
Figure 1. Feynman diagrams for self-energy in an external potential. The x represents exter-
nal potentials like the one created by a the charge distribution of other electrons, a magnetic
field (Lande factor) or a nucleus magnetic moment (hyperfine structure). The left diagram is
the wave function correction, the middle one represent the reducible contribution (change in
the bound state energy) and the third one os the vertex correction.
354 P. INDELICATO ET AL.
factors, following [50]. One of us (J. B.) implemented the latter operator in the
GRASP code [51] for comparison.
Table I. Contributions from single, double and triple substitutions to the cor-
relation energy in Be-like Cs, calculated by MCDF (eV), and comparison with
RMBPT. All configurations involving virtual orbitals up to 5g were considered.
CC: Coulomb correlation; BC: Breit correlation. Number in 0 are power of 10
1.0E+02 r - - - - - -- -- - - - - - - ,
1.0E+OI
>
....
~
-+-(KrJ
~
5c I.OE+OO - ---IArI
---(A11
1
o
......... (Mgl
-I e)
.S I.OE-OI ---(Bel
-+-(He)
~
Iii
o
1.0E.()2 1-----------'~,_-- 1
I.OE-03 I----.---_-~-~--.____l
2p 3d 4f Sg 7i
Size of the vinual orbital space
Figure 2. Convergence of correlation as a function of (n, e).
([He]: 0.15 eV; [Be]: ~ 0.01 eV). The main source of difference between MCDF
and RMBPT results is the incompleteness of the correlation calculation when many
electrons are present (for MCDF), since uncorrelated Coulomb and Breit contribu-
tions agree always within 0.1 eV. In order to show the rate of convergence and
the present limitations in MCDF values, we plotted in a logarithmic scale the
dependence of the Coulomb correlation as a function of the most excited virtual
orbital included in the variational process. The values presented in Figure 2 are the
increment in the modulus of Coulomb correlation, since this gives a good idea of
356 P. INDELICATO ET AL.
Table Ill. Comparison between theory and experiment (eV). RMBPT correlation for
Cs, MCDF correlation for Ge are included. The third column gives the sum of the
Coulomb and Breit correlation for comparison. Experimental values from [2, 3] and
G. Douyset (private communication)
the speed of approach of the "real" value. We also studied in detail the influence of
single and triple excitation as there are fundamental differences in the way those
correction are produced in MCDP and RMBPT case (Table I). More details on the
Cs case can be found in [58].
Por some Cs ions, the experimental values of some binding energies can be
determined from the comparison of the values from [2, 3], allowing for the com-
parison with theoretical predictions, as presented in Table III.
The detailed comparison we have performed shows that at the present stage of
experimental accuracy, the Dirac-Pock approach is more than sufficient to allow
RELATIVISTIC CALCULATIONS FOR TRAPPED IONS 357
for an accurate detennination of atomic binding energy from highly charged ions.
In Table IV we present binding energies of a number of ions of experimental in-
terest, evaluated in the DF approximation. In the near future however, correlation
energies accurate to the second order of perturbation theory will be needed.
-0 .000013 r - - - - - - - - - - - - - - - - - --------,
-0.000014
-0 . 000015
\~~,~
, "
- FBSCpalr1
-FBSCnp
.. ..
-0.000016 """
: ---- esc pair
--\:--'\----.~',.~ -::-:::.";"... ----- -----
~
.;, ----- -..--GAASP
-0.000017 - - 1Mlf'T
-- -+'-- EJp
-0.000018
-0.000019
shell
Figure 3. Convergence of correlation contribution to the Lande factor in lithium. FBSC: the
full Breit operator made self-consistent (with contribution from pairs); CSC: only Coulomb
interaction is made self-consistent; np: no-pair wave function is used; GRASP: the code
from [51] is used, equivalent to CSC, pair; RMBPT: [59].
-0.003030
-0.003040
-0.003050 --~\----
"• 1 ... ...
-0.003060 \'\~ ......~
" ..... , ...- --·- -FBSC
~
-
., -0.003070 .,
f \\ "" ---- ..... ----------- - FBSCnp
, -0.003080
0, \\ --... -- csc np j
Cl -0.003090 - ...... - GRASP
-0_003100 -
\
-0_003110 ,
-0.003120
-0.003130
3 5 7
shell
Figure 4. Convergence of correlation contribution to the Lande factor in lithiumlike Ca.
Contributions identical to those of Figure 3.
RELATIVISTIC CALCULATIONS FOR TRAPPED IONS 359
,3.
7000
6000
..
'
, .
sooo
4000
11.
3000
2000
1000
Energy (eV)
(a)
'ie
(m·') 13.
12.
".
e,.Ox10 "
, .
(b)
Figure 5. Comparison between experimental (a) [8] and theoretical (b) spectra for Ar ions in-
side an ECR ion source. Theory includes Ar9+ to Ar 16+ ions calculated with: K excitation and
single ionization only (- - -); double KL ionization only (~ .. ~); sum of all contributions (-).
The use of available literature has not allowed for a complete understanding of the
spectra obtained by using a rather high-resolution X-ray spectrometer [8]. We thus
used our MCDF code to provide transition energies, radiative and Auger transition
probabilities, electron-impact excitation [13]. The Dirac-Fock electron kinetic en-
ergies were used to provide single ionization cross section in the BEB formalism
of Kim and Rudd [60], while the double ionization cross sections were obtained
360 P. INDELICATO ET AL.
(m" )
Figure 6. Theoretical spectra in the energy region between 3080 and 3100 eV, including the
following processes: K excitation and single ionization only (- - -); double KL ionization from
Ar12+ Is2 2s2 2p2 configuration leading to Ar14+ Is 2s 2p2 (a); double KL ionization from
Ar 13 + Is22s22p configuration leading to Ar IS + Is2s2p (b); double KL ionization from
Ar12+ Is2 2s2 2p2 configuration leading to Ar14+ Is 2s2 2p (c); sum of all contributions (-);
experimental data [8] (x).
in the formalism of Shevelko and Tawara [61]. All ions in the range Ar9+ to Ar 16+
were calculated. With this calculation it was possible to explain quantitatively all
features of the experimental spectrum (Figure 5), including the ones not under-
stood before (see, e.g., Figure 6), thus comforting the determination of the plasma
properties in the source done in [8].
5. Conclusion
Acknowledgements
Financial support for this research has been provided by the European Community
under TMR contract number FMRX-CT97-0144 (EUROTRAPS). Prof. 1. Berg-
strom, Dr. C. Carlberg, Dr. G. Douysset and T. Fritioff are gratefully thanked
for discussions and communication of experimental data prior to publication. We
are very much indebted to Dr. Jean-Paul Desclaux for his substantial help in the
implementation of self-consistent treatment of the Breit operator. One of us (J. B.)
acknowledges support from the bilateral polish-french Polonium project no. 98113
during the completion of this work.
References
1. DiFilippo, F., Natarajan, v., Boyce, K. R. and Pritchard, D. E., Accurate atomic masses for
fundamental metrology, Phys. Rev. Lett. 73 (1994), 1481-1484.
2. Carlberg, C., Fritioff, T. and Bergstrom, I., Determination of the 133Cs and proton mass ratio
using highly charged ions, Phys. Rev. Lett. 83 (1999), 4506-4509.
3. Bradley, M. P., Porto, J. v., Rainville, S., Thompson, J. K. and Pritchard, D. E., Penning trap
measurements of the masses of 133Cs, 87,85Rb, and 23Na with uncertainties 0.2 ppb, Phys.
Rev. Lett. 83 (1999), 4510-4513.
4. Hermanspahn, N., Haffner, H., Kluge, H.-J., Quint, w., Stahl, S., Verdli, J. and Werth, G.,
Observation of the continuous Stem-Gerlach effect on an electron bound in an atomic ion,
Phys. Rev. Lett. 84 (2000),427-430.
5. Weiss, D. S., Young, B. C. and Chu, S., Precision measurement of the photon rccoil of an atom
using atomic interferometry, Phys. Rev. Lett. 70 (1993), 2706-2709.
6. Mohr, P. J. and Taylor, B. N., CODATA recommended values of the fundamental physical
constants: 1998, Rev. Modem Phys. 72 (2000),351-495.
7. Van Dyck, R. S., Jr., Schwinberg, P. B. and Dehmelt, H. G., New high-precision comparison of
electron and positron g factors, Phys. Rev. Lett. 59 (1987), 26-29.
8. Douysset, G., Khodja, H., Girard, A. and Briand, J. P., Highly charged ion densities and ion
confinement properties in an electron-cyclotron-resonance ion source, Phys. Rev. E 61 (2000),
3015-3022.
9. Beiersdorfer, P., Knapp, D., Marrs, R. E., Elliot, S. R. and Chen, M. H., Structure and Lamb-
shift of 2slj2-2 P3 /2 levels in lithiumlike U 89 + through neonlike U 82 +, Phys. Rev. Lett. 71
(1993),3939-3942.
10. Beiersdorfer, P., Osterheld, A., Elliott, S. R., Chen, M. H., Knapp, D. and Reed, K., Structure
and Lamb shift of 2s1 /2-2P3/2 levels in lithiumlike Th 87 + through neonlike Th80+, Phys. Rev.
A 52 (1995), 2693-2706.
11. Beiersdorfer, P., Osterheld, A. L., Scofield, J. H., Crespo L6pez-Urrutia, J. R. and Widmann, K.,
Measurement of QED and hyperfine splitting in the 2SI/2-2P3/2 X-ray transition in Li-like
209Bi80+, Phys. Rev. Lett. 80 (1998),3022-3025.
12. Santos, J. P., Marques, J. P., Parente, F., Lindroth, E., Boucard, S. and Indelicato, P., Multicon-
figuration Dirac-Pock calculation of 2s1 /2-2 P3 /2 transition energies in highly ionized bismuth,
thorium, and uranium, European Phys. 1. D 1 (1998), 149-163.
13. Martins, M. c., Costa, A. M., Santos, J. P., Indelicato, P. and Parente, F., Interpretation of
X-ray spectra emitted by Ar ions in an electron cyclotron resonance ion source, 1. Phys. B 34
(2001),533-543.
14. Mohr, P. J., Plunien, G. and Soff, G., QED corrections in heavy atoms, Phys. Rep. 293 (1998),
227-372.
362 P. INDELICATO ET AL.
39. Indelicato, P., Correlation and negative continuum effects for the relativistic M1 transition in
two-electron ions using the multiconfiguration Dirac-Fock method, Phys. Rev. Lett. 77 (1996),
3323-3326.
40. Derevianko, A., Savukov, 1. M. and Johnson, W. R., Negative-energy contributions to transition
amplitudes in heliumlike ions, Phys. Rev. A 58 (1998), 4453-4461.
41. Savukov, 1. M., Derevianko, A., Berry, H. G. and Johnson, W. R., Large contributions
of negative-energy states to forbidden magnetic-dipole transition amplitudes in alkali-metal
atoms, Phys. Rev. Lett. 83 (1999), 2914-2917.
42. Desclaux, J. P., A multiconfiguration relativistic Dirac-Fock program, Comput. Phys. Commun.
9 (1975),31-45.
43. Desclaux, J. P., A relativistic multiconfiguration Dirac-Fock package, In: E. Clementi (ed.),
Methods and Techniques in Computational Chemistry, Vol. A. STEF, Cagliary, 1993.
44. Indelicato, P., Parente, F. and Marrus, R., Effect of the hyperfine structure on the 2 3 PI and the
2 3 Pl lifetime in heliumlike ions, Phys. Rev. A 40 (1989), 3505-3514.
45. Boucard, S. and Indelicato, P., Relativistic many-body and QED effects on the hyperfine
structure of lithium-like ions, European Phys. J. D 8 (2000), 59-73.
46. Schafer, A., Soff, G., Indelicato, P., Miiller, B. and Greiner, W., Prospects for an atomic parity-
violation experiment in U 90 +, Phys. Rev. A 40 (1989), 7362-7365.
47. Marques, J. P., Parente, F., Indelicato, P. and Desclaux, J. P., Estimation of the ratio of double
and single Auger transition rates for the L shell of Kr, Nb and Gd, J. Phys. B. 31 (1998),
2897-2901.
48. Santos, J. P., Marques, J. P., Parente, F., Lindroth, E., Indelicato, P. and Desclaux, J. P., Rela-
tivistic 2SI/2 (Ll) atomic subshell decay rates and fluorescence yields for Yb and Hg, J. Phys.
B. 32(1999), 2089-2097.
49. Santos, J. P., Parente, F. and Indelicato, P., Application of B-splines finite basis sets to rela-
tivistic two-photon decay rates of 2s level in hydrogenic ions, Europeran Phys. J. D 3 (1998),
43-52.
50. Cheng, K. T. and Childs, W. J., Ab initio calculation of 4 j n 6s 2 hyperfine structure in neutral
rare-earth atoms, Phys. Rev. A 31 (1985), 2775-2784.
51. Parpia, F. A., Froese Fischer, C. and Grant, 1. P., 1996.
52. Persson, H., Salomonson, S., Sunnergren, P. and Lindgren, 1., Radiative corrections to the
electron g-factor in H-like ions, Phys. Rev. A 56 (1997), R2499-R2502.
53. Beier, T., Lindgren, 1., Persson, H., Salomonson, S., Sunnergren, P., H. Haffner and N. Her-
manspahn, g j factor of an electron bound in a hydrogenlike ion, Phys. Rev. A 62 (2000),
032510.
54. Blundell, S. A., Cheng, K. T. and Sapirstein, J., Radiative corrections in atomic physics in the
presence of perturbing potentials, Phys. Rev. A 55 (1997), 1857-1865.
55. Persson, H., Schneider, S. M., Greiner, W., Soff, G. and Lindgren, 1., Self-energy correction to
the hyperfine structure splitting of hydrogenlike atoms, Phys. Rev. Lett. 76 (1996),1433-1436.
56. Shabaev, V. M., Shabaeva, M. B., Tupitsyn, 1. 1., Yerokhin, V. A., Artemyev, A. N., Kiihl, T. and
Tomaselli, M., Transition energy and lifetime for the ground-state hyperfine splitting of high-Z
lithiumlike ions, Phys. Rev. A 57 (1998), 149-156.
57. Indelicato, P. and Mohr, P. J., Quantum electrodynamic effects in atomic structure, Theoret.
Chem. Acta. 80 (1991), 207-214.
58. Rodrigues, G. c., Ourdane, M. A., Bieron, J., Indelicato, P. and Lindroth, E., Relativistic and
many-body effects on total binding energies of cesium ions, Phys. Rev. A 63 (2001), 012510.
59. Lindroth, E. and Ynnerman, A., Ab initio calculations of gj factors for Li, Be+ and Ba+, Phys.
Rev. A 47 (1993), 961-970.
60. Kim, Y.-K. and Rudd, M. E., Binary-encounter-dipole model for electron-impact ionization,
Phys. Rev. A 50 (1994),3954-3967.
61. Shevelko, V. P. and Tawara, H., J. Phys. B. 28 (1995), L589.
..... Hyperfine Interactions 132: 365-368,200l. 365
© 2001 Kluwer Academic Publishers.
"
Abstract. We derive the electroweak radiative corrections to the basic PNC atomic transition am-
plitude for highly charged hydrogenlike ions. In the case of highly charged ions (HCI) effects of
m;
strong fields are reflected by the momentum transfer q2 involved. It is of the order q2 ;::,; in HCI,
while q2 ;::,; 0 in neutral atoms. This may open the possibility to search for "new physics" beyond the
Standard Model and to test the Standard Model in experiments with HCI.
1. Introduction
where s2 = sin2 ew
~ 0.2394 - with the Weinberg angle ew,
N and Z denote the
numbers of neutrons and protons, respectively.
Radiative electroweak corrections were calculated previously for neutral atoms
by Lynn and Sandars [1] in the factorized form
~rad
HPNC = ~Arad
U PNC
H~
PNC· (3)
M
8p = Mi
2 1 ~ 0.0880, (4)
[M~(q = 0)] 2 -
w
where P == Pw(q ).
~ 2
~
In [1] electroweak radiative corrections to Pw were presented in the form:
Pw = p:jjl + 8;cak-anapole + 8~-boson vertex + 8~oxes (5)
together with the various "specific" PNC radiative corrections 8~).
In the case of HCI we have to consider the basic PNC transition amplitude, be-
cause the PNC matrix element, that describes the interaction between the electron
and nucleus via a Z-boson in principle has no physical meaning at all [5]. The basic
atomic PNC amplitude is described by the expression
A*
Mo = e(n"Sl/21A eln' PI/2)(n' PJ/21 HpNC InsI/2)~ol, (6)
where ~oJ is the energy shift between the n' PI/2 and the nSI/2-state.
To deduce the radiative corrections to the amplitude Mo we have to evaluate the
corresponding amplitude M~~d in the case of strong field and MLW in the low-field
limit, respectively. Accordingly, any deviation of the ratio
Irad = M SF
rad
-
M0
(7)
M LF
rad - M0
from unity will describe the strong field effects in comparison with the situation in
neutral atoms.
The influence of strong fields on the PNC radiative correction in the HCI is
essentially carried by the electron-loop corrections to the wave functions and the
renonnalization corrections to the Z-coupling. The electron-loop corrections to
the wave functions were not considered in the neutral atoms [1]. The other field-
dependent corrections are suppresed by the factor (l - 4s 2 ).
We write the generic results of our calculations in the fonn
(8)
+ 8Z-vertex(frad - 1) (9)
P Z-vertex .
Here ~ Pw describes the difference between the corrections for HCI and neutral
atoms, the various I are defined in Equation (7).
4. Conclusion
Two reasons for the investigation of PNC effects in HCI should be addressed: Due
to the very small energy denominator ~o and the large nuclear charge Z the PNC
effects are expected to be strongly enhanced in Ga62+ and Eu 61 +. The complete
set of the PNC radiative correction can contribute about 10% in the total PNC
effects [5]. For a consecutive treatment of PNC radiative corrections in HCI we
should consider all of the radiative corrections of first order in C( to the basic PNC
transition amplitude.
At present the complete numerical calculations for field dependent PNC radia-
tive corrections are carried out for 8~·f. [5]. Next steps will be to derive all radiative
368 I. BEDNYAKOV ET AL.
Acknowledgements
The authors are indebted to M. G. Koslov, V. M. Shabaev and I. B. Khriplovich
for useful discussions. L. L. and I. B. are grateful for financial support from RFBR
(grant No. 99-02-18526) and I. B. is grateful also for a ISSE grant.
References
1. Lynn, B. W. and Sandars, P. G. H., 1. Phys. B 27 (1994), 1469.
2. Labzowsky, L., Klimchitskaya, G. and Dmitriev, Yu., Relativistic Effects in Atomic Systems,
lOP, Bristol, 1993.
3. Labzowsky, L. N., Nefiodov, A. v., Plunien, G., Soff, G., Marrus, R. and Liesen, D., Parity-
violation effect in heliumlike gadolinium and europium, Phys. Rev. A 63 (2001), 054105.
4. Indelicato, P., Birkett, B. B., Briand, J.-P., Charles, P., Dietrich, D. D., Marrus, R. and
Simionovici, A., Phys. Rev. Lett. 68 (1992), 1307.
5. Bednyakov, I., Labzowsky, L., Plunien, G., Soff, G. and Karasiev, v., Phys. Rev. A 61 (1999),
012103.
Hyperfine Interactions 132: 369-374, 2001. 369
© 2001 Kluwer Academic Publishers.
Abstract. We present results for the two-photon vacuum-polarization screening correction in He-
like heavy ions in the range Z = 20--100. Values for the ground state as well as for low-lying excited
states are given.
Key words: bound-state QED, heliumlike systems, strong fields, vacuum polarization, two-time
Green function.
1. Introduction
* Corresponding author.
370 T. BEIER ET AL.
the (1s2p3/2h,2 states), Throughout the paper, we use relativistic units (Ii = c =
me = 1).
2. Calculation formulae
The diagrams under consideration are shown in Figure 1. The graphs in the upper
row are these where the vacuum polarization loop modifies one of the electron
propagators whereas in the diagram of the lower row the photon propagator is
affected. To obtain expressions suitable for calculating the energy shifts we have
applied the method of the two-time Green function [10, 11] with its possibility
to consider quasidegenerate states [12, 13]. For the present case, this is important
since the states (ls2pl/2)1 and (lS2p3/2)1 are strongly mixed for low and intermedi-
ate Z. A detailed derivation for the diagrams of the present article is given in [14].
The final expressions are
H . -
lk -
H(irred)
ik
+ H(red)
ik' (1)
where
1
Hi~rred) = '2 I)-I)P[{8Pi 1 Pi 2 jI(lll) + I(ll2)jk 1k2}
P
and
1
Hi~ed) = '2 L(-I)P {[(pi]IU~pjPil) - (klIU~plk]}](PilPi2II'(lll)lklk2)
p
Figure 1. Vacuum polarization screening diagrams. Double lines denote bound fermions,
wavy lines denote photons.
VP SCREENING CORRECTIONS IN HE-LIKE IONS 371
Here, kl and k2 denote the two electrons in the initial state k, i 1 and i2 the electrons
in the final state i, P is the permutation operator given by
The notation 81a) refers to the perturbation of a state la) by the vacuum polarization
potential,
f
00
utp(x) ~fdy_l- dwTr(G(w,y,y»), (5)
2m Ix - yl
-00
where
G(
w, x, y
)= L o/(x)o/t(y)
(1 '0)
n W - Cn -1
is the Coulomb-Green function, The notation I (~) in Equations (2) and (3) in-
dicates the photon propagator together with the corresponding matrices for the
vertices,
I (w) == 4:rraaia~Dllv(w), (6)
Here all == yOyll = (1, a), D llv is the photon propagator given by
exp (iJ
w 2 - IL2 + iO Ix - yl)
Dllv(w, x - y) = g l l v - - - ' - - - - - - - - - (7)
4:rrlx - yl
in the Feynman gauge C:sJw2 - IL2 + iO > 0) with I (w) I (-w), l' (w) =
-I'( -w), and ~l = CPi) - ck), ~2 = cPi2 - ck z '
The contribution from the diagram with the modified photon propagator is ob-
tained by taking the expression for the one-photon exchange diagram, Figure 2 [14],
1
Hi~l) = 2 L(-I)P[(PilPi21I(~1)lklk2) + (PilPi21/(~2)lklk2)]' (8)
P
f f
utp(c, x, y)
z ll)a2 v exp(ilcIIY- Z21)
= -a2100dw dZ l al-ll-exp(ilcllx-
dz 2- '--=----------=---"----
2:rri -00 Ix - zll Iy - z21
Thus, we have
1
H (m.ph.p.)
ik 2: ~)-1)P[(Pi1Pi2Iutp(~1)lk1k2}
p
Equations (2), (3), and (10) provide the matrix elements between the one-determi-
nant wave functions
1
Ui = M L(-1)PIPi1 Pi2)· (11)
.y 2 p
For a singly excited state, the full unperturbed wave functions are written as
(12)
To obtain the matrix elements between these wave functions, the equations have
to be multiplied by the corresponding Clebsch-Gordan coefficients and the sums
over the projections of the one-electron angular momenta have to be carried out.
The contributions (2), (3), and (10) are ultraviolet divergent. The renormaliza-
tion of these contributions is performed in the same way as in [5, 15] by splitting the
vacuum-polarization potential into a Uehling part which is renormalized separately
and a Wichmann-Kroll part.
Acknowledgements
The work of A. N. A., V. M. S., and V. A. Y. was supported by the Russian Foun-
dation for Basic Research (Grant No. 98-02-18350) and by the program "Russian
Universities. Basic Research" (project No. 3930). T. B. is grateful to the EU-TMR
programme (contract No. ERB FMRX CT 97-0144). The support by BMBF, by
GSI, and by DFG is also acknowledged.
References
I. Marrs, R. E., Elliott, S. R. and SWhlker, Th., Phys. Rev. A 52 (1995), 3577.
2. Blundell, S., Mohr, P., Johnson, W. and Sapirstein, J., Phys. Rev. A 48 (1993),2615.
3. Lindgren, 1., Persson, H., Salomonson, S. and Labzowsky, L., Phys. Rev. A 51 (1995), 1167.
4. Yerokhin, V. A, Artemyev, A N. and Shabaev, V. M., Phys. Lett. A 234 (1997), 361.
5. Artemyev, A. N., Shabaev, V. M. and Yerokhin,V. A, Phys. Rev. A 56 (1997),3529.
6. Persson, H., Salomonson, S., Sunnergren, P. and Lindgren, 1., Phys. Rev. Lett. 76 (1996),204.
7. Persson, H., Salomonson, S., Sunnergren, P., Lindgren, I. and Gustavsson, M. G. H., Hyp.
Interact. 108 (1997), 3.
8. Asen, B., Salomonson, S. and Lindgren, I., Hyp. Interact. 127 (2000), 319.
9. Mohr, P. J. and Sapirstein, J., Phys. Rev. A 62 (2000), 052501.
10. Shabaev, V. M., In: U. 1. Safronova (ed.), Many-Particle Effects in Atoms, AN SSSR, Nauchnyi
Sovet po Spectroskopii, Moscow, 1988, p. 15; Soviet. Phys. J. 33 (1990), 660.
374 T. BEIER ET AL.
Abstract. Atomic mass differences are influenced by QED corrections, and a reliable understanding
of these corrections is therefore of importance for the current and next generation of high-precision
mass determinations based on Penning traps. We present a numerical evaluation of the self-energy
correction, which is the dominant contribution to the Lamb shift, in the region of low nuclear charge.
Our calculation is nonperturbative in the binding field and has a numerical uncertainty of 0.8 Hz in
atomic hydrogen for the ground state and of 1.0 Hz for L-shell states (2SI/2, 2PI/2, and 2P3/2).
states. The self energy is an effect of order a because of the two electron-photon
vertices and it obtains four additional powers of Za due to the renormalization and
the Ward-Takahashi identity,
a (Za)4 2
8EsE = ---3-meC F(nlj, Za), (1)
rr n
where the indices n, land j refer to the bound state quantum numbers, and the
parameter Za indicates the dependence on the nuclear charge and the fine structure
constant, which enters only through the product of the two quantities. The function
F (nl j, Za) is dimensionless. For an S state, we have the following semi-analytic
expansion of F in powers of Za and In Za,
The A-coefficients carry two indices the first of which indicates the power of
Za, the second corresponds to the power ofln[(Za)-2]. The nonperturbative self-
energy remainder function GSE(nSlj2, Za) summarizes the nonlogarithmic terms
of relative order (Za)2 (the so-called A6o -coefficient) and higher-order terms. We
have,
(3)
For P states, the A 41 , A50 and A62 -coefficients vanish, and we have
The P state wavefunction is less singular at the origin than the S state. Leading
corrections to the Lamb shift, which correspond to the singular logarithmic terms,
are proportional to the square of the wavefunction at the origin and therefore absent
for P states.
Numerical values for the self energy of K and L shell states have been obtained
in the range Z = 1-5 and are summarized in Table 1. The numerical values for
the self-energy remainder function G SE can be determined from the values of the
dimensionless self-energy function F defined in Equation (1) by subtracting the
contribution of the known lower-order analytic terms given by the A-coefficients
from Equations (2) for S states resp. (4) for P states.
The values given in Table I constitute the first direct calculations in this Z-range
for the self energy. The numerical uncertainty of the values is ::;; 1.0 X Z4 Hz,
thereby matching the accuracy of the best current measurements. A comparison
of numerical and analytic approaches, which is discussed in detail in [5], leads
to very satisfactory consistency of the two principal approaches (numerical and
CALCULATION OF QED EFFECTS IN HYDROGEN 377
Table I. Numerical results for the scaled self-energy function F and the self-energy remainder
function GSE for K and L shell states in the low-Z region
analytic) to the Lamb shift problem, which has attracted attention for more than
five decades. The reliable understanding of quantum electrodynamic radiative cor-
rections, which contribute to atomic binding energies, is important for current and
projected accurate mass measurements based on Penning traps.
Acknowledgements
U. D. J. acknowledges support from the Deutscher Akademischer Austauschdienst
(DAAD). p. J. M. acknowledges the Alexander von Humboldt Foundation for con-
tinued support. G. S. wishes to acknowledge support from BMBF, DFG and from
GSI.
References
1. Niering, M., Holzwarth, R., Reichert, J., Pokasov, P. Udem, Th., Weitz, M., Hansch, T. W.,
Lemonde, P., Santarelli, G., Abgrall, M., Laurent, P., Salomon, C. and Clairon, A., Phys. Rev.
Lett. 84 (2000), 5496.
2. de Beauvoir, B., Nez, F., Julien, L., Cagnac, B., Biraben, F., Touahri, D., Hilico, L., Acef, 0.,
Clairon, A. and Zondy, J. J., Phys. Rev. Lett. 78 (1997), 440.
3. Jentschura, U. D., Mohr, P. J. and Soff, G., Phys. Rev. Lett. 82 (1999), 53.
4. Jentschura, U. D., Mohr, P. J., Soff, G. and Weniger, E. J., Comput. Phys. Commun. 116 (1999),
28.
5. Jentschura, U. D., Mohr, P. J. and Soff, G., submitted (2000).
Hyperfine Interactions 132: 379-383, 2001. 379
© 2001 Kluwer Academic Publishers.
V. G. PAL'CHIKOV
National Research Institute for Physical-Technical and Radiotechnical Measurements - VNIIFTRI,
Mendeleevo, Moscow Region, 141570 Russia; e-mail: vitpai@mail.ru
Abstract. The wavelengths of the Isnp 1 PI-Is 2 1So transitions in He-like Mg XI, F VIII (n = 4-8)
and Al XII (n = 6,9) have been calculated in the framework of the 1/ Z expansion method includ-
ing relativistic effects and QED contributions. It is found that QED corrections to the ground-state
ionization energy are significant at the present level of experimental accuracy.
1. Introduction
Here, the subscripts LS and JJ in (1) denote the coupling scheme in which the corre-
sponding matrix elements are evaluated, R is the JJ-LS recoupling-transformation
matrix [3], the matrix ~ is the correction and is counted twice in the first two terms.
HNR is the non-relativistic two-body Hamiltonian, Bp consists of the electron-
electron interaction in the Breit-Pauli approximation, Ho is the sum of one-electron
Dirac Hamiltonians, HM is the mass polarization correction. V 12 in Equation (1) is
the pure Coulomb potential, B is the Dirac form of the Breit interaction including
retardation effects. The lowest order QED-correction EQEO in (1) is given by
where (8 (rd + 8 (r2») is the total electron density which can be obtained from the
variational wave-functions [3]. For the ground state the Bethe logarithm (Ln(EnLS»
in (2) was calculated from the screened H-like approximation. The J-independent
part of the two-electron correction is presented in [3]. E~EO in (1) contains higher
order diagonal one-electron QED corrections as calculated by
a 3 Z4
~En = - - 1 Fn(aZ), (3)
rrn-
where
The numerical values for C;, and Cij are tabulated in [4-6]. Higher-order vacuum-
polarization contributions in Equation (4) are presented in [7].
I'1EQED a(aZ)2 n2
-~-
I'1E
=2 7T
-2--F(aZ),
n - 1
(7)
where F(aZ) is a slowly varying function of Z having values of 4.65 and 1.86
for Z = 10 and Z = 50, respectively [6]. In our case of He-like Mg ions this
ratio is approximately equal to 1.5 x 10-4 and, as follows from Table I, the I'1EQED
contribution is within the experimental accuracy.
Table II gives the corresponding contributions to the transition energies as well
as the experimental estimates for I'1E;"lo. As it is shown in the Table II the QED
contributions to the transition energies are almost independent of n and originate
mainly from the Is orbital of the Isnp configuration. By applying the same ap-
proach for ions Al XII (n = 6, 9), we obtain the following final result: I'1E~X:o(n =
6) = -2.68(0.9) . lO3 cm- I and I'1E~X:o(n = 9) = -2.68(0.7) . 103 cm- i . The
corresponding theoretical result for I'1EQED is I'1E~ED = -2.662· 103 cm- 1 [3].
382 V. O. PAL'CHIKOV
Table II. Nonrelativistic and QED contributions to the transition energy of He-like Mg
(in cm- I ). l1E~ED = -1.996.10 3 cm- I [3]
Ground state ionization energies of Mg XI, F VIII ions deduced from experi-
mental [1] and theoretical data are:
[exp(F VIII XI) = 953.96 ± 0.11 eV,
[th(F VIII XI) = 953.9035 ± 0.11 eV,
(8)
[exp(Mg XI) = 1761.88 ± 0.15 eV,
[exp(Mg XI) = 1761.8065 ± 0.15 eV.
References
I. Pal'chikov, V. G., Faenov, A. Ya., Skobclev, 1. Yu., Magunov, A. 1., Pikuz, T. A., Rosmej, F. B.,
Hoffmann, D. H. H., Suss, W., Geissel, M., Biemont, E., Palmeri, P., Quinet, P., Sasaki, A. and
Utsumi, T., 1. Phys. B (submitted).
2. Osterheld, A. L., Magunov, A. 1., Dyakin, V. M., Faenov, A. Ya., Pikuz, T. A., Skobelev, 1. Yu.,
Pisarczuk, T., Parys, P., Wolowski, 1., Makowski, 1., PikUL, S. A. and Romanova, V. M., Phys.
Rev. A 54 (1996), 3971.
3. Drake, G. W. F., Can. 1. Phys. 66 (1988). 586.
4. Shabaev, V. M., Phys. Rev. A 57 (1998), 59.
5. Yerokhin, V. A. and Shabaev, V. M., Phys. Rev. A 60 (1999), 800.
6. Pal'chikov. V. G. and Shevelko, V. P., Reference Data Oil Multicharged Ions, Springer, Berlin,
1995.
