Nuclear Instruments and Methods in Physics Research 219 (1984) 569-574 569

North-Holland, Amsterdam

A MASS SPECTROMETER FOR FISSION FRAGMENTS BASED ON TIME-OF-FLIGHT

A N D ENERGY M E A S U R E M E N T S

A . O E D a n d P. G E L T E N B O R T
Physikaliscbes lnstitut der Universitiit Tiibingen, Auf der Morgenstelle D, D- 7400 Ttibingen, Fed. Rep. Germany

R. B R I S S O T , F. G O N N E N W E I N a n d P. P E R R I N
lnstitut Max oon Laue - Paul Langevin, 156 X, 38041 Grenoble Cedex, France

E. A K E R a n d D. E N G E L H A R D T
Institut fi~r Experimentelle Kernphysik, Universitiit Karlsruhe, Kaiserstrasse 12, D-7500 Karlsruhe, Fed. Rep. Germany

Received 24 February and in revised form 14 June 1983

With the aim of determining masses of fission fragments by measuring velocities and kinetic energies, specially adapted
time-of-flight detectors and ionization chambers were developed. The start and stop detectors for the time-of-flight sense the electrons
emitted from a thin foil upon passage of fragments. The time resolution achieved is = 100 ps. The ionization chambers have the
electric field arranged parallel to the particle trajectory, and with isobutane as the counting gas, the intrinsic energy resolution for
fragments with m -- 100 amu is typically 400 keV. The combination of both the time-of-flight and ionization chamber, to produce an
energy-time-of-flight spectrometer, was tested with fragments from the 235U(n, f) reaction on an external thermal neutron beam at the
ILL/Grenoble. In the light group of fission fragments, all masses are resolved individually with a mass resolving power m / S m = 170
for m = 95.

I. Introduction mass resolution due to the velocity spread introduced

by neutron boil-off. In fact, it has recently been experi-
Since the discovery of nuclear fission, the measure- mentally shown that if and only if one samples those
ment of fragment masses and energies has been of rather rare fragmentations where virtually no neutrons
major concern in fission research. Up to very recently are emitted, a mass resolution well below one amu is
the most reliable mass distributions came from radio- achievable [8]. The next step was to replace gaseous
chemistry [1]. Studies of the energetics of the fission ionization chambers by semiconductor detectors. The
process started with the ionization chamber [2,3]. The reasonable energy resolution, the large solid angles and
correlated energies of both fragments as obtained in the ease of handling made energy correlation measure-
back-to-back ionization chambers, at least yielded infor- ments on binary fission events with those detectors very
mation on the gross features of mass distributions and successful. Even the first applications of semiconduc-
mass-energy correlations, by taking into account mass tor-detectors were devoted to studied as diverse as the
and m o m e n t u m conservation to calculate masses. The search for fine structure in mass and energy distribu-
poor energy resolution of these early ionization cham- tions [9], mass and energy distributions in ternary fis-
bers and above all the uncertainties in their energy sion [10], charged particle induced fission [11] and
response to heavy ions, prompted the development of spontaneous fission of transuranium elements [12].
time-of-flight (TOF) techniques to investigate the veloc- Again, neutron evaporation and the limited resolving
ity distribution of fragments [4]. The introduction of power of the detectors prevented a one-by-one mass
time-zero detectors [5], being capable of delivering a resolution for fission fragments to be achieved. In order
start signal for fission fragments with almost negligible to overcome the problem with the mass resolution be-
distortion to the velocity or energy characteristics, was coming degraded by neutron boil-off, one has to per-
especially fruitful. Correlated velocity studies became form correlated velocity-energy measurements on single
possible by this technique [6,7]. Again applying mass fragments. The first experiment of this type made use of
and m o m e n t u m conservation, fragment masses prior to a time-zero detector for the start signal of a fragment
prompt neutron emission became accessible. However, traversing a flightpath of known length and derived
in this method there is an unavoidable limitation to the simultaneously the stop signal and an energy propor-

0167-5087/84/$03.00 © Elsevier Science Publishers B.V.

