RD50 Status Report May 2017 Radiation hard semiconductor devices for very high luminosity colliders

Transcription

1 LHCC 10.May 2017, CERN OUTLINE: RD50 Status Report May 2017 Radiation hard semiconductor devices for very high luminosity colliders Gianluigi Casse Michael Moll University of Liverpool, UK CERN, Geneva, Switzerland & FBK-CMM, Trento, Italy. RD50 Collaboration Scientific results (some highlights) Defect and Material Characterization Detector Characterization New Detector Structures Full Detector Systems RD50 Achievements & Work Program 2017/2018 Spare slides (RD50 common projects, additional material)

2 [I. Dawson, P. S. Miyagawa, Sheffield University, Atlas Upgrade radiation background simulations] [W.Riegler, RD50 Workshop, 12/2015] Motivation and Challenge LHC upgrade LHC upgrade towards High Luminosity LHC (HL-LHC) after LS3 (~ ); expect 3000 fb -1 (x6 nominal LHC) FCC Future Circular Collider.later than 2035 Radiation levels innermost pixel layer (30ab -1, without safety factor): 7x10 17 n eq /cm 2, 200MGy Semiconductor detectors will be exposed to hadron fluences equivalent to more than n eq /cm 2 (HL-LHC) and more than 7x10 17 n eq /cm 2 (FCC) detectors used now at LHC cannot operate after such irradiation RD50 : mandate to develop and characterize semiconductor sensors for HL-LHC.. and FCC? G.Casse and M.Moll, RD50 Status Report, May

3 The RD50 Collaboration RD50: 55 institutes and 327 members 46 European institutes Belarus (Minsk), Belgium (Louvain), Czech Republic (Prague (3x)), Finland (Helsinki, Lappeenranta ), France (Paris, Orsay), Germany (Dortmund, Erfurt, Freiburg, Hamburg (2x), Karlsruhe, Munich(2x)), Italy (Bari, Perugia, Pisa, Trento, Torino), Kroatia (Zagreb) Lithuania (Vilnius), Netherlands (NIKHEF), Poland (Krakow, Warsaw(2x)), Romania (Bucharest (2x)), Russia (Moscow, St.Petersburg), Slovenia (Ljubljana), Spain (Barcelona(3x), Santander, Valencia), Switzerland (CERN, PSI), United Kingdom (Birmingham, Glasgow, Lancaster, Liverpool, Oxford, RAL) 7 North-American institutes Canada (Montreal), USA (BNL, Brown Uni, Fermilab, New Mexico, Santa Cruz, Syracuse) 1 Middle East institute Israel (Tel Aviv) 1 Asian institute India (Delhi) Detailed member list: G.Casse and M.Moll, RD50 Status Report, May

4 RD50 Organizational Structure Co-Spokespersons Gianluigi Casse and Michael Moll (Liverpool University, UK & FBK-CMM, Trento, Italy) (CERN EP-DT) Defect / Material Characterization Ioana Pintilie (NIMP Bucharest) Detector Characterization Eckhart Fretwurst (Hamburg University) New Structures Giulio Pellegrini (CNM Barcelona) Full Detector Systems Gregor Kramberger (Ljubljana University) Characterization of microscopic properties of standard-, defect engineered and new materials pre- and postirradiation DLTS, TSC,. SIMS, SR, NIEL (calculations) Cluster and Point defects Boron related defects Characterization of test structures (IV, CV, CCE, TCT,.) Development and testing of defect engineered silicon devices EPI, MCZ and other materials NIEL (experimental) Device modeling Operational conditions Common irradiations Wafer procurement (M.Moll) Acceptor removal (Kramberger) 3D detectors Thin detectors Cost effective solutions Other new structures Detectors with internal gain LGAD: Low Gain Avalanche Det. Deep depleted Avalanche Det. Slim Edges HVCMOS LGAD (S.Hidalgo) HVCMOS (E. Vilella) Slim Edges (V.Fadeyev) LHC-like tests Links to HEP(LHC upgrade, FCC) Links electronics R&D Low rho strips Sensor readout (Alibava) Comparison: - pad-mini-full detectors - different producers Radiation Damage in HEP detectors Timing detectors Test beams (M.Bomben & G.Casse) Collaboration Board Chair & Deputy: G.Kramberger (Ljubljana) & J.Vaitkus (Vilnius), Conference committee: U.Parzefall (Freiburg) CERN contact: M.Moll (EP-DT), Secretary: V.Wedlake (EP-DT), Budget holder & GLIMOS: M.Moll & M.Glaser (EP-DT) G.Casse and M.Moll, RD50 Status Report, May