TESTING OF QED-THEORY AND PRECISE MEASUREMENTS 383
Abstract. Dielectronic recombination of Pb 53 + has been studied and a resonance is detected only
~0.1 meV above the ground state. The possibility to determine the 4Plj2 - 4S1/2 energy splitting
with a similar accuracy from the determination of the resonance position is discussed. Such a pre-
cision can only be achieved by calculations which treat QED in a many-body environment at levels
which can still not be reached. A fully relativistic many-body calculation of the splitting is described
and the uncertainties are discussed.
1. Introduction
In electron-ion collisions, a free electron may be captured by an ion which is
simultaneously excited. In this inverse Auger process a doubly excited state is
created. If it decays radiatively the new charge state is stabilized and Dielectronic
Recombination (DR) is completed. DR is a fundamental recombination process
important for modeling of astrophysical and fusion plasma.
Accurate measurements of recombination resonances can be used for critical
tests of calculations in order to improve our understanding of formation of atoms
and the binding of their constituents. With the new super-conducting cooler at the
CRYRING storage ring in Stockholm the electron temperature is so low that reso-
nances of the electrons with the stored ions can be seen that are only meV or less
above threshold. It is then possible to determine energy splittings with the same ac-
curacy and in highly charged ions tests of predictions of quantum-electrodynamical
(QED) effects with very high precision are possible. The inclusion of radiative
effects as self energy and vacuum polarization, effects appearing only with a QED
treatment, in many particle theory is still in the development phase and the preci-
sion needed here goes beyond what is achievable with approaches [1-3] usually
applied to systems with more than three electrons.
Resonances situated closely above the ionization threshold exist in many ions
and it should be possible to find several species with resonances extremely close
to threshold. According to the experiment performed at CRYRING, Pb53+ is such
an ion. The first suspicion that Pb 53+ was a good candidate rose when it was dis-
386 M. TOKMAN ET AL.
10
8
. -.---
'VI
~
•
\
"?O
-
0
E
6
•
\
Q)
u 4
:E
Q)
0
°
!!
1';1
a: 2
Figure 1. The experimental recombination rates for Pb 53 + is plotted against relative energy.
covered that the storage ring lifetime for this ion was extremely short; 100 times
shorter than for Pb54+ [4]. If electrons, due to their velocity distribution, happen to
be in resonance during cooling the enhancement of the recombination rate could
cause a lifetime effect. In the recombination spectra presented here a first resonance
seems to appear just around 10-4 e V, see Figure 1.
Pb53+ has 29 electron. In the ground state 28 of these fill all shells up to 3d lO ,
and the last electron occupies the 4S I / 2 orbital. Dielectronic recombination ofPb53 +
with a low energy electron can occur through the following steps
Pb53 +(4s 1/ 2 ) + e- ---+ Pb52 +** (4pjnC j ) ---+ Pb52 +* + fiw, (1)
where only active electrons are given explicitly. Pb52 +** denotes the resonant dou-
bly-excited state and Pb52 +* any state bound below the ionization threshold. From
the difference in binding energy of a 4S 1/ 2 and a 4pl/2 electron, see Table I, one
finds that the lowest energy doubly excited states 4 PI/2nC j which are above the
ionization threshold, and can contribute to recombination, will have n = 18.
The recombination resonances will appear for electron energies
(2)
and if it is possible to calculate b..E very accurately the E(4pj) - E(4s l / 2) split-
ting can be obtained from the energy positions of the recombination resonances.
Accurate calculations of b.. E are indeed possible since the binding energy of such
a highly excited electron as n = 18 is close to hydrogen-like, especially for the
higher C-states, and the small contribution from electron correlation (in the me V re-
gion) to the E(4pl/218C j) energy level can be calculated by relativistic many-body
ACCURATE CALCULATIONS ON DlELECTRONIC RECOMBINATION RESONANCES 387
Table I. The first three columns show the contributions to the binding energy of the last electron in
Pb53 + (4S1/2), Pb53+ (4Pl/2) and Pb53+ (4P3/2) respectively in a.u. The last two columns show the
4Pl/2 - 4S1/2 and 4P3/2 - 4S1/2 splittings in eV. 1 a.u. = 27.211396 M/(M + me) eV
a From Table II, plus the additional contributions due to the DFB-potential, see text.
perturbation theory to an accuracy within at least 10-4 eV. It is also possible to get a
consistency check of the many-body calculation since there are several recombina-
tion resonances which all in principle can be used to deduce the E(4pj) - £(4S1/2)
splitting. Here we concentrate on the precision with which one today can predict
the £(4pj) - £(4S I / 2 ) splitting. This is investigated in Section 2. In Section 3 we
outline how the resonances are calculated. The comparison between theory and
experiment is discussed in Section 4.
4S I / 2 , 4PI/2 and 4p3/2, respectively, and are added to the second order values in
Table I. The differences E (.e max) - E (f max - 1) are fitted by a polynomial a / f~ax +
b/.e~ax + c/f~ax and summed up to f max = 00. In Table II we show the results of
such summation from three-point fits. The final sum was obtained by Aitkens 82 _
process. The complete second order Coulomb-Coulomb correlation is calculated
as a sum of the second order Coulomb-Coulomb core-core correction and this final
sum. The uncertainties of the extrapolation were estimated as half the difference
between the final sum and the core-valence contribution with f max = 10. The
uncertainties for the 4PI/2-4s l / 2 and 4p3/2-4s l / 2 splittings are even smaller since
the contributions for higher partial waves are always negative and of the same order
of magnitude for all the n = 4 states.
The result is compared against an earlier calculation by Johnson et al. [6]. Ref-
erence [6] used a procedure similar to the present one, although the partial wave
expansion was truncated at f max = 8 instead of at f max = 10. The second order
Coulomb contributions from [6] differ by less than lOme V from the present result.
The differences are probably caused by the partial wave extrapolation which is the
main source of uncertainty in both calculations.
The scheme of obtaining the final result for the second order Coulomb-Breit
correlation is the same as for the second order Coulomb-Coulomb one. The con-
tribution obtained within the Dirac-Fock approximation when the Breit interaction
is mixed with the Coulomb interaction (also called Breit-RPA contributions) is
listed in Table II as '(DFB-DF)-Breit 1st'. The final results differ from [6] with
around 10 meV, which is probably due to the different treatment of the Breit-
RPA contributions. The RPA-diagrams are in [6] evaluated explicitly. It should be
remembered that Breit-Coulomb correlation as well as Breit-RPA are approxima-
tions of two-electron two-photon (or more) contributions which should ultimately
be calculated directly from QED. Small differences when the Breit-interaction is
treated in different ways indicates probably that the approximation as such should
not be trusted to more than these differences.
The fifth row in Table I corresponds to higher order correlation, which is calcu-
lated by RMBPT in an all order formulation, as described in [7, 8]. These values
were obtained as the difference between the all-order and the second order con-
tributions with the partial wave expansion up to emax = 8 evaluated in the DFB-
potential. The calculation by Johnson et at. is restricted to third order contributions
only.
The final values for the RMBPT calculations are given in Table I. The uncer-
tainty given comes from the I-extrapolation. Radiative contributions, self-energy
and vacuum polarization for the splittings in Pb 53+ have been calculated by Blun-
dell [1] including the so called screening of these effects, i.e., the calculation
accounts for the fact that the virtual photons are emitted and absorbed in the poten-
tial from both the nucleus and the core of electrons. In Table I we include results
by Indelicato [9] for each n = 4 level. The uncertainty of these results comes from
ACCURATE CALCULATIONS ON DIELECTRONIC RECOMBINATION RESONANCES 389
Table I/. The contributions to the second-order correlation energies in Pb 53 + (4S I/2),
Pb53+ (4PI/2) and Pb53+ (4P3/2), respectively, in a.u.
the difference with Blundell's calculation, mainly due to the difference in included
diagrams, and from the estimation of numerical errors given in [1].
As it has been shown above the main uncertainties in the 4pI/2 - 4S I / 2 and
4p3/2 - 4S I / 2 energy splittings arise due radiative contributions and due to two-
electron contributions which would require a full QED treatment. If fj.E in Equa-
tion (2) can be calculated with an accuracy within "'-'I meV for the recombination
390 M. TOKMAN ET AL.
resonances the 4pl/2 - 4S 1/ 2 splitting can be determined within the same accuracy
which would put a stringent bound on the QED contributions.
5. Conclusion
Narrow recombination resonances very closely above the ionization threshold in
highly charged ions open a possibility to determine energy splittings with an accu-
racy which can only be matched by calculations which treat radiative effects with
a precision beyond what is possible with methods available today.
ACCURATE CALCULATIONS ON DIELECTRONIC RECOMBINATION RESONANCES 391
Acknowledgements
The authors would like to thank P. Indelicato for communicating unpublished re-
sults. Financial support for this research was received from the Swedish Natural
Science Research Council (NFR). M. Tokman is grateful for support from the
Swedish Royal Academy of Science.
References
I. Blundell, S. A., Phys. Rev. A 47 (1993), 1790.
2. Indelicato, P. and Mohr, P. J., Theor. Chem. Acta 80 (1991), 207.
3. Lindgren, 1., Persson, H., Salomonson, S. and Ynnerrnan, A., Phys. Rev. A 47 (1993), R4555.
4. Baird, S. et al., Phys. Lett. B 361 (1995), 184.
5. Johnson, W. R. and SofT, G., At. Data and NucZ. Data Tables 33 (1985),405.
6. Johnson, W. R., Blundell, S. A. and Sapirstein, J., Phys. Rev. A 42 (1990), 1087.
7. Salomonson, S. and Oster, P., Phys. Rev. A 41 (1990), 4670.
8. Lindroth, E. and Hvarfner, J., Phys. Rev. A 45 (1991), 2771.
9. Indelicato, P., private communication.
10. Zong, W. et al., Phys. Rev. A 56 (1997), 386.
11. Lindroth, E., Hyp. Interact. 114 (1998), 219.
12. Mannervik, S. et at., Phys. Rev. Lett. 81 (1998), 313.
~ Hyperfine Interactions 132: 393-396, 200l.
,, 393
© 2001 Kluwer Academic Publishers.
1. Introduction
In energetic atomic collisions between highly charged high-Z ions and low-Z tar-
get atoms, radiative electron capture (REC) is one of the most important reaction
channels. In the limit of a loosely bound target electron, REC is identical with
radiative recombination. Reactions of this type have been extensively studied in
recent years for heavy highly charged projectiles up to bare uranium. The rel-
ativistic theory of REC in the one-electron approximation is well established at
present (see [1, and references therein]), and results of numerical calculations are
in excellent agreement with experiment [2]. While radiative recombination of an
electron with a bare nucleus is well understood theoretically, the process involving
an ion with several electrons is complicated by the interelectronic interaction. The
REC process into the L-shell of He-like uranium was measured in [3].
In this work, we present a systematic investigation of the interelectronic-inte-
raction effect on radiative recombination of an electron with a heavy helium-like
ion. In our system the number of the electrons (N = 3) is much smaller than
the nuclear charge number Z and, therefore, the interaction of the electrons with
each other is much smaller (by factor 1/ Z) than the interaction with the Coulomb
field of the nucleus. In zeroth approximation we can neglect the interelectronic
394 V. A. YEROKHIN ET AL.
interaction and consider that the electrons interact only with the Coulomb field
of the nucleus. So, to zeroth order the process under consideration is equivalent
to radiative recombination of an electron with a bare nucleus, studied thoroughly
in [1]. In the present work, we investigate the first-order (in l/Z) correction due to
the interelectronic interaction in this process.
the electron shell. This approximation was employed, e.g., in [3] for the theoretical
description of radiative recombination of an electron with a heliumlike uranium. In
what follows, we refer to that method as the effective-nuclear-charge approxima-
tion. The second method is slightly more sophisticated. In it, the influence caused
by the closed electron shell is approximated by a spherical symmetric screening
potential. While the screening potential can be chosen in different ways, we define
it as
V,er(x) = 2et{ ~ 1x
dy l[gis(Y) + II2,(y)] + 100
dy y[gfs(Y) + II2,(y)] }, (1)
where gls and lIs are the upper and the lower components of the radial wave func-
tion of the ground state, respectively. In this screening-potential approximation, the
Table I. Zeroth-order total cross section u(O) and the first -order interelec-
tronic-interaction correction in different evaluations. in barns. U ~~~ denotes the
interelectronic-interaction correction calculated in the effective-nuclear-charge ap-
proximation with parameter Zeff = 90.3, us~) corresponds to the screening-potential
approximation. ui~lt) indicates the results of the rigorous relativistic treatment
zeroth-order description can be used as well, with the nuclear potential modified by
Vscr(x). It can be shown that the definition (1) corresponds to taking into account
only the direct Coulomb part of the interelectronic interaction.
The numerical results for the interelectronic-interaction correction to the total
cross section of radiative recombination of an electron with He-like uranium are
presented in Table 1. The calculation is carried out in the laboratory frame for
capture into the 2s, 2Plj2, and 2P3j2 states of He-like uranium and for projectile
energies of 10-700 Me V per nuclear mass unit. The results of the rigorous relativis-
tic treatment are compared with the calculations based on the effective-nuclear-
charge approximation and on the screening-potential approximation. The com-
parison shows a decreasing accuracy of the approximate methods for increasing
projectile energy. They also yield better results for capture into excited states than
for the ground state. In average, the screening-potential approximation is found to
be more reliable than the effective-nuclear-charge approximation. Its typical devia-
tion from the rigorous treatment is about 10-20% of the interelectronic-interaction
correction, i.e., about 1-2% of the cross section of the process.
In summary, we have carried out a systematic investigation of the effect of the
interelectronic interaction on the process of radiative recombination of an electron
with He-like uranium. The applicability of frequently used approximate methods
was studied and compared with the rigorous relativistic treatment. The screening-
potential approximation, in which an interaction of an electron with a He-like ion is
replaced by an interaction with a modified potential, is shown to be a reliable tool
for estimating both total and differential cross sections of the process for projectile
energies far from resonance. Still, full relativistic calculations are needed to obtain
an accuracy better than a few percent of the cross section of the process.
Acknowledgements
We want to thank Th. Stbhlker for valuable discussions. Financial support by the
DFG (Grant No. 436 (RUS) 113/479) and by the RFBR (Grant No. 98-02-04111) is
gratefully acknowledged. Th. Beier acknowledges also support from the EU-TMR
programme (contract No. ERB FMRX CT 97-0144).
References
1. Eichler, J. and Meyerhof, W., Relativistic Atomic Collisions, Academic Press, San Diego, 1995.
2. StOhlker, Th., Ludziejewski, T., Bosch, F., Dunford, R. W, Kozhuharov, C., Mokler, P. H.,
Beyer, H. F., Brinzanescu, 0., Franzke, F., Eichler, J., Griegal, A., Hagmann, S., Ichihara, A.,
Kramer, A., Liesen, D., Reich, H., Rymuza, P., Stachura, Z., Steck, M., Swiat, P. and
Warczak, A., Phys. Rev. Lett. 82 (1999), 3232.
3. StOhlker, Th., Geissel, H., Irnich, H., Kandler, T., Kozhuharov, C., Mokler, P. H., Miinzen-
berg, Nickel, F., Scheidenberger, C., Suzuki, T., Kucharski, M., Warczak, A., Rymuza, P.,
Stachura, Z., Kriessbach, A., Dauvergne, D., Dunford, B., Eichler, J., Ichichara, A. and
Shirai, T., Phys. Rev. Lett. 73 (1994), 3520.
4. Yerokhin, V. A., Shabaev, V. M., Beier, T. and Eichler, J., Ph),s. Rev. A 62 (2000), 042712.
Hyperfine Interactions 132: 397-400, 2001. 397
© 2001 Kluwer Academic Publishers.
Abstract. We report on the recent evaluation of the two-photon electron self energy to all orders
in the interaction with the Coulomb field of the nucleus. With the present results at hand the major
theoretical uncertainty is diminished, which provides predictions of the ground-state energy with a
relative accuracy of about 10- 6 for the hydrogenlike uranium and lead systems. This allows for
high-precision tests of quantum electrodynamics (QED) in strong fields that are expected to be
experimentally available in the near future.
Key words: atomic binding energy, highly charged ions, quantum electrodynamics.
An ideal scenario to test quantum electrodynamics (QED) in the strong field limit
is provided by the strong electric field of the nucleus in highly charged ions, e.g.,
by measurements of the Lamb shift at utmost precision. Therefore, at the SISIESR
facilities in Darmstadt one is aiming for an accuracy of about 1 eY in measurements
of the ground-state Lamb shift for hydrogenlike uranium in the near future [1].
Theoretical evaluations on the same level of accuracy require calculations of the
complete set of radiative corrections of order a 2 (a is the fine structure constant)
but to all orders in the coupling constant Za to the Coulomb field of the nu-
cleus. The set of these second-order diagrams includes all various combinations of
the first-order self-energy (SE) and vacuum-polarization (YP) graphs. The present
status of the theoretical predictions for the Lamb shift in different one-electron
ions is presented in [2]. Most of these diagrams have already been calculated in
recent years. However, calculations of the most difficult set, the second-order self-
energy correction (SESE) are yet incomplete. In this paper we report the present
status of this challenging theoretical problem for the most interesting cases of the
hydrogenlike uranium and lead systems.
The general renormalization scheme for the two-photon self energy was con-
sidered in more detail in [3]. The loop-after-Ioop diagram SESE (a) can be divided
into an irreducible and a reducible part. Por an elegant way deriving the corre-
sponding energy shifts in reducible as well as irreducible diagrams of bound-state
QED we refer to the two-times Green-function method [4]. The irreducible part
398 I. GOIDENKO ET AL.
P EP
A A
A E SESE
ren
- -Lnl'A
+
P 8
BE
A A
p-p
A
A
+ + _ 1:(1)
A A
additional
SESE b) SESE c) countertenn
Figure I. The graphical representation of the partial wave renormalization approach. The bar
in the irreducible part denotes the exclusion of the state A in the sum over the intermediate
states n. The double and ordinary solid lines with the cross denote the quadratic denominators
in the bound and free electron propagators. The triangles represent Fourier expansion of the
bound state A wavefunction into free electron states.
SESE (a) (irred) is invariant under covariant gauges [5] and therefore this part can
be renonnalized and calculated separetely. In [6] the corresponding energy shift
of the SESE (a) (irred) diagram has already been evaluated for the nuclear charge
numbers Z = 70, 80,90 and 92, and in [7, 8] for arbitrary nuclear charge numbers
in the range 3 ~ Z ~ 92. In the high-Z limit the results of [6, 7] and [8] are
in coincidence with each other (~E~;SE(a)(irred) = -0.97 eV for uranium). But in
the low-Z limit a disagreement has been obtained. This discrepancy is a subject of
several controversial statements made in a series of subsequent papers [8-12]. For
more details on this point we refer also to [13]. We would like to emphasize that
while the discrepancy between the different calculations of the SESE (a) (irred)
contribution for low-Z values still needs to be resolved, for the high-Z region all
the calculations [6-8] give identical results.
Now we shall consider the SESE (a) (red) as well as the SESE (b), (c) di-
agrams. The renormalized expression for the SESE-contribution is depicted in
Figure I. In the Feynman gauge only the sum of the graphs in the renonnalized
expression in Figure 1 is ultraviolet as well as infrared convergent. Therefore these
diagrams should be calculated together. According to the partial wave renormal-
ization method r14, 15] all single terms, including the mass counterterms, are
decomposed in partial waves. Accordingly, the renormalized expression of the two-
photon electron self energy can be expressed as a double sum over two partial
EVALUATION OF THE TWO-PHOTON SELF-ENERGY CORRECTION 399
U 12 =0 12 = I 12 =2 12 =3 Pb 12 =0 12 = I 12 =2 12 =3
I] =0 0.699 0.384 0.106 -0.059 I] =0 0.439 0.158 0.051 -0.04
I] = I 0.384 -0.188 0.563 I] =1 0.186 -0.092 0.228
I] = 2 -0.427 -0.376 I] =2 -0.135 -0.124
I] = 3 0.501 I] =3 0.172
waves I] and h:
=L L
00 00
The decisive advantage of the PWR approach is that every single partial wave is
already UV- as well as IR-finite.
For the ground state IA) = lIs) of uranium and lead we were able to compute
4 partial waves I], I2 = 0, 1,2,3 with the limitation I] + l2 ~ 3 for the sum.
The individual terms ,6,Ei~I.l2) of the double partial-wave expansion for uranium
(Z = 92) and lead (Z = 82) ions are listed in Table I.
The inaccuracy of our calculations is determined by the unstability of the nu-
merical results with the change of the number of grid points within the B-spline ap-
proximation [16] for the solution of the radial Dirac equation from N = 23 to N =
46. We estimate this inaccuracy as 12%. The final value ,6,E~;SE(a)(red),(b),(c)(ren) was
obtained by an extrapolation from the numbers given in Table I. Accordingly, we
evaluated the accumulated sums
1]+12=1
SI = L ,6,Ei~lh) (2)
1],12
It is interesting to note that the so-called sign approximation [17] gives already
~40% of the exact result. The limit of the number of partial waves was set by
400 I. GOIDENKO ET AL.
the extremely large computer time required. The calculations were performed at
the computer center of the Technical University of Dresden on the CRAY-T3E
supercomputer with 32 parallel processors. The inclusion of 4 partial waves 11, 12 =
0, I, 2, 3 in both partial wave expansions with the limitation 1\ + 12 (; 3 required
more than 20 thousand single-processor CPU hours for each ion (U,Pb). The in-
accuracy assigned to our results for SESE (a) (red) and SESE (b), (c) corrections
remains the main source of the total error in the theoretical Lamb shift predic-
tion. We expect that the inaccuracy can be substantially diminished within the
framework of the method described above.
Acknowledgements
I. G., L. L., and A. N. are grateful to the Technische Universitat Dresden and
the Max-Planck-Institut ftir Physik komplexer Systeme (MPI) for the hospitality
and for financial support from the MPI, DFG, and RFBR (grant no. 99-02-18526).
G. S., G. P., and S. Z. acknowledge financial support from BMBF, DFG, and GSI.
References
1. SWhlker, T. et al., The Is lamb shift in hydrogenlike uranium measured on cooled. decelerated
ion beams, preprint (2000), submitted to Phys. Rev. Lett.
2. Mohr, P., Plunien, G. and Soff, G., Phys. Rep. 293 (1998), 227.
3. Labzowsky, L. N. and Mitrushenkov, A. 0., Phys. Rev. A 53 (1996), 3029.
4. Shabaev, V. M., Phys. Rev. A 49 (1994), 4489;
Shabaev, V. M. and Fokeeva, I. G., Phys. Rev. A 50 (1994), 4521.
5. Blundell, S. A., Phys. Rev. A 47 (1993), 1790.
6. Mitrushenkov, A. 0., Labzowsky, L. N., Lindgren, I., Persson, H. and Salomonson, S., Phys.
Lett. A 200 (1995), 51.
7. Mallampalli, S. and Sapirstein, J., Phys. Rev. Lett. 80 (1998), 5297.
8. Goidenko, I. A., Labzowsky, L. N., Nefiodov, A. v., Plunien, G. and Soff, G., Phys. Rev. Lett.
83 (1999), 2312.
9. Yerokhin, V. A., Phys. Rev. 62 (2000), 012508.
10. Eides, M. A., Grotch, H. and Shelyuto, V. A., Theory of light hydrogen like atoms, hep-
phl0002158.
11. Manohar, A. V. and Stewart, I. W., Logarithms of IX in QED bound states from the renormal-
ization group, hep-phl0004018.
12. Goidenko, I. A., Plunien, G., Nefiodov, A., Zschocke, S., Labzowsky, L. N. and Soff, G.,
No regularization corrections to the partial-wave renormalization procedure, hep-phl0006220;
to appear in Phys. Rev. A (2000).
13. Goidenko, I. A., Labzowsky, L. N., Nefiodov, A. V., Plunien, G., Zschocke, S. and Soff, G.,
In: Proc. to the Satellite Meeting afthe 10th ICAP COIlf., Florence, Italy, 1-3 June 2000.
14. Persson, H., Lindgren, I. and Salomonson, S., Phys. Scripta T46 (1993), 125;
Lindgren, I., Persson, H., Salomonson, S. and Ynnerman, A., Phys. Rev. A 47 (1993), R4555.
15. Quiney, H. M. and Grant, I. P., Phys. Scripta T46 (1993), 132; 1. Phys. B 27 (1994), L299.
16. Johnson, W. R., Blundell, S. A. and Sapirstein, J., Phys. Rev. A 37 (1988),307.
17. Goidenko, I., Labzowsky, L., Nefiodov, A., Plunien, G., Soff, G. and Zschocke, S., Hyp.
Interact. 127 (2000), 293.
Hyperjine Interactions 132: 401-407,2001. 401
© 2001 Kluwer Academic Publishers.
P. H. HEENEN
Service de Physique Nucleaire Theorique, U.L.B.-C.P.229, 8-1050 Brussels, Belgium
Abstract. Several methods based on effective interactions or Lagrangians are available today. Al-
though different in many respects (use of zero range or finite range interactions, relativistic or non
relativistic framework, different treatments of pairing correlations), their applications to nuclei far
from stability have shown converging results which still have to be incorporated in macroscopic
approaches. Many efforts are also actually devoted to the improvements of the effective interactions,
especially of the pairing force. Finally, developments are performed to include in a microscopic
framework correlations beyond a mean-field (in particular, the correlations generated by rotation and
vibration in the deformed nuclear potential). I shall review some key aspects of these developments
and show how they affect the determination of nuclear masses in particular at the limits of stability.
1. Introduction
Thanks to the developments of new techniques and to the increase of computa-
tional power, mean-field methods based on effective 2-body interactions are now
systematically applied to the studies of a wide range of nuclear properties. In this
talk, I will limit myself to a review of the methods whose only phenomenological
ingredient is an effective nucleon-nucleon interaction or an effective Lagrangian
(in the case of the relativistic mean field approach). Moreover, I will consider
effective forces adjusted on a few specific properties of nuclear matter and of finite
nuclei and are not modified according to the region of the nuclear chart which is
studied. The method set up by Pearson et al. (see [1] and the talk of Pearson at this
conference) enters into this category but since it is discussed in another talk, I will
not review its specificities. Rather, I shall focus on methods which have not been
specifically designed to reproduce nuclear masses and which have been applied to
a wide range of nuclear structure problems.
2. The methods
The microscopic mean field methods can be classified according to the way the
interactions between the nucleons is treated. The general framework can be either
non relativistic or relativistic. In the first case, the nucleon-nucleon interaction can
be either zero range or finite range.
402 P.H.HEENEN
Two different languages are used to describe zero range nucleon-nucleon in-
teractions. The most widely used is based on the Skyrme force. After the first
developments by Vautherin and Brink [2], it has been applied to a large variety of
problems already in the 70's (for a review, see [3]). The microscopic mass formula
of Pearson et ai. [1] belongs to this family, with the 10 parameters of the Skyrme
interaction adjusted with a constraint on the reproduction of the masses of known
nuclei. There are now several parametrizations of the Skyrme force available on
the market. One of the most recent systematic study has been performed by Cha-
banat et ai. [4]. The set of parametrizations that he has introduced are called SLy
followed by a number. It differs from previous adjustments by a very careful study
of nuclear and neutron matters, with the hope to develop an interaction valid also
for applications in astrophysics. The main constraints that have been used to built
the SLy forces are:
• a good reproduction of the saturation point of symmetric infinite nuclear mat-
ter (E/A around -16 MeV, ro around 0.16 fm- 3 ),
• a compression modulus of nuclear matter around 230 MeV,
• a symmetry energy around 32 MeV,
• a constraint on the reproduction of a fundamental equation of state of pure
neutron matter,
e
• the binding energies and rms radii of doubly magic nuclei 6 0, 40.48Ca, 56Ni,
132Sn,208Pb).
The spin orbit interaction has been adjusted on 208Pb and not on \60 as for the
interactions which were determined in the 70's and the 80's.
Several groups perform systematic applications of the HF + BCS or HFB meth-
ods with Skyrme forces. The codes used differ mainly by the symmetries which
are broken: shape symmetries (to treat axial or triaxial quadrupole deformations
or octupole deformations) or time reversal invariance (to describe odd and rotating
nuclei).
The second way to introduce zero range interactions is used by Fayans and
coworkers (see [5]). It is based on the Landau theory of Fermi liquids. Although
the codes developed by Fayans group are limited to spherical symmetry, the appli-
cations that have been performed have been put into evidence some important key
properties of effective interactions.
The systematic use of finite range interactions is limited to the use of the Gogny
force [6]. Although very time consuming, the Gogny interaction has been applied
to describe nuclei with the same flexibility to break symmetries than the Skyrme
forces.
The most recent mean-field methods are based on relativistic Lagrangians (see,
for instance, [7]). The relativistic mean-field introduced by Ring and coworkers
is a Hartree method (the exchange terms are thus neglected). The non-relativistic
.. ,it has been shown to be very close to a conventional HF method with a Skyrme
force. One of the advantage of the RMF method is that the spin orbit is introduced
naturally and is not put by hand. The non-relativistic reduction of the spin orbit
MASS PREDICTIONS FROM MEAN-FIELD CALCULATIONS 403
term has been shown to have an isospin dependence different from the conven-
tional spin orbit parametrization used in both Gogny and Skyrme interactions. This
isospin dependence has been introduced in some new Skyrme parametrizations.
However, one must note that the spin orbit derived from a relativistic Hartree-
Fock method would still be different. Unfortunately, there are presently only a very
limited number of applications of the relativistic HF method.
The Gogny interaction is the only one to be used to describe at the same time
mean-field and pairing correlations. This is due to the fact that a zero range inter-
action like the Skyrme force has an unphysically large strength for high particle
momenta and that up to now, the pairing interaction has only been introduced non
relativistically in RMF calculations.
The most commonly pairing interaction used with a Skyrme force nowadays is
a zero range density dependence force:
Vp = -(1 (PCrl») (_ _)
Vo - PrJ) 1 - - - 8 rl - r2 . (2.1)
2 Pc
Vo and Pc are two parameters and P (r) is the total local single-particle density in
coordinate space. In order to avoid divergences, the definition of the force involves
also an energy cut-off parameter in the valence single-particle space to limit the
active pairing space above the Fermi level.
The parameter Pc determines the spatial dependence of pairing. In the limit
where Pc goes to infinity, we recover the usual delta force. In this case, the matrix
elements of Vp depend on properties of the wave functions in the whole volume
of the nucleus. This leads to volume active pairing fields. On the other hand, for
Pc equal to the nuclear saturation density Po, the interaction is weak inside the nu-
cleus and strongly attractive outside, leading to a surface active pairing field. In the
Landau-Migdal approach of Fayans et al. [5], the interaction in the pairing channel
contains additional terms with derivatives of the nuclear density. The introduction
of these density dependent terms is important to describe correctly the isotopic
dependence of charge radii around closed shells.
The pairing energy is a small fraction of the total energy of a nucleus (of the
order of 10 Me V to 1 Ge V). It is, therefore, hard to find a simple quantity which
depends mainly on pairing correlations. The usual method to define such quantities
is to perform energy differences between odd and even nuclei. Satula et al. [8] have
recently reanalyzed the adjustment of pairing and compared different combinations
of energies of even and odd nuclei to HF + BCS calculations. They showed that
the best indicator of pairing seems to be:
1
,6.3(N) = --(B(N -1) + B(N + 1) - 2B(N») (2.2)
2
at odd N values and that the fourth order difference ,6.4(N), currently used, is much
more sensitive to mean-field effects. A consequence of their analyzes is that they
obtain a ,6. nearly constant with mass and not the usual 12/ A 1/2 law.
404 P.H.HEENEN
3. Applications
There have not been many systematic calculations of nuclear masses starting di-
rectly from effective two-body interactions. Patyk et al. [9] have tested the Gogny
force, several Skyrme forces and relativistic Lagrangians on 116 even-even spher-
ical nuclei. The best results give rms deviations with the experimental data around
2 MeV, without large differences between the methods. This is significantly larger
than the deviation around 0.7 MeV found by Pearson et al. [1] with a force adjusted
specifically to reproduce masses. It is hard to trace the origin of the difference
between Pearson's result and the other calculations. One obvious difference is
related to the value of the effective mass which is equal to the nucleon mass in the
Pearson et ai. adjustment. This mass is significantly smaller in all other cases, as a
result of the other constraints imposed on the interactions. It still remains to check
whether with this choice, satisfactory results can be obtained for the description
of other phenomena, like the super deformed rotational bands known in different
regions of the mass table and for which several other mean field interactions are
very successful.
The mean-field calculations of energy differences are in much better agreement
with data than masses. In Figure 1 are plotted the Qa energies as a function of the
neutron numbers for several super heavy nuclei [10]. The agreement between the
118
14 SLy4
13
12
11 Hs
-> 10
<J.)
6 9
(j
a 8
7
6 Fm
5 Cf
• exp
4 o th Cm
== Qa. .
912' [6(4) ~-o.14
W;1: [~l
312+ t61 d =IO. 16 MeV
112+ [6111 285 Exp: 9.89(5) MeV
112173 30
~=0. 1 7
281 11°171
1112+ [606J Q =9 3M V
312' [611 J ~=0.2 1 al - .4 e
1312' [71 6] - - Q.a=9.2SMeV
§~: 1~1 :::: Exp: 9.0 1 (1~) MeV
1m t61 1 m l08'69 1.6 min
Figure 2. Predicted structure of the lowest one-quasiparticle excitations in the 2~l} 114 a-decay
chain. Each excitation is characterized by the Nilsson quantum numbers of the odd neu-
tron. The deformations of each isotopes is also given, together with the Qa value for the
ground-state to ground-state transition and the experimental Qa value (from [10]).
