Journal of Environmental Radioactivity 149 (2015) 158e163

Contents lists available at ScienceDirect

Journal of Environmental Radioactivity

journal homepage: www.elsevier.com/locate/jenvrad

Quantifying radionuclide signatures from a geg coincidence system

Richard Britton*, Mark J. Jackson, Ashley V. Davies
AWE, Aldermaston, Reading, Berkshire, RG7 4PR, UK

a r t i c l e i n f o a b s t r a c t

Article history: A method for quantifying gamma coincidence signatures has been developed, and tested in conjunction
Received 13 May 2015 with a high-efficiency multi-detector system to quickly identify trace amounts of radioactive material.
Accepted 23 July 2015 The geg system utilises fully digital electronics and list-mode acquisition to timeestamp each event,
Available online 4 August 2015
allowing coincidence matrices to be easily produced alongside typical ‘singles’ spectra. To quantify the
coincidence signatures a software package has been developed to calculate efficiency and cascade
Keywords:
summing corrected branching ratios. This utilises ENSDF records as an input, and can be fully automated,
Gamma-spectrometry
allowing the user to quickly and easily create/update a coincidence library that contains all possible g and
Coincidence measurements
Coincidence library
conversion electron cascades, associated cascade emission probabilities, and true-coincidence summing
True-coincidence summing corrected g cascade detection probabilities. It is also fully searchable by energy, nuclide, coincidence pair,
CTBTO g multiplicity, cascade probability and half-life of the cascade. The probabilities calculated were tested
using measurements performed on the geg system, and found to provide accurate results for the nu-
clides investigated. Given the flexibility of the method, (it only relies on evaluated nuclear data, and
accurate efficiency characterisations), the software can now be utilised for a variety of systems, quickly
and easily calculating coincidence signature probabilities.
Crown Copyright © 2015 Published by Elsevier Ltd. All rights reserved.

1. Introduction has included Compton suppression systems (Britton et al., 2013,

2014), cosmic veto systems (Burnett and Davies, 2012, 2013),
The Comprehensive Nuclear-Test-Ban Treaty (CTBT) bans nu- and a multi-detector geg coincidence system (Britton et al.,
clear explosions for military and civilian purposes (United Nations, 2015a).
1996), and treaty compliance (once ratified) will be monitored For multi-detector systems, detection of coincident radiation
using a network of detection stations (the International Monitoring can be used to extract rare events, or to remove contributions from
System e IMS). Up to 80 stations will perform g spectrometry on air radiation that interacts with multiple detectors (such as g cascades
filter samples, and monitor for 83 radionuclides that are indicative and scattered radiation). Quantification of such measurements,
of (and can be used to discriminate between) nuclear weapons tests however, is extremely challenging. Firstly, the probability of the
and other reactor based incidents. Key to the monitoring system is target nuclide emitting a particular combination of emissions must
the establishment of certified and accredited laboratories that can be calculated, and then the efficiency of detecting these in a multi-
re-measure IMS samples, confirming and quantifying any result detector system must also be known.
produced by the automated stations. This report presents a methodology for calculating both the
Currently, such laboratories utilise a large HPGe crystal, sur- probability of a specific coincidence signature occurring, and the
rounded by an extensive, multi-layered shield to reduce back- subsequent probability of detecting the g emissions in multiple
ground radiation and therefore improve the Minimum Detectable detectors. All values are then combined with a ROOT (Brun and
Activity (MDA). Recent efforts at GBL15 (the UK IMS laboratory), Rademakers, 1997) based analysis package to directly quantify the
have focussed on highly efficient coincidence systems, whereby nuclide from the coincidence signature in question.
simultaneous detection of multiple decay signatures can be uti-
lised to dramatically increase the sensitivity of the system. This
2. Experimental setup

* Corresponding author. The geg coincidence system utilised for this work consists of
E-mail address: [email protected] (R. Britton). two large area, planar HPGe detectors (model BEGe-6530, from

http://dx.doi.org/10.1016/j.jenvrad.2015.07.025
0265-931X/Crown Copyright © 2015 Published by Elsevier Ltd. All rights reserved.
 R. Britton et al. / Journal of Environmental Radioactivity 149 (2015) 158e163 159

