AGATA: Gamma-ray tracking in segmented HPGe detectors

Abstract

The next generation of radioactive ion beam facilities, which will give experimental access to many exotic nuclei, are presently being developed. At the same time the next generation of high resolution gamma-ray spectrometers, based on gamma-ray tracking, for studying the structure of these exotic nuclei are being developed. One of the main differences in tracking of $\gamma$ rays versus charged particles is that the gamma rays do not deposit their energy "continuously" in the detector, but in a few discrete steps. Also, in the field of nuclear spectroscopy, the location of the source is mostly well known while the exact interaction position in the detector is the unknown quantity. This makes the challenges of gamma-ray tracking in germanium somewhat different compared to vertexing in silicon detectors. In these proceedings we present the methods for determining the 3D interaction positions in the detector and how these are used to reconstruct the gamma-ray tracks in the AGATA detector array. We also present preliminary simulation results of a proposed in-beam method to measure the interaction position resolution in the germanium detectors. Comment: 10 pages, 6 figures. To appear in the proceedings of 17th International Workshop on Vertex detectors July 28 - 1 August 2008 Uto Island, Sweden

Full text

Page 1

AGATA: Gamma-ray tracking in segmented HPGe
detectors
P.-A. Söderström∗, A. Al-Adili, J. Nyberg
Department of Physics and Astronomy, Uppsala University, SE-75121 Uppsala, Sweden
E-mail: P-A.Soderstrom@physics.uu.se
F. Recchia
INFN Legnaro, 35020 Legnaro (Padova), Italy
E. Farnea
INFN Padova, 35122 Padova, Italy
A. Gadea
IFIC CSIC, Valencia, Spain
The next generation of radioactive ion beam facilities, which will give experimental access to
many exotic nuclei, are presently being developed. At the same time the next generation of high
resolution γ-ray spectrometers, based on γ-ray tracking, for studying the structure of these exotic
nuclei are being developed. One of the main differences in tracking of γ rays versus charged
particles is that the γ rays do not deposit their energy “continuously” in the detector, but in a
few discrete steps. Also, in the field of nuclear spectroscopy, the location of the source is mostly
well known while the exact interaction position in the detector is the unknown quantity. This
makesthechallengesofγ-raytrackingingermaniumsomewhatdifferentcomparedtovertexingin
silicon detectors. In these proceedings we present the methods for determining the 3D interaction
positions in the detector and how these are used to reconstruct the γ-ray tracks in the AGATA
detector array. We also present preliminary simulation results of a proposed in-beam method to
measure the interaction position resolution in the germanium detectors.
17th International Workshop on Vertex detectors
July 28 - 1 August 2008
Utö Island, Sweden
∗Speaker.
c ? Copyright owned by the author(s) under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike Licence.
http://pos.sissa.it/
arXiv:0812.0091v1 [nucl-ex] 29 Nov 2008