Skyrme HFB calculation and the experimental data is excellent, giving confidence
in the ability of mean field calculations to predict Qa energies in unknown regions
of the nuclear chart.
In Figure 2 are given the spectra of odd nuclei along the chain of alpha emission
starting at the Z = 114 isotope recently detected at Dubna [11]. Here also the ener-
gies of the Qa particles are correctly predicted by the calculation. In this case, the
odd nuclei are calculated fully self consistently, a feature which requires a specific
calculation for each of configuration of each isotope. This feature is necessary to
determine which state in each nucleus is the most favorable one for the emission of
the a particle, some low lying states being excluded by a spin mismatch with the
daughter nucleus.
4. The future
Several methods have been developed to introduced correlations beyond a mean
field approach. Valor et ai. [12] (for Skyrme interactions) and Egido et af. [13]
(for the Gogny force) have very recently introduced similar methods which extend
largely the range of applications of mean-field methods and should be extensively
used in the next years. The principle of the methods is the same and is illustrated
in Figure 3 in a calculation of 32Mg with a Skyrme interaction.
First, mean field wave functions are generated with a constraint on a collective
variable (the axial quadrupole moment in this case). The variation of the energy of
32Mg as a function of q is represented by a dotted curve in Figure 3. The minimum
of the energy is obtained for the spherical configuration. A shallow secondary
406 P.H. HEENEN
-235
-240
..
:> -245
~
w -250
-260
o 50 100 150 200 250 300 350
Q(fm2)
Figure 3. Mean-field and projected energies obtained for 32Mg as a function of the axial
quadrupole moment qo. The pairing interaction is a zero range interaction without density
dependence. The Skyrme parameterization is SkM*. The dotted line correspond to the HF +
BCS + LN energies. The energies obtained by projecting on angular momentum the intrinsic
wave functions are plotted in full lines for 0+,2+ and 4+, in ascending order, as a function of
the quadrupole moment of the intrinsic wave function. The energies obtained for each angular
momentum in the configuration mixing calculation are represented by horizontal bars centered
at the value of qO, where the respective collective wave functions are maximum.
Acknowledgement
This research was supported in part by the PAI-P3-043 of the Belgian Office for
Scientific Policy.
References
1. Tondeur. F.. Goriely. S .. Pearson. 1. and Onsi. M .. Phys. Rev. C 62 (2000). 24308.
2. Vautherin. D. and Brink. D. M .• Phys. Rev. C 5 (1972). 626.
3. Aberg, S .• Flocard. H. and Nazarewicz. w.. Ann. Rev. NucZ. Part. Sci. 40 (1990). 439.
4. Chabanat. E .• Bonche. P.. Haensel. P.• Meyer. 1. and Schaeffer. R .. Nuclear Phys. A 635 (1998).
231.
5. Fayans. S. A .• Tolokonnikov. S. v.. Trykov. E. L. and Zawischa. D .. Nuclear Phys .• in press.
6. Decharge. 1. and Gogny. D .• Phys. Rev. C 21 (1980). 1568.
7. Ring. P.• Progr. Part. NucZ. Phys. 37 (1996). 193.
8. Satula. W.• Dobaczewski. 1. and Nazarewicz. W.• Phys. Rev. Lett. 81 (1999). 3599.
9. Patyk. Z .• Baran. A .• Berger, 1. F., Decharge, 1.. Dobaczeswki. 1.. Ring, P. and Sobiezewski. A..
Phys. Rev. C 59 (1999). 704.
10. Cwiok, S., Nazarewiez. W. and Heenen, P.-H., Phys. Rev. Lett. 83 (1998), 1108.
11. Oganessian, Y. T. et aI., Phys. Rev. Lett. 83 (1999). 3154.
12. Valor, A., Heenen, P.-H. and Bonehe. P., Nuclear Phys. A 671 (2000). 145.
13. Rodriguez-Guzman. R .• Egido. 1. L. and Robledo. L. M .. Phys. Lett. B 474 (2000), 15.
14. Bonche, P.. Dobaezewski. 1.. F1ocard. H.. Heenen, P.-H. and Meyer. 1., Nuclear Phys. A 510
(1990). 466.
Hyperfine Interactions 132: 409-416,2001. 409
© 2001 Kluwer Academic Publishers.
TAKAHARU OTSUKA
Department of Physics, University of Tokyo. Hongo, Bunkyo-ku, Tokyo 113-0033, Japan;
e-mail: otsuka@phys.s.u-tokyo.ac.jp
Abstract. The nuclear mass calculation is discussed in terms oflarge-scale shell model calculations.
First, the development and limitations of the conventional shell model calculations are mentioned.
In order to overcome the limitations, the Quantum Monte Carlo Diagonalization (QMCD) method
has been proposed. The basic formulation and features of the QMCD method are presented as well
as its application to the nuclear shell model, referred to as Monte Carlo Shell Model (MCSM). The
MCSM provides us with a breakthrough in shell model calculations: the structure of low-lying states
can be studied with realistic interactions for a nearly unlimited variety of nuclei. Thus, the MCSM
can contribute significantly to the study of nuclear masses, An application to N ~ 20 unstable nuclei
far from the J'i-stability line is mentioned.
1. Introduction
The nuclear shell model was conceived by Mayer and Jensen in 1949 [1] as a
single-particle model. Afterwards, many valence particles are treated in the shell
model, which then became a many-body theory or calculational method. A good
example can be found in the sd shell [2]. The nuclear shell model has been suc-
cessful in the description of various aspects of nuclear structure, partly because it
is based upon a minimum number of natural assumptions, and partly because all
dynamical correlations in the model space, beyond the mean-field calculations, can
be incorporated appropriately. Although the direct diagonalization of the Hamil-
tonian matrix in the full valence-nucleon Hilbert space is desired, the dimension
of such a space is too large in many cases, preventing us from performing the full
calculations.
In order to overcome this dimension problem, quantum Monte Carlo approaches
have been introduced. As one of them, the Shell Model Monte Carlo (SMMC)
method has been proposed successfully [3]. However, the SMMC is basically re-
stricted to ground-state and thermal properties. Moreover, the SMMC suffers from
the so-called minus sign problem for realistic interactions [3, 4]. As a completely
different approach, the Quantum Monte Carlo Diagonalization (QMCD) method
has been proposed for solving quantum many-body systems with a two-body in-
teraction [5-8]. The QMCD can describe not only the ground state but also excited
states, including their energies, wave functions and hence transition matrix ele-
410 T. OTSUKA
ments, through which one can examine the validity of the effective interaction.
Thus, based upon the QMCD method, we introduce the Monte Carlo Shell Model
as its application to the nuclear shell model calculation. The MCSM has become a
new tool for clarifying the structure of the ground and low-lying states of nuclei,
providing us with a quite powerful tool for calculations of the nuclear mass.
*****
**** o
***
H = ** diagonalization
o
Conventional Shell Model
all Slater determinants
diagonalization
all matrix elements are calculated, one should diagonalize the matrix. A schematic
description of this process is shown in the upper part of Figure 1. So, if the shell
model dimension is large, the number of matrix elements is certainly much larger
(up to about half of the square of the dimension), and the actual calculation be-
comes very difficult. By recent (conventional) shell-model codes like ANTOINE
by Caurier, VECSSE by Sebe or MSHELL by Mizusaki, one can handle up to
shell model dimension ~ 100 million in extreme cases, while practical calculations
up to a few ten million can be accomplished. Recent examples can be found in
[9, 10], where the levels of 52Fe and 50Mn are calculated by the codes ANTOINE
and VECSSE, respectively.
Although the conventional shell model calculation has thus been developed
significantly, the required dimension can be much larger in many real nuclei and
is indeed much beyond the reach of the future development. For instance, certain
unstable nuclei to be discussed in this paper require calculations with more than
10 billion dimension. This is already very far beyond the limit of the existing shell
model codes.
(3.1 )
where CJ means a set of random numbers which control the basis generation, 1<P (CJ))
is a basis state, 1\11(0)) is the initial state. Here h(CJ) is so-called one-body Hamil-
412 T. OTSUKA
drastic excitations within a nucleus. For instance, one can describe spherical yrast
states, deformed rotational bands and nearly superdeformed bands with the same
Hamiltonian in the same model space. An example can be found in 56 Ni where
those three kinds of states are nicely presented by an MCSM calculation [12] with
the full pf shell and the g9/2 orbit. This kind of description over a wide variety
of states may be characterized as the feasibility along the energy axis. The yrast
states and (normally) deformed band can be described basically within the pf shell
[8, 13], while the g9/2 orbit is needed for the description of more deformed states.
The second major advantage of the MCSM calculation is the feasibility of
handling many valence particles. The maximum number of valence particles is
rather limited in conventional calculations. However, if one wants to describe a
long chain of isotopes entering the region of exotic nuclei far from the ,B-stability
line, the number of particles should change significantly. So, this capability plays
an indispensable role in studying the structure of such exotic nuclei. This feature
can be characterized as the feasibility along the isospin axis in the sense that one
can study nuclei with large variation of isospin due to neutron or proton excess,
and will be discussed in the next section.
20
15
"""'
>(l)
~ 10
'--'
=:
C')~
20 22 24
N
Figure 2. Two-neutron separation energies in the N ~ 20 region. The filled (open) squares,
triangles, diamonds, and circles denote the experimental (calculated) values for 0, Ne, Mg
and Si isotopes, respectively. Experimental values are taken from [15, 16].
414 T. OTSUKA
15
,--...,
10
:>
(l)
~
'-"
s:: 5
~(";j
0
100
(b)
80 OpOh
,--...,
~ 60
'-"
.D
0 40
I-<
~ . . ....4p4h
........
20 ......................
.... --~.-:-.~-
08
10 12 14
z
Figure 3. (a) Two-neutron separation energies of N = 20 isotones. Triangles denote the
experimental values in [IS], while the square for 30Ne is a recent datum [16]. Solid and
dashed lines correspond to the full and OpOh calculations [17] with the present interaction,
respectively. The cross for 29F means the OpOh + 2p2h truncation. (b) Probabilities of OpOh,
2p2h and 4p4h configurations in the ground state of N = 20 isotones, indicated by dotted,
solid and dashed lines, respectively.
state can be of quite exotic nature. In this situation, the above two feasibilities
combined together play really crucial roles in clarifying the structure of exotic
nuclei far from the ,B-stability line. As an example, we shall discuss the structure
of nuclei in the vicinity of 32Mg.
Figure 2 indicates two-neutron separation energies of even-A 0, Ne, Mg and Si
isotopes obtained by MCSM calculation [14] compared with experiment [15, 16].
The calculated results reproduce experimental trends rather nicely. This calculated
result is an aspect of a systematic calculation for N ~ 20 exotic nuclei [14].
Figure 3 indicates some details of the structure of the ground state of unstable
nuclei with N = 20. The two-neutron separation energy (S211) in Figure 3 decreases
as Z decreases. This trend is quite natural, because the nucleus moves away from
the stability line in this case as Z decreases. The S2n shows considerable differences
between the OpOh and fully mixed calculations. The latter can be done only by the
MCSM systematically, because the dimension can be extremely large in conven-
MONTE CARLO SHELL MODEL MASS PREDICTIONS 415
6. Summary
The Monte Carlo Shell Model (MCSM) is very useful and relevant to the study
of the nuclear mass. There have been a systematic calculation for N ......, 20 exotic
nuclei, where a number of important characters have been clarified through the
anomalous features ofthe nuclear mass. We are now working on another systematic
study on pf-shell nuclei, and have succeeded in reproducing the nuclear mass for
84 nuclei in the middle part of the pf shell with the accuracy of 134 keV. Thus, the
calculation of the nuclear mass will be advanced significantly in the next decade.
Acknowledgements
The MCSM calculations were performed in part by the Alphleet computer in
RIKEN. This work was supported in part by Grant-in-Aid for Scientific Research
(A)(2) (10304019) from the Ministry of Education, Science and Culture.
References
1. Mayer, M. G., Phys. Rev. 75 (1949), 1969;
Haxel, 0., Jensen, J. H. D. and Suess, H. E., Phys. Rev. 75 (1949), 1766.
2. Brown, B. A. and Wildenthal, B. H., Ann. Rev. Nucl. Part. Sci. 38 (1988), 29.
3. Koonin, S. E., Dean, D. J. and Langanke, K., Phys. Rep. 278 (1997),1 and references therein.
4. Otsuka, T., In: H. Rebel, D. Poenaru and J. Wentz (eds), Proc. of the Advanced Study [nst. 2000,
Predeal, Kluwer Academic, Dordrecht, 2001.
5. Honrna, M., Mizusaki, T. and Otsuka, T., Phys. Rev. Lett. 75 (1995),1284.
6. Mizusaki, T., Honrna, M. and Otsuka, T., Phys. Rev. C 53 (1996),2786.
7. Honrna, M., Mizusaki, T. and Otsuka, T., Phys. Rev. Lett. 77 (1996), 3315.
8. Otsuka, T., Honma, M. and Mizusaki, T., Phys. Rev. Lett. 81 (1998), 1588.
9. Ur, C. A. et al., Phys. Rev. C 53 (1996), 2786.
10. Schmidt, A. et aI., Phys. Rev. C 62 (2000), 044319.
416 T.OTSUKA
11. Otsuka, T., Mizusaki, T. and Honma, M., 1. Phys. G 25 (1999), 699.
12. Mizusaki, T., Otsuka, T., Honma, M. and Brown, B. A., In: T. Otsuka, H. Sakurai and 1. Tanihata
(eds), Proc. of RIKEN Symp. Shell Model 2000, Nuclear Phys. A (2001).
13. Mizusaki, T., Otsuka, T., Utsuno, Y., Honma, M. and Sebe, T., Phys. Rev. C 59 (1999), R1846.
14. Utsuno, Y., Otsuka, T., Mizusaki, T. and Honma, M., Phys. Rev. C 60 (1999),054315.
15. Audi, G. et al., Nuclear Phys. A 624 (1997), 1.
16. Sarazin, F. et al" Phys. Rev. Lett. 84 (2000), 5062.
17. Utsuno, Y., Otsuka, T., Mizusaki, T. and Honma, M., Phys. Rev. C 64 (2001), 011301(R).
Hyperfine Interactions 132: 417-424,200l. 417
© 2001 Kluwer Academic Publishers.
Abstract. The rp-process is thought to result from explosive stellar conditions of high temperature
and density typically in an accreting binary star system. These conditions favor the successive proton
captures leading to the synthesis of heavier nuclei beyond 56 Ni and may provide an explanation for
the observed high abundances of some p-nuclei. Simulations of this process are an important com-
ponent of understanding the pathways and the resulting abundances of the process. The most crucial
input parameters include nuclear mass differences, tJ-decay half-lives, deformations, and especially
reaction rates. Here is a brief description of a semi-empirical way of predicting the properties of
nuclei far from stability that may lie on the rp-process path.
1. Introduction
Thennonuclear runaways on the surface of neutron stars were suggested to be
responsible for the observed type I X-ray bursts [1-3]. The explosive conditions
resulting from the thennonuclear runaway are thought to tip the balance between
rapid sequences of successive proton captures and the subsequent ,B-decays re-
sulting in the creation of nuclei far beyond 56Ni all the way to the proton rich
regions [4] of the chart of nuclides. The explosion and the ensuing thennonu-
clear runaway are the results of hydrogen ignition at high temperature (1.9 OK)
and density conditions (1.1 x 106 g/ cm3 ) making proton and two-proton capture
rates favorable with respect to the competing ,B-decay. A possible site where the
rp-process may be taking place is on X-ray bursting binary star systems. All of
the observed X-ray bursting systems involve neutron stars with accretions from a
nearby companion star of hydrogen and helium rich material at a rate of 10-9 solar
masses per year. A video [5] of the motion of the two such stars can be viewed on
line at http://universe . gsfc. nasa. gov/videos/millisecond. html.
X-ray bursts are observed frequently [5], yet the nucleosynthesis and the corre-
lated energy generation is not completely understood. In addition to the uncertainty
of the astrophysical conditions, network simulations of the rp-process are hindered
by the lack of experimental infonnation on the structure and reaction rates of nuclei
along the rp-process path [4, 6]. Nuclei of particular interest to the rp-process are
418 A. APRAHAMIAN ET AL.
the N Z waiting point nuclei. These are the nuclei where further p-capture
typically leads to a p-unbound isotone or where the Q value for the (p, y) reac-
tion is low making the reverse reaction possible. In order for the rp-process to go
on, it needs to wait for ,B-decay or for the 2p-capture to jump over the waiting
point. Some of the waiting point nuclei above the Ni and Fe regions to the next
closed shell include; 68Se, 76Sr, 80Zr, 84Mo, 88Ru, 92Pd and 96Cd. The time spent
at these N = Z waiting points is strongly dependent on the photodisintegration
rates, the ,B-decay lifetimes, and the masses of nuclei along the path. Important
questions remain on the rp-process pathway and the endpoint of the process. It
has recently been shown [7] that the process may in fact end in the Te/Sn region.
The rp-process may make a significant contribution to the unexplained high solar
system abundances of some p-nuclei if an ejection mechanism can be found for the
burned material [7].
This paper deals primarily with the development of a semi-empirical approach
to the prediction of nuclear deformations and nuclear masses far from stability us-
ing the NpNn-parametrization [8] and the P-parametrization [9] where information
about known nuclei can be used to predict the structure of the ones far from stabil-
ity. There are several models available for reliable calculation of nuclear properties
near stability. Typically, a large divergence is seen between the various models
for nuclei far from stability. Our approach is inherently an interpolation between
known nuclei rather than an extrapolation. The NpNn parametrization yields infor-
mation on deformation of nuclei. The deformation in turn has significant impact
on the Gamow-Teller strength and level densities. The Gamow-Teller strength
may influence the competition between p-capture and ,B-decay in the rp-process.
Reaction rates, on the other hand, area strongly dependent on the mass differences
or Q-values of the reaction and the level densities at a fixed temperature. Reaction
rates tend to be strongly temperature dependent whereas ,B-decay is not.
Casten et al. [10] have shown that complex nuclear structure properties are corre-
lated with the product of the valence number of protons and neutrons for a given
nucleus NpNn. Np or Nn is counted as the number of valence protons or neutrons
up to the middle of the shell and as the number of holes beyond it. For example,
the nucleus IOOSn would have an NpNn value of zero while 72Kr with N = Z = 36
would have an NpNn value of 64. Information on known B(E2) values, deforma-
tion parameters, and energy ratios can be used to predict the structure of nuclei
that are presently unknown. We have used this same parametrization to predict the
structure of nuclei in the A = 80 region along the rp-process path. As stated above,
the deformations of nuclei along the rp-process nucleosynthesis path can result in
very different Gamow-Teller strength distributions and affect the balance between
capture and ,B-decay.
NUCLEOSYNTHESIS IN THE RP-PROCESS 419
All the known B(E2: 2+ ---+ 0+) values [11] for the Z = 28-50; N = 28-50
major shells can be plotted as a function of NpNn. When the data points for each
major shell are fit with a simple weighted linear function, the empirical relationship
of B(E2: 2+ ---+ 0+) values to the parameter NpNn for the N = 28-50 shell is
found to be
We have used the fits to the known data as a first order preliminary guess to
interpolate the B(E2) values of unknown nuclei along the rp-process path. The
measured or predicted B(E2) values can be used to calculate nuclear shape related
parameters such as the intrinsic quadrupole moment Q [12], where
16n
Q= - 2 (B(E2») (2)
5e
or the intrinsic coordinates f3 and y along with the three Euler angles el , (h, and e3
that describe its orientation in space.
The transition probability for an inband transition where the rotation-vibration
coupling is turned off and the nucleus is y-symmetric is given [13] by
(3)
where
2 3ZeRo2
A =--- (4)
4n
and
a = 0.36f32. (5)
Ze is the nuclear charge and Ro = 1.2 A l / 3 fm. A compilation of B(E2) values has
simply extracted f32 values using the relation
where A 2 is the same as above. f32 values were calculated using both equations.
The differences in the extracted deformation parameters are at most 10%. We have
therefore chosen to display for comparison the f32 values calculated by the latter
equation in order to be compatible with the data compilation [11].
Figure 1 shows the deformation parameters f32 calculated for various isotopes
of interest in the region on both sides of stability [14].
The results of our predictions as well as those from other models were compared
to each other in terms of their influences on the modelling of the rp-process via
extensive network calculations [6].
420 A. APRAHAMIAN ET AL.
48 ll071122J.1731.175I.l771·1861·1911·1911·1921·1711·166 [108
Cd
46 Pd 1.1151.1461.1971.2091.2291.2431.2571.2161.1571.201 I
44 Ru 1.1581.1951.2171.2441.2741.2831.2961.3031.3011
42 Mo EE 1.1511.1721.1681.2311.32713261.3531.3531
B. B
l
40 ZrB·250 1.1491.187 1. 103 1.090 1.081 320 1.421
38 Sr 1.35513731
~ &13~1_13cI2~12~1~1 1·~FMI3ul
34 Se 1.2931.207130213091.271 1.2321.1941 B
35 40 45 50 55 60 65 70 75
N
Figure 1. Predicted deformation parameters fh on both sides of the valley of stability. The
bold squares show predictions in comparison with deformation parameters extracted from
experiments in the regular boxes.
Masses are perhaps one of the most fundamental nuclear properties where a nu-
cleus A made up of N neutrons and Z protons has a mass that is measurably
different than the sum of the free nucleon masses. The bulk of the nuclear mass
is due to gross nuclear binding properties. We have attempted to isolate the struc-
ture component of the nuclear mass by using a combination of measured masses
and calculated ones from a mass model [15]. Brenner et al. [9] had shown that
the structure component of nuclear masses can be correlated with P, the ratio of
the product over the sum of the valence number of protons and neutrons. That is
P = NpNnj(Np + Nn). The method involves taking the measured masses from
the compilation of Audi and Wapstra [16] and subtracting away the macroscopic
nuclear features by the Macroscopic term of the FRDM+folded Yukawa single
particle mircroscopic model [15]. What remains behind is a semi-empirical mass
value which is correlated with P. Figures 2 and 3 show plots of SEM vs. P for the
N = Z = 28-50 and the Z = 28-50 with N = 50-82 shells. The range of the
SEM is a few MeV compared to the range of the entire nuclear mass on the order
of 60 MeV. It is evident from the figure that there is a smooth correlation with P,
there are some variations from the overall agreement for nuclear masses near the
end of the closed shell. We suspect that these deviations are due to pairing effects
since they seem to separate by Z with decreasing slope as one approaches P values
of approximately 3. The decrease in slope is itself quite a smooth function of Z.
The relationship of the SEM with P is exploited to predict masses of nuclei along
the rp-process path that are presently unknown.
This semi-empirical model may be applied widely across the chart of nuclides
but we expect limited applicability in very light nuclei. Also, no wide-scale com-
parisons have been made with the predictions of this approach with other models
NUCLEOSYNTHESIS IN THE RP-PROCESS 421
SEM VS P for Z=28-50 & N=28-50
..
5
.-,
~
- , If
..
+ ~
,.,. .. ,
:! -
......
+ - +
4 ....
+'0 ,0 lib J. -,
0
, ......
iI.~ +- 0
.•
.. .. .
0 0
3
• 0
.-
0
0
.
l:.
•
6 0
••
•
2 · 0 6 Cu
*
l:.
..• Zn
<:1
••
<:1 <:1
..
.. <:1 0 Ga
:;- • +
.,. ..
Ge
GI .... IJ.. 'V 0
.~ As
~ 0 11>
+ Sa
+
-1 . " 0<:1 Br
W
(J) lIE Kr
-2 • + -, Ab
Sr
0 y
-3 0 Zr
Ll.
Nb
-4 Mo
<:1
0 Tc
-5 + Au
.. ..
..
.. ... + +
4 ~~''''
+ lIE +
2
• Se
• Br
o .. Kr
.. Rb
Sr
-2 + Y
Zr
w Nb
en -4
lIE
- Mo
, Tc
+
-6 o Ru
+ + o Rh
Ll. Pd
+
-8 <:1 Ag
+
o Cd
+ In
o 2 3 4 5 6
p
Figure 3. Semi-empirical mass values in MeV as a function of P for the Z = 28-50 and the
N = 50-82 shell.
422 A. APRAHAMIAN ET AL.
2.5 S Index
0 dM p=M-M
p exp CU(1-14)
2.0 • dM lh =M 1h -Mexp Zn(15-34)
Ga(35-53)
Ge(54-71)
1.5
As(72-87)
>
<l> 1.0
Se(88-99)
Br(100-1 13)
~
Kr(1 14-127)
oS 0.5
Rb(128-140)
~
"0 Sr(141-151)
a. 0.0 Y (152-160)
~
"0 Zr(161-167)
-0.5 Nb(168-173)
Mo(174-179)
-1.0 Tc(180-183)
Ru(184,185)
-1.5
0 15 30 45 60 75 90 105 120 135 150 165 180
index
Figure 4. The errors between experimental mass excess values and the calculated ones from
P-parametrizations (solid squares) and the FRDM values (unfilled circles).
in regions that do not apply for the rp-process. Figure 4 shows a comparison of the
errors between the measured mass excess values and those calculated using this
semi-empirical approach and the FRDM approach.
4. Example
An interesting application can be made for the waiting point nucleus 68Se with N =
Z = 34. Figure 5 shows the relevant section of the chart of nuclides. There have
been several attempts [17-19] to measure the half-life of 69Br by fragmentation
studies at GANIL and MSU. No 69Br was observed but an upper limit of 150 ns
was set for the half-life. The implication is that 69Br should be proton unbound.
This implies that the rp-process should proceed via 2p-capture or ,B-decay of 68Se.
There are some new mass measurements [20] of several N = Z nuclei in the
A = 80 mass region from GANIL. The measurement involves the use of the CSS2
cyclotron at GANIL as a high resolution mass spectrometer. The nuclei of interest
were produced in a fusion evaporation reaction and injected into the cyclotron.
The relative mass differences of unknown and known nuclei are determined from
their differences in times of flight in the CSS2. The mass excess measured for
68Se was -52.347+/-0.080 MeV. The Audi and Wapstra extrapolations predict
a mass excess that is 2 MeV lower at -54.150 MeV. If one now uses the mea-
NUCLEOSYNTHESIS IN THE RP-PROCESS 423
68 Se Waiting Point
68 Se(~u) 68As T1/2=36 s
68 70
Se(2p,y) Kr II (40)
•
Y (39)
5«38)
Rb (37)
Kr (38)
81 (35)
Se(34)
(P'Y)2
As (33)
Ge(32)
Ga(31)
(p,y) 1 Zn(30)
Cu(29)
56
Ni
Irom He
Figure 5. A section of the chart of nuclides showing a possible rp-process path beyond 56Ni
through various N = Z waiting point nuclei.
10 '
....., ..
-- IV,.
-
10'
- ....
10 ~ 10
J10" ! 10 '
10· 10 ·
10'"
0.0 1000 2000
TIm4I(.j
300. '000 ..... 10 ·
TItM(.'
... 0
(a) (b)
Figure 6. The abundances calculated in a type I X-ray burst using FRDM predicted masses (a) and
the new measurement for the mass of 68Se (b).
sured 68Se mass, the implication is that 69Br should be bound by greater than
1 Me Y. This result is of course inconsistent with the fragmentation experiments.
The calculated proton separation energy for 69Br is estimated to be -100 keV
by the FRDM [21], -330 keY by the Moller-Nix-Meyers-Swiatecki model [15],
-660 keY by Ormand [22]. Our estimate is +40 keY. The new mass measurement
for 68Se has profound implications to the abundances of the elements produced in
the rp-process. Figure 6 shows a comparison of the abundances calculated with the
FRDM masses, and the experimental value of 68Se. There are sizeable differences
in the resulting abundances specifically with the depletion of 68Se.
424 A. APRAHAMIAN ET AL.
Acknowledgement
The support of the National Science Foundation under grant PHY 99-01133 is
gratefully acknowledged.
References
1. Woosley, S. E. and Taam, R. E., Nature 263 (1976), 101.
2. Maraschi, L. and Cavaliere, A., Highlights of Astronomy 4 (1977), 127.
3. Joss, P., Nature 280 (1977),310.
4. Van Wormer, L. et al., Astrophys. 1. 432 (1994), 326.
5. Bildsten, L. and Strohmayer, T., Physics Today 40 (February 1999).
6. Schatz, H. et al., Phys. Rep. 294 (1998),167.
7. Schatz, H. et al., Phys. Rev. Lett. 86 (2001), 3471.
8. Casten, R. E et al., Phys. Rev. Lett. 58 (1987), 658.
9. Haustein, P., Brenner, D. S. and Casten, R. E Phys. Rev. C 38 (1988), 467.
10. Casten, R. E, Nuclear Structure from a Simple Perspective, Oxford, 1990.
11. Raman, S., ADNDT 36 (1987), 1.
12. Nazarewicz, W. and Ragnarsoon, 1., In: D. Poenaru and W. Greiner (eds), Handbook of Nuclear
Properties, Oxford, 1996.
13. Eisenberg, J. M. and Greiner, W., Nuclear Models, North-Holland, Amsterdam, 1987.
14. Aprahamian, A., Gadala-Maria, A. and Cuka, N., Rev. Mex. Fis. 42 (1996), 1.
IS. Moller, P., Nix, J. R., Myers, W. D. and Swiatecki, W. J., ADNDT 59 (1995), 185.
16. Audi, G. and Wapstra, A. H., Nucl. Phys. A 595 (1995), 409.
17. Blank, B. et at., Phys. Rev. Lett. 66 (1991), 1571.
18. Morrissey, D. J. et al., In: Proceedings of the International Conference on Exotic Nuclei and
Atomic Masses (1995), p. 303.
19. Blank, B. et al., Phys. Rev. Lett. 74 (1995), 4611.
20. Lalleman, A. S., PhD Dissertation, GANIL (2000) and this issue, p. 315.
21. Moller, P., Myers, W. D., Swiatecki, W. J. and Treiner, J., ADNDT 39 (1988), 225.
22. Ormand, W. E., Phys. Rev. C 55 (1997), 2407.
Hyperjine Interactions 132: 425-432, 2001. 425
© 2001 Kluwer Academic Publishers.
Abstract. A highly extrapolatable semiempirical shell model mass equation applicable to translead
elements up to Z = 126 is presented. The equation is applied to the recently discovered superheavy
nuclei 293 118 and 289 114 and their decay products.
Key words: atomic mass, nuclear binding energy, shell model, superheavy elements.
1. Introduction
A recent experiment [11 is consistent with the formation of the nucleus 293 118 and
its sequential decay down to 265Rf (Z = 104). The a-decay energies vary rather
smoothly along the chain, precluding the traditional macroscopic-microscopic
[2, 3] Z = 114 as a major magic proton number in these nuclei. Recent phenom-
enological studies ofB(E2) [4] and Wigner term [5] systematics indicate Z = 126
as a plausible next spherical proton magic number after lead. Recent self-consistent
and relativistic mean field calculations [6-8] variously predict proton magicities for
Z = 114, 120, 124 and 126, depending on the interaction used.
Contrary to most of the above findings, the semiempirical shell model mass
equation (SSME) [9] is based on the assumption that Z = 114 is the next proton
magic number after lead, and it stops there. Moreover, the quality of its agreement
with the data starts deteriorating already beyond Hs (Z = 108). (See Figure 1
and [1, Figure 4].) One has to find a substitute for the equation in the neighbour-
hood of Z = 114 and beyond.
In the early stages of developing the SSME [10] both Z = 114 and Z = 126,
then considered possible alternative candidates for the postle ad proton magic num-
ber, were tried as an upper shell-boundary for translead elements. The agreement
with the data was about the same for both choices, and considering the prevailing
view in the mid nineteen-seventies Z = 114 was chosen for the SSME mass table.
We have recently [11] established a high predictive power of the early Z = 126
results in the interior of the shell region with Z ;? 82 and 126 ~ N ~ 184 (called
>'
CD
1.0
•
!, ••
=0.5
ICL
•
::::IE
<:1
u
I •
•
• •
ii 0.0
u
::::IE
<:1
• •
Atomic number Z
Figure 1. Deviations of the mass predictions [9] from the data for the 56 new masses in
region B measured after the original adjustments were made.
here region B) by comparing them to the newer data measured since then, and pro-
posed using them as a substitute for the SSME [9] in superheavy elements (SHE)
research. We have also [12] established a high predictive power for Qa values of
the early Z = 126 results in the interior of the shell region with 82:( Z, N :( 126
(called region A). However, the quality of the predicted masses and other mass
differences worsened much compared to that of the original adjustment. Therefore
we readjusted the coefficients which largely cancel in Qa, and proposed using the
resulting equation as a substitute for the SSME [9] in the interior of region A.
In Section 2 we give the Z = 126 equation in the two regions and briefly discuss
its predictive power. In Section 3 we apply and comment on it in relation to SHE
research in region B.
ECoul(N, Z)
(2Z
A o)
1/3
[a C + f3 C (Z - ZO) + y C (Z - 2
ZO) ], (2)
Epair(N, Z)
SEMIEMPIRICAL SHELL MODEL MASSES 427
l-(-l)N l-(-1)Z l_(_l)NZ]
+ 2
8 1 + 2
82 + 2
fJ, (4)
for region B,
with
with
>' 1.01- I
GI 00
o N = 126-128
~
• N ~ 129
-L
-10 -5 o 5
Neutrons from ~·stability NFS
Figure 2. Deviations of the mass predictions Equation (1) from the data for the 56 new masses
in region B measured after the original adjustments were made. The deviations are plotted as
function of the variable NFS = N - Z - 0.4A 2 ;(A + 200) [14].
