ITS Upgrade Proposal CERN LHCC Upgrade Session 20

ITS Upgrade Proposal CERN, LHCC Upgrade Session, 20 March 2012 L. Musa for the ALICE Collaboration Outline Overview on ALICE Upgrade Strategy ITS Upgrade Proposal o Physics motivations and design goals o Detector performance studies o Benchmark channels o Technical implementation and R&D roadmap o Conclusions 20 March 2012 L. Musa 1

Beyond the ALICE approved physics programme Progress on the characterization of QGP properties is made by studying multi-differential observables: ¦ Flavour, Centrality, Transverse momentum, Reaction plane, … ¦ This requires high statistics (high luminosity) ¦ Physics plans focused on physics observables where ALICE is unique at (PID, low material thickness, precise vertexing down to low pt, …), such as • precision measurements of spectra, correlations and flow of heavy flavour hadrons and quarkonia at low transverse momenta • precision measurements of low-mass lepton pairs emitted from the QGP • energy loss and flavour tagging of partons in the QGP via g-jet and jet-jet with hadron PID • search for the existence of exotic objects such as the H dibaryon or L-neutron bound states and systematic study of production of anti-matter This requires high statistics and precision measurements Standard trigger strategy not applicable in most cases 20 March 2012 L. Musa 2

ALICE Upgrade Strategy “ALICE at High Rate” submitted to LHCC o A document has been prepared defining the physics goals and the experimental approach for a run of at least 10 nb-1 with Pb. Pb • run ALICE at high rates, 50 k. Hz Pb-Pb (i. e. L = 6 x 1027 cm-1 s-1), with minimum bias (pipeline) readout ( max readout with present ALICE set-up ~500 Hz ) • Improve vertexing and tracking at low pt The Pb-Pb run would be complemented by p-Pb and pp running o It entails building • New beam pipe • New silicon tracker (scope and rate upgrade) • High-rate upgrade for the readout of TPC, TRD, TOF, CALs, Muons, DAQ/HLT ¦ This will allow a data driven readout architecture with continuous readout and event selection done by software algorithms in the online systems (DAQ/HLT) o Targets LS 2 20 March 2012 L. Musa 3

ALICE at High Rate - Readout Architecture • 50 k. Hz Pb–Pb collisions inspected with the least possible bias with online event selection based on topological and PID criteria max readout with present ALICE set-up ~500 Hz • HI run 2011: online cluster finding data compression factor of ~5 for TPC • Two HLT scenarios for the upgrade: 1. Partial event reconstruction: Factor of 5 (in use) → Rate to tape: 5 k. Hz 2. Full event reconstruction: Overall data reduction by a factor of 25 → Rate to tape: 25 k. Hz – min. bias event size ~20 MB ~4 -1 MB after data reduction – throughput to mass storage: 20 GB/s. • Event rate reduction from 50 k. Hz to 5 k. Hz (25 k. Hz) can only be reached with online event reconstruction and selection 20 March 2012 L. Musa 4

Readout and Online Systems Architecture Fast Trigger Processor (FTP) inter. t 0 Event Processors (EPN) Readout scheme L 0 L 1 L 2 L 3 t 0+ ~1 s t 0+10 s t 0+ ~1 s t 0+ >10 s FTP Fast Trigger Processor FLP First Level Processor EPN Event processor Node 20 March 2012 L. Musa 5

ALICE Upgrade Strategy o Contextually, submitted to LHCC the CDR for the ITS upgrade which is an essential part of the General Strategy o Furthermore, three major proposals are under consideration to extend the scope of ALICE: VHMPID, MFT, and FOCAL (a decision will be taken by September) — New high momentum PID capabilities — b-tagging for low pt J/psi and low-mass di-muons at forward rapidities — Low-x physics with identified g/ 0 o Meanwhile, R&D program continues for the different proposed upgrades, and the negotiations with the Funding Agencies to define the resource boundaries have been launched 20 March 2012 L. Musa 6