(North-Holland Physics P, lblishing Division)
570 A. Oed et al. . Mass spectrometer jor fisston fi'agments.

tional pulse from a surface barrier detector [13]. The fragment flight path and the secondary electrons ejected
mass resolution (mainly limited by the energy response from the foil upon passage of the heavy ions are accel-
of the detector), observed in this experiment was - 2 erated onto an electron multiplier. The multipliers are
amu for the light mass group and about - 3 amu for the built from two cascaded multichannel plates mounted
heavy mass group. In the mean time, this technique has coaxially to the fragment beam. A central hole in the
found a widespread application in heavy ion physics plates allows the fission fragments to pass through the
[14]. detector. For the present case the detectors had a cir-
To our knowledge, with the exception of the study cular entrance aperture of 10 mm in diameter. Without
cited above [8], the only successful attempts reported in any further amplification the electronic signals are fed
the literature where a mass resolution of less than 1 amu into fast discriminators and further into a time-to-pulse
for fission fragments has been obtained by physical height converter. It was shown that this detection sys-
methods, resorted to traditional mass spectrometry. Very tem is capable of a timing precision of 6t ~-100 ps
early in the history of fission research, mass spectrome- (fwhm) for the TOF device. For the chosen flightpath of
try was combined with radiochemical methods to look 107 cm, the accuracy of a velocity measurement for
into fine structure of mass distributions [15], or energy typical fission fragments approaches the 0.1% level. The
distributions of selected isotopes [16]. A fission frag- excellent timing performance is advantageous but not
ment recoil mass spectrometer of outstanding perfor- crucial, since by adjusting the length of the flight path
mance is the L O H E N G R I N facility installed at the one could accommodate for any velocity resolution re-
High Flux Reactor of the Institute Laue-Langevin in quired, obviously at the expense of counting rate. The
Grenoble [17]. In this instrument a magnetic field more important features of the above detectors are the
analyzes the momentum and an electrostatic condenser use of only a single foil perpendicular to the beam and
field the kinetic energy E of the particles. Mass and the absence of any additional grids to accelerate the
energy resolving powers of m / S r n = 103 and E / a E = electrons. In this way, energy losses and above all
103 are routinely achieved. A hybrid setup with the energy straggling of the heavy ions are kept to a mini-
velocities of fragments being measured by a TOF sys- mum. Energy losses in the start foil must be taken into
tem and the problem of energy resolution being solved account in calculating the true initial velocities of the
by the use of an electrostatic analyzer is H I A W A T H A fission fragments, whilst the combined energy losses in
[18]. A mass resolution better than 0.5 amu has been the start and stop foils enter as a correction to the
found with complete separation for the light fragments kinetic energy as seen by the energy detector. In all
and good separation for the heavy ones. cases, the absolute energy losses should be minimized in
The spectrometer reported here resembles mostly order to prevent the effects of their mass-energy depen-
that of Williams, Kiker and Schmitt [13] measuring dence becoming appreciable. Energy straggling in the
velocities and kinetic energies of fission fragments with start foil is sensed both by the TOF system and the
electronic detectors. Significant improvements in detec- energy detector and thus will not effect the mass resolv-
tor technology give this approach a new quality since it ing power of the spectrometer. However, energy strag-
is nowadays feasible to reach mass resolutions below 1 gling in the stop foil has to be considered more seriously
amu by this method. The spectrometer forms part of the since it effectively contributes to the resolution of the
COSI F AN TUTTE project at the Institute Laue-Lan- energy detector. Thus, the thinness and homogeneity of
gevin [19]. the electron emitting foils is of importance. The foils at
our disposal most suited to this purpose were AI203
foils (thickness 600 A), coated with AI (thickness 100 A)
2. Experimental setup
for the start and stop system.
The basic experimental setup comprises a start and The energy ~tetector stopping the fission fragments is
stop TOF system and an energy detector to measure in an ionization chamber. A more thorough description of
coincidence the velocities and kinetic energies of single this chamber has been published previously [21]. Basi-
fission fragments. Fragments emitted from a fission cally it is what may be called an 'axial' chamber with
source first pass a time-zero detector initiating a start the electric field lines arranged parallel to the particle
signal, travel down a flight path of given length, then hit trajectory. The entrance window separating the gas
a second time-zero detector to deliver a stop signal and volume from the high vacuum in the TOF drift tube is
finally enter an ionization chamber. The fission source mounted onto a supporting metallic grid and forms the
and the TOF measuring device are housed in a common cathode, while the anode and its Frisch grid shield are
vacuum. The start and stop detectors are of identical just outside the range of the most energetic fragments.
design and have been specially developed to meet the For a total length of the chamber of 5 cm and with
requirements,for the application at hand. They have isobutane as counting gas, stopping the fragments is
already been described in more detail elsewhere [20], ensured by a gas pressure of 160 mb. The gas pressure
Very briefly, a thin foil is placed perpendicular to the and the gas flow (30 standard cm3/min) are regulated.
 A. Oed et aL / Mass spectrometer for fission fragments. 571