5 Defect & Material Characterization Some highlights G.Casse and M.Moll, RD50 Status Report, May

6 [Elena Donegani et al, UniHH, Trento Meeting, 2017] Defect Characterization Aim of defect studies: Identify defects responsible for Trapping, Leakage Current, Change of N eff, Change of E-Field Understand if this knowledge can be used to mitigate radiation damage (e.g. defect engineering) space charge trapping leakage current Deliver input for device simulations to predict detector performance under various conditions Method: Defect Analysis performed with various tools inside RD50: C-DLTS (Capacitance Deep Level Transient Spectroscopy) TSC (Thermally Stimulated Currents) PITS (Photo Induced Transient Spectroscopy) FTIR (Fourier Transform Infrared Spectroscopy) EPR (Electron Paramagnetic Resonance) TCT (Transient Current Technique) CV/IV (Capacitance/Current-Voltage Measurement) MW-PC (Microwave Probed Photo Conductivity) PC, RL, I-DLTS, TEM, and simulation RD50: several hundred samples irradiated with protons, neutrons, electrons and 60 Co-g B i O i Example: TSC measurement on defects produced by 23 MeV, 188 MeV and 23 GeV protons G.Casse and M.Moll, RD50 Status Report, May

7 Some identified defects Summary on defects with strong impact on device performance after irradiation Phosphorus: shallow dopant (positive charge) positive charge (higher introduction after proton than after neutron irradiation, oxygen dependent) positive charge (higher introduction after proton irradiation than after neutron irradiation) Leakage current: V 3 leakage current & neg. charge current after g irrad, V 2 O (?) Reverse annealing (negative charge) Boron: shallow dopant (negative charge) A table with levels and cross sections is given in the spare slides. Trapping: Indications that E205a and H152K are important (further work needed) Converging on consistent set of defects observed after p, p, n, g and e irradiation. Defect introduction rates are depending on particle type and particle energy and (for some) on material! G.Casse and M.Moll, RD50 Status Report, May

8 Nitrogen enriched Silicon New defect and material engineering approach: Nitrogen enriched silicon Produced in collaboration with wafer foundry TOPSIL, Denmark/Poland 4 materials used: Floating Zone x 10 (Standard, Nitrogenated, Oxygenated) diode/w cm 2 x 10-7 FZ Magnetic Czochralski NIT Processing at CNM Barcelona Spain Sensors ready in 4/2017 (in excellent quality) being distributed for radiation testing now J reverse [A/cm 2 ] DOFZ MCZ Diode current [A] J reverse [A/cm 2 ] RD50 NitroStrip p-on-n: Run 9802 first measurements 3/ diode/wafer - area = cm 2 10nA diode/w cm 2 FZ NIT DOFZ MCZ Wafer 1 FZ Wafer 2 FZ Wafer 3 FZ Wafer 4 FZ Wafer 5 FZ Wafer 6 FZ Wafer 7 NIT Wafer 8 NIT Wafer 9 NIT Wafer 10 NIT Wafer 11 NIT Wafer 12 NIT Wafer 13 DOFZ Wafer 14 DOFZ Wafer 15 DOFZ Wafer 16 DOFZ Wafer 17 DOFZ Wafer 18 DOFZ Wafer 19 MCZ Wafer 20 MCZ Wafer 21 MCZ Wafer 22 MCZ Wafer 23 MCZ Wafer 24 MCZ J reverse [A/cm 2 ] diode/w cm 2 FZ NIT DOFZ MCZ 1000V 10nA Reverse voltage [V] wafer wafer 1000V wafer G.Casse and M.Moll, RD50 Status Report, May

9 Device Characterization and Device Simulations - selected topics - G.Casse and M.Moll, RD50 Status Report, May

10 TCT Transient Current Technique TCT: Pulsed laser induced generation of charge carriers in the detector Study of: electric field in sensor, charge collection efficiency, homogeneity,.. Benchmarking of simulation tools, measure physics parameters from mobility to impact ionization New TCT technology: TPA-TCT Two Photon Absorption TCT Concept: TPA TCT TCT (red) short penetration length (650nm = 1.9eV) carriers deposited in a few mm from surface front and back TCT study electron and hole drift separately 2D spatial resolution (5-10mm) TCT (infrared) long penetration (1064nm = 1.17 ev) similar to MIPs (though different de/dx) top and edge-tct 2D spatial resolution (5-10mm) TPA-TCT (far infrared) No single photon absorption in silicon 2 photons produce one electron-hole pair Point-like energy deposition in focal point 3D spatial resolution (1 x 1 x 10 mm 3 ) Example: HV-CMOS 100x100 mm 2, 10 mm depleted [M.Fernandez, TCT Workshop, Ljubljana, 10/2016] G.Casse and M.Moll, RD50 Status Report, May

