1

Conseil Scientifique et Technique du SPhN

LETTER OF INTENT

Title: SOLAR Project

Experiment carried out at: SPIRAL2/GANIL, FAIR/GSI, HISOLDE/CERN, RIKEN

Spokesperson(s): David JENKINS (York), Emanuel POLLACCO (IRFU)

Contact person at SPhN: Emanuel POLLACCO

Experimental team at SPhN: Bart BRUYNEEL, Anna CORSI, Alexandre OBERTELLI,

Stefano PANEBIANCO, Emanuel POLLACCO

List of IRFU divisions and number of people involved:

SPP(1), Sedi/LDEF (2), Sedi/LILAS (2), SACM (5), SIS/LCAP (1) – Still to be fully

defined

List of the laboratories and/or universities in the collaboration and number of

people involved:

University of York, UK (2), STFC Daresbury Lab (2), University of Liverpool (1),

Manchester University (2)

– Still to be fully defined.

SCHEDULE Estimated total duration of the proposed project: 6 years

Possible starting date of the experiment: 2017

Expected duration of the data analysis: 2 years

ESTIMATED BUDGET Total investment costs for the collaboration: Demand ERC for a total of 15M€

Share of the total investment cost for SPhN: None

Total Travel Budget for SPhN: None

2

SOLAR Project

Letter of Intent for CSTS approval - 9th Nov 2012 Shebli Anvar1, Guy Aubert1, Andrew Boston2, Philippe Bredy1, Bart Bruyneel1, Antoine Chancé1 , Anna Corsi1, Antoine Dael1, Eric Delagnes1, Dominique Yvon1, David Jenkins3, Ben Kay3, Marc Labiche4, Ian Lazarus4, Alexandre Obertelli1, Stefano Panebianco1 , Emanuel Pollacco1 , Lionel Quettier1 and Jean-Michel Rifflet1

1 IRFU, C.E.A.-Saclay, 2University of Liverpool, UK, 3 University of York, UK 4 STFC Daresbury Laboratory, UK

Synopsis Physics and technological headway has often been driven by advancements in instrumentation and technique developments in Nuclear Physics. This document outlines the case for a powerful detector facility called SOLAR, based on a superconducting solenoid, which is robust, versatile and addresses a wide class of experiments foreseen for present RI beam facilities as well as RI beam facilities coming online in the next decade. Importantly, the instrument will have strong synergies with current hot topics in medical imaging such as simultaneous PET/SPECT and MRI, as well as other areas of Physics. An approach to funding such a system would be through a bid to the ERC synergy scheme drawing on the expertise of key partners in the UK and France. The instrument itself, however, would be of clear benefit to a very large community of scientists engaged both in cutting-edge nuclear physics and topical areas of applications. This document is intended to give an overview of the Physics and the key technical challenges which will be addressed. A preliminary discussion of the budgetary, management, risk and safety issues concerned with such a project is presented. In addition, annexes contain more details on the detector subsystems and details of simulation work which has already been carried out. The intention would be, following CSTS approval, to take this work forward to a much more detailed level in preparing an ERC bid for January 2013. We can then reprioritize different portions of the SOLAR package according to the advice we receive.

3

Section 1: Program Introduction Historically, nuclear physics has advanced through developments in accelerator technology, production of radio-isotopes and improvements in detector technology. These developments have led to important synergies with societal needs in areas as diverse as cancer therapy and medical imaging. Nuclear Physics is presently undergoing a renaissance with the advent of accelerated radioactive ion beams (RIBs) from a wide range of facilities worldwide and exciting prospects at facilities presently under construction. The often low intensity of such beams compared to stable accelerated beams demands sophisticated detector systems that can maximize the experimental sensitivity. The present proposal foresees a highly versatile system that can be moved between different classes of radioactive beam facility and operated successfully both with intense radioactive beams and with very low intensity beams. The demands on this system will lead to developments having strong synergy with the requirements of both medicine and other areas of Physics such as astrophysical plasmas. The heart of the system, which we call SOLAR, is a large superconducting solenoid. This solenoid will contain different detector subsystems that can allow it to operate in different modes to service the broad Physics program intended. These different modes are outlined briefly below and described in detail in section 2.

- SOLAR-TPC – An active target – time projection chamber, TPC system which can image light ions

- SOLAR-HELIOS – A solenoidal spectrometer mode where light ions perform helical trajectories in the field and are detected along the axis

- SOLAR-GAMMA – A scintillator array for gamma-ray detection - SOLAR-CHYMENE – A cryogenic solid hydrogen target

1.1: Nuclear Physics Program Nuclear physics is at an important crossroad as intense radioactive beam facilities come on-line allowing us to address many of the key questions in the subject in more detail as well as opening up avenues of investigation for the first time. There are many documents available which provide a comprehensive evaluation of such physics programs, including the NuPECC long-range plan of 2010, and documents supporting the design and mission of the various RIB facilities. Rather than trying to be complete, we will provide specific examples which justify the design of the SOLAR facility and which can be done, in principle, better or more sensitively than with existing or intended

instrumentation. The following motivations are partly (and freely) inspired from Letters of Intent for SPIRAL2 [LoISn,LoIKr,LoIActar].

Nuclear structure Historically, nuclear models were able to provide a good description of stable nuclei by starting from basic foundations such as the properties of doubly-closed shell nuclei. State-of-the-art nuclear models are found to be strongly challenged or to fail, however,

4

as we approach the limits of nuclear stability. This breakdown has been attributed to the possible roles played by three-body forces, coupling to the continuum and the tensor term in the nucleon-nucleon interaction. This motivates the focus of the present project so far as nuclear physics is concerned on shell-gap evolution and the unusual quantum phenomena encountered in exotic nuclei. The key physics questions are: what is the structure of exotic nuclei? How do the shell gaps evolve for weakly-bound nuclei? We will want to study low-lying bound and unbound states in exotic nuclei, their configurations and spectroscopic factors. To extract structure information relevant to the signature of new shell effects (modification of known magic numbers, new shell gaps in neutron-rich region, etc.), the measurement should be as complete as possible (elastic and inelastic scattering, transfer reactions to the main channels) in order to fix the coupling between the main processes to be included in the coupled-reaction scheme. Key observables are the energy and width of the excited states, and the angular distributions of direct cross sections to the single-particle states, giving access to transferred angular momentum. The principal tool for obtaining precise spectroscopic information is the light-ion transfer reaction, including reactions such as (p,d), (p,t), (d,p) and (t,p). This technique is currently undergoing a renaissance as it comes to be applied to studies in inverse kinematics with radioactive ion beams. In principle, it would be desirable to study extended isotopic chains and push to the very limit of beams available at a given facility. The conventional approach to the study of such reactions in inverse kinematics is using a barrel-like configuration of silicon detectors such as TIARA or T-REX. There are clear limitations in this approach in terms of required beam intensity – a realistic lower limit for studies with ISOL beams being around 104 particles/s. Secondly, systems using silicon detectors at a fixed lab. angle suffer from kinematic compression in the excitation energy spectrum. A way of avoiding the kinematic compression is to use a solenoidal spectrometer such as the pioneering HELIOS spectrometer at Argonne National Laboratory [Lig10, Wuo07]. In this mode, light ions from transfer reactions in inverse kinematics perform helical orbits in the high magnetic field and are detected in a silicon detector array around the solenoid axis. Different particle types can be identified from their cyclotron period. This would be SOLAR operating in the HELIOS mode. To access, the most exotic cases with intensities below 104/s, we would use SOLAR in active target – TPC mode.

The region around 132Sn The magic numbers are the foundation of nuclear structure and our understanding of the nuclear properties. The simplest excitations in doubly-closed shell nuclei consist of particle-hole (p-h) states in which particles are excited across the energy gap defining the closed shell. The nuclear shell effects are revealed by identification of the single-particle (-hole) and 2-particle (-hole) states and by determination of their spectroscopic factors. The location of such states is particularly interesting in the region near closed-shell nuclei. In the region above A=100, there are only two doubly-magic nuclei - stable 208Pb and the neutron-rich radioactive nucleus, 132Sn. The latter will be a strong focus of investigation. Spectroscopic information deduced on the various isotopic chains for neutron-rich unstable nuclei around Z=50, N=82 will be used to reconstruct the single-particle spectrum for neutrons and protons in this region. This study is not only of interest to nuclear structure as modifications of shell structure have a strong influence on the modeling of astrophysical processes. A shell quenching at N=82 (i.e. a shell gap less pronounced than expected through macroscopic-microscopic mass models), and along

5

N=50 and 82 will modify the expected abundances of heavy nuclei. The structure and spectroscopic information for unstable neutron-rich nuclei are also required as inputs for r-process calculations (e.g. solar abundances, neutron-capture rates). In the past, level schemes for the stable tin isotopes have been determined through a complete set of reaction data: Coulomb excitation, in-beam γ-ray spectroscopy, heavy-ion induced reactions, inelastic scattering of protons, deuterons, α particles and one- and two-nucleon transfer reactions like (d, t), (3He,n) (t,p) or (p,t) [Bla02]. The levels of the neutron-rich tin isotopes (132-135Sn) have been studied by beta-decay studies, mainly at ISOLDE [Bjo86,Fog95,Omt95]. A ground-breaking 132Sn(d,p) study was made at the HRIBF facility at Oak Ridge but only the bound states were studied and the experimental resolution was somewhat poor [Jon10]. A more complete data set should be collected to fully understand the structure of 133Sn and the location of single-particle strength in this region. For 134Sn, only the energy of the bound states are known, and spin and parities were assumed. For 135Sn, only the ground state has so far been observed. A clearer understanding of the interplay between shell effects around N=82 and correlation effects should be drawn from new data collected for the low-lying states of tin isotopes at N=84,85. We propose to study the shell structure around N = 82 and the low-lying excited states of the N=85, 86 tin and antimony isotopes using the (d,p), (p,d) and (p,t) transfer reactions using beams of 134,135Sn and Sb from HIE ISOLDE or SPIRAL2. The kinematics of such reactions is shown in figure 1.1. Angular cross sections for transfer reactions to the excited states at least up to 6 MeV are of the order of few mb/sr in the low energy regime available with HIE ISOLDE and SPIRAL2 beams (4-10 MeV/u). Where beam intensities exceed 104 /s, we will use SOLAR in the HELIOS mode. To push to the most exotic cases, we would need to use the active target-TPC mode of SOLAR. The addition of a gas target would allow us to also carry out (3He,d) studies, for example.

Figure 1.1: Energy versus lab. angle of the light charged particles produced from (d,p), (d,d’), (d,t) and (d,3He) from 134Sn at 10 MeV/nucleon [LoISn].

Shape coexistence in the proton-rich and neutron rich Krypton isotopes A remarkable feature of the atomic nucleus is its ability, under certain circumstances, to take on different mean-field shapes for a small cost in energy compared to the total binding energy of the nucleus. A spectacular example of this phenomenon is in the light lead nuclei, for example, 186Pb where two excited 0+ states associated with a prolate and oblate deformed minimum coexist with a spherical ground state within an excitation energy range of less than 1 MeV.