Canberra UK in Harwell, Oxfordshire) situated in a face-to-face events. As they are in a matrix, time and energy gates can also be set
configuration (both surrounding a central source), and enclosed to extract information from the dataset.
within a 100 mm Pb cave to minimise terrestrial radiation. Each
detector signal is passed through a preamplifier (model 2002), and 4. Quantifying radionuclides
to two separate acquisition systems for redundancy and testing
purposes. The primary signal is passed to separate channels on a To quantify a radionuclide signature, several factors must be
DT5724D Desktop Digitiser (CAEN S.p.A., Italy), which can accom- used to correct the measured source activity. These are summarised
modate up to 4 inputs. CAEN supplied MC2A acquisition software is in Equation (1), below.
used for both set up and list-mode acquisition. Digitisation of the
pulse, amplification and shaping is all performed using this module Npeak
Activity ¼ (1)
with optimised parameters. As the unit is self-contained, each bεct
channel uses a common clock, and so no further time synchroni-
Here, Npeak is the number of measured counts in the peak, b
sation is necessary. The second signal is passed into two LYNX™
represents the branching ratio (the proportion of decays that emit
units from Canberra, with the internal clocks synchronised via an
the g radiation of interest), ε is the efficiency of the detector to g
external cable, and all data collected in list-mode. A schematic of
radiation of this energy, and c is the correction factor to account for
the detector system is shown below, in Fig. 1.
true-coincidence summing (whereby the signal of interest can be
lost/augmented by multiple g emissions interacting in the same
3. Measurement & analysis detector).
For coincidence systems, the same formula applies, however the
Calibration sources were chosen to cover the 50e2000 keV correction factors become far more complex to calculate. In a
energy range to fully characterise each system. Several single g cascade of three g emissions (g1, g2, g3), where two are of interest
emitters were used, as well as a NIST (National Institute of Stan- (g1, g2), the coincidence signature can be extracted by creating a
dards and Technology) traceable complex g source. The isotopes coincidence matrix, and projecting all events that are in coinci-
contained within these sources included 241Am (59.54 keV), 109Cd dence with g1 in each detector. Analysing the resulting spectra for
(88.03 keV), 57Co (122.06 keV), 139Ce (165.86 keV), 113Sn g2 will yield Npeak. The branching ratio (b) will need to include the
(391.68 keV), 137Cs (661.67 keV), 54Mn (834.84 keV), 88Y (898.04 total probability of seeing both g1 and g2 at the same time, which
and 1836.06 keV), 65Zn (1115.54 keV), and 60Co (1173.23 and may be present in multiple cascades. The efficiency will need to
1332.49 keV). account for the fact that each g can be detected in multiple de-
Analysis was completed using a ROOT & cþþ based post- tectors, and the true-coincidence correction factor (c) will need to
processor for sorting of the data, matrix manipulation and peak account for the probability of detecting g1 or g2 in the same de-
searching/fitting. Coincidences were identified by searching for all tector as g3.
events in the secondary detector with a time delay (t)
between 4 ms and þ4 ms (with t ¼ 0 defined as the interaction time
4.1. Calculating a decay cascade ‘branching ratio’
in the primary detector). This covers both the coincidence peak
(seen between 1.2 ms and 1.2 ms), and a region of ‘background’
Given that the number of energy levels in a nuclide can be
(accidental) coincidences outside this range, which is used for the
extremely large, the resulting number of g cascades may quickly
coincidence background subtraction detailed in section 5. These
become unmanageable when trying to calculate all possible paths
coincidences were then recorded into a 3D matrix with the energy
via which an isotope may decay. A MonteeCarlo method of calcu-
deposited in the primary detector, the energy deposited in the
lating the decay cascades was therefore selected, and to aid auto-
secondary detector, and the time delay on each axis (Eg1, Eg2, Dt).
mation, ENSDF (Evaluated Nuclear Structure Data Files) (Bhat, 1992)
This allows a delay spectrum to be created, with peaks identifying
selected as the input.
characteristic delays where there are a large number of coincident
The resulting package is known as the ‘Randomised Iterative
MonteeCarlo Model for ENSDF Records’ (RIMMER), and is split into
three main sections; the first loads each ENSDF record and creates
the decay scheme in the programs memory. The strict formatting
requirements of ENSDF records makes them ideal for such an input,
and they contain all the information required to calculate the
probabilities for each g cascade, including decay branches, energy
levels and g transitions. The second part runs the MonteeCarlo
model, which (based upon the level feeding distribution specified
in the ENSDF records) randomly feeds a level. The proportion of
level feeding is calculated from the Normalisation record and the
relevant Beta/Electron Capture (EC) records for each level. For a
typical level, the proportion of 100 decays that is populated via beta
decay (Pb) is:

