Ternary ssion yields of 241Pu(nth,f)

oster a 1, H. Faust b, G. Fioni c, T. Friedrichs b d,

U. K ; ;

M. Gro e and S. Oberstedt b

a Technische Universitat Munchen, Physik-Department, 85748 Garching, Bavaria,
Germany
b Institut Laue Langevin, 38042 Grenoble, France
c CEA Saclay, DSM/DAPNIA/SPHN, 91191 Gif-sur-Yvette, France
d Technische Universitat Braunschweig, Institut fur Metallphysik und Nukleare
Festkorperphysik, 38106 Braunschweig, Germany
e Ludwig-Maximilians-Universitat, Sektion fur Physik, 85748 Garching, Bavaria,
Germany

Abstract
Ternary events in the thermal neutron induced ssion of 241 Pu(n,f) were studied
with the recoil separator LOHENGRIN at the high- ux reactor of the Institut Laue
Langevin in Grenoble. Yields and energy distributions could be determined for most
isotopes of the elements hydrogen to oxygen. Also several heavier nuclei up to 30 Mg
could be observed. Yields were measured for 42 isotopes, for further 17 isotopes
upper limits could be deduced. For the rst time the halo nuclei 11 Li, 14 Be and 19 C
were found in neutron induced ssion with yields of some 10,10 per ssion.

Key words: neutron induced ssion, ternary ssion, 241 Pu(n,f), recoil separator
PACS: 25.85.-w, 25.85.Ec, 27.90.+b, 29.30.Aj, 29.30.Ep,

Introduction

More than fty years ago rst experimental evidence for so-called \long range
alpha particles" 2 (LRA) stemming from ternary ssion was found [1{3]. Since
1 Corresponding author: Ulli Koster, Email: [email protected]
2 This notation serves for dierentiation against less energetic alphas from radioac-
tive alpha decay.

Article published in Elsevier Preprint NPA (1999) 1{22

then detailed studies of ternary ssion 3 were done with dierent coincidence
set-ups, see e. g. [4,5] for an overview. Yields and energy distributions of var-
ious ternary particles were measured for several thermal neutron induced s-
sioning systems: 233 U(n,f) [6], 235 U(n,f) [7{9], 239 Pu(n,f) [10] and 241 Am(2n,f)
[8{12]. However for other ssioning systems only scarce data existed, which
made a systematic study of the dependence of yields and energy distributions
of ternary particles on the ssility and neutron excess of the compound system
dicult.
There is still no theory which could reproduce the individual yields and en-
ergy distributions of ternary ssion fragments from rst principles. Only in
the special case of alpha accompanied ssion of 236 U a numerical solution of
the Schrodinger equation describing \alpha decay during ssion" was possi-
ble and gave roughly correct values for the absolute emission probability and
energy distribution of the alpha particles [13]. A similar approach had been
used by C^arjan to reproduce average energy and emission probability of LRA
in 236 U [14,15]. For cold spontaneous 10 Be accompanied ssion of 252 Cf the
relative probabilities of the associated fragment mass splits have been calcu-
lated recently [16], but the inverse problem, i.e. the calculation of individual
ternary particle yields with a xed or averaged mass split for the light and
heavy fragments, has not been attacked yet. The random neck rupture model
of Rubchenya and Yavshits [18] can explain the LRA emission probability in
dierent ssion systems, but to reproduce the yields of individual ternary par-
ticles a t to experimental data is used. Several other semi-empiric ternary
ssion models exist [17,19{22], but they also require a basic set of yield data
to t the free constants.
Very low yields of ssion products can be measured with a recoil separator.
To date the most sensitive recoil separator is LOHENGRIN at the high- ux
reactor of the Institut Laue Langevin (ILL) in Grenoble. Recently a series
of measurements was performed at LOHENGRIN to study the ternary s-
sion of: 229 Th(n,f) [23], 233 U(n,f) [24,25], 239 Pu(n,f) [23,25], 241 Am(2n,f) [26],
245 Cm(n,f) [25,26] and 249 Cf(n,f) [27].

Here we will report on the investigation of 241 Pu(nth ,f), where before only data
for the two most abundant and technologically important ternary particles
existed. The yield of 4He had been determined to (1:86  0:05)
10,3 per
binary ssion and the yield of 3H to (1:41  0:06)
10,4 per binary ssion or
(7:61  0:19) % of the 4He yield [28].
3 In fact, here we are not discussing \true" ternary ssion with a breakup in three
fragments of about equal size, but a process where additionally to two heavy nuclei
with masses comparable to binary fragments a light charged particle is emitted. For
brevity we will call this process of \light charged particle accompanied ssion" just
ternary ssion or tripartition.

