HE CALORIMETER DETECTOR UPGRADE RD Y Onel for

HE CALORIMETER DETECTOR UPGRADE R&D Y. Onel for University of Iowa Fairfield University of Mississippi SLHC Workshop at FNAL November 19 – 21, 2008

HCAL Upgrades • • • 1 st Phase of R&D 2 nd Phase of R&D – Light enhancement tools: Zn. O, PTP – Radiation damage tests on Quartz and PTP 3 rd Phase of R&D – Alternative readout options: • PIN Diode, APD, Si. PMT, MCP • Microchannel PMT, MPPC – Radiation Hard WLS Fiber options • Quartz core sputtered with Zn. O • Sapphire fibers First Phase of the R&D 1. 2. Show that the proposed solution is feasible Tests and simulations of QPCAL-1 2

The “Problem” and the “Solution” • • As a solution to the radiation damage problem in Super. LHC conditions, quartz plates are proposed as a substitute for the scintillators at the Hadronic Endcap (HE) calorimeter. Quartz plates will not be affected by high radiation. But the number of generated cerenkov photons are at the level of 1% of the scintillators. Rad-hard quartz – Quartz in the form of fiber are irradiated in Argonne IPNS for 313 hours. – The fibers were tested for optical degradation before and after 17. 6 Mrad of neutron and 73. 5 Mrad of gamma radiation. – Polymicro manufactured a special radiation hard anti solarization quartz plate. 3

Radiation Damage Tests on Quartz with 24 Ge. V protons K. Cankocak et al. , NIM A 585, 20 -27, 2008 • We investigated the darkening of two high OH content quartz fibers irradiated with 24 Ge. V protons at the Cern PS facility IRRAD. • The two tested fibers have a 0. 6 mm quartz core diameter, one with hard plastic cladding (qp) and the other with quartz cladding (qq). • These fibers were exposed at about 1. 25 Grad in 3 weeks. • The fibers became opaque below 380 nm and in the range 580– 650 nm. • The darkening under irradiation and damage recovery after irradiation as a function of dose and time are similar to what we observed with electrons. • The typical attenuation at 455 nm are 1: 44 0: 22 and 2: 20 0: 15 d. B=m at 100 Mrad for qp and qq fibres, respectively. The maximum damage recovery is also observed near this wavelength. 4

Damage Recovery on Quartz Along and after irradiation the quartz exhibits a peculiar behaviour in the transmission of the blue light (near 450 nm) compared to the transmission of UV light and near 600 nm: • Quite low absorption near 450 nm compared to high absorption below 380 nm and near 600 nm. • Almost full recovery of the radiation damage near 450 nm and no recovery after high dose irradiation, below 380 nm and above 580 nm. 5

1 st Paper : R&D Studies on Light Collection • As a solution to the radiation damage problem in Super. LHC conditions, quartz plates are proposed as a substitute for the scintillators at the Hadronic Endcap (HE) calorimeter. • The paper (CMS-NOTE 2007/019) summarizing the First Phase of the R&D studies has been published : • F. Duru et al. “CMS Hadronic End. Cap Calorimeter Upgrade Studies for SLHC Cerenkov Light Collection from Quartz Plates” , IEEE Transactions on Nuclear Science, Vol 55, Issue 2, 734 -740, Apr 2008. • With these very nice comments from the editor and the refrees: • “The paper is very interesting and clearly proves that a solution exits for calorimeters in the SLHC era with similar light collection. ” • “The authors are to be thanked for a very interesting piece of work” 6

1 st Paper : R&D Studies on Light Collection • We have tested/simulated different fiber geometries in the quartz plates, for their light collection uniformity and efficiency. • Wave. Length Shifting (WLS) fiber, Bicron 91 a, is embedded in the quartz plate. Quartz plates are wrapped with reflecting material of 95 % efficiency. • The Cerenkov photons reaching the Photo. Multiplier. Tube (PMT) are counted. • Cerenkov Photons are shown in green. Photons emitted by WLS process are shown in red. • At the test beams we compared the light collection 7 efficiencies with that of original HE scintillators.

