Blair Ratcliff5 th SuperB Workshop, Paris May PID Related R&D at SLAC: The Focusing DIRC and a Fast TOF Using Cherenkov Light. Blair Ratcliff Blair Ratcliff Representing: I. Bedajanek, J Benitez, D.W.G.S. Leith, G. Mazaheri, J. Schwiening, K. Suzuki, J. Uher, J. Va’vra.

Blair Ratcliff5 th SuperB Workshop, Paris May Focusing DIRC Prototype Goals Work with manufacturers to develop and characterize one or more fast, pixelated photon detectors including; basic issues such as cross talk, tube lifetime, and absolute efficiency operation in 15 KG field Measure single photon Cherenkov angular resolution in a test beam use a prototype with a small expansion region and mirror focusing, instrumented with a a number of candidate pixelated photon detectors and fast (25 ps) timing electronics  Much lower background sensitivity with somewhat improved performance parameters demonstrate correction of chromatic production error via precise single photon timing demonstrate that the performance with mirror focusing is as expected. measure N 0 and timing performance of candidate detectors.

Blair Ratcliff5 th SuperB Workshop, Paris May Focusing DIRC Prototype Optics Radiator 3.7m-long bar made from three spare high-quality BABAR-DIRC bars Expansion region coupled to radiator bar with small fused silica block filled with mineral oil (KamLand experiment) to match fused silica refractive index include optical fiber for electronics calibration would ultimately like to used solid fused silica block Focusing optics spherical mirror from SLD-CRID detector (focal length 49.2cm) Photon detector placed in fixed slots allowing easy replacement. typically using 2 Hamamatsu flat panel PMTs and 3 Burle MCP- PMTs in focal plane readout to CAMAC/VME electronics with 25 ps resolution. Limited number of channels available.

Blair Ratcliff5 th SuperB Workshop, Paris May Focusing DIRC prototype photon detectors Nucl.Inst.&Meth., A 553 (2005) 96 T iming resolutions were obtained using a fast laser diode in bench tests with single photons on pad center. T iming resolutions were obtained using a fast laser diode in bench tests with single photons on pad center.  narrow ≈70ps time (ns)  narrow ≈140ps time (ns)  narrow ≈220ps time (ns) 1) Burle MCP-PMT (64 pixels, 6x6mm pad,  TTS ~50-70ps) 2) Hamamatsu H-8500 MaPMT (64 pixels, 6x6mm pad,  TTS ~140ps) 3) Hamamatsu H-9500 Flat Panel MaPMT (256 pixels, 3x12mm pad,  TTS ~220ps)

Blair Ratcliff5 th SuperB Workshop, Paris May Beam Test Setup 10 GeV/c e- beam in End Station A at SLAC. Beam enters bar at 90º angle. 10 Hz pulse rate, approx. 0.1 particle per pulse Bar contained in aluminum support structure Beam enters through thin aluminum foil windows Bar can be moved along long bar axis to measure photon propagation time for various track positions Trigger signal provided by accelerator Fiber hodoscope (16+16 channels, 2mm pitch) measures 2D beam position and track multiplicity Cherenkov counter and scintillator measure event time Lead glass calorimeter selects single electrons Mirror and oil-filled detector box: Movable bar support and hodoscope Start counters, lead glass Hodoscope Scintillator Cherenkov counter Calorimeter Radiator bar in support structure Prototype e – beam

Blair Ratcliff5 th SuperB Workshop, Paris May GeV/c electron beam data. 10 GeV/c electron beam data. ~ 200 pixels instrumented. ~ 200 pixels instrumented. Ring image is most narrow in the 3 x 12 mm pixel detector. Ring image is most narrow in the 3 x 12 mm pixel detector. Hamamatsu H-8500 Hamamatsu H-9500 Burle Cherenkov ring in pixel domain: Cherenkov photons in time domain: Cherenkov Photons In Time & Pixel Domains

Blair Ratcliff5 th SuperB Workshop, Paris May Cherenkov angle production controlled by n phase (cos  c = 1/(n phase  :  c (red) <  c (blue) Propagation of photons is controlled by n group (v group = c 0 /n group = c 0 /[n phase -  phase   ): v group (red) > v group (blue) Principle of chromatic correction by timing: TOP = time of propagation of photon in the bar TOP/Lpath = 1/v group ( ) The DIRC as a Spectrometer

Blair Ratcliff5 th SuperB Workshop, Paris May  ll detectors: PMT with 3mm pixels only: Structure in the uncorrected distributions are caused by the coarse pixelization and the finite beam phase space. The time correction smears this effect away. Structure in the uncorrected distributions are caused by the coarse pixelization and the finite beam phase space. The time correction smears this effect away. Smaller radial pixel size (3mm) helps to improve the Cherenkov angle resolution as expected. Smaller radial pixel size (3mm) helps to improve the Cherenkov angle resolution as expected. Correction off: Correction on:  ~ 5.0 mrad  ~ 5.1 mrad  Chrom +  Pixel Correction off: Correction on:  ~ 5.0 mrad  ~ 5.1 mrad  Chrom +  Pixel Resolution Improvement with Chromatic Correction