11 Charge in 25ns [a.u.] Charge in 25ns [a.u.] TPA TCT on Irradiated Sensors Problem: Radiation creates defects absorbing far-infrared light! Worry: TPA method compromised by radiation damage? Solution: Measure both SPA and TPA and correct data d dt N( r, z) 2 I( r, z) I ( r, z) 2 SPA TPA+SPA n eq /cm 2 TPA [ ] TPA [ ] SPA [ ] Si SPA [ ] n eq /cm 2 Example: Diode irradiated to n eq /cm 2 Laser power[a.u.] Laser power[a.u.] SPA correction n eq /cm 2 Conclusion: TPA-TCT can be applied to highly irradiated silicon sensors! [M.Fernandez, Seminar, Geneva Uni, Nov.2016] G.Casse and M.Moll, RD50 Status Report, May

12 RD50 Simulation Working Group Aim of TCAD Device Simulations: Understand and predict sensor operation & avoid design mistakes Simulate performance of irradiated silicon sensors and performance predictions under various conditions Challenge for irradiated sensors: Correct implementation of bulk and surface damage by defect levels Defect concentration depends on fluence, particle type, material, annealing,., i.e. very complex! Close collaboration with CMS, AIDA and ATLAS sensor simulation working groups Simulation Roadmap: Measure macroscopic damage parameters using test-structures under well controlled conditions Measure defect parameters as guidance for input parameters (often obtained at lower particle fluences). Simulate devices and optimize simulation input parameters to match macroscopic results. Use simulations to optimize new specific sensor designs. Produce new sensors, test them and iterate on the simulation. Status: Good results; Input parameters (defect levels) need further improvement! Example: V dep after proton irradiation G.Casse and M.Moll, RD50 Status Report, May

13 What is different from TCAD tools and why needed? Solve Poisson Equation for an input N eff distribution (rather than calculating N eff and hence the electric field distribution from defect levels like in TCAD) Charge drift considered in static electric field induced signal from Ramo Theorem / Weighting field Higher flexibility (source code in hands) and much faster than TCAD tools Allow for fast simulation of multi-electrode systems; fitting to experimental data Plug and play (easy to install and to use - GUI), free of charge Summary of UFSD test results and comparison with simulation Development within RD50 UFSD test beam data vs. Simulation 200 TRACS (CERN, Santander) KDetSIM (JSI) Weightfield 2 (Torino, UCSC-SCIPP, ) others Successful in optimizing sensor concepts and understanding signal formation Example: Optimization of fast timing sensors Signal Simulation Tools Resolution [ps] Thickness [µm] Data - CFD [CNM G ~ 10, 5x5 mm] Data - CFD [CNM G ~ 10, 3x3 mm] Data - CFD [CNM G ~ 20, 3x3 mm] Data - CFD [FBK G ~ 15, 1x1 mm] Data - CFD [CNM G ~ 5, 1x1 mm] Data - CFD [ CNM G ~ 20, 1.2x1.2 mm] WF2 [G = 10, 5x5mm] WF2 [G =10, 2 pf] [LGAD sensors 50,75, 300 mm] [H.Sadrozinski et al., 2017, ] G.Casse and M.Moll, RD50 Status Report, May

14 [G. Kramberger et al JINST 8 P08004] [ TM.Mikuz, Torino, RD50 Workshop 6/2016] Motivation: Investigation on extreme fluences Future detectors (e.g. LHC forward calorimeters, FCC detectors) will face levels of 1x10 17 n eq cm -2 and beyond! Linear extrapolations of the parameterizations (e.g. depletion, leakage, charge) from lower fluences fail! Need to understand limits of our physics parameters, parameterizations and modelling! Need new models! Evidence of a different regime of sensor operation CCE (Sr 90 ) on pad sensors gives surprisingly high charge (1000 electrons after cm -2 ), low current How to characterize sensors/materials at cm -2 (e.g. defects?) Working on new parameterization/model for Electric field, leakage, trapping, charge collection Strong indication that the mobility decreases with fluence (e.g. factor 6 at cm -2 )! G.Casse and M.Moll, RD50 Status Report, May

15 New structures..optimizing for Radiation hardness Time resolution Cost effectiveness LGAD, APD (Sensors with intrinsic gain) HVCMOS (towards monolithic sensors) 3D (sensors with vertical electrodes) G.Casse and M.Moll, RD50 Status Report, May

16 Sensors with intrinsic gain Exploit impact ionization (charge multiplication in high field regions) to achieve faster ( improve timing performance) and radiation harderer ( mitigate trapping) sensors. Main focus: LGAD (Low Gain Avalanche Detectors) gain O(10) Use of LGAD proposed & developed at CNM within RD50; 3 suppliers (CNM,FBK,HPK) today Some work on DD-APD s gain O(500) at 1800V, 1 supplier (RMD USA) LGAD structure: Core Region: Uniform electric field, high enough to activate impact ionization Termination: Confining the high electric field of the core region Optimization: TCAD simulations and studies on prototype devices Considerations for optimization: Gain versus V breakdown trade-off, timing performance Thin detector integration, radiation hardness Proportional Response (linear mode operation) Better S/N ratio (small cell volumes and fast shaping times) [G. Pellegrini et. al., NIM A 765, 2014, 12 16] G.Casse and M.Moll, RD50 Status Report, May