Page 2

AGATA: γ-ray tracking in segmented HPGe detectorsP.-A. Söderström
1. Introduction
It is sometimes said that we now are on the brink of the fourth revolution within nuclear spec-
troscopy [1]. The first revolution was the discovery of NaI(Tl) scintillator detectors, by which
one could start to measure quite accurately the energy and intensity of γ-ray transitions in various
radioactive nuclei. These measurements later radically improved when one started using semicon-
ductor detectors of germanium instead of NaI(Tl) crystals. The second revolution came with the
possibilities to use in-beam γ-ray spectroscopy together with heavy-ion nuclear reactions. This
made it possible to study specific levels of the nuclei of interest. The latest revolution is the de-
velopment of sophisticated high-resolution spectroscopy arrays of high-purity germanium (HPGe),
crowned by EUROBALL [2] and GAMMASPHERE [3]. These gave access to very weak signals
fromhigh-spinstates, unravelingmanynewnuclearstructurephenomena. Andnow, justaroundthe
corner, awaits the second generation facilities for radioactive ion beams (RIB) like NuSTAR/FAIR
[4, 5] and SIPRAL 2 [6] as the fourth revolution. These facilities will make it possible to study
very short lived exotic nuclei with extreme values of isospin, located in the terra incognita far from
the line of β stability. Indeed there will be very interesting times for nuclear structure physics in
the years to come.
These new exotic nuclei will be produced at the RIB facilities with very low cross sections and
in a high γ-ray background environment, which makes the weak γ-ray transitions extremely diffi-
cult to detect with existing spectrometers. In order to distinguish these rare events, new instruments
with higher efficiency and resolving power are required. To meet the new requirements 45 institutes
in 12 European countries (as of fall 2008) are collaborating to build the Advanced Gamma Track-
ing Array (AGATA) [7, 8]. A similar device, the Gamma Ray Energy Tracking Array (GRETA), is
also being built in the USA [9].
2. AGATA
The AGATA γ-ray spectrometer will be built in stages and moved between different host labo-
ratories in order to maximally exploit the strengths of the different RIB facilities as they come into
operation. A first version of AGATA, the AGATA Demonstrator consisting of 1/12 of the full array,
will begin operation at Laboratori Nazionali di Legnaro (LNL) in Italy in the beginning of 2009.
The final version of AGATA will consist of a HPGe shell with an inner radius of 23 cm, made
of 180 crystals assembled into 60 triple cluster detectors. The total amount of HPGe in AGATA
will be 362 kg. Each HPGe crystal has a hexaconical shape and is of the closed-end coaxial type,
with the electric field along the radial direction. The crystals are electrically segmented into 36
individual segments, six azimuthal and six axial, plus the central contact (core). The total HPGe
solid angle coverage will be 82 % of 4π. The full-energy peak efficiency of the 4π array will be
∼ 45 % for γ rays of multiplicity Mγ= 1 and energy 1 MeV, a factor of about four improvement
over existing arrays, and ∼25 % for Mγ=30, which will give several orders of magnitude increase
of the resolving power compared to existing arrays [7].
This gain in detection power is due to novel techniques to reconstruct the tracks of the γ rays
that scatter within the array. In order to do this, a three dimensional position resolution of ? 5 mm
of the interaction points (IPs) within the crystal is required. Such a position resolution corresponds
2

Page 3

AGATA: γ-ray tracking in segmented HPGe detectorsP.-A. Söderström
Figure 1: Signals from the core, the segment with the primary hit (Seg 4), and from the mirror charges for
a γ-ray interaction in a six fold segmented HPGe detector. Figure from [7].
to an angular resolution of about 1 degree. This is achieved using digital electronics to sample the
pulse shapes and advanced pulse-processing techniques to locate the IPs.
2.1 Pulse-shape analysis
Inordertodeterminetheinteractionposition, pulse-shapeinformationfromthesegmentwhich
was hit, the mirror charges in the neighboring segments, and the core signal is used, as illustrated in
fig. 1. The method to use pulse-shape analysis (PSA) to determine interaction position is, however,
not unique for AGATA and GRETA. This is for example done also in MINIBALL, where the
azimuthal position is obtained by comparing the amplitudes of the mirror charges [10]. The radial
position is obtained from the shape of the signal of the hit segment. This is good enough to correct
for Doppler effects, but is does not give enough resolution to do γ-ray tracking. The PSA algorithm
that will be implemented in AGATA is the grid search algorithm [11], where the sampled pulse
shapes are compared to a database of pulse shapes for different interaction positions.
One of the main problem of PSA algorithms is to obtain the database of pulse shapes. When
delivered, the crystals are scanned using radioactive sources [12]. This process gives a very precise
database of pulse shapes but it is unfortunately a very slow process. To complement the scanned
pulse shapes, codes for calculating them are being developed [13]. These calculations are much
faster, but the main challenge is to obtain realistic pulse shapes, while dealing with large uncertain-
ties in the impurity concentration and in the modeling of charge mobility [14].
3. Tracking of γ rays
Once the interaction points and the corresponding energies are identified, the different γ rays
should be disentangled from the “world map”, see fig. 2. Since the primary interaction mechanism
at these energies is a series of Compton scatterings followed by a photoelectric effect, the tracking
3