2 nd Paper : Quartz Plate Calorimeter Prototype - I The first quartz plate calorimeter prototype (QPCAL - I) was built with WLS fibers, and was tested at CERN and Fermilab test beams. Hadronic Resolution 8

What is missing on the 1 st Phase? - The WLS fibers used in QPCAL are BCF-12 by Saint Gobain (old Bicron) are not radiation hard. -The radiation hardness tests performed on BCF-12 shows that they are not very different than Kuraray 81 (current HE fibers). -The studies shows that BCF-12 can be more radiation hard with the availability of oxygen. 9

Second Phase of the R&D • • 1. How can we solve the fiber radiation problem? a) Use engineering designs b) Light enhancement tools (Zn. O, PTP, etc. ) 2. Radiation Damage Tests a) On Quartz b) On PTP Engineering Options • Current BCF-12 WLS fiber is not very radiation hard, but it can still be used • • *) We can engineer a system with fibers continuously fed thru a spool system. Iowa has built the source drivers for all HCAL (Paul Debbins), we also have expertise on site; Tom Schnell (University of Iowa Robotic Engineering). We have shown that a set of straight (or a gentle bend) quartz plate grooves allow WLS fibers to be easily pulled out and replaced. • • *) Different approach could be to use radiation hard quartz capillaries with pumped WLS liquid. We have the expertise; B. Webb (Texas A & M), E. Norbeck (Iowa) and D. Winn (Fairfield). This has been studies at Fairfield. The liquid (benzyl alcohol + phenyl naphthalene) has an index of 1. 6 but the attenuation length is still somewhat too short, possibly because of a too high WLS concentration. 10

Engineering Options Current BCF-12 WLS fiber is very radiation hard, but it can still be used *) We can engineer a system with fibers continuously fed thru a spool system. Iowa has built the source drivers for all HCAL (Paul Debbins), we also have expertise on site; Tom Schnell (University of Iowa Robotic Engineering). We have shown that a set of straight (or a gentle bend) quartz plate grooves allow WLS fibers to be easily pulled out and replaced. *) Different approach could be to use radiation hard quartz capillaries with pumped WLS liquid. We have the expertise; B. Webb (Texas A & M), E. Norbeck (Iowa) and D. Winn (Fairfield). This has been studies at Fairfield. The liquid (benzyl alcohol + phenyl naphthalene) has an index of 1. 6 but the attenuation length is still somewhat too short, possibly because of a too high WLS concentration. 11

Light Enhancement Tools Proposed Solution *) Eliminate the WLS fibers: Increase the light yield with radiation hard scintillating/WLS materials and use a direct readout from the plate. Questions… Questions … *) What is out there to help us? PTP (o. TP, m. TP, p. QP), and/or Zn. O can be used to enhance the light production. • How to apply them to the plates? and what thickness? • Which one work better? • Which is more radiation hard? 12

Quartz Plates with PTP • At Fermilab Lab 7, we have covered quartz plates with PTP by evaporation. We deposited 1. 5, 2, 2. 5, and 3 micron thickness of PTP. 13

Quartz Plates with PTP evaporation setup, and quartz plate holder 14

Quartz Plates with Zn. O • We also cover quartz plates with Zn. O (3% Ga doped), by RF sputtering. 0. 3 micron and 1. 5 micron. • We are currently working on 100 micron thick quartz plates, we’ve deposited Zn. O on each layer and bundle the plates together, for a radiation hard scintillating plate Fermilab Lab 7, Zn. O sputtering system and guns. 15

Test Beams for PTP and Zn. O We have opportunity to test our Zn. O and PTP covered plates, at CERN (Aug 07), and Fermilab MTest (Nov 07, and Feb 08). Blue : Clean Quartz Green : Zn. O (0. 3 micron) Red : PTP (2 micron) 16