2
Experiments

Targets

In total four dierent targets were used. They were prepared by the IRMM
(Institute for Reference Materials and Measurements) of the Joint Research
Centre of the European Commission in Geel. The plutonium oxide was de-
posited within an area of 70
5 mm2 on a platinum-coated 4 titanium backing.
A \thin" target with 208 g Pu per cm2 thickness was used to determine dur-
ing a measurement of binary ssion [29] the yield of 10 Be as a reference point
for further ternary ssion measurements. Another \thin" target (100 g/cm2
Pu) was used to measure exactly the energy distribution of the helium iso-
topes down to 3 MeV. Two \thick" targets with 379 and 362 g Pu per cm2
thickness served for the measurement of ternary ssion yields relative to 10 Be.
The rst three targets had in November 1990 an isotopic composition of 87 %
241 Pu, 1.6 % 239 Pu and 11 % other plutonium isotopes with low ssion cross
sections. At the time of the measurements about 20 %, 31 % and 28 % respec-
tively of the 241 Pu had decayed to 241 Am, but as discussed in [30] during two
weeks of measurement 5 the contribution of other ssion systems (e.g. from
breeded 242 Am(n,f)) will stay below 6 %. The last target was used 9 months
after plutonium separation, thus limiting contributions from other ssion sys-
tems to below 2 %. The targets were covered with a 2500  A thick nickel foil
to reduce loss of target material by sputtering and thus guarantee a smoother
burnup.

Separator

The targets are brought in an evacuated beam tube of the ILL high- ux reactor
(58 MW thermal power and 1:5
1015 cm2 s,1 maximal unperturbed thermal
ux). At the target position, about 50 cm from the reactor core, the thermal
neutron ux is about 5:3
1014 cm2 s,1. The epithermal neutron ux is more
than two orders of magnitude smaller and the fast neutron ux is more than
three orders of magnitude smaller. Both contributions can be neglected for our
purposes. Recoiling ssion products leave the target with a small energy loss
and y as highly charged ions through the beam tube. While binary fragments
have typically ionic charges of 20 to 25, light ternary particles are mostly
fully stripped. After 8 m the ions enter a horizontally de ecting homogeneous
magnetic sector eld, separating ions according to the ratio of momentum
4 The 400 g per cm2 thick Pt layer reduces the diusion of plutonium into the Ti
backing.
5 The second target was kept not even 48 h in the neutron ux.

3
to ionic charge p=q. Subsequently they pass a vertically de ecting cylindric
condenser, which provides a separation according to the ratio of kinetic energy
to ionic charge E=q. Both elds are arranged to the double focusing parabola
spectrometer LOHENGRIN [31,32], separating ions of the same A=q ratio
onto a parabola, on which the position is given by the kinetic energy of the
ion [33]. For cases where no ultimate energy resolution is needed, but rather
a high beam intensity or transmission should be achieved, a second magnet
was designed which focusses about 40 cm of the mass parabola to a small area
(some cm2), see gure 1. This so-called RED (Reverse Energy Dispersion)
magnet [34] was used for the measurements with the \thick" targets, while
the measurements with the \thin" targets were made without RED magnet.

Fig. 1. Arrangement of the recoil separator LOHENGRIN.

Whereas the usable range of magnetic elds (up to 0.24 T for the rst magnet
and maximal 1.6 T for the RED magnet) is sucient to cover most of the
required separator settings, the maximal high voltage of the condenser limits
the E=q ratio to about 6.5 MeV, thus excluding the measurement of high
kinetic energies at low charge states.
The ight time of the fragments from the target to the detector is below
2 s. Thus decay losses of the studied particles with hal ives larger than
milliseconds can be neglected.

Particle identication

LOHENGRIN only separates particles according to their A=q and E=q ratios.
For an unambiguous determination of the fragment mass it is still necessary
to measure directly the kinetic energy of the fragments. With the known set-
tings for E=q and A=q thus the ionic charge q and nally the mass A can
be determined. The kinetic energy is measured with an ionization chamber

4
placed in the focus of the separator. Using a split anode for the readout of
the energy signal, additionally the nuclear charge Z can be deduced from the
specic energy loss in the rst section of the ionization chamber, the so-called

E part. The used ionization chambers are described in detail in references
[35,36]. The counting gas (isobutane) pressure was adapted in the range of
15 to 160 mbar for best Z identication. In all cases the Z -resolution of the
ionization chamber was by far sucient for a clear identication of the ternary
particles.

Calibrations

Magnet calibration
During the measurement the magnetic elds are controlled with NMR probes
and stabilized to some 10,5. However the NMR probe measures the eld only
locally, whereas the particles are de ected according to the integrated eld
along their ight path. Hysteresis and remanence eects may change the cal-
ibration of the magnet. The optimum magnetic eld settings are found by
scanning the beam horizontally over the detector placed in the focus, while
keeping the electrical eld xed. The so determined magnet constants, called
 (for the rst magnet) in [37] and  (the ratio of elds of RED to rst mag-
net), stayed constant within a range of  2
10,4 during the measurement
periods of about 2 weeks.