A difficult problem with ionization chambers having signal of the ionization chamber being digitized and
external sources, is the entrance window. The best stored in a memory. Both, single spectra (for the TOF
window tested previously was made from a Parylene C and the energy) and correlated spectra (64 x 128 chan-
pellicle, 30 # g / c m 2 in thickness, supplied by the Union nels) for given windows in TOF and energy were taken
Carbide Corporation. With this window an energy width simultaneously.
6E of below 500 keY (fwhm) for light mass fission
fragments with energies around 100 MeV could be
reached [21]. The difficulty with windows made from 3. Experimental results
the above pellicles is that they are not sufficiently
gastight to maintain with reasonable pumping effort the The setup described was first tested with 252Cf
high vacuum (about 10 -6 mb) required for the safe spontaneous fission fragments. The results were encour-
operation of the channelplates of the stop detector just aging but suffered from poor statistics. A very similar
in front of the chamber. Therefore, entrance windows experiment has become known to us in the meantime
either made from propylene stretched uniformly under [231.
heat down to about 40 lag/cm 2 or made from Lexane The experiment we would like to report has been
has to be used. The Lexane pellicle [22] is composed of performed for fragments from the thermal neutron in-
three layers with a total thickness of about 20 ~tg/cm2. duced fission of uranium-235. To that purpose a uranium
The Lexane window proved to be far superior. On an target (thickness 100 # g / c m 2 or 350 # g / c m 2 235U, 10
energy measurement, the total energy straggling both in mm in diameter) was positioned in an external thermal
the Lexane entrance window of the chamber and the neutron beam (capture flux 5 × 109 n / c m 2 s) of the
stop foil of the TOF system limits the effective energy High Flux Reactor at the Institute Laue-Langevin. The
width to 8E = 500 keV (fwhm) for typical light fission TOF-energy spectrometer was mounted near to the
fragments and 8 E - - 5 5 0 keV for the heavy ones. The source and aligned to intercept those fragments being
window has a diameter of 15 mm and has to be sup- emitted perpendicular to the source surface. Count rates
ported. The electroformed Ni-mesh (thickness 5 ~m, of 2.5 and 8.5 fission events/s were observed for the
linewidth 25 ttm, spacing 340 /~m, transmission 87%) two targets under study.
used in the present investigation may not have been the The results obtained are summarized in the following
optimum choice since a non-negligible amount of figures. Fig. 1 shows the TOF spectrum for fission
scattered events showed up in the single spectrum for fragments from the thinner 235 U(n th, f) target. The qual-
the energy distribution of fission fragments. ity of the data is demonstrated by the clearcut separa-
The electronics of the experiment were quite conven- tion of the TOF peaks for the mean light mass (to the
tional, with the output signal of the time-to-pulse height left) and mean heavy mass fragments (to the right). The
converter of the TOF device and the amplified charge peak to valley ratio P / F for the light mass peak to the

235 2
time 3~8 h U lO0/ug/cm

- 1000
¢-

o
0

I
0 500 1000 1500 ch.n.
Time of flight 53ps/ch.
Fig. 1. Time-of-flight spectrum of fission fragments from 235U(n, f) for thermal neutrons. Target thickness 100 #g/cm 2.
 572 A. Oed et al. / Mass spectrometer fl)r fission fragments.

235U 100 ~ g / c m 2 time 3 7 8 h

lOgO

,,-- coinc.window
ch.n. 640-768

I
~J I
0 2 O0 4 O0 600 800 ch.n
Energy 140keV/ch
Fig. 2. Energy spectrum of fission fragments from 235U(n, f) for thermal neutrons. Target thickness 100 p.g/cm 2.