Pb ¼ BR:NB:IB (2)
Where BR is the branching ratio multiplier for converting in-
tensity per 100 decays through this decay branch to intensity per
Fig. 1. A schematic of the geg system, with the detector source separation set to 5 cm. 100 decays of the parent radionuclide, NB is the multiplier for
The HPGe crystals are shown in green, surrounded with copper crystal holders and a converting relative b, bþ and EC intensities to intensities per 100
copper cold finger (both in red). The aluminium casings are shown in light grey, with
carbon fibre detector faces and the surrounding lead shield (outlined in dark grey). (For
decays through this decay branch, and IB is the intensity of the b
interpretation of the references to colour in this figure legend, the reader is referred to decay branch. Note that this is the same nomenclature used in the
the web version of this article.) ENSDF manual (Tuli, 1987), and the reader is referred to this for
160 R. Britton et al. / Journal of Environmental Radioactivity 149 (2015) 158e163

further information. For nuclei that decay via positron emission, the 4.3. Calculating the decay cascade ‘true-coincidence correction
511 keV annihilation radiation may be seen in coincidence with the factor’
g cascade. To account for this, an additional level is added 511 keV
above the level in which the positron decay leaves the daughter Depending on the multiplicity of the cascade, the correction
nucleus. Assigning a conversion coefficient of 0.5 to this pseudo factor must account for the possibility of detecting additional g
level allows the program to take into account both of the annihi- radiation in one or both of the detectors when detecting the (g1, g2)
lation photons (Gilmore, 2008). signal. In a typical detector, the probability of detecting multiple g
The code then decides (based upon the transition probabilities emissions is vanishingly low, however this is not true for a high-
to each level and a random number generator) which transition to efficiency system. For high-multiplicity cascades in particular, the
make. The transition intensities (TI) for each g are given in the probability of detecting one of the additional g emissions may be
Gamma records that are present for each Level record in the ENSDF substantial. Consider the (g1, g2, g3) cascade mentioned earlier. For
file. The probability for each transition can be simply calculated two detectors, the corrected efficiency (εcorr) for detecting the (g1,
from the relative fraction of the TI s for that level: g2) signal will need to be modified to account for the probability of
not detecting g3:
TIg
Pg ¼ P (3) h  i
TI εcorr ¼ εtotal 1  εD1 D2
g3 þ ε g3 (6)
Note that for some Gamma records in the ENSDF records the TI
The correction required can therefore be calculated solely from
values are not available, and instead the RI (relative intensities) and
the total detection efficiency for g3. A subtlety for this calculation is
CC (conversion coefficients) values are provided. In this case:
that the efficiency for g3 must account for the total efficiency of
  detection, not just the peak efficiency (any interaction that deposits
TIg ¼ RIg 1 þ CCg (4)
energy will be enough to destroy the coincidence signature). In a
The MonteeCarlo simulation then uses the CC values to more general form, (and substituting εg ¼ εD1 D2
g þ εg ), the corrected
randomly decide whether the transition will be accomplished via g efficiency can be calculated as follows for a cascade of N gs:
decay or internal conversion. The resulting g or conversion electron
for the chosen transition is recorded, and the program ‘moves’ to Y
i¼N 
the next level, selecting another transition. This process is repeated εcorr ¼ εtotal 1  ε gi (7)
i¼3
until the code reaches the ground state or an isomeric state, at
which point the program outputs the total half life of the cascade, This is subject to several assumptions; the first and foremost of
the final state of the isotope (ground or isomeric), the multiplicity which is the lack of any angular correlation between the emissions.
of the cascade, and a complete list of the gs and conversion elec- Due to the close geometry and large solid angle covered by this
trons emitted. By repeating the simulation multiple times, a list of system, however, the effect of any angular correlations will be small
all possible cascades are generated; these are then counted and (Warren et al., 2006). Summing of signals due to X-rays and con-
normalised to the total number of MonteeCarlo events to calculate version electrons is treated in a simplistic way; while the proba-
a cascade probability. Each cascade and associated information (the bility of a transition emitting a conversion electron (and therefore
probability of the cascade being emitted, multiplicity, half-life of reducing total probability of seeing the full g cascade) is taken into
the cascade, whether it ends in an isomeric state) is then recorded account, summing due to the resulting conversion electrons and X-
into a library. rays is not. If these effects began to cause issues, their contributions
The third and final part of the RIMMER package is a Python could be mitigated with the use of thin absorbers between the
script, which loads all decay paths and corresponding probabilities source and each detector. Backscattering of photons between the
for each nuclide that has been simulated. This can then search the detectors was also considered, however this effect will not
MonteeCarlo results by nuclide, g energy, g multiplicity, half-life of contribute to the summing out of potential signals in coincidence
the cascade, and probability of cascade, filtering results from either mode. By definition, a backscattered photon must interact in one of
the full library or from the previous search results. When searching the detectors before possibly being detected in the other. Any
results for a specific (g1, g2) coincidence signature, both individual backscattered photon will therefore not contribute to the peak ef-
cascades that include the signature and the total probability of ficiency used to calculate εtotal, nor cause an additional summing
seeing that signature are returned. Efficiencies (peak and total) for event from additional g emissions as the scattering of these is
each detector can be folded in, and used to produce a cascade already accounted for. Backscattering of photons from shielding
summing corrected coincidence signature branching ratio via the materials may cause some loss of coincidence signatures, however
method described in the next sections. this can be mitigated by maximising the distance between the
detectors and the shielding.
4.2. Calculating the decay cascade ‘detection efficiency’ As the efficiency correction depends on the number of simul-
taneous g emissions, it is computed individually for each cascade
For a simple case with one g emission, the peak efficiency of within which the (g1, g2) signal is present. The resulting factors can
detection is a combination of the solid angle subtended by the then be applied to the cascade probabilities, and summed to
detector and the intrinsic efficiency of the crystal over a range of calculate the efficiency and true-coincidence corrected branching
energies. For the case of two emissions, the efficiency for a single ratio for the (g1, g2) signal. All that is then needed is Npeak and the
detector measuring both is simply a product of the two peak effi- total acquisition time (t) to determine the activity of the isotope
ciencies. For the case of two detectors (D1 & D2), the efficiency of (see Equation (1)).
two emissions (g1 & g2) entering separate detectors (εtotal) can be
calculated as: 5. Verifying the correction factors