ALICE New Inner Tracking System Conceptual Design Report 20 March 2012 L. Musa 7

Physics measurments – examples where ITS plays a key role 1. Study thermalization of heavy quarks in the QGP • Measurement of baryon/meson for charm (Lc /D) and possibly for beauty (Lb /B) for pt > 2 Ge. V/c • Elliptic flow for B and HF baryons (Lc , Lb ? ) for pt > 2 Ge. V/c • Possible in-medium thermal production of charm quarks (D down to pt = 0) 2. Study of the quark mass dependence of energy loss in the QGP • Nuclear modification factors RAA of the pt distributions of D and B mesons separately o Beauty via displaced D 0 K o Beauty via displaced J/y ee 20 March 2012 L. Musa 8

Heavy Flavour – Experimental measurement Topological Identification of open charm Particle Decay Channel c ( m) D 0 K + (3. 8%) 123 D+ K + + (9. 5%) 312 K+ K + (5. 2%) 150 p K + (5. 0%) 60 D 0 K- + • Analysis based on decay topology and invariant mass technique • Essential selection cuts - impact parameter, distance of closest approach - distance of secondary vertex to primary vertex - pointing angle • High precision tracking (ITS+TPC) • p/K/ identification (TPC+TOF) reducing background at low p. T 20 March 2012 L. Musa Impact parameter resolution 9

Heavy Flavour – Experimental measurement Current detectors: – ALICE uniqueness: PID ( charm); low pt (low material budget); – Present limits: • charm difficult for pt 0 (background is too large); • resolution not sufficient for charmed baryons (Lc ct=1/2 D 0=1/5 D+); • Lc impossible in Pb-Pb collisions , at the limit in pp (only high pt); • Lb impossible in Pb-Pb collisions (insufficient statistics and resolution) • B/D separation difficult, especially at low pt (e PID + vertexing); • indirect B measurement via electrons; – CMS limits: • much larger material thickness (even after upgrade) • minimum pt at about 5 -6 Ge. V/c? • no PID low pt charm and Lc very difficult; 20 March 2012 L. Musa 10

Design goals 1. Improve impact parameter resolution by a factor of ~3 • Get closer to IP • Reduce material budget • Reduce pixel size 2. High standalone tracking efficiency and pt resolution • Increase granularity • Increase radial extension 3. Fast readout • readout of Pb-Pb interactions at > 50 k. Hz and pp interactions at > 2 MHz 4. Fast insertion/removal for yearly maintenance • possibility to replace non functioning detector modules during yearly winter shutdown 20 March 2012 L. Musa 11

How to improve the impact parameter resolution I) Get closer to the IP • radius of innermost pixel layer is constrained by central beam pipe Present beam pipe: ROUT = 29. 8 mm, DR = 0. 8 mm New reduced beam pipe: ROUT = 19. 8 mm, DR = 0. 8 mm II) Reduce material budget (especially inner layers) • present ITS: X/X 0 ~1. 14% per layer • target value for new ITS: X/X 0 ~0. 3 – 0. 5% per layer (STAR HFT 0. 37% per layer) reduce mass of silicon, electrical bus (power and signals), cooling, mechanics III) Reduce pixel size • currently 50 m x 425 m monolithic pixels O(20 m x 20 m), hybrid pixels O(30 m x 30 m), state-of-the-art O(50 m x 50 m) 20 March 2012 L. Musa 12

How to improve standalone tracking efficiency and pt resolution Higher granularity • increase number of layers in the outer region (seeding) and inner region (high occ. ) o present detector: 6 layers, optimized for track matching with TPC o new detector: 7 layers (assuming 95% efficiency) • increase granularity of central and outer layers o pixels 20 m x 20 m o Combination of pixels (20 m, 20 m) and strips (90 m, 20 mm) Increase radial extension • present detector: 39 mm – 430 mm • new detector: 22 mm – 430 mm(*) (CDR value) (*) increasing outer radius to 500 mm results in a 10% improvement in p 20 March 2012 L. Musa t resolution 13

