Challenges with decay vertex detection in CBM using an ultra-thin pixel detector system M. Deveaux, S. Amar-Youcef, C. Dritsa, J. Heuser, I. Fröhlich, C. Müntz, C. Schrader, F. Rami, S. Seddiki, J. Stroth, T. Tischler, C. Trageser and B. Wiedemann on behalf the CBM collaboration. : Outline: The CBM experiment (a reminder) Requirements for the CBM vertex detector Technology choices and consequences Running conditions Summary and Conclusion M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

CBM@FAIR CBM: Fixed target heavy ion experiment 10-45 AGeV beam energy RICH TRD TOF ECAL MVD+STS Beam and Target CBM: Fixed target heavy ion experiment 10-45 AGeV beam energy Located at GSI/FAIR, Darmstadt, Germany First collision ~2015 M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

The Physics case of CBM Explore the nuclear phase diagram in the region of high net baryon density CBM M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Observables of interest Onset of chiral symmetry restoration at high B in-medium modifications of hadrons (,, e+e-(μ+μ-), D) Deconfinement phase transition at high B excitation function and flow of strangeness (K, , , , ) excitation function and flow of charm (J/ψ, ψ', D0, D, c) charmonium suppression, sequential for J/ψ and ψ' ? The equation-of-state at high B collective flow of identified hadrons particle production at threshold energies (open charm) QCD critical endpoint excitation function of event-by-event fluctuations (K/π,...) flow charm fluctuations dileptons Our topic for today M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Global layout of the CBM experiment electron ID: RICH & TRD  p suppression  104 muon ID: absorber + detector layer sandwich  move out absorbers for hadron runs MVD + STS aim: optimize setup to include both, electron and muon ID → would be the best crosscheck of results we can ever do! MVD: Micro Vertex Detector, STS: Main Silicon Tracking System RICH + TRD: electron identification, TOF: Time of Flight system for hadron identification, MUCH Muon detector (replacing the RICH), ECAL M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

The open charm challenge D : c ~ 312 m D+  K- + + (9.5%) D0 : c ~ 123 m D0  K- + (3.8%) D0  K- + + - (7.5%) Ds : c ~ 150 m D+s  K+ K- + (5.3%) Multiplicities / central collision SIS18 SIS100/ 300 +c : c ~ 60 m +c  p K- + (5.0%) [W. Cassing, E. Bratkovskaya, A. Sibirtsev, Nucl. Phys. A 691 (2001) 745] High interaction rate required => CBM is designed as fixed target experiment Beam intensity provided by SIS300: Up to 109 ions/s (slow extraction) Needs excellent secondary vertex resolution ~50 µm M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

How to get the resolution Hybrid pixels CMOS-Sensors Assuming first station 5cm behind the target. Simulation of the sec. vtx resolution of the MVD C. Dritsa Sec. vtx. resolution demands for thin sensors like CMOS-Sensors (MAPS) Can’t tolerate material of the beam pipe. Vertex detector must be placed in (moderate) vacuum. M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Minimum Ionizing Particle MOS Active Pixel Sensor CMOS-sensors (MAPS) Minimum Ionizing Particle MOS Active Pixel Sensor Features of the MIMOSA – detectors: Single point resolution 1.5µm - 2.5µm Pixel – pitch 10 - 40 µm Thinning achieved 50 - 120µm S/N for MIPs 20 – 40 Radiation hardness: 1MRad ; 1 x 1013 neq/cm² Time resolution ~ 20 µs (massive parallel readout) MIMOSA IV We follow the R&D for SoI-Detectors and 3D-VLSI-detectors M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