start stop valley in between the peaks is P / V = 365 _+ 30. The

235U th ~] ~] ] ionization energy spectrum is given in fig. 2. Again, the light mass
peak to the right is well separated from the heavy mass
peak to the left, the peak to valley ratio being P ~ V = 17
+ 1. As may be seen on the figure some low energy
tailing is present which is presumably due to fragments
being scattered by the wires of the supporting grid. In a
ch.n previous experiment where a different grid was used this
60' effect did not appear [20].We do not aim at evaluating
mean values for the velocities and energies as obtained
here, since this is the purpose of a more specialized
experiment utilising a thinner fissile source. The main
x~ objective of the present study was to look into the
R feasibility of resolving single fragment masses by corre-
4o lating velocity and kinetic energy data. The two-dimen-
sional matrix data of T O F and energy are depicted as a
contour plot for the 350 / s g / c m 2 target in fig. 3. The
"6 data shown are for the central part of the light mass
group near to the peak yields of velocity and energy
20 within the channel ranges indicated by vertical hatched
lines in figs. 1 and 2. It should be stressed that these are
raw data without any corrections. The individual fission
fragment masses should appear in the figure as diagonal
lines running from the top left (long T O F or equivalent
low velocity and low energy) to the bottom right (short
640+0 50 100 ch n T O F and high energy) with the mass number increasing
Energy 140 keV/ch. from bottom left to top fight. Indeed, the data gather
around those lines which shows that individual masses
Fig. 3. Contour plot of correlated time-of-flight and energy
data from 235U(n, f) for thermal neutrons. Target thickness 350 have been resolved. This becomes perhaps even more
/xg/cm2. The plot is for window settings on the time-of-flight evident from fig. 4 where a horizontal cut through the
and the energy as indicated by hatched lines in figs. 1 and 2. plot of fig. 3 is given. The figure represents an energy
The data correspond to the light mass group. The insert shows spectrum of fission fragments gated by a time window
schematically the setup. width of - 53 ps of the T O F spectrum. The peaks in the
 A. Oed et aL / Mass spectrometerfor fission fragments. 573

235
U 350 AJg/cm2
Time of flight w i n d o w ch .n. 6/~9
time 44.6 h

j
50

0
(..)

I I i
640,, 0 50 100 ch.n.
Energy 140 keV/ch.

Fig. 4. Energy spectrum of fission fragments from 235U(n, f) for thermal neutrons with a time window of 53 ps set on the
time-of-flight. The peaks in the spectrum correspond to individual masses in the light mass wing of fragments.

spectrum correspond to single masses covering the mass 4. Conclusion

range from about 85 through 103 amu. Unfortunately,
at the moment the energy calibration of the ionization It has been shown that a fission fragment mass
chamber is not accurate enough to assign mass numbers spectrometer built from a T O F system and an ioniza-
uniquely. Nevertheless, the mass resolution 8 m / m can tion chamber for the measurement of the correlated
be judged from the peak to valley ratios in fig. 4 to be velocities and kinetic energies of fragments, is capable
8 m / m = 0.6% for a mean mass m = 95 amu. This figure • of resolving individual masses in the light mass group. A
should be compared to the mass resolution as calculated mass resolution 8 m / m = 0.6% could be reached in the
from the known resolution 8 E / E of the chamber and case of thermal neutron induced fission of 235U. The
the resolution 8t/t of the T O F device. With m = 2 E / v 2 masses seen by the spectrometer are those after prompt
and v = L / t the mass resolution becomes neutron evaporation but before delayed neutron emis-
sion. The advantage of this method over the double
8 m / m = [(BE~E) 2 + ( 2 ~ t / / ) 2] 1/2
velocity technique [8] is that it is not limited to high
With the energy and time width, 8 E and St, respectively kinetic energy events. As compared to spectrometers
given above, a mean kinetic energy E = 100 MeV, a using magnetic a n d / o r electric fields [17,18] the mass
mean velocity v = 1.42 c m / n s (typical fragment of the resolution is certainly worse. However, the spectrometer
light mass group of the fissioning nucleus under study) reported here is much more compact and does not rely
and a flightpath L = 107 cm the experimentally achieved on an inpile target facility as is necessary for
mass resolution comes very close to that expected. For L O H E N G R I N or H I A W A T H A . For some purposes the
the present setup, the mass resolution is mainly limited fact that in the latter facilities the ionic charge of the
by the response of the energy detector. The worse ions enters as an additional parameter and that data
energy resolution 8 E / E for the heavy mass fragments have to be taken pointwise is a disadvantage. With the
due to 8E being increased and E decreased makes a spectrometer proposed here all fission events are treated
one-by-one mass resolution difficult for these fragments equally and therefore fine structures or mass-energy
by the present technique. Masses of approximately m = correlations may be directly read from plots similar to
130 amu with an energy of E = 80 MeV could just be the one in fig. 3. Conversely, it is feasible to install
resolved, the peak-to-valley ratio in a plot similar to fig. additional detectors on e.g. L O H E N G R I N to identify
4 barely exceeding 1.1. nuclear charges, i.e. the elemental composition of given
574 A. Oed et aL / Mass spectrometer.[or fission jragments.