Upgrade options Two design options are being studied A. 7 layers of pixel detectors • better standalone tracking efficiency and pt resolution • worse PID B. 3 inner layers of pixel detectors and 4 outer layers of strip detectors • worse standalone tracking efficiency and momentum resolution • better PID Option B 4 layers of strips Option A 7 layers of pixels 3 layers of pixels Pixels: O(20 x 20µm 2 – 50 x 50µm 2) > 685 krad/ 1013 neq per year Includes safety factor > 4 20 March 2012 L. Musa Pixels: O( 20 x 20µm 2 – 50 x 50µm 2) Strips: 95 µm x 2 cm, double sided 14

Improvement of impact parameter resolution MAPS Case Simulations for two upgrade layouts Option A: “All New” – Pixels (7 pixel layers) radial positions (cm): 2. 2, 2. 8, 3. 6, 20, 22, 41, 43 • Resolutions: srf = 4 m, sz = 4 m for all layers • Material budget: X/X 0 = 0. 3% for all layers Same for both layouts Option B: “All New” Pixel/Strips (3 layers of pixels + 4 layers of strips) • Resolutions: srf = 4 m, sz = 4 m for pixels srf = 20 m, sz = 830 m for strips • Material budget: X/X 0 = 0. 3% for pixels X/X 0 = 0. 83% for strips 20 March 2012 L. Musa 15

ITS standalone tracking efficiency MAPS Case Central events at sqrt(s. NN) = 5. 5 Te. V Simulations for two upgrade layouts radial positions (cm): 2. 2, 2. 8, 3. 6, 20, 22, 41, 43 Option A (all pixel layers) • Resolutions: srf = 4 m, sz = 4 m for all layers • Material budget: X/X 0 = 0. 3% for all layers Same for both layouts Option B (3 layers of pixels + 4 layers of strips) • Resolutions: srf = 4 m, sz = 4 m for pixels srf = 20 m, sz = 830 m for strips • Material budget: X/X 0 = 0. 3% for pixels X/X 0 = 0. 83% for strips 20 March 2012 L. Musa 16

Performance studies for heavy flavor detection Benchmark channels presented in the CDR • Charm meson production via D 0 → K−π+ • Charm baryon production via Λc → p. K−π+ • Beauty production via B → D 0 (→ K−π+) • Beauty production via B → J∕ψ (→ e+e−) • Beauty production via B → e+ Simulation method • Fast simulation scheme based on MC productions including detailed geometry and response of current ALICE detector. • Impact of new ITS by recomputing reconstructed track parameters by means of a simple scaling of the residuals of the impact parameters in rf and z, as well as of the transverse momentum, with respect to their true values (Fast Estimation Tool) 20 March 2012 L. Musa 17

Some examples of physics performance (D 0) Further optimization In progress Benchmark channel at very low pt: D 0 K S/B improved x 5 - Assuming ~ 109 central events: Significance better than 100 in all pt bins 20 March 2012 L. Musa 18

Some examples of physics performance (Lc) Lc p. K as benchmark channel in Pb-Pb Further optimization In progress Assuming ~ 1. 7 x 1010 central events (10 The background level in the Lc mass region will be determined precisely using the invariant mass side-bands nb-1) in centrality class 0 -20% pt(Ge. V/c) significance (3 s) Lc production measurable down to 2 Ge. V/c: 2 -4 4 -6 6 -8 v 2 for pt > 4 Ge. V/c in semi-central collisions (e. g. 30 -50%) 20 March 2012 7 40 53 RCP and Lc /Dfor pt> 2 Ge. V/c in central collisions v 2 for pt > 2 Ge. V/c in peripheral (e. g. 60 -80%) L. Musa 19

Detector technologies Several technologies are being considered for pixel detectors Hybrid pixel detectors • Edgeless sensors (100 m) + front-end chip (50 m) in 130 nm CMOS Monolithic pixel detectors • • • MIMOSA like in 180 nm CMOS INMAPS in 180 nm CMOS Le. Pix in 130 nm CMOS Option B 4 layers of strips Option A 7 layers of pixels 3 layers of pixels Pixels: O(20 x 20µm 2 – 50 x 50µm 2) > 685 krad/ 1013 neq per year Includes safety factor > 4 20 March 2012 L. Musa Pixels: O( 20 x 20µm 2 – 50 x 50µm 2) Strips: 95 µm x 2 cm, double sided 20