Page 4

AGATA: γ-ray tracking in segmented HPGe detectorsP.-A. Söderström
Figure 2: Simulated interaction points of 30 γ rays of energy Eγ= 1.33 MeV in the (θ,φ sinθ) plane of an
ideal germanium shell with an inner radius of 15 cm and an outer radius of 24 cm. Circles are correctly and
squares incorrectly identified clusters. Figure from [7].
algorithm is based on the Compton scattering formula
E?
γ=
Eγ
1+
Eγ
mec2(1+cosθ)
.
(3.1)
In order to detect higher and lower energy γ-rays efficiently the algorithm should also take into
account single photoelectric absorption as well as electron-positron pair-production. It is also im-
portant that the algorithm can discriminate between γ rays and other types of interactions, like
neutrons [15, 16]. Two main tracking algorithms to be implemented in parallel has been developed
and compared carefully [17].
3.1 Clusterisation or Forward tracking
One of the algorithms to be used is the clusterisation algorithm, developed by the group at
INFN Padova [18]. In this algorithm the interactions caused by a certain γ ray are assumed to
be well localized, which means that the IPs from the same γ ray should make up clusters in the
(θ,φ sinθ) plane, see fig. 2. After these clusters have been identified, the source location is as-
sumed as the zeroth IP. A first and a second IP is chosen and the γ-ray energy after scattering is
determined from the measured energy depositions in the cluster. This energy is compared to the
energy calculated from the Compton scattering formula in eq. (3.1) and a figure of merit based on
the agreement between the two energies is calculated and if necessary the clusters are rearranged.
All possible permutations are evaluated and the one with the best figure of merit is chosen. This
procedure is repeated with the identified first IP as a starting point until all IPs in a cluster have
been assigned to the track.
4

Page 5

AGATA: γ-ray tracking in segmented HPGe detectorsP.-A. Söderström
3.2 Backtracking
The backtracking algorithm was developed by the group at KTH, Stockholm [19, 20]. This
algorithm uses the information that the final IP most probable has an energy deposition between
100 to 250 keV. Starting from this assumed final IP, other IPs are searched for within a distance
based on the interaction length in germanium for γ rays of that energy. This procedure is repeated
until the track is terminated by the source location. It is then repeated until no more suitable IPs
are available.
4. In-beam measurements of position resolution
As mentioned in sec. 2, the position resolution within the crystal is of fundamental importance
in order to do γ-ray tracking. Thus this is an important parameter to measure experimentally, which
is planned to be done during the commissioning phase at LNL. One way to do this directly is to
use a collimated beam of γ rays in order to get a well defined first IP. But this requires a very
narrow beam in order not to be limited by the collimation. The consequence of this would be a
very time consuming measurement in order to get good statistics. Instead indirect measurements
based on imaging techniques [21] or in-beam measurements of Doppler correction capabilities are
being examined.
The contribution to the full width at half maximum (FWHM) of the full-energy peak in a γ-ray
spectrum (see fig. 3) can be divided into four parts
W2=W2
i+W2
vr+W2
θr+W2
θγ,
(4.1)
whereWiis the intrinsic resolution of the detector,WvrandWθrare the contributions from the uncer-
tainties in the velocity and angle of the γ-ray emitting nucleus (recoil), and Wθγis the uncertainty
in the emission angle of the γ ray relative to the beam axis. The last three terms in eq. (4.1) are due
to effects from the Doppler shift,
E?
γ= Eγ1−β cosθ
?1−β2
≈ Eγ1−β cosθγ
?1−β2.
(4.2)
where β = vr/c and θ is the angle between the recoil and γ-ray directions. The last expression
is valid if the recoil direction is close to the beam direction (0 degrees). Since the γ-ray emission
angle is determined by the position of the first IP in the detector, there is a direct correspondence
between Wθγand the position resolution, p, while Wi, Wvrand Wθrare independent of p. Since Wθγ
depends on the Doppler correction of the γ-ray energy from Eγto E?
γas
Wθγ=∂E?
γ
∂θγ∆θγ= Eγ
β sinθγ
?1−β2∆θγ,
(4.3)
with ∆θγbeing the angular resolution, one would like to have a setup with the detector places as
close as possible to the target, as high β as possible, and the detector placed at an angle close to 90
degrees. See fig. 4 for an example of a simulation, using the AGATA GEANT4 code [22], showing
how W depends on p for two different nuclear reactions.
5