6

In the past, shape coexistence has largely been inferred from in-beam studies, where the layout of levels and the presence of multiple excited bands has been attributed to different mean-field configurations. In certain special cases, for example, 72Kr, electron spectroscopy has been used where the lowest excited state has, highly unusually, spin/parity of 0+ [Bou03]. Laser spectroscopy has also been used to map out changes in nuclear radius and deformation but this is only applicable to the ground state or long-lived metastable states. A breakthrough in the study of shape coexistence has come about through the availability of intense re-accelerated radioactive beams. Such beams can be employed in Coulomb excitation studies. Coulomb excitation is a very useful technique for examining nuclear collectivity through extraction of transition matrix elements. Such matrix elements can also be extracted from lifetime measurements but the techniques are complementary as states are excited from the ground state up in Coulomb excitation, often leading to a very different level population from in-beam lifetime studies. The principal advantage of Coulomb excitation, however, is the possibility to extract diagonal matrix elements through the second-order process of reorientation. Such matrix elements allow the sign of the spectroscopic quadrupole moment to be extracted and, hence, the sign of the associated nuclear deformation. An excellent example of this technique is the Coulomb excitation of 74Kr and 76Kr [Cle07], where a large number of matrix elements were extracted (see figure 1). The data were such as to allow discrimination between competing mean-field descriptions [Cle07]. There are two regions of the nuclear chart which we could address in order to better understand shape coexistence. The first is the region close to the N=Z line for A~70. Here there is a rapidly changing nuclear shape evolving from the spherical nucleus, 56Ni with doubly-closed shell, through nuclei like 64Ge which are suggested to be triaxial, to 68Se [Fis00] and 72Kr [Fis01,Fis03] which are believed to exhibit a close competition between oblate and prolate minima [Pet02,Lal99,Naz85]. The shape coexistence in this region is said to be driven by increasing occupation of the g9/2 orbital. The second region of interest would be the neutron-rich Zr, Sr and Kr isotopes around N=60 which has attracted a great deal of theoretical and experimental activity over recent years. These nuclei represent some of the best examples of interplay between single-particle and collective modes of excitation. In such regions, large variations in the observed spectroscopic properties as a function of proton or neutron number make the theoretical interpretation particularly challenging. It is well established that the neutron rich Sr and Zr isotopes are characterized by a sudden onset of quadrupole deformation at neutron number N = 60. This becomes already evident from the excitation energies of their first excited states. Measurements of B(E2) values in the Sr and Zr chain confirm the sharp change in collectivity at N=60. A first spectroscopic quadrupole moment was obtained for the 2+1 of 96Sr (N=58) by Coulomb excitation reaction at REX-ISOLDE. The shape change seems to be very located in proton and neutron number. The Mo isotopes (Z=42) do not present such a sharp transition and might indicate the upper Z limit of the phenomenon. Recent in-beam experiments as well as mass or <r2> measurements in 96Kr lead to inconsistent conclusions. Since excited 0+ states are a key indicator of shape coexistence, the (p,t) reaction would

be important since it has been shown to be an excellent tool to populate low-lying 0+ states with a two-particle-two-hole structure (see for example [Duv83, Bor77, Ard78]). The total angular momentum, L, of the final state is clearly identified from particle-angular distributions. 0+ states of the final nucleus are favorably fed whenever they have a shape similar to the one of the incident nucleus (see figure 1.2). This experimental technique has been extensively used for stable Ge-

7

Se-Kr nuclei in direct kinematics. The ratio, R = (0+2)/(0+1) was found to be extremely sensitive to shape transitions and shape coexistence [15]. We propose to perform a similar systematic study in the case of unstable heavy Kr isotopes. In the case of a shape transition from spherical deformation in 94Kr to prolate shapes for both 96Kr and heavier isotopes, as suggested by microscopic theories, the (p,t) reaction from 96Kr to 94Kr is expected to populate significantly the (prolate) unknown first 0+ excited state in 94Kr, and a large R value is expected. In a similar scenario, 94Kr(p,t)92Kr should weakly feed the (unknown) second 0+ state of 92Kr. These reactions would be studied using SOLAR in the HELIOS mode. The rigidity of the tritons dictates the long homogeneous field region and high maximum field (4-T) demanded of the solenoid used for SOLAR.

Figure 1.2: Comparison of triton angular distributions for the population of a 0+ state (L=0 transfer) and the population of a 2+ state (L=2 transfer) at different incident energies for the reaction 96Kr(p,t)94Kr. Excited states are considered at 1 MeV in both cases.

The second area of interest is to obtain a more microscopic understanding of the shape coexistence phenomenon in this region. This could be achieved by carrying out a study of the 74Kr(d,p) and 76Kr(d,p) reaction in inverse kinematics (see figure 1.3). Existing potential energy surface calculations tally with the findings from Coulomb excitation of substantial deformation in this region and suggest that the lowest prolate deformed orbits are 3/2+[431], 3/2-[312] and 3/2-[301] Nilsson configurations which are dominated by components of the g9/2, f5/2 and p3/2 spherical shell model states respectively. Identification of the g9/2 strength in particular should reveal a great deal about the nature of the deformation encountered. The intruder nature of this orbit means that the deformed orbits stemming from it maintain their single-particle purity. Transfer on deformed nuclei is more challenging but still very valuable as “fingerprints” of deformation can be obtained [Elb69].

8

Figure 1.3: The left-hand panel shows the angular distribution and cross-section dependence expected for the 76Kr(d,p) reaction as a function of c.m. angle. The right-hand panel shows the trajectories of different light ions for (d,p) and (d,d’) reactions as a function of c.m. angle.

Spectroscopy of the most exotic species The most neutron rich nuclei are currently produced at relativistic energy at fragmentation facilities such as RIKEN and, in the future FAIR. Nuclei of large interest regarding the shell structure of exotic nuclei will be produced at intensities of about 1 particle per second or less. Only the use of very thick targets to counterbalance the low-beam intensity allows to reach an acceptable luminosity. An innovative detection system called MINOS is under development to allow the spectroscopy of the most exotic nuclei at RIKEN from 2014 and FAIR from 2018. MINOS is a thick liquid hydrogen target of about 150 mm coupled to a compact vertex tracker surrounding the target. MINOS has been designed to be coupled to a gamma spectrometer for in-beam gamma spectroscopy. It has been funded by Europe program through a ERC starting grant (PI A. Obertelli, CEA Saclay). The MINOS physics program could be highly enriched by the coupling to the SUPHeX device for particle spectroscopy of bound and unbound states in very exotic nuclei via (p,2p) measurements, extending therefore the physics program of MINOS to particle spectroscopy of unbound states. A new MINOS TPC-tracker will be used to measure the proton momentum. A very-high precision on the energy of the protons will be reached from their curvature with no need for a calorimeter. The setup should allow an energy resolution of <2 MeV FWHM, mainly governed by the angular straggling of protons in the thick hydrogen target, providing a unique device for the particle spectroscopy of very neutron rich nuclei produced at intensities of less than a particle per second.

Nuclear astrophysics Lorem ipsum

Fission Studies The understanding of the dynamics of fission process, leading to the split of the compound

nucleus, is probably one of the most challenging problems in nuclear physics. While, from

one side, the properties and structure of the compound nucleus are clearly important, only a

precise knowledge of the final state can lead to a reliable description of the main features of

the process. Nuclear fission is probably the only process deeply involving all the different

aspects of nuclear physics, from the basic nuclear structure to the N-body collective

interactions. Therefore it constitutes an ideal “nuclear physics laboratory” whose results have

a direct impact on a wide number of nuclear physics fields. Moreover, the development of a

new generation of nuclear energy plants is presently suffering from a poor or absent

9

knowledge of nuclear data on the fission of more “exotic” actinides or at higher neutron

energy. The most advanced experimental techniques which have been developed for the study

of fission rely on the identification of the final state particles. This is important in fission

cross section measurements in order to tag the fission process over all the possible reactions.

Secondly, the study of the fission final state needs very selective identification of low energy

ions within a wide range of nuclear charge and mass, from very light (hydrogen isotopes and

alphas as in ternary fission) to heavy ions (fission products). Indeed, a TPC, coupled with

external magnetic field, can be a very suitable device to provide systematic measurement of

fission cross sections together with a full characterization of the final state particles (mass,

nuclear charge and kinetic energy of fragments) produced by neutron or photon induced

fission of actinides in a large range of excitation energies and for a large number of fissioning

nuclei. This systematic approach is fundamental both for fundamental physics and

applications. The development of fission models, based on phenomenology and systematics

or on more microscopic physics ingredients, needs a large set of experimental data to improve

their predictive power and reliability. For example, the correlation of the total kinetic energy

release with the fragment mass gives information concerning the properties of the fissioning

nucleus at the scission point. Although a lot of data are available at thermal energies, there is

a lack of knowledge on the evolution of the fission properties with neutron energy (from

thermal to MeV range) and only few fissioning actinides have been studied in the keV-MeV

range. The latter are very important for the design of future fast reactors (Generation IV) and

for the optimization of Radioactive Ion Beam facilities. The advantage of using a compact

and versatile setup is that it can easily fit to different neutron facilities by providing coherent

data at different energy ranges.

In this framework, a R&D project called FIDIAS (FIssion Detector at the Interface with

Astrophysics) has started at Irfu at the beginning of 2010 to evaluate the feasibility of a multi-

purpose TPC based on the Micromegas detector for the full reconstruction and identification

of low energy heavy ions. The detector is arranged as a double-sided TPC, with the actinide

target in the middle, in order to detect both fragments in coincidence. The main advantage of

this setup is its full angular coverage, leading to high detection efficiency. This has to be

compared with a “standard” two-arm setup [14] which has an efficiency rarely higher than

10%. The TPC ionization products are detected by a Micromegas detector [5]. This micro-

pattern gaseous detector has been chosen because, thanks to its localized electron cloud and

its fast ion evacuation provided by the presence of a micromesh, it has been proven to achieve

very high space resolution, even at high rate. Moreover, the Micromegas is quite insensitive

to gammas and its material budget is very low, providing radiation hardness. The ionizing gas

foreseen for the FIDIAS TPC is He, since at atmospheric pressure it gives the longest path for

the fission fragments. However, thanks to the high granularity achievable with the

Micromegas detector, a low-pressure operation mode is also possible in order to increase the

energy resolution and therefore the charge identification capabilities. Several TPC prototypes

have been realized and tested at Saclay, showing the tracking capabilities of the detector in

different gas mixture. In the near future, the low-pressure operating mode will be tested in

Saclay and at the Lohengrin spectrometer of ILL (Grenoble). The magnetic field required for

tracking purposes is between 2 and 5 T and a permit uniformity is mandatory within the

whole active area. Therefore, the characteristics of the solenoidal field of SOLAR fully

satisfies the detector requirements and providing a very performing tool to investigate the

fission process.

10

1.2 APPLICATIONS

Medical Application program

1.2.1 PET/MRI and SPECT/MRI There is strong interest in the medical community in combining imaging methodologies such as PET (positron-emission tomography – typically using 18F as a tracer) or SPECT (single-photon emission computed tomography - typically using 99Tcm gamma-decaying tracer) imaging with MRI imaging. This is of high value since MRI provides anatomical detail of the full body or part of the body, while PET and SPECT provide functional information where radioisotope tracers are selectively taken up into diseased organs [Bol09]. Simultaneous imaging is important due to the difficulties of co-registration i.e. the patient inevitably moves between the two successive imaging techniques thus blurring the images. There is also a subsidiary advantage that since for PET the emitted positrons follow helical trajectories in the magnetic field their effective range is reduced which means that the annihilation photons are emitted closer to the position of the decaying atom. Developing PET/MRI systems is very challenging as conventional PET systems rely on photomultiplier tubes (PMTs) for light collection and these do not function in high magnetic field environments. Solutions investigated involving splitting the field in the middle of the magnet, placing tubes in the fringe fields, or replacing the PMTs with novel light collection devices like APDs or silicon photomultipliers. More work has been done on PET/MRI and some prototypes of animal scanners exist [Jud08]. SPECT is a methodology much more commonly used than PET, though the tracer used, 99Tcm can be milked from a 99Mo source and does not need to be produced using a cyclotron. The efficiency of SPECT is much lower, however, since it typically uses a physical collimator to locate the emitting source. The alternative approach is to use electronic collimation – the so-called Compton camera approach. This approach has been investigated by the University of Liverpool and Daresbury NP groups using planar germanium strip detectors in an MRI magnet. The associated project is known as ProSPECTus [Hark11]. Development of new PET isotopes: Radioactive beam facilities such as ISOLDE have the capability to produce high intensities of radioactive isotopes both as low energy (40 keV) and re-accelerated beams. Isotopes are available which could not be practically produced with a standard medical cyclotron. This may motivate testing and development of novel PET or SPECT methodologies. More detail study of the three-photon annihilation will be investigated.