17 LGAD within RD50 Collaboration 2010: LGAD activities started with CNM project 2012: First 4 inch wafer, 300 µm thick, LGAD 2014: First 200 µm thick LGAD. First Gallium Process 2015: First PiN on 6 inch wafer. First Inverse LGAD 2016: First 50 µm thick LGAD for Timing applications (ATLAS HGTD and CMS CT-PPS. SOI and SOS, 4 inch wafer). Gallium and Carbon Processes 2017: First LGAD on 6 inch wafer 25 Fabrication processes. 200 Wafers processed at CNM LGAD ilgad HGTD/CTPPS CTPPS FBK, Trento: 1 st batch of LGAD in 2016, 2 nd in 2017 (processed, being characterized) p-side and n-side segmentation, pixel, strip, AC coupling, (see Annex) Today: CNM Spain, FBK Italy and HPK Japan produced LGAD s (3 suppliers) [G. Pellegrini et al., Low Gain Avalanche Detectors (LGAD), Vertex 2016 G.Casse and M.Moll, RD50 Status Report, May

18 LGAD for Timing Applications LHC experiments starting to implement/plan on Si based timing detectors with RD50 technology 6 x 12 mm 2 CTPPS CMS-TOTEM Precision Proton Spectrometer Installed in 2016 YETS in two Roman pots two planes of segmented LGAD Time resolution measured by CTPPS (at high voltage!) 45 mm LGAD on SOI 1.7 mm 2 27ps single device; 16 ps with 3 layers (@ 230 V ) ATLAS AFP - Forward Proton 30 ps up to 3x10 14 n eq /cm 2 57 ps at n eq /cm 2 (@ 620 V) ATLAS HGTD - High Granularity Timing Detector CMS Timing layer Planning for hermetic timing detector for phase II; CB approved a full coverage eta 0 3 timing layer for MIPs, using crystals in the barrel and LGAD as baseline in the endcap; Expect:10 14 n eq /cm 2 to n eq /cm 2; 9m 2 silicon sensor surface; (tentative: 40-50mm thick, 1x3mm 2 pads, >95% fill factor) CTPPS: 45mm thick LGAD (SOI) HGTD LGAD on 50 mm SOI CMS Timing layer G.Casse and M.Moll, RD50 Status Report, May

19 Time resolution & radiation damage N. Cartiglia et al: LGAD AFP test beam, ATLAS/RD50 Device: SOI, 45 mm thick, 1.3 x 1.3 mm 2 area ~ 30ps time resolution up to about 3x10 14 n eq /cm 2 57 ps for sensor irradiated to n eq /cm 2 (620V) J.Lange et al LGAD test beam, CTPPS/RD50 SOI, 45 mm thick, 1.7 mm 2 16ps with 3 layers (230V) Radiation hardness strategy for LGAD LGAD loosing gain with increasing radiation (Boron removal and/or compensation effect) Going towards thin layers (50 mm) Gain-layer engineering [devices produced, testing starting]: (a) Gallium implanted gain layer (b) Carbon co-implantation to Boron G.Casse and M.Moll, RD50 Status Report, May

20 CMOS & monolithic devices RD50 started to work on HVCMOS device characterization in 2014 close collaboration with ATLAS HVCMOS group and other collaborations RD50 focus on characterizing radiation damage Typical (HV-)CMOS device Depleted active pixel detectors in CMOS process Sensor element is deep n-well in (usually) low resistivity (~10 Ωcm) p-type substrate 60 V ~ 10 μm depleted ~1000 electrons Lower resistivity offers bigger depleted volume Characterization with edge-tct within RD50 Charge collection profiles LFoundry (150nm, 2 kw-cm) I.Mandic, RD50 Workshop, Nov G.Casse and M.Moll, RD50 Status Report, May

21 c [cm 2 ] [Igor Mandic et al, Ljubljana, RD50 Workshop, November 2016] Acceptor removal - Boron related defects? Acceptor removal Effective doping N eff is composed of Boron (negative charge) and radiation induced charged defects! N eff = N Boron exp( c Φ) +. [simplified] Acceptor removal rate p-type sensors of different resistivity show different rate in acceptor removal!? G. Kramberger et al, 10 th Trento Meeting, Feb , 2015 Irradiation changes: effective doping depletion depth signal Example: Change of depletion depth in CMOS X-FAB AMS CHESS1 AMS CHESS1 AMS CHESS2 Blue marker charged hadron irradiated Red marker neutron irradiated So-called acceptor removal responsible for: Gain degradation in sensors with intrinsic gain Good performance of low resistivity CMOS sensors after high irradiation. Why not studied more intensively before? Focus was on high resistivity and on n-type not on low resistivity! What is the origin? Boron removal kinetics (e.g. B i B i O i ) and/or compensation effects!? Need more work on this! G.Casse and M.Moll, RD50 Status Report, May