Energy calibration
The absolute energy calibration of the separator can be checked with direct
(n,X)-reactions producing monoenergetic particles. The electric and magnetic
elds are scanned simultaneously to nd the high energy edge of the energy
distribution, which has typically a low energy tail due to energy loss in the
target and scattered particles, see gure 2.
Three reactions were used:
(1) 6Li(n,)t giving tritons with 2.73 MeV kinetic energy and alphas with
2.05 MeV. These calibrations are done \o-line" by introducing a thin
6 Li target at the target position [37].
(2) 59 Ni(n,) giving alphas with 4.78 MeV kinetic energy. The unstable 59Ni
is produced \on-line" by neutron capture on 58 Ni present in the nickel
foil which is used to cover the target.
(3) Decay alphas of 242 Cm with up to 6.11 MeV. The 242 Cm is here produced
by beta-decay of 241 Pu to 241 Am which is subsequently transmuted to
242 Am by neutron capture and then decays to 242 Cm. Also this reaction

5
 5
10

count rate (a.u.)

4
10
4He

3
10

4 4.2 4.4 4.6 4.8 5 5.2 5.4 5.6 5.8 6 6.2

Energy (MeV)

Fig. 2. Energy spectrum of 4 He used for on-line energy calibration. The peak around
6.1 MeV is due to decay alphas from 242 Cm and the peak at 4.78 MeV comes from
59 Ni(n,). Shown is a raw spectrum without correction for energy loss in the target
and cover foil.
is used \on-line" without target change.

Calibration of the ionization chamber

The energy and nuclear charge calibration of the ionization chamber was done
with reference spectra using separator settings A=q = 3. Here the most abun-
dant ternary particles (3H, 6 He, etc.) and scattered stable isotopes (e.g. 12 C,
24 Mg, 27 Al) are easy to identify and make the calibration obvious, see gure
3.

Data evaluation

Background
Instabilities of the high tension or sparks in the condenser can cause back-
ground in the
E/E scatter plots. Scattered binary particles occur especially
with separator settings close to typical values for binary particles (A=q = 4 to

6
 A/q = 3
E/q = 3.75 MeV
Si
Al
E (channels)

Mg
Na
Ne
F
O
N
C
B
Be
Li
A= 9 12 15 18 21 24 27 30

Etotal (channels)

Fig. 3.
E-E-scatter plot with separator setting A=q = 3 and E=q = 3:75 MeV. The
measurement time for this spectrum was 6.1 h. The horizontal scale is proportional
to the particle energy (and due to a xed A=E ratio also to the mass), whereas
the vertical scale is roughly proportional to the nuclear charge Z . Scattered binary
particles create background close to the diagonal in the upper part of the spectrum.
Background in the lower part is due to pile-up (from abundant 3 H and 6 He) and
particles scattered in the entrance window of the ionization chamber (tails going to
the top and left which can be seen at 6 He, 9 Li and 12 C). One channel corresponds
approximately to 75 keV.
5, E=A around 1 MeV).
High background from natural alpha decay and (n,) reactions is present for
4 He below 6.1 MeV. Also for 3 H below 2.7 MeV signicant background arises
from 6 Li(n,) reactions. This 6 Li is sputtered o from calibration targets (see
above) and sticks to the target holder or the beam tube. For these cases only
the data points with higher kinetic energy are evaluated.
Some stable particles (12C, 14N, 16 O, etc.) occur very abundantly. They are
knocked out by ssion fragments from the target and its cover, from the resid-
ual gas in the vacuum chamber or from diaphragms in the separator. Their
abundance drops rapidly down towards higher kinetic energies. However, if
the contribution of such background gets too high, the events stemming from
ssion cannot be identied and no ssion yield can be determined for these
stable isotopes.

7
Typically the background conditions are more favorable for isotopes with a
large neutron excess than for proton-rich isotopes. This explains partly the
great dierences in measured upper limits for rare ternary isotopes on the
neutron-rich and the neutron-decient side.

Separator acceptance
The energy dispersion coecient 6 of the separator is 7.2 m, i.e. a 1 % large
slice of the energy distribution is spread out over 7.2 cm of the mass parabola.
The accepted energy range is therefore proportional to the chosen energy. To
correct this eect, all count rates are divided by the set energy for normaliza-
tion.
In mass direction (i.e. perpendicular to the mass parabola) the dispersion co-
ecient 7 is 3.24 m. Due to the dierent mass defects, the dierence between
ionic mass and atomic mass depending on the charge state and relativistic
corrections not all isotopes with the same A=q ratio lie exactly on the mass
parabola. The actual transmission is a complicated function, which could be
calculated by folding the size of the target with the transmission through in-
termediate diaphragms and vacuum chambers, the edge focussing by the RED
magnet and the acceptance of the circular entrance window of the ionization
chamber. It can also be experimentally deduced from the scan curve of the
main magnetic eld (see gure 4). The FWHM is typically 0.2 to 0.3 % of the
magnetic eld. For small deviations from the ideal eld the yield of a given
isotope is corrected for the transmission, in case of larger deviations (e.g. for
the light particles 3 H, 6 He, etc. in a spectrum with A=q = 3) two spectra
with dierent magnetic eld settings have to be taken. The scan curve of the
RED magnet is much broader (FWHM some %), therefore no correction is
necessary.