Page 6

AGATA: γ-ray tracking in segmented HPGe detectorsP.-A. Söderström
(keV)
γ
E
100010201040106010801100112011401160
1
10
2
10
3
10
9Be(80Se,3n)86Sr @ 40 mm, Smearing 5 mm
W = 5.1 keV
= 3.5 keV
γθ
= 3.2 keV
r
W
= 1.8 keV
i
W
W
Figure 3: Photo peak from a simulation of the reaction9Be(80Se,3n)86Sr using the clusterisation algorithm
for tracking, a position resolution of 5 mm and a distance between the detector and the source of 40 mm.
Here W2
r=W2
vr+W2
θr.
Figure 4: Simulations of the total width, W, of the full-energy peak in an AGATA triple cluster detector
placed at a distance of 15 cm from the target to the front of the detector and at 90 degrees relative to the
beam as a function of the position resolution. In the left panel the reaction d(51V,n)52Cr and in the right
panel the reaction d(37Cl,n)38Ar is shown. Figure from [23].
There is some previous experience using this method to estimate the position resolution of a
detector. Kröll et al. have done a similar measurement on the MARS detector [24, 25], an Italian
pre-AGATA detector. The GRETA collaboration has also made measurements of the position res-
olution of their detectors, using the reaction12C(82Se,4n)90Zr at 385 MeV. They obtained a p of
4.7 mm [26]. Two experiments have also been made with AGATA detectors, one using a single
AGATA crystal [27] and another using a prototype triple cluster detector [21, 28]. Both of these
experiments were performed at the tandem accelerator laboratory at University of Cologne.
In the experiment with a single AGATA crystal, the fusion-evaporation reaction d(37Cl,n)38Ar
6

Page 7

AGATA: γ-ray tracking in segmented HPGe detectorsP.-A. Söderström
(deg)
rec
θ
00.511.522.5
β
0.066
0.068
0.07
0.072
0
50
100
150
200
250
300
350
Figure 5: Simulation of the angular and velocity distribution of the recoils in the reaction9Be(80Se,3n)86Sr.
at 70 MeV, and in the experiment with a triple cluster detector, the fusion-evaporation reaction
d(48Ti,p)49Ti at 100 MeV was used. The results gave a p of 4.3 mm FWHM at a γ-ray energy
of 2168 keV and a p of 4.4 mm FWHM at a γ-ray energy of 1382 keV in the two experiments,
respectively. The use of protons as the evaporated particle complicated the setup and the analysis,
resulting ultimately very time consuming, time that will not be available during the commissioning
phase. Furthermore, the main uncertainty in the estimation process was coming from the modeling
of the reaction by the Monte Carlo simulation used to translate the FWHM of the full-energy peak
to position resolution in the detector.
The lessons learned from this experiment is that for commissioning, the experimental setup
must be simplified even further in order to decrease the analysis time. One also needs to find a
Monte Carlo independent way to translate a measured FWHM of the full-energy peak to position
resolution. The way to deal with the first issue is to choose another reaction, that evaporates neu-
trons instead of protons, since these do not have to cross any Coulomb barrier and thus will have
lower average energy and the angular deviation of the recoils from 0 degrees with respect to the
beam will be as small as possible. A possible reaction with such properties is9Be(80Se,3n)86Sr.
Results of a simulation of this reaction are shown in figs. 3 and 5. As seen, the recoils are well
localized in the (θr,β) plane.
A new strategy has been proposed in order to obtain a Monte Carlo independent estimation of
the position resolution [21]. The idea is that instead of taking a measurement at one distance and
compare it to a Monte Carlo simulation, measurements of the total widthsWcloseandWfarare made
at two distances dcloseand dfar. The position resolution can then be estimated through
p2=1
k2(Wclose−Wfar)
?
1
dclose−
1
dfar
?−1
,
(4.4)
with k being a constant independent of W and d. Such a setup has been simulated for different
reactions of which preliminary results are shown in fig. 6. The deviation of the estimated values
7