The deviations of the remaining 49 nuclei with N ;? 129, which do not follow
the N = 126 boundary but extend into the interior of the shell region, seem more
random and they are smaller, with respective average and rms deviations of - I and
155 keY.
The above deviations are as a rule about twice smaller than those of several
recent mass models. This is presumably due mainly to the inclusion in Equa-
SEMIEMPIRICAL SHELL MODEL MASSES 429
:> I I I I
.2
~
. 0 mass known in 1973 0 Q}
-15 -10 -5 0
Neutrons from ~-stability NFS
Figure 3. Deviations of the mass predictions Equation (I) from the data for all the 150
presently known masses in region A. The deviations are plotted as function of the variable
NFS = N - Z - 0.4A 2 /(A + 200) like in Figure 2.
tion (1) of the particle-hole (p-h) symmetric configuration interaction terms Edef,
Equation (8). A more detailed discussion is given in [11].
The situation in region A is less simple. The experimental data used in the ad-
justment [10] included 29 masses and 62 Qa values connecting unknown masses.
The respective average and rms predicted deviations for the presently known 121
new masses which became available after the previous adjustment increase dras-
tically to -807 and 1008 keY, as compared to -29 and 146 keY in [10]. For the
31 new Qa values, though, the respective deviations are only 40 and 89 keY, as
compared to 5 and I 03 keY.
In order to restore the new mass predictions to the same quality as the old pre-
dictions had, while at the same time retaining the high quality of Qa predictions,
we made [12] a least-squares adjustment of Equation (1) to all the 150 known
masses, with only four adjustable parameters a, c, 8 and K (Equation (3)) which
largely cancel in Q,n while the other seven coefficients were held fixed on their old
values [10]. These are the values given in the table.
The readjusted values of a, 8 and K are higher and that of c is lower than
in [10], indicating smaller overall binding, smaller symmetry energy coefficient,
and increased pairing energies in proton-rich nuclei away from stability.
Figure 3 shows the deviations from the data of the predictions of the readjusted
Equation (1) for all the 150 experimentally known masses. The respective aver-
age and rms deviations are 2 and 246 keY. The corresponding deviations of the
predicted Qa values are 2 and 99 keY.
A more detailed discussion is given in [12].
3. Applications to SHE
Panel a of Figure 4 shows the chain of measured a-decay energies starting from
293 118 [1], and the values predicted for them by Equation (1) when the data are
interpreted as g.s. transitions of the assigned nuclei. The figure shows as well the
430 S. LIRAN ET AL.
12
-o-Data
- Present work
9
- - .. _ . Smolariczuk
12
11
~
!..10
1:1
a
9 - 0 - Data
- - - - Cwiok et al.
8 - - .. -. MOiler et al.
._ ..... - Aboussir et al.
predictions of [15] which motivated the search undertaken in [1]. The respective
average and rms deviations of the predicted values from the data are -197 and
308 keY for Equation (1) and -154 and 357 keY for [15]. The rms deviation
of Equation (1) is consistent with the deviations of the new nuclei in Figure 2
considered above, but the average deviation is too negative. The respective average
and rms deviations of the predictions of Equation (1) from those of [15] are -43
and 369 keY.
The variation ofthe predicted values of Equation (1) along the chain is smoother
than that of the data, where there are kinks at Z = 112 and 116 presumably
representing submagic numbers or other structure effects. In the SSME such ef-
fects are assumed [9, 13] to have been smoothed out by configuration interaction,
represented by the terms E def . The SSME is inadequate to describe non-smoothed
abrupt local changes associated with subshell structure [9].
On the other hand, the microscopic energies calculated in [15] are basically
sums of (bunched minus unbunched) single nucleon energies, and as such have
SEMIEMPIRICAL SHELL MODEL MASSES 431
(magic and) submagic gap effects built in. The corresponding predicted line in
Figure 4 has kinks at Z = 110 and 116, corresponding to predicted submagic
numbers Z = 108 and 116.
Most of the smoothing effect of configuration interaction is missing in macro-
scopic-microscopic Strutinsky type and in self-consistent mean field calculations.
The included T = 1, J = 0 pairing correlations seem not to be enough. This
might result in calculated submagic gaps and associated kinks which are too large
compared to the data. Panel b of Figure 4 shows such predicted large kinks [16-18].
As a second application we address the a-decay chain observed in [19], which
is considered a good candidate for originating from 289 114 and its sequential decay
to 285 112 and 281 110. The respective average and rms deviations of the predictions
of Equation (1) from the measured energies are 847 and 905 keY, which consider-
ably exceed the deviations expected from Figure 2 for g.s. transitions. If the above
assignments are confirmed, the large deviations might indicate that the decay chain
does not go through levels in the vicinity of the g.s.
It might also be worthwhile mentioning, that for the conceivable parents 288 112
or 291 113 which can be obtained from the compound nucleus 292 114 by respective
1a or 1p evaporation, the corresponding average and rms deviations of Equation (1 )
from the measured energies are -181 and 366 ke V and -242 and 417 ke V, which
are more than twice smaller than for the parent 289 114.
Finally we address the bearing of the Z = 118 results on the question of magic-
ity of Z = 126 versus 114. The energy Equation (1) comprises a part E pair + ECouJ,
Equations (2) and (4), which is a quadratic parity-dependent function of the valence
particle numbers, and a part E def , Equation (8), which is a p-h symmetric function
of both particle and hole numbers. For different assumed upper shell boundaries the
hole numbers are different and E def is a different function of Nand Z. The superior
agreement of Equation (1) with the data in Figure 4, as compared to the deteriora-
tion of the SSME [9] near Z = 114 mentioned in the introduction, demonstrates
the superiority of proton-hole numbers defined with respect to Z = 126 as a proton
magic number rather than Z = 114. This conclusion has also been arrived at in [4]
on the basis of B(E2) systematics. It is important to emphasize, though, that the
above rather suggestive results are not a proof of superior magicity of Z = 126 as
compared to the recently predicted Z = 120 or 124, because no comparative mass
studies of this kind were made.
Acknowledgement
We thank Stelian Gelberg and Dietmar Kolb for help with the calculations.
References
1. Ninov, V. et ai., Phys. Rev. Lett. 83 (1999). 1104.
2. Moller, P. and Nix, J. R., J. Phys. G 20 (1994), 1681.
3. Smolanczuk, R., Phys. Rev. C 56 (1997),812.
432 s. LIRAN ET AL.
Towards a Skyrme-Hartree-Fock-Bogolyubov
Mass Formula
Abstract. In this work, we explain our astrophysical motivations for deriving a mass formula based
on HFB calculations with a Skyrme interaction. We give an overview of existing mass formulae and
present briefly the last HF+BCS mass formula [1]. The Skyrme force MSk7 [1] is considered in the
study of shell effects at N = 82, in the neutron-rich region far from stability, within the HFB and
HF+BCS theories, and compared with results obtained using the forces Skp8 and Skp 8p [2].
1. Introduction
1.1. ASTROPHYSICAL MOTIVATIONS
In order to explain the origin of about half of the heavy nuclei observed in nature
(neutron-rich nuclei with A > 60), nuclear astrophysicists invoke a nucleosyn-
thesis process called the r-process, or rapid neutron capture process. This process
has the specificity to produce exotic neutron-rich nuclei far from the valley of
,B-stability. The properties of these nuclei are not known experimentally, so that
theoretical extrapolations are needed. To put the description of the r-process of nu-
cleosynthesis on safer grounds, constant effort must be made to improve the mass
predictions [3]. The more fundamental the theory, the more reliable (hopefully) the
predictions will be, especially for nuclei near the neutron drip line.
Until recently the atomic masses were calculated on the basis of one form or
another of the liquid-drop model, the most sophisticated version of which is the
"finite-range droplet model" (FRDM) including improved pairing and shell cor-
rections [4]. This model has the quality to reproduce experimental data with a
rms error of about 0.670 MeV for the 1888 (Z, N ;;::: 8) nuclei with known
434 M. SAMYN ET AL.
masses, but suffers from shortcomings like (among others) the incoherent link
between the macroscopic part and the microscopic corrections [3, 5]. A first step
in the development of a mass formula that is more closely connected to the ba-
sic nuclear interaction is found in the "Extended Thomas-Fermi plus Strutinsky
Integral" model (ETFSI) [6, 7]. The ETFSI method is a high-speed macroscopic-
microscopic approximation to the Hartree-Fock (HF) method based on Skyrme
forces, with pairing correlations generated by a 8-function force treated in the usual
BCS approach. The Skyrme interaction underlies both macroscopic and micro-
scopic parts of the model, avoiding the incoherence of the droplet-like approaches.
Following this work, a microscopic mass formula based on the HF+BCS method,
HFBCS-l, has been developped recently [1, 8]. The corresponding lO-parameters
Skyrme force (MSk7) along with a 4-parameters 8-function pairing force (in the
BCS approach with blocking for odd nuclei) are obtained by a fit to the masses
of 1719 nuclei (A > 36), both spherical and deformed. The HF+BCS model
also includes corrections to the spurious rotational energy [1], centre of mass mo-
tion [9], and Wigner anomalies. The final rms deviation is 0.738 MeV for the 1888
(Z, N ;?: 8) nuclei with known masses. All details can be found in [1, 3, 5, 8].
Much work remains to be done to improve the reliability of the mass extrapolations
towards the neutron drip line. In particular, the HF+BCS model predicts a gas of
nucleons outside the nucleus, that increases for nuclei close to the neutron drip line
where nucleon pairs are scattered into the continuum [10]. This problem is avoided
in the HF-Bogolyubov (HFB) method [11], which puts the pairing correlations into
the variational function without any approximation (the BCS approach consists of
neglecting the nondiagonal matrix elements of the pairing tensor). Therefore, if we
are to have any confidence in the extrapolations out to the neutron drip line, it is
essential that the HF+BCS method used here be replaced by the HFB method.
where Vo is the pairing strength parameter and Pais the two-body spin-exchange
operator. It appears from Figure 1 that the HFB and HF+BCS shell gaps follow the
same behaviour, that is a reduction of shell effects when Z decreases; the deviations
between the three methods are relatively small, especially close to the neutron drip
line at Z = 34. To check the accuracy of the calculations in the configuration
436 M. SAMYN ET AL.
f-FB
tm! ---1(---
~dK. ' , l lm l -
HF .. BCS dx.'.1
,.
HF"t-BCS lbaae ... • ...
0
:lO 50 010 4lI .I!' 60 as 10
(a)
•
..
•
!'i
....
~
10
~
3
(b)
Figure 2. (a) Same as Figure 1, with Skyrme force Skp8; (b) same as Figure 1, with Skyrme
force SkP8P .
space, the mesh size has been decreased to 0.8 fm for the shell gaps at Z = 38: the
difference with the shell gap obtained with a mesh size 1.1 fm is negligible.
Figure 2(a) represents the shell gaps obtained with the Skyrme force SkP8 [2].
Again, no significant variations can be seen between the methods. In this case
however, the first unbound nucleus of isotones N = 82 is 116Se, which means
that the neutron drip line with this force will go through 118Kr (Z = 36) instead
of going through 116Se (Z = 34) if MSk7 is used (considering even Z only). Shell
gaps are smaller with this force than with MSk7 for Z < 50, and larger for Z > 50.
Figure 2(b) shows the shell gaps obtained with the Skyrme force SkP8p [2]. In
this case, the pairing interaction is density-dependent,
so that the pairing is confined to the surface. As seen in Figures 1 and 2, the results
are not very sensitive to the method. Shell gaps disappears faster for Z < 50 than
when use is made of Msk7 force and shell effects nearly vanish at Z = 36. We
note that our HFB results (for the two SkP forces) are in close agreement with
those obtained in [2].
3. Conclusions
In this work, we use three methods to solve the HF equations: the HF+BCS method
with a basis expansion, and the HF+BCS and HFB methods in the configuration
space. In addition, we consider three Skyrme forces (MSk7, SkP o, SkPoP ) to test
the sensitivity of the shell gap predictions far away from stability. We show that the
HFB treatment does not lead to critical changes in the behaviour of N = 82 shell
effects than when treated in the HF+BCS approximation: the shell-quenching of
astrophysical interest (defined as the vanishing of shell effects for nuclei far from
stability) is not a Bogolyubov effect, i.e., no significant differences appear between
HF+BCS and HFB predictions, for all forces considered here.
References
1. Tondeur, F., Goriely, S., Pearson, J. and Onsi, M., Phys. Rev. C 62 (2000), 02430R.
2. Dobaczcwski, J., Nazarewicz, W. and Werner, T., Phys. Scripta T56 (1995), IS.
3. Goriely, S., this issue, p. 105.
4. MOller, P., Nix, J. and Swiatecki, w., At. Data Nucl. Data Tables S9 (1995), 185.
5. Pearson, 1., this issue, p. 59.
6. Aboussir, Y., Pearson, J., Dutta, A. and Tondeur, F., Nuclear Phys. A S49 (1992), 155.
7. Aboussir, Y., Pearson, J., Dutta, A. and Tondeur, F.,At. Data Nucl. Data Tables 61 (1995),127.
8. Goriely, S., Tondeur, F. and Pearson, J., At. Data NucZ. Data Tables 77 (2001),311.
9. Butler, M., Sprung, D. and Martorell, 1., Nuclear Phys. A 422 (1984), 157.
10. Dobaczewski, J., Flocard, H. and Treiner, J., Nuclear Phys. A 422 (1984), 103.
11. Ring, P. and Schuck, P., The Nuclear Many-Body Problem, Springer, Berlin, 1980.
12. Pearson, J., Nayak, R. and Goriely, S., Phys. Lett. B 387 (1996), 455.
13. Vautherin, D., Phys. Rev. C 7 (1973), 296.
14. Bonche, P., Flocard, H., Heenen, P.-H., Krieger, S. and Weiss, M., Nuclear Phys. A 443
(1985), 39.
15. Terasaki, J., Heenen, P.-H., Flocard, H. and Bonche, P., Nuclear Phys. A 600 (1996),371.
Hyperjine Interactions 132: 439-442, 2001. 439
© 2001 Kluwer Academic Publishers.
Abstract. The masses of ions of some keY can be determined in mUlti-pass time-of-flight mass
analyzers [1,2] with high precision. The mass accuracies thus achieved are sufficient to determine the
proton and neutron numbers for most short-lived and stable nuclei [3,5]. Recording IX- or y-radiation
of the investigated nuclei in delayed coincidence to the ion arrival, one thus can perform nuclear
spectroscopy of selected nuclei.
Reflecting low energy ions repeatedly between grid-free electrostatic mirrors one
can increase the overall flight time. If this flight time is energy isochronous, i.e.,
if ions of the same mass but increased energies are guided along longer paths
so that they have the same flight time as ions of lower energies, the masses of
individual ions can be determined very precisely. For the isotopes of krypton mass
accuracies of about 100 keV were achieved [3-5]. Two such arrangements with
electric mirrors we have considered, both of which are illustrated in Figure 1. Such
multi-pass time-of-flight mass spectrometers we have built in three versions:
1. A V-shaped time-of-flight mass analyzer (see Figure la) as a ~1.0 m long
system [1]. Such a system uses one large and two small ion mirrors. The two
smaller mirrors can both be switched off at predetermined times so that ions
can enter and leave the time-of-flight system. For most of the time, however,
they reflect the ions back and forth and thus achieve a flight path of 10m or
more. The larger ion mirror is operated by static potentials. In combination
the three ion mirrors ensure energy-isochronicity by guiding more energetic
and thus faster ions on detours in that they penetrate deeper into the carefully
designed large repeller fields.
2. A V-shaped time-of-flight mass analyzer as a 1.2 m long system [5] whose
appearance is similar to Figure lao The fact that in this system the initial and
final ion beams move along the same axis, simplified the alignment procedure
considerably and thus increased the mechanical precision of the time-of-flight
mass analyzer. Since the system of [5] was build lO years later than the sys-
tem of [1], also much better electronics and more stable power supplies could
440 H. WOLLNIK AND A. CASARES
reflector I a
reflector II
f?:} IIIIII
source \\ \\ \ \
\\\\\\\\\\
IIlLII~~~iJ//,w//ml!I!'
IllJl
~\. \\\\\\
reflector III
reflector I reflector II b
r!.:].fI //// \ \ \\\\"I~
source \ \ \ \ \ \ / / / / / / detector
Figure 1. Two versions of multi-pass time-of-flight mass analyzers are shown that both use
two switchable electric ion mirrors. (a) In the first system the ions are guided along V-shaped
trajectories. In this system one static ion mirror is used additionally to the two switch able
ones. (b) In the second system the ions are guided on more or less straight trajectories.
400
200
>
~
E
0 f I
I
1
<l
-200
-400
82 83 84 85 86
mass number
Figure 2. Obtained mass accuracies for the four more abundant Krypton isotopes are shown,
whose masses have been measured in the system of Figure I b after 15 passes through the
'racetrack' .
---- ----
buncher E::~~ -1Il1ll -1Il1ll
III
-- --
-- - - reflector I
---- ----
--
-- -- - - reflector II
oB
~
~
~
€I
~
tE 8
I r-~
L-'
=
.c=
Figure 3. A system is shown that allows nuclear spectroscopy of mass tagged ions. For this
purpose the ion-arrival time must be recorded very precisely so that the neutron and proton
numbers of an individual ion can be determined as well as a slightly later occurring y-decay.
If additionally also the ion-position is recorded as well as the origin of the y coincident
,a-particle, a correlation between the y-decay coincidence can be attributed to the before
recorded ion.
442 H. WOLLNIK AND A. CASARES
1 ms can be detennined. The mass accuracies achieved in such systems so far are
inferior to those obtained in Penning traps [6] for ions of half lives ~50 ms [7].
However, multi-pass time-of-flight mass analyzers should be able to detennine
nuclear binding energies of exotic nuclei with errors around 100 ke V, which makes
such measurements useful for astrophysical calculations.
Besides providing a system to detennine nuclear masses of short-lived nuclei,
the electric mUlti-pass time-of-flight mass analyzers here described could be used
as isotope selective tagging devices that would allow nuclear spectroscopic in-
vestigations of nuclei of specified neutron and proton numbers. Such a system is
indicated in Figure 3.
Starting from a bunching ion source the ions arrive in such a system at some
ion detector after a flight time proportional to the square root of the ion mass. This
detector could be built as a position sensitive flight-time detector similar to the one
[8, 9] that was developed for mass measurements with TOFI [10] in Los Alamos
and with the ESR [11] of the GSI in Darmstadt. In both these detectors the ions
impinge on some foil in which secondary electrons are formed, accelerated, and
isochronically transported to some multi-channel-plate detector (MCP) indicated
in Figure 2. The amplified electron signal then records the electron-arrival time
and position, which reveal the ion-arrival position as well as the ion-arrival time.
This ion-arrival time characterizes the neutron and proton numbers of the ion in
question, while the ion-arrival position can be correlated to the position of a later
emitted ,B-particle which in tum can be recorded in coincidence with some y-decay.
Thus nuclear spectroscopy can be performed of ions of 'precision mass-tagged'
ions, i.e., of ions of known neutron and proton numbers.
Acknowledgements
For fruitful discussions the authors are grateful to T. Horvath and A. Piechaczek
and for financial support to the German Ministry for 'Bildung und Forschung'.
References
1. Wollnik, H. and Przewloka, M., Int. 1. Mass Spectr. Ion Proc. 96 (1990), 267.
2. Piyadasa, C. K. G., Haakenson, P. and Ariyaratne, T. R., Rapid Commun. Mass Spectrom. 13
(1999), 620.
3. Casares, A. et ai., In: Proc. 47th ASMS Cant Dallas, 1999.
4. Casares, A. et al., In: Proc. 48th ASMS Can! Long Beach, 2000.
5. Casares, A., Kholomeev, A. and Wollnik, H., Int. 1. Mass Spectr. 206 (2001).
6. March, E. and Todd, F. J., Practical Aspects of Ion Trap Mass Spectrometry, CRC-Press, New
York, 1995.
7. Bollen, G., this issue, p. 215.
8. Kraus, R. H. et ai., Nuci. [nstr. Meth. A 264 (1988), 327.
9. Troetscher, J. et al., [nst. Pins. Can! Ser. 132 (1992),959.
10. Wouters, J. et al., Nuc/. [l1str. Meth. 240 (1985), 77.
11. Stadlmann, J. et al., AlP Can! Proc. 512 (2000), 305.
Hyperfine Interactions 132: 443--450, 200l. 443
© 2001 Kluwer Academic Publishers.
N. I. TARANTIN
Flerov Laboratory of Nuclear Reactions, lINR. 141980 Dubna, Moscow, Russia
Abstract. Data on nuclear masses provide a basis for creating and testing various nuclear models.
A tandem system comprised of the U-400M cyclotron, the COMBAS magnetic separator and the
mass spectrometric ion trap of an 'in-flight capture' type is considered as a complex for producing of
the short-lived nuclei by heavy ions in fragmentation reactions and for precise mass measurement of
this nuclei. The FLNR plan scientific and technical research includes a project DRIBs for producing
accelerated beams of radioactive nuclear reaction products and photofission fragments. This project
proposes also precise mass measurements with the help of ion trap.
1. Introduction
Data on nuclear masses provide a basis for creating and testing various nuclear
models. Comparing the values of the masses predicted by theoretical models with
those determined as a result of measurement is of special interest for short-lived
nuclei far from the beta-stability line. For mass measurement of those nuclei car-
ried out on the beam of radioactive nuclear reaction products or fission fragments,
a Penning ion trap can be used. Two facilities are described:
(A) The large solid angle and momentum acceptance of projectile-like fragment
separator COMBAS [1] is being designed for experiments at intermediate-energy
heavy ions accelerated on the U-400M cyclotron of the FLNR. The COMBAS
system is a achromatic separating magnetic line with the double spatial focusing of
nuclear reaction products. A tandem system comprised of the U-400M cyclotron,
the COMBAS magnetic separator and the ion trap of an 'in-flight capture' type [2]
can be used for rapid and precise mass measurement of the recoil short-lived nuclei
produced by heavy ions in the fragmentation reactions.
This tandem system which is shown in Figure 1 is to perform the following
operations:
444 N. !. TARANTIN
D3
D4
D5
DB
PT
Figure 1. Lay-out of the COMBAS magnetic separator (dipoles DI-D4 and DS-DS grouped
into two sections) and mass spectrometric ion trap of an 'in-flight capture' type (PT). The
U-400M heavy ion cyclotron is situated in the same hall. Magnetic dipoles Dl and D2 (DS
and 07) are the analysing section of the separator. Dipole D 1 (DS) defocuses in the radial plane
and focuses in the axial plane. The action of dipole D2 (D7) on the charged particles is reverse.
Magnetic dipoles 03 and 04 (D6 and 05) are correcting and steering pair. The induction sign
of 04 (DS) is opposite to that of DI-D3 (DS-D6). FI is the intermediate dispersive focus, F2
is the exit achromatic focus.
(i) Accepting the recoil nuclei produced in the target (T) by the magnetic sepa-
rator and decreasing their kinetic energy down to 1-2 A Me V by degraders.
At such energy, the light recoil reaction products have a total ionic charge.
In a general way, the reduction of the kinetic energy of nuclei by degraders
results in their relative energy dispersion increasing. This disadvantage is sub-
stantially offset by monochromatizing degraders of a wedge type [3], which
make use of the energy-dependent position dispersion of charged particles in
the angular focusing plane (FI or F2).
As an alternative to wedge degraders, we proposed [4] to use a degrader of
a new type: a uniform flat plate placed at the site of the beam transformation
of 'point in parallels' (between magnetic dipoles DI and D2 or between D7
and D8) and inclined at some angle to the optical axis. A flat plane degrader
is very simple to fabricate.
(ii) In-flight transportation of the retarded nuclei to the magnetic solenoid of the
ion trap and transforming their remaining longitudinal kinetic energy into an
azimuthal rotation by the fringing magnetic field of the solenoid according to
Bush's theorem [5].
STORAGE ION TRAP OF AN 'IN-FLIGHT CAPTURE' TYPE 445
(iii) Confinement and accumulation of the rotating ions in the trap by using their
reflection by the electrostatic field of the end cap electrode of the Penning trap
and their repelling by the magnetic field of a magnetic ventil with unidirec-
tional transportation of moving ions located at the entrance to the solenoid.
(iv) Cooling and bunching the orbital rotation of the captured ions by a high-
frequency azimuthal electric field or cooling by a retardation of the ions in
the gas.
(v) Precise mass measurement of the ions in the trap.
(B) One of the FLNR's research projects is dedicated to producing beams of
radioactive nuclear reaction products and fission fragments. This plant, named the
DRIBs project [6] (Dubna radioactive ion beams), in particular, includes:
(i) Production of fission fragments in the photofission of uranium on the MT-25
microtron. Paper [7] reports the experimental integral photofission yields for
the whole bremsstrahlung spectrum and the calculated photofission cross sec-
tions for nuclei of 232Th, 238U and 249Cf. (The measurement of the photofis-
sion cross section of 249Cf reported in [7] was made for the first time, 232Th
and 238U being used as reference.)
From the results of [7] it follows that the total flux of fission fragments from a
238U thick target (30 g/cm2) for a 20 /LA beam of electrons from MT-25 with
energy of 25 MeV is equal to 7 . 10 II fragments/so
(ii) Rapid diffusion and effusion of the fission fragments from the hot target and
the ionisation of the atoms leaving the heated target in a special ion source.
(iii) Preliminary mass selection of the fission fragments with a magnetic dipole
separator and subsequent precise mass measurement of the separated fission
fragments with a mass spectrometric ion trap of an 'in-flight capture' type.
1
(a) 2
3
Rmox
Rmin
BzIO.Z)
z
Bo
(b)
Z
Figure 2. (a) Fragment of a magnetic solenoid. 1 - Iron armour of the solenoid, 2 - excitation
electric current coil, 3 - projections of the charged particle trajectory in the magnetic field
on the r, z- and cp, r-coordinate planes. (b) Distribution of the magnetic induction component
Bz(O, z) along the solenoid axis z.
where a = (z - zo)le, e - the extent of the fringing magnetic field. Formula (1)
describes the distribution of the magnetic induction Bz over the interval e and
its vanishing derivatives of the first, second and third orders at the ends of the
interval e.
Charged particles entering the magnetic field of the solenoid by the off-axial
entrance with the parameter ro i=- 0, the fringing magnetic field transforms the
particle longitudinal momentum into an orbital rotation. This rotation is defined, as
indicated above, by known Bush's theorem [5]. But the monograph [8, pp. 85-89]
states this theorem in the more simple form.
To simplify the consideration, we admit that the magnetic field in the inter-
nal space of the solenoid is homogeneous and the external Bz-component sharply
decreases to zero at the entrance (z = -I) boundary of the solenoid. It means
that dBz(cp, 0, -l)/dz = Bo8(z + I), where 8(z) is Dirac's function. Then the Br-
component of the fringing magnetic field at the entrance boundary (z = -I) found
by the known method (see, for example, [8, p. 28, 291 is
r Bo8(z + I)
Br(cp, r, -I) = - 2 . (2)
The projection of the particle trajectory on the cp, r-plane after the passage of
the particle through the entrance fringing field we obtain by time integration of the
equations of motion with using (2). After the first step integration for the initial
STORAGE ION TRAP OF AN 'IN-FLIGHT CAPTURE' TYPE 447
where the symbol (') denote the time derivatives, cpi is the angular velocity after
the passage of the particle through the entrance fringing field, W B = q Bo / m is
the numerical value of the cyclotron angular velocity of the charged particle in a
homogeneous magnetic field Bo; m and q are the mass and electric charge of the
particle.
We recall that the angular velocity cp' (t) is called angular velocity (frequency)
of Larmore for charged particle in a homogeneous magnetic field Bo: 1/1 = 2cp,
1/1' = 2cp'.
Figure 2(a) shows the projection of the positive charged particle trajectory in
the main magnetic field of the solenoid on the r, z- and cp, r-coordinate planes. We
see that a charged particle moves in a circular orbit of radius rB = ro/2, which
crosses the z-axis at cp = IT /2.
The result obtained for an ideal magnetic field and given by Equations (3) is
correct, as for the first time was shown in monograph [8, pp. 88, 89], for a real
magnetic field with an extended fringing field 0), which is shown in Figure 2(b),
if the concept of the effective boundary of a magnetic field is used.
The axial component of the particle kinetic energy after passing through the
entrance fringing field of the solenoid we obtain from equation of motion as
(4)
where To = mv 2/2 = m(w B R B )2/2 is the initial kinetic energy of the particle, v
is the velocity of the particle, !:1 To is the disappearing initial longitudinal kinetic
energy.
As can be seen, the electrostatic reflector with the value U q equal to the lon-
gitudinal kinetic energy Tl determined by formula (4) reverses all ions into main
solenoid volume (see Figure 3).
The acquired kinetic energy of the rotation of the particle in the magnetic field
is
m(WB R o)2
T2 = 8
and !:1To = T2; (5)
Figure 3. Electrostatic reflector of the charged particles in the storage trap. 1 - Iron armour of
the solenoid, 2 - excitation electric current coil, 3 - electrostatic ring quadrupole electrodes,
4 - projection of the charged particle trajectory in the magnetic and electrostatic fields on the
r, z-coordinate plan, 5 - the end cap electrode of the quadrupole hyperboloid.
From Equations (5) and (6) it follows that at ro = 2RB = Rmax the solenoid's
fringing field completely decreases the longitudinal particle motion, transforming
it into rotary motion. This result is used in our case for the realisation of a repelling
magnetic ventil with unidirectional transportation of moving ions. The ventil is
located before the entrance of the trapping solenoid and in Figure 2 is not shown.
Thus the continued in-flight capture of the nuclear reaction products in the trap is
realised.
As shown in [9, 10] the single possible decomposition of the motion of a charged
particle in the crossed homogeneous magnetic field B z = Bo and volume quadru-
pole electric field Er = Grr of the Penning trap is two circular motions: circular
rotating motion Wro! and circular drift motion Wd. The cyclotron radius RB (6) is
the radius of the rotating circular and drift motions: R rot = RB and Rd = R E . The
maximal angular velocity of the charged particle in the Penning trap - so-called
reduced cyclotron frequency w~G' is the angular velocity of the rotating circular
motion: W ral = w~G' Then the drift angular velocity is Wd = w~G - so-called
magnetron frequency. This result is the reproduction in the new consideration of
the known result given, for example, in the very thorough review article [11].
particles in an ion trap and their detection. However, Tarantin [12] used the inverse
problem solution method to analytically obtain the fields of a prescribed structure
for various electrical systems. Within this method, the prescribed basic field is
expanded by integration of Maxwell's equations div E = 0, rotE = O.
Using the inverse problem solution method [12] and the azimuthal non-uni-
formity of the basic magnetic field Eq; in the plane z = 0,
the following potential distribution we found for an ion trap possessing azimuthal
electric field component (7):
4. Summary
The parameters for proposed ion trap of an 'in-flight capture' type for mass mea-
surement of separated photofission fragments to be used possibly, for example,
within the framework of the DRIBs project are as follows. The kinetic energy of
accelerated and mass-separated fission fragments is 15 keY. This value of accel-
erating voltage has been chosen from the condition of the optimal transportation
the ions into the ICR source for breeding their charges. The magnetic induction of
the superconductive solenoid is up to Bo = 4 T in order to be able to capture such
heavy singly-charged fission fragments as 132Sn, 133Sb and so on to 141CS, 144Ba.
The length of solenoid is L = 100 cm, and its bore diameter is D = 30 cm. The
off axial input is at ro = 10 cm. The stop potential of the end cap electrode (5) is
Vstop = 2.5 kV at bora = Rmax - Rmin = 1 mrn. The minimal electric potential of the
quasi-hyperboloidal circular and ellipsoidal ring electrodes of the linear quadrupole
cell is Vmax = -200 V.
450 N. I. TARANTIN
The response frequency function of the ion trap for atomic mass measurement
of ions is equal to wtG + wiG = WB, which provides the most precise mass mea-
surement, for example, with relative accuracy of up to 10-7 -10- 8 , on the analogy
of the results of R. S. Van Dyck et ai., S. Rainville, G. Bollen, G. Savard et ai.
presented at the APAC 2000 Conference, Cargese, France.
We note that the developed storage mass spectrometric ion trap of an 'in-flight
capture' type result also as a sufficient effective accumulator-collaider device for
generation of the neutrons in vacuum collisions of the accelerated deuterium and
tritium ions. This device can be used as driver of a subcritical nuclear fission reactor
for the production of commercial electric energy (see [13]).
References
1. Artukh, A. G. et aI., Nucl. Instrum. Meth. A 306 (1991), 123; 426 (1999), 605.
2. Artukh, A. G. and Tarantin, N. I., Nucl. Instr. Meth. B 126 (1997), 246.
3. Peterson, J. M., IEEE Trans. Nucl. Sci. NS-30 (1983), 2403;
Hoath, S., Nucl. Instr. Meth. A 248 (1986), 287;
Dufour, J. P. et al., Nucl. Instr. Meth. A 248 (1986), 267.
4. Tarantin, N. I., FLNR, JINR Sci. Report 1993-1994, Dubna, 1995, p. 207.
5. Bush, H., Ann. Phys. 81 (1926), 974.
6. Oganessian, Yu. Ts., Scientific prospect and further development of the accelerator complex of
FLNR, Report to the 9th meeting of PAC for Nuclear Physics, Dubna, November 23-29, 1998.