fission fragment masses [24]. For a mass separated [10] H.W. Schmitt, J.H. Neiler, F.J. Walter and A. Chetham-
b e a m this i n f o r m a t i o n can also be obtained with an Strode, Phys. Rev. Lett. 9 (1962) 427.
axial ionization c h a m b e r in parallel to the energy [25]. [11] H.C. Britt, H.E. Wegner and J.C. Gursky, Phys. Re','. 129
Before i m p l e m e n t i n g this technique to the spectrometer (1963) 2239.
[12] R. Brandt, S.G. Thompson, R.C. Gatti and 1.. Philips,
discussed here, the mass resolution has p r o b a b l y still to
Phys. Rev. 131 (1963) 2617.
be improved. As has been stressed the main limitation is
[13] C.W. Williams, W.E. Kiker and H.W. Schmitt, Re,,'. Sci.
the energy straggling in the stop foil a n d the entrance Instr. 35 (1964) 1116.
window of the chamber. Work is in progress to test still [14] A. Gamp, W. Bohne, P. Braun-Munzinger and CK.
t h i n n e r pellicles on an improved ionization chamber. It Gelbke, Nucl. Instr. and Meth. 120 (1974) 281.
is envisaged that a n o t h e r 100 keV or even more is to be [15] H.G. Thode and R.L. Graham, Can. J. of Res. A 25 (1947)
gained in energy width BE. This would m e a n that in a 1.
plot like the one of fig. 4 a peak-to-valley ratio of 10 to [16] B.L. Cohen, A.F. Cohen and C.D. Coley, Phys. Rev. 104
1 could be reached for adjacent masses in the light (1956) 1046.
f r a g m e n t group. It is therefore felt that the potential [17] E. Moll, H. Schrader, G. Siegert, M. Asghar, J.P. Bocquet.
G. Baillard, J.P. Guntheron, J. Greif, G.I. Crawford, C.
capabilities of the present spectrometer for studies of
Chauvin, H. Ewald, H. Wollnik, P. Armbruster, G. Fiebig,
the fission process are very appealing.
H. Lawin and K. Sistemich, Nucl. Instr. and Meth. 123
(1975) 615.
T h a n k s are due to the B u n d e s m i n i s t e r i u m ftir For- [18] G. Di loro and B.W. Wehring, Nucl. Instr. and Meth. 147
schung und Technologie, Bonn, for financial support. (1977) 487.
[19] A. Oed, P. Perrin, F. GOnnenwein, H. Nifenecker, P.
Geltenbort, G. Barreau and C. Ristori, Proc. of Int. Symp.
References on Nuclear Fission and Related Collective Phenomena
and Properties of Heavy Nuclei, Bad Honneff (Springer,
[1] E.A.C. Crough, At. Data and Nucl. Data Tables 19 (1977) Berlin, 1982).
419. [20] A. Oed, G. Barreau, F. GOnnenwein, P. Perrin, C. Ristori
[2] W. Jentschke and F. Prankl, Z.f. Physik 119 (1942) 696. and P. Geltenbort, Nucl. Instr. and Meth. 179 (1981) 265.
[3] D.C. Brunton and G.C. Hanna, Can. J. of Res. A 28 [21] A. Oed, P. Geltenbort, F. GOnnenwein, T. Manning and
(1950) 190. D. Souque, Nucl. Instr. and Meth. 205 (1983) 455.
[4] R.B. Leachman, Phys. Rev. 87 (1952) 444. [22] H. Huizenga, J.A.M. Bleeker, W.H. Diemer and A.P.
[5] W.E. Stein and R.B. Leachman, Rev. Sci. Instr. 27 (1956) Huben, Rev. Sci. Instr. 52 (1981) 673.
1049. [23] W. Schuwerk, U. Quade and K. Rudolph, Jahresbericht
[6] W.E. Stein, Phys. Rev. 108 (1957) 94. (1981) Beschleuniger Laboratorium Universitiit and T.U.
[7] J.C.D. Milton and J.S. Fraser, Phys. Rev. 111 (1958) 877. Mtinchen, p. 26.
[81 C. Guet, M. Asghar, P. Perrin and C. Signarbieux, Nucl. [24] W. Lang, H.G. Clerc, H. Wohlfarth, H. Schrader and K.H.
Instr. and Meth. 150 (1978) 189. Schmidt, Nucl. Phys. A345 (1980) 34.
[9] W.M. Gibson, T.D. Thomas and G.L. Miller, Phys. Rev. [25] A. Oed, P. Geltenbort and M. GiSnnenwein, Nucl. Instr.
Lett. 7 (1961) 65. and Meth. 205 (1983) 451.