Page 8

AGATA: γ-ray tracking in segmented HPGe detectorsP.-A. Söderström
MGT Smearing (mm)
246810
Position resolution (mm)
0
2
4
6
8
10
12
1440 mm - 140 mm
9Be(80Se,3n)86Sr
2H(81Br,1n)82Kr
9Be(136Xe,3n)142Ce
Figure 6: Position resolution as a function of smearing in the tracking program for the reactions
9Be(80Se,3n)86Sr,2H(81Br,n)82Kr and9Be(136Xe,3n)142Ce for the two distances dclose= 40 mm and
dfar= 140 mm.
from the simulated ones is due to the fact that AGATA Demonstrator placed close to the target is
covering a wide angle around 90 degrees. Another contribution to this deviation is the uncertainty
in the distance to the front face of the detector, due to its curvature radius, when the detector is
placed closer to the target than the nominal distance of 235 mm. Future investigations are foreseen
to refine the estimation procedure. For more details, see ref. [29].
5. Summary
Inordertostudyweaktransitionsinnucleiwithextremevaluesonspin, isospinortemperature,
new radioactive ion beam facilities are now being built. These new facilities will be complemented
by detector arrays with a great increase in resolving power. The novel technique implemented
in these arrays are that of γ-ray tracking using pulse-shape analysis with digital electronics. The
efficiency of the tracking depends, however, critically on the obtainable position resolution of the
detector. Some preliminary results on simulations to determine the position resolution of the the
AGATA Demonstrator γ-ray tracking array have been presented.
Acknowledgments
We would like to acknowledge the AGATA collaboration for all the effort that has been put into
this project. This work was partially supported by the Swedish Research Council and by the the
EURONS project RII3-CT-2004-506065, JRA02: Advanced Gamma Tracking Array (AGATA).
8

