MEDIPIX: A Technology Developed at CERN, That Can Be Developed As An Active Real-time Space Radiation Dosimeter Lawrence Pinsky University of Houston Note that Michael Campbell, CERN (The Medipix2 Consortium Spokesperson) Is Also Present at this meeting... Slide Courtesy of Michael Campbell Slide 2 Important Disclaimers . Note that ALL DISCLOSURES regarding the Design and Performance of the MEDIPIX technology made in this presentation is to be considered CONFIDENTIAL to the extent that subsequent patent applications submitted. . The University of Houston is NOT presently a member of the Medipix Consortium, but rather we have beenformally invited by that Collaboration to join for the purpose of pursuing the adaptation of this technology to Space Radiation Dosimetry… Slide 3 The Medipix2 Consortium . Institut de Fisca d'Altes Energies, Barcelona, Spain . University of Cagliari and INFN Section thereof, Italy . CEA, Paris, France . CERN, Geneva, Switzerland, . Universitat Freiburg, Freiburg, Germany, . University of Glasgow, Scotland . Universita' di Napoli and INFN Section thereof, Italy . NIKHEF, Amsterdam, The Netherlands . University of Pisa and INFN Section thereof, Italy . University of Auvergne, Clermont Ferrand, France, . Laboratory of Molecular Biology, Cambridge England . Mitthogskolan, Sundsvall, Sweden, . Czech Technical University, Prague, Czech Republic . ESRF, Grenoble, France . Academy of Sciences of Czech Republic, Prague . Universität Erlangen-Nurnberg, Erlangen, Germany . University of Californiz-SSL, Berkeley, USA Slide 4 WHAT IS MEDIPIX2 DETECTOR? Medipix2 is a pixel-based detector technology that can be employed to measure charged particles, photons (visible through gammas), and neutrons. It is based on a read-out chip that embeds the electronics for each pixel within the pixel’s footprint! Outline of This Talk . The Medipix2 Chip and Readout System . Recent Heavy Ion Beam Medipix Exposures . Timepix—An Evolution Within Medipix2 . Medipix3—The Next Generation Medipix . Where Do We Go From Here With Medipix Slide 5 Hybrid Pixel Detector diagram of Hybrid Pixel Detector with labels: p-type silicon layer, high resistivity n-type silicon, sensor chip (e.g. silicon), aluminium layer, single pixel read-out cell, electronics chip, fligh chip bonding with solder bumps Detector and electronics readout are optimized separately Slide 6 Hybrid Pixel Detector - Cross Section Slide 7 Current Medipix2 Cell Schematic Charge sensitive preamplifier with individual leakage current compensation 2 discriminators with globally adjustable thresholds 3-bit local fine tuning of the threshold per discriminator 1 test and 1 mask bit External shutter activates the counter 13-bit pseudo-random counter 1 Overflow bit Analog section; Input; Ctest; Test Input; Testbit; Preamp; Disc1; Disc2; Vth Low; Vt hHigh; Maskbit Maskbit; 3 bits threshold; 3 bits threshold; Preamp; Disc1; Disc2; Digital section: Double Disc logic; 13 bits Shift Register; Shutter; Mux; Mux; ClockOut; Previous Pixel; Next Pixel; Con f8 bits configuration; Polarity; Double Disc logic; Slide 8 Current Medipix2 Cell Layout labels 55 µm; 55 µm; 503 transistors per pixel; 33M per Chip Slide 9 image of Bumps on the readout side (MCNC-RDIx200) 200 µm Slide 10 image of Bumps on the readout side – close up (MCNC-RDIx1.00k) 50.0 µm Slide 11 image of Medipix2 Si Assembly Slide 12 A new USB based Medipix2 Readout System USB1 compatible Developed by S. Pospisil et al. CTU, Prague images of Medipix2 Slide 13 Image of a dried anchovy W-tube, 35kV, 2.5mm Al, 20mGy Medipix Collaboration, CERN, 2004 Slide 14 Image of a fly W-tube, 14kV, 125µm Al, 5mm PMMA Medipix Collaboration, CERN, 2004 Slide 15 Image of a leaf (55Fe) Image taken ove r14 Hours (5.9 KeV e-) Note celophanetape used to holdl eaf