MUST II: Large solid angle light charged particle telescope
for inverse kinematics studies with radioactive beams.
E. Pollacco°, E. Atkin°, F. Auger°, P. Baron°, J.P. Baronick, Y. Blumenfeld,
A. Boujrad, A. Drouart°, P. Edelbruck, L. Lavergne, L. Leterrier,
L. Olivier, B. Raine, A. Richard, M. Rouger°,
P. Roussel-Chomaz, F. Saillant, M. Tripon, E. Wanlin
°DSM/DAPNIA/SPhN, CEA Saclay, 91191Gif-sur-Yvette, France
Institut de Physique Nucléaire, IN2P3-CNRS, 91406 Orsay, France
*GANIL, IN2P3-CNRS / DSM-CEA , BP 55027, 14021 Caen Cedex 5, France
Abstract. Over the past four years we have studied (p,p'), (d,p) ,(d, 3He) and other reactions using radioactive beams in inverse
kinematics to obtain spectroscopic information for nuclei away from the valley of stability After a general overview of the
experimental method we will describe our ongoing MUST II development. This is to build a very compact (1000cm 3) three stage
telescope with an active area of 100cm² with position resolution of 0.70.7 mm² and time of flight measurement. The mass
identification and energy dynamic range is of 0.4 to 80 MeV.A up to alpha particles. The compactness of the array is assured
through the use of an ASIC development to measure the time of flight and energy. The large solid angle coverage of 2.6sr and
compactness of this array will allow it to be used in particle-gamma coincidence experiments.
INTRODUCTION important to increase the detector geometrical
Direct reactions with light ions (protons, deuterons, Also of importance is the beam emittance because
… alphas) on stable nuclei were shown to be an of kinematic effects. To obtain reasonable energy
important tool in nuclear spectroscopy. Reactions like resolution the angular resolution better than 0.5° is
elastic and inelastic give matter radii and deformation. required. For available beams this implies that beam
The transfer of one or two nucleon (d,p), (d, 3He), tracking to better than 1 mm is vital. A final comment
(t,p)… give spin and parity as well as spectroscopic about exotic beams, is their purity which is not often
strength. These elements were central in building our guaranteed. Therefore the coincident detection of the
present macro and micro vision of the nucleus. It is projectile-like fragment in a spectrometer, is frequently
therefore natural to extend this method to unstable crucial to obtain a clean background in the light ion
species. Unstable nuclei far outnumber stable ones and energy spectra. In pickup reactions, like (p,d), the
we enquire to what extent do our present models, tuned kinematics for the detected light ejectiles are forward
on stable systems, will apply to highly neutron rich or focused making it relatively easy to cover a large
proton rich ensembles. Inklings of modified structure is fraction of the solid angle. This is not so true with the
given by the neutron rich Li and Be isotopes . interesting (d,p) reaction. The p cover a wider angular
Neutron haloes in nuclei like 6He, 11Li and 19C are domain with steep kinematic variations. Thus for an
structures that were not expected and still need to be acceptable telescope position resolution, the distance to
understood . target has to be increased at the expense of solid angle.
Direct reaction measurements with exotic nuclei are Energy resolution deteriorates with the combined
performed in inverse kinematics using CH2 or CD2 effects of target thickness, beam emittance and total
targets. Liquid or solid H2 and gaseous T2 have been angular resolution which are limiting factors for this
employed. The beams of course are not intense. With experimental technique. Values better than 300 keV are
cross sections of a few mb/sr and reasonable target difficult to achieve even when applying thin targets and
thickness, present detection systems require beams beam tracking. A solution is to perform gamma-particle
intensities better than 104 particles/s . Hence it is coincidence measurements. Highly efficient Ge
detectors like ExoGam coupled to position sensitive
detectors will be used at GANIL. Thick targets can then The standard flight distance is 15cm and the
be employed which offsets the loss in efficiency. geometry of the telescope ensemble is highly dictated
Elastic and inelastic scattering measurements are by this choice. Thus, the mechanics of the telescopes is
fundamental and are often a prerequisite to transfer a truncated pyramid with a base 1313cm² with a
reaction analysis. The complication with this reaction is vertex at 15 cm and an “active” face of 1111cm². To
the relatively low threshold that is required and values allow set-up flexibility in gamma-particle
of 0.4 MeV.A with particle identification, is a must. In measurements, the CsI can be removed. In
our experiments we have opted for a time of flight back/forward angle measurements a typical ensemble
method. has high and smooth solid angular coverage as shown
In many ways MUST II is defined to have features in fig. 1.
very similar to that of MUST . Our original ambition
was to increase the active area to cover symmetrically
and widely the forward/backward angles without
modifying the electronic structure. However with the
opportunity to found the electronics in ASIC
(applications specific integrated circuit) form presented
a significant reduction in the volume occupied by the
electronics behind the telescope and the number of
cable/connectors runs. This solution opens the
possibility to perform particle-gamma, measurements
that permit only limited volume around the Ge clovers.