Blair Ratcliff5 th SuperB Workshop, Paris May Main contributions to the  c resolution: Main contributions to the  c resolution: - chromatic smearing: ~ 3-4 mrad - chromatic smearing: ~ 3-4 mrad - 6mm pixel size: ~5.5 mrad - 6mm pixel size: ~5.5 mrad - optical aberrations of this particular design: - optical aberrations of this particular design: grows from 0 mrad at ring center to 9 mrad in outer wings of Cherenkov ring grows from 0 mrad at ring center to 9 mrad in outer wings of Cherenkov ring  c resolution - all pixels:  c resolution - 3mm pixels only: Data / GEANT 4 MC Comparison

Blair Ratcliff5 th SuperB Workshop, Paris May Prototype’s Npe_measured and Npe_expected are consistent within ~20%. Prototype’s Npe_measured and Npe_expected are consistent within ~20%. Hamamatsu H-9500 MaPMTs: Hamamatsu H-9500 MaPMTs: We expect No ~ 31 cm -1, which in turn gives Npe ~ 28 for 1.7 cm fused silica bar thickness, and somewhat better performance in pi/K separation than the present BaBar DIRC. Burle-Photonis MCP-PMT: Burle-Photonis MCP-PMT: We expect No ~ 22 cm -1 and Npe ~ 20 for B = 0kG. BaBar DIRC design: BaBar DIRC design: No ~ 30 cm -1, and Npe ~ 27. Expected performance of a final device: Focusing DIRC prototype bandwidth: Performance Estimates of this Design for Different PMT Choices (JV model)

Blair Ratcliff5 th SuperB Workshop, Paris May Goal: Develop a Cherenkov Based TOF with total timing resolution <~15 ps. Bench Tests of Main Elements of a Fast TOF

Blair Ratcliff5 th SuperB Workshop, Paris May Laser Test Stand PiLas laser head:Parameter SLAC tests Laser diode source PiLas Wavelength 635 nm TTS light spread (FWHM) ~ 35 ps ~ 35 ps Fiber size 62.5  m Start Calibration of a fast detector:

Blair Ratcliff5 th SuperB Workshop, Paris May Single Photon Timing Resolution (  TTS ) 10  m MCP hole diameter 10  m MCP hole diameter B = 0 kG B = 0 kG 64 pixel devices, pad size: 6 mm x 6 mm. 64 pixel devices, pad size: 6 mm x 6 mm. Phillips CFD Phillips CFD PiLas red laser diode operating in the single photoelectron mode (635 nm). PiLas red laser diode operating in the single photoelectron mode (635 nm).  TTS <  ( ) = 26 ps (Npe = 1)  TTS <  ( ) = 26 ps (Npe = 1) Ortec VT120A amplifier ~0.4 GHz BW, 200x gain + 6dB Fit: g + g Burle/Photonis MCP-PMT (ground all pads except one) Fit: g + g Hamamatsu C amplifier 1.5 GHz BW, 63x gain

Blair Ratcliff5 th SuperB Workshop, Paris May Bench Tests With Fast Time Digitization  PiLas ~15 ps/  N pe Trigger TTL NIM Disc  PiLas_trigger ~? Pulser START STOP 14 bit ADC 114 TAC 566  Pulser + TAC_ADC ~ 3.2 ps (My measurement)  Fiber ~?  Delay ~?  MCP-PMT  =  {   MCP-PMT +   Fiber +   Amp_CFD +   Delay +   PiLas +   Pulser+TAC_ADC +   PiLas_trigger } + Systematic effects: laser & temperature drifts, ground loops, etc.  Pulser_TAC_ADC ~ 3.2 ps Ortec 9236 Amp/CFD  Amp_CFD ~ 6 -7 ps (Manufacturer) ManufacturerMy measurement

Blair Ratcliff5 th SuperB Workshop, Paris May Timing Resolution Versus Npe for different R/O Npe = for 1cm-thick Quartz radiator + window & with Burle Bialkali QE. Npe = for 1cm-thick Quartz radiator + window & with Burle Bialkali QE.  < 15 ps seems feasible.  < 15 ps seems feasible. The Ortec 9327-like performance is adequate. The Ortec 9327-like performance is adequate.