in place… And air bubble Surrounding leaf… Medipix Collaboration, CERN, 2004 Slide 16 High rate images X-Ray movie at 5.5fps 512x512 pixels Uses 4-chip Quad detector Slide 17 Neutronography & Neutron Dosimetry Lukas Tlustos, Czech Technical University, Prague (To be developed by the University of Houston) • Detection of light elements due to different attenuation of neutrons in matter, strong attenuation by H -> organic materials • Conversion of thermal neutrons to heavy charged particles in 6Li converter layer • Reaction: 6Li + n -> sigma(2.05 MeV) + 3H (2.72 MeV) Cross section: 940 barns (0.0253 eV) image and diagram of Netron with labels: Neutron, Converter, Detector chip, Bump-bonding, Readout chip Slide 17 Monte-Carlo efficiency simulations Detection efficiency was simulated in dependence on the converter thickness • The neutron transport simulated using MCNP • Energy deposition and ionization computed by TRIM/SRIM graphs with labels: 6LiF, enrichment 90%; 940 barns (0.0253 eV); Amorphous 10B, enrichment 80%; 3840 barns (0.0253 eV) Slide 19 Spatial resolution estimation Spatial resolution is affected by: • Range of heavy charged particles in converter material – depends on density 6LiF (rho=1.6 g/cm3): R sub Triton=52µm, R alpha=10µm 10B (r=1.2 g/cm3): R sub Li=5µm, R sub alpha =7µm • Range in silicon 6LiF: R sub Triton=44.1µm, R sub alpha=8.6µm. 10B: R sub Li=3µm / 2.7µm, R sub alpha=5.4µm / 5.2µm • Charge sharing effect ? graphs and diagram with labels: p[gcm-3]; R [..m]; triton ; (2.72MeV, LiF); alpha (2.05MeV, LiF); R [µm]; lithium (1.01MeV); lithium (0.84MeV); alpha (1.78MeV); alpha (1.47MeV); alpha particle; Pixel row; Charge spread Slide 20 6LiF converter • Sensor covered by 6LiF layer (3mg/cm^2). Detection efficiency is about 3% • High energy of alpha particles and tritons is deposited near detector surface => charge sharing is significant. • Each hit creates signal in cluster of pixels. • Cluster size limits spatial resolution in integrating regime. • Cluster size can be decreased by high threshold at the expense of efficiency. • Using event-by-event acquisition and finding centroids of clusters it is possible to reach subpixel spatial resolution (approximately half of pixel) graph of Cluster size distribution for 6LiF converter. Exposition=50 x 0.001x, Vfbk=250, Vthl=205 graph of Clusters of 8LIF converter (Exposition=0.001s, Vfbk=250, Vthl=200) Slide 21 Amorphous 10B converter • Energy of heavy charged particles is lower than in case of 6Li converter => smaller clusters are produced. clusters are produced. • From gamma interactions electrons are generated => electron tracks are present. Spatial resolution is deteriorated by electron tracks. • Energy of electrons is lower then energy of heavy particles => electron tracks can be suppressed by suitable threshold selection. graph of Cluster size distribution for 10B converter (Exposition= 50 x 0.001s, Vfbk=250, Vthl=200) graph of Clusters of 10B converter (Exposition=0.001s, Vfbk=250, Vthl=200) Slide 22 Recent Heavy Ion Measurements with Medipix2 (@ HIMAC & TAMU) . In collaboration with Jack Miller’s Group at LBL, and Eric Benton, we recently made measurements with a Medipix2 of tracks in beams of: . Fe @ 500 Mev/A . O @ 290 MeV/A . Si @ 800 MeV/A . Ne@ 390 MeV/A . This past weekend, in collaboration with NASA/JSC/SRAG, measurements were made at Texas A&M’s cyclotron in beams of: . Xe@ 12 & 24 MeV/A . p @ 20, 30 & images of computer and lab Slide 23 Heavy Ion Images in Medipix2 imageof GUI window with label 60 Degree Incidence O @ 290 MeV/A graph with label: Image Slice Size v. LET for beams taken • Pixels indicate above lower threshold & below upper threshold... • These are effectively the energy slices through the track structure... Slide 24 Medipix Movie of Xe Beam from