Target burnup
The ssion rate of the target decreases with time. This is mainly due to the
nuclear burnup in the high neutron ux. Additional losses of target material
occur from sputtering by ssion fragments, evaporation from the heated target
and diusion into the target backing. The slight reduction of the neutron ux
(about 3 % during one reactor cycle of 52 days) can be neglected. The decrease
6 Particles with a kinetic energy E0 +
E are separated in the plane of dispersion
by a distance
x from the reference particles with energy E0 . For small deviations
(
E  E0 ) the linear relation with the energy dispersion coecient DE applies:

x = DE
EE0 .
7 In analogy to the energy dispersion the mass dispersion coecient Dm is dened
as:
x = Dm
m .
m

8
 1

0.8

Relative count rate

0.6

0.4

0.2

0
0.998 0.999 1 1.001 1.002
Detuning of main magnet (B/B0)

Fig. 4. Measured transmission in dependance of a detuning of the main magnetic

eld.
of the ssion rate is monitored with a reference spectrum taken once to twice
a day and can be described with an exponential decay curve.
It should be noted that plutonium targets can show a less regular burnup
characteristic than other actinide targets. This might be due to a rapid mi-
gration into the target backing or sidewards out of the target area (see also
[30]). The former explanation is supported by diering burnup times for heav-
ier and lighter isotopes and a broadening of the \natural" width of the energy
distribution. These eects are dicult to quantify, but contribute to the sys-
tematic errors. For very precise measurements plutonium targets cannot be
used \completely", but should be changed after some days.

Energy loss
To deduce the original kinetic energy distribution of the ssion fragments from
the measured one, the energy loss in the target and in the cover foil has to
be taken into account. The energy loss for all measured isotopes and kinetic
energies was calculated with the Monte-Carlo program TRIM 8 [38]. Note that
8 Since TRIM does not contain transuranium elements as target, the calculation
was performed for uraniumoxide and scaled. The resulting uncertainty is negligible.

9
the energy loss in the entrance window of the ionization chamber occurs after
the mass separation and has therefore not to be considered here.

Ionic charge distribution

Due to the limitations of the separator elds and the available beam time it is
not possible to measure the yields of a given isotope at all possible ionic charge
states. Therefore a model has to be applied to inter- and extrapolate the ionic
charge state fractions to other energies and isotopes. Several empiric formulae
are available to predict the mean ionic charge states, the width of the charge
state distribution and thus allow to calculate the equilibrium charge state dis-
tributions (ECSDs) [39{47]. However, the ECSDs used for these formulae were
measured under well dened experimental conditions, mostly by transmitting
low charged ions from tandem accelerators through free-standing foils made
from pure materials. On the other hand at LOHENGRIN the charge state
before passing the foil is not well known. Moreover the target cover foils are
quite fragile (0.25 m of nickel) and coated with a protective lm of acrylic or
polystyrene for better handling. The nuclear heating in the reactor will bring
the target to above 600 C and the coating evaporates. Still some atomic lay-
ers of e.g. carbon may be left and then in uence the ECSDs. Therefore the
constants of such formulae have to be determined for each target by measuring
in-situ the ECSDs of appropriate isotopes (10Be, 14 C, 20 O).
None of the investigated formulae was able to reproduce the ECSDs of all
elements with a given set of constants or with one single t parameter. For
better comparison with other LOHENGRIN measurements we nally used a
modied reduced chi-square distribution with a parametrisation in analogy to
the data evaluation of [9,12,24,26].
The mean charge is calculated as:
 1  0 0  1  0:5 ,0:4811
q = Z + Z 2
@1 , exp @,2:7284
vion
Z + Z 2 AA (1)

and the width of the charge state distribution is:

 !
,0:475  1 0:447, 12
1
q = 0:426 , 
vion
Z + Z
Z+Z (2)

with two t parameters  and  depending on the element and target. vion is
given in units of cm/ns. This expression for q is, with exeption of the small
modications by  and 1=Z , identical to that of [43]. q is similar to formula
(6) in [44].

10
The fraction of an individual charge state can be calculated with a \reduced
chi-square distribution" [44], modied by a factor exp(,a).

F (q) = c
t(q) n2 ,1
exp(,a) (3)

2 n2
,( n2 )
exp( t(2q) )

with , being the gamma function and the constants:

c = 2
(Z ,2 q + 2) ;
q
t(q) = c
(Z , q + 2);
n = c
(Z , q +p 2);
 vZ ,q+1  1+ Z
a= v

where exp(,a) gives the stripping probability for the (Z -q+1)th electron. Its
orbital velocity in the Bohr model vZ ,q+1 is calculated from the qth ionization
energy given in [49].
The parameters  and  had to be adapted for each element. It appeared,
that  diered strongly between the third and fourth target 9 . This could be
explained by dierent evaporation residues on the target cover foil. In the
former case the nickel foil was covered by an acetate lm, in the latter by a
polystyrene lm. On the other hand  showed only a weak variation (ranging
from 0.02 to maximal 0.025). It is possible to t the constants by using the
measured charge state fractions of dierent isotopes of one element up to
about oxygen. Figure 5 shows one of the tted ECSDs for 14C. For these light
elements the ECSDs formula is mainly used for interpolation of the measured
charge state fractions and simplies a transfer of these ratios to other isotopes
of the same element. The systematic errors are therefore acceptable. However,
for lack of experimental data 10 , the trends of the charge state distribution
have to be extrapolated towards heavier elements. Here it had to be assumed
as in [26], that, compared to the big dierences for light elements, the variation
9 With the second target only helium isotopes were measured. The contributing
charge states (2+ and below 2.5 MeV per nucleon also 1+) can be measured directly
and no elaborate ionic charge correction is required.
10 It is in principle possible to measure also the ECSDs of elements like Mg, Si,
etc. in a dedicated experiment at LOHENGRIN. Instead of ternary particles, stable
isotopes of these elements would be used which are knocked out from an additional
layer containing the elements in question between the actinide and the cover foil by
collisions with binary fragments. In this way the constants could be pinned down
much better even for heavier elements, reducing the systematic errors signicantly.