εtotal ¼ εD1 D2 D1 D2
g1 $εg2 þ εg2 $εg1 (5) To verify the technique described above, several measurements
were made of typical IMS sources. These were previously
 R. Britton et al. / Journal of Environmental Radioactivity 149 (2015) 158e163
Fig. 2. The raw, total coincidence, energy/time gated coincidence spectra, and coincidence background subtracted energy/time gated coincidence spectra for an IMS source dominated by 140La. The RIMMER package for 140La suggests
that the 1596.21 & 487.02 keV coincidence pair is the most abundant, and so the energy gate was placed around the 1596.21 keV transition in each detector, and the resulting coincidence spectra projected and summed to estimate
Npeak. The gated coincidence spectrum greatly reduces the contribution from additional decays, and clearly reveals the additional decay lines from 140La (labelled in black).

161
162 R. Britton et al. / Journal of Environmental Radioactivity 149 (2015) 158e163

Table 1
A summary of the activity concentrations calculated in both singles and coincidence mode, compared to the accepted reference activities (corrected for radioactive decay
between the previous and current measurement). The singles mode activities were calculated from the decay lines present in the spectrum, while the coincidence mode
activities were calculated with the coincidence pair specified. For two key radionuclides (133Ba and 88Y), duplicate entries are included to demonstrate the effect of neglecting
X-ray summing from the analysis; entries for these radionuclides without the X-ray summing corrections are denoted by an ‘*’. Errors ranged from 4% to 6%, and are quoted at 1
sigma.