This Past Weekend at A& M Cyclotron Slide 25 100 Sec Integrated Medipix2 Images 1 image On my lap in a 777 airliner at 34,000 Feet over Anchorage, Alaska on the flight to Japan… 2nd image On the 15th floor of the Mitsui Garden Hotel in Chiba, Japan... Slide 26 alpha Source Slide 27 Adding Pulse Height Capability... Timepix Schematic for next the version of Medipix2 Analog: labels: Preamp; Disc; THR; Input; Ctest; Testbit; Test Input; Mask; 4 bits thr Adj; Polarity; Digital: labels: 14 bits Shift Register; Mux; Mux; Clk_Read; Previous Pixel; Next Pixel; Conf; 8 bits configuration; Clk_Count; Timepix Synchronization Logic; Clk_Countb; P0; P1; Shutter; Ovf Control; Clk_Read; Shutter_in; t Slide 28 Analog pixel summary Amplifer Gain ~18mV/Ke Peaking Timei 90ns…140ns (IPreamp) Pixel noise ~75e-sub rms Preamp DC Level (FBK) 800mV (e-) | 1.4V (h+) Threshold dispersion ~170e- Adjusted Threshold dispersion ~25e- Voltage linear range 0 to 50 Ke-(< 2%) TOT linear range >200Ke- Time Walk ~25ns (2Qth to infinite) TOTgain ~55ns/Ke-(Ikrum=5nA) Analog Pixel consumption (Max) 2.9µA x 2.2V = 6.38 µW (30% less than Mpix2MXR20) All these values are extracted from simulations !!! Slide 29 Timepix Synchronization Logic control Mask P1 P0 Mode 0 0 0 Masked 0 0 1 Masked 0 1 0 Masked 0 1 1 Masked 1 0 0 Medipx 1 0 1 TOT 1 1 0 Timepix-1hit 1 1 1 Timepix - Use of 3-bit High threshold adjustment bits for : 4th equalization bit and P0, P1. - Each pixel can be configured independently in 5 different independently in 5 different modes. - This logic needs 104 Trts(Mpix2MXR20 had 92 Trts) - Logic only consumes power only when a hit is present Slide 30 Timepix Mode (P0=1,P1=1) diagrams of of timepix at 10MHz and 100MHz Slide 31 TOT Mode (P0=1,P1=0) diagrams of TOT Mode at 10MHz and 100MHz Slide 32 Timepix proposed Floorplan labels: 55µm; 55µm; Preamp; Discriminator + 4bits; TSL + Clock Buffering; 14bit counter + Overflow control; 8-bits configuration register; Slide 33 Timepix - Summary . Timepix development is driven by TPC readout for the ILC (EUDET consortium) . Timepix will act as proof of principle for concept using existing Medixix2 readout system and software . Initially foreseen for Time of Flight, the chip is now programable to measure Time over Threshold… . Chip design is in final stages . Submission planned end of May, 2006 Slide 34 Medipix3 – The Next Generation . New pixel electronics taking care of charge diffusion –event-by-event clustering . Flexible architecture – 2 counters per pixel . Higher acquisition and frame rate with dead time free readout possible . Small prototype just in testing . Uses 0.13µm CMOS . Medipix3 will attempt to resolve many of the limitations of the Medipix2 system . Collaboration agreement ready for signature . Could be made Rad-Hard… Slide 35 Comparison Medipix2/Medipix3 - Si graph of Si, 55µm square x 300µm, 10 keV with labels: single pixel 1003-; 2x2 summing 200e-; au; energy [keV] graph of Si, 110µm square x 300µm, 10 keV with labels: single pixel 1003-; 2x2 summing 200e-; au; energy [keV] µt ~ 5.8e3 cm Vbias = 120V Slide 36 Where Do We Go From Here? . The University of Houston would like to join the Medipix Consortium for the purpose of developing a Space Radiation Dosimeter based on the Medipix Technology. Funding from NASA is needed to initiate this effort . The imminent availability of Timepix version of the Medipix2 chip offers an immediate opportunity to develop prototype flight hardware for evaluation. . Longer term, we would like to also join the new Medipix3 Consortium to participate in the the development of a robust versatile portable personal active Space Dosimeter… Slide 37 Thank You For Your Attention.. Real-time demo to follow… Time permitting Slide 38 Medipix2 Chip Architecture labels: IO Logic; LVDS