Pixel Technologies Hybrid pixels ¦ 2 components: CMOS chip and high-resistivity (4 -8 k. Wcm) sensor connected via bump bonds ¦ Optimize readout chip and sensor separately with in -pixel signal processing ¦ Charge collection by drift Monolithic pixels ¦ All-in-one, detector-connection-readout ¦ sensing layer (resistivity ~1 k. Wcm epitaxial layer) included in the CMOS chip ¦ Charge collection mostly by diffusion (MAPS) Figure - Rossi, L. , Fischer, P. , Rohe, T. & Wermes, N. (2006). Berlin: Springer. intrinsically more prone to signal loss due to charge trapping by radiation induced defects … but some development based on charge collection by drift ¦ Made significant progress, soon to be installed in STAR (Heavy Flavor Tracker) 20 March 2012 L. Musa Figure Stanitzki, M. (2010). Nucl. Instr. and Meth. A doi: 10. 1016/j. nima. 2010. 11. 166 21

Monolithic Pixels - MIMOSA ¦ State-of-the-art (MIMOSA family – IPHC Strasbourg) uses rolling-shutter readout • Continuous charge collection (mostly by diffusion) inside the pixel • Charge collection time ~200 ns • Pixel matrix read periodically row by row: column parallel readout with end of column discriminators (integration time readout period ) • Integration time (~180 s for ULTIMATE chip) ¦ Pixel size ~ 20 µm ¦ Low power consumption: only one row is powered at time ¦ Total material budget X/X 0 ~ 0. 3 % per layer (STAR HFT detector) ¦ R&D for ITS upgrade: MISTRAL MImosa Sensor for the TRacker of ALICE ULTIMATE sensor for STAR HFT further development of the low-power rolling shutter architecture of ULTIMATE o reduce readout time (20 -40 s) improve radiation resistance by a factor ~3 AMS 0. 35 m Tower. Jazz 0. 18 m o o Target power consumption: < 250 m. W / cm 2 20 March 2012 L. Musa 22

Monolithic Pixels – MISTRAL Tower-Jazz CMOS 0. 18 m - submissions • 2 designs for MPW run completed: – MIMOSA 32 designed by IPHC submitted in Nov. 2011 – delivered (Jan 2012) – Monalice. T 1 test chip designed by CERN/CCNU/IPHC to be submitted in Feb 2012 • Goal for 2012: – Evaluation of the technology (detection efficiency, S/N, quadrupole-well, . . ) – Test of radiation hardness 20 March 2012 L. Musa 23

Monolithic Pixels – Evaluation of Tower. Jazz technology MIMOSA 32 (IPHC) • Digital and analog blocks • Analog blocks (2 T and 3 T structures with various diodes) • Source tests and testbeam • 100 circuits delivered January 2012 5. 7 mm MONALICET 1 (CERN/CCNU) 3. 7 mm • Single transistors • Digital structures 4. 1 mm • Memories 2. 2 mm • Breakdown structures • Shift register 20 March 2012 CMOS test structures w/o deep p-well L. Musa Breakdown diodes 24

Monolithic Pixels - INMAPS ¦ In-pixel signal processing using an extension (deep p-well) of a triple-well 0. 18 m CMOS process developed by RAL in collaboration with Tower. Jazz (owner of the technology) • CMOS Imaging Sensor with additional deep p-well implant • CMOS electronics in the pixel compatible with charge collected 100% by in-pixel diode • charge collection (mostly) by diffusion New development dedicated to ITS upgrade started in 2012 (Daresbury, Birmingham, RAL, … - ARACHNID Collaboration) ¦ R&D objectives TPAC prototype 50 µm pixel - over 150 CMOS transistors • Verify radiation resistance • Reduce power consumption exploiting detector duty cycle (5% for 50 k. Hz int. rate) • Develop fast sparse readout RAL Irradiation test structrues (Single transistors w/o deep p-well) 20 March 2012 L. Musa 25