1.2.2 Solenoid The main direction of MRI is presently towards higher fields with 7-T becoming a new standard. The solenoid to be constructed for this project has interesting features in that it exceeds the typical design envelope of an MRI magnet in terms of length of the homogeneous field region by a factor of two, although the homogeneity obtained will be at the 10-4 level. We are investigating at least some portion of the magnet having sufficient homogeneity to allow MRI to be carried out in that volume. A second interesting feature of the solenoid is the capacity to introduce different devices into the center of it through chimneys where there are gaps in the coil construction. A

11

more speculative application would be to perform imaging while a proton or 12C beam was being introduced into a patient. High-energy beams of these species produce PET isotopes in situ like 11C or 12N. These could be imaged and overlaid with MRI images to show exactly where the treatment was taking place. The idea would be to use the chimneys in the solenoid to introduce ion beams and/or gamma beams.

1.3 Plasma Application program The idea to include Plasma physics as a possible user of the facility is a recent idea. We have approached interested parties at York Plasma Institute - http://www.york.ac.uk/physics/ypi/ consider that interests lies particularly in the “astrophysical plasmas” domain. The combined CHYMENE-Solenoid is an attractive combination. Labs other than YORK are interested in the facility.

1.4 Electronics In order to support the very large number (120k) of electronic channels required for SOLAR, the electronics will need to be highly-integrated and make extensive use of ASIC technology. This has interesting implications for transfer to the medical sector. The system being considered will integrate the needs of the Nuclear, Medical and other application requirements through multi-purpose architecture.

12

Section 2: SOLAR – Instrumentation As far as instrumentation is concerned, SOLAR is a facility made up of a magnet and an evolved Data Acquisition system, DAQ. Both of these core elements are conceptually conceived to include multi-functionality. The magnet will house several detector subsystems, linked to the same DAQ. A number of such subsystems will make up the project. The whole assembly will be transportable and is designed to be compatible with beams at ISOLDE/CERN, n-ToF/CERN, SPIRAL2/GANIL, FAIR/GSI and RIKEN. This means that the design will need to take account of the required floor space, integration within other instruments and safety requirements. Where and when possible, the SOLAR system will be made available to conduct experiments, which do not require beam but a radiation, controlled area.

CORE ELEMENTS

2.1 SupHeX We call the component of SOLAR, SupHeX, derived from SUPra HElmoltz for nuclear eXotic nuclei. SupHex differs from a typical solenoid used as the basis of an MRI scanner in several respects. Firstly, the homogeneous volume is double the length of a typical MRI solenoid but is homogeneous only to better than 1 part in a thousand, compared to 10-6 for an MRI solenoid. Secondly, SupHex will have 12 ports in it where various auxiliary instruments and detectors may be inserted. SupHex will have a bore of 1 m and a length of 2.5 m. The magnetic field is generated by six superconducting Helmholtz coils to provide a maximum field of 4 T. A calculated field map is given in figure 2.1 and details are given in ANNEX 1: SupHeX. It should be mentioned that the present plan does not consider active shielding of the magnet to reduce stray field in the surrounding environment. Options to provide shielding are presently under discussion. IRFU will be responsible for the conceptual design, construction and testing of SupHeX. SupHeX is devised to receive ion beams parallel (or close to parallel) and perpendicular to the symmetry axis. Targets can be placed inside and outside the magnetic field volume. In particular, we intend to be able to employ an advanced version of CHYMENE at one of the three possible positions.

13

Figure 2.1: SupHeX magnet calculated field map – see annex. The red boxes are the position of the superconducting coils. The ellipses show the region with field homogeneity better than 10-3.

2.1.1 Mechanics Considerations - SupHeX A requirement of SOLAR is that it should be easily reconfigureable both in terms of orientation and detector subsystems. Further, the transportability of the system will require adequate consideration to be given to issues such as alignment, which will need to be taken care of in the final mechanical design. In particular, the cryostat associated with SupHeX will be built so as to be able to hold vacuum/alignment (SOLAR-HELIOS configuration) in the bore and to be pressurized to a maximum of 2 bars (SOLAR-TPC configuration) by placing two large caps on either end of the magnet.

2.1.2 Beam Trackers RIBs often have poor emittance qualities, so it will be important to track the beam onto the target to obtain kinematic corrections and a measurement of time-of-flight. In the optimum position, the trackers will have to be placed close to the target and so will need to be relatively insensitive to a field perpendicular to the tracker surface. It will be necessary to develop beam trackers which are relatively insensitive to magnetic fields. In particular, solutions for low energy beams based on Micromegas and PPAC will be investigated. High-energy beams present significantly less difficulties. The group at SPhN have strong experience in the development of beam trackers and presently have a running program in this domain [Ott99].

2.1.3 Front-end and DAQ Electronics today is undergoing rapid change. It is therefore essential that we future proof the system and consider longer term developments. Thus within the conception period it will be a playground of ideas so far as developing an architecture that is both novel and makes allowances for future progress. With this in mind, we are envisaging a step-change in electronics design from the historic approach of the nuclear physics where we are accustomed to relatively small dedicated systems, to a system which is far more ambitious and versatile. At the same time, we have to be mindful that the time and engineering support available for this task are limited.

14

The developments associated with the electronic systems for R3B(STFC/UK) and GET(IRFU) [Pol12b] will provide a strong starting point, making it possible to concentrate on specific elements of advanced technology that will lead to the generic character of the electronics we foresee for SOLAR. In particular, the new system will include relatively large dynamic range and saturation-free operation (TPC). The dynamic range of the pre-amplifier/filter and the capacitive array will be software adjusted to allow for coupling to DSSSD, LaB3(APD), Gas ion chambers and micro-channel devices (PET). The GET and R3B frame-works will be enhanced to include a fully comprehensive and all encompassing generic approach. This will be achieved by the choice of telecommunication standards equipment (eg -TCA) and a fully integrated free-ware approach. ANNEX 4: Front-End and DAQ delimits some of the highlights expected in this area. Ten FTC are requested (ERC) to cover the advanced ASIC development and Generic firmware development. The R&D as well as the production workloads associated with this task will be carried by IRFU and STFC.

SUB-SYSTEMS SOLAR will provide three detector sub-systems which will operate within SupHeX. Namely,

1. SOLAR-TPC, 2. SOLAR-HELIOS & SOLAR-GAMMA 3. SOLAR-MEDICAL

Further, to supplement (2) and possibly (1), SOLAR will include a LaBr3 gamma hodoscope. A detailed description of the TPC, HELIOS, SOLAR-MEDICAL and the hodoscope are given in the correspondingly named ANNEXs. For continuity, we give a condensed description below.

2.2.1 SOLAR-TPC Kinematic curves for different nuclear reactions at different incident energies are presented in figure 2.2. It is clear that a broad dynamic range is mandatory to cover the diversity of reactions of interest; also noticeable are the very steep gradients with azimuthal angle. Superimposed on the figure are overlapping colored domains delimiting differing experimental methods. These methods have particle ID, energy and position resolution necessary for spectroscopy studies. Hence, for a given experiment a variety of experimental set-ups are necessary. Not superimposed on this figure are the needs to have wide solid angle cover coupled with a high luminosity of the order of 1020 or higher. The latter may be incompatible once spectroscopy (50 to 300 keV) resolution is requested.

15

Figure 2.2.1 Kinematics for direct reactions overlaid with different experimental method. After many years where it was hardly used, the ‘imaging’ of reactions in a gas (or LH2) is back in vogue. In recent years, there has been extensive work at GANIL, GSI and Texas A&M making use of TPC-like devices for spectroscopy [Dem07] and decay [Bla08] studies. To-date, essentially three ‘imaging’ detector and configurations have emerged and are on the drawing boards:

I. where the gas (H2, D2, 3He or 4He) is used as a target and detection medium, referred to as Active Target, AT,

II. where the gas (eg P10) is used to stop and detect the mother nuclei as well as the subsequent [Pol12a] ,protons or alphas decays, and

III. where a solid target is placed at the entrance of the TPC to the study of nuclear equation of state. This is of course classical and comes directly from particle physics. However other reactions of this type more resemble spectroscopy studies (e.g. AX+208Pb A-1X+p+208Pb to study GMR) are in the making.

By placing 1, 2 or 3 in a uniform magnetic (3-4 T) field (dipole at RIKEN and solenoid at NSCL) the dynamic range and PID is extended considerably AT and TPC detectors have a number of interesting advantages; a. inherently low thresholds. In principle, gas ionization (30eV) is detectable with

some efficiency, thus we propose to employ this technique to perform some very exciting experiments , such as ( ) to study GM excitations in neutron-rich nuclei or to measure low-energy proton or alpha decay spectra.

b. energy, particle ID and angular resolutions are good c. 4 detectors for relatively low recoiling energies d. for a detector of say 100cm in length, the luminosity can be factor of five to ten

higher than solid targets with the same resolution.

16

The number of channels required for such TPCs is no trivial affair by nuclear physics standards. To allow for a measure of the length of the trace (energy), pad densities approaching 25pads/cm2 are required. Hence, 100,000 channels is easily reached. As shown in the ANNEX, the SOLAR-TPC gives a nearly complete coverage of all these requirements and would therefore be an optimum tool to complement the investment in state-of-the-art RIB facilities. SOLAR-TPC and SOLAR-HELIOS provide complementary instrumentation in that the SOLAR-TPC is particularly effective for low intensity beams (< 104 /s) while SOLAR-HELIOS covers intensities above 104 /s. Given that we do not have a complete simulation of SOLAR-TPC that includes the pad structure, amplification, electronics and the data extraction it is difficult to give values of the resolution and efficiencies over the available phase space cover to be covered for reactions of interest. Nevertheless in the figure 2.2.2a and b below we make a comparative simulations for ACTAR-TPC and MUST2 for energies for SPIRAL2 energies. The conclusions are evident as far as the resolutions is concerned. ACTAR-TPC does not reside in a magnet and hence it has a limited gas dynamic range. Thus comparing SOLAR-TPC with GASPARD, for non- coincident gamma ray measurements the SOLAR-TPC is a favorable choice. It is to be stressed that each detector system has it salient points and SOLAR-TPC is of interest once the energy loss of the quasi-target is small and the luminosity is limited.

Figure 2.2.2a ACTAR-TPC simulations of the reaction 78Ni(d,p) at 8Mev/A on D2 at STP (ACTAR-TPC collaboration).