22 RD50 HV-CMOS submission RD50 design for R&D on HV-CMOS sensors funding assured, submission target: end 2017; 150 nm HV-CMOS from LFoundry Target: - Improve the timing resolution of HV-CMOS sensors with different solutions implemented at the readout circuit level - Study new sensor cross-sections Novel ideas for increasing the device area (above the limitation of the reticule size. - Study pre-stitching options more details in spare slides G.Casse and M.Moll, RD50 Status Report, May

23 RD50 main achievements & links to LHC Experiments Some important contributions of RD50 towards the LHC upgrade detectors: p-type silicon (brought forward by RD50 community) and now used for the ATLAS and CMS Strip Tracker upgrades n- MCZ and oxygenated Silicon (introduced by RD50 community) might improve performance in mixed fields due to compensation of neutron and proton damage: MCZ is under investigation in ATLAS, CMS and LHCb Double column 3D detectors developed within RD50 with CNM and FBK. Development was picked up by ATLAS and further developed for ATLAS IBL needs, followed by AFP and TOTEM and now also within CMS/ATLAS upgrades. RD50 results on highly irradiated planar segmented sensors demonstrated: devices are a feasible option for LHC upgrade RD50 data and damage models are essential input parameters for operation scenarios of LHC experiments and their upgrades (evolution of leakage current, CCE, power consumption, noise,.) and sensor design (TCAD parameters). Charge multiplication effect observed for heavily irradiated sensors (diodes, 3D, pixels and strips). Dedicated R&D launched in RD50 to understand underlying multiplication mechanisms, simulate them and optimize the CCE performances. Evaluating possibility to produce fast segmented sensors? Fast timing sensors: LGAD (Low Gain Avalanche Detectors) were developed within the RD50 community New characterization techniques and simulation tools for the community: Edge-TCT, Alivaba readout, TPA-TCT, partly available through spin-off companies. Signal simulation and TCAD tools. Close links to the LHC Experiments: Only few RD50 groups are not involved in ATLAS, CMS and LHCb upgrade activities (natural close contact). Common projects with Experiments: Detector developments; Irradiation campaigns, test beams, wafer procurement and common sensor projects. Close collaboration with LHC Experiments on radiation damage issues of present detectors. G.Casse and M.Moll, RD50 Status Report, May

24 Workplan for 2017/2018 (1/2) Defect and Material Characterization (Convener: Ioana Pintilie, Bucharest) Consolidate list of defects and their impact on sensor properties (Input to simulation group) including introduction rates & annealing for different type of irradiations and materials Extend work on p-type silicon including low resistivity material Understand boron removal in lower resistivity p-type silicon: Performance of MAPS, CMOS sensors, LGAD adding new macroscopic measurements Working group on acceptor removal formed! Characterization of Nitrogen enriched silicon (starting project, wafers ready) Detector Characterization (Convener: E.Fretwurst, University of Hamburg, Germany) TCAD sensor simulations Cross-calibration of different simulation tools (ongoing) and comparison of TCAD models Refine defect parameters used for modeling (from effective to measured defects) Extend modeling on charge multiplication processes Surface damage working group Extend use of signal simulation tools towards fitting of measured data Extend experimental capacities on e-tct equipment Parameterization of electric field (fluence, annealing time, etc.) Exploit full potential of Two Photon Absorption for sensor characterization; build setup at CERN Continue parameterization of radiation damage (performance degradation) of LHC like sensors! Explore fluence range to10 17 cm -2 and beyond (to prepare for future needs in forward physics and FCC) G.Casse and M.Moll, RD50 Status Report, May

25 Workplan for 2017/2018 (2/2) New structures (Convener: Giulio Pellegrini, CNM Barcelona, Spain) Continue work on thin and 3D sensors (especially in combination with high fluence) Continue characterization of dedicated avalanche test structures (LGAD, DD-APD) Understand impact of implant shape and other geometrical parameters on avalanche processes Study of Gallium based amplification layers and impact of Carbon co-implantation LGAD, DD-APD: intensify evaluation of timing performance and radiation degradation (Where are the limits? How to overcome radiation damage? How much gain is optimum?) HVCMOS Continue characterization of existing devices (close collaboration with ATLAS HVCMOS group) End of year: submission of first RD50 devices in an engineering run Full detector systems (Convener: G.Kramberger, Ljubljana University, Slovenia) Further studies of thin (low mass) segmented silicon devices Study performance of thin and avalanche sensors in the time domain (Fast sensors!) Long term annealing of segmented sensors (parameterize temperature scaling) Continue study on mixed irradiations (segmented detectors) Continue RD50 program on slim edges, edge passivation and active edges Merging of RD39 into RD50 Cryogenic operation at high fluences? Links with LHC experiments and their upgrade working groups Continue collaboration on evaluation of radiation damage in LHC detectors Continue common projects with LHC experiments on detector developments G.Casse and M.Moll, RD50 Status Report, May