11
of the tting constants from element to element continue to decrease and the
\constants" can be kept constant. The uncertainty of this extrapolation creates
the dominating part of the systematic errors.

80%

70%

Charge state fraction q=4

60% q=6
50%

40%
q=3
30%
q=5
20%
q=2
10%

0%
0 5 10 15 20 25 30 35
E (MeV)

Fig. 5. Measured charge state fractions and tted charge state distribution for 14 C.

Kinetic energy distribution

The total yield of an isotope is determined by integration over its yield at
dierent kinetic energies. In cases where, due to limitations in electric eld or
beam time, not enough dierent energies could be measured to determine the
energy distribution, the mean energy and/or the width of the energy distri-
bution had to be xed for the tting. For 4He we used the mean energy and
width from [28]. For 3 H we also used the mean energy from [28], but kept the
width as a free t parameter. For 2H we employed also the mean energy of
3 H, which is justiable regarding the systematics of other ssion systems [4].
For the rare isotopes we used an extrapolation which is consistent with the
systematics in other ssion systems and the results of trajectory calculations
for the nearby ssion system 242 Am(n,f) [9,21]. In all cases where the energy
parameters were xed for the t, an additional systematic error was added to
account for the uncertainties of these values.
Practically all energy distributions are well reproduced by a Gaussian. An ex-
eption is 4 He. It is well known [50], that the low energy part of the 4 He spec-
trum shows an enhancement over a symmetric Gaussian distribution, probably
due to contributions from the breakup residues of the particle unstable iso-
topes 5 He and 8 Be [5]. Therefore we used only the data points with energies
above 10 MeV for the integration of the yield.

12
Results

Fig. 6. Overview of the measured ternary particles arranged in form of a nuclide

chart. The numbers mark the mass of the lightest and heaviest measured isotope of
each element.
Figure 6 shows an overview of the measured isotopes. The detailed data are
given in tables 1 and 2. Emean and EFWHM are obtained by tting a Gaus-
sian distribution to the data points. Cases where these values were taken from
systematics are marked by the label \x" in the row of the corresponding
uncertainty. For better comparison to other ssion systems the yields are nor-
malized to a standard of 10000 for 4He. Note that the reference point itself has
a rather large error since the yield of 4He can only be determined from a small
energy range (10 to 13 MeV) 11 . An additional systematic error (dicult to
estimate and not included in table 1) could occur if the energy distribution
showed a non-Gaussian behaviour in the t range. The given errors are the
individual errors for each isotope and do not include the error of the reference
point.
In LOHENGRIN measurements the light charged particle is detected with-
out coincidence to the heavy fragments. Its origin from ternary ssion 12 has
therefore to be deduced from its kinetic energy distribution. Tables 1 and 2
show the maximal kinetic energy Ebin which the light charged particle could
attain in case it is produced in a binary reaction, e.g. 241 Pu(n,)238 U. A dash
indicates that the Q-value of such a direct reaction is negative, i.e. the reac-
tion is impossible with thermal neutrons. No entry is made in cases where the
11 The upper limit is given by the maximally achieved voltage in the LOHENGRIN
condenser.
12 It could also come from quaternary ssion, however such an event was never ob-
served and should be much rarer than ternary ssion. Breakup fragments from decay
of excited ternary 7 Li and of 8 Be which can give experimentally the impression of
quaternary ssion are discussed in [52,53].

13
mass of the potential heavy binary partner is unknown [51]. Especially for the
lighter isotopes it is evident from the high measured energies that they are
indeed produced in ternary ssion.
Figure 7 shows a plot of the individual isotopic yields. The strong proton even-
odd eect which favors emission of ternary particles with even Z is obvious
(see also gure 9). Also the less pronounced neutron even-odd eect is clearly
visible (staggering within one element).
4
10

10
3 He

2
10 H Be
Yield (relative to 104 4He)

10 C
Li O
1

-1
B Ne
10 N
-2
Mg
10 F
-3
10 Na
5 10 15 20 25 30
Mass
Fig. 7. Measured yields of ternary particles normalized to 104 for 4 He. The isotopes
of each element are connected by a line, dashed for odd Z and solid for even Z .
Upper limits are marked with an arrow. Some upper limits have been omitted for
sake of clarity.
The integrated mass and elemental yields are shown in gures 8 and 9 respec-
tively. For complete summing the missing individual yields (e.g. for mass 13
nuclei) were interpolated from neighbouring yields according to the ratios pre-