The time resolution of MAPS vs. collision rate On - chip zero suppression Discriminators Sensor array Bonding pads + output Amplis. Consecutive readout of pixels required => expected time resolution ~10 µs (~100 kFrame/s) Readout bus Pixel column Expected collision rates: Up to 109 heavy ions / sec (up to uranium @ 35 AGeV) 1% - target => Up to 107 collisions / s Expect pile up of nuclear collisions in the vertex detector M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Running conditions (1) ~130 hits / evt Y (cm) X (cm) Average hit density from nuclear collisions (@ 10cm) Hits/mm²/coll. x 1000 X (cm) Y (cm) ~130 hits / evt Typical non-central collision M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Hot spots up to 1 hit / evt / mm2 Running conditions (1) Mean hit density from nuclear collisions (@ 10cm) Incl. delta electrons from the target ~700 hits / evt Hits/mm²/coll. x 1000 Rate Hits (Max.) 100 kHz => 0.7k (1/mm²) 1 MHz => 7k (10/mm²) 10 MHz => 70k (100/mm²) Adapt beam to abilities of the detector! Hot spots up to 1 hit / evt / mm2 M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Running conditions Hits/event distance between station and target[cm] The occupancy is dominated by delta electrons generated in the target. Handling them needs detector with very high granularity M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Running conditions and consequences Occupancy estimate: Fired pixels / hit (from Mimosa16 beam test) Pixel pitch: ~ 15 µm => 4.4k pixel /mm² Assume collision rate of ~ 1MHz for the first years + 5 fired pixels / hit: 1.2 % occupancy @ 10cm 4.0 % occupancy @ 5cm IPHC-Strasbourg CEA-Saclay 3.5 mV 7.8 mV A. Besson (modified) Data flow estimate: Assume on-chip zero suppression: 40-80 bit/cluster (including addresses) 1 MHz collision rate ~ 5-10 GB/s @ 10 cm ~ 8-16 GB/s @ 5 cm Tracking? Start in fast detectors of the STS, Extrapolate tracks to the MVD. Validated excluding Delta-electrons To be confirmed accounting for them M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Radiation dose and tolerance Full collision rate => few 1015 neq/year Radiation tolerance of MAPS Mi-9 Mi-15 Mi-18 z=5 cm z=10 cm Assumption today: Detector may tolerate 1013 neq/cm² => two weeks beam on target (1 MHz) ~ 5 x 1011 coll. @ 5 cm ~ 1 x 1012 coll. @ 10 cm => Replace detector after one beam time M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Artistic view on the CBM-MVD How it might look like Artistic view on the CBM-MVD Power: ~ 1W /cm² S. Belogurov et. al Cool with heat conduction: TPG - ~1400 W/(m K), 3-4 times better than Cu M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Preliminary Material budget? Assuming 15 µm pitch Acceptable area More speed needs more power => More material for cooling & cables Larger diameter needs more pixels => Even more power & cables Larger diameter needs longer ways for heat conduction => More material Time resolution might need fine tuning accounting for material budget M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Some simulation results (25 AGeV) Multiplicity based on HSD model M ~ 4.17 10-5 BR = 9.51% ε D+ = 1% and S/B2σ = 0.4 We expect ~ 5k (up to few 10k) reconstructed D+ within 2 weeks of measurement (1012 coll. at 106 coll./s, 25 AGeV). λC can be seen. Expect better results for 35 AGeV (higher multiplicities) But: Material budget of the detector must not exceed ~0.3% X0 per layer M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Research lines for the MVD mm Dose [neq / cm2 / coll. ] BEAM Delta electrons produced in the target Mechanical system integration R&D @ Frankfurt Simulation studies @ GSI + IPHC The CBM Micro Vertex Detector Design studies Electronic system integration R&D @ Frankfurt MAPS R&D @ IPHC Rad. hardness studies: Frankfurt, IPHC, FRM II M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Summary and conclusion The CBM experiment aims for a first time to measure open charm being emitted from heavy ion collisions at beam energies between 10 and 45 AGeV This requires a vertex detector with: very low material budget (~0.3 % X0 /layer) good radiation hardness (minimum 1013 neq/cm², optimal > 1015 neq/cm²) good time resolution (minimum ~10µs, optimal < 100 ns) good spatial resolution (~ 5 µm or better) vacuum compatible design and cooling Currently we consider CMOS sensors (by IPHC Strasbourg) as guideline technology Excessive R&D seems required to match the requirements. R&D on radiation hardness of CMOS sensors, mechanical and electronical detector integration have started at GSI (Darmstadt) and Goethe University Frankfurt We will presumably build our detector in two phases: A CMOS sensor based vtx detector operating roughly 2015 (min. performance) A novel technology based detector operating few years later Challenges in vertex detector design? We have plenty of them  We are open to alternative technologies, advices and collaborators. M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