Further, although ASIC developments are costly, the
cost per channel is inexpensive, making the prospect
for future increase of solid angle possible. Presently,
MUST II is an ongoing project where we have opted
for an ASIC solution for the front-end electronics. It
consists of six telescopes that multiply the solid angle
coverage of the MUST ensemble by a factor of three.
The large phase space coverage will make it possible to
measure low yield reactions and open the prospect to
study several reactions simultaneously ((p,d), (p,t),
(p,2p) etc to bound or unbound states.
Each telescope consists of a Si double-sided strip
detector, Si(strip), followed by a Si(Li) and CsI crystal.
The Si(strip) is of dimension 1010cm² with 128 strips
on either face. The crystal is an n-type low resistivity
(~ 6 KOhm-cm) to be biased to twice the depletion
voltage to allow a high field strength over the full
thickness of 300µm. The masks for this detector are not
much different from those of MUST with the exception
that the inter-strip will be 56 µm. Overall energy and Fig 1. Geometry of telescope (top) and Efficiency vs
time resolution expected are 50 keV and 250 ps for laboratory angle for the Si(Strip)
alphas of 5.48 MeV. Two Si(Li) detectors of thickness
4.5 mm will be used to cover the 100 cm². Each crystal ELECTRONICS
of ~105 cm² will be segmented into 8 pads.
Resolution aimed for is 120 KeV and a dynamic range The electronic hardware of MUST II consists of
for protons up to 32 MeV. The CsI crystals are three basic units. MATE, MUFEE and MUVI. MATE,
segmented to shadow the Si(Li) from a point target and (Must Asic for Time and Energy) delivers the E and T
are of length 3 cm to stop 80 MeV protons. The light from the detectors. A total of 18 MATEs/telescope are
output is read by 22cm² photodiodes. An energy distributed on two quasi identical card MUFEE (MUst
resolution about 6% is expected for alphas of 5.48 Front End Electronics) connected with the detectors via
MeV. 8 cm Kapton bands. Data transfer, high tension and
communication are done through 25 channel pulser inputs, gain, shaping and inverse current
connectors. A single width unit MUVI, (MUst in VXI) measurement) are satisfied via the industrial protocol
in VXI standard, assures the slow control and data I²C.
coding for the six telescopes. With the exception of The principal results derived from simulation for
MUVI, the general philosophy is that each telescope in the strip detector are as follows (capacitance 65pF, dc
the reaction chamber is electrically independent. 20 namp,). (The slow controlled dynamic range and
resolution for the Si(Li) and CsI are given in italics);
MATE Energy range: 50 MeV, 250 MeV
Energy resolution (fwhm): 16 KeV, 90 KeV (filter
The ASIC MATE has 16 channels per chip and rc-cr 1µs, 3µs) Track and Hold, T/H.
process signals delivered from silicon strip detectors, TAC range 300 ns.
Si(strip) pads and photodiodes. The architecture Time jitter (fwhm): 240 ps (protons 6 MeV filter
delivers three types of information for each channel: rc-cr 30 ns)
1.Value of the energy losses from particles hitting Threshold range: ± 1.0 MeV, on 8 bits quantization
the telescope. Power consumption: 35 mW.
2.Value of time of flight from a leading edge Readout: 2MHz serial. All channels read at request.
discriminator with adjustable threshold and Time to MATE uses a BICMOS technology A.M.S. 0.8µm. The
amplitude converter, TAC, with a common stop. first submission was in May 2002 and characterization
3.Value of DC leakage current for monitoring will start in October 2002.
The choice of discriminators is a leading edge. Current MUFEE
pulses were simulated and showed that the time
resolution is sufficient to separate the 3He and 4He over The main function of MUFEE is to process the
the Si (strip) E-dynamic range. Namely, the differential physical signals from the detectors; each MUFEE
walk for different particle types of the same energy is processing 128 strips of one side of the Si(strip) and 16
negligible in comparison to the time of flight. physical signals from the Si(Li) or CsI detector.
I leak i selidi
Rf Filtre &
Filter & energy i seleni
Idl cf Ampli Hold
Si in VIC
Strip in i
time i Th
Filter +discri Or
in Res Thr Thr
seuilp seuilm side hyst i
inhibit Tstart ResetStart
Fig 2. MATE schematic diagram
In spite of the relatively large dynamic range MUFEE has I²C driven internal pulse generator to
requested, MATE process bipolar signals in energy and allow the different functions to be tested and the
time channels and therefore software adaptable for both physical calibration of the E and T channels. A single
sides of the strip detector. The programming functions injection capacitor is used so the injected charge is the
of MATE (discriminator levels, inhibits of channels, same for all channels of ASIC. An external pulse
generator input is also made available. Numerical
information for the slow control of the MATEs is and is specifically studied to allow a high density of
carried on the standard serial bus, I²C. channels with a minimum of acquisition dead time, DT.