Test Beams for PTP and Zn. O Mips from plain quartz plate. Mips from 0. 3 micron thick Zn. O (3% Ga) sputtered quartz plate. Mips from PTP evaporated quartz plate. 17

Test Beams for PTP and Zn. O We evaporated PTP on quartz plates in IOWA and tested them in MTest. Different deposition amounts and variations Were tested. 18

PTP Radiation Damage Tests • Sr-90 activated scintillation light output of the different p. TP samples which are saturated in toluene. • The toluene makes no measurable scintillation contribution. • Protons were done at CERN and Indiana Cyclotron. • The neutron data from Argonne. 19

What is learned from Phase II ? • The PTP and Ga: Zn. O (4% Gallium doped) enhance the light production almost 4 times. • OTP, MTP, and PQP did not perform as well as these. • PTP is easier to apply on quartz, we have a functioning evaporation system in Iowa, works very well. We also had successful application with RTV. Uniform distribution is critical!! • Zn. O can be applied by RF sputtering, we did this at Fermilab- LAB 7. We got 0. 3 micron, and 1. 5 micron deposition samples. 0. 3 micron yields better light output. • In light of these results we focused our efforts to Summer 08 Cern Test Beam. 20

Cern Test Beam – Summer 2008 • We have constructed and tested the QPCAL-II, with PTP deposited quartz layers. • The 20 cmx 5 mm, GE-124 quartz plates are used. • 2 μm PTP is evaporated on every quartz plate at Fermilab Lab 7. • The readout has been performed with Hamamatsu R 7525 PMTs. • For hadronic configuration 7 cm iron absorbers used between layers. • No WLS fiber! This is the second prototype “QPCAL-II” 21

Cern Test Beam – Summer 2008 • We also have tested different thickness of Zn. O and PTP deposited plates for mips. • Micro channel PMT prototype • Also HF PMT tests are performed by the same team. 22

Cern Test Beam – Summer 2008 • The “new plate” with stack of seven 100 μm thick quartz plates, each sputtered Zn. O on. This can give us a very radiation hard scintillating quartz plate. As a by product of our work . 23

QPCAL-II Hadronic Resolution - We have taken data with 30, 50, 80, 130, 200, 250, 300, and 350 Ge. V Pion beam. - Hadronic resolution is better than 12% at E > 350 Ge. V. *) At QPCAL-I the hadronic resoution was 18% at 300 Ge. V. 24

QPCAL-II Hadronic Response Linearity Very linear response and A nice signal distribution 25

QPCAL-II EM Configuration - After hadronic configuration the iron plates between layers reduced to 2 mm thickness for Electron runs. - We call this EM configuration. - We took data with 50, 80, 100, 120, and 200 Ge. V electrons 26

QPCAL-II EM Configuration We had very good EM resolution As well, but at higher electron Energies the signal started to deviate. 27

QPCAL-II Muon Response 225 Ge. V Muon signal on QPCAL-II 28

Different PTP Thickness on Quartz - The plate prepared in IOWA vacuum chambers performed really well - Since we don’t have a thickness gauge we cannot tell the thickness, but it has 1 gr PTP deposited on 15 cmx 15 cm quartz. We did not see drastic variations between 2, 2. 5, and 3 micron PTP deposited plates 29

7 layer 700 micron Zn. O plate - We have deposited 0. 2 micron Zn. O (%4 Ga) to “ 100 micron” thick quartz plates. - This sandwich structure with 0. 7 mm total thickness is placed in an aluminum frame and tested for mips on this test beam for the first time. - We got very promising results, for both pion and electron beams. We need to work on this technique to develop future “radiation hard scintillators”. 30

off-axis beam vs co-axis beam The pyramids are positioned so the PMTs Are not aligned with the beam. When they are aligned with beam, we observe Cerenkov form pmt window. 31