26 RD50 common projects RD50 common projects Part of the RD50 R&D is performed within RD50 common projects Projects involving several RD50 institutes, one lead institution, evaluated by RD50, co-funded from RD50 fund, well defined time-line and objectives targeting a specific R&D question Most recent projects UBM for Avalanche Sensors (Giulio Pellegrini, CNM, Spain) Avalanche PAD sensors (Giulio Pellegrini, CNM, Spain) Sensors with ADVACAM (Anna Macchiolo, MPI, Munich) RD50 common test beam (Marco Bomben, LPHNE, Paris ) Investigation of the properties of thin LGAD (Nicolo Cartiglia, Torino, Italy) D sensors for HL-LHC (Marcos Fernandez Garcia, Santander, Spain) Evaluation of Two Photon Absorption - TPA (Ivan Vila, Santander, Spain) Doping profiling of LGAD and other devices (Hartmut Sadrozinski, SCIPP, USA) NitroStrip project: Nitrogen doped silicon (Alexander Dierlamm, Karlsruhe, Germany) Gallium doping of LGAD Sensors (David Flores, CNM, Barcelona) Acceptor removal in p-type Silicon (Gregor Kramberger, Ljubljana) TPA edge-tct studies in Bilbao Laser Center (I.Vila, Santander) LGAD based on EPI wafers (G.Pellegini, CNM, Barcelona) TPA TCT on CMOS sensors (I.Vila, Santander) LGAD fabricated with epitaxial layer (G.Pellegrini, CNM, Barcelona) RD50 CMOS submission (Gianluigi Casse, Liverpool, UK / Vitaliy Fadeyev, SCIPP, USA) G.Casse and M.Moll, RD50 Status Report, May

27 Spare Slides Some spare slides More details on G.Casse and M.Moll, RD50 Status Report, May

28 Segmented Sensors with read-out at the n + contact (n-in-p or n-in-n) G.Casse and M.Moll, RD50 Status Report, May

29 Thin p-type pixel sensors Thin FZ p-type pixel sensors: 75 to 300 mm with 450 mm edge (MPI/CIS/VTT) ATLAS FEI4, 25 MeV protons, 800 MeV protons, neutrons, data obtained with beta source Detectors irradiated with n eq /cm 2 100mm thick sensors give more charge than 75mm thick sensors, both saturate with voltage 200 mm thick sensors give more charge than 300 mm thick sensors at moderate voltage Beam tests show 97-99% hit efficiency (thickness tested μm, 500V) S. Terzo, 24 th RD50 Workshop June 2014 Detectors irradiated up to n eq /cm 2 Sensor modules still functional (even if in homogeneously irradiated) B. Paschen, 25 th RD50 Workshop Nov G.Casse and M.Moll, RD50 Status Report, May

30 Summary on defects with strong impact on device performance after irradiation Most important defects [for details and references see JAP 117, , 2015] G.Casse and M.Moll, RD50 Status Report, May

31 p-type strip sensors with n + readout (brought forward by RD50) are now the sensor choice for ATLAS and CMS Tracker upgrades Collected Charge [10 3 electrons] Reminder: Segmented sensors: n + vs. p + readout p-in-n-fz (500V) n-in-p-fz (800V) eq [cm -2 ] n-in-p-fz (1700V) n-in-p-fz (500V) M.Moll - 09/2009 FZ Silicon Strip Sensors n-in-p (FZ), 300mm, 500V, 23GeV p [1] n-in-p (FZ), 300mm, 500V, neutrons [1,2] n-in-p (FZ), 300mm, 500V, 26MeV p [1] n-in-p (FZ), 300mm, 800V, 23GeV p [1] n-in-p (FZ), 300mm, 800V, neutrons [1,2] n-in-p (FZ), 300mm, 800V, 26MeV p [1] n-in-p (FZ), 300mm, 1700V, neutrons [2] p-in-n (FZ), 300mm, 500V, 23GeV p [1] p-in-n (FZ), 300mm, 500V, neutrons [1] References: [1] G.Casse, VERTEX 2008 (p/n-fz, 300mm, (-30 o C, 25ns) [2] I.Mandic et al., NIMA 603 (2009) 263 (p-fz, 300mm, -20 o C to -40 o C, 25ns) n + -electrode readout ( natural in p-type silicon ): favorable combination of weighting and electric field in heavily irradiated detector [3] [1] [2] [3] n/n-fz, 3D, Diamond p/n-fz, double 285mm, 300mm, [RD42 sided, (-10 (-30 Collaboration] 250mm o C, C, 40ns), 25ns), columns, pixel strip 300mm [Casse [Rohe substrate et 2008] al. 2005] [Pennicard 2007] electron collection, multiplication at segmented electrode Situation after high level of irradiation: p + readout p + readout small holes E E w n + readout n + p + large E S. Wonsak, 25th RD50 Workshop Nov E w electrons [G. Kramberger, Vertex 2012] n + readout G.Casse and M.Moll, RD50 Status Report, May