The TPC aspect falls naturally in the development program that the Saclay group has undertaken over the last five years or so through GET (General Electronics for TPCs), AstroBoX and MINOS. The GET program is supported by an ANR (E. Pollacco et al., 2009) and MINOS by an ERC Starting Grant (Obe10). The SOLAR-TPC will be used to perform experiments covering a broad energy range of incident beams for nuclear spectroscopy and equation of state studies. In particular it will be able to operate with hydrogen and helium gases and, hence, to operate in an Active Target mode. On the nuclear spectroscopy side, which is the SPhN group’s interest, we would be able to study direct reactions, in particular (’), (p,2p), (p,np) or (p,) reactions for RIBF/Riken and Fair/GSI beam energies experiments. At SPIRAL2 and HIE-ISOLDE energies, the SOLAR-TPC will be able to cover direct reactions and reactions strongly associated with

17

astrophysics interests. The wide dynamic range that this tool provides is principally due to the large volume, the high field of the solenoid (4-T) and the small size of the anode pads. The anode will be covered with 120,000 channels. The experience gained by SEDI in developing MICROMEGAS [Gio96] devices will be exploited. Existing home-delivered IPs will be extended to include high dynamic range charge amplification. Hence the instrument integration and analysis of the SOLAR-TPC will be assisted via a three-year post-doc request (ERC). IRFU will be responsible for the conceptual design, construction and testing of the SOLAR-TPC. The SOLAR-TPC in its presently conceived geometry will allow the European community to have an instrument compatible with RI beam facilities being built today. TPC instrumentation are today being built in major non-European facilities, namely the SAMURAI-TPC at RIBF and the AT-TPC at FRIB/NSCL/MSU. It is to be noted that GET will provide the electronic system for these instruments and hence we are in the foreground of these initiatives.

Figure 2.2.2b Detector like MUST2 simulations of the reaction 134Sn(d,p) at 5Mev/A [LoISn].

18

2.2.2 SOLAR-HELIOS The HELIOS spectrometer at Argonne National Laboratory has pioneered a new way of studying light-ion transfer reactions with stable and ISOL beams [Lig10,Wuo07]. In such a spectrometer illustrated in figure 2.3.1, the radioactive beam is incident on a thin target e.g. deuterated plastic, at the centre of a large solenoidal magnet of the type used for MRI scanners. Light ions from single-particle transfer reactions follow helical orbits in the magnetic field and are recorded in a compact, linear silicon array along the beam axis. This design of spectrometer has three significant advantages over conventional charged-particle spectroscopy using silicon detectors, beyond the solenoid representing a much simpler and less complex detector system. Firstly, better energy resolution where it is not limited by target energy loss effects. This arises due to the linear function between center-of-mass energy and the measured position and lab energy measured at the axis of the solenoid. Essentially, the lab ion energy resolution is identical to the CM energy resolution. Secondly, the linear nature of this relationship also means that the dispersion different excited states in the ion-energy and Q-value spectra are the same. In the conventional approach using Si at fixed angles, the non-linear relationship between proton energy and angle can mean that when moving from an ion energy spectrum to excitation energy, peaks become compressed by factors of up to three, degrading the effective Q-value resolution. Even for experiments where target energy-loss effects are important in the ion-energy resolution, the resulting Q-value spectrum with a solenoid still benefits from this lack of compression. As an example, a recent d(86Kr,p) measurement achieved an excitation energy resolution of ~70 keV. Some conventional approaches use γ-ray measurements to recover the excitation energy resolution. This necessarily introduces an additional efficiency factor of up to 10% due to the coincidence requirement but this can be mitigated to some extent by using a thicker target. A solenoidal system allows good resolution, sufficient for many purposes, from the measurement of outgoing ions alone and thus avoids the efficiency hit in yield for many experiments.

Figure 2.3.1 Schematic of a solenoidal spectrometer. The beam enters from the left hand side and is incident on a target at the center. Light ions are detected along the solenoid axis. The present system at Argonne National Laboratory is largely dedicated to (d,p) studies and also the suite of radioactive beams is limited. For reactions with outgoing tritons, which are very rigid, the homogeneous volume of the magnet is not long enough and this motivates the present design of the SOLAR solenoid, which has double the length of the homogeneous field volume and a higher maximum field (4-T) which ensures that tritons from e.g. the 96Kr(p,t) reaction can be detected on axis.

19

The silicon array used with HELIOS at Argonne comprises a four-sided box of detectors each with an active area of 50-mm long and 9-mm wide, and is 700-μm thick. Position-sensitivity is

achieved by resistive charge division. Undoubtedly, better resolution could be obtained with a

next generation silicon array using silicon strip technology. Simulations have shown that a 10-3 field homogeneity introduces typical position uncertainty of light ions on axis of less than 0.5 mm. This practical consideration dictates both the necessary homogeneity of the SOLAR solenoid as MRI field precision i.e. 10-6 homogeneity will be swamped by the position uncertainty, as well as a practical dimension for silicon strips of 0.5 mm. In addition, a silicon array for SOLAR would need to be significantly longer to profit from the larger field volume and the variety of the Physics cases.

2.2.3 SOLAR-GAMMA The gamma-ray detector array for SOLAR will strive for high energy resolution, high timing resolution and high granularity. It would be operated in both the “HELIOS” mode and in conjunction with the TPC. Given the diversity of Physics cases and the clear need for flexibility, the gamma-ray array should be highly modular. This favours scintillator technology both in terms of physical size but also achievable efficiency and cost compared to germanium detector arrays, To operate scintillators successfully in a high-field environment, a technology other than conventional photomultiplier tubes will need to be used. The York group has recently been investigating novel photosensors like APDs and SiPMs coupled to scintillators. They have tested prototype detectors in various environments including a 3T MRI magnet at the York Neutroimaging Centre. This work was supported as part of the GANAS programme funded through the NuPNET scheme. We will profit extensively from this work in the present project. Our concept is for novel scintillator detectors like lanthanum bromide coupled to an avalanche photodiode (APD) readout. In this sense, the array would have strong synergies with present directions in combined imaging, generally exceeding what is needed for the medical application. We envisage a modular array of 100 elements of 1” cubic crystals of Lanthanum Bromide each coupled to an APD. Recent tests indicate that an energy resolution better than 6% should be achievable for such a configuration and time resolution of the order of 1 ns.

Figure 2.4: Cartoon showing giving a possible set-up of SOLAR-GAMMA within the SOLAR-HELIOS.

20

2.2.4 SOLAR-CHYMENE SOLAR-CHEMENE will complement the HELIOS detection system by allowing for a thin (30-100 micron) windowless pure hydrogen and deuterium solid target. This will provide an enhanced yield (by a factor of 3) for a given energy resolution and remove contamination due to fusion-evaporation on carbon in solid plastic targets. When operating in HELIOS mode, thin targets are central to achieving the best energy resolutions. SOLAR-CHYMENE is an extension of the present study and will need to be modified in terms of length to support its insertion into the magnet. In addition, careful choice of materials will be necessary given the operation in a high magnetic field environment. R&D tasks will be required to cover these specifications. At the present time, CHYMENE is supported by an ANR Gil12] and results to date are encouraging. Once this program has been completed, a full study to implement CHYMENE towards SOLAR-CHYMENE will be undertaken.

Figure 2.5: CAD drawing of CHYMENE To date we have not considered a polarized H2 target to operate in or close to SOLAR. This possibility will no doubt be invoked at a later stage. The CHYMENE model can be modified to employ other elements and compounds.

2.3 Installing SOLAR at different laboratories SOLAR will be installed in different laboratories with programs covering approximately two to three years. The intended programs cover both nuclear physics and applications. The nuclear physics aspects will be pursued with radioactive beam facilities but there is also scope for experiments without beams such as fission fragment studies with strong sources. It is clear that nuclear physics experiments have evolved strongly since the advent of RI beams. Instrumentation projects for RIB-based experiments must take account of non-ideal beam properties including energy spread, emittance and low intensity. This is compounded with the need to use thin targets when the cross-sections are relatively low. Further, to draw maximum use from the low intensity beam, the general trend is to set-up experiments as fully exclusive as possible. As a result, multi-subsystems are being innovated to cover as many exit channels as possible (gamma, neutron, light charged particle, electrons and projectile-like fragments). Furthermore, to address the beam characteristics, most experimental setups track the incident particles, event-by-event.

7

the cryostat: Ø 250 mm, L 550 mm

Turbo pump with PV

Thermal screen (bottom part)

Nozzle

Bellow with CF 100 flange

Cold head

Extruder driver

CHyMENE Cryostat

Feed through

700 mm Beam axis

900 mm

Ø 90mm

21

Figure 2.6.1: Schematic of a possible setup for SOLAR at a high-energy beam facility.

Figure 2.6.2: Cartoon of possible setups for low-energy beams, in a HELIOS mode with a CHYMENE target. The addition of a recoil separator depends on the Physics case and is not mandatory. In figures 2.6.1 and 2.6.2, we consider possible layouts required for standard experiments at high (>200MeV/A) and low (<20MeV/A) beam energies. The footprint of SOLAR is approx. 5m by 3m. As with most set-ups involving RI beams the beam is tracked several times before reaching and after the target, in the case of a high-energy beam. For low energy, this is quite delicate because of the energy and angular straggling involved. Developments of tracking techniques associated with SOLAR have to be reviewed to take into account the fringing field. Hence, a sub-task for tracking will have to be set-up. Neutron detectors are indispensable tools when dealing with neutron-unbound nuclei produced in knockout reactions. The invariant-mass technique, combined with the use of high energy beams and thick targets, allows to reach the more exotic nuclei. The invariant mass of the system is determined by measuring the four momenta of all decay products in the laboratory frame. This quantity gives information about the invariant mass of the unbound state before the decay, which is equal to the mass of the unbound system in its rest frame. [T.Baumann et al., Rep. Prog. Phys. 75 (2012) 036301 as a reference for invariant/missing mass measurements, both status and techniques] . The neutrons emitted in the decay of the knockout residue have energies between few tens of MeV and hundreds of MeV, depending on the beam energy. Therefore, most RIB facilities today have a large solid angle neutron wall typically made of plastic scintillators

22

(MoNA/NSCL/MSU, NeutronWall/RIBF/RIKEN and LAND/FAIR/GSI). Also associated with neutron walls are wide acceptance magnetic sweepers to allow for charge particle to be removed from the neutron wall acceptance and deflected in the direction of a charged-particles array. Hence it is important to foresee installing SOLAR where such facilities are available. LAND and GLAD at FAIR/GSI and SAMURAÏ at RIBF/RIKEN are examples. It is clear that this combination maximizes the exclusive mode criteria placed in nuclear physics experiments today. For low energy beam facilities at SPIRAL2 and ISOLDE sweepers will be a challenge to deploy however large acceptance neutron detections are in the making. Facilities today have magnetic spectrometers which are tuned to detect specifically the quasi-projectile (SHARAQ or ZeroDegree/RIBF/RIKEN, VAMOS or SPEG/GANIL, S800/NSCL, etc.). It is essential that SOLAR could be coupled with such instruments. The present section focuses the discussion towards experiments in existing and future RIB facilities. However, this does not exclude the possibility to install SOLAR in facilities where gamma, neutron and stable beams can be deployed. Gamma detection with something other than the SOLAR-GAMMA array is likely to be difficult to implement because of the relatively large target-to-detector distance and the strong magnetic field. However the use of AGATA, PARIS or equivalent, for example could be foreseen at the focal plane of detectors of the stopped beam or a forward station displaced with respect to SOLAR. The SOLAR collaboration has not so far sought approval to bring SOLAR to the facilities mentioned above since this would be premature. The initial installation could be foreseen to be in a European laboratory that can provide exclusively stable beams. Studies and adopted agreements will be worked on once the CSTS has been accepted.