14
dicted by the Faust formula [22]. Contributions from particle unstable nuclei
are neglected. The mass yields were converted to absolute yields (per ssion)
with the ternary to binary ratio from [28]. For comparison the binary yields of
the light fragment group from [30,54] are also plotted. The strong oscillation
(see gure 8) of the ternary yields can be understood from the fact that most
ternary mass yields are dominated by a single isotope. Thus high yields of
individual isotopes (e.g. 10 Be, 14C, . . . ) lead to a signicant staggering of the
yields.
The detection limit for ternary particles is around 10,10. Figure 10 shows
that isotopes with large neutron excess can, despite their extremely low yield,
be clearly and unambiguously identied in the
E-E-scatter plot. Thus also
the neutron halo nuclides 11Li, 14 Be and 19 C could be identied with absolute
yields of 8:4(28)
10,10, 5:0(19)
10,10 and 4:7(26)
10,10 per ssion respectively.

-1
10
-2
10
-3
10
Yield (per fission)

-4
10
-5
10
-6
10
-7
10
-8
10
-9
10
-10
10
20 40 60 80 100
Mass
Fig. 8. Absolute mass yields from 241 Pu(nth ,f) in events per ssion.

15
 -3
10

-4
10
Yield (per fission)

-5
10

-6
10

-7
10

1 2 3 4 5 6 7 8
Z
Fig. 9. Elemental yields of ternary particles in events per ssion. A possible contri-
bution of low-energy (polar) protons to the hydrogen yield was neglegted.
Summary and Outlook

Ternary ssion of 241 Pu(nth ,f) was studied and yields and energy distributions
could be determined for most isotopes of the elements hydrogen to oxygen.
Also several heavier nuclei up to 30 Mg could be observed. Yields were mea-
sured for 42 isotopes, for further 17 isotopes upper limits could be deduced.
The measured yields and energy distributions give a good overview on the
ternary ssion of 241 Pu(n,f). In future this survey could be completed by mea-
suring some \missing" isotopes at the masses A=13, 23, etc. and improving
the statistics for the heaviest isotopes. Together with recent measurements of
other ssion systems this enlarged data set of ternary ssion yields will help
to improve the current ternary ssion models.

16
 400

350
A/q = 7/2 28
E/q = 4 MeV Mg
300 28
Na
E (channels)
250 21
F
21
200 O
14
N 21
N
150
14
C
100 14
B
14
50 Be
7
Li
0
0 50 100 150 200 250 300 350 400 450 500
Etotal (channels)

Fig. 10. Identication of 14 Be in a
E/E scatter plot taken during 19.5 h with the
separator set to A=q = 3:5 and E=q = 4 MeV.
The ssion system 242 Pu is particularly interesting since a direct comparison
between ternary yields from thermal neutron induced ssion 241 Pu(n,f) and
spontaneous ssion of 242 Pu (see e.g. [55{57]) will become possible once the
latter has been studied in more detail 13 . A comparison of the binary ssion
yields of 241 Pu(nth ,f) and 242 Pu(sf) [58] showed signicant dierences due to
a 6 MeV dierence in excitation energy.

Acknowledgements

We are grateful for fruitful discussions with Michael Hesse and Marcus Wost-
heinrich. Thanks for support by the Accelerator Laboratory of the TU and
LMU Munchen.
13 Several ternary yields are known in two other cases of spontaneous ssion:
248 Cm(sf) and 252 Cf(sf), but it is still dicult to nd sucient quantities of the
corresponding target nuclides 247 Cm and 251 Cf for a detailed study of the neutron-
induced ssion yields.

17
References

[1] T. San-Tsiang et al., C. R. Acad. Sci. Paris 223 (1946) 986.