The CBM collaboration 46 institutions > 400 members Russia: IHEP Protvino INR Troitzk ITEP Moscow KRI, St. Petersburg China: CCNU Wuhan USTC Hefei Croatia: RBI, Zagreb Portugal: LIP Coimbra Romania: NIPNE Bucharest Poland: Krakow Univ. Warsaw Univ. Silesia Univ. Katowice Nucl. Phys. Inst. Krakow LIT, JINR Dubna MEPHI Moscow Obninsk State Univ. PNPI Gatchina SINP, Moscow State Univ. St. Petersburg Polytec. U. Ukraine: Shevchenko Univ. , Kiev Cyprus: Nikosia Univ. Univ. Mannheim Univ. Münster FZ Rossendorf GSI Darmstadt Czech Republic: CAS, Rez Techn. Univ. Prague France: IPHC Strasbourg Germany: Univ. Heidelberg, Phys. Inst. Univ. HD, Kirchhoff Inst. Univ. Frankfurt Univ. Kaiserslautern Hungaria: KFKI Budapest Eötvös Univ. Budapest India: VECC Kolkata SAHA Kolkata IOP Bhubaneswar Univ. Chandigarh Univ. Varanasi IlT Kharagpur Korea: Korea Univ. Seoul Pusan National Univ. Norway: Univ. Bergen Kurchatov Inst. Moscow LHE, JINR Dubna LPP, JINR Dubna 46 institutions > 400 members Strasbourg, September 2006 M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Backup M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

First vacuum compatibility tests High-tech prototype ladder for vacuum tests (silicon glued on TPG) M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Radiation hardness of MAPS M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Outlook Radiation hardness Non-ionizing radiation hardness: Limitation: Slow charge collection in the sensor + reduced lifetime of signal electrons Approach (1): Speed up charge collection: Diode Diode P+ => P- P+ P- P+ P+ P- P+ Pixel with gradient doping. Attractive build in potential? M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Outlook Radiation hardness Approach (2): Restore lifetime of electrons Recombination destroys the signal Efficient approach for depleted N-doped detectors. BUT: MAPS are P-doped and undepleted. M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

The power dissipation of MAPS On - chip cluster-finding processor Discriminators Sensor array Bonding pads + output Amplis. Readout bus Power scales with occupency. Neglegted in the follwing Pixel column Only one pixel / col. ac-tive, low power dissipation Permanently biased, high consumption. 1 cell : „Pcol“ Basic assumption: Power scales with number of the end of column logics (discriminator + cluster finding processor) M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

The power dissipation of MAPS f pixels / s p p Time resolution of MAPS Readout freq. (Pixels / time) Power of one logic cell The surface of one collumn is: With: pixels M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

The material buget of the MVD Details: See talk of M.Deveaux, MVD-working group M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Correlations of the thickness (Simplified) Big stations are thick stations Small pixels make thick Fast pixels save material Fast frame rates make thick stations Time resolution of MAPS Readout freq. (Pixels / time) Pixel pitch M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Correlation of thickness Material budget (revised) Material budget according to earlier studies 3000 D0/lifetime, S/B=0.3, Det. Eff. 0.7% @ Au+Au 25AGeV (preliminary) z = 10 cm M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Electronic system integration studies (Readout concept) Demo-Aux-PCB @ End of the ladder Voltage regulation Signal buffering Single ended signal => differential signal Mi17 Clock, bias, slow control Sync. signals Analogue output 16 x 50 MHz <=> 1.2 GB/s FPC Up to 280 pins Analog output Sync Clock JTAG ADC + FPGA Bias (8V DC) PC Data transfer Online monitoring Reset Data processing/reduction M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

How to study things more in detail => R&D in Frankfurt MAPS clock, fast control MAPS Analog signal MAPS slow control MAPS sync 4 x ADC 12-bit, 50 MHz Digital front-end Virtex 4 FPGA 512 MB SDRAM Bases on existing USB-FPGA board (IPHC, Strasbourg) Allows for: FPGA-based data sparsification (CDS, clustering, data compression) Pipeline readout of MIMOSA-17 (~ 1 ms time resolution) without dead time Add-on board 12 Gbit/s Optical link 2 Gbit/s TRB-II Sparsified data DAQ M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN

Needs still to be scaled with pile – up (up to factor 100) Running conditions Needs still to be scaled with pile – up (up to factor 100) M. Deveaux, Vertex 2008, 27 July 2008 to 01 August 2008 , Utö, Stockholm, SWEDEN