2.4 SOLAR-MEDICAL As discussed, there are strong synergies between interests in medical imaging and the present SOLAR project. The gamma-ray hodoscope, SOLAR-GAMMA, can be used to investigate some of the possibilities. We also consider a more ambitious and novel aspect for gamma-ray detection, particularly relevant to PET . PET imaging conventionally uses scintillator detectors connected to photomultiplier tubes. For some prototype PET inserts used with MRI systems, APDs have been used for scintillation detection. An important aspect for such systems is RF-shielding since there are strong RF fields connected with the operation of MRI scanners. As part of SOLAR, we will investigate whether it is possible to introduce gamma-ray emitting in the far UV coupled to high gain gas amplifiers [Gio96]. If we reach sufficiently high efficiency they will yield good position and time resolutions that are very important development for PET. Naturally, it would also function straightforwardly in a high magnetic field. In principle, very fast timing resolution is possible, superior to conventional PMT readout and this could open the door to the investigation of multi-gamma PET using multilateration to determine the point of emission. It is not clear if we need to carry out NMR simultaneously to validate the approach. This is possible since SupHex will have some small localized regions where the field homogeneity is high enough. In this case, shimming and infrastructure would have to be added. To date the contribution to this section by SPhN/IRFU has been limited. However we realize that within the CEA a number of groups should be consulted and possibly integrated in this study. Areas of direct interest are of course in the MRI and applications in development of high efficiency gamma detection (see ANNEX CALIPSO) for example.

23

2.5 SOLAR-TESTBED

To allow the project to do Physics from day one, we will profit from the employment of a 3-T solenoidal magnet recovered “for free” from the University of Nottingham. This magnet has only half the length of homogeneous volume compared to the SOLAR magnet and would not have the facility to incorporate a solid hydrogen target. Its utility would therefore be largely restricted to studies of (d,p) reactions in the “HELIOS” mode, which are undemanding in terms of homogeneous field volume.

It is foreseen that this system (see figure 2.8.1) could be deployed at GANIL in 2014 or HIE-ISOLDE (from 2015) and be superseded by SOLAR when it comes into commission. A simple silicon system is available for loan from Argonne National Laboratory that would allow basic Physics to take place early in the intended program. This removes the gap in the timeline before the Physics program can begin. This system would also serve as a useful test-bed for the other detector subsystems for SOLAR proper as they come to be constructed.

Figure 2.8.1: Mechanical design for testbed solenoidal spectrometer (courtesy of Andrew Smith, University of Manchester)

24

Section 3: ERC Planning, time scales, budget, safety assessment and risk register

3.1 Rationale for submission to the ERC Synergy scheme

The ERC synergy program is intended to support ambitious programs in the sciences which have a strong element of “synergy”. The SOLAR project has synergies at many levels. Firstly, there is a synergy between those wishing to carry out ambitious experimental work and those who have the capacity to produce state-of-the-art equipment to meet the needs of this exciting program. For example, Saclay is a world leader in the design of superconducting magnets – such a device lying at the core of SOLAR. Moreover, the key technical developments in TPC and solid hydrogen target technology at Saclay may be redirected to this present project.

The second area of synergy is between the demands of the cutting-edge of nuclear physics and those of societal applications including, importantly, medical imaging, as well as other areas of physics such as the study of astrophysical plasmas. The third synergy is external and that is the link with the wide community of experimental nuclear physicists since SOLAR will be a device which will on completion be the “property” of the community to be supported at different facilities in the manner of other travelling facilities such as AGATA. An ERC synergy grant should have 2-4 PIs where one is primus inter pares. David Jenkins at the University of York, Emanuel Pollacco from SPhN/IRFU will be PIs. We have approached Andy Boston from University of Liverpool to be a third PI bringing in expertise in medical imaging and accessing a wider community linked more closely to the medical needs. There is scope to add a fourth PI – possibly in the area of astrophysical plasmas. Within this structure, there is license to define who will handle specific tasks. For example, within the UK umbrella, we could also bring in specific expertise from STFC Daresbury. In fact, Marc Labiche from Daresbury has performed the SOLAR-HELIOS simulations. In order to ensure the strong synergy required for this program we are considering that David Jenkins would be affiliated to SPhN/IRFU Saclay for the period of the ERC contract.

3.2 Schedule Figure 3.2 gives an overview of the timeline that we foresee for a six-year programme. In essence, we anticipate a scenario where the non-nuclear physics program can be covered over the whole period through the employment of the testbed facility at CERN or elsewhere. The nuclear physics R&D and realization will allow for a two-year physics experimental program to

perform proof-of-principle physics experiments.

25

Figure 3.2: Projected timeline for SOLAR

3.3 Budget estimation and personnel requests to IRFU The funding request to the ERC will cover all of the instrumentation costs and part of the personnel. A rough breakdown is given in figure 3.3. By far the largest part of the budget goes towards the R&D and building of SupHeX. Hence effort will be directed towards reducing this line as far as possible. The costs for SupHex include all the associated manpower costs. Personnel costs represent the second largest expenditure. The ERC will employ three postdocs for three years at SPhN/IRFU that will cover the physics, simulation, instrumentation and medical-physics tasks. A three-year and two-year CDD will cover the front-end and DAQ requests. An equivalent request will be posted by the UK institutions. Personnel will amount to an estimate total of 2M€. Other elements are discussed in the adjoining text. The total request to the ERC will be of the order of 15M€. The personnel requested in units of FTE is given in the table below. It amounts to a total FTE of 23 FTE . We note that this is a significant demand on the resources of IRFU.

2013 2014 2015 2016 2017 2018

SupHeX

TPC

Helios

Applica ons/Medical

Gamma

Experiments/ISOLDEISOLDE/CERN

FAIR/R3B

SPIRAL2/VAMOS

RIKEN/SAMURAÏ

Labs2015++

CHYMENE2

Test-Bench

26

Figure 3.3: Approximate breakdown of the SOLAR budget

Engineering requirements from IRFU

SupHeX and CHyMENE (SACM) SupHeX 10, FTE CHYMENE 1,5 FTE

Electronics System (Sedi)

FEE 3,0 FTE DAQ 1,5 FTE Dedicated Software 1,5 FTE

TPC (Sedi) Mech design 1,0 FTE Gas Handling 0.4 FTE Detector R&D 3,0 FTE System Integr. 1,0 FTE

Total Engineering Request from IRFU 23 FTE.

3.4 Safety There are a number of safety aspects to be considered. The SupHeX solenoid itself will contain 2000 liters of liquid helium in an appropriate cryostat. Large overhead volume will be needed to cope with the unlikely potential for a magnet quench. The magnet will also have a strong stray field, up to 20 gauss at a distance of 10 m that will have to be accommodated unless active shielding is included into the design. Safe working can be achieved by marking appropriate safety lines. Very close to the magnet there is risk from “missile hazard” which may imply the need for a physical barrier.

0-1M€ 1-2M€ 2-3M€ 3-4M€ 4-5M€ 5-6M€

SupHeX

TPC

Helios

Medical

g

Labs2015++

T

Personnel

Test-Bench

27

The SOLAR TPC/Helios has particular hazards. The active target can operate with 2 bar of hydrogen gas. The TPC will have a maximum voltage of 50 kV and will need to be surrounded by 2 bar of SF6 for insulation purposes. CHYMENE employs hydrogen and hence subject to security requirements. Gas at high pressure and use explosive gases are a hazard which our instrumentation groups often tackled in the past. Other safety issues are generic to such experiments and include the potential for fire in the electronics vault, for example or suffocation due to release of large volumes of SF6 or He.

3.5 RISK Scenario At this stage we have not conducted a risk-factorization analysis of the program. However, the risk is mitigated to some extent in that we will capitalize from instrumentation developments which have been, by-and-large, proven to be successfully operational and where we understand the exclusive risks involved. Nonetheless, there are factors that remain to be investigated, as denoted in the different ANNEXs. We raise the following as potential break- through which will have to be achieved;

I. development of a magnet within a reasonable budget that will has a relatively small fringe field to allow it to operate at different labs

II. development of the magnet to contain a large enough uniform volume that will allow MRI to be conducted (B/B<10-5)

III. development of electronics that is immune to high frequency perturbation from the MRI.

IV. development of a large dynamic range gas amplifier for SOLAR-TPC

4 RESUME In resume we note that that this program has an extensive scientific Nuclear Physics program achievable in a number of European laboratories. The program includes other domains of physics, in particular Medical Physics. Hence it corresponds to a synergy imitative. With regards the instrumentation it is ambitions in that it extends well beyond what the community of Nuclear Physics in general has been attached to in recent years. SOLAR will designed to ease the limiting constrains in developing complex instrumentation. The development of SupHeX and the data capture system stands to be paragons in the domain.

28

References [Lig10] J. C. Lighthall et al., NIM A 622, 97 (2010). [Wuo07] A.H. Wuosmaa et al., NIM A 580, 1290 (2007). [LoISn] SPIRAL2 letter of Intent, LoI_SP2_Ph2_33, V. Lapoux et al. [LoIKr] SPIRAL2 letter of Intent, LoI_SP2_Ph2_32, E. Clément et al. [LoIActar] SPIRAL2 letter of Intent, LoI_SP2_Ph2, G. Grinyer, E. Pollacco et al. [Bla02] J. Blachot, Nucl. Data Sheets 97, 593 (2002). [Bjo86] T. Bjornstad et al., Nucl. Phys. A 453, 463 (1986). [Fog95] B. Fogelberg, et al., Phys. Rev. Lett. 73, 2413 (1994). [Omt95] J.P. Omtvedt et al., Phys. Rev. Lett. 75, 3090 (1995). [Jon10] K.L. Jones, K. L. et al. Nature 465, 454–457 (2010). [Duv83] P. D. Duval, D. Goutte, M. Vergnes, Phys. Lett. B 124, 297 (1983). [Bor77] M. Borsaru et al., Nucl. Phys. A 284, 379 (1977). [Ard78] D. Ardouin et al., Phys. Rev. C 18, 2739 (1978). [Cle07] E.Clément et al., Phys. Rev. C 75, 054313 (2007). [Fis00] S.M. Fischer et al., Phys. Rev. Lett. 84, 4064 (2000). [Fis01] S.M. Fischer et al., Phys. Rev. Lett. 87, 132501 (2001). [Fis03] S.M. Fischer et al., Phys. Rev. C 67, 064318 (2003). [Pet02] A. Petrovici, K. W. Schmidt and A. Faessler, Nucl. Phys. A 710, 246 (2001). [Lal99] Lalazissis et al., Nucl. Data Tables 71, 1 (1999). [Naz85] W. Nazarewicz et al., Nucl. Phys. A 435, 397 (1985). [Elb69] B. Elbeck and P.O. Tjom, Adv. in Nucl. Phys. 3, 259 (1969). [Jud08] M.S. Judenhofer et al., Nature Medicine, 14, 459 (2008). [Hark11] L.Harkness et al. NIM A638 (2011)67–73; [Bol09] N.E. Bolus et al., J. Nucl. Med. Tech. 37, 63 (2009). [Ada09] R.W. Adams et al., Science 323, 1708 (2009). [Rae09] Electron avalanches and breakdown in gases. Heinz Raether, Butterworths 1964; The Mechanism of the Electric Spark By Leonard Benedict Loeb, John M. Meek. Stanford University Press, 1941; High Voltage Engineering; M S Naidu, V Kamarju. Tata McGraw-Hill Education, 2009

[Riegler]: “Particle Detection with Drift Chambers” , Walter Blum · Werner Riegler and Luigi Rolandi, ISBN: 978-3-540-76683-4 [Gil12] A. Gillibert et al., accepted for publ. EPJA (2012), ANR-CHYMENE Funding [Obe11a] A. Obertelli and T. Uesaka Eur. J A 47, 105 (2011) [Obe10] A. Obertelli, ERC Starting Grant MINOS-258567 [Ott99] S. Ottini-Hustache et al., NIM A431 (1999) 476.

[Pan12] J.Pancin et al., JINST 7, P01006 (2012).

[Dem07] C.E. Demonchy et al, NIM A573(2007)145

T.Roger et al. Nucl. Instrum. Meth. Phys. Res. A 638, 134 (2011)

M.Camaano et al. Phys. Rev. C 78, 044001 (2008),

I.Tanihata et al. Phys. Rev. Lett. 100, 192502 (2008)

[Bla08] B. Blank et al., NIM B266(2008)4606

[Gio96] Y. Giomataris et al., NIM A376(1996)29, T.Zerguerras et al., NIM A608(2009)397.