Blair Ratcliff5 th SuperB Workshop, Paris May Summary Pixelated Photon detectors continue to be improved by manufacturers. The performance expected from a small SOB DIRC with a MaPMT H-9500 tube should be similar to the BaBar DIRC in Npe with ~ 20% better resolution with the 3 mm pixels. Need to continue to work with manufactures in order to improve the photon detectors. We have demonstrated the principle of chromatic correction to the Cherenkov angle using timing. Bench tests of fast timing PMTs are encouraging to date. Best results with the laser diode: Best results with the laser diode: -  ~ 12 ps for Npe = (as expected from 1cm thick Cherenkov radiator). -  TTS < 26 ps for Npe ~ 1. - Upper limit on the MCP-PMT contribution :  MCP-PMT < 6.5 ps. - TAC/ADC contribution to timing:  TAC_ADC < 3.2 ps. - Total electronics contribution at present:  Total_electronics ~ 7.2 ps. More to come soon….We hope to continue Fast DIRC detector work and confirm TOF performance in a test beam run this year….However, the future of test beams at SLAC is quite uncertain.

Blair Ratcliff5 th SuperB Workshop, Paris May Burle MCP-PMT bialkali photocathode 25μm pore MCP gain ~5×10 5 timing resolution ~70ps 64 pixels (8×8), 6.5mm pitch Typical Scanning System results (Burle )

Blair Ratcliff5 th SuperB Workshop, Paris May Typical Scanning System Results (Hamamatsu H-8500) Hamamatsu H-8500 Flat Panel Multianode PMT bialkali photocathode 12 stage metal channel dynode gain ~10 6 timing resolution ~140ps 64 pixels (8×8), 6.1mm pitch

Blair Ratcliff5 th SuperB Workshop, Paris May Timing versus Beam Position Hit time distribution for single PMT pixel in three positions. Position 1 direct mirror reflection Position 4 Position 6 hit time (ns) Mirror Expansion region Position 1 Position 4 Position 6

Blair Ratcliff5 th SuperB Workshop, Paris May Chromatic Broadening ΔTOP (ns) hit time (ns) Example: chromatic growth for one selected detector pixel in position 1 75cm path 870cm path σ narrow ≈ 170ps σ narrow ≈420ps First peak ~75cm photon path length Second peak ~870cm photon path length Important: careful calibration of all TDC channels to translate counts into ps Use accelerator trigger signal as event time Calculate the time of propagation assuming average ≈410nm Plot ΔTOP: measured minus expected time of propagation Fit to double-Gaussian Observe clear broadening of timing peak for mirror-reflected photons calculate from reco

Blair Ratcliff5 th SuperB Workshop, Paris May Burle MCP-PMT with 10 micron holes: sensitivity to magnetic field angular rotation wrt z axis ( B = 15kG)

Blair Ratcliff5 th SuperB Workshop, Paris May Hamamatsu H-9500 Hamamatsu H-9500 Flat Panel Multianode PMT bialkali photocathode 12 stage metal channel dynode gain ~10 6 typical timing resolution ~220ps 256 pixels (16×16), 3 mm pitch custom readout board – read out as 4×16 channels σ narrow ≈220ps Efficiency relative to Photonis PMT, 440nm, H-9500 at -1000V

Blair Ratcliff5 th SuperB Workshop, Paris May Timing in Magnetic Field (B=15 Kg)

Blair Ratcliff5 th SuperB Workshop, Paris May Chromatic Effects Chromatic effect at Cherenkov photon productioncos  c = 1/n(λ)  n(λ) refractive (phase) index of fused silica n=1.49…1.46 for photons observed in BABAR-DIRC (300…650nm) →  c γ = 835…815mrad Larger Cherenkov angle at production results in shorter photon path length → 10-20cm path effect for BABAR-DIRC (UV photons shorter path) Chromatic time dispersion during photon propagation in radiator bar Photons propagate in dispersive medium with group index n g for fused silica: n / n g = 0.95…0.99 Chromatic variation of n g results in time-of-propagation (ΔTOP) variation ΔTOP= | –L d  / c 0 · d 2 n/d 2 | (L: photon path, d  wavelength bandwidth) → 1-3ns ΔTOP effect for BABAR-DIRC (net effect: UV photons arrive later)

Blair Ratcliff5 th SuperB Workshop, Paris May Reconstruction Precisely measured detector pixel coordinates and beam parameters. → Pixel with hit (x det, y det, t hit ) defines 3D propagation vector in bar and Cherenkov photon properties (assuming average )  x,  y, cos  cos  cos  L path, n bounces,  c,  c, t propagation

Blair Ratcliff5 th SuperB Workshop, Paris May PID Software R&D Studies are Needed Need high quality physics MC simulations to demonstrate need for detector upgrades base-lined. (Realism is important, not only for PID simulation (both ID and Mis-ID rates), but also for interactions in material, tracking, CAL performance, etc.) Will need to be able to study physics tradeoffs between an added PID device and other detector systems. Is a separate system necessary or should dE/dx be revisited? Need specific physics channels that demonstrate compelling need for forward PID. Geometrical coverage plus momentum range: In particular, are there important physics channels that would benefit from enhanced PID momentum coverage range?