32 TCAD - Simulations Device simulation of irradiated sensors Using: Custom made simulation software and Silvaco & Synopsis TCAD tools RD50 simulation working group Good progress in reproducing experimental results on leakage current, space charge, E-Field, trapping.. Enormous parameter space ranging from semiconductor physics parameters and models over device parameters towards defect parameters Tools ready but need for proper input parameters! Working with effective levels for simulation of irradiated devices Most often 2, 3 or 4 effective levels used to simulate detector behavior Introduction rates and cross sections of defects tuned to match experimental data Measured defects TCAD input G.Casse and M.Moll, RD50 Status Report, May

33 Models of radiation damage in TCAD [G.Kramberger, VERTEX 2016] G.Casse and M.Moll, RD50 Status Report, May

34 UFSD Roadmap at FBK In close collaboration with TIFPA, UNI TN and INFN Turin Early st Batch on 300um Si - technology assessment - microfabrication process tuning - Validation of the simulations Mid 2016 Early nd Batch on 50um Si 1st batch characterization Mid nd batch characterization - Improve timing performance - Improve the overall detectors performance - Test new methods to increase the radiation hardness G.Casse and M.Moll, RD50 Status Report, May

35 Batch 1: segmentation Three different approaches for device segmentation have been investigated multiplication junction 1. N-side segmentation: both n+ and the gain layers are segmented (some concerns about E field uniformity) P++ ohmic contact multiplication junction 2. P-side segmentation: the p layer opposite to the gain layer is segmented P++ ohmic contact multiplication junction 3. AC coupling: The signal is frozen on the resistive sheet, and it s AC coupled to the electronics P++ uniform layer G.Casse and M.Moll, RD50 Status Report, May

36 570 µm 50 µm UFSD Batch 2 A new production batch on 50 µm thick substrates Modify the process to increase the UFSD radiation: Gallium for gain layer formation Carbon doping Moptimization of Boron profile High-resistivity silicon wafer Low-resistivity handle wafer G.Casse and M.Moll, RD50 Status Report, May

37 Front-Side Back-Side The inverse LGAD ilgad Decouple multiplication layer from segmented electrodes Strip, pixel and pad structures with this configuration produced in 2016 First wafers show good electrical performance (CV,IV), intensive testing ahead [G.Pellegrini, CNM, May 2016, private communication] G.Casse and M.Moll, RD50 Status Report, May

38 [G.Kramberger, RD50 Workshop, June 2014 ] LGAD Radiation Damage Radiation Hardness is limited: Acceptor removal problem Collected Charge is degrading with fluence (i.e. we are loosing the gain ) TCAD simulations and CV, e-tct, TCT measurements point out two mechanisms: Boron removal [Boron bound into defects, i.e. de-activation of shallow dopant] Space charge build-up due to increased hole current from amplification Experimental approaches in new production runs: Use Gallium instead of Boron to impact on defect kinetics [wafers ready] Use Carbon co-implant to protect Boron from removal [to be done] Gallium doped wafer produced,..to be tested G.Casse and M.Moll, RD50 Status Report, May

39 [G.Pellegrini, RD50 Workshop, 12/2015] [I.Lopez et al., IFAE,UAB, RD50 Workshop, 12/2015] 3D detectors Development of 3D silicon sensors for innermost pixel layers at HL-LHC Low depletion voltage (low power dissipation) Low drift length (reduced trapping) Testbeam data Testbeam [ATLAS-ITK; FE-I4 3D modules] 97.5% hit efficiency at 9x10 15 n eq /cm 2 3D sensor concept Implementation as double and single sided (see below) process Several recent production runs [RD50 projects] joint 3D MPW pixel run with ATLAS, CMS, LHCb Various layouts, small pitch [mm 2 ] 50x50, 25x100, 50,125,.. technology studies on SOI wafers small holes: down to 8um diameter at 230um depth or 5 um in single side 3D on SOI (50, 100 or 150 um layer) temporary metal layers for pixel testing before UBM implemented G.Casse and M.Moll, RD50 Status Report, May