[Sau97] F. Sauli, NIM A386(1997)531

[Dai] Daisuke Suzukiet al., www.nscl.msu.edu/exp/sr/attpc

[Pol12a] E.C. Pollacco, et al., to be published NIM

[Pol05] E.C. Pollacco, et al., Eur. Phys. J. A 25, s01, 287–288 (2005),

[Wal07] M.S. Wallace et al., NIM A 583(2007)302

[Pol12b] E.C. Pollacco et al., TIPP proceedings – Chicago 2011 GET, ANR-GET Funding

[Del12] E. Delagnes, Private communication (2012).

29

ANNEX 1: SupHeX (Antoine Dael et al.,)

BASIC DESIGN OF THE MAGNET The function of SOLAR is to characterize particles created by the interaction of the beam with the target. These secondary particles are curved helicoidally and follow large trajectories inside the magnet. The design guidelines have been the following:

1. To optimize a set of solenoids according to a requirement on field homogeneity of 5. 10-3 in two ellipsoids of 1. m in length and 0.8 m in width

2. To compute the trajectories in the optimized field map to evaluate the resulting errors both in position and in time.

Optimization of the magnetic configuration: The present basic requirements for the SOLAR magnet are summarized hereafter:

- No active magnetic shielding - Engineering current density : 75 A/mm2 - Inner winding diameter Ri=560 mm - |B/B|<5.10-3 in a volume defined by 2 ellipsoids with dimensions 2a=1000

mm and 2b = 800 mm :

- Forbidden areas: z=z0±75mm where z0 = 0. and z0=±900±100 mm

These requirements are implemented in a very sophisticated set of software developed by Pr Guy Aubert in the frame of scientific developments for high field and MRI magnets. This set of software consists in the analytical calculation of solenoid field everywhere in the space and in different steps of coil optimization using non linear optimization methods. This very powerful tool has led to the possibility to accommodate both a forbidden section in the center for the target and two additional forbidden sections in the ends, also for the target. The reference design is presently with 6 coils presented below (New geometries will be studied including an active shield).

Computation of the trajectories: The total charge of the particles is negligible and the trajectories can be transported independently by a step by step integration with a classical algorithm implemented in Mathematica. The magnetic field map is established analytically and therefore with a very high precision in a grid of 3 m x 0.56 m with a point every 2 mm (421781 points). The magnetic field components are interpolated along the trajectory at each time step. The position of each trajectory is computed when they are crossing the longitudinal axis. This position is then compared with the value obtained theoretically. Three cases have been studied:

- d(132Sn,p)133Sn at 10MeV/u with B= 1.0 tesla - p(32Mg,t)30Mg at 10MeV/u with B=4.0 teslas - t(50Ca,p)52Ca at 10MeV/u with B=2.0 teslas

After compromising between the time step and the duration of the simulation the chosen time steps are the following: 0.5 fs for Sn and 0.3 fs for Mg & Ca.

30

The correlation between the kinetic energy and the emission angle is given below. The squares correspond to points which are either lost on the solenoid coils or with an error on the interception z greater than one millimeter.

The error on the interception z coordinate on these three examples of particles is small enough to be totally acceptable (see figure below).

Design of the magnet The superconducting magnet is made of 6 solenoids distributed along the z axis as shown in figure below. This distribution allows a radial access, both in horizontal and vertical directions, in three planes at z = 0, z= 1.132 m and z = -1.132 m.

31

Figure Magnetic configuration and magnetic field map. Coils are indicated in red

The distance between coils is such that a cylindrical access with a diameter of 100 mm is possible. The inner radius of each coil is 0.56 m which gives enough space for mechanical support, thermal screens, multilayer insulation and vacuum vessel and leaves a free inner diameter of 1 m. The maximum outer diameter of the coils is about 1.4 m. The outer diameter of each coil is chosen to achieve the magnetic field homogeneity. All six coils are powered in series by a current supply connected in “driving mode”. Its stability should be about 1 ppm in order to ensure the temporal stability of the magnetic field. Depending of the experiment carried out, the current in the magnet can be varied. This current change will be possible within about 20 minutes. The main parameters of the magnet are:

coils inner free diameter 1120 mm Coils outer diameter (max) 1400 mm Engineering current density 75 A/mm2 Current at nominal field 500 A (1) Nominal Central field 4. T Peak field on conductors 6.8 T Stored energy 28 MJ

(1) This arbitrary value could be increased up to 1000 A, depending on the final

superconductor, which should be available on the shelf

The cooling of the superconducting coils is ensured by outer aluminum shells which are in turn cooled by a thermo-siphon. The cooling shells in aluminum shrink more than the coils. They are then used to pre-stress the coils and to sustain the radial magnetic forces acting on the coils (which are mainly exploding forces). The axial magnetic forces, which tend to push the coils toward the magnet center, are taken by the support structure which links coils together and which can be placed in the angular space left free by the horizontal and vertical radial access.

32

Cryogenic design

Magnet cold mass: The cold mass is cooled with a serial of aluminum heat exchangers welded on the external diameter of the cylinder. The cooling fluid is a two-phase flow of 4.2 K Gas/Liquid Helium at 1 bars which is moved thanks to the well-known thermo-syphon loop principle (see CMS, ALEPH, ATLAS CS). This pre-filled closed loop also contains a phase separator located outside the magnet cryostat in a dedicated small cryostat (cryogenic satellite) where helium coming from the magnet cryostat is re-condensed thanks to the cold power available on the second stages (4.2 K) of two dedicated cryocoolers (type SHI TDK 415).

Thermal shielding: The thermal shielding of the cryostat is realized with two levels of temperature. The two successive aluminum shields are cooled by conduction respectively around 77 K and 20 K thanks to one or two “shield coolers” (type SHI CH100) installed on the upper sides of the magnet cryostat at a distance where the stray field is low enough. It is mainly the temperature homogeneity of the shields (in particular the “20 K” one) which can lead to the installation of two shield coolers. The first “77 K” aluminum thermal shield is also equipped with a heat exchanger able to be supplied with liquid nitrogen for the first cooling down from 300 K and also in case of long stand-by without the shield coolers under operation (maintenance period). The LN2 inlet and GN2 outlet are installed in the cryogenic satellite.

Shield cooler 1

110 W @ 77 K

6 W @ 20 K

Re-condensers 4 K

35 W @ 50 K

1.5 W @ 4.2 K

X 2?

Shield cooler 2

110 W @ 77 K

6 W @ 20 K

I (A)

Phase separator LHe/GHe

LN2/GN2 Spare circuit for cold stand-by

Current leads (Cu/HTc)

Power supply

LHe supply line

Cryogenic and electrical satellite

Thermal shield “77 K”

Thermal shield “20 K” LHe/GHe collecting pipe

LHe distributing pipe

Preliminary design of cryogenics

for SOLAR 22/10/2012 PhB

Sc coils

Magnet cryostat (without H2 target)

33

Electrical current leads and superconducting bus bars: The two electrical current leads needed for the “driving mode” current supplying the coils are installed in the cryogenic satellite. These two current leads are based on the well-known hybrid configuration including a resistive copper length, between 50 and 300 K, and a High Temperature Superconducting (HTS) between 50 and 4.2 K with a low heat leak at this temperature. If needed, a part of the resistive copper length could be cooled using the LN2 circuit. The bus bars between the feet of the current leads and the coils will be made of NbTi conductor and cooled along the thermo-syphon branches. If the second stages of the cryo-coolers, installed in the satellite (in a low magnetic stray field), are used for the re-condensing of helium, the first stages will be used both for a unique thermal shielding of the satellite and for the shield of transfer line between satellite and magnet cryostat.

Tie rods suspension (or equivalent) : Between magnet cryostat vacuum vessel and cold mass, eight tie rods are used for the cold mass suspension. Each tie rods (G10 or TA6V) will be thermalized on the two thermal shield levels (77 and 20 K).

Cooling-down: From 300 K, the cooling down will be done by using LN2 in the “77K” thermal shield and also by using LN2 in the thermo-syphon loop above 90 K (specific ports installed on the satellite). Below this level of temperature, the N2 will be purged by pumping and replaced by LHe coming from a temporary Dewar (around 1000 litres). When the 4 K level rising and the thermo-syphon priming are performed, the phase separator will be filled with LHe and definitively sealed by starting the 4K cryo-coolers.

Maintenance of cryocoolers : Each cryo-cooler (thermal shield and re-condensers) will be equipped with a tight sleeve permitting a temporary removal of the cold head without breaking the vacuum. In these periods, the magnet will be kept cold at low temperature (< 80 K) thanks to the LN2 circuit. Heat loads: A first estimation is summarized in the following table.

Temperature 77 K 20 K 50 K

(satellite) 4 K

Radiation 100 6 20 0.8 Supports 2 1 3 0.4

Current leads tbd - 75 0.5 Heat loads (W) 102 7 98 1.7

Accommodation of the slush H2 target The magnet and the cryostat are designed to allow accommodation of the slush target in three locations: z0 = 0. and z0=±900±100 mm

34

Additional target cryostat

The removable H2 target cryostat (to be installed in one of the three upper ports) uses an additional cryo-cooler able to maintain a thermo-syphon loop around 11K for realizing the operation conditions for extruding the H2 slush. The additional cryostat has a 100 mm outer diameter rod containing the extruding system and, thanks to the thermo-syphon principle, the cryo-cooler is sufficiently deported from the high magnetic field area.

Utilities: Electrical 2x 9 kW + 2 x 8 kW for cryocooler compressors (+ 1 x 9 kW for H2 target) Water cooling for compressors 2 x 7 + 2 x 9 l/min at < 28°C (+ 1 x 7 l/min for H2 target) LN2 : maximum 20 l/h (for cooling down)

Shield cooler 1

110 W @ 77 K

6 W @ 20 K

Re-condensers 4 K

35 W @ 50 K

1.5 W @ 4.2 K

X 2?

Shield cooler 2

110 W @ 77 K

6 W @ 20 K

I (A)

Phase separator LHe/GHe

LN2/GN2 Spare circuit for cold stand-by

Current leads (Cu/HTc)

Power supply

LHe supply line

Cryogenic and electrical satellite

Thermal shield “77 K”

Thermal shield “20 K” LHe/GHe collecting pipe

LHe distributing pipe

Preliminary design of cryogenics

for SOLAR 22/10/2012 PhB-JMG

H2 nozzle

Extruder

“11 K” GHe

thermosiphon

for H2 extruder

30 W @ 50 K

15 W @ 11 K

Sc coils

Magnet cryostat with H2 target

35

ANNEX 2: SOLAR-TPC

Gas Amplifiers A detailed description of the functioning is not given as there is extensive literature on the subject. Briefly, gas amplifiers consist of a ‘PPAC’ with a plate separation of 50-100m but having one plane highly transparent to the primary electrons. The gains are high (104 and more). This allows for highly pixelated ‘imaging’ for low density tracks (minimum ionization particles) and even single electrons. Large areas can now made (>0,2m2) at an affordable budget. State-of-the-art TPCs-like devices in particle physics employ MICROMEGAS [Gio96] and/or GEM [Sau97] gas amplifiers. As shown in the references, the gains are high and the charge and position resolutions reached are indeed very good.