Results from Cern TB 08 • We had very successful test beam, performed various tests at a very short time. • QPCAL-II with PTP deposited plates and performed better than QPCAL-I (with. WLS fibers). With the obtained hadronic resolution of better than %13, we successfully finished the 2 nd phase of our R&D. • As one of the many spinoffs of this R&D , we showed than stacking very Thin Zn. O treated quartz plates, we can Get “new rad-hard scintillators”. 32

Third Phase of the R&D Alternative Readout Options : APD, Si. PMT, PIN diode, Micro-channel PMT, MCP, MPPC • Which one is better? Wavelength response? Surface area? • Are they radiation hard? Developing Radiation Hard Wavelength Shifing Fibers • Quartz fibers with Zn. O covered core. • Sapphire fibers

New Readout Options We tested; *) Hamamatsu S 8141 APDs (CMS ECAL APDs). The circuits have been build at Iowa. These APDs are known to be radiation hard; NIM A 504, 44 -47 (2003) *) Hamamatsu APDs: S 5343, and S 8664 -10 K *) PIN diodes; Hamamatsu S 5973 and S 5973 -02 *) Si PMTs 34

New Readout Options • Si. PMT has lower noise level. • For all of these readout options we designed different amplifier approaches: • 50 Ohm amplifier. • Transimpedance amplifier. • Charge amplifier. 50 Ohm Amplifier circuit design. 35

New Readout Options The speed of the readout is essential. The pulse width of the optical pulses from the scintillator limits the selection of photodiode or APD used. A bandwidth of 175 MHz is equivalent to a rise and fall time of 1. 75 nsec. Topology Price Speed (Rise time) Input Equiv. Noise Comments Photodiode with 50 Ohm amplifier Low Fast (< 1 nsec) ~ 50 p. W/√Hz Simple circuitry Photodiode with fast transimpedance amplifier Low Moderate (< 3 nsec) ~ 10 p. W/√Hz Simple circuitry APD with 50 Ohm amplifier Moderate Fast ~ 250 f. W/√Hz (Gain of 50) Drift with temperature High voltage Moderate complexity for HV APD gain from 25 to 150 APD with 50 Ohm amplifier Moderate (<3 nsec) ~ 50 f. W/√Hz Drift with temperature High voltage Moderate complexity for HV APD gain from 25 to 150 Silicon PMT Moderate Fast ~. 1 f. W/√Hz Simple to moderate complexity 36

New Readout Options We have tested ECAL APDs as a readout option. 2 APD connected to plain quartz Plate yields almost 4 times less light than fiber+PMT combination. 37

So far what is learned from Phase III ? • Single APD or Si. PMT is not enough to readout a plate. But 3 -4 APD or Si. PMT can do the job. • Si. PMTs have less noise, higher gains, better match to PTP and Zn. O emission λ. • As the surface area get bigger APDs get slower, we cannot go above 5 mm x 5 mm. • The PIN diodes are simply not good enough. • The ECAL APDs are claimed to be radiation hard • Feed the linear arrays of Si. PMT or APD to the system, arranged as a strip of 5 mm x 20 -50 cm long… engineering !!… • A cylindrical HPD, 5 -6 mm in diameter, with a sequence of coaxial target diodes anodes on the axis, 20 -50 cm long, and a cylindrical photocathode. 38

Developing new technologies • We propose to develop a radiation hard readout option. – Microchannel PMT (Onel-Winn) – MCP-PMT ( Hamamatsu and Photonis) – MPPC (Multi Pixel Photon Counter) • We also propose to develop a radiation hard WLS fiber option. – Doped sapphire fibers. – Quartz fibers with Zn. O sputtered on core. 39

Radiation hard readout option “Microchannel PMT” *) Fairfield and Iowa have focused on revolutionizing photomultiplier technology through miniaturization coupled with the introduction of new materials technologies for more efficient photocathodes and high gain dynode structures. *) Miniaturization enables photomultipliers to be directly mounted on circuit boards or silicon for interfacing directly with readout circuits. *) Fast response time, high gain, small size, robust construction, power efficiency, wide bandwidth, radiation hardness, and low cost. 40