40 3D Detectors installed in experiments at CERN 3D detectors based on double sided technology developed within the RD50 collaboration. Atlas IBL (Insertable b-layer): Installed in 2014 (produced at CNM and FBK) AFP (ATLAS Forward Proton): 1 st Arm assembled and installed in Roman Pots in February 2016, 2 nd arm will be installed in (produced at CNM) CT-PPS (CMS-TOTEM Precision Proton Spectrometer): under installation. (produced at CNM) IBL Insertion CT-PPS Detectors G.Casse and M.Moll, RD50 AFP Status detectors Report, with May slim 2017edges -41-

41 New 3D detector productions for the Atlas and CMS upgrade 3D detectors are the baseline for the innermost layer of the Atlas and CMS detectors. Total area foreseen for the upgrade is approximatively 2 m 2. 3D detectors with 50x50um 2 geometry and 25x100 um 2 are a reliable option for the upgrade. Thin 3D detectors based on single side technology is ready and the thickness of the substrate can be adapted to the requirements of the experiments (most probable 150um thick). First results on small pitch pixels are very promising showing improving operation voltage at 5E15 n eq /cm 2 from 120V to 40V. At 9E15n eq /cm 2 the voltage needed for 97% efficiency is 100V compared with IBL-type 160V. Last productions show better yield and good performance at wafer level. Simulation is an important tool to study 3d detectors radiation damage. More details in: D. Vazquez- Beam test measurements of irradiated 3D pixel sensors G.Casse with 50x50 and M.Moll, μm² pixel RD50 size Status TREDI Report, meeting, May

42 [A.Affolder et al., 2016 JINST 11 P04007, Strip CMOS collaboration] [M.Fernandez-Garcia et al, 2016 JINST 11 P02016] HVCMOS CCE as function of fluence 350nm neutron irradiation AMS, 350nm, CHESS-1, 20 Wcm ( cm -3 ) 2mm x 2mm passive sensor (400 pixel) Sr 90, 25 ns shaping, 120 V 180nm neutron irradiation AMS, 180nm, HV2FEI4, 10 Wcm ( cm -3 ) 100mm x 100 mm passive pixel IR-laser, 5 ns integration CCE rising above initial value for fluences in order of some n eq /cm 2 Up to about n eq /cm 2 CCE decreases [diffusing charge gets trapped] Above about n eq /cm 2 CCE rises above initial value [reduction of N eff, acceptor removal ] Above some n eq /cm 2 CCE finally degrading [trapping, increase of space charge N eff ] G.Casse and M.Moll, RD50 Status Report, May

43 HV-CMOS submission Floorplan Target: - Improve the timing resolution of HV-CMOS sensors with different solutions implemented at the readout circuit level - Study new sensor cross-sections - Study pre-stitching options Technology: nm HV-CMOS from LFoundry Design effort: - IFAE (R. Casanova) - Uni. Barcelona (O. Alonso) - Uni. Liverpool (S. Powell, E. Vilella and C. Zhang) Test structure 1 Test structure 2 Test structure 3 Test structure 4 Test structure 5 Test structure 6 Simple CMOS capacitors to study oxide thickness 10 x 10 matrix of very small pixels with passive readout 10 x 10 matrix of very small pixels with 3T-like readout Small matrix of pixels for TCT, e-tct and TPA-TCT measurements Single pixels for sensor capacitance measurements 44/4 G.Casse and M.Moll, RD50 Status Report, May

44 HV-CMOS submission - Matrix 1 Matrix with analog sampling circuit: - Chain of delay elements to generate WR 0, WR 1,, WR n (signals to enable writing the sensor analog value, i.e. the voltage) when there is an event - Programmable delay - Analog memories based on metal-insulator-metal capacitances (< 5 per pixel) - Analog serializers to send the data off-chip - Off-chip ADCs - Off-chip processing to determine t 0 (time of event) 45/4 G.Casse and M.Moll, RD50 Status Report, May

45 HV-CMOS submission - Matrix 2 Matrix with in-pixel Time-to-Digital Converter (TDC): - Global time-stamp coarse time measurement - TDC fine time measurement 46/4 G.Casse and M.Moll, RD50 Status Report, May

46 HV-CMOS submission - Matrix 3 Matrix with very fast amplifier: - Option 1 Amplifier with continuous slow reset + switched fast reset - The output of the comparator is used to enable the fast reset when there is an event - The current consumption is high only when there is an event (average value is low) - Total recovery time < 50 ns (independent of input energy) - Option 2 Fast amplifier with continuous reset - Total recovery times is < 60 ns (dependent on input energy) 47/4 G.Casse and M.Moll, RD50 Status Report, May