Isotope Coincidence Pair (keV) Reference Activity (Bq.m3) Ratio to reference Value

Singles mode Coincidence mode

60
Co (1173.23, 1332.49) 1.968 ± 0.032  103 1.006 ± 0.036 0.994 ± 0.050
88
Y* (898.04, 1836.06) 2.950 ± 0.050  103 1.053 ± 0.055 0.908 ± 0.046
88
Y (898.04, 1836.06) 2.950 ± 0.050  103 1.053 ± 0.055 0.966 ± 0.048
103
Ru (53.29, 557.06) 3.660 ± 0.081  103 1.007 ± 0.049 0.994 ± 0.052
133
Ba* (81.00, 356.01) 2.953 ± 0.050  103 1.034 ± 0.052 0.749 ± 0.050
133
Ba (81.00, 356.01) 2.953 ± 0.050  103 1.034 ± 0.052 0.999 ± 0.050
133
Ba (160.61, 276.40) 2.953 ± 0.050  103 1.034 ± 0.052 0.989 ± 0.050
140
Ba (162.66, 304.85) 8.556 ± 0.179  103 0.978 ± 0.048 1.004 ± 0.050
140
Ba (29.97, 537.26) 8.556 ± 0.179  103 0.978 ± 0.048 0.993 ± 0.052
140
La (815.77, 1596.21) 9.658 ± 0.208  103 1.024 ± 0.053 1.038 ± 0.053
140
La (487.02, 1596.21) 9.658 ± 0.208  103 1.024 ± 0.053 0.994 ± 0.052

characterised with a validated and CTBTO (Comprehensive Nuclear- cascade intensity (as a result of electron emission) was included in
Test-Ban Treaty Organisation) accredited detector system. The ra- the calculations. An accurate activity estimate was obtained for the
dionuclides included isotopes such as 133Ba and 140La, which emit (276.4 keV, 160.6 keV) coincidence pair in 133Ba, however the far
multiple g emissions in coincidence. Both detectors in the geg more abundant (81.00 keV, 356.01 keV) signal initially biased ~25%
system are LabSOCS LYNX™ units characterised, allowing the Ge- low. Detailed analysis of the spectra revealed significant summing
ometry Composer module of the GENIE 2000 LYNX™ units soft- between the g emissions and 133Cs Ka/Kb X-rays; once these were
ware to produce accurate (uncorrected) peak and total efficiency included in both the energy gate and subsequent peak analysis, the
characterisations. activity estimate was found to be in excellent agreement with the
Fig. 2 shows an example coincidence spectrum, and the level of reference value.
88
background reduction achieved at each stage of preparation (note Y also exhibits similar problems, as large amounts of 88Sr X-
that during routine operation, all stages are completed in a single rays were seen in summation with the coincidence pair. By again
step, and automated). Firstly the delay gate is applied to select widening the energy gate to include the X-ray/g sum peak, and
coincident events from the (Eg1, Eg2, Dt) matrix, producing the ‘Total combining the peak areas from the ‘pure’ and X-ray summed g
Coincidence Spectrum’. An energy gate is then applied to refine the photopeaks in the projected spectra, the correct activity could be
spectrum further, producing the ‘Preliminary Gated Coincidence calculated. Positron emission (and subsequent 511 keV g emission
Spectrum’. This produces a very clean spectrum, however there are following b/bþ annihilation) was also included in the calculations
a small number of contaminants that are the result of random for 88Y, however it was not found to substantially affect the coin-
summing; these are real coincident events, however they are due to cidence signature.
chance coincidences rather than a g cascade. To minimise the effect These results validate not only the detector system, but the
of random summing, a secondary delay window can also be method for calculating coincidence correction factors that are
extracted that is not part of the prompt coincidence peak, and is crucial for accurate quantification of radionuclide signatures. The
therefore representative of the random coincidence background authors would like to note that while the calculation works with
(Britton et al., 2015b). This can then be subtracted from the ‘Pre- these radionuclides, it cannot yet be assumed that this is true for
liminary Gated Coincidence Spectrum’ to produce the ‘Final Gated untested isotopes due to current limitations of the model, and
Coincidence Spectrum’. Note that in the final gated coincidence further work is under way to test the system against a wide range of
spectra, the background is dramatically reduced, almost all con- CTBT relevant radionuclides.
taminants are removed or reduced, and the coincidence signals
remain unaffected. This will allow large MDA improvements for a 6. Conclusions
number of nuclides.
Using the RIMMER package, the most probable g cascades were A method for quantifying coincidence signatures has been
calculated for each radionuclide. By then incorporating the effi- developed, and tested in conjunction with a high-efficiency geg
ciency characterisations for the detector system, the most probable system to quickly identify trace amounts of radioactive material.
geg detections were also identified. These signatures were The geg system utilises fully digital electronics and list-mode
extracted from the coincidence matrix, and the calculated factors acquisition to timeestamp each event, allowing coincidence
applied to obtain activity estimates. Table 1 documents the activity matrices to be easily produced alongside typical ‘singles’ spectra.
concentrations calculated using the geg system in both ‘singles’ This enables fast and reliable identification of nuclides, however to
mode (where no coincidence information is used) and full coinci- quantify these signatures a software package has also been devel-
dence mode. oped to calculate efficiency and cascade summing corrected (g1, g2)
Excellent agreement with the reference values are seen for both coincidence signature probabilities (branching ratios). This utilises
operational modes of the geg system. Certain nuclides, however, ENSDF records as an input, and can be fully automated, allowing the
were notably more difficult to analyse than others. Specifically, user to quickly and easily create/update a coincidence library that
nuclei that decay via transitions with large conversion coefficient contains all possible g and conversion electron cascades. Cascade
(CC) values such as 133Ba can be problematic due to the summation emission probabilities are automatically calculated and included
of conversion electrons/X-rays and g radiation. While this effect within the library, and the software can also fold in efficiency
was not explicitly accounted for, the corresponding reduction in g characterisations to calculate true-coincidence summing corrected
 R. Britton et al. / Journal of Environmental Radioactivity 149 (2015) 158e163 163