Monolithic Pixels – Le Pix Le. PIX: monolithic detectors in advanced CMOS Scope: • Develop monolithic pixel detectors integrating readout and detecting elements by porting standard 90 nm CMOS to wafers with resistivity of about 1 k. W cm • Reverse bias of up to 100 V to collect signal charge by drift Key parameters: • Pixel ~10 x 10 m 2 • Very large signal to noise ratio • Signal processing at end of column Track in telescope of 4 planes (beam test at PSI – Nov 2011) • ORTHOPIX readout architecture: multiple projections in a pixel matrix beyond X and Y. The method compresses the hit information by a factor of about 100 to a fixed data size and moves it to the periphery immediately (within one clock cycle) Low power consumption and reduced digital circuitry 20 March 2012 L. Musa 26

Le. PIX: Fe-55 spectrum and charge collection § Fe-55 spectrum at room temperature § Clearly two peaks: (5. 90 & 6. 45 ke. V) § Sigma ~ 140 e. V or 40 e RMS (for a 10 microsecond integration time) § Illustrates S/N potential of the device § Illuminating the back side with a laser confirms that signal charge is collected only from the pixel center, not from the pixel corners. This leads to a 70 % efficiency for MIPs. Simulations and focused ion beam repair are in progress to fully understand fix this issue § The collected charge increases with reverse bias 20 March 2012 L. Musa 27

Hybrid Pixels and Ongoing R&D ¦ Current flip-chip bonding technology permits pitch ~ 50 µm 30 µm seems achievable but R&D is needed ¦ Total material budget target X/ X 0 < 0. 5% (100 µm sensor, 50 µm chip) ¦ High S/N ratio, ~ 8000 e-h pairs/MIP S/N > 50 ¦ Edgeless sensors to reduce insensitive overlap regions (~ 20µm) ¦ Possible to operate at room temperature in high radiation environment up to 10 13 neq ¦ Cost driven by fine-pitch bump-bonding of sensor to readout chip • Could be considered for the three inner layers ¦ R&D: ¦ Assembly of ultra-thin components ¦ Finer pitch bump bonding ¦ Edgeless detectors ¦ FEE chip floor-plan optimization ¦ Power/Speed optimization Sensor 100 m, readout chip 50 m, glass carrier 300 m possible (shaping time O(µs)) to reduce power budget 20 March 2012 L. Musa 28

Hybrid Pixels – example of R&D on thin assemblies sensor 100 m FEE chip 50 m glass carrier FEE CHIP 50 m glue glass carrier 20 March 2012 L. Musa 29

Strip Sensor Design The sensor of the present detector Layout: 300 µm thick, double-sided, 768 strip/side, 35 mrad stereo angle Sensor area: 0. 0028 m 2 -- Cell size: 95 um x 40000 um The new strip sensor layout has been designed (Trieste) decrease the strip length from ~40 mm to 20 mm cell size ~ -50% Cstrip ~ -50% 2 x # of channels same cluster size 2 rows of strips per sensor side d t f ra o New front-end chip with integrated ADC 20 March 2012 L. Musa 30

Preliminary Studies on the mechanical structure CONSTRAINTS TO THE MECHANICAL DESIGN External Detector integration and accessibility New Beam Pipe side C Internal X 0 radiation length Sensor Type Sensor Power Dissipation Sensor Distribution The new ITS design should ensure a rapid accessibility to the inner detectors ❶ No TPC moving side A ❷ Services only at one side, side A ❸ Detector in two half 20 March 2012 L. Musa 31

Preliminary Studies on Mechanical Structure Several concepts are being studied STAR PXL detector centered in the Middle Support Cylinder single end support allows rapid insertion and removal ALICE New ITS Ultra-Light Concept ALICE New ITS Clam-shell Concept Cooling and support mechanics strongly coupled 20 March 2012 L. Musa 32

Mechanical Structure – foam shell concept • Developed on a solid cooling concept extensively tested in similar application (Atlas, Panda) • Modularity stave level • Stiff ends • Carbon fiber + Carbon foam • Various option for pipe material Carbon fiber Carbon foam Sensor + electrical bus Prototype 20 March 2012 L. Musa 33