7. Tarantin, N. T. and Men, Kim Su, In: Proc. of Workshop on Application of Microtrons in
Nuclear Physics, Plovdiv, 22-24 September 1992, D-93-80, p. 75.
8. Tarantin, N. I., Magnetic Static Analyzers of Charged Particles. Fields and Linear Optics,
Energoatomizdat, Moscow, 1986 (in Russian).
9. Tarantin, N. I., Nucl. Instr. Meth. B 126 (1997), 392.
10. Tarantin, N. I., In: Proc. of 3rd Intern. Workshop on Application of Lazers in Atomic Nuclei
Researh, Poznan, 3-5 February 1997, p. 149.
11. Brown, L. S. and Gabrielse, G., Rev. Mod. Phys. 58 (1986), 233.
12. Tarantin, N. 1., JINR, P9-88-149, Dubna, 1988.
13. Tarantin, N. I., In: Proc. of III Scientific Seminar in Memory of V. P. Sarantsev, Dubna, 22-23
September 1999, D9-2000-69, p. 180 (in Russian).
~ Hyperfine Interactions 132: 451--456,2001.
,, 451
© 2001 Kluwer Academic Publishers.
Abstract. The results of a numerical simulation of heavy ion beam motion in a multiturn mass-
spectrometer are presented. It consists of a magnet system with an azimuthal variation of magnetic
field. two dees and an open type arc discharge ion source. The calculations have shown that the mass
resolution of ion beams is above 104 .
1. Introduction
There is a number of important problems in the field of experimental atomic and
nuclear physics concerning very precise measurements of atomic masses. Analy-
sers with static magnetic or electric fields are widely used [1], but such devices
based on radio-frequency electromagnetic fields are not so often used [2]. One can
refer cyclotrons to charged particles mass-spectrometers of the latter type [3, 4]
where the resonance condition of acceleration (or deceleration) of an ion with mass
A and charge q in the magnetic field B is determined by the formula Wrf = hwo,
with Wo = qe B / Arne the revolution frequency, Wrf the generator frequency and
h the harmonic number of rf voltage. The phase shift of particle about rf voltage
during acceleration (or deceleration) over N turns is due to the difference of Wo
from the resonance value
(1)
It could be caused by the difference of the average magnetic field distribution from
the isochronous value (f!..B = B - B is ), accelerating frequency error, and time
instability of the guiding field. Let us suppose that the stability of the above values
lies within 10- 6 _10- 5 . Thus, we get the phase shift
. f!..(q/A)
f!.. sm <p = 2rr N h (q / A) . (2)
* This presentation was partially supported by the Russian Foundation for Basic Research, grant
No. 18400.
452 O. S. KOZLOV AND S. I. KOZLOV
The total resolution of a cyclotron (i.e., the phase shift of particle over a range of
±900) for the same charge q is equal to S = AI D-A = rr Nh. For the isochronous
cyclotron U-200 [5] FLNR, JINR, Dubna, the resolving power is higher than 1300
(h = 2, N = 200). One can increase a high S-value by choosing the corresponding
hand N. A proposal to build a specific isochronous cyclotron for particle mass
analysis with a resolution of up to 30000 has been put forward [3]. The method
of ion analysis and storage at a maximum cyclotron radius has been considered, as
well [6]. The generation of radioactive accelerated beams includes a preliminary
target irradiation either in the reactor [7] or in the accelerator [8] and the following
separation of the required isotope by a special optical system.
2. Mass-spectrometer scheme
In this paper, we describe the development of a storage ring of the cyclotron type
for mass-spectrometry purposes. In the presented device, the analysed beam with
definite ratio Am I q e and energy W circulates along the track in the sector-focusing
magnetic field. The orbit radius corresponds to the ion beam rigidity. The ions are
extracted from an arc discharge source of the open type [9], which is at positive
potential to the ground Vs (W = qeVs ), by an extracting electrode - a slit (6, Fig-
ure 1). The pulse operating cycle of the ion source ('" 1 ms) requires pulse gas
delivery and accelerating voltage Vs.
The investigated substance is supplied to the anode cone in the gas state (53, Fig-
ure 1). Aiming to obtain the ions from the solid state, for example, to analyse the
radioactive products from the irradiated targets, one can set the targets in different
ways [10], in particular, on the anticathode (52, Figure 1).
The ion source can be kept open for the analysed beam passage. In this case, one
of the lateral walls is a target, irradiated by the beam from another accelerator, and
the opposite one is a collector of nuclear reaction products. Then, these products are
directed to the discharge, ionized and extracted to the track for the analysis and fol-
lowing deflection towards the detector, or injected to the other postaccelerator [11].
An ion source of this type, used for the heavy ion magnetic mass-separator [12],
has a high efficiency of substance use. The residual pressure in the spectrometer
chamber must be within 10- 8 _10- 7 Torr. This requires powerful vacuum pumping
in the ion source location. In the considered mass-spectrometer, the magnetic field
distribution is formed by an iron yoke and a core excitation coil (Figure O.
- - - - MP
)~
Figure I. Mass-spectrometer scheme: 1 - magnet pole; 2 - coil; 3 - yoke; 4 - sector plates;
5 - arc discharge ion source: 51 - cathode, 52 - anticathode, 53 - anode cone, 54 - filament;
6 - extracting electrode; 7 - dees; 8 - limiting slit; 9 - vacuum chamber; 10 - deflector;
11 - magnetic channel; MP - median plane.
where Eo = 3.1 kGs is a constant field along the radius and coefficient b is equal
to 0.1.
The trajectories of two particles of the same charge q = 1 and the difference
in mass by unity during the 20th tum are shown in Figure 2 (Ao = 101, W =
16 keY, R = 59 cm). The maximum amplitude of the coherent oscillation of the
equilibrium ions with mass A = 102 is about 6 mm. These ions can be separated
by the slit arranged at half-tum. The mass resolution of such a system is about 200.
To increase the resolution, the dees with angle extent C{Jd are located between
the sector plates (Figure 1). The harmonic number h is chosen from the condition:
2krr
h=--, (4)
C{Jd
~W =4q eVd sm 2
. (hC{Jd) =0 (5)
454 O. S. KOZLOV AND S. I. KOZLOV
59.5 r--------~---~---___,
59.0
~58.5
~
58.0
57.5L----~---~---~---~
and stay on the circular track while the ions with a difference either in mass or in
charge are accelerated or decelerated. This leads to radial oscillation excitation and
losses of nonresonant ions on the limiting slit (Figure 1).
The number of turns required for the beam loss on the slit can be estimated as
/j.R W 2 A
Nm = - - - - - - - - , (6)
R qeVd h /j.A
where /j.A is the mass difference from the isochronous one and /j.R the coordinate
of the slit relative to the equilibrium orbit.
A numerical simulation of the spectrometer operation with the above parameters
of the magnetic structure has been carried out. The azimuthal dee extent ({Jd = 45°
(h = 8, k = 1) and the dee voltage Vd = 5 kV were used for the calculations.
The trajectories were obtained for ions with q = 1 since the particles with other
charges are easily separated from the main ones.
Figure 3 presents the dependences of the particles radii versus tum number
N for the isochronous ion of mass Ao = 101 and the ions with different mass
deviations /j.A = A - Ao.
The calculation of the beam envelope for emittance E = 25 mm . mrad shows
a perfect separation for the particles with Ao = 10 1 and A = IO 1.1 by the 20th
tum. It is in good agreement with the N m estimate from the above formula. The
separation for the particles with /j.A = 0.01 requires about 200 turns.
The extracting electrode (6, Figure 1) is one of the radially limiting slit in the
given spectrometer scheme, and another one is arranged at the half-tum down-
stream the source. A low resolution regime with a high beam intensity or, on the
contrary, regimes with different slit widths can be achieved here. The analysis for
other masses is tuned by varying the resonant condition defining stable orbits.
The resolution of the considered mass-spectrometer is not limited by a value
of S = 104 and can rise by increasing rf voltage harmonics and the number of
A SMALL ISOCHRONOUS STORAGE RING FOR SPECTROMETRY 455
61
60
----------M,;:O:2.L.-.-.--.-:~1·0-1·
._._. ___ ._._._-_._.- nu-
59
58~------~------~------~------~
o 5 10 15 20
N
Figure 3. Dependence of the ion radius on the number of turns for AA = 0.01,0.1,0.5, 1.
turns (N = 102 -104 ) under rigid requirements for the stability of the spectrometer
parameters bo V, boB, bof and vacuum conditions.
4. Conclusion
The presented analyser type offers wide possibilities. When higher harmonics of
accelerating voltage are used (h = 102 -10 3 ) together with a considerable decrease
of the quarter-wavelength resonator size, one could obtain a resolution of about
104 -105 . The well-known analysing properties of charged particle accelerators in-
crease in the considered mass-spectrometer by using an ion source located on the
circular orbit. The arc discharge ion source allows one to vary the analysed particle
charge over a wide range (together with the corresponding variation of V,) and,
thus, to resolve the particles with close values AI q. In this case, the molecular com-
pounds contributing a considerable background fall into pieces. The backgrounds
from isobars (for example, BlO-BelO, which differ in mass by units of 10- 5 ), mole-
cules (C 12 H 2 -C 14 ), scattered ions and residual gas are removed in this device. This
fact is of great importance for low energy mass-spectrometers.
We believe that this analyser can be successfully applied to the detection of
10- 16 _10- 9 rare atom isotopic proportions for a stable isotope of the same ele-
ment, to the solution of many important problems in geology, archaeology, cos-
mochronology and paleontology where radionuclei are used as tracers and chrono-
meters of the event happening hundreds of thousands or some million years ago,
and to the analysis of long-lived cosmic radioisotopes as well.
References
1. Tarantin, N. t, Magnetostatic Analysers of Charged Particles, Atomizdat, Moscow, 1986.
2. Lunney, D., Current status of the MISTRAL at ISOLDE, these proceedings.
3. Cleark, D. J., In: X International Conference on Cyclotrons, USA, 1984, p. 534.
456 O. S. KOZLOV AND S. I. KOZLOV
4. Chartier, M., Mass measurements at GANIL using the CSS2 and CIME cyclotrons, this issue.
5. Shelaev, I. A., JINR rapid communication, 9-3988, Dubna, 1968.
6. Kozlov, S. I., JINR rapid communication, 9-87-896, Dubna, 1987.
7. Belmont, J. L., The Grenoble project of radioactive beam facility, ISN, 91-100.
8. Tanihata, I., Phys. Rev. Lett. 55 (1985), 2676.
9. Livingstone, M., Rose, M. N. and Namiac, M., Cyclotron, OGIZ, Moscow, 1948, p. 38.
10. Kutner, V. B. et at., JINR rapid communication, 9-11281, Dubna, 1978.
11. Agapov, N. N., Butenko, A. v., Dinev, D., Khodzhibagiyan, H. G., Kovalenko, A. D., Ko-
zlov, O. S., Mikhailov, V. A., Monchinsky, V. A., Smimov, A. A. and Volkov, V. I., In: Rapid
Cycling Superconducting Booster Synchrotron, VII European Particle Accelerator Conference,
Vienna, Austria, 2000.
12. Ivanov, N. S., Kabatchenko, A. P., Kuznetsov, I. V. and Tarantin, N. I., JINR rapid communica-
tion, P13-9645, Dubna, 1976.
Hyperfine Interactions 132: 457-461, 2001. 457
© 2001 Kluwer Academic Publishers.
Abstract. HITRAP is a planned ion trap facility for capturing and cooling of highly charged ions
produced at GSI in the heavy-ion complex of the UNILAC-SIS accelerators and the ESR storage ring.
In this facility heavy highly charged ions up to uranium will be available as bare nuclei, hydrogen-
like ions or few-electron systems at low temperatures. The trap for receiving and studying these ions
is designed for operation at extremely high vacuum by cooling to cryogenic temperatures. The stored
highly charged ions can be investigated in the trap itself or can be extracted from the trap at energies
up to about 10 keY jq. The proposed physics experiments are collision studies with highly charged
ions at well-defined low energies (eV ju), high-accuracy measurements to determine the g-factor of
the electron bound in a hydrogen-like heavy ion and the atomic binding energies of few-electron
systems, laser spectroscopy of HFS transitions and X-ray spectroscopy.
1. Introduction
Highly charged ions (HCI) play an important role in different fields of science,
e.g., in atomic physics, surface science, plasma physics, nuclear physics and astro-
physics. Recent years have seen a breakthrough of the technological developments
in the field of HCI research. The most advanced and most intense source of heavy
HCI world-wide is the accelerator facility of the heavy-ion research institute GSI,
where stable as well as radioactive heavy HCI can be investigated at MeV ener-
gies. At the HITRAP facility, a broad spectrum of physics experiments with heavy
HCI up to bare uranium (U92+) at low energies « 1 e V lu) is planned which can
presently not be performed at any other institution using light or medium-heavy
HCI or heavy HCI at MeV energies [1].
in a foil to the desired charge state and injected into the Experimental Storage
Ring (ESR). For bare or hydrogen-like ions, energies of a few hundred MeV lu are
required. In the ESR the ions will be decelerated to an energy of 3 MeV lu. It is
planned to demonstrate this low-energy operation mode of the ESR during 200l.
They are then extracted in a fast-extraction mode as short ion bunches.
After beam deceleration and pulsed extraction from ESR, the ions are still too
fast to be directly captured in an ion trap. Therefore, it is necessary to decelerate the
extracted ions. The ions should attain a final energy of the order of 30 ke V I q after
post-deceleration. For post-deceleration two options are presently discussed: (i) a
radiofrequency quadrupole structure (RFQ), and (ii) a linear interdigital H-mode
(IH) drift tube structure. A RFQ decelerator has been commissioned recently at the
Antiproton Decelerator (AD) at CERN to allow antiprotons to be decelerated from
an energy of 5.3 MeV down to below 100 keY [2].
After deceleration to energies below 30 keV I q the ions are captured into a first
Penning trap and cooled (Figure 2). In a Penning trap highly charged ions are stored
in a combination of a homogeneous magnetic field and an electrostatic field [3]. In
the cooling trap the highly charged ions are cooled with a combination of electron
or positron cooling and resistive cooling to a temperature of T = 4 K. The vacuum
enclosure of the cooling trap is cooled to 4 K with liquid helium in order to achieve
a very low background gas pressure and long storage times [4]. The electrodes of
the cooling trap have a cylindrical shape with an inner diameter of about 50 mm.
Two resonant circuits are connected to these electrodes for electronic detection and
resistive cooling of the axial motion and of the cyclotron motion of the trapped
highly charged ions. After cooling to cryogenic temperatures the highly charged
ions can be extracted and transferred to physics experiments. The HITRAP facility
will provide about 106 charges per second, i.e., about 104 ionsls in the case of
NOVEL
INSTRUMENTATION
FOR
ilaser spectroscopy
mass spectrometry
rf spectroscopy
nuclear spectroscopy
Figure I. Schematic of the GSI accelerator complex and the planned HITRAP facility.
Highly charged ions are extracted from the ESR storage ring at 3 MeY ju, decelerated in a
post-decelerator to some keY ju, and then transferred into a Penning trap system.
HITRAP: A FACILITY FOR EXPERIMENTS WITH TRAPPED HIGHLY CHARGED IONS 459
____ , l~I'
~~ ___ ...!l=
_._-------
a) ion enter trap
.....
...
c::
Q)
d) cooling
-15 -10 -5 0 5 10 15 20 25 30 35 40
distance from trap center (cm)
Figure 2. HITRAP facility: Highly charged ions are captured in a Penning trap. The trapped
ions are cooled in collisions with simultenously stored electrons or positrons. The cooled He!
can be non-destructively detected with an electronic detection circuit.
u92+ or 106 U 9 2+ ions in a single shot every 100 s. HITRAP will be a unique
facility enabling novel experiments in many fields of physics.
tance of the beam of slow heavy HCI extracted from the cooler Penning trap will
make it possible to perform a new generation of collision experiments with sub-
stantially increased momentum resolution and kinematically complete detection
of the reaction products. The formation of "hollow atoms", i.e., multiply excited
states of HCI, which are created when HeI approach a surface at close distance, is
of particular interest [7]. The encounter of a HeI with a surface leads to multiple
emission of particles like electrons and photons.
In addition to the collision experiments described above, high-accuracy atomic
physics experiments on trapped HeI are part of the physics programme of the HI-
TRAP facility. The measurement of the g-factor of the bound electron in hydrogen-
like ions is a stringent test of Quantum Electrodynamics (QED) calculations in
extremely strong electromagnetic fields [8]. A precision Penning trap has been
developed in a joint effort of the University of Mainz and GSI [9]. First g-factor
measurements on light hydrogen-like ions have been performed with an accuracy
of 10- 9 in this precision Penning trap [10]. At the HITRAP facility, such g-factor
measurements can be performed up to the heaviest hydrogen-like ions, where the
product of proton number and fine-structure constant (Za) approaches unity. Also
atomic binding energies of HeI will be determined by measuring the masses of
highly charged ions in different charge states with an accuracy of 10- 10 . This will
provide a crucial test of QED calculations and atomic structure theories in extreme
electromagnetic fields [11].
Moreover, precision X-ray spectroscopy of exotic atomic systems which enters
deeply into the fundamental aspects of atomic structure will be performed, e.g., for
He-like uranium ions. Investigations of such ions along the isoelectronic sequence
of ions with different nuclear charge Z probe our understanding of correlation,
relativistic and quantum electrodynamic effects in a unique manner. It should be
pointed out that for high-Z He-like systems such information cannot be achieved at
existing facilities. Cooled and trapped ions will provide ideal experimental condi-
tions for almost Doppler-shift free spectroscopy. This advantage, combined with
a new generation of position-sensitive solid-state detectors for X-ray radiation,
promises a considerable improvement of the currently reached precision in spec-
troscopy studies of high-Z ions. Construction of a new position-sensitive X-ray
detector (for energy range 5-100 keY), is also crucial for precise estimation of
lifetimes of exotic states in highly charged ions.
Finally, high-accuracy laser spectroscopic studies of trapped HCI are foreseen
at the HITRAP facility. The measurement of the hyperfine level Zeeman splitting
in the ground state of hydrogen-like ions will yield nuclear g-factors with high
accuracy. Measurements of the hyperfine splitting (HFS) in hydrogen-like ions
[12] in combination with the measured nuclear g-factor will give information about
the distribution of the nuclear magnetization within the nucleus. The comparison
of nuclear g-factors in hydrogen-like ions with the nuclear g-factors in neutral
atoms will make it possible to test calculations of the diamagnetic shielding cor-
rection for the first time. Moreover, the manipulation of trapped radioactive ions
HITRAP: A FACILITY FOR EXPERIMENTS WITH TRAPPED HIGHLY CHARGED IONS 461
with laser light opens up possibilities to study questions of the Standard Model
of fundamental interactions in a unique way at the HITRAP facility. By optical
pumping within the HFS levels of the ground state, the nuclear spins of radioactive
nuclides can be polarized with high efficiency. The detection of the asymmetry of
beta decay, for example, will allow one to explore the V-A-structure of the weak
interaction and to set limits for the masses of heavy bosons which are not included
in the Standard Model. These experiments will give important contributions at the
border-line between atomic and nuclear physics.
References
1. http://www.gsi.de/eurotraps/hitrap.htm.
2. Lombardi, A. M., In: Proc. 1998 Linear Accelerator Conj., Chicago, 1998, p. 377.
3. Brown, L. S. and Gabrielse, G., Rev. Modern Phys. 58 (1986), 233.
4. Diederich, M., Haffner, H., Hermanspahn, N., Immel, M., Kluge, H.-J., Ley, R., Mann, R.,
Quint, W, Stahl, S. and Werth, G., Hyp. Interact. 115 (1998),185.
5. Cassimi, A., Phys. Scripta T80 (1999), 98.
6. Arnau, A., Aumayr, F., Echenique, P. M., Grether, M., Heiland, W., Limburg, J., Morgen-
stern, R., Roncin, P., Schippers, S., Schuch, R., Stolterfoht, N., Varga, P., Zouros, P. J. M.
and Winter, H., Suif. Sci. Reports 27 (1997), 113.
7. SchlathOiter, T., Narmann, A., Robin, A., Winters, D. F. A, Marini, S., Morgenstern, R. and
Hoekstra, R., Phys. Rev. A 62 (2000), 042901.
8. Quint, W, Phys. Scripta T59 (1995), 203.
9. Hermanspahn, N., Haffner, H., Kluge, H.-J., Quint, W, Stahl, S., Verdu, J. and Werth, G., Phys.
Rev. Lett. 84 (2000), 427.
10. Haffner, H., Hermanspahn, N., Kluge, H.-J., Quint, W, Stahl, S., Verdu, J. and Werth, G., Phys.
Rev. Lett. 85 (2000), 5308.
11. Haffner, H., Hermanspahn, N., Indelicato, P., Kluge, H.-J., Lindroth, E., Natarajan, v.,
Quint, W, Stahl, S., Verdu, J. and Werth, G., Hyp. Interact. 127 (2000), 271.
12. Seelig, P., Borneis, S., Dax, A., Engel, T., Faber, S., Gerlach, M., Holbrow, c., Huber, G.,
KUhl, T., Marx, D., Meier, K., Merz, P., Quint, w., Schmitt, F., Tomaselli, M., Volker, L., Win-
ter, H., WUrtz, M., Beckert, K., Franzke, B., Nolden, F., Reich, H., Steck, M. and Winkler, T.,
Phys. Rev. Lett. 81 (1998), 4824.
Hyperfine Interactions 132: 463-468,2001. 463
© 2001 Kluwer Academic Publishers.
Abstract. The ion trap facility SHIPTRAP is being set up to deliver very clean and cool beams
of singly-charged recoil ions produced at the SHIP velocity filter at OSI Darmstadt. SHIPTRAP
consists of a gas cell for stopping and thermalizing high-energy recoil ions from SHIP, an rf ion
guide for extraction of the ions from the gas cell, a linear rf trap for accumulation and bunching
of the ions, and a Penning trap for isobaric purification. The progress in testing the rf ion guide is
reported. A transmission of about 93(5)% was achieved.
PRELIMINARY PURIFIED
ION BUNCHES ION BUNCH
""3m
Figure 1. The overall SHIPTRAP configuration.
axis by an rf field while the noble gas is pumped away. An axial field within the
structure guides the ions along the axis towards the bunching system. The recent
development for the SHIPTRAP stopping chamber and the extraction RFQ is de-
scribed elsewhere [2]. In the ion bunching system, aIm quadrupole rod structure
immersed in a low-pressure buffer gas, the ions are trapped by a proper choice
of axial dc and transverse rf fields and cooled in collisions with the buffer gas.
The RFQ-buncher has been set up and is currently being tested. The purification
system into which these preliminary bunches are collected is based on a Penning
trap similar to the one used for this purpose at the ISOLTRAP facility at ISOLDE.
In such a system the contaminating isotopes are very effectively suppressed due to
the high mass resolving power of the cooling process.
6
,
~,.
•
\
•I.•
,••
<f s I
E.
.~
c
!• •
...... 4
••
Q)
I•
::J
u
c 3 • ~
.Q
/ •..
\
""C
.... •
••
d)
~ (J)
2
•II \
•.,
•
.......cm
•
I
~
I
•••
0
0,0 02 0,4 0 ,6 0,6 1 ,0 1,2
r.f. amplitude (q-parameter)
Figure 2. Transmission plot in rf-only mode for Ar+ ions.
for Ar+ions. The driving field frequency was set to Wrf = 600 kHz. In simulations
it was shown that the left edge of the transmission curve is affected by beam di-
vergence, whereas the right edge is rather unaffected [3]. A lower transmission for
low q values as beam divergence grows can be explained by the fact that ion entry
angle grows; hence the probability of hitting the rods increases. The right edge
is affected by beam width effects, becoming less abrupt as beam width grows. In
simulations it was shown that this effect is more pronounced when the ion beam is
not centered or annular. As a conclusion our transmission curves are indicating an
off-axis ion source with low emittance.
Another interesting point becomes evident when we zoom into the center region
of the stability diagram. In Figure 3 we benefit from a segmented Faraday Cup
that gives us a closer insight in the shape of the extracted ion beam of Figure 2.
The Faraday cup is divided into three segments, a 3 mm diameter inner section,
a 3-8 mm diameter second ring, and an outer ring from 8 to 12 mm diameter.
In Figure 3 one can see the detected ion currents. The sequence of maxima and
minima on the central plate and the second ring shows how the beam emittance is
affected by the operating point. The current on the third plate is low and not shown.
The fundamental task of the RFQ buncher at SHIPTRAP is to cool the ions from
the stopping chamber and to collect them, Since the RFQ buncher will therefore
be operated under buffer gas it is necessary to investigate the influence of gas on
the ion motion. The average effect of ion collisions with buffer gas molecules can
be approximated by a frictional drag force. This leads to the usual form of the
Matheu equation but with an added velocity dependent term. Figure 4 shows three
measured transmission curves at different pressures. One can see a tendency that
466 G. MARX ET AL.
«"
.£.4
....I: i~i
i ~./
r\ J-"\ i-\ ,.\
• i·"
I 1.;1
_ inner section
......
Q)
i3 3 Ii.,.
....\ fl .~
: .... \
\. 1 \ I 1·\~ ~ ~ .:-__
0:/\ JI ~\rr111°t\~ /I F"Lr~~
I:
.2
'"C
\ /\ <tP. m-
....
....
Q)
'E
I/)
2
2 0d <edi"
.......lU
I:
<0\
~ <D \J
ro
bJ
0
m \/
<D \F
m
0,2 OJ o~ o~ 0,6
rJ. amplitude (q-parameter)
Figure 3. Transmission plot in rf-only mode for Ar+ ions. Various maxima and minima on
a segmented Faraday cup behind the RFQ showing a dependence of the emittance on the
rf amplitude.
1,0
~
::J
0,8
~
....c
....
d)
.... 0,6
::J
U
c
.9 0,4
]
:::
'EI/)
0,2
c
........lIS 0.0 -
0 ,0 0.1 0,2 0,3 0,4 0,5 0 ,6 0.7 0,8 0,9 1 JJ 1,1 1,2 1,3
the right edge of the stability diagram is shifted to higher q-values. Due to the
damping of the ion motion one can apply higher quadrupole field strength until
the ion motion is unstable. On the other hand it should be noted that the pressure
at which the stability region starts to broaden is typically near the minimum in
the Paschen curve for voltage breakdown of the gas being used. In fact the range
of investigation shown in Figure 4 did not reach the minimum in the Paschen
curve since the used gas fed cross beam ion source is affected by an increase of
the buffer gas pressure. We will repeat this investigations with a slightly changed
setup that will provide us with better differential pumping between ion source and
buncher. The information how much the stability region is increased under buffer
gas operation is important since one tries to use the limited mass resolution of
the RFQ in rf only mode to suppress contamination of a lighter ion species. The
qualitative overall effect of space charge on an ion beam in an RFQ rod structure,
neglecting the size of an ion beam due to thermal motion of the ions, will push
out the ions until those at the periphery of the beam experience an effective radial
electric force from the confining field that balances the radial expansion force of
the space charge. Assuming a uniform charge distribution along the beam axis, the
space charge electric field at the edge of a beam of radius r is balanced by the
effective trapping field electric field. This results in a charge density
4. Summary
The SHIPTRAP facility at GSI Darmstadt is designed to slow down heavy-ion
projectiles from the velocity filter SHIP to thermal energies, to accumulate and cool
them in an ion trap system and to deliver these ions as isobarically pure ion bunches
with low emittance to different physics experiments. After an intense simulating
and construction phase first components are set up and under test. SHIPTRAP will
start operation in Summer 2001. The experimental programme which is envisaged
by the SHIPTRAP user community promises to give new insights into the nuclear,
atomic and chemical properties of the transeinsteinium elements.
Acknowledgement
We acknowledge financial support by the European Union (Network EXO-
TRAPS).
References
1. Miinzenberg, G. et al., Nucl. Instrum. Methods 161 (1979),65.
2. Engels, O. et al., this issue, 505.
3. Muntean, F., Internat. 1. Mass Spectrom. Ion Processes 151 (1995), 197-206.
4. Moller, P. et al., At. Data Nucl. Data Tables 59 (1995), 185.
5. Dilling, J. et al., this issue, p. 331, p. 495.
Hyperfine Interactions 132: 469-472, 200l. 469
© 2001 Kluwer Academic Publishers.
Abstract. The beam quality of radioactive ion beams produced by present target ion source technol-
ogy is often not sufficient for direct post-acceleration. Furthermore. pulsed beams insure a more
efficient use of an accelerator. In the case of REX-ISOLDE, the post accelerator at the CERN
ISOLDE facility, a gas-filled Penning trap (REXTRAP) has been chosen for accumulation of the
radioactive ions and conversion into cooled bunches. Radial centering of the ions is achieved by
applying an rf field with a frequency equal to the cyclotron frequency of the desired ion species.
The efficiency achieved in the first tests with different isotopes covering nearly the entire mass range
was already> 20%. Going to total numbers of > 105 stored ions in the trap a shift of the centering
frequency could be observed, which is most likely due to space charge effects. Despite this, it was
possible to accumulate up to 107 ions and deliver them as cooled bunches.
Key words: Penning trap, ion beam cooling, space-charge, radioactive ions, post acceleration.
1. Introduction
The possibility of reducing the transversal phase space volume occupied by ions
confined in a Penning trap has been studied and experimentally proven in [1-3].
The principle is to reduce the ion energy via the dissipative force provided by the
interaction with a buffer gas. To avoid an increase of the radius of the unstable
magnetron motion (11->-) it is coupled to the reduced cyclotron motion (w+) by
applying an azimuthal quadrupolar rf field at the cyclotron frequency We = W+
+ W_. This results in a reduction of both radii and a centering of the ions in the
trap. As this process relies on the ions' cyclotron frequency it is mass selective
and can be used for mass purification. It can be described analytically if only a
small number of ions is involved and no interaction between these ions has to be
considered [4]. For a higher number of ions or dense ion clouds such dependencies
can be found by ray tracing for individual ions and scaling of the Coulomb force
[5] or by methods of plasma physics [6].
* Corresponding author.
470 F. AMES ET AL.
REXTRAP, a penning trap operating under such conditions, is the first part of
the REX-ISOLDE project [7]. It is used to cool and fonn bunches of radioactive
ions to prepare them for the injection into an electron beam ion source (EBIS) and
further post acceleration. The ions are delivered as a continous beam at 60 ke V
from the mass separator ISOLDE at CERN. Capturing and accumulation is done
by operating the trap at high voltage for an electrostatic deceleration and providing
the final energy loss by a buffer gas.
35000
30000
(/)
r:: 25000
.2
'C
~ 20000
u
Q)
'Q)
15000
'0
t
Sl 10000 -0-- p=8.6·10" mbar
E
:::I
r::
5000
-+--- p=6.2·10" mbar
- - - p=4.5·10" mbar
0
0 5 10 15 20 25 30 35 40
tcoo,[ms]
Figure I. Number of detected 133Cs+ ions as function of the cooling time for different Ar
buffer gas pressure in the trap.
REXTRAP 471
::i
~ +
+
I/)
c
.2
+
+
.
'0
CD
.Q
+
+
E ~+
~
c +~
~~
342000 343000 344000 345000
V R• [Hz}
Figure 2. Number of 133Cs+ ions ejected from REXTRAP as a function of the applied radio
frequency.
356000
354000
~ 352000
N
:I:
;: 350000
0
c
~ 346000
cr
~ 346000
344000 I I II
342000
10' 10' 10· 107
number of ions
Figure 3. Center frequency of the 133Cs+ cooling resonances as shown in Figure 2 as function
of the total number of ejected ions, the bars represent the half-width of the resonance line.
amount of ions near to the trap center and thus a reduction of transversal phase
space volume. Taking the full width at half maximum from this curve a mass
resolution of about 300 is found. From the analytical treatment done in [4] it is
mainly determined by the strong damping force at this buffer gas pressure. This
measurement has been done with a total number of ions below 104 within one
pulse. Going to higher ion numbers space charge effects start playing an important
role. As shown in Figure 3 a significant broadening and shift in the frequency
dependence is observed.
The efficiency of the set-up could be determined to up to 30% for ISOLDE
beams and up to about 50% for ions from the test ion source. It was found that the
efficiency also depends on the number of accumulated ions. For higher numbers
the efficiency decreases, indicating an increase of the ion clouds radius.
472 F. AMES ET AL.
Acknowledgements
This work is supported by the "Bundesministerium flir Bildung und Forschung"
(BMBF) of the Federal Republic of Germany with grant number 06MZ866I and
by the European EXOTRAP network.
References
1. Bollen, G. et ai., 1. Appl. Phys. 68 (1990), 4355.
2. Savard, G. et aI., Phys. Lett. A 158 (1991), 247.
3. Raimbault-Hartmann, H. et aI., Nucl. Instrum. Methods B 126 (1997), 378.
4. Konig, M. et al., Internat. 1. Mass Spectr. Ion Pro 142 (1995), 95.
5. Beck, D. et aI., this issue, p. 473.
6. Dubin, D. H. E. and O'Neil, T. M., Rev. Modern Phys. 71 (1999), 87.
7. Habs, D. et aI., NucZ. Instrum. Methods B 126 (1997), 218.
8. Ames, F. et aI., In: B. M. Sherril, D. J. Morrissey and C. N. Davids (eds), Proc. Exotic Nuclei
and Atomic Masses, Bellaire, USA, 1998, AlP Conf. Proc. 455, Amer. Inst. Phys., New York,
1998, p. 927.