[2] T. San-Tsiang et al., Phys. Rev. 71 (1947) 382.
[3] G. Farwell, E. Segre and C. Wiegand, Phys. Rev. 71 (1947) 327.
[4] Ternary Fission, C. Wagemans, chapter 12 in The Nuclear Fission Process, ed.
by C. Wagemans, CRC Press, Boca Raton (1991) p. 545.
[5] Particle-accompagnied ssion, M. Mutterer and J.P. Theobald, chapter 12 in
Nuclear Decay Modes, ed. by D.N. Poenaru, IOP Publishing, Bristol, 1996, p.
487.
[6] A. Vorobyov et al., Phys. Lett. 30B (1969) 332.
[7] A. Vorobyov et al., Phys. Lett. 40B (1972) 102.
[8] W. Baum, PhD thesis, TH Darmstadt, 1992.
[9] W. Baum et al., 6th Int. Conf. on Nuclei far from Stability and 9th Int. Conf. on
At. Masses and Fund. Const., Bernkastel-Kues, ed. by R. Neugart and A. Wohr,
Inst. Phys. Conf. Ser. No. 132, IOP Publishing, Bristol, 1992, p. 477.
[10] A. Vorobyov et al., Sov. J. Nucl. Phys. (1975) 248.
[11] F. Gonnenwein et al., 6th Int. Conf. on Nuclei far from Stability and 9th Int.
Conf. on At. Masses and Fund. Const., Bernkastel-Kues, ed. by R. Neugart and
A. Wohr, Inst. Phys. Conf. Ser. No. 132, IOP Publishing, Bristol, 1992, p. 453.
[12] Berthold Borsig, PhD thesis, Universitat Tubingen, 1993.
[13] R. Schafer and T. Fliessbach, J. Phys. G21 (1995) 861.
[14] N. C^arjan, A. Sandulescu and V.V. Pashkevich, Phys. Rev. C11 (1975) 782.
[15] N. C^arjan, J. Phys. (Paris) 37 (1976) 1279.
[16] A. Sandulescu et al., J. Phys. G24 (1998) 181.
[17] G.V. Valskii, Sov. J. Nucl. Phys. 24 (1976) 140.
[18] V.A. Rubchenya and S.G. Yavshits, Z. Phys. A329 (1988) 217.
[19] G.A. Pik-Pichak, Phys. At. Nucl. 57 (1994) 906.
[20] M. Hesse and F. Gonnenwein, Workshop on nuclear ssion and ssion product
spectroscopy, Seyssins, ed. by H. Faust and G. Fioni, ILL report 94FA05T (1994)
p. 18.
[21] Boualem Bouzid, PhD thesis, Univ. des Sciences et de Technologie Houari
Boumedienne, Alger, 1994.

18
[22] H. Faust and Z. Bao, Proc. Seminar on Fission \Pont d'Oye III",
ed. by C. Wagemans, Euratom Geel, EUR16295 EN (1995) p. 220.
[23] M. Wostheinrich et al., Acta Phys. Slov. 49 (1999) 987.
[24] M. Wostheinrich et al., 2nd Int. Workshop on Nucl. Fission and Fission-Product
Spectroscopy, Seyssins, ed. by G. Fioni et al., AIP Conf. Proc. 447, p. 330.
[25] Ulli Koster, PhD thesis, TU Munchen, 1999.
[26] Michael Hesse, PhD thesis, Universitat Tubingen, 1997.
[27] Marco Davi, PhD thesis, Universitat Mainz, 1997.
[28] C. Wagemans et al., Phys. Rev. C33 (1986) 943.
[29] T. Friedrichs et al., 2nd Int. Workshop on Nucl. Fission and Fission-Product
Spectroscopy, Seyssins, ed. by G. Fioni et al., AIP Conf. Proc. 447, p. 231.
[30] Thomas Friedrichs, PhD thesis, TU Braunschweig, 1997.
[31] E. Moll et al., Int. conf. on electromagn. isot. sep., Marburg, ed. by H. Wagner
and W. Walcher, report BMBW-FB K70-28 (1970) p. 241.
[32] E. Moll et al., Nucl. Instr. Meth. 123 (1975) 615.
[33] S. Neumann and H. Ewald, Z. Phys. 169 (1962) 224.
[34] G. Fioni et al., Nucl. Instr. Meth. A322 (1993) 175.
[35] J.P. Bocquet, R. Brissot and H. Faust, Nucl. Instr. Meth. A267 (1988) 466.
[36] M. Hesse et al., Workshop on High Resolution Spectroscopy of Fission
Fragments, Neutrons and rays, ed. by H. Maerten et al., report FZR-93-08
(1993) p. 31.
[37] H. Faust et al., ILL Report 81FA45S (1981).
[38] J.P. Biersack and L.G. Haggmark, Nucl. Instr. Meth. 174 (1980) 257.
[39] L.C. Northclie, Ann. Rev. Nucl. Sci. 13 (1963) 67.
[40] H.H. Heckman, E.I. Hubbard and W.G. Simon, Phys. Rev. 129 (1963) 1240.
[41] H.D. Betz et al., Phys. Lett. 22 (1966) 643.
[42] V.S. Nikolaev and I.S. Dmitriev, Phys. Lett. A28 (1968) 277.
[43] K.X. To and R. Drouin, Nucl. Instr. Meth. 160 (1979) 461.
[44] Y. Baudinet-Robinet, Nucl. Instr. Meth. 190 (1981) 197.
[45] Y. Baudinet-Robinet, Phys. Rev. A26 (1982) 62.
[46] K. Shima, T. Ishikara and T. Mikumo, Nucl. Instr. Meth. 200 (1982) 605.
[47] K. Shima et al., At. Data and Nucl. Data Tables 51 (1992) 173.