g cascade detection probabilities. It is also fully searchable by en- identification of g coincidence signals with required multiplicities
ergy, nuclide, coincidence pair, g multiplicity, cascade probability and branching ratios, the identification of the optimal coincidence
and the half-life of the cascade. signatures to measure for a particular system, and the calculation of
Routines have been developed for the automatic sorting and cascade summing corrections for single detector systems.
processing of list-mode coincidence data, allowing the user to
easily project energy and time gated spectra. Automatic reduction
of the coincidence background from ‘random’ coincidences is also References
incorporated. The probabilities calculated were tested using mea-
Bhat, M., 1992. Evaluated nuclear structure data file (ENSDF). Nucl. Data Sci. Tech.
surements performed on the geg system, and found to provide 817e821.
accurate results for all nuclides investigated. Currently the software Britton, R., Burnett, J., Davies, A., Regan, P.H., 2013. Nucl. Instrum. Methods Phys.
does not account for angular correlations between emissions, Res. A 729, 64e68.
Britton, R., Burnett, J., Davies, A., Regan, P.H., 2014. Nucl. Instrum. Meth Phys. Res. A
however the effect in this particular system is minimised due to the 762, 42e53.
large geometrical efficiency achieved. Summing between g emis- Britton, R., Burnett, J., Davies, A., Jackson, M., 2015. J. Environ. Radioact. 146, 1e5.
sions and X-rays is not currently accounted for in the detection Britton, R., Burnett, J., Davies, A., Regan, P.H., 2015. Nucl. Instrum. Methods Phys.
Res. A 769, 20e25.
probabilities, however these events can be easily identified from Brun, R., Rademakers, F., 1997. Nucl. Instrum. Methods Phys. Res. A 389, 81e86.
the coincidence signatures and used to obtain reliable, accurate Burnett, J., Davies, A., 2012. J. Radioanal. Nucl. Chem. 292e3, 1007e1010.
estimates of activity concentrations. Work is also under way to Burnett, J., Davies, A., 2013. J. Radioanal. Nucl. Chem. 298e2, 987e992.
Gilmore, G.R., 2008. Practical Gammaeray Spectroscopy, second ed. John Wiley &
incorporate X-ray summing explicitly in the calculations.
Sons, Chichester, England.
Given the flexibility of the method described in this work, (it Tuli, J.K., 1987. Evaluated Nuclear Structure Data File, a Manual for Preparation of
only relies on evaluated nuclear data, and accurate efficiency Data Sets. BNL-NCS-51655-Rev. 87, NNDC.
characterisations), the software can now be utilised for a variety of United Nations, 1996. The Comprehensive Nuclearetesteban Treaty, Resolution
CTBT/MSS/RES/1.
systems, quickly and easily calculating coincidence signature Warren, G., Smith, E., Aalseth, C., Ellis, E., Hossbach, T., Valsan, A., 2006. Nucl. Ins-
probabilities. Additional uses for the software include the fast trum. Methods Phys. Res. A 560, 360e365.