9. Schmidt, P. et al., Nue!. Phys. A, in press.
Hyperfine Interactions 132: 473-478, 200l. 473
© 2001 Kluwer Academic Publishers.
Abstract. Mass selective buffer gas cooling is a technique used for ions that are stored in a Penning
trap. The technique can be applied to all elements and the mass resolving power achieved has proven
to be sufficient to resolve isobars. When not only a few but 106 and more ions are stored at the same
time, space charge starts to playa dominant role for the spatial distribution. In addition, the observed
cyclotron frequency is shifted. This work investigates these effects by numerical calculations.
Key words: Penning trap, buffer gas cooling, numerical calculation, space charge.
8 8
E
~6
en
:::>
is
~ <4
z
~
2
2 2
(a) (b)
Figure 1. Calculated mean radius of 106 ions (a) and the width of the ion cloud in the axial direction
(b) as a function of the excitation frequency.
are calculated at the same time so that each ion interacts with all the others. The
equations of motion are solved by numerical integration. Due to limited CPU time
this method is only applicable up to a few thousand ions. In order to get results for
up to 107 stored ions the Coulomb force is scaled as
nstore
Fe.eompute = . Fe,
neompute
where Fe,compute is the scaled Coulomb force used in the calculation, Fe is the true
Coulomb force between two ions, nstore is the number of stored ions and ncomputc
is the number of particles used in the calculation. For the computation, a fifth-
order Runge-Kutta method with adaptive step size control was adapted from [6].
The calculations were performed on two PC clusters at K.U. Leuven, Belgium and
at GSI-Darmstadt, Germany. Around 40 CPUs have been used for approximately
8 weeks. The uncertainties in this work represent the statistical error.
To test the scaling approach different numbers of particles in a series of cal-
culations were used to extrapolate to 106 stored ions. A cloud of particles with
homogeneous density and a diameter of about 1.0 mm was placed in the trap and
allowed to expand for 2 ms. Then, the width of the ion cloud in the radial and axial
direction was determined. Table I shows the results. There is a good agreement
for the different computed particle numbers, although the size of the ion cloud for
small particle numbers could be slightly underestimated.
Table I. Test of the scaling behavior of the Coulomb force to 106 stored ions. The
first column gives the number of particles in the calculation. The scaled Coulomb
force is given in the second column. The width of the ion cloud in the radial and
axial direction after 2 ms is shown in the last two columns
30
25 ;[ 25
...-
...-
~20
E
...- ...- ...-
E
...- _-1
';15
:r ...- ...-
...-
...
~
10 ~
...-
/ /
5~ 5~
ol!- O~--------~----~--~----~~
0 2e+07 4e+07 68+07 80+07 1.+08 o 2e+07 4e+07 68+07 88+07 1e+08
I OF STORED IONS I OF STORED IONS
(a) (b)
Figure 2. Calculated size (FWHM) of the ion cloud in the radial (a) and in the axial direction (b) as
a function of ion number without excitation after 10 ms.
NO EXCITATION
Starting from a distribution with ~r.z.FWHM = 0.5 mm in the radial and in the axial
direction the ion cloud was allowed to evolve for 10 ms. Then, the width of radial
and axial extension is determined. It can be seen from Figure 2 that the ion cloud
expands with increasing ion number. Similar investigations have been done for Paul
traps [7].
Starting with a size of ~r.z.FWHM = 0.5 mm in the radial and in the axial direction
the ions were excited for 10 ms with a dipole field in the radial plane and driven to
larger radii. When calculating the mean radius as a function of excitation frequency
a resonance signal is obtained. The center frequency results from a Gaussian fit to
the data points. Figure 3 shows the center frequency as a function of the ion number
for excitation of the magnetron and cyclotron motion. The observed frequency
shifts are smaller than 100 Hz.
476 D. BECK ET AL.
'N 34SOOO
N' .:.
>-
~ ~ u
z
....
344950
~ rJ.-'i-. :::>
.., '!IOII
....0II: 344900
:::>
..,o ... T __ _
---I
...
II: z
0 344850
/1.. - - __
!!:
-I
Z
o
II:
WiG
....
0
u 344800
Z >-
u
~ ..,d 344750
II:
(a) (b)
Figure 3. Dipole excitation. Calculated magnetron (a) and reduced cyclotron (b) frequency as a
function of ion number obtained from resonance signals (see text).
3!>4000 3!>4000
N' /
N
'">-
~352000
'">-
~352000
/
...
/
u u
..,z ....:::>z
:::>
I
...~J50000
II:
~ _..... -- 8350000
...
II:
'"~
J.
~ 348000 ~ 348000 I
u u
~
.L
.1
346000 346000
0 1.+07 2.+07 38+07 0 1.+07 2.+07 38+07
, or STORED IONS , OF STORED IONS
(a) (b)
Figure 4. Quadrupole excitation. Calculated shift of cooling resonances as a function of ion number
(see text). The initial distribution in the radial and axial distribution is 15 mm (a) and 5 mm (b).
Cooling resonances were calculated for different ion numbers. The left graph in
Figure 1 gives an example for such a cooling resonance. Here, 106 ions were
excited for 20 ms. Two series of resonances were calculated, one with an initial
distribution of b.r,z,FWHM = 5 mm and one with b.r.z.FWHM = 15 mm. The result
of the calculation is shown in Figure 4. It can be seen that the observed resonance
frequency shifts to higher values with increasing ion number. If one compares the
shift for the two different radial distributions, one observes a larger shift for the
smaller distribution that has a higher space charge density.
Another information that can be extracted is the size of the ion cloud in reso-
nance after cooling. Figure 5 shows the width in the radial and axial direction as a
function of ion number. It can be seen that the extension of the ion cloud increases
with ion number. However, one should pay attention to the fact that the extension
SPACE CHARGE EFFECTS IN A GAS FILLED PENNING TRAP 477
15 15
I
I
_-ID
,..m-
(a) (b)
Figure 5. Quadrupole excitation. Calculated size of the ion cloud in resonance after 20 ms. Shown
is the radial (a) and axial (b) width as a function of ion number for two different initial distributions
(circles: L'.r.z,FWHM = 15 mm, squares: L'.r,z.FWHM = 5 mm).
in the radial as well as in the axial direction depend on the excitation frequency as
shown in Figure I.
If 106 _107 ions are stored in a trap at the same time, the repulsive Coulomb force
can lead to a expansion of the ion cloud if the initial dimension is small and no
rf-field is applied. The final size of the ion cloud depends on the ion number.
When exciting the ion cloud with a radial dipole field the calculated magnetron
(reduced cyclotron) resonance frequencies for 107 stored ions shift by less than
30 Hz to lower (higher) values for the parameters used here.
Compared to the two cases above, the charge density of the ion cloud is in-
creased when quadrupole excitation is used to center the ions in the trap. For
the parameters used here, without introducing any trap imperfection except the
Coulomb interaction of many ions, the observed frequency shifts of the cooling
resonances are in the order of a few kHz. The size of the shift as well as its sign
agree well with measurements that were made with the REXTRAP set-up [4]. The
frequency shift does not only depend on the number of the stored ions but also
on the size of the ion cloud at the beginning of the cooling process. Comparing
the sum of the frequency shifts ~w+ and ~w_ for the two radial eigen motions
with the calculated sum frequency shift ~wc, no agreement is observed. However,
one has to keep in mind that the size of the ion cloud, and by this also the charge
density, is much different for dipole and quadrupole excitation.
To conclude, extending mass selective buffer gas cooling to large ion numbers
can only be done if large final ion clouds and frequency shifts are tolerated.
478 D. BECK ET AL.
References
1. Savard, G. et al., Phys. Lett. A 158 (1991), 247.
2. Konig, M. et al., Internat. J. Mass Spectr. Ion. Proc. 142 (1995), 95.
3. Ames, F. et al., In: Proc. Exotic Nuclei and Atomic Masses, Bellaire, USA, 1998, AlP Conf.
Proc. 445, Amer. Inst. Phys., New York, 1998, p. 927.
4. Ames, F. et aI., this issue, p. 469.
5. Brown, L. S. and Gabrielse, G., Rev. Modern Phys. 58 (1986), 233.
6. Press, W. H. et al., Numerical Recipes in C: The Art of Scientific Computing, Cambridge
Univ. Press.
7. Meis, C. et al., Appl. Phys. B 45 (1988).
Hyperfine Interactions 132: 479-484, 200l. 479
© 2001 Kluwer Academic Publishers.
Abstract. New developments in ion cooling and ion bunching allow the trapping of radioactive ions
from a low energy beam, in a very short time scale and with very good efficiency. The performances
of Radio Frequency Quadrupole (RFQ) coolers, recently developed in several laboratories in Europe
and USA, are now better known. In the present study, a RFQ device will be used to inject 6He+ ions
in a transparent Paul trap. The ion beam will be delivered by the SPIRAL facility at GANIL. The
careful measurement of the tJ-recoil ion coincidence spectrum is sensitive to the angular correlation
parameter a, which depends on the coupling constants of the weak hamiltonian. It should be equal
to -1/3 if the interaction is only of axial vector type (V-A theory). Deviation from this value would
imply a new tensor-like interaction involving a new exchange boson thus introducing new physics
beyond the Standard Model.
Key words: ion trapping and ion cooling techniques, tJ-v angular correlations.
1. Introduction
We intend to trap 6He+ ions delivered by the future radioactive beam facility
SPIRAL installed at GANIL in an original transparent Paul trap to study the Ga-
mow-Teller (GT) ,B-decay spectrum. The shape of this spectrum is sensitive to
the ,B-v angular correlation parameter, which contains information on the nature
of the weak interaction. In general, due to Lorentz invariance, the interaction may
be of five different types respectively denoted Scalar (S), Vector (V), Axial (A),
Tensor (T) and Pseudoscalar (P). Data suggest a strong dominance of two of these
interactions (V for Fermi transition and A for Gamow-Teller transition) mediated
by the gauge bosons W's. The angular correlation parameter a is sensitive to two
other types of interaction (S for Fermi transition and T for GT transition) which
could be present at a small but sizeable level. They could be mediated by new
bosons, leptoquarks [1]. Such small effects require a very precise measurement of
the kinematics, hence the use of a Paul trap to study the decay of 6He+ nearly at
rest.
480 P. DELAHAYE ET AL.
It is now possible to inject radioactive ions at low energy in a trap with high
efficiency. The RFQ cooler is an ion cooling device capable to slow down ions from
a few ke V to less than leV while keeping a reasonable beam emittance. Three of
these devices have already been built and are in operation at Argonne (CPT penning
trap), at JYFL (Finland) and ISOLTRAP (CERN) [2]. In our case, the transparent
Paul trap will be at the center of the experimental set-up: it will allow to store
radioactive ions at low energy in a well defined volume, and will be surrounded by
detectors for both the recoil ion and the electron. At least, two parameters for each
event will be measured simultaneously so that the whole kinematics can be recon-
structed. It should be noted that the only experiment where the whole kinematics
of each fJ decay was measured so far, was at TRIUMF where coincidences were
measured for the Fermi transition of 38m K, using a Magnetic Optic Trap (MOT) [3].
Other previous experiments concerning the measure of fJ decay spectrum and fJ-v
angular correlations consisted on measuring only a direct or indirect integrated
recoil ion energy spectrum. Therefore, the coincidence technique is expected to
give a better sensitivity on the determination of a.
(1)
where E e, Pe, m are respectively the total energy, the momentum and the mass
of the electron, and E v , Pv, Q ev are respectively the energy of the neutrino, its
momentum and the angle between the v and the fJ particle. Wo is the phase space
distribution as given by the Fermi theory of fJ decay. Last, b is the so-called
Fierz interference term and a is a function of the different coupling constants C i
associated with the various terms of the weak Hamiltonian discussed above.
For a pure GT transition, one has:
(2)
Q/l=3.5097MeV
6L·
3 I
Figure 1. The pure GT decay scheme of 6He. Note the life time of about 800 ms which is
long enough for cooling and trapping.
WEAK INTERACTION STUDIES USING A PAUL TRAP 481
The V-A theory [1] assumes time reversal invariance, maximal parity violation
and the non existence of scalar or tensor-like interactions, thus: with C i and C; the
coupling constants of parity conserving and not conserving Hamiltonian, C; = C;
are real quantities, C A ;:; -Cv = 1 and C s = C T = 0 so aGT = -1/3. The present
constraints on exotic tensor contributions are [5, 6] for a compilation: ICT/CAI <
0.13 and IC~/C~I < 0.12. These limits were obtained studying the pure GT decay
of 6 He.
We intend to search for the existence of tensor interaction by measuring again
the decay of this nucleus once trapped inside a Paul trap. The main advantages of
studying 6He are:
The kinetic energy of the recoil nucleus is large enough leading to a good
detection efficiency.
The transition is a pure GT (see Figure 1) with a branching ratio to the ground-
state close to 100%.
3. The injection from SPIRAL to the Paul trap: the RFQ cooler buncher
The injection of 6He+ from the future low energy beam line of SPIRAL to the Paul
trap requires specific attention, since the ions are trapped with very low energy (less
than 1 e V) and thus must be carefully decelerated by several tens of keY. With
the assumed characteristics of the beam (109 per second 6He at 30 ke V with an
emittance of 80n mm rnrad), the strong deceleration of the ion beam from 30 ke V
to less than 1 eV would cause beam emittance spreading by a factor close to 200.
This excludes efficient trapping. Thus, the use of some cooling device is absolutely
necessary.
A RF quadrupole cooler buncher is now being built at LPC Caen [7]. The
principle of this apparatus is shown in Figure 2. It consists of a segmented RF
quadrupole filled with a so-called buffer gas for thermalisation [8]. A set of DC
potentials creates a potential well in the axial direction on the different segments.
Thus, the ions are cooled and bunched inside the device. After this cooling step,
another set of DC potential allows to extract the ions.
Figure 2. The principle of the RFQ cooler. The main functions of this apparatus are ions
guiding (RF quadrupole potential), cooling (thermalisation gas) and bunching (DC potentials
on segments creating a potential well).
482 P. DELAHAYE ET AL.
This device would reduce considerably the beam emittance during the last de-
celeration step (ions are decelerated from 100 eV to trapping and thermalisation)
and group them as bunches before the trap. For a good trapping efficiency (about
20% has been obtained by simulation), the ions must be thermalised (less than
1 e V) in the quadrupole, and grouped as bunch of 1 ~ in a short time less than
a few hundreds ms. The new performances recently obtained with such an appa-
ratus in other labs of the EXOTRAP Network show that the present device would
correspond to these requirements [9].
The major difficulty of this ion cooling method is the choice of the thermali-
sation gas: the atoms of buffer gas must be much lighter than the ion to be cooled
(here the 6He) for fast efficient cooling. Simulations with Monte Carlo microscopic
approach including charge exchange cross section exclude the use of 4He as the
thermalisation gas [7], the resonant charge exchange between the two isotopes
reducing the transmission to zero. The only relevant buffer gas is thus H 2 , for which
a transmission of about 10% is expected [7, 10].
ozr
The RFQ cooler and decelerating electrode
in the high voltage platform
8 r= 1148ro
gas inlet
30kV
Figure 3. The whole decelerating set-up.
483
.• .
WEAK INTERACTION STUDIES USING A PAUL TRAP
SSD
Paul trap
~~.~
..... ' .. ..... . .~ ... .
IlCP
I a-.
I-----i ~ telescope
Figure 4. The detection set-up. The energy and the position of the electron are measured by a
telescope consisting of a striped silicon detector followed by a plastic scintillator. The recoil
ion position and time of flight (whose start signal is delivered by the electron detector) are
measured by a set of MCP and delay lines anode. The angle is extracted from the position of
the ion cloud and the two relative positions of the particles.
parameters. More, the residual energy of these ions is predicted to be less than
one e V. This residual energy is quite small compared to the energies involved in
the decay (6He maximal recoil energy is 1.4 keV). The first trapping experiment
with a laser plasma source shows that the trap has a capacity of at least 103 -1 04
ions.
6. Conclusion
Progress in the techniques of ion and atom cooling in the last ten years has made
possible new precision experiments on spectroscopic or decay properties of exotic
484 P. DELAHAYE ET AL.
nuclei. For the first time, kinematically complete fJ decay spectra can be measured
so that new accurate measurements on the parameters predicted by the Standard
Model description of the weak interaction can be undertaken. We reported here on
our progress in the design of an experiment dedicated to the fJ-v angular corre-
lations measurement. The good expected efficiencies of ion cooling and trapping,
evaluated from simulation for the ion 6He and experiments in other labs for other
ions, combined with the high production rate of 6He by the SPIRAL source let us
hope that very small statistical error on the a parameter could be obtained. To be
competitive, the uncertainty on a must be less than 1% including both statistical
and systematic errors. Intensive simulations are being undertaken to estimate the
experimental conditions under which such an accuracy could be reached.
The tests of the RFQ cooler buncher will start at the end of this year with
H2 cooling. In parallel, tests of the electron telescope will be soon performed
and the efficiencies of the detection set-up will be evaluated on the basis of both
simulations and experiments.
Acknowledgements
The development of the RFQ cooler buncher has been made possible by the col-
laboration with the EXOTRAP Network.
References
1. Langacker, P. (ed.), Precision Tests of the Standard Electroweak Model, World Scientific,
Singapore, 1998.
2. This issue.
3. Gorelov, A. et aI., Hyp. Interact. 127 (2000), 373.
4. Jackson, J. D., Treiman, S. B. and Wyld, H. w., Jr., Nuclear Phys. 4 (1957),206.
5. Johnson, C. H., Pleasonton, F. and Carlson, T. A., Phys. Rev. 132 (1963), 1149.
6. Boothroyd, A. I., Markey, J. and Vogel, P., Phys. Rev. C 29 (1984),603.
7. Szerypo, J. et al., LPC Caen report LPCC-99-18, 1999 (unpublished).
8. Lunney, M. D. and Moore, R. B., Intemat. 1. Mass Spec. 190 (2000), 153.
9. Nieminen, A. et aI., to be published in NIM.
10. Ban et aI., in preparation.
11. Zhao, W. Z. et aI., Hyp. Interact. 108 (1997), 483.
Hyperfine Interactions 132: 485-490,2001. 485
© 2001 Kluwer Academic Publishers.
Abstract. A new HIGISOL chamber devoted to the study of short-lived products from heavy-ion-
induced fusion-evaporation reactions is proposed. It enables, via the extraction of ions by means of
a SPIG (SextuPole rf Ion Guide), to improve the mass resolving power by a factor 2.5 compared to
the previous system using a skimmer-ring assembly. The gas cell was also equiped with an electric
field for faster transportation of recoiling ions to the nozzle where they are ejected with the gas jet.
The first results obtained both with a radioactive a-source and cyclotron beam will be reported.
Key words: ion guide, mass separator, nuclear spectroscopy, radioactive nuclei.
1. Introduction
The HIGISOL system implemented at the K = 130 cyclotron in JyvaskyUi since
1996 gives readily, for short-lived products of heavy ion (H.I.) induced fusion-
evaporation reactions, a yield of '"'-' 100 ions/(s mbarn J..LApart) [1]. The technique is
fast (ms range), applicable to all elements, but, as the overall efficiency is rather
low ('"'-' 1%), its use is presently restricted to very refractory elements where no
ISOL techniques can compete. The survival time of singly-charged ions is reduced
to a few ms in high pressure He buffer gas needed to stop high energy recoils.
Therefore, it is of utmost importance to diminish the mean evacuation time of
reaction products and for this purpose, the addition of an electric field guidance
has proven to be an appropriate solution [2].
2. Experimental results
2.1. DESCRIPTION OF THE NEW CHAMBER
Figure 1 shows the new target- and recoil chamber assembly installed on-line at the
Jyvaskyla cyclotron. The main differences, regarding our previous set-up described
486 R. BERAUD ET AL.
in [1] are both a larger He pressurized recoil chamber with its axis parallel to the
He flow and a primary beam stopper outside (replacing the channel). According to
the He flow simulations using the techniques described in [3], this long chamber
should provide much more homogeneous He evacuation than our previous one [4].
In counterpart, the effective stopping volume is reduced from "-' 100 to "-'60 cm3 .
In order to prevent the cyclotron H.I. beam from interacting with buffer gas, we
have installed a cylindrical beam stopper (<Pe = 7 mm) placed at a 5 mm distance
from the havar entrance window (thickness = 2 mg/cm2 ). We have shown that the
overall efficiency of the system remains constant within a large range of 36 Ar beam
intensity (0.8-4.0 . 10 12 pps). This clearly demonstrates that the "plasma effect"
(if any!) is now kept independent of the primary beam intensity. Moreover, this
intensity is no longer limited by the exit window survival as it was in our previous
chamber with the channel.
The transportation of ions from the pressurized chamber to the high vacuum region
using a radiofrequency sextupole was proposed in [5] and also in [6] at JyvaskyHi.
Such a device has been successfully installed at Leuven [7]. The SPIG (SextuPole
Ion Guide) allows a much larger pumping speed of buffer gas in the extraction
region providing lower density and therefore less collisions between gas atoms and
ions to be guided towards the acceleration stage. Consequently, this results in two
advantages:
(i) a much smaller energy spread ("-' 1 eV) compared to the skimmer system
("-'100 eV), and
RIB
1 diaphragm
2 targec
3 beam stopper
4 havar w·ndoll'
5 SPIG
(ii) the possibility to use higher buffer gas pressure, increasing thus the stopping
efficiency of the cell.
In case of 500 hPa He pressure, the Mass Resolving Power (MRP) obtained with
the SPIG is about llO0 whereras it was only ",400 using the skimmer. More-
over, the MRP is independent of He pressure. It is the result of a soft collisions
regime in the extraction-SPIG region and this could also explain the large amount
of molecular ions, formed in the recoil chamber, which can fly through the RF
sextupole without dissociation. Adding a DC voltage (up to 225 V) to the RF one
(URF = 600 V, VRF = 2.2 MHz) did not help to dissociate them. For Ce and La
elements the amount of CeO+ and LaO+ ions compared to Ce+ and La+ ones was
strongly dependent on the grade of the gas used. The ratio M+ /(M+ + MO+) could
vary in a very broad range (95-5%).
insulator
5 electrodes
Figure 2. Drawing of the electrodes system and mapping of equipotential curves and ion
trajectories.
488 R. BERAUD ET AL.
10 , - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ,
o>-
c: 0.1
JI!
~
g 0.01
::J
"0
'"
.c
(1) 0.001
- - V electrodes = 0 vo~s
_ Velectrodes =-100 volts
0.0001 1....-_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _- - - - - - - - - - - - - '
o
Distance (cm)
Figure 3. Absolute efficiency for 219Rn recoils as a function of exit hole-source distance.
the electric field to extract, with constant efficiency, mass-separated ions 19 Rn) e
which are thermalized far away from the exit hole (up to 4 cm). However, the drop
of efficiency at short distance (compared to the "no field" case) should be avoided,
providing a more focusing field very close to the nozzle.
3. Conclusion
In this work we have shown the advantages of the RF sextupole extraction system
which provides much better MRP and that the stopper enables the use of very
STATUS OF HIGISOL 489
'.... f
Yield vs electrode voltage
A; 219, PHe=500mbars, V DC SPIG ; 225 v
di lance Rn source· exit hole =l5mm
Uloo ~
I
.. I
o
51
J!J '''' .
c
::I
'
a
u
'0
intense H.I. beams. The electric field allows to collect ions in a much larger volume,
a revised version of our electrodes system (under construction) should avoid losses
at short distance and therefore improve efficiency in the near future. In presence of
beam, the situation is much more complex due to field perturbations induced very
likely by the secondary particles, non-suppressed beam and subsequent ionisation
of the buffer gas.
Acknowledgement
This work has been supported by the Human Potential Programme and Access to
Research Infrastructures of the European Union.
References
1. Dendooven, P., Beraud, R., Chabanat, E., Emsallem, A., Honkanen, A., Huhta, M., Jokinen, A.,
Lhersonneau, G., Oinonen, M., Penttilii, H., Periijiirvi, K., Wang, J. C. and Aystti, J., Nucl.
Instrum. Methods A 408 (1998), 530.
2. Backe, H., Eberhardt, K., Feldmann, R, Hies, M., Kunz, H., Lauth, w., Martin, R, SchOpe, H.,
Schwamb, P., Sewtz, M., ThOrle, P., Trautmann, N. and Zauner, S., NucZ. Instrum. Methods B
126 (1997), 406.
3. Perajiirvi, K., Dendooven, P., Huikari, J., Jokinen, A., Kolhinen, V. S., Nieminen, A. and
Aystti, J., Nucl. Instrum. Methods A 449 (2000), 427.
4. Beraud, R, Bouvier, R, Canchel, G., Chabanat, E. and Emsallem, A., Report LYCEN 2000-49,
IPN Lyon 98-99, p. 52.
5. Xu, H. J., Wada, M., Tanaka, J., Kawakami, H., Katayama, 1. and Ohtani, S., Nucl. Instrum.
Methods A 333 (1993), 274.
6. Lindell, A. and Morita, K., Report 2.16, Annual report, JYFL, 1993, p. 52.
490 R. BERAUD ET AL.
7. Van den Bergh, P., Franchoo, S., Gentens, J., Huyse, M., Kudryavtsev, Y. A., Piechaczek, A.,
Raabe, R., Reusen, 1., Van Duppen, P., Vermeeren, L. and Wohr, A., Nucl. Instrum Methods B
126 (1997), 194.
8. Beraud, R., Emsallem, A., Astier, A., Bouvier, R., Duffait, R., Le Coz, Y., Morier, S., Woj-
tasiewicz, A., Lazarev, Y. A. Shirokovsky, 1. v., Izosimov, I. N., Barneoud, D., Genevey, J.,
Gizon, A., Guglielmini, R., Margotton, G. and Vieux-Rochaz, J. L., Nucl. Instrum. Methods A
346 (1994), 196.
Hyperfine Interactions 132: 491-494,200l. 491
© 2001 Kluwer Academic Publishers.
Abstract. Crystal spectrometry and absorption edge technique have the capability to overcome the
gap in accuracy between experiment and theory in the strong field domain of QED. New results are
presented which indicate the capacity of these methods to measure the energies of X-rays emitted
by highly charged heavy ions at modern storage rings with a precision sensitive to second order
corrections to the Lambshift in H-like very heavy ions.
Key words: absorption edge technique, crystal spectrometry, highly charged heavy ions, Lambshift,
Ly-a lines, micro-strip Ge-detector, QED, storage ring, time-projection chamber.
1. Introduction
QED represents one of the most successful theories in physics and is very well
tested in the case oflight atoms [1, 2]. However, a test in the strong field domain of
highly charged heavy atoms with strongly pronounced higher order effects in ex . Z,
where ex is the fine structure constant and Z the nuclear charge, is still outstanding.
Taking advantage of cooled brilliant ion beams of the heaviest, fully stripped
atoms with corresponding small momentum spreads and low emittance, in the
storage ring ESR of the GSI Darmstadt it became possible to increase the accuracy
of QED tests in the strong field domain of the heaviest few electron systems by one
order of magnitude [3, 4]. However, these experiments with an accuracy of 10-4 do
* Corresponding author.
492 M. CZANTA ET AL.
not yet attain the accuracy of theory which is on the level of 10-5 for the ground-
state energy of H -like U, for example [5]. Therefore, new experimental approaches
are indispensable. The current program at the ESR storage ring at GSI for the
spectroscopy of Ly-a lines emitted by very heavy H-like systems includes different
strategies. Aside from the usage of calorimeters [6], they are crystal spectrometry
and the absorption edge technique, which are the subject of this article.
2. Crystal spectrometry
A new kind of X-ray spectrometer in the focusing compensated asymmetric Laue
geometry focuses the radiation of a certain wavelength from an extended source to
a single line on the Rowland-circle [7]. It serves for a measurement of small wave-
length differences between a fast moving X-ray source and a stationary calibration
source. For a certain wavelength, two Bragg reflections arising from a cylindrically
curved Si-crystal are measured by appropriate position-sensitive detectors. The
design is optimized for photon energies from 50 up to 100 ke V with an efficiency
of ~ 10- 7 .
For a first test of the crystal optics no position-sensitive detector was used.
Instead an attached slit of 100 !-lm width was used to scan the spectrum. An X-ray
spectrum from a stationary 169Yb source is depicted in Figure 1. The observed
line separation is in accordance with the calculated dispersion of the spectrometer.
An experiment at the ESR with Pb 82 + has shown, that it is feasible to measure
the lines of H-like heavy ions at theoretically expected rates. In this experiment, a
position-sensitive gas counter filled with Xe at high pressure of ~20 bar based on
/..{pm)
24.2 24.4 24.6 24.8 25.0 25.2
500rT~~~~~~~~~~~~~~
169Ybsource
400
Qj
c
~ 300
.s=.
_ > ~_ 60eV
~ 156 J.l.m
C 200
::J
o
o 100
o~~--I~~~~~~~j~'~-=~~
127 128 129 130 131 132 133
Position{mm)
Figure 1. X-ray spectrum from a stationary 169Yb source. The Tm-Ka lines near 50 keY are
measured by scanning a slit in front of a Ge(i) detector across the lines. The measured line
width of 60 eV nicely agrees with the theoretical one of 63 eV resulting from X-ray line width
(30 eV), crystal (41 eV) and slit width of 100 flm (38 eV).
SPECTROSCOPY OF LY-Q' LINES 493
the principle of the time-projection chamber (TPC) has been used [8, 9, 12]. An
unexpected large background from the counter itself reduced the signal-to-noise
ratio drastically. In order to overcome this problem, a new micro-strip germanium
detector for position sensitive hard X-ray spectroscopy with an intrinsic higher en-
ergy resolution is presently developed at FZ Jtilich in collaboration with GSI [10].
In combination with such a detector, the crystal spectrometer promises to overcome
the steady conflict between accuracy and sensitivity at low background which can
even be reduced by the measurement of coincidences between X-rays and H-like
ions.
700,------,(f---c-------------~----------_,
600
I•
-
experimentaldata
expecledbehavior
~
.. . .
Figure 2. Comparison of the measured and expected intensity variation of the Ly-a2 line
along the Er absorption edge at 145 0 emitted by 42.3 MeV lu Pb 82 + ions. Shape and height
of the measured and the calculated intensity step nicely agree with each other.
494 M. CZANTA ET AL.
intensity step nicely agree with each other. The jump ratio is, besides the thickness
of the absorber, mainly determined by the intensity of closely spaced X-ray lines.
An analysis of the data [12] leads to the conclusion that the finally achievable
accuracy will mainly depend on the uncertainties of the observation angle and the
beam velocity, being t:,.e "'-' 0.3 0 and t:,.f3 / f3 "'-' 10-4 in this experiment.
As a consequence, an experimental setup was designed in which an rigid geo-
metrical determination allows the control of the geometrical conditions. Making
use of six absorbers, three drift-chambers measured the Ly-a2 line of H-like Au
at three angles simultaneously for two different ion beam velocities. The data of a
recent beam time are presently analysed.
4. Conclusion
Two methods for measuring the energy of Ly-a lines at high precision are presented
which profit from the advantages in storage ring technology. Both of them are char-
acterized by a substantially higher spectral resolution than former concepts. The
challenge to be met here is the inherent low efficiency of the crystal spectrometer
of 10- 7 and the control of the kinematic parameters in the case of the absorption
edge technique. Nevertheless, first results have shown that these methods have the
capability to achieve an experimental accuracy of the order of the theoretical one
and that tests of the QED predictions in the strong field domain become possible.
References
I. Weitz. M. et ai., Phys. Rev. Lett. 72 (1994), 328.
2. Bourzeix, S. et al., Phys. Rev. Lett. 76 (1996), 384.
3. Beyer, H. F. et ai., Z. Phys. D 3S (1995),169.
4. St6hlker, Th. et ai., GSI scient. rep., 1997, p. 99.
5. Beier, Th. et ai., Phys. Lett. A 236 (1997), 329.
6. Beile, A. et ai., GSI scient. rep., 1998, p. 180.
7. Beyer, H. E, Nucl. Instrum. Methods A 400 (1997), 137.
8. Badura, E. et ai., GSI scient. rep., 1997, p. 181.
9. Czanta, et ai., GSI scient. rep., 1999, p. 207.
10. St6hlker, Th. et ai., GSI scient. rep., 1999, p. 206.
11. Schmieder, R. W. et ai., Nucl. Instrum. Methods 110 (1973), 459.
12. Strietzel, c., Thesis, Universitat Siegen, 2000.
Hyperjine Interactions 132: 495-499,200!. 495
© 2001 Kluwer Academic Publishers.
Abstract. SHIPTRAP will allow direct measurement of masses of transuranium nuclides. The
method of choice is a Penning trap spectrometer coupled to the SHIP (Separator for Heavy Ion
Products) facility at GSI, Darmstadt. In this paper the impact of the SHIPTRAP facility, with its
capability of systematic mass measurements with high precision, is explored. Rather few masses
of nuclides above uranium are presently known experimentally. In the region of nuclides above
Z = 100 no ground state masses were measured directly. SHIPTRAP will play an important role
in systematically mapping out this area. Possible candidates for direct mass measurements, even
with small or very small production cross sections, are presented.
Key words: nuclear binding energy, Penning trap mass spectrometer, superheavy elements.
1. Introduction
SHIPTRAP is a facility that is currently under construction at GSI, Darmstadt. As
one major feature, it will allow one to measure the atomic masses of exotic nuclides
by use of the Penning trap technique as applied at ISOLTRAP [1] at the on-line
isotope separator ISOLDE, Geneva. The novelty of SHIPTRAP is the coupling of
an ion trap to an in-flight facility. The exotic nuclides are produced by an primary
ion beam impinging on a thin target. The recoil particles are separated in-flight at
typical energies of 20-500 keY lu using the velocity filter SHIP [2]. These ions
are stopped in a gas cell and transferred to the ion trap system. At ISOL-facilities
typical secondary beam energies are between 30-60 ke V, and electrostatic retar-
dation techniques can still be employed for transferring the ions into an ion trap.