Cooling – liquid with polymide micro-channels Multilayer polyimide composite • A layer of Pyralux LF 110 at the bottom • Photoimageable PC 1020 layer in the middle • Pyralux LF 7001 layer glued on the top. The multilayer PC 1020, which is 200 μm thick, is glued on the LF 110. The rectangular pattern which defines the channels is created with a photolithography process, and the foil of LF 7001 is hot pressed on the top of the substrate to cover the channels. 20 March 2012 L. Musa 34

Cooling – Evaporative with silicon micro-channels Cooling technique developed by CERN/PH-DT with Microsystems Laboratory (EPFL) and CSEM (Neuchatel) Considered by NA 62 for Giga-Tracker detector o Detector and cooling structure are both silicon reduce mechanical stress due to thermal stress o Micro-channels (100 x 200 mm 2) etched on a silicon substrate (130 mm thick) closed by a cover-layer fusion bonded on top 20 March 2012 L. Musa 35

Mechanical Structure – U-light shell concept • Structure based on CFRP with possibility to integrate different cooling concepts (silicon micro-channels, polymide channels) • Light structure with openings along the sensitive region • Modularity from stave level can be reduced to half barrel with gain in stiffness and reduction in material Sensor Electrcial bus cooling Prototype under construction 20 March 2012 L. Musa 36

Mechanical Structure – U-light shell concept …to Half Barrel from Single Stave to Half Layer • The external (L 3) end-stiffening surface shall be a complete honeycomb shell (half cylinder) • This external shell, linked through the endcap to the different layers, provides the overall structure stiffness 20 March 2012 L. Musa 37

Mechanical Structure – U-light shell concept Endcap cooling ONLY Based on sensor final choice, i. e. minimum power dissipation, the possibility to have a cooling system only in the ITS Layer endcap is being considered. Heat will be tranferred by high thermal conductive material along the stave 20 March 2012 L. Musa 38

Mechanical Structure – Clam shell Concept • Move all support to the outside shell(s) • Clam-shell structure (i. e. two halves) • Low-z materials – carbon fiber composite (Xo=25 cm) – Airex (Xo=1326 cm) – Tedlar film (Xo=26. 2 cm) • air cooling scheme 20 March 2012 L. Musa 39

Installation mechanics • Three beams made of carbon fiber are fixed to the inner surface of the TPC to support/guide the insertion of the outer barrel • The outer barrel (split in two halves) is mounted outside the TPC before being moved to final position • The inner barrel (also in two halves) is installed at the end Inner barrel insertion Cut view after installation 20 March 2012 L. Musa 40

Upgrade Studies - organization Four working groups to study the upgrade and coordinate R&D activties I) Physics Motivations and Detector Functional requirements (convenors: A. Dainese, G. Usai) II) Detector Specifications and Performance Simulations (convenors: G. Bruno, M. Sitta) III) Detector design and implementation (convenors: P. Riedler, A. Rivetti, G. Contin) IV) Cooling, Cabling, Services, mechanics and integration (convenors: A. Tauro, R. Santoro) CDR preparation editorial board: L. Musa, V. Manzari, G. Usai, A. Dainese, G. Bruno, P. Riedler, G. Contin, A. Rivetti, R. Santoro, A. Tauro, R. Lemmon, S. Rossegger contributing authors: C. Di Giglio, M. Kweon, M. Mager, A. Mastroserio, S. Moretto, A. Rossi, C. Terrevoli, S. Bufalino, S. Piano, F. Prino, S. Senyukov, R. Shahoyan, L. Bosisio, M. Campbell, C. Cavicchioli, T. Kugatashan, W. Snoeys, M. Winter, G. Aglieri Rinella, R. Turchetta, C. Pastore, I. Sgura, E. Da Riva, C. Bortolin, A. Mapelli, S. Coli 20 March 2012 L. Musa 41

Institutes that participate in the upgrade studies Institutes in the present ITS Project • About 10 new Institutes joined the ITS project to participate in the upgrade studies and the preparation of the CDR CERN, INFN, St-Petersburg, Kharkow - long standing expertise in • ASIC design • construction of detector ladder and support mechanics • manufacturing of composite materials • integration and characterization of hybrid pixel and microstrip IPHC, RAL - among world leaders • Monolithic pixels detectos 20 March 2012 L. Musa 42