For RIB-induced reaction or decay studies, the ionization per unit length in a gas can be very different between the quasi-target and quasi-beam, thus the amplifier has to cope with large dynamic ranges. Simply, but not realistically, it can be as high as 1:10,000. The Raether limit (1.6pC/mm) (ref Rae09) could be reached rather rapidly, resulting in sparking/non-linearities for high Z nuclei. Therefore, to render AT and TPC devices all-inclusive for Light Charged Particles and quasi-projectiles, we will have to overcome a number of technical difficulties. Work in this domain is being done at GANIL (Ref. 6) and NSCL(Ref. 7), IRFU and CERN (Rui de Oliviera et al. ) by employing the modification of the micro-pattern gas amplifier structure or adopting a ‘mechanic’ ploy to protect the amplifier. Further, electronic front-end have to be adopted to Z2 increase in dynamic range and pre-amp blocking. These aspects are a source of interesting and provide an extensive playground for development. IRFU/SEDI have been in the forefront of Micromegas development and collaborate strongly with CERN and specifically with GANIL with respect to the dynamic range problematics. Thus our present project will be situated at the heart of these initiatives. Significant R&D is today being done at IRFU (MINOS) GANIL (ACTAR-TPC) to allow pad density of the order of 25 pads/cm2. Thus solutions to allow low capacitive pad-to-electronics are in the making. SOLAR- TPC will extend these technical achievements to cover of the order of 120,000 channels and will most likely consider direct pre-amplifier+filters+ADC at the pad level. Such ambitions, recently achievable (eg MINOS or ACTAR-TPC developments) will be studied and applied such that they can be reproduced within a generic approach and manufactured within industrial context. Final pad sizes, shapes and their layout on the Anode has not been fixed. To take advantage of the very good position resolution of Micromegas concept, random shape as well as non-standard shapes will be simulated. This will allow for a reduced number of channels and avoid pad-alignment effects. Field Cage and Laser Calibration. The field cage which couples the Cathode and Anode is not necessary a simple affair. Models based on AT-TPC [Dai] and MINOS will be considered. The tensions that we stand to operate are relatively high and reach values of 50KV. Thus the cage will be inserted in SupHeX as for AT-TPC (see figure A2.1) and will sit in an isolating gas of SF6. The orientation of SOLAR-TPC with respect to the incident beam is to allow for the track of the quasi-projectile to lie on several pads. Alternative solutions will no doubt be looked into.

36

Figure A2.1: AT-TPC development at NSCL/MSU. Operational in 2014. Typically laser beams are used to calibrate and monitor the functionality of the TPC. A standard system will be employed for SOLAR-TPC.

SOLAR-TPC simulations (Bart Bruyneel, SPhN/IRFU, CEA Saclay) To make an estimate of the dynamic range of the TPC, we will consider the case of protons in Hydrogen gas at different pressures of 0.2, 1.0 and 1.7 bar corresponding to the minimum, the nominal and the maximum pressure for operating the TPC. In figure A2.2 the stopping power is shown for protons as function of energy. The limit upon which the traces can be detected is depending on the gain of the TPC and is shown in the figure by red dashed lines. The detection limit is formed by the condition

(

)

which states that the electron signal at the readout plane should be a minimum of five times the electronic noise (ENC = 2000) in order to be detected. The average energy W needed for the creation of one ionization electron in H2 gas is 36.4 eV and on the order of 26 eV in Ar gas, depending on the quencher admixture [Riegler]. Gains of the order of 103 are easily obtainable using bulk micromegas without sparking (a sparking rate of <1/hr). Higher gains can also be achieved, although sparking can become of a problem. A recent solution to this problem can be the use of resistive layers on the readout plane, which render the spark effects negligible. Operating the TPC at

37

gains of 104 will allow for single electron detection, and for the detection of fast protons up to 1GeV.

Fig. A2.2: Stopping power of protons in H2 gas at different operating pressures. Dynamic range estimations can be made by looking at the proton range as function of energy. This is shown in figure A2.3. Without the magnetic field, the dynamic range would be simply given by the dimensions of the TPC. In the z-direction, one could stop up to 5 MeV protons, while in radius, one would be limited up to 3 MeV for the highest pressures. At low energies, one is limited by the fact that the traces need to be distinguishable from the primary beam trajectories. Therefore, assuming minimum 5 pads need to be hit in order to identify the proton trace, one obtains lower energy thresholds of 50keV for the lowest pressures up to 300keV for the highest pressures. If a 4-T magnetic field is used, the upper limits of the dynamic range can be pushed extensively. In fact, for protons with energies above 2 MeV perpendicular to B, the protons trajectory will be near spherical, loosing only a small fraction of its energy per completed circle. The effective range of the particles at elevated energies and perpendicular to the B-field is therefore given by two times the radius of curvature of the particle. This effective range is shown in figure A2.3 by the black curve. Using the magnetic field, energies up to 30 MeV perpendicular to B can be contained entirely within the TPC. This limit is independent of the pressure and is a minimum of one order of magnitude larger than what was possible without B-field.

38

Fig. A2.3: Range of protons in H2 gas in the absence of a magnetic field. In black: the range of protons in the x,y plane in a 4-T field and in absence of the gas. However for a TPC operated within a magnetic field, the particle does not need to be stopped inside the active volume of the TPC in order to determine its energy. This is contradictory to TPCs operated without magnetic field, where the energy is determined from the track length. Using a magnetic field, the initial energy can also be determined from the curvature of the track. An estimate to how far it would be possible to push the dynamic range can be made using the sagitta method. This method is illustrated in figure A2.4. In the high energy limit, the trajectories in x and y will describe nearly perfect circles. Let us suppose that we can only observe a section of length l of the trajectory; l will thus be of order of the TPC radius: l = 40 cm. The bending of the trajectory can then be expressed by measuring the sagittal, s. The radius of curvature R and hence the momentum in the x,y direction can then be determined from

In the limit of R>>s, and assuming that the error of R will be entirely due to the error in s, one observes from this that the relative error in R will equal the relative error in s:

δR = δs Let us assume now that we want to determine the momentum px,y with a resolution of minimum 1%. Since the position resolution of the TPC is 0.6 mm, we will need measuring a minimum sagitta of smin=6cm. This corresponds to a radius of a 81MeV proton. In reality, however, one determines s from a fit through the trajectory. Note that for l=40cm, more than 200 pads will be hit. One expects from fitting that the error on s

39

will reduce proportional to the square of the number of pads hit. Assuming that Δs through fitting is ten times less than the position, one could measure 3-GeV protons, provided the detection limit is not reached. Knowing the momentum perpendicular to the field also allows the reconstruction of the momentum in the z-direction from the measureable angle of the momentum. A dynamic range up to 1 GeV should therefore be feasible for the high pressure modes combined with the highest gains.

Fig. A2.4: The sagitta method allows the determination of the radius of curvature, R, from the observation of the sagittal, s, of only a small section, l, of the circle. From these simple estimates we note that dynamic ranges reachable by SOLAR-TPC are quite wide covering 0.3MeV to 3000MeV. Namely four orders of magnitude. These estimates will be certified by a full simulations that include the pad structure, amplification and electronics. The calculations will also be able to delimit the efficiencies. To be noted that the present estimates would imply that through SOLAR-TPC we would be able to cover the direct reactions and resonant measurements at SPIRA2/HISOLDE and measurements like (p,2p), (p,np) and (p,p) at GSI/RIKEN. In the latter we expect reactions we will employ the high angular resolution (not discussed here) and the medium resolution energy resolution (<3%) to perform knock-out type spectroscopy.

40

ANNEX 3 – Simulations SOLAR-HELIOS (Marc Labiche, STFC UK) A Monte-Carlo Simulation and Analysis framework for SOLAR A simulation package using the NPTool framework has already been developed for SOLAR at Daresbury Laboratory and validated. The NPTOOL framework is particularly suitable for nucleon-transfer reactions into bound or unbound states and elastic or inelastic scattering. It has been developed originally at the IPN Orsay and is built on the GEANT4 Monte-Carlo transport code and the ROOT analysis package. The framework is user-friendly with an executable file that only requires two independent input files. One is an input file for the reaction to be simulated and contains the following parameters: ion projectile, ion target, target-like fragment, projectile like fragment, beam properties, excitation energy for recoil and a reaction angular cross-section file name; and the other input file is for defining the geometry of the setup including: the target characteristics, the magnetic field, the detector type and the detector positions. Multi-detector setup can also be defined in the geometry input file. The user can choose the option to track all the reaction products or one individual fragment during a run. The output file is a ROOT tree containing a branch with some of the input information and, for each type of detector defined in the setup input file, a corresponding branch with the measured observables. The reaction kinematics is calculated at the beginning of each event by choosing randomly an interaction point in the target and taking into account of the beam emittance (beam energy and angular spread) and the beam energy loss in the target up to the interaction point. Only the reaction angular cross section file must be produced from an external reaction mechanism model such as a Distorted Wave Born Approximation model. To show the quality of the simulations possible, we consider the case of 136Xe(d,p) which was studied by Kay et al. using the HELIOS spectrometer at Argonne National Laboratory. Figure A3.1 shows the experimental spectrum obtained in this measurement. Figure A3.2 shows a visualization of the proton-tracks in a simulation of this measurement, while figure A3.3 shows a simulation of the excitation energy spectrum.

41

Figure A3.1: Experimental measurement taken from [PRC 84 024325 (2011)]

Figure A3.2: SOLAR-Helios setup as defined in the simulation package and proton trajectories for 100 events in the reaction 136Xe(d,p) at 10 MeV/u

42

Figure A3.3: Simulation of the 136Xe(d,p) reaction in the same conditions as the experimental measurement in [PRC 84 024325 (2011)] but considering the same flat angular cross-section distribution for all main excited states (up to the 2.61 MeV excited state)

The HELIOS simulation package will be developed further towards a more complete simulation and analysis framework. The aim is to provide a framework for SOLAR that will allow both real and simulated data to be analyzed using the same tools. The implication is for the simulation output data file to resemble raw data files provided by the DAQ system. The simulation package will also be completed with the two other detection systems of SOLAR, the TPC and the PET-NMR, as well as the gamma hodoscope used to complement HELIOS and TPC. The SOLAR magnetic field map (either measured or calculated) will be introduced in the simulation as applied to the different detection systems. Finally, the development toward a time based simulation system will also be considered in order to simulate realistic data stream. This system will allow the simulated data to be time stamped, for instance according to the beam time structure, and will facilitate the merging of simulated background data with simulated data of reactions of interest. Angular and energy distributions of charged particles calculated using the PACE algorithm will be used as input for the simulation while the fission reaction channel will be simulated using the appropriate physics list already available within the GEANT4 distribution (ex: QGSP_INCL_ABLA). For medical applications, the STFC Daresbury Laboratory already runs the GATE simulation toolkit. GATE has been developed for PET and SPECT applications and is also a GEANT4-based framework. The developments made within the NPTool framework should therefore be transferrable to the GATE toolkit. We will investigate the implementation of a system like SOLAR into GATE in order to facilitate the design of new medical imaging devices and the development or assessment of image reconstruction algorithms and correction techniques.