At SHIPTRAP transuranium nuclides can be investigated which are inaccessible
at ISOL-facilities. In this contribution a feasibility study for mass measurements at
SHIPTRAP is carried out in order to determine the number of particles which are
available for measurements at SHIPTRAP.
* Corresponding author.
496 J. DILLING ET AL.
Table I. Budget for the time required for the manipulation of the SHIP beam and budget
for the transfer efficiencies. The SHIP transfer efficiency is reaction depending and here a
minimum value is taken. The values marked (*) are approximations based on simulations.
All other values are taken from other experimental set-ups. For the gas cell two types are
listed, where the explanation is given in Section 3
2. Setup
The set-up of SHIPTRAP consists of three basic elements. These are the stop-
ping gas cell with extraction radiofrequency quadrupole (RFQ), the beam cooler
and buncher RFQ trap and the Penning trap unit. The exotic beam from SHIP
is stopped in the gas cell and extracted as a quasi-continuous low energy beam
(Ekin ~ 100 eV) using the gas flow through a nozzle and electric fields. The
extraction RFQ transfers the ions to the buncher trap. Here the ions are cooled by
collisions with buffer gas. From the buncher trap the ions are extracted in bunched
mode. The ion stack is transported to the first Penning trap (purification trap) where
the ensemble is purified via mass selective buffer gas cooling. In the next step the
selected ions are injected into the Measurement Trap where the mass is determined.
The experiment is described in more detail in [3,4]. Table I lists the time required
for manipulating the secondary ion beam as well as the expected efficiencies for
transmission in the different components.
3. Feasibility study
For mass measurements in a Penning trap the accuracy obtained is determined by
the measurement time, which is limited by the nuclear half-life, and the statistics,
which is limited by the available intensity. Therefore, specific limitations arise for
exotic nuclides. Here an investigation of all known transuranium nuclides is done
A PHYSICS CASE FOR SHIPTRAP 497
Region 1. The nuclides in this region are produced in collisions of relatively light
projectiles on heavy targets. A typical reaction would be 12C (7.3 MeV lu) + 234U
(at rest) :::::} 246Cf (17 keV lu). Due to the low kinetic energy of the recoil ions the
entrance window of the stopping gas cell should be very thin. Possibly a carbon-
foil with 20 !-lgl cm2 can be employed. The required gas pressure would be of the
order of 10-4_10- 5 mbar.
Region 2. In this case heavy ions impinge on heavy targets like 40Ca (6.4 MeV lu)
+ 205Ti (at rest) :::::} 245Md (170 ke V lu). To stop the products a thicker window
on the gas-cell is used, for example, a 100 !-lg/cm2 Ni-foil. The gas pressure is
~100 mbar.
Region 3. The stopping scenario would be the same as for region 2 for a typical
reaction like 54Cr (5.6 MeV lu) + 208Pb (at rest) :::::} 262Sg (240 keY lu). The differ-
ence is that for these nuclides the production rate is so low that an extremely sen-
sitive and non-destructive detection method, e.g., Fourier Transform Mass Spec-
trometry (FTMS) [7] has to be employed, allow for a reduction of the required
number of ions by at least a factor of 10.
4. Outlook
The SHIPTRAP facility will open up the field for direct mass measurements of
transuranium nuclides. Quite a number of very interesting cases were found, where
some nuclides also have in addition Qa-links. For example 254No can be linked
down to 234U. Accurately measured masses of, e.g., 257Rf, 260Db and 265Sg com-
bined with Qa values determined from a-decay chains at nuclides 269 110, 272111
and 277112 [8, 9] allow for an independent evaluation of the masses of the heaviest
elements synthesized at SHIP. Other nuclides which are not listed, because their
498 J. DILLING ET AL.
Table II. Candidates for mass measurements by SIDPTRAP where the conditions of
Tl/2 > 500 ms and 8m > 50 keY are fulfilled and which are produced with sufficient rates.
Shown is the isotope, its half life and mass uncertainty estimated from systematic trends [5],
the cross section at SHIP and the corresponding number of ions at SHIPTRAP. For the different
regions, the primary beam intensities, the transfer efficiency, etc.
Region 1
239Cm 3h 100 12 ~b 3.5
239Bk ?s 290 18 ~b 6.5
240Bk 5m 159 16 ~b 6
239Cf 39 s 150 8 ~b
240Cf 1m 200 12 ~b 3.5
241Cf 3.8m 250 10 ~b 3
245Cf 43.6m 100 370 nb 0.05
Region 2
245Fm 4.2 s 280 10 nb 0.001
246Fm Us 200 20 nb 0.002
249Fm 2.6m 140 200nb 0.Q2
250Md 52 s 300 10 nb 0.001
251Md 4m 200 80 nb 0.01
252Md 2.3m 200 800nb 0.08
253Md 6m 210 800nb 0.08
254m No 0.78 S 100 500 nb 0.05
259No 58 m 100 600nb 0.06
257Rf ?s 270 14 nb 0.001
259Rf 3s 70 10 nb 0.001
Region 3
260Db 1.5 s 230 ""'10 nb 0.001
261Db 1.8 s 230 ""'10 nb 0.001
262Db 34 s 180 ""'20 nb 0.002
263Sg 0.9 s 70 ""'10 nb 0.001
263mSg ?s 120 ""'10 nb 0.001
265Sg 20 s 100 ""'10 nb 0.001
264Bh 740 ms 280 100pb 0.00001
265Bh 1s 350 200pb 0.00002
mass uncertainty is lower than 50 keY, are interesting to be measured directly, since
"ir present value was derived indirectly by alpha-spectroscopy. Shell correction
energy calculations [10] predict a minimum value (shell closure) at neutron number
N = 162 and proton number Z = 108 for deformed nuclear shapes. A possibility
A PHYSICS CASE FOR SHIPTRAP 499
References
1. Bollen .• G. et al., NucZ. Instrum. Methods A 368 (1996), 675.
2. Miinzenberg, G. et ai., Nucl. Instrum. Methods 161 (1979), 65.
3. Dilling, J. et al., Hyp. Interact. 127 (1999), 491.
4. Marx, G. et ai., this issue, p. 463.
5. Audi, G. and Wapstra, A. H., Nuclear Phys. A 595 (1995), 409.
6. Dilling, J., PhD Thesis, University of Heidelberg, in preparation.
7. Bradley, M. P. et al., Phys. Rev. Lett. 83 (1999), 4510.
8. Hofmann, S. et ai., Z. Phys. A 350 (1995), 281.
9. Hofmann, S. et ai., Z. Phys. A 354 (1996), 229.
10. Patyk, Z. and Sobiczewski, A., Nuclear Phys. A 533 (1991),132.
11. Geisel, H. et at., Nuclear Phys. A, submitted.
Hyperjine Interactions 132: 501-504,2001. 501
© 2001 Kluwer Academic Publishers.
Abstract. Heavy element chemistry is related to the fundamental interest that lies in exploring
the upper limits of the periodic table. Chemical properties of the heaviest elements have already
been studied at single atoms in aqueous solutions and in the gas phase up to an atomic number
Z = 107. These techniques allow to study nuclides with half lives as short as about I s. Next
generation chemistry experiments could be envisaged with an ion trap technique already developed
for stable isotopes. At very low production rates in the order of I per 100 s and/or half lives as short
as about 10 ms, the ion-molecule reactions can be studied in a buffer gas cell, in which the heavy
elements are stopped and thermalize with a high probability as singly charged ions. Ion-molecule
reactions with well defined buffer gas admixtures, as, e.g., 02, H20, CH4, C02, are identified by
mass selective detection.
1. Introduction
2. Experimental setup
Ion chemical reactions of 252Md (T1/2 = 2, 3 m), 255No (T1/2 = 3, 1 m), and 256Lr
(T1/2 = 25, 9 s) with various reactant gases as O2, H 2, or others will be investigated
with the apparatus shown in Figure 1.
The heavy elements can be produced with cross-sections in the order of fLb
by the reactions 238Uct9F,5n)252Md, 208Pb( 48 Ca, n?55No, and 209Bi(48 Ca, n) 256 Lr.
They will be separated from the primary beam by the velocity filter SHIP. A large
fraction of the fusion reaction products come to rest as ions in buffer gas (300 mbar
Argon), to which reaction gas is added.
From the gas phase reaction A + +X 2 -+ AX+ +X of heavy element ions A +
with reaction gas X 2 of known concentration [X 2] the reaction constant k, defined
from -Cl[A +]/Clt = k[A +][X 2] , can be determined from a measurement of [A +] or
[AX+] as a function ofthe reaction time. To achieve this, the ions will be extracted
after different variable reaction times. The concentration [X2] must be chosen in
such a manner that the expectable reaction time is long in comparison with the
transport time, which may be as short as about 1.5 ms with a width of about
0.5 ms [5]. The ions will be separated from the buffer gas in a differential pumping
section. Their masses will be detected employing the Ion-Guide Quadrupole Mass
Separation (lGQMS) technique [5, 6].
The investigation of various reaction channels and reaction constants provides
detailed information on electron configurations [7].
OPtical cell
'\
T Ext r anC T i o Bunching,
system cooling system
\ Collection and---
purification
system
Channe ~roo
aJID
c>-~
c:rr:o
r p -lD-'mba.
p . 1mba. P - T1D-' mba.
p -1"'mbar Preliminary Purified
ion bunches ion bunch
-1.5 m
3. Reference measurements
In a first step, the method must be established on chemical homologue elements.
To achieve this, ion-molecule reactions will be studied with a Fourier Transform
Mass Spectrometer (FTIMS) under well defined experimental conditions [8].
~ryostat I
LHe
Quench-
Gas --+ ------~
' 00 ,00
lomsl Isomsl
eo eo
£., 60 60
.."
c
~ '0 '0
il
«
20 20
' 00
'so ,. 'oo '65 170 ", '81)
,., ,,., '95 200
'00
'so ,., '80 '65 170 '75 '00
"
"
' &5
,,., 195 200
l:
., eo
""
"c:
"'c:" '0
I ~ 00
~
20
I III 20
.' ~ ~
Figure 3. Er+ and ErO+ mass spectra at reaction time as indicated. The groups around mass number
167 and 183 belong to nat. Er+ and nat. ErO+ , respectively.
504 A. DRETZKE ET AL.
1.0 •
0.8
S
....
0." ~
W
-::: 0,'&
0,2
Figure 4. Exponential decay of the Er+ to (Er+ + ErO+) ratio as function of the reaction time.
In first experiments, erbium reactions with O2 will be investigated. The ions are
created by laser ablation from a solid target using a pulsed CO2-laser (Figure 2).
The ions are quenched to the electronic grounds tate by an argon push and stored
in a FT-ICR cell ("Source Cell") at a constant O2 gas pressure. With an ejection
sweep, the cell is cleared from all other ions, with the exception of erbium ions,
which react with oxygen. For high resolution mass detection, the ions can also be
transferred after a given reaction time in a second FT-ICR cell ("Analyzer Cell").
For an example see Figure 3, which shows Er+ and ErO+ mass spectra obtained
after 0, 50, 500, and 1800 ms reaction time with oxygen. From the exponential
decay of the Er signal (Figure 4), a reaction constant k = (1.4 ± 0.6) . 10- 10 cm3
molecule- l S-l can be deduced.
The results from FT-ICR experiments must be compared with results obtained
in the inert buffer gas cell to which the reaction gas is admixed. To achieve this, the
chemical homologues will directly be implanted as energetic ions from a Tandem
Van de Graaff accelerator into the buffer gas cell.
References
1. Hofmann, S. and Mlinzenberg, G., Rev. Modern Phys. 72 (2000), 733.
2. Schadel, M. et at., Nature 388 (1997), 55.
3. Fricke, B. et at., Radiochim. Acta 62 (1993), 17.
4. Eliav, E., Kaldor, U. and Ishikawa, Y, Phys. Rev. Lett. 74 (1995), 1079.
5. Backe, H. et at., Nucl. Instrum. Methods B 126 (1997), 406.
6. Iivonen, A., Saintola, R. and Valli, K., Phys. Scripta 42 (1990), 133.
7. Armentrout, P. B., Science 251 (1991),175.
8. Marshall, A. G. and Schweikhard, L., Internat. 1. Mass Spectrom. Ion Proc. 118/119 (1992), 37.
.... Hyperfine Interactions 132: 505-509,200l. 505
© 2001 Kluwer Academic Publishers.
"
Abstract. SHIPTRAP is an electromagnetic transport and trapping system to provide very clean and
cold beams of singly-charged recoil ions from the SHIP facility at GSI. The different components
of the system are currently under development in Munich (gas cell and extraction RFQ) and GSI
(Buncher RFQ and Penning traps) [I]. Design and manufacturing of the prototype buffer gas cell and
the extraction RFQ based on a wide range of simulations have been completed. The results of these
simulations together with the first measurements will be reported.
1. Introduction
The present state of ion trap technology allows the coupling of a trap system to ra-
dioactive ion beams of higher energies. Such a project is SHIPTRAP at GSI which
is presently under construction to stop, cool and deliver heavy radionuclides for
downstream experiments. The energetic radionuclides are delivered by the SHIP
velocity filter [3]. The scientific case as well as a detailed description of the whole
SHIPTRAP facility are given elsewhere [4, 5].
Stopping ions in a gas cell filled with buffer gas has been investigated for several
years. The aim of most of these projects in the field of nuclear physics is the
production of ions directly in the gas cell, by nuclear reactions of an energetic
primary beam with appropriate targets placed inside the cell. This technique has
been very successfully used in JyvaskyUi. [6], Leuven [7], and Mainz [8]. A new
generation of projects has started recently in which the reaction takes place outside
the gas cell. The reaction products are injected and stopped in a gas cell. Groups at
ANL [9], RIKEN [10], NSCL-MSU [11], and Munich [12] are currently working
in that field.
The method of ion extraction from the gas cell differs for the cells in use or
under development, depending on the boundary conditions. The concepts reach
506 O. ENGELS ET AL.
from an extraction via gas flow and via gas flow combined with electrostatic fields
to the extraction via gas flow plus a combination of electrostatic and rf-fields. After
their extraction, the ions have to be separated from the gas and guided into a high
vacuum region. This is done by an extraction RFQ, a longitudinally segmented rf
quadrupole structure. This structure is similar to ion guides, emittance improvers
and ion bunchers, described elsewhere [13-15].
2. Setup
In the SHIPTRAP project the gas cell design together with the extraction RFQ
is expected to be the most critical part. Therefore a gas cell study program was
initiated in Munich and a prototype gas cell has been designed for tests at the
Tandem Accelerator in Garching and for first tests at SHIP.
2.1. CONCEPT
The prototype of the gas cell for SHIPTRAP follows the concept of extracting the
ions via the gas flow and guiding electrostatic fields. These two mechanisms can be
simulated separately in first order, due to the special feature of a high transmission
grid in front of the nozzle. This grid prevents a loss of ions in the defocusing field
inside the nozzle and improves the focusing inside the cell due to a spherical shape.
The extracted ions out of the gas cell are then guided via a linear RFQ into a good
vacuum region. This extraction RFQ separates the ions from the neutral gas.
_ _ _ _ _ _-----, EI
E2
Extraction-RFQ
Cl=:l(:=lC-,,..,....---,~~-Jt:::J....--, ,.--------,
Gas cell Gri ozzlc QMS and Dc.e<:tion
r---"=:J--'r,r-'CJCJDCJ~ L -_ _ _ _......
J 120mm
100mm
Figure 1. The SHIPTRAP test set-up consisting of the gas cell with the three electrodes (El,
E2, E3) to create the accelerating and guiding field and a nozzle covered by a high transmission
grid, the extraction RFQ as a differential pumping section to guide the ions to a region with
high vacuum and the QMS and detection section for a mass selective ion detection. The ions
are injected radially into the gas cell.
a pressure of 10- 2 mbar at the position of the extraction RFQ, the 400 lis pump
is situated at the QMS and Detection section. In addition to the clean environment
in the cell, the buffer gas is purified by a getter based system, suppressing the
impurities in the gas to the level of < 1 ppb thus preventing molecule formation.
The whole setup can be heated up to 250 0
•
3. Simulations
After a wide range of simulations the design and manufacturing of the prototype of
the gas cell is completed. The simulations covered the following topics: stopping
of the ions in the gas (SRIM); drag of the ions via the electrical field in the cell
towards a supersonic nozzle (SIMION); drag of the ions through the nozzle via
the gas flow (VARJET, solving the full system of time-dependent Navier-Stokes
equations [17]).
First stopping simulations indicate that, whatever gas is used, a majority
(50-90%) of most radionuclides of interest could be stopped within a spheroid
of 40 diameter and 120 mm length (for 232Th at 100 keV /u in He at 100 mbar). To
these values one has to add the horizontal and vertical SHIP-beam dimensions of 50
and 30 mm. These simulations imply the minimum dimensions for the innermost
electrode of the final gas cell for SHIPTRAP.
The ion motion caused by the electrostatic field was simulated extensively with
SIMION. With a maximum voltage of 700 V at electrode 3 an average extraction
time of 30 ms is expected.
VARJET calculations showed that the gas velocities in the cell are negligible
« 1 m/s), with the exception of the area close to the nozzle in the gas cell and
between gas cell and RFQ. In order to determine the exact motion of the ions in
this area one would need a combined code for gas- and electrodynamics. Such a
508 O. ENGELS ET AL.
code is under development using the electrostatic field information of SIMION and
the gasdynamic parameters of VARlET. Introducing the parameter
G _ V·[on;r,Z - vgas;r,z
r,Z - K (1)
lead us to the following equations of motion for ions in a combination of gas and
electric field:
q
a r + -(G r - Er) = 0, (2)
m
where Gr,z are the gasdynamic analogon of the electric field components Er,z.
Vion;r,zand vgas;r,z are the radial and longitudinal components of the gas and ion ve-
locity, respectively, and K is the ion mobility; ar,z are the acceleration components
and q and m are the charge and mass of the ion.
Beside the gas cell, the jet of the ion-gas mixture expanding into the extraction
RFQ is a critical point for the design of the cell and the RFQ. Therefore the gas
flow and the cooling of the ions in the extraction RFQ were measured.
4. Measurements
First measurements with the gas cell showed a clear dependence of the ion focusing
towards the nozzle and the grid in front of the nozzle. The grid allows to achieve
faster extraction times due to higher voltages applicable in the cell. These higher
voltages lead to a stronger defocusing of the ions. However, this drawback can be
compensated by the better focusing properties of the grid.
Figure 2 shows as an example a comparison of a VARJET simulation and the
measurement of the impact pressure in the RFQ. The pressure in the gas cell was
100 mbar of argon, with a pumping speed of 100 1/s at the position of the extraction
RFQ. The pressure in the RFQ was measured with a pitot probe. The measured
pressure is an impact pressure, caused by the gas jet impinging the probe. The
10
9
118
.D
3 7
~ 6
~ 5
ti.4
~ 3
.§2
30 40 50 60 70
distance from the nozzle [mm]
Figure 2. Simulation and measurement of the impact pressure inside the RFQ.
FIRST MEASUREMENTS WITH THE GAS CELL FOR SHIPTRAP 509
5. Outlook
After completion of the tests at the Tandem Laboratory in Garching/Munich the
prototype of the gas cell will be tested at GSI with a SHIP beam in the middle
of 2001. In parallel, the final gas cell is under development and will be ready for
the first tests end of 2001. The goal for the final version is to be able to stop all
incoming ions and extract the ions in an average time of 30 ms.
Acknowledgements
Work supported by the BMBF with grant number 06LM973 and by the European
Union (RTD project EXOTRAPS).
References
1. Marx, G. et ai., this issue, 463.
2. Bollen, G., Nuci. Instrum. Methods A 368 (1996), 675.
3. Miinzenberg, G. et ai., Nucl. Instrum. Methods 161 (1979),65.
4. Dilling, J. et ai., Hyp. Interact. 127 (2000), 491-496.
5. Proposal for SHIPTRAP: A capture and storage facility at GSI for heavy radionuc1ides from
SHIP, GSI, 1998.
6. Dendooven, P., Nucl. Instrum. Methods B 126 (1997), 182.
7. Van den Berg, P., Nucl. Instrum. Methods B 126 (1997), 194.
8. Backe, H. et ai., Phys. Rev. Lett. 80 (1998), 920.
9. Savard, G., In: Proc. IGISOL-7, Mainz, 1999.
10. Wada, M., In: Proc. IGISOL-7, Mainz, 1999.
II. Bollen, G., private communication, 2000.
12. Engels, O. et ai., Annuai Report, GSI, 1999.
13. Arje, J. et ai., Phys. Rev. Lett. 54 (1985), 99.
14. Lunney, M. D. etai., Internat. 1. Mass Spectrometry 190 (1999),153.
15. Herfurth, F. et ai., submitted to Nucl. Instrum. Methods (April 2000).
16. Naaman, R. et ai., Rev. Sci. Instrum. 67 (1996),3332.
17. Varentsov, V. L. et ai., Nucl. lnstrum. Methods A 413 (1998), 447.
Hyperfine Interactions 132: 511-515,2001. 511
© 2001 Kluwer Academic Publishers.
Abstract. With the Penning trap mass spectrometer ISOLTRAP, close to 200 nuclides have already
been investigated and their masses determined with a typical relative precision of 8mjm = 10- 7 .
Recently, ISOLTRAP's beam preparation system was replaced by an RFQ ion beam cooler and
buncher. The principle and the characteristics of this new beam preparation system will be presented.
It is planned to use ions of various carbon clusters C;; (n > 1) as reference ions for mass measure-
ments. Apart from negligible molecular binding energies, these clusters have masses that are exact
mUltiples of the unified atomic mass unit. This will allow ISOLTRAP to carry out absolute mass
measurements as well as to investigate possible mass-dependent systematic errors. The results of
tests of the production, transport, and trapping of such carbon clusters will be presented.
Key words: atomic masses, ion guide, ion trap, ion cooling, on-line mass spectrometry, radioactive
ion beams.
1. Introduction
The ISOLTRAP experiment is a Penning trap mass spectrometer that is installed at
the ISOLDE facility at CERN (Geneva). The mass measurement is carried out by a
determination of the cyclotron frequency We = (q / m) B of an ion with mass m and
charge q in a magnetic field of known strength B. Using this technique, more than
200 nuclides have already been investigated. A relative precision of 8m / m ~ 10- 7
and a resolving power of R ~ 107 are routinely reached.
ISOLTRAP consists of a beam preparation trap and a tandem Penning trap
setup. The tandem Penning trap configuration is made up of a cylindrical Penning
trap for the cooling and isobaric cleaning of the ion bunch and a hyperbolical
* Corresponding author.
512 A. KELLERBAUER ET AL.
Penning trap for the precision mass mesurement [1]. In ISOLTRAP's original
configuration, the ion beam was deposited on a rhenium foil at the entrance side
of the cooling Penning trap and later released towards the trap by evaporation
and surface ionization. The surface ionization technique restricted the applicability
of ISOLTRAP to the nuclides of surface-ionizable elements, namely, the alkali
metals, the alkali earths, and the rare earths. This restriction was overcome by the
implementation of a very large buffer-gas-filled Paul trap as the beam collection
device [2]. Recently, this Paul trap was replaced by an RFQ ion beam cooler and
buncher (RICB) whose acceptance was more carefully matched to the phase space
of the ISOLDE beam, thereby increasing the efficiency of the beam preparation
stage by about three orders of magnitude. It has been used very successfully in
several on-line experiments in 1999 [3-5] and 2000.
Since in practice the magnitude of the magnetic field inside the precision Pen-
ning trap is not known to high precision, an actual measurement consists of com-
paring the cyclotron frequency of the ion of interest to that of a nuclide with a
well-known mass. In the case of ISOLTRAP, such reference masses are, for in-
stance, 208Pb, 133Cs, 85Rb, and 36 Ar. In the future, it is planned to replace these
reference nuclides with carbon cluster ions which have practically no mass uncer-
tainty and are available over a very large mass range. For this purpose, a carbon
cluster ion source is presently under development.
~
(first part)
UL-A--L-.A..-.-I . . • D H
I,~~~~~I ••• D~~~~~ H~______
lium gas pressure of about one pascal. Adequately low pressures in the neighboring
regions are obtained by differential pumping through small orifices. The ISOLDE
beam enters the structure through the deceleration and the focusing electrodes. In
the ion guide, the decelerated beam is radially confined by the pseudo-potential of
the RFQ field while it is dragged along the axis of the structure by a small axial
DC field. The cooled ions are finally trapped in an axial potential well near the
end of the rod structure. When the ions are ejected, the ion bunch is accelerated to
a few keV as it enters the pulsed cavity and then leaves the cavity without being
accelerated further. The ions are then transferred to the tandem Penning trap setup
for measurement.
An analysis of the shape of the ion pulse ejected from the beam buncher shows
that the ions are cooled to room temperature in the ion trap within a cooling time of
only a few milliseconds [7]. From the time and energy spreads of the pulse, a lon-
gitudinal emittance of about 10 eV ~s can be extracted. The transverse emittance
of the ejected ion pulse at an energy of 2.5 ke V was also measured using a beam
observation system. It was found to be less than IOn mm mrad. This corresponds to
a transverse emittance of less than 2n mm mrad for a 60-keV beam and represents
a substantial improvement over the ISOLDE beam with an emittance of typically
35n mm mrad.
The total efficiency of the beam preparation system in bunching mode is a
quantity that is crucial for the performance of the whole ISOLTRAP apparatus,
especially when investigating very rare or short-lived nuclides. It was measured
using an ISOLDE beam of stable ions whose current was carefully chosen to be
large enough for a current measurement at a Faraday cup just before the system,
but small enough to still allow the counting of single ions on a micro-channel-
plate (MCP) detector on the ejection side just beyond the pulsed cavity (typically
a few pA) [7]. The detection efficiency of the MCP, which is a function of the ion
energy and mass [8], was assumed to be about 30% for 2.5-keV ions of moderately
high mass (about 100 u). It was found that the total efficiency of the RICB is
between 12 and 15% for the various xenon isotopes which were used in this study.
The microscopic mass standard is the "unified atomic mass unit" u, which is de-
fined as one twelfth of the mass of the 12C atom. Therefore, the masses of clusters
of 12C atoms are just multiples of u. (The molecular binding energies, which are
at the level of eV and thus by more than two orders of magnitude smaller than the
precision of ISOLTRAP, can be safely neglected.)
The mass of an ion is obtained from the comparison of its cyclotron frequency
with the cyclotron frequency of a "reference ion". The uncertainty in the mass of
the reference ion therefore contributes to the relative uncertainty of the mass of the
ion of interest. If, however, the reference particle is a cluster of 12C atoms, the mass
514 A. KELLERBAUER ET AL.
600
.....,
(/)
§ 400
o
U
200
OLL~~~~~~~~~~~
10 20 30 40 50 60 70 80 90 100
TOF [115] TOF [115]
Figure 2. Time-of-f1ight mass spectra of carbon cluster ions and some charged contaminants
produced by laser-induced desorption and fragmentation of fullerenes at 532 nm with a laser
light f1uence of about 10 mJ/cm 2 .
of the ion of interest can immediately be expressed in terms of u and the error due
to the uncertainty of the reference mass is eliminated.
Carbon cluster ions will also permit a systematic study of possible mass-depen-
dent errors in the determination of the cyclotron frequency. Such errors become
important when the mass number of the reference atom is far removed from the
mass number of the atom whose mass is to be measured. In the past, the mass-
dependent error has been determined to be well below 2 . 1O- 9 /u [9]. Recent
tests comparing the cyclotron frequencies of 85Rb and 133Cs reference ions suggest
that the error may be as low as IO- Io lu. It is desirable to confirm these results
with measurements over a large region of mass and using ions that have no mass
uncertainty.
As a first step of an ongoing development whose goal it is to make carbon
cluster ions available for reference measurements in the precision Penning trap, a
carbon cluster ion source based on the technique of laser-induced fragmentation,
desorption, and ionization was mounted upstream of the cooling Penning trap of
the ISOLTRAP setup [10]. Charged clusters were detected with an MCP mounted
directly in front of the trap. Figure 2 shows the carbon cluster ions produced at a
laser light ftuence of about 10 mJ/cm 2 . All even-n fragments down to n = 32 as
well as the light fragments ranging from the monomer C up to C 19 are observed.
Coalescence products appear in the mass spectrum around C 120 . Thus the whole
mass spectrum of the chart of the nuclides from mass number A = 12 up to
A = 276 can be covered.
The carbon cluster ions thus produced were then injected into the cooling Pen-
ning trap and stored [10]. By a choice of the timing of a pulsed cavity between
the ion source and the Penning trap that decelerates the ions from the transport
energy of about 2.7 keY to the trapping energy of a few tens of eV, ions werect
selected for capture into the cooling Penning trap. Helium buffer gas was present
in the trap, but no RF excitation was applied. The storage time in the trap was
IMPROVEMENT OF ISOLTRAP 515
varied between 5 and 390 ms. After ejection out of the cooling Penning trap, the
ions were detected with an MCP detector placed above the cooling trap. A time-
of-flight spectrum obtained of ions ejected after 370 ms storage time showed that
ct ions formed the most prominent peak. Some smaller clusters C 2 ,3,4, produced
either from fragmentation of hot heavier clusters or from dissociation by buffer gas
collisions, as well as some larger clusters C6 ,7 whose origin is yet to be determined,
were also observed. After a storage time of 35 ms, the width of all peaks in the TOF
spectrum reached a constant minimum value, indicating that the cooling process
had reached a state of thermal equilibrium.
References
1. Bollen, G, et aI., NucZ. lnstrum. Methods A 368 (1996), 675,
2. Schwarz, S., PhD Thesis, Johannes-Gutenberg-Univ. Mainz, 1999, ISBN 3-8288-0735-6.
3. Herfurth, F. et aI., submitted to Phys. Rev. Lett.
4. Schwarz, S. et al., submitted to Nuclear Phys. A.
5. Dilling, J. et aI., this issue, p. 331.
6. Kellerbauer, A. et al., accepted for publication in Nucl. lnstrum. Methods A; available as CERN
preprint EP/2000-048.
7. Herfurth, F. et al., accepted for publication in NucZ. lnstrum. Methods A; available as CERN
preprint EP/2000-062.
8. Brehm, B. et aI., Measm. Sci. Techno!. 6 (1995), 953.
9. Beck, D. et al., Nucl. lnstrum. Methods B 126 (1997),374.
10. Scheidenberger, C. et al., In: Proc. of the RNB2000 Conference, Divonne, France, April 2000;
to be published in Nuclear Phys. A.
Hyperfine Interactions 132: 517-520,2001. 517
© 2001 Kluwer Academic Publishers.
Abstract. Absolute cross sections of electron capture and dissociative excitation for the Na+ -H2'
N2 and K+ -H2, N2 pairs are determined. The high intense hydrogen and nitrogen atomic lines
HI (121.6 nm) NI (120.0 nm), have been observed. For the Na+ -H2 and K+ -H2 pairs the qualitative
interpretation of experimental results in the framework of quasidiatomic approximation are carried
out.
Key words: dissociative capture process, electron capture cross-section, potassium, quasimolecule,
sodium.
In the present work the values for absolute cross sections of electron capture and
dissociative excitation processes at collisions of Na+ and K+ ions with H2 and N2
molecules in 0.7-10 keV energy range are given. The measurements were carried
out by optical and collision spectroscopy methods. The experimental set-up and
calibration procedure for the determination of absolute value of cross sections of
electron capture and excitation processes have been described in details [1-3]. An
estimation of uncertainties of the absolute value of all cross sections given here is
about 20--25% and accuracy of relative measurements did not exceed 4-5%.
In Figures 1-3 the dependencies of absolute cross sections from the collision
energy for the excitation and electron capture inelastic processes are presented.
The comparison of results for dissociative excitation processes of hydrogen
molecule by K+ and Na+ -ions (Figures 2 and 3, respectively) shows, that the
excitation of La-line occurs more effectively in the case of K+ -H2 pair, while
in case of K+ -N2 and Na+ -N2 (Figures 1 and 2) cross section of excitation of
nitrogen line (120.0 nm) is of the same order.
The experimental results for K+ -H2 and Na+ -H2 pairs are interpreted on a
basis of the quasimolecular representation in framework of the quasidiatomic [4]
approximation by using the schematic correlation diagrams of molecular orbitals
(MO) of colliding particle system. An important point in this connection is that
it is necessary to take into account the orientation of molecule axis relative to the
direction of the ion beam [5]. This is achieved by using schematic MO correlation
diagrams for two different particular the C 2V (isosceles triangle) and Coov (linear
conformation) symmetries of the triatomic molecule. Analysis of correlation dia-
518 M. GOCHITASHVILI ET AL.
M
S 10 or
oj NI(120.0nm)
!)( t
u
t-- =::::M(124.3nm)
....-t
'0 1 l(116.4-116.8nm)
....-t "r ...............-rU(113.4 nm)
d ~"""-.--II"""
o
:=
u 0,1 ,
•
./' ......., KI(766.5 nm)
Q)
r.Jl
•
r.Jl
/"--+ NII(500.5nm)
~ ---+-+
~ 0,01 .-4 .....
i-<
U
o 2 4 6 8 10 12
Energy,keV
Figure 1. Absolute excitation cross sections for nitrogen atomic (A = 120.0 nrn, transi-
tion 3s4P_2p34 S, A = 124.3 nrn, transition 3s,2D-2p32D, A = 113.4 nrn transition
2p4 4 p _2p3 4S, A = 116.4-116.8 nrn, transition 3d 2D, 2p_2p3 2D); nitrogen ionic (A =
500.5 nrn, transition 3d 3p _3p 3D), potassium (A = 766.5 nrn, transition 4p 2p-4s 2S) lines
and total electron capture cross section (5 c for K+ -N2 collisions.
l/
N ...--HI(121.6nm)
S .----.----- -.-. -.-.
u ...... J<O"c 0 0 0
r:--- 0-0-0"'--
~0 1 /HI(656.3nm)
....-t • ~• .---.___e-.