Page 9

AGATA: γ-ray tracking in segmented HPGe detectorsP.-A. Söderström
References
[1] R. A. Ricci, Phenomenological nuclear spectroscopy (a personal recollection), In A. Covello, et al.
(editors), Proceedings of the International School of Physics ’Enrico Fermi’, volume 169 Nuclear
Structure far from Stability: new Physics and new Technology, page 29, SIF, Bologna and IOS Press,
Amsterdam (2008).
[2] J. Simpson, The Euroball Spectrometer, Z. Phys., A358 (1997) 139.
[3] I.-Y. Lee, The GAMMASPHERE, Nucl. Phys., A520 (1990) c641.
[4] B. Rubio and T. Nilsson, NuSTAR, Nucl. Phys. News Intl., 16 (2006) 9.
[5] W. F. Henning, FAIR and its experimental program, J. Phys. G, 34 (2007) 551.
[6] S. Gales, SPIRAL2 at GANIL: A world leading ISOL facility at the dawn of the next decade, Prog.
Part. Nucl. Phys., 59 (2007) 22.
[7] J. Gerl and W. Korten (editors), Technical Proposal for an Advanced Gamma Tracking Array for the
European Gamma Spectroscopy Community (2001).
[8] J. Simpson and R. Krücken, AGATA: The Advanced Gamma Tracking Array, Nucl. Phys. News Intl.,
13 (2003) 15.
[9] I.-Y. Lee, M. A. Deleplanque, and K. Vetter, Developments in large gamma-ray detector arrays,
Rept. Prog. Phys., 66 (2003) 1095.
[10] N. Warr, et al., MINIBALL: The first gamma-ray spectrometer using segmented, encapsulated
germanium detectors for studies with radioactive beams, Eur. Phys. J., A20 (2004) 65.
[11] R.Venturelli and D. Bazzacco, Adaptive Grid Search as Pulse Shape Analysis Algorithm for
gamma-Tracking and Results, In D. R. Napoli, et al. (editors), LNL Annual Report 2004, page 220,
Legnaro, Italy (2005).
[12] L. Nelson, et al., Characterisation of an AGATA symmetric prototype detector, Nucl. Instrum. Meth.,
A573 (2007) 153.
[13] Th. Kröll and D. Bazzacco, Simulation and analysis of pulse shapes from highly segmented HPGe
detectors for the γ-ray tracking array MARS, Nucl. Instrum. Meth., A463 (2001) 227.
[14] B. Bruyneel, Characterization of segmented large volume, high purity germanium detectors, Ph.D.
thesis, University of Cologne (2006).
[15] J. Ljungvall and J. Nyberg, A study of fast neutron interactions in high-purity germanium detectors,
Nucl. Instrum. Meth., A546 (2005) 553.
[16] J. Ljungvall and J. Nyberg, Neutron interactions in AGATA and their influence on γ-ray tracking,
Nucl. Instrum. Meth., A550 (2005) 379.
[17] A. Lopez-Martens, et al., γ-ray tracking algorithms: a comparison, Nucl. Instrum. Meth., A533
(2004) 454.
[18] D. Bazzacco, The Advanced Gamma Ray Tracking Array AGATA, Nucl. Phys., A746 (2004) 248c.
[19] J. van der Marel and B. Cederwall, Backtracking as a way to reconstruct Compton scattered γ-rays,
Nucl. Instrum. Meth., A437 (1999) 538.
[20] L. Milechina and B. Cederwall, Improvements in γ-ray reconstruction with positive sensitive Ge
detectors using the backtracking method, Nucl. Instrum. Meth., A508 (2003) 394.
9

Page 10

AGATA: γ-ray tracking in segmented HPGe detectorsP.-A. Söderström
[21] F. Recchia, In-beam test and imaging capabilities of the AGATA prototype detector, Ph.D. thesis,
Università degli Studi di Padova (2008).
[22] E. Farnea, et al., To be published in Nucl. Instrum. Meth..
[23] A. Al-Adili, Simulations of Doppler Effects in Nuclear Reactions for AGATA Commissioning
Experiments, Masters thesis. In manuscript.
[24] Th. Kröll, et al., Gamma-ray Tracking With The MARS Detector, AIP Conf. Proc., 656 (2003) 357.
[25] Th. Kröll, et al., In-beam experiment with the γ-ray tracking detector MARS, Nucl. Instrum. Meth.,
A586 (2008) 421.
[26] M. Descovich, et al., In-beam measurement of the position resolution of a highly segmented coaxial
germanium detector, Nucl. Instrum. Meth., A553 (2005) 535.
[27] J. Eberth and J. Simpson, From Ge(Li) detectors to gamma-ray tracking arrays – 50 years of gamma
spectroscopy with germanium detectors, Prog. Part. Nucl. Phys., 60 (2008) 283.
[28] F. Recchia, et al., In-beam test of the AGATA prototype triple cluster, To be published in Nucl.
Instrum. Meth.
[29] P.-A. Söderström, et al, In manuscript.
10