1. TDR for the PANDA Forward Spectrometer Calorimeter (Shashlyk) Pavel Semenov IHEP, Protvino PANDA Collaboration Meeting, Giessen 20 March 2015

2. Pavel Semenov, PANDA Collaboration Meeting, Giessen2 Outline FSC TDR structure FSC requirements FSC module and detector design Readout electronics Monitoring system and FSC calibration procedure Test beam results Resources and timelines

3. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen3 TDR structure PANDA detector overview FSC requirements FSC module and detector design including selection of the photodetector, high voltage base design and three FSC prototypes design Short chapter on readout electronics Discussion on calibration procedures and description of the light monitoring systems for the FSC Simulations based on PandaRoot Detector prototypes performance (testbeam results) –Protvino –Mainz 2012, Mainz 2014 Project Management

4. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen4 PANDA Detectors view - sizes are determined by a position at ~7 meters from the IP, opening in the Forward EMC Endcap (covers ~8% of solid angle) - depth by the space between Forward TOF and forward Muon Range system

5. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen5 DPM simulations (photons into the FSC) All photons energy over photons at the Shashlyk aperture (color curves) from interactions with 5 and 15 GeV/c beam As a part of the PANDA calorimeter complex FSC should measure photons from 10 MeV to 15 GeV

6. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen6 Energy resolution requirements Precise measurement of electrons and positrons energy and efficient recognition of light mesons requires 3% / √E and ~1% of constant term Depth of the detector of 20 radiation length Energy registration threshold (single cell) 3 MeV

7. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen7 FSC rate capability requirements DPM simulation for the time between events in FSC cell shows mean rate of 1 MHz near the center of the detector and 300 kHz at the detector vertical edge

8. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen8 Table of the FSC requirements Rates and doses calculations are based on luminosity 2x10 32 cm -1 s -1

9. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen9 FSC design The fine segmented sampling calorimeter with light collection by WLS fibers going through the detector media (shashlyk) can provide all the required features. Sampling electromagnetic calorimeters with alternating layers of scintallator and absorber are widely used in HEP experiments. PANDA FSC design based on the KOPIO project at BNL with very fine sampling fraction to improve the energy resolution. IHEP was involved in prototype production for KOPIO and still has the required expertise The cell size was optimized for PANDA

10. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen10 Compilation of the FSC mechanical properties

11. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen11 FSC detector general view 378 modules are divided into two sections On the side of the each section crates with readout electronics are installed The support frame is installed on the rollers Back side zone provides space for the photodetectors, HV bases, cables and optical fibers of the light monitoring system

12. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen12 Individual module design

13. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen13 Scintillator tiles and lead plates The lateral dimensions of a single cell (55x55mm 2 ) are close to the Moliere radius of 59.8 mm. The scintillator material is made of polystyrene doped with 1.5% paraterphenyle and 0.04% POPOP. The tiles are produced at IHEP Protvino exploiting the injection moulding technology and have 36 holes for the light collecting WLS fibres. The quadratic lead absorber plates are common for four cells. The 0.275 mm thick lead sheets are doped with 3% of antimony to improve their rigidity and they provide fixing holes for four individual scintillator tiles The scintillation light of a single FSC cell is collected and accumulated by 36 wavelength shifting (WLS) multi-cladding fibres running axially through the sandwich structure and form 18 loops per cell at the front side of the module. To make the loop with the required small radius of curvature, additional care with thermal treatment was taken during the bending of fibres.

14. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen14 FSC Prototypes Type-1 KOPIO mold was used to produce 110x110 mm 2 tiles, Bicron BCF91A fibres, Tyvek wrapping around the module, four fixing pins per tile (correct spacing between plastic surface and lead for the full internal reflection), no Tyvek beween scintillator and lead, fiber loop radius is 28 mm Type-2 KOPIO size tiles were milled in four quadratic parts (55x55mm 2 ), Bicron BCF91A fibres, 1 fixing pin per tile (incorrect spacing between plastic surface and lead), painted side faces of the tile, no Tyvek between scintillator and lead, fiber loop radius is 14 mm Type-3 55x55 mm 2 scintillator tiles were produced with new mold, Kuraray Y11 fibres, four fixing pins per tile, painted side faces of the tile, sheets of Tyvek between scintillator and lead, fiber loop radius is ~14 mm

15. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen15 Photodetectors Wide dynamic range, fast signal and position outside the magnetic field allows the use of photomultiplier as a FSC photodetector Several PMTs were selected from the manufacturers datasheets including Hamamatsu R7899, which is used for the same type of the calorimeter by LHCb Cosmic muon test setup measurements with those PMTs showed similar performance, but taking into account the price, R7899 was selected

16. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen16 HV bases To fulfill the requirements of high rate capability and wide dynamic range Cockcroft-Walton type (CW) of HV base was designed and tested during testbeam runs and on the dedicated test setup For the Type-3 prototype tapered CW was designed to increase the output current linearity up to the 100 mA To test the performance of the PMT and CW base the dedicated test setup was build at IHEP

17. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen17 PMT + HV base test setup

18. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen18 PMT and bases performance Linearity was tested with the set of neutral filters with calibrated optical density, measured linearity >10 4 For the rate performance two LEDs setup was used and slightly different CW bases (80 kHz, 100kHz, 130 kHz) were tested as well as passive (resistorized) HV base. Almost full range signal (4 V of LED amplitude at 50 Ohm) was used to detect gain change and backgroud LED simulated 800 MeV signal at several frequencies). Measured gain drop at 2MHz load is 6-10% for different versions of CW base

19. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen19 PMT and CW base temperature dependence Three digital thermo sensors (DS18B20) were used The sensors provided the temperature measurements with a resolution of 0.05 degree C and were installed at two points of the CW base inside the PMT compartment and outside the PMT compartment to measure the external temperature. With the emulated most heavy load expected at the PANDA FSC (2 MHz) the temperature increase at the hottest point in the PMT compartment was < 5 C relative to the external temperature. The measured dependence of the gain was < 0.2 %/ C.

20. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen20 DCS

21. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen21 Light monitoring system design Simple front side monitoring system – LED is installed in the front cover of the each module and illuminates fout channels. Useful to check is the channel is alive, HV is on, etc) High precision back side monitoring system – quartz fiber to each module from high stability light source

22. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen22 FSC support frame Modules of the complete FSC arranged in 14 layers with 27 modules in each layer To frame support consists of two sections, each can be assembled and transported to the beam zone separately Frame support must be strong enough to carry the weight of the modules (~4300 kg) and keep the stable position within 200 micron tolerance Back plate provides the positioning of the modules, fixation of the PMT compartments and serves as a heat transfer media

23. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen23 Stress and deflection calculations Deflection calculations : < 150 microns (200 microns allowed) Stress < 70 MPa (240MPa allowed)

24. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen24 FSC cooling The only heat source is PMT HV bases Our CW bases power dissipation is 100 mW per channel (measured during testbeam) Metal part in included in the design of the plastic PMT compartment to transfer heat outside

25. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen25 FSC cooling calculations Only heat conduction through the FSC detector back plate was taken into account Event with steel back plate the maximum temperature difference is 12 degrees C

26. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen26 Assembly and installation procedure Assembly of the two FSC sections support frames with additional temporary sidebeams Stacking of the modules applying pressure on each raw of modules with side pressure bars Installation of the PMT compartments, PMTs, bases Mounting horizontal cable trays Mounting crates for the readout electonics Connecting cables and fibres of monitoring system Transportation of the FSC sections to the FS platform or directly to the beam hall Moving the FSC sections close to the final position near the beam pipe and dismounting the temporary sidebeams Adjustment the sections position relative to the beam pipe Fixing the sections together by fastening elements

27. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen27 FSC maintenance Maintenance can be done in the beam hall after the installing the additional supports at the left and right side of the FS platform FS rail system is extended to move FSC sections to those additional support platforms No need to remove the whole FS platform or the beam pipe

28. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen28 Readout electronics Similar to the target EMC readout the FSC readout chain contains digitizer, data concentrator and the Compute Node Digitizer processes data from the SADC (base line calculation, hit detection, feature extraction, pule up detection) Data concentrator collects data from several digitizers, pre-process data and provides connection to the PANDA time distribution system

29. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen29 FSC Sampling ADC design The design is under development by Uppsala group 14 bit ADC, 250 MS/s, cover wide range from 3 MeV to 12 GeV signals ADC input range (+/- 1V ) will be adapted to the relatively high amplitude of the PMT output (up to the 5 V) Dual range ADC configuration is possible Analog shaping at the input of ADC (from 50 ns to 100 ns) 16 LEMO inputs to one ADC mezzanine card microTCA carrier board Two competing ADC chips AD9250 or LTC2123

30. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen30 Calibration procedures Periodical calibration of each FSC cell and monitoring of the PMT gain changes to compensate the calibration coefficients change between subsequent calibrations Calibration procedures divided into three levels – pre- calibration, on-line calibration and off-line calibration with increasing demands for the calibration time and thus the precision Pre-calibration with cosmic muons is not precise enough (not better than 4%), but useful for the detector commissioning phase and the maintenance Fine calibration with neutral pions uses invariant mass peak to calibrate the scale of the calorimeter Fine calibration with electrons based on measured by other detectors electron momentum

31. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen31 Testbeam results (Type-1, Protvino) Type-1 prototype 1-19 GeV electron beam at 2b beam line at U-70 accelerator Measured energy resolution: stochastic 3.51%, constant 1.3% Measured position resolution: stochastic 3 mm, constant 15.4 mm

32. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen32 Testbeam results (Type-2, Protvino) Measured energy resolution: stochastic 3.15%, constant term 1.37% Measured position resolution: 2.6 mm resolution for the 19 GeV (5 mm for the Type-1)

33. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen33 Testbeam results, Type-2, MAMI Measured energy resolution: stochastic 4.21%, constant 3.82% Measured position resolution strongly depends on the position over the cell 8-13 mm for the 800 MeV (without beam spot size of 10 mm correction) Unexpected large non-uniformity of the light output over the cell surface(+/- 15%)

34. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen34 Testbeam results, Type-3, MAMI Improved light output non-uniformity (+/- 4.5%) Similar to Type-2 energy resolution (constant term improved) Similar position resolution from 7 to 12 mm (without beam spot size of 10 mm correction) at 800 MeV Time resolution with a stochastic term of 100 ps

35. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen35 Project management IHEP, Protvino will fulfill the main part of the workpackages SADC development will be done at Uppsala by Pawel Marcinievsky Pre-assembly can be done at Julich Resources at IHEP: 28 people in total 17 FTE –14 experts from the IHEP PANDA (0.3 to 1 FTE) two full-time designers, two full-time engineers –5 FTE technicians from the IHEP Scintillator Department. –5 technicians from the IHEP Mechanical Workshop will be part- time involved Our cost estimate has shown that we can provide all the workpackages for 1.352.000 Euro (Costs in 2005) as it is fixed in PANDA CostBook

36. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen36 Schedule Autumn of 2015 - get the FAIR approval of the TDR; End of 2015 - get a signed Collaboration contract of IHEP-Protvino with FAIR to produce the FSC at IHEP Protvino; January 2016 - December 2019 - the four years Collaboration contract of IHEP with FAIR on the FSC by using the Russian contributed money into FAIR; July 2016 - July 2018 - manufacture parts and assemble all modules; January 2016 - July 2017 - manufacture all the mechanical support parts in Russia; July 2016 - purchase the photomultipliers and ship them directly from the manufacturer to Juelich, Germany; 2017 - purchase the readout electronics (Uppsala University) and ship it to Juelich, Germany; 2017 - October 2018 - shipping the modules to Germany; July 2017 - End of 2017 - shipping mechanical support to Germany; November 2017 - End of 2018 - pre-assembly and on-site tests of the FSC modules at Juelich; 2019 move the whole assembled and tested FSC from Juelich to the PANDA experimental hall at FAIR, install the FSC in the PANDA experimental hall

37. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen37 Shashlyk timelines

38. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen38 Acknowledgements We acknowledge the dedicated financial support from the Russian State Corporation ROSATOM" over the years 2008-2013. We gratefully acknowledge the generous support by the A2 Collaboration in providing the tagged-photon beam time at MAMI,Mainz, Germany Many thanks for the members of the Editorial Board for the great job of improving the TDR text

39. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen39 Backup slides

40. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen40 Charmonium study, 6.27 GeV/c beam Energy of photons from charmonium decays going to FSC is relatively small, registration threshold should be similar to EMC

41. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen41 FSC moving system A dedicated moving system is required to install the two FSC sections in the beam position and to move them out for maintenance Commercial rollers from INDUSTRIAL LIFTING company can be used for the FSC moving system

42. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen42 Monitoring system Prototype had a light distribution uniformity of 2% (RMS) and stability of 0.1% over one week of operation

43. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen43 Simulations FSC detector geometry is implemented in the PandaRoot framework Cluster finding and bump splitting Digitization of the FSC signal

44. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen44 Energy deposition non-uniformity Fiber positions Non-uniformities in the order of 1% at the fiber positions because of Cherenkov light in the fiber

45. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen45 Testbeam results for 1.5 m (more sophisticated analysis) Shower profile fit, Charged hadrons removed (drift chamber), Rough calorimeter calibration π 0 1-2 GeV, σ m 12.5 MeV

46. 20 March 2015Pavel Semenov, PANDA Collaboration Meeting, Giessen46 Prototype tests (Type-1, Protvino)