_________ ---
Jr'w: _____
§~ KI(766.5nm)
:a
~ 0,1 /HI(486.1 nm)
00 . ~~-A-.
~
8
/
. . . .. . .
~----""
U 0,01+-~~~~~~~.-~--~~~~
o 2 4 6 8 10
Energy,keV
Figure 2. Absolute excitation cross sections for hydrogen atomic (A = 121.6 nrn, transition
2p2p-1s2p, A = 656.3 nrn, transition 3d 2D_2p2p, A = 486.1 nrn, transition 4d 2D_2p2p);
potassium (A = 766.5 nrn, transition 4p 2p-4s 2S) lines and total electron capture (5 c cross
section for K+ -H2 collisions.
grams shows that for dissociative excitation processes the perpendicular orientation
of molecule axis plays a dominant role in formation of excited hydrogen atom
H(2p) (Figure 4). In the case of K+ -H2 pair excitation of the hydrogen line La
after decay of excited molecular states of H20s0'2pO') or Hi(2pJT u ) are reached
by interaction (rotational interaction) between initial K+ -H20 SO')2 and final in-
elastic channels of K+ -H2(1SO'2pO'), K(4s) - Hi(2pJT u ) (Figure 4(b». In the case
of Na+ -H2 the initial molecular term is promoted to the continuous spectrum and
hence for the excitation of H 2 0 SO' 2pO') molecular state the process of ionization
ELECTRON CAPTURE AND DISSOCIATIVE EXCITATION 519
NI(120.0nm)
.~-.--.
N
8u 1
t--
..-
'0
...-l
_____
.-------
HI(121.6nm)
~Nal(589.0-689.6nm)
~-... -~-... -..t.
~
/.:.--"
-
0
~
u 0,1
<U
Yl HI(656.3nm)
Yl
Yl .... -?'....
0
;.., .......---......---
u
0,01
0 2 4 6 8 10
Energy,keV
Figure 3. Absolute excitation cross sections for nitrogen (J.. = 120.0 nm, transItion
3s 4p_2p3 4S); hydrogen (J.. = 121.6 nm, transition 2p 2P-ls 2S; J.. = 656.3 nm, transition
3d 2D -2p 2p) and sodium (J.. = 589.0--589.6 nm, transition 3p 2p-3s 2S) atomic lines for
Na+ -H2 and N2 collisions.
+ + 2
AI (3s3d) Na·H (1so) SC+(3d4S)
- 2
Figure 4. Schematic correlation diagrams for (Na H2)+ and (K H2)+ systems.
is competitive (Figure 4(a)). This is the reason for the reduction of the excitation
cross section of line La in the last case.
The qualitative analysis in the case of K+ -N2 pair for intense line NI (A =
120.0 nm) is carried out as a typical example. Emission of this line can be con-
nected with the decay of excited intermediate molecular states of N~ or Ni*. In
particular, it is caused both by excitation of Rydberg states of molecule N2 with
core (C2~: or D2 rIg) and of the hole excited state 2ag-l of Ni in different inelastic
processes: K+ + N2 ~ K+ + N;*~ K+ + N* + N (direct excitation), K+ + N 2---+
K+ + Ni*(2ag- 1 ) + e ~ K+ + N* + N+ + e (ionization), K+ + N2 ~ KO +
Ni*(2ag- 1) ~ KO + N* + N+ (charge exchange). In the case of decay of Ni mole-
520 M. GOCHITASHVILI ET AL.
cular ions the intense ionic line of N+ must be observed too. For example, in the
collision e + N2 (in the ionization process) and He+ -N2 (in the charge exchange
process) [3], we have observed a lot of ionic lines. In particular ionic line 108.4 nm
NIl (2p3 3D -+ 2p2 3p) is the most intensive. We should emphasise that relation
of emission cross section of NI (A. = 120.0 nm) and NIl (A. = 108.4 nm) lines at
collisions of e + N2 and He+ -N2 is the same and does not depend from collision
energy. This means that there is strong correlation between molecular states of
Ni* which after decay gives N* and N+* excited products. Therefore we suppose
that in the case of decay excited state of Nt the intensive line 120.0 nm of N is
accompanied by intensive ionic line 108.4 nm of N+ too. In the K+ + N2 collision
we have not observed any ionic lines. Therefore excited products (atoms) in the
dissociative processes must be formed in direct excitation processes of N2 only:
K+ + N2 -+ K+ + N~* -+ K+ + N* + N.
We suppose that the "core" ion model [6] is adequate to interpret the experimen-
tal results. The molecule in high-excited states, the "core" of which is decaying
spontaneously in Franck-Condon region, gives a major contribution to the for-
mation of excited atomic products. The weakly bonded electron follows to the
atomic ion and is captured in excited states. For the explanation of the dissociation
mechanism we have suggested that the spin conservation rule is valid because the
spin-orbital interaction is ignored in our case. Therefore, in case of the excitation
of more intense nitrogen atomic lines NI (A. = 120.0 nm) both fragments of dis-
sociation must be formed in quartet states in the following way K+(3p6 IS) + N2
(Xl ~:) -+ K+ + N(3s 4p) + N(2p3 4S). In the framework of the core ion model,
formation of these lines is caused by excitation of intermediate molecular singlet
state of N~* in which ion core Ni is in the states C2~: or D2 n g . Dissociation limit
of this state is N(2p3 4S) + N+(2p2 3p). For the C 2 ~: state this dissociation limit
is reached by the predissociation 4nu state. In this case, weakly bonded molecular
electron is captured by the N+(2p2 3p) ion, forming nitrogen atom in the excited
NI(3s 4p) state.
Acknowledgement
This work was supported by INTAS grant N 99-01326.
References
1. Gochitashvili, M. et ai., Zh. Techn. Fiz. 49 (1993), 2338.
2. Kikiani, B., Lomsadze, R. and Mosulishvili, N., Bulletin of Georgian Academy of Sciences 3
(1995), 394.
3. Gochitashvili, M., Jaliashvili, N., Kvikhinadze, R. and Kikiani, B., 1. Phys. B 281 (1995), 2453.
4. Yenen, 0., Jaecks, D. H. and Martin, P. J., Phys. Rev. A 35 (1987),1517.
5. Dowek, D., Dhuicq, D., Sidis, V. and Bafat, M., Phys. Rev. A 26 (1982),746.
6. Schulman, M. B., Sharpton, F. A., Chung, S., Lin, C. C. and Anderson, L. W., Phys. Rev. A 32
(1985), 2100.
Hyperfine Interactions 132: 521-525, 2001. 521
© 2001 Kluwer Academic Publishers.
Abstract. An ion catcher as presented in this contribution is able to create cooled and very clean
singly-charged ion pulses out of a 'hot' beam within a very short period of time. Precision measure-
ments on shortlived radioactive nuclides become possible. This contribution describes experiments
with a 252Cf fission source at the 'gas-cooler' at ATLAS (Argonne Tandem Linac Accelerating
System) at the Argonne National Laboratories (ANL), Argonne, USA [1]. The system consists of a
gas-cell to stop and thermalize the ions, two extraction radio frequency quadrupole structures (RFQ)
to separate the ions from the buffer gas and a buncher RFQ to cool and accumulate the ions. The
system and its performance is investigated with two independent measurements. The transported
activity was measured to determine the efficiency of the system and time of flight measurements
(TOF) were performed to determine the transported masses with respect to the transported activity.
Key words: stopping, trapping, cooling, 252Cf fission fragments, ion catcher, gas-cell.
1. Introduction
well as the total efficiency for the transport of the fission fragments. Time-of-flight
measurements were carried out to determine the transported masses and compared
to those of the transported activity.
2. Setup
For the experiments a 4ILC 252Cf fission source placed in front of the gas-cell win-
dow was used as shown in Figure 1. The gas-cell is 170 mm long and has a open
diameter of 50 mm. The entrance window is made of Havar (a metal alloy) with a
density of 8.3 mg/cm 3 and a thickness of 2.3 mg/cm 2. The nozzle has an open di-
ameter of 1.7 mm. The gas-cell is segmented to apply electric fields for guiding the
ions towards the nozzle. The cooler consists of three sections as shown in Figure 1.
The two extraction RFQs are used to separate the ions from the buffer gas and to
maintain the right pressures to allow for differential pumping and the third section
is the buncher RFQ, where the cooling, accumulation and trapping of the ions takes
place. The three sections combined consist of 41 segments each 20 mm long and
18 mm in diameter. The open diameter between segments is 12 mm. The diameter
of the apertures between the different sections are 2, 2 and 6 mm. The maximum
pressure in the cell, limited by the stability of the entrance window, is 200 mbar
leading to 8 x 10- 5 mbar in the buncher. Behind this system a transfer line connects
the gas-cooler to the Canadian Penning Trap (CPT) where high-accuracy mass
measurements can be performed. At two positions in the transfer line a detector
system is mounted each consisting of a CsI detector for activity measurements and
a micro channel plate (MCP) for time of flight (TOF) measurements. A third CsI
detector was temporarily installed after the gas-cell to determine the total efficiency
of the gas-cell.
3. Experiments
The stopping properties of the gas-cell were simulated with the Monte Carlo Code
SRIM [6]. These results were used to calculate the stopping efficiency depending
on the buffer gas used, the pressure in the cell, the fission fragment and the angle
STOPPING OF RADIOACTIVE FISSION FRAGMENTS IN A GAS-CELL 523
III
100 100
II
/
III 00
~ ~
.~
I j oo
}.
(0
1-0
~
20
0
~II
I
I
20
i.- ][ I
I
.m ..0() .3) 0 ;a) .c) Ell .fiO -40 ·20 0 20 40 00
Angle IdcgJ AngJeldegl
Figure 2. The two fission fragments 143Cs and l08Tc stopped in Argon 75 Torr as a function
of the entering angle calculated with SRIM [6].
Table I. The different efficiencies obtained with measurements and SRIM simu-
lations. The total efficiency was measured behind the cell nozzle, the stopping
efficiency was calculated using SRIM and the extraction efficiency combines the
two by assuming that only stopped ions are extracted
Argon 67 0.62(7)
Argon 133 1.0(1) 65(10) 1.3(3)
Argon 200 1.3(2)
Helium 67 0.14(5)
Helium 133 0.44(8) 1.9(3) 24(6)
Helium 200 0.7(1) 2.9(5) 24(5)
of entrance. The final positions of the ions inside the cell can be used to optimize
the gas pressure to direct the ions as close as possible towards the nozzle. Due to
the different angles the effective window thickness changes and thus the ions have
to travel through different amounts of material. This is the reason for the three
maxima in the angle distribution in Figure 2.
At the given geometry and source strength the total efficiency of the gas-cell
for transported activity can be determined, by measuring the transported activity
behind the gas-cell nozzle. With the calculated stopping efficiency from the sim-
ulations and the strength of the source, the extraction efficiency of the cell can
be deduced by comparing the simulated number of stopped fission fragments per
time unit to the count rate behind the nozzle. The results for several pressures with
helium or argon as buffer gas are shown in Table 1. The difference in the stopping
efficiency of the two investigated gases is due to the different stopping powers
(mass-dependent) of the two gases. In the case of extraction it is not understood
why there is a factor of twenty difference in efficiency.
524 M. MAIER ET AL.
PI ~u
1 c.J"b
I ..... hon DD Arg(lrl
PI ~u
t(l n.,
- pt nt ze
1011dRF~
PI
I
.or " ry tr:lft~por' _30 ptimlzed RF forupt"C'CM mas
aCI II
.ll'.icl., ~~"
I',. TIl"
PI . tI
PI k a
.,
.t ouplimized NF for II b Itr mllU!H
Pi II. '*
[[1r 'I I I
-- _
" .. _.
11 am. i<
101JS . Ar Peak
36.Sm
Corre:sponding COU Il I rlt e@ Csi. 1H'l e.I. 0.8H< ell: O.41lz
Figure 3. Time-of-flight measurements for different mass ranges and the corresponding count rates
at the CsI detector. The left picture shows the calibration with ionized argon gas and the three right
pictures the transported masses depending on the RF settings. The two dots indicate the two expected
main fission fragments, 143Cs with 78.3 MeV and 108Tc with 103.5 MeV.
4. Conclusion
The efficiency for stopping, extracting and transporting ions in a gas-cell was
determined using a off-line 252Cf source. The stopping process seems to be well
understood, however the extraction efficiency has to be further investigated to un-
derstand the differences between helium and argon. As a possible improvement for
the off-line tests a collimator could be used to investigate the efficiency dependence
on the entering angle and therefore position of the stopped ion inside the cell. The
total efficiency of the gas-cell has to be investigated with additional electrical fields
to guide the ions. The attempts to break up the molecules by applying accelerating
fields failed due to discharge problems inside the system. To prevent the forming
of molecules is the better approach. Therefore the next step is to get the system as
clean as possible.
STOPPING OF RADIOACTIVE FISSION FRAGMENTS IN A GAS-CELL 525
References
1. Savard, G., Canadian Penning trap at ANL, Oral contribution to APAC' 2000.
2. Paul, M. et aI., Heavy ion separation with a gas-filled magnetic spectrograph, Nucl. Instrum.
Methods A 277 (1989), 418-430.
3. Geissel, H. et aI., Ions penetrating through ion-optical systems and matter, Nucl. Instrum.
Methods A 282 (1989), 247-260.
4. Savard, G. et aI., The accuracy of heavy-ion mass measurements using time of flight ion
cyclotron resonance in a Penning trap, J. Appl. Phys. 68(9) (1990).
5. Dawson, P. H., Quadrupole Mass Spectroscopy and Its Applications, Elsevier Science, New
York,1976.
6. Ziegler, 1. F. et al., Stopping and ranges in matter (SRIM2000), http://www.research.
ibm.com/ionbeams/.
Hyperjine Interactions 132: 527-530, 2001. 527
© 2001 Kluwer Academic Publishers.
Abstract. A beam cooler for low-energy ion beams was constructed to improve the ion optical
properties of radioactive ion beams produced at the IGISOL facility in Jyvaskyla. The beam cooler is
a buffer gas filled RF-quadrupole. The delay properties and the possibility to accumulate a continuous
IGISOL beam and release it in short bunches is discussed.
Key words: collisional cooling, RF-quadrupole, ion-guide, on-line mass separation, Penning trap.
Low-energy radioactive ion beams are produced at the IGISOL facility [1] in
JyvaskyHi by the ion guide method [2] utilizing fission, light ion and heavy ion
reactions. The ion guide method is chemically non-selective and very fast. It there-
fore allows for studies of isotopes far from stability with half-lives of the order of
1 ms. The chemical non-selectivity makes it possible to study radioactive isotopes
of even refractory elements. It also creates a problem in the form of isobaric conta-
mination especially when using a non-selective reaction like fission. As a solution
a Penning trap system is being built which will comprise of two traps inside the
same 7 T superconducting solenoid. First one is a cylindrical Penning trap for
isobar separation in order to produce isotopically pure ion beams. The second trap
will be used for a high resolution mass spectroscopy of radioactive isotopes. To
be able to inject an ion beam into the trap, the beam has to be of good ion optical
quality and/or bunched. The IGISOL beam has typically a better emittance than
beams from conventional ion sources, but it has a relatively large energy spread
of >50 eY. This is far too large to be injected efficiently in continuous mode [3]. To
overcome this problem, an ion beam cooler [4, 5] was constructed. The beam cooler
can reduce the energy spread of the mass-separated IGISOL beam from'" 100 to
below 1 eV with 60% efficiency allowing a more efficient continuous injection. The
cooler can also be used as a buncher, where the continuous beam is accumulated in
a trapping region at the end of the device and released in short bunches. This mode
* Corresponding author.
** Present address: EP-division, CERN, CH-1211 Geneva, Switzerland.
528 A. NIEMINEN ET AL.
40 kV
39 kV
o ground
makes it possible to use a pulsed injection into the Penning trap [3]. A similar
installation is operational also at the ISOLTRAP facility at CERN [6].
The beam cooler for low-energy radioactive ion beams was constructed in the
end of 1999 and has been operational since then. It was built in one of the three
beam lines of the IGISOL facility. The three lines are at 30° angles relative to each
other served by an electrostatic switchyard. The central line is the dedicated nuclear
spectroscopy beam line and it also connects to the collinear laser spectroscopy
beamline which is upstairs from the IGISOL area. The cooler is installed in the
left beamline on a 40 kV HV platform. The Penning trap is being installed on
the same platform and the transport of ions between cooler and trap is taken care
of by a 1 kV transfer line. There already exists a connection from the cooler-trap
beamline to the central/laser beamline. The beam is extracted from the cooler and
accelerated to the 1 ke V transport energy. The beam is deflected by 90° and after
that further accelerated to the full 40 ke V energy, see Figure 1. A cooled beam has
already been successfully transported upstairs to the laser spectroscopy setup.
The cooler is a radiofrequency quadrupole with buffer gas load where the ions
are kept confined in the RF-quadrupole field as they loose energy in collisions with
the buffer gas atoms eventually ending up in thermal equilibrium with the buffer
gas and the driving RF-field. After they have been cooled, the ions are essentially
on the central axis of the quadrupole and their forward momentum is lost leaving
the ions to move back and forth along the axis by diffusion and space charge effects.
Unless additional electric fields are applied, the evacuation of the cooler is solely
governed by these two effects and can be very slow. As the length of the cooler is
40 cm and the evacuation through the injection aperture is prevented, in principle,
the longest distance the ions have to travel by diffusion is 80 cm in 0.1 mbar helium.
The speed of the cooling process is important when isotopes with short half-lives
are studied. The IGISOL method is very fast and it can produce radioactive ion
beams of isotopes having half-lives of the order of 1 ms. In order to preserve access
to these species, the beam cooling has to be as fast. The cooling process can be sped
TIME CHARACTERISTICS OF THE ION BEAM COOLER-BUNCHER 529
1000
800
CI)
..-
c:: 600
:::J
0
0
400
200
Number of segments Lengths of the segments (cm) DC voltage of the segments (V)
1 ~ 0
2 17.5,22.5 2, 0
3 17.5,12.5, IO 4,2,0
4 17.5, 12.5,5,5 4,2, 1,0
up by applying an electric field in the axial direction towards the exit. The axial
field to speed up the evacuation from the cooler was implemented by segmenting
the 40 cm long RF-quadrupole rods into shorter segments and applying a different
dc voltage on the segments so that when approaching the exit of the cooler the dc
level is dropping from 4 to 0 V. The speed of the cooler was studied by injecting a
pulsed separator beam into the cooler and observing the change in the pulse shape
after the cooler. The injected pulse had a rise and fall time of 10 I1-s. The cooled
ion pulse shape was measured by a micro sphere plate (MSP) by counting the
ions vs. time in a cycle TDC spectrum. Typically 1000 cycles were recorded with
100 I1-s/channel. Figure 2 shows typical TDC spectra from these measurements.
Four different segmentation schemes were investigated, varying the length of
the segments and the voltage on the segments. The values are listed in Table I. The
cooler evacuation times are shown in Figure 3 with different segmentation schemes
numbered as in Table I.
Without axial field, i.e., with one segment, the evacuation time is relatively
long, over 60 ms. Only dividing the one segment into four gives already evacu-
530 A. NIEMINEN ET AL.
70 r-~--'---~-.--~--.-~---r--~-'--~
~.---------.---------~
60
.
50
40
30
• ---' -_._- 2
20 • ---- _.--. --- - - - - . .
10
...
4 3
3
... . ........ •
. ... .............. • ............ 4
2
0
0 .09
• 0 .10 0 .12 0 .13 0 .14 0 .15
PHe (mbar)
Figure 3. The cooler evacuation times with different segmentation schemes.
ation times close to 1 ms level which is short enough not to reduce the selection
of radioisotopes available for study from the IGISOL. For transporting the beam
into a Penning trap for high-precision mass measurements, the cooler speed is not
as essential, since the mass measurement itself sets a limit for the half-life. The
shortest half-lives of a radioactive isotopes in a Penning trap mass measurement
have been the mass of 33Ar (173 ms) [7] and recently the mass of 74Rb (65 ms) [8]
using the ISOLTRAP spectrometer.
The other possible injection scheme into the Penning trap is the pulsed injection
where the trap potential can be dynamically adjusted to capture the arriving ion
cloud. IGISOL beam bunching was also tested preliminary by applying similar
segmentation as in Table I, number of segments = 4, with the exception that the
last two segments were 7.5 and 2.5 cm long. A trapping potential was created by
applying a positive voltage on an end plate electrode located after the last segment.
It was shown that the beam can be accumulated in the potential minimum and
subsequently released as short bunches with length of 20 ~s (FWHM). With a
better optimization of trapping potentials at the end of the RF-quadrupole this value
is expected to reduce to 5 ~s.
References
I. PenttiHi, H. etai., NucZ. Instrum. Methods B 126 (1997). 213-217.
2. Arje, J. et ai., Phys. Rev. Lett. 54 (1985),99.
3. Raimbault-Hartmann, H. et al., NucZ. Instrum. Methods B 126 (1997), 378-382.
4. Nieminen, A. et aI., Hyp. Interact. 127 (2000),507-510.
5. Nieminen, A. et al., NucZ. Instrum. Methods A (2000), in press.
n. Herfurth, F. et al., NucZ. Instrum. Methods A (2000), in press.
,. Herfurth, F. et ai., submitted to Phys. Rev. Lett. and this issue, p. 309.
8. Herfurth, F., private communication.
Hyperjine Interactions 132: 531-534,200l. 531
© 2001 Kluwer Academic Publishers.
Abstract. A new concept for direct mass measurements of short-lived nuclei with an electrostatic
time-of-flight mass spectrometer is described. The spectrometer can be coupled to an in-flight separa-
tor such as SHIP or FRS via a gas stopping cell and a gas-filled linear Paul trap. The time required for
mass measurement is of the order of a few milliseconds allowing one to investigate nuclei with these
short half-lives. A mass-resolving-power m / L'lmpWHM = 50000 has been reached. First results of
the range-focusing technique are presented, which is under development for the efficient stopping
and extraction of relativistic exotic nuclei.
ectrum
Start
Figure 1. Schematic view of an electrostatic time-of-flight mass spectrometer, which can be coupled
to an in-flight separator via a gas stopping cell and a gas-filled linear Paul trap system.
end. In this linear Paul trap the ions can be accumulated, stored, cooled, and
bunched. The device acts as ion source, from which the ions are injected into
• an energy-isochronous (multiple-) reflection time-of-flight mass spectrome-
ter with electrostatic ion mirrors [1]. In these devices the time-of-flight be-
tween ion injection (ejection from the Paul trap) and their arrival at the micro-
channel plate stop detector (MCP) is independent of their kinetic energy and
depends only on their mass. The mass of an ion is obtained by the comparison
of its time-of-flight with that of at least two known reference masses.
Figure 1 shows a schematic view of a system which can be used for multiple
reflections. With such a spectrometer of 35 cm length and ions with kinetic energies
of 6 keY up to hundred reflections have been observed at a residual gas pressure
of the order of 10-7 mbar, the efficiency was 20% [2]. The remarkable property
and the potential of these spectrometers is that a mass resolving power of 50000
is reached after 31 reflections and the measurement time is less than 800 ).!S for
ions with mass number A = 28 [2]. Using different stable Kr and Pb isotopes a
relative mass accuracy 8m/m = 2-3 . 10-6 corresponding to 100-600 keY has
been reached. Presently several applications are considered and their feasibility is
under investigation:
• The spectrometer can be used for mass determination and identification of the
extracted ions and/or molecules created in the stopping cell. It can also serve
as mass filter, when the ejection mirror is properly pulsed to transmit only ions
within a certain mass range.
• Direct mass measurements on short-lived nuclei with half-lives of the order
of milliseconds are feasible because the time for stopping, extraction, cooling,
and mass determination is of the same order of magnitude [3-5]. Coupled for
instance to the SHIPTRAP [6] facility, the spectrometer can complement the
planned Penning trap mass measurements on transuranium elements.
• Decay-spectroscopy experiments with mass-separated exotic nuclei are possi-
ble if a position-sensitive detector is used as stop detector and the ions' flight
time and position is recorded event-by-event.
A NEW CONCEPT FOR TIME-OF-FLIGHT MASS SPECTROMETRY 533
One source for short-lived nuclei are fragmentation and/or fission reactions of rel-
ativistic projectiles and in-flight separation with, e.g., the FRS [7]. The fragment
beams are characterized by a broad energy distribution due to the nuclear reaction
process and the energy-loss straggling in the production target. For these nuclei
range focusing [8], which is achieved by using a dispersive magnetic dipole stage
and a wedge-shaped degrader, is essential in order to reach fast slowing down and
efficient extraction from the gas stopping cell. Typically they have specific kinetic
energies of several hundred MeVju and ranges of several gjcm2 . For instance, a
monoenergetic beam of 360 Me V ju 56Ni28+ ions has a calculated mean range of
4150 cm and a range straggling of 14 cm (FWHM) in Ar gas at normal temper-
ature and pressure. The range-focusing technique [8] has been developed at the
FRS [9], which yields a small spatial distribution of the stopped fragments, see
Figure 2. This is an essential prerequisite to obtain high collection and extraction
efficiency. As can be seen from Figure 2, the range-focusing technique can reduce
the range straggling by a factor five as compared to the conventional slowing
down in homogeneous degraders. In this measurement the 56Ni fragments had a
relative momentum spread of 1% and the achieved range straggling of the stopped
56Ni fragments of 14.8 mgjcm 2 in P-1O gas at normal temperature and pressure
corresponds to 9 cm. The same value has been measured with a 58Ni primary beam
having a momentum width of less than 10-3 .
. . ., ..;.."""".............
0.6 0.25
0.20
-z o.
- ",
~ 0.15
'-
z 6
-oli§ 0.10
0.2 ~
El. __ .. ~~=76. 6mg/cm?
0.05
h.
o0 ..........-'-.o-L-'--'-'--'---'-"'...... ~-'-----'-'-"'-'~---<D OOO ~
· ~~~~~~~~~~.o-L~
130 132 134 13 6 138 100 142 1.. 146 148 130 132 13. 136 138 140 142 144 146 148
degrader thickness x [arb. u.l degrader thickness x [arb. u.l
(a) (b)
Figure 2. (a) Number-distance curve of 360 MeVju 56Ni fragments stopped in P-IO gas (90% Ar,
10% CH4) at normal temperature and pressure [9]. In one case (open squares, dotted line) the ions are
slowed down with a homogeneous degrader, in the other case (full dots, solid line) the ions are slowed
down using the range-bunching scheme described above. The fragments cover a momentum width of
±0.5%. The range straggling of the ions (b) is obtained from the derivative of the number-distance
curve shown in (a).
534 C. SCHEIDENBERGER ET AL.
References
1. Wollnik, H., 1. Mass Spectrom. 34 (1999),991.
2. Casares, A. et al., Internat. 1. Mass Spec. Ion Proc., accepted for publication.
3. Savard, G. et al., this issue, 223.
4. Maier, M. et aI., this issue, 521.
5. Engels, O. et al., this issue, 505.
6. Marx, G. et al., this issue, 463.
7. Geissel, H. et al., Nucl. Instrum. Methods B 70 (1992),286.
8. Weick, H. et al., Nucl. Instrum. Methods B 164/165 (2000), 168;
Scheidenberger, C. et al., In: Proc. STORI'99, Bloomington, IN, AlP Conference Proceedings
512 (2000), p. 275.
9. Nankov, N., Thesis in preparation, University of GieBen.
10. Lunney, D. et al., this issue, 299.
11. Hausmann, M. et al., this issue, 291.
Hyperjine Interactions 132: 535-539,200l. 535
© 2001 Kluwer Academic Publishers.
Abstract. We have used REXTRAP at ISOLDE to test the feasibility of in-trap electron spec-
troscopy. The results of calculations, experiments with various electron sources as well as a first test
with trapped radioactive ions are presented.
Key words: conversion electrons, trapped ions.
1. Introduction
The thickness of radioactive sources is a limiting factor for high-resolution electron
spectroscopy, due to scattering in the source material. The detection of conversion
electrons from the decay of trapped radioactive ions may provide new possibilities
in electron spectroscopy since the trapped ions constitute a source of electrons
without energy loss or scattering.
In this contribution we report the first in-trap measurements performed in REX-
TRAP. REXTRAP is a large Penning trap which is used for bunching and cooling
radioactive ion beams for the post accelerating step of REX-ISOLDE [1]. The trap
is situated at a high-voltage platform in order decelerate the ion-beams coming
from ISOLDE. Strong magnetic field and electric potentials trap the decelerated
ions. The trapped ions are cooled by collisions with argon buffer gas molecules.
The introduction of a buffer gas in the center of the trap requires that this part of
the trap is separated by a set of diaphragms to provide differential pumping. For a
detailed description of REXTRAP see [2] and references therein.
The strong magnetic field inside REXTRAP (up to 3 T in the trap center) will
cause the electrons emitted by trapped radioactive ions to be transported within
a limited volume. Thus, the major part of the electrons emitted in the forward
direction will pass the ejection diaphragm (Figure 1(a)).
2. Experiments
2.1. MEASUREMENTS WITH ELECTRON SOURCES
The highly efficient electron transportation within the trap makes it possible to use
small-area high-resolution Si detectors. A RD EB IOGC-500P detector assembly,
536 L. WEISSMAN ET AL.
a.
1.0 ij I
0.8-r
0.9
·i lIli
~'
c
~
(j
. c~
CD
0.7
0,6
'"'ii os -0 0.5
IE
UJ 04
03
01 0.2
0.1
"a L.....~..J......~.....J..~~L..o...~"'O"""~.......L- O. OL.....~~.L......o.~~.L......o.~~=:. · ...:...""
·;.;. • •"'-'..>
. ~ a.....-.,
•• ...,. ......
.
, Zll 011: !D: IIIl 1]]] 0 200 400 600 BOO 1000
Energy [kEV] Energy [keV]
b. c.
Figure 1. (a) General overview of the inner part of the REX-TRAP together with an example of a
simulated trajectory for a 400 keY electron. (b) Calculated probability for electrons emitted from
the trap center to reach the detector (source diameter is 1 mm). (c) Experimental (circles) and the
calculated (solid line) detection efficiencies for in-trap measurements. The squares and the dashed
line represent the experimental intrinsic detector efficiency measured in the test chamber.
the source was chosen to be the same as the distance from the trap center to the
diaphragm. The in-trap spectra have almost two times worse energy resolution and
higher continuum background (Figure 2(b». This is due to the larger spread in
electron energy loss in the source material as the electrons emitted at large angles
with respect to the axis of the magnet are also transported to the detector. Moreover,
/
X-ray in the test chamber
6000 ..... .. _. . . . . . ... _..................................................................... .
4000
<II 30000
'5o
U 200 00
10000
116mT_
200 ·K ........ ·· .. "'" ........ ·..
trapped ions
.. 150
1:
"o
(.) 100 :. ·::::::2~~.~~~~~~~~:::::· . ·. . :. : )':nj~:":":::":':'
:::::::::::::::::.::J\
50 ..... :~:~ ................... ............. -.............................. ~ .............. .
oW-_r~'r~~~~~~~"'~~~~l~~~I\~~~
o 20 40 60 80 100 120 140 160 180 200
c. E(keV)
Figure 2. Comparison of the 131 Ba implanted source spectrum measured in (a) the test cham-
ber (no magnetic field); (b) the same source measured in the trap (strong magnetic field); and
(c) the spectrum measured for the trapped 116m In ions.
538 L. WEISSMAN ET AL.
during the measurements with high tension applied at the platform we observed a
background associated with stray electrons from the external beam line acceler-
ated by the applied high voltage (60 keV) and transported towards the detector
(Figure 2( c».
The measurements with the calibrated sources allow to determine the absolute
detection efficiency as a function of electron energy. The results are summarized
in Figure l(c), where the experimental efficiency (full circles) is compared to a
calculated efficiency curve (solid line). To calculate the latter we convoluted the
intrinsic efficiency of the detector, as measured in the test chamber (full squares and
dashed line), with a calculated transmission probability of electrons from a 1 mm
diameter source placed in the trap's center (Figure l(b». The result is corrected
for a higher probability of electron backscattering from the detector and a higher
probability for electrons to escape in vicinity of the detector edge in the magnetic
field.
3. Conclusion
We have demonstrated the feasibility of in-trap electron spectroscopy. In spite of
the fact that we did not observe an improvement in energy resolution in the first
measurements, the advantages of in-trap experiments as the high detection effi-
ciency and significantly better peak to background ratio are demonstrated. There
FEASIBILITY OF IN-TRAP CONVERSION ELECTRON SPECTROSCOPY 539
Acknowledgement
L. W. wishes to acknowledge Dr. M. Dietrich for technical assistance and help in
the preparation of the l3lBa source.
References
1. Habs, D. et aI., NucZ. Phys. A 616 (1997), 29-38.
2. Ames, F. et ai., this issue, 469.
* Hyperfine Interactions 132: 541-543,2001. 541