Project timeline (to be adapted according to LHC schedule) 2012 – 2014 R&D o 2012 finalization of detector specifications evaluation of detector technologies (radiation and beam tests) first prototypes of sensors, ASICS, and ladders (demonstrators) o 2013 selection of technologies and full validation engineered prototypes (sensors, ASICs, ladders, data links) engineered design for support mechanics and services Technical Design Report 2014 final design and validation start procurement of components o 2015 -16 2017 2018 20 March 2012 production, construction and test of detector modules assembly and pre-commissioning in clean room installation in ALICE L. Musa 43

Conclusions In line with the ALICE general upgrade strategy, we propose to build a new ITS based on 7 silicon layers characterized by • Continuous readout • Factor ~3 improvement in impact parameter resolution • Very high standalone tracking efficiency down to low pt (> 95% for pt > 200 Me. V/c) • Fast access (winter shutdown) for maintenance interventions After a couple of years of studies, we are confident that this ambitious proposal can be turned into a real detector to be ready for physics in 2019 Strong support from Funding Agencies for R&D phase We wish it can be reviewed and approved by the LHCC such that the we can secure the necessary resources 20 March 2012 L. Musa 44

BACKUP SLIDES 20 March 2012 L. Musa 45

ITS PID performance 4 layers silicon strips 7 layers of MAPS 4 layers of Hybrid + 3 layers of strips Pion to kaon separation (black circles) and proton to kaon separation (red triangles) in unit of sigma in the case of 4 layers of 300 μm (left panel), 7 layers of 15 μm (central panel) and 4 layers of 100 μm + 3 layers of 300 μm (right panel) silicon detectors. The horizontal lines correspond to a 3 sigma separation. 20 March 2012 L. Musa 46

Some examples of physics performance (B D 0) Further optimization In progress This will provide a direct measurement of RAA for charm (prompt D) and b (non-prompt D) Measuring D mesons from pt=2 Ge. V/c gives access to B meson from pt=0 Ge. V/c (due to the decay kinematics) 20 March 2012 L. Musa 47

Requirements on integration time and time resolution At high interaction rate (e. g. p-p at 2 MHz, Pb-Pb at 50 k. Hz) significant pile-up in the detector occurs depending on the integration time (or time resolution). This causes two separate problems for the reconstruction and analysis of collected data: 1. the reconstruction efficiency drops due to the ambiguity of track-cluster association at inner layers 2. distinguishing triggered event from surrounding pile-up interactions Reconstruction efficiency for triggered Pb-Pb central collision with increasing number of pile-up collisions 20 March 2012 L. Musa Same as the plot on the left but with three fast layers: e. g. 1, 2 and 6 48

Requirements on integration time and time resolution, cont’d Tracks can be correctly assigned to the vertex of triggered interaction if • the vertices of pile-up events are at a distance larger than 1 mm from the triggered vertex • comparing its “time stamp” to trigger Further constraints are imposed by the extraction of signal from different off-vertex decays. • for short lived particles, e. g. Λc or D 0, same ∼ 1 mm isolation should be enough • B-mesons will require ∼ 1 cm isolation of the triggered vertex Problem can be solved if each track has at least one point with precise time stamp associated to trigger. For 2 MHz p-p interaction rate one (100% efficient) layer with ~1μs time resolution allows to eliminate the effect of pile-up. 20 March 2012 L. Musa 49

Improvement of impact parameter resolution HYBRID PIXELS (state-of-the-art) and comparison with MAPS radial positions (cm): 2. 2, 2. 8, 3. 6, 20, 22, 41, 43 Simulations for two upgrade layouts Same for both layouts Layout 2: Pixel/Strips (3 layers of pixels + 4 layers of strips) • Resolutions: srf = 12 m, sz = 12 m for pixels srf = 20 m, sz = 830 m for strips • Material budget: X/X 0 = 0. 5% for pixels X/X 0 = 0. 83% for strips 20 March 2012 L. Musa 50