Radiation hard readout option “Microchannel PMT” *) Photograph of a micromachined PMT in engineering prototype form. *) The metal tabs for the dynode and focusing voltages, signal, cathode. *) 8 stage device is assembled from micromachined dynodes which exhibits a gain of up to 2 -4 per stage onsingle stage. *) The total thickness < 5 mm. *) 8 x 4 pixel micro-dynode array is shown *) The layers are offset relative to each other to maximize secondary electron emission collisions. 41

Hamamatsu MPPCs Hamamatsu released a new product. Multi Pixel Photon Counter, MPPC. We purchased this unit, working on tests, but it is simply an array of APDs. It is not the same thing with our proposed “microchannel PMT”. 42

Ti: Sapphire looks promising But not so fast !! Remember, there is NO quick solution to our problem 43

Doped Sapphire !! A. Alexandrovski et al. Ti 3+ Cr 3+ 44

Ag-Sapphire ? ? A recent study shows that the Ag ions can be implanted into sapphire in the ke. V and Me. V energy regimes. The samples implanted at 3 Me. V shows a large absorption peak at the wavelengths ranging from 390 to 450 nm when heated to temperatures higher than 800◦C. Y. Imamura et al. 45

What can be done with sapphire? • Sapphire optical fibers are commercially available in standard lengths of 200 cm x 200 micron diameter. Cheaper stock fibners are 125 micron diameter x 125 cm long. These fibers are of use in Ti: sapphire fiber lasers, and sensors. • A large variety of dopants are possible in sapphire, covering a large wavelength interval. • Under the right conditions, the Ti+4 ion (40 ppm) in heat treated sapphire absorbs in the UV and emits in the blue, with a time constant 5 -7 ns. it is reasonably (50%90% or more) efficient. At 1 ppm the shift is at 415 nm - even at 1 ppm the fluorescnece is visible to the human eye. At 40 ppm it shifts to 480 nm. Fe 2+ and Mg 2+. Other Ti charge states and other dopants absorb in the UV-Blue and emit in the yellow and red. • We propose to investigate these and similar inorganic fibers, grown mainly for fiber lasers, but with dopants adjusted for fast fluorescence (rather than forbidden transistion population inversions), and to test the rad hardness. 46

What about treating quartz fibers? • Heterogenous nanomaterials Scintillating glass doped with nanocrystalline scintillators has also been shown to be a good shifter. We propose (i) testing radiation hardness and (ii) to investigate doping quartz cores with nanocrystalline scintillators (Zn. O: Ga and Cd. S: Cu). The temperatures involved are very reasonable. • Thin film fluorescent coatings on quartz cores 250 -300 nm UV has been shown to cause 5 -10 ns fluorescence in Mg. F 2, Ba. F 2, Zn. O: Ga. We propose coating rad-hard quartz fibers with a thin film, and then caldding with plastic or fluoride doped quartz. CVD deposition of Doped Zn. O is now a commercial process, as it is used to make visible transparent conducting optical films as an alternative to indium tin oxide, as used in flat panel displays and solar cells. 47

Damage of HE scintillators R (cm) 5 layers of HCAL megatiles will be affected severely Tolerable level of the absorbtion dose for HCAL scintillators is order of 5· 104 Gy (5 Mrad) After 5· 105 pb-1 integral luminosity (10 years LHC) After 1. 5· 106 pb-1 integral luminosity (10 years LHC + 1 year SLHC) Z (cm) 48

Thick Gas Electron Multiplier (THGEM) Proposal from Russian Colleagues The idea of the THGEM – to obtain gas amplification inside «GEM mesh» VD - 2000 V VU Drift gap Vd Induction gap VI - 1800 V - 200 V -0 V Readout Schematic view of the THGEM chamber 49