19
[48] Marcus Wostheinrich, PhD thesis, Universitat Tubingen, 1999.
[49] CRC Handbook of Chemistry and Physics, ed. by D.R. Lide and
H.P.R. Frederikse, CRC Press, Boca Raton, 1997.
[50] F. Catucoli et al., Z. Phys. A298 (1980) 219.
[51] G. Audi et al., Nucl. Phys. A624 (1997) 1.
[52] S.K. Kataria, E. Nardi and S.G. Thompson, Proc. 3rd Symp. on Phys. and
Chem. of Fission, Rochester, 1973, IAEA-SM-174/63, p. 389.
[53] N. Feather, Proc. R.S.E. (A) 71 (1974) 323.
[54] P. Schillebeeckx et al., Nucl. Phys. A580 (1994) 15.
[55] P. Schillebeeckx et al., Nucl. Phys. A545 (1992) 623.
[56] J.H. Hamilton et al., Prog. Part. Nucl. Phys. 35 (1995) 635.
[57] Y.X. Dardenne et al., Phys. Rev. C54 (1996) 206.
[58] E. Allaert, C. Wagemans and G. Wegener-Penning, Nucl. Phys. A380 (1982)
61.

20
Table 1
Measured yields and energy parameters for hydrogen to carbon isotopes. \u.l." is
an upper limit for isotopes where no or too few events were found to determine a
reliable yield.
Isotope Emean (Emean ) EFWHM (EFWHM ) Yield (Yield) Ebin
(MeV) (MeV) (MeV) (MeV) (MeV)
2H 8.4 x 6.3 0.3 42 4 -
3H 8.4 x 6.1 0.3 786 42 -
4He 15.9 x 9.9 x 10000 700 11.1
6He 11.3 0.2 10.3 0.4 260 30 1
8He 8.7 0.2 9.5 0.4 15 1 -
7Li 14.7 0.5 9.9 1.1 6.7 0.6 3.7
8Li 15.7 1.1 11.8 1.9 4.2 0.6 -
9Li 12.6 0.6 12.4 0.9 8.3 0.8 -
11Li 11 x 11 x 0.0045 0.0015 -
7Be 20 x 13 x 0.2 u.l. 1
9Be 17.9 0.6 12.9 1.4 4.4 0.5 10.5
10Be 18 0.7 16 2.5 46 6 12.4
11Be 17.5 0.7 12 2 5.9 1.7 6.7
12Be 14.3 0.8 13.4 1.8 2.8 0.3 4.8
14Be 13 x 13 x 0.0027 0.0010 -
10B 22 x 18 x 0.03 u.l. 9.4
11B 21 x 17 x 1.6 1.0 15.7
12B 20 x 16 x 1.0 0.4 13.4
14B 19 x 15 x 0.13 0.04 8
15B 17.7 0.6 12.9 1.1 0.046 0.010 5.8
17B 16 x 13 x 0.001 u.l. -
14C 20.8 0.6 21.6 1.9 12.6 0.8 27.4
15C 18.6 1.3 23.1 2.7 4.3 0.4 22.5
16C 16.6 2.7 26 6 5.0 0.9 22.1
17C 18 x 20 x 0.64 0.15 16.7
18C 17.7 1.7 15.2 1.9 0.28 0.05 16
19C 17 x 17 x 0.0025 0.0014 10.1
20C 16 x 16 x 0.0036 0.0028 8.4
21
Table 2
Measured yields and energy parameters for nitrogen to silicon isotopes. \u.l." is
an upper limit for isotopes where no or too few events were found to determine a
reliable yield.
Isotope Emean (Emean ) EFWHM (EFWHM ) Yield (Yield) Ebin
(MeV) (MeV) (MeV) (MeV) (MeV)
15N 22 x 18 3 0.044 u.l. 29.3
16N 21 x 20 x 0.079 0.032 26.2
17N 20 x 20 x 0.34 0.10 27.3
18N 19 x 18.6 2.0 0.16 0.04 24.3
19N 18 x 18 x 0.26 0.08 24.7
20N 17.5 x 17.5 x 0.0029 0.0016 21
21N 17 x 17 x 0.0025 0.0014 20.6
15O 24 x 20 x 0.12 u.l. 23.6
19O 23 x 20 x 0.26 0.12 34.5
20O 22.1 0.9 18.5 2.1 1.10 0.12 37.5
21O 21 x 16.9 3.5 0.23 0.06 35.1
22O 20 x 16 x 0.12 0.04 37.4
24O 18 x 18 x 0.08 u.l. 33.2
19F 26 x 22 x 0.020 0.016 35.9
20F 25 x 23 x 0.002 u.l. 36.9
21F 23 x 19.2 7.5 0.021 0.006 40.3
22F 23 x 23 x 0.30 0.21 40
24F 21 x 22.3 3.3 0.030 0.016 40.9
24Ne 25 x 20 2.8 0.040 0.012 52.8
27Ne 24 x 22 x 0.008 u.l. 48.4
24Na 28 x 24 x 0.001 u.l.
27Na 26 x 24 x 0.005 u.l. 57.7
28Na 26 x 24 x 0.006 0.004 55.9
30Na 24 x 24 x 0.0008 u.l. 53.1
27Mg 30 x 26 x 0.0012 u.l.
28Mg 30 x 26 x 0.018 u.l. 67.4
30Mg 28 x 26 x 0.009 0.006 67.9
30Al 32 x 28 x 0.002 u.l.
34Si 36 x 30 22 x 0.001 u.l. 58.3
35Si 35 x 30 x 0.003 u.l. 78.4
36Si 33 x 30 x 0.004 u.l. 80.4