New Beam-pipe Two scenarios are being studies RI/RO [mm/mm] Tolerance [mm] N 1 [σ] 1 19/19. 8 6 16 2 16. 2/17 5 11. 8 Length: 5500 mm scenario 1 (adopted for the CDR): • first pixel layer at 22 mm (radial distance to outer beampipe ~2 mm) scenario 2: • same n 1 as for ATLAS and CMS upgrade • More aggressive value for mechanical tolerances (5 mm) 20 March 2012 L. Musa 51

New Beam-pipe Specifications for scenario 2 Quantity Value Comment Fabrication tolerance 0. 5 mm CMS quotes 0. 4 mm Sag 0. 5 mm Unsupported length 4335 mm + counterweight Mechanical adjustment precision 1. 0 mm Very demanding! Survey to beamline uncertainty 1. 5 mm Provided by survey Quad fiducial to beamline uncertainty 0. 5 mm L 3 movement <0. 5 mm B field movement <0. 5 mm Linear sum 5. 0 mm n 1 11. 8 2. 36σ/mm RI/RO 16. 2/17 mm/mm Thickness 800 um Measured values Beampipe length: 5500 mm 20 March 2012 L. Musa 52

Mechanical Structure – Clam shell concept Layout of 3 inner pixel detector layers Rbp=19. 8 R 1=22, 20 March 2012 Layout of the innermost gas cooled pixel ladders around the beam-pipe R 2=27. 5, R 3=37 L. Musa 53

Mechanical Structure – Clam shell concept Layer 3 half support with • Ruby ball pins or ladder mounting • Gas input manifold Gas supply pipes will be connected to the individual ladders 20 March 2012 L. Musa 54

Mechanical Structure – Clam shell concept Layer 3 half “clam-shell” ( X/Xo= 0. 2%) Outer Cylinder CF+Airex mechanically stable supporting sandwich shell Conical end-cap with the gas manifolds and blank end-cap One of the pixel ladder mounted Layer 3 fully equipped with ladders 55 20 March 2012 L. Musa 55

Mechanical Structure – Clam shell concept Full assembly of 3 detector layers Grand total: 0. 94 X/Xo One half of the conical end-cap support Total material budget - 100 m carbon fiber wall thickness shell (0. 16% of X/Xo) - 50 m thick MAPS (0. 053% X/Xo), facing the cooling panel with the holes - Kapton+Al multilayer cable (0. 056+0. 035% X/Xo) - Total radiation length per layer ~0. 3% X/Xo 20 March 2012 L. Musa 56

Stave layout and material budget Component Support Structure Glue Pixel module Flex bus Total Stave material budget Material budget X/X 0 (%) Notes 0. 07 – 0. 22 3 different options are being considered: carbon foam, polyimide and silicon 0. 045 0. 053 – 0. 16 0. 15 2 layers of glue 100 µm thick each Monolythic (50 µm thick) – hybrid (150 µm thick) single layer flex bus 0. 32 – 0. 58 Support Structure Carbon foam structure Sketch of building blocks constituting a generic stave 20 March 2012 Material budget X/X 0 (%) 0. 22 Notes PEEK cooling tube Polyimide micro -channel structure 0. 085 – 0. 13 Different coolant: H 20 or C 6 F 14 Silicon microchannel structure 0. 07 – 0. 11 Different layout: sideline or distributed micro-channels L. Musa 57

Support structure (example of conceptual design) Inner barrel: 3 layers of pixels Outer barrel: 4 -layer structure • 4 pixel/strip layers mounted on 2 barrels • 3 beams made of carbon composite or beryllium are fixed between the two structures to provide rigidity and support/guide the inner part insertion Outer barrel inner barrel 20 March 2012 L. Musa 58

Ultra-light innermost layer • Very light structure with almost no material (only silicon) in the active area • Very light stave without glue layers, electrical bus, etc. – Such a design rely on the use of large silicon structures that integrate the electrical bus for the distribution of signals and power (stitching fabrication process) • No overlap to simplify the geometry X/X 0 ~ 0. 1% • Air cooling to avoid extra material Layer 0 mechanical structure 20 March 2012 L. Musa Layer 0 conceptual design 59