To insure good immunity against electromagnetic Each telescope is independent and connected to (one of
perturbations, all control signals are carried in Low six) cards CAS residing in MUVI. CAS delivers the
Voltage Differential Signal (LVDS) except the STOP- slow control for discriminator, pulser, current reading
TDC signal that is in Low Voltage Positive Emitter etc. It receives the hit signal and distributes the stop
Coupled Logic (LVPECL) to minimize the timing signal for the TACs, T/H, clock readout and pulser
jitter. For the same reason, the two analog lines, trigger. Four lines are dedicated to analogical
carrying E and T travel in differential current through differential outputs at 2 MHz that carry the signal train.
twisted pair. The fan-in to form the trigger and the The architecture of CAS is presented in fig 3. The
distribution of the detector tensions, STOP-TDC and coding block allows the pattern of signals on four
T/H are also on the card. channels to be coded. Each channel is composed of 2
Of importance is the relatively large thermal energy differential converters giving a numeric resolution of
(~15Watt/telescope) generated by the electronics in 14 bits in 400 ns. The FPGA+DSP in CAS will cover
vacuum. This is drained via a heat exchanger important functions, such as the suppression of zero
sandwiched by the two MUFEE boards. The MATEs readings, the I²C communication, and sliding scale.
are cooled directly by a conductive interface material Three trigger modes are available. The common DT
(Gap Pad). The temperature is monitored via a sensor mode is estimated to correspond to10% dead time at
on the card. Mechanically, MUFEE is of 1KHz. The mode semi autonome allows the liberation
dimension130130 mm². of all telescopes that are not triggered. This option cuts
down the acquisition time by a factor of two. Finally
MUVI the mode autonome, is characterized by removal of
GMT function and each CAS functions independently.
The ensemble of the E and T information (3072 for For the last two modes the reconstruction of each event
6 telescopes) is sent to acquisition system based on is done through the time stamping of CENTRUM The
VXI-C. For MUST II and two tracking detectors CATS triggers and acquisition with CENTRUM will facilitate
 the configuration will include a MUVI, a trigger the integration of MUST II with other detection
(GMT) coupled to a bit pattern and scalars unit (U2M). systems.
CENTRUM provides a time and event stamping. The
back plane data transfer for CENTRUM and CONCLUDING REMARKS
distribution of resources for the visual inspection of
signals is done by GAMER. Four QDCs (XDC3214) Direct reactions are an important tool and the
will code the CATS data. The slot is coupled to a development of MUST II will make it possible to
VIC8250 for connection to VME or VXI or coupling to exploit lower yield reactions in this domain. Other
a CPU allowing an Ethernet access. reaction studies that require similar specifications will
benefit from such a development. Examples are
resonant elastic scattering , breakup reactions 
MUST2 / MUVI / CAS: block diagram
and energy measurement in magnetic spectrometers.
STOP Logic LVPECL
STOP Logic LV
Trigger signals + STOP
TRIG We hope to perform our first tests in 2003 with MUST
SIGTRIG II and first measurement in 2004.
JTAG DSP JTD
Logic LV I2Cint Logic LV
Slow control JTAGs
FROM / TO MUVI
FROM / TO MUFEE
Logic LVDS ETAT Logic LV
MUFEE Acquisition Acquisition signals
FPGA 1. M. Labiche et al., Phys. Rev. C 60 (1999) 027303
DAG EG DSP LP DSP 2. A. Lagoyannis et al., Phys.Lett. B518 (2001) 27
I to V + ADC + EG
3. Y. Blumenfeld et al., NIM A366 (1999) 298
ADINSANA Logic LV
INSLOG2 INSLOG 1&2
INSLOG 4. S. Ottini-Hustache et al., NIM A431 (1999) 476
I_TST 5. J. Gómez del Campo et al., Phys. Rev. Lett. 86
INSANA 1 & 2 (2001) 43, reference therein.
V_TST, SYN_TST VTST 6. A. Wuossma et al., Ann. Rev. Nucl. Part. Sci.
Astrophys. 45 (1995)1
7. E. C. Pollacco et al., contribution to CAARI 2002
Fig 3. MUVI schematic diagram.
The units mentioned above, with the exception of
MUVI, correspond to basic elements of the GANIL
acquisition system. MUVI is a MUST II development