Input; LVDS Output; 32-bit CMOS Output; 256-bit Fast Shift Register; 3328-bit Pixel Column-0; 13 8-bit DACs; 14111 µm; 3328-bit Pixel Column-1; 3328-bit Pixel Column-1; 16120 µm; 10 ms readout time; 300 µs readout time Slide 39 Tests of Medipix2 with alpha particles • Medipix-2 without converter layer • Alpha particles: 5.6 MeV(241Am) • Short exposition time • Circular clusters observed Cluster size distribution (241Am source, alphas of 5.6 MeV energy) graph with labels: Mean: 12.0286; Sigma: 1.10985; Cluster size [pixels] = energry Dependence of cluster size on threshold level graph with labels: Cluster size [pixel]; Vthl [bit] Slide 40 Tests with Thermal Neutrons . NEUTRA station of spallation neutron source SINQ in Paul Scherrer Institute, Villigen, Switzerland . Intensity about 3·10^6 neutrons/cm^2s at proton accelerator current of 1mA and proton energy of 590 MeV . Beam Cross section: 40 cm in diameter . Horizontal channel of the LVR-15 nuclear research reactor at Nuclear Physics Institute of the Czech Academy of Sciences at Rez near Prague. . Intensity is about 10^7neutrons/cm^2s (at reactor power of 8MW) . Beam Cross section: 4 mm (height) x 60 mm (width) . The divergence of the neutron beam is < 0.5° Slide 41 6LiF converter Spatial resolution – Edge response Tilted cadmium edge profile graph with Fit by ERF: sigma=0.83 pixel=> LSF FWHM=107 µm Edge blurring is caused by clusters => Spatial resolution is dependent on the threshold level Spatial resolution is limited by size of clusters and range of product particles in silicon (R sub Triton=44µm, R alpha=8.9µm) graph of Dependence of edge profile sigma on Vthl with labels Edge porfile sigma [pixel]; Vthl [bits] Slide 42 10B converter Spatial resolution – Edge response Tilted cadmium edge profile graph with labels Rel. Signal, column [pixel]; LSF. Fit by ERF: sigma=0.35 pixel=> LSF FWHM=µm Heavy charged particles emitted by 10B converter have shorter ranges then in case of 6Li. Their energies are also lower so charge sharing is less important. => Spatial resolution is better But lower number of particles can penetrate to depleted volume of the detector. => Efficiency is lower (approx. 2 times) Slide 43 113Cd converter • Only conversion electrons are usable for imaging. • Resolution highly deteriorated • Using event-by-event acquisition and robust track analyzing algorithm it is probably possible to increase resolution. graph of Simulation of electron interactions in Si in MCNP graph of Tracks from 113Cd converter Exposition=0.001s, Vfbk=250, Vthl=200 Slide 44 113Cd converter Spatial resolution – Edge response Tilted cadmium edge profile graph with labels: Rel. Signal; Column [pixel]; LSF. Fit by ERF: sigmas=13.1 pixel => LSF FWHM=1.7 mm 113Cd as converter in combination with Si detector is not good choice for position sensitive detection of neutrons. But thanks to the large cross section it can reach good detection efficiency especially in case of CdTe detectors. graph of Simulation of electron interactions in CdTe in MCNP. => Estimated LSF FWHM is 440 µm Slide 45 Comparison of Medipix-2 with other neutron imaging detectors Tested: - CCD camera with scintilator containing 6Li (pixel size 0.139 mm) - Imaging plate (excitation by neutrons, deexcitation by laser scanner followed by light emission, scanner pixel size 50µm) - Medipix-1 device with 6LiF converter - Medipix-2 device with 6LiF converter table and graph of Resolution of imagers Imagers Resolution (FWHM^1 Resolution^3 of LSF^2) [µm] [lp/mm] Medipix-1 device 370 2.5 Medipix-2 device 108 8.5 CCD camera 824 1.1 Imaging plate 124 7.3 Medipix-1 Pixel size: 170 µm Resolution: 370µm Medipix-2 Pixel size: 55 µm Resolution: 108 µm CCD + scintillator Pixel size: 139 µm Resolution: 824 µm Imaging plate Pixel size: 50 µm Resolution: 124 µm