The 10 x 10 cm THGEMs were manufactured with standard PCB technology by precise drilling and Cu etching, out of double-face Cu-clad G-10 plate, of 0, 4 -0, 8 mm thickness A gap of 0, 1 mm was kept between the rim of the drilled hole and the edge of the etched Cu pattern Pitch, a=0, 8 mm Hole, d=0, 4 mm Rim, r=0, 1 mm A microscope photograph of a THGEM 50

Effective gain of THGEM thickness=0, 4 mm, hole diameter d=0, 3 mm and pitch a=0, 7 mm The pulse-height dependence on the event rate of the THGEM for two different gains. thickness=0, 4 mm, hole diameter d=0, 3 mm and pitch a=1, 0 mm THGEM 740 Torr 51

Issues of THGEM Production Technology Y R Electric field (V/m) Different types of THGEM shape simulated using Maxwell (Sergey Vasilyev) Electric field at the age of the Rim: ~150 -200 k. V/cm Rim Hole Electric field (V/m) Distance along R, on THGEM surface (mm) Electric field inside hole: ~40 k. V/cm Multiplicatio Inductio n Drift: n: 1 -2 3 -5 Kv/cm Distance along Y across THGEM detector (mm) corona at the boundary of Cu The shape of the rim have an impact on clad Technological issues to avoid sharpness at the age of the rim is under 52 study

THGEM test pcb 8 holes d=0. 8 mm Rim around drilled holes r 8 holes d=3. 1 mm p Boards were tested at HV = 2 KV No sparks ~10 mm d 100 mm Double side G 10 plate Cu Drilled foil hole t HV inputs from both sides of the Cu plate Etched Cu diameter 0, 5 mm 100 mm ~ 5 mm 1. Sensitive area 100 x 100 mm 2 2. Material: double-face Cu clad G 10 plate Thickness: t = 0, 4 ÷ 0, 8 mm Diameter of the hole: d = 0, 4 mm Pitch: p = 0, 8 ÷ 1, 2 mm Rim around hole r = 0, 1 mm 53

Participation in the HCAL beam test The idea is to put the box of THGEMs: 4 planes ~30 x 30 cm separated by absorber (equivalent of 4 HCAL layers) in front of HCAL prototype on H 2 beam Goal – to measure energy resolution of HE with and without THGEM 2 options of R/O electrodes: § R/O pads 10 x 10 cm 9 ch/plane x 4 planes = 36 R/O channels = 6 QIE boards § R/O pads 15 x 15 cm 4 ch/plane x 4 planes = 16 R/O channels = 3 QIE boards Total 36 R/O channels are needed = 6 QIE boards 54

Schedule – Quartz Plates v Purchase/Delivery of Quartz Plates: 4 -6 months v Coating of the plates by PTP or Zn. O(Ga): 4 -6 months v Photodetector testing: 3 months v Final system assembly testing: 3 months Total: 18 months 55

Back-up 56

Radiation Hard WLS fibers: Sapphire Fibers Sapphire is a very radiation hard material and it can be brought into fiber form. But by itself It has very little absorption and florescence. • Tong et. al. , Applied Optics, 39, 4, 495 Absorption in Sapphire can be provided by; – conduction to valence band in UV – multiphonon in mid-IR – native defects • vacancies, antisites, interstitials, – Impurities !!!! • e. g. transition metals: Cr, Ti, Fe, … 57

Problems with Ti: Sapphire • There are some crystals used for lasers, but no fiber, yet. • The Ti: Sapphire has a luminescence lifetime of 3. 2 microsec!! And looks like this is temperature dependent (Macalik et. al. Appl. Physc. B 55, 144 -147). • off “resonant” absorption significant • sum of several species can contribute to absorption at given l • Redox state important – e. g. a[Ti 3+] a[Ti 4+] – annealing alters absorption without altering impurity concentrations • Impurities do not necessarily act independently – Al : Ti 3+ : Ti 4+ : Al Al : Ti 3+ : Al : Ti 4+ : Al – absorption spectra at high concentrations not always same as low 58