43

ANNEX 4: SOLAR-Front-end and DAQ In figures A4.1 and A4.2 we present overviews of the GET [Pol12b] and R3B Si tracker electronics. The systems are both driven by specific mixed-mode analogue/digital ASICs. The ASICs are assigned to perform signal-to-information extraction (Charge, Time, Shape or Position) with internal (R3B) or external (GET) ADCs. To allow a widely versatile application the pre-amplifiers and filter can be outside and/or inside the chip through an external chip lying close to the detector. The internal shaping filters (R3B and AGET) are programmable over a wide variety of shaping times for different applications. Sampling speed (AGET) can also be programmed. Time-stamping lies at the mixed-mode ASIC level because it comes from the discriminators in the analogue part of the ASICs. Digital time stamp data is inserted by the ASIC alongside the energy and position data to attain trigger-less configurations. GET by definition has a common or quasi-common dead time operation because of the TPC function. The four-level trigger (MUTANT) is fully numeric and with L3 is based on pattern recognition of the event. Hence in principle all the pads can take part in the event selection before data processing. The R3B Si tracker may be operated free running with all data passed to a software trigger process but it also has the facility to require a hardware coincidence (in the ASIC) or timestamp coincidence in the readout FPGA. Typically the coincidence requirement would come from other parts of R3B electronics such as the CALIFA calorimeter which surrounds the tracker. The post-ASIC multiplexing-concentration tasks are assigned to FPGA . The FPGA handles the configuration and data reduction and formatting processes and resides in telecommunication compatible off-the-shelf equipment (-TCA). Hence economically very attractive with optimum throughput and market standardized. GET uses home built -TCA units while R3B has opted for commercial units which have FMC slots where home-built interfaces using the FMC standard are mounted to connect the R3B ASIC data stream to the processing FPGA on the -TCA motherboard.

44

Figure A4.1: Hardware and Data flow for GET.

Figure A4.2: Hardware and Data flow for R3B Si tracker.

High through-put 10Gb/s

Full Numeric Trigger

Selective Readout Zero Suppress

Base-Line Correction Time Stamp

Automated Calibration

Beam

MeasureQ(t),X,YperPadSamplingADC

45

The two groups have a significant experience in developing systems for the nuclear physics community and the division and sharing of labor is relatively easy and highly profitable. This task will profit significantly from investment already made in present developments and will allow advanced front-end and DAQ to be reached within the time available for the SOLAR-Front-end and DAQ.

Microelectronics for Nuclear Physics TPC. (Eric Delagnes, Irfu, CEA Saclay) The readout of large highly segmented TPCs is requiring a very dense electronics so that frontend multichannel custom integrated circuits (ASICS) are widely used to read them. For High Energy Physics, several ASICs have been designed specifically for this purpose [1], [2],[3] integrating 16 to 72 channels. Usually they are integrating charge sensitive amplifiers followed by filters optimizing the signal to noise ratio which output waveform is sampled and digitized inside or outside the chip. To minimize the number of interconnections, the channel outputs are multiplexed on one of few pins before to be sent towards either a DAQ FPGA or an external ADC. For each pad, the knowledge of the pulse waveform is necessary for the reconstruction of tracks with low angle with the drift field in the TPC. The charge can be measured by the peak value or the integral of the shaped signal and the timing - giving the third coordinate of the track voxel - can be calculated easily using digital timing techniques (digital LE discriminator, dCFD, peak interpolation). The detector we propose will be embedded in a large modulated magnetic field. To minimize effects related to induction on cables; we propose to reduce the number and length of all the connections. For this reason, the front end has to be directly plugged on the detector –without any interface cable- and output data, coming from the frontend electronics, highly multiplexed, need to be digital and sent on only few cables. To achieve the required density a minimum of 64 channels must be integrated on the front-end chip. Such a front-end electronics [4] with a density of ~8 channels/cm2, thus corresponding to the need of SOLAR, but optimized for High Energy Physics and for low rate applications has already been successfully designed by our laboratory. Nuclear Physics experiments have special constraints: on one hand the charged generated inside the detector by some particles of interest may be very tiny so that the electronics must have a low noise front-end part. On the other hand, heavy ions studied during Nuclear Physics experiments can deposit a large amount of charge so that the dynamic range to measure can cover more than 4 decades, which is more than one order of magnitude larger than what is usual for TPC used in High Energy Physics. In some recent chips, [3], it is possible to program a gain for each electronic channel so that the dynamic range can be statically adapted to zones of the detectors with higher charge depositions (beam), but it requires an a priori knowledge of these zones. New solution using dynamically changing gains [5] can be used to solve this issue. At last, the modern microelectronics technologies are using decreasing power supply voltages which are also reducing the available dynamic range. Based on our former experience [2], [3],[4] we propose:

- To design a new 64 low noise channel chip, with self-triggering capability, zero

suppression, multiplexing and on-chip digitization, providing the digitized

waveform for each channel. It will be fully programmable as [3] and compatible

with the full dynamic range of a Nuclear Physics TPC. It will be designed on a

0.13-0.18µm CMOS technology depending on result obtained on former

prototypes with these technologies.

- To design front-end electronics boards achieving the required density and

compatible with the high magnetic field.

References

46

[1] Bosch R.E.,” The ALTRO chip: a 16-channel A/D converter and digital processor for gas detectors “IEEE Trans. Nucl. Sci., ., vol. 50, Issue6, part 2, pp 2460-2469, 2003. [2] P. Baron et al., “AFTER, an ASIC for the Readout of the Large T2K Time Projection Chambers”, IEEE Trans. Nucl. Sci., vol. 55, no. 3, pp. 1744-1752, June 2008. [3]P. Baron et al., “AGET, the GET front-end ASIC, for the readout of the Time Projection Chambers used in nuclear physic experiments “,in Proc. IEEE Nuclear Science Symposium and Medical Imaging Conference (NSS/MIC), 2011 IEEE. [4] D. Attié, D. Calvet et al, “Micromegas-based Large Time Projection Chamber Prototype for the International Linear Collider” in Proc. IEEE 18th Real-Time Conference, Berkeley, USA, 11-15 June 2012. [5] Ph. Vallerand, E. Delagnes, “ Floating Point Charge Sensitive Amplifier for the focal plane of S3”, ANSIP 2011 Workshop.

Firmware: Developments beyond GET (Shebli Anvar, Irfu, CEA Saclay) This document discusses possible future software developments beyond those of the GET (General Electronics for TPCs) project. The GET project has delivered a number of electronic, firmware and software modules that address many issues of genericity and reuse encountered in today’s typical nuclear physics experiments (GET activity reports). The software modules include layers pertaining both to the control and configuration of electronic systems and the actual acquisition, processing, monitoring and storage of the data produced the AGET Asic over TPC detectors, although some layers have reached a more advanced development stage than others. The management and control of the data flow produced by the detector, based on the Narval framework, implements a multi-process architecture that addresses the issues of online farm processing and storage of the data; however the actual functionalities implemented are still very limited, some of them no more than stubs. This leaves room for the development of more advanced features such as parallelizable, real-time data processing architectures, patterns allowing the integration of modules programmed in different languages, advanced GUIs for the global control of very large systems (thousands of network nodes), etc. An efficient, binary, general purpose data format called MFM has been designed together with a C++ reference framework implementing the encoding and decoding of raw data structures produced by detection systems. Although the MFM framework has reached a satisfactory level of maturity, it is believed that its real-time performance might still be optimized and improved. Also, implementations in other programming languages such as Java or Python might be of interest. Improving on developments made for high energy physics experiments the CompoundConfig (CCfg) configuration framework (C++ and Java) has reached a high level of maturity, both over general-purpose and embedded systems. However, many desirable improvements relating to GUI-based tools and widgets for easy and intuitive configuration use cases can still be developed. One of the most original and advanced software developments is the MDaq C++ framework which covers important issues relating to the control of distributed, heterogeneous, specific hardware modules. It allows users to dynamically develop over standard workstations “remote drivers” for specific hardware with a minimal need for recompilation, namely reading and writing over low-level hardware registers using arbitrary protocols through the use of configuration files. Although the MDaq framework thoroughly meets the demands of the GET project, many improvements pertaining either to new features or architecture design have been revealed during the actual development. These include features such as the remote and dynamic

47

management of arbitrarily long registers, the possibility of fine-grained tailoring for the optimization of memory imprint on embedded systems, advanced GUIs for the management of very large systems (e.g. tens of thousands of registers), Web-based rich GUIs, central database for wide-used protocols such as I2C, SPI, etc. Based on MDaq, CCfg, MFM, a very interesting development would be a generic framework for the development of prototype and production electronic test benches. It would consist of all the software bricks needed to build any test bench needed for the development or production of specific electronic objects (such as boards or Asics). The framework would have to be based on an open architecture allowing for improvement through the addition of specific modules or plugins written in multiple programming languages. Another development of high interest would relate to modules implementing the “virtual observatory” concept, i.e. non-localized control and configuration of experiments using advanced GUIs over diverse platforms such as workstation browsers, PDAs, tablets… This would also include addressing security issues without compromising ease of use and ease of installation.

48

ANNEX 5: SOLAR-Gamma We consider using a new generation scintillator such as lanthanum bromide or cerium bromide. The former has a reliable 3% energy resolution for 667-keV gamma rays but this is achievable only with a photomultiplier tube. Tests with APDs suggest that 6-7% is more realistic for a detector which can be used within a high magnetic field environment. This suggests cerium bromide with its slightly worse energy resolution (~4 %) is worthy of consideration – it may also be considerably cheaper. However, there are presently issues with reproduceability. Typical crystals and APDs are shown in figure A5.1.

Figure A5.1: top) 1/2" lanthanum bromide crystal. middle) typical Hamamatsu APDs with dimensions of 5 X 5 mm and 10 X 10 mm. bottom) prototype detector assemblies

49

ANNEX 6: SOLAR-HELIOS Estimate for DAQ & some basics (Lolly)

Most likely 100 DSSSDs Of 60X60 Si 300micron to 600micron Low resistivity Si (Time & energy measure) Strips 0,5mm Total N° of channels 24000 channels To have the Front-end up to ADC (12 bit) in vacuum Sampling might be necessary if we consider that we need PID in the Si. In this case it might be necessary to have the a pre-amp plus two filters ( note increase in cables is dramatic) ie numeric out and cooling Possibility to have 1mm thick back-up detector in pad form. Mechanical – with interleafing to avoid dead areas.

Figure 6.1 Possible disposition (cross-section) of DSSSDs on the “trumpet” detector assembly (Lolly)

Beam

Tube

Target

50

ANNEX 7: SOLAR- MEDICAL

Develop innovative positron detectors for PET like instrumentation The CaLIPSO project develops a new, high sensitivity, high spatial resolution PET-scan (Positron Emission Tomography) technology. The project focuses on the development of an innovative calorimetric detector for the gamma rays coming from the positron annihilation. Beyond the measurement of Gamma photon energies, this detector allows locating photon interactions in the detector in three dimensions, with accuracy of the order of 1 mm3. The CEA Technosanté steering committee considers this project as a technological breakthrough. A patent has been accepted (WO2011/117158), and is being extended to international.

The detector uses a “heavy” organometallic liquid (TriMethyl Bismuth, 82% by weight of Bismuth), to convert photons with energies of the order of 1 MeV. An ionisation chamber filled with ultra-pure TMBi is being developed: the detection of electrons allows the interactions to be located, in the plane of charge collection. The collected charge also codes the energy deposited in the detector. In addition, a common physical property of materials subjected to ionising radiation is light production, generally with a poor efficiency. The detection of optical photons will “trigger” the

detector, and provide accurate dating of the interaction time. The drift time of the electrons measures the distance covered through the thickness of the detector. The expected performances are far above existing techniques. For a 511 keV photon, computation give a photoelectric efficiency of 57%; and an energy resolution of ~ 10%; a temporal resolution of few 100 ps, allowing Time Of Flight reconstruction, and interaction localisation in the detector accurate to within 1 mm. Such a detector will improve the time resolution for material science analysis by nearly a factor two.

Operating principle of the CaLIPSO detector. The charge liberated by the ionising radiation (measurement of the energy, location in 2D), and the scintillation photons (detector triggering, interaction time) are both detected. Cross analysis of the two informations gives us access to the drift time of the electrons, and thus to the interaction location within the thickness of the detector.

ions

e-

-1O kV O kV

Photo-detectors PMTs trigger, timing

Pixellised Charge collection (Energy, 2D positioning)

Ionisation

Chamber

TMBi

51