1. Radiation tolerant silicon detectors G. Casse 1 G. Casse - Sensing in Extreme Environments - London, 7 May 2015 RD50

2. Outline 2 Silicon detectors are the devices of choice for vertex and tracker systems in high energy and (increasingly) nuclear physics experiments. This is due to their speed, low mass, relatively low cost and their surprising radiation tolerance. These characteristics have evolved over several order of magnitude since the introduction of segmented silicon sensors for tracking in the 1980’ in CERN experiments. Their evolution is not slowing down. A review of the state-of-the-art and of current trends is given from a Particle Physics perspective. Requirements for detectors for future hadron supercolliders: Radiation tolerance probably the more stringent one. Definition of the challenges: speed and radiation hardness Status Other future colliders Vertex and tracker sensors common requirements Summary 2 G. Casse - Sensing in Extreme Environments - London, 7 May 2015

3. 3 The rise of silicon detectors Surface # of channels If surface is saturating, certainly the granularity is not. 2022 (ish) Up to by a factor of 10-40

4. RD50 The RD50 Collaboration RD50: 50 institutes and ~ 300 members 38 European and Asian institutes Belarus (Minsk), Belgium (Louvain), Czech Republic (Prague (3x)), Finland (Helsinki, Lappeenranta ), France (Paris, Orsay), Germany (Dortmund, Erfurt, Freiburg, Hamburg (2x), Karlsruhe, Munich (2x)), Greece (Democritos), Italy (Bari, Florence, Padova, Perugia, Pisa, Torino), Lithuania (Vilnius), Netherlands (NIKHEF), Poland (Krakow, Warsaw(2x)), Romania (Bucharest (2x)), Russia (Moscow, St.Petersburg), Slovenia (Ljubljana), Spain (Barcelona(2x), Santander, Valencia), Switzerland (CERN, PSI), Ukraine (Kiev), United Kingdom (Glasgow, Liverpool) 7 North-American institutes Canada (Montreal), USA (BNL, Fermilab, New Mexico, Santa Cruz, Syracuse) 1 Middle East institute Israel (Tel Aviv) 1 Asian institute India (Delhi) Detailed member list: http://cern.ch/rd50 -4-

5. RD50 RD50 Organizational Structure Co-Spokespersons Gianluigi Casse and Michael Moll (Liverpool University) (CERN PH-DT) Defect / Material Characterization Mara Bruzzi (INFN & Uni Florence) Detector Characterization Eckhart Fretwurst (Hamburg University) Full Detector Systems Gregor Kramberger ( Ljubljana University) Characterization of microscopic properties of standard-, defect engineered and new materials pre- and post- irradiation DLTS, TSC, …. SIMS, SR, … NIEL (calculations) WODEAN: Workshop on Defect Analysis in Silicon Detectors (G.Lindstroem & M.Bruzzi) 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) Simulations (V.Eremin) 3D detectors Thin detectors Cost effective solutions Other new structures Detectors with internal gain (avalanche detectors) Slim Edges 3D (R.Bates) Semi 3D (Z.Li) Slim Edges (H.Sadrozinski) LHC-like tests Links to HEP Links electronics R&D Low rho strips Sensor readout (Alibava) Comparison: - pad-mini-full detectors - different producers Radiation Damage in HEP detectors Test beams (G.Casse) New Structures Giulio Pellegrini (CNM Barcelona) Collaboration Board Chair & Deputy: G.Kramberger (Ljubljana) & J.Vaitkus (Vilnius), Conference committee: U.Parzefall (Freiburg) CERN contact: M.Moll (PH-DT), Secretary: V.Wedlake (PH-DT), Budget holder & GLIMOS: M.Glaser (PH-DT) -5-

6. Possible 20 Year LHC Schedule 6 G. Casse - Sensing in Extreme Environments - London, 7 May 2015

7. Radiation Background Simulation For strips 3000fb -1 × 2 implies survival required up to ~ 1.3 × 10 15 n eq /cm 2 At inner pixel radii - target survival to 2-3 × 10 16 n eq /cm 2 I. Dawson et al. 7 G. Casse - Sensing in Extreme Environments - London, 7 May 2015

8. Radiation levels expected with sLHC Radiation hardness requirements (including safety factor of 2) 2 × 10 16 n eq /cm 2 for the innermost pixel layers 1 × 10 15 n eq /cm 2 for the innermost strip layers Dominated by pion damage Dominated by neutron damage [M.Moll] 8 G. Casse - Sensing in Extreme Environments - London, 7 May 2015

9. Status of radiation hardness studies Crucial point for Vertex and Tracker sensors Silicon still the forefront runner, no serious possibility to change track within the timescale (possible exception, diamond for innermost layer....). Dedicated R&D activity within the upgrading experiments and a dedicated community (CERN- RD50: http://rd50.web.cern.ch/rd50/). Most advanced results from this community. G. Casse - Sensing in Extreme Environments - London, 7 May 2015 9

10. Defect characterization – Aim: 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) Deliver input for device simulations to predict detector performance under various conditions – Method: Defect Analysis on identical samples performed with various tools inside RD50: C-DLTS (Capacitance Deep Level Transient Spectroscopy) I-DLTS (Current Deep Level Transient Spectroscopy) TSC (Thermally Stimulated Currents) PITS (Photo Induced Transient Spectroscopy) FTIR (Fourier Transform Infrared Spectroscopy) RL (Recombination Lifetime Measurements) PC (Photo Conductivity Measurements) EPR (Electron Paramagnetic Resonance) TCT (Transient Current Technique) CV/IV (Capacitance/Current-Voltage Measurement) – RD50: several hundred samples irradiated with protons, neutrons, electrons and 60 Co-  … significant progress on identifying defects responsible for sensor degradation over last 5 years! Defect Characterization Example: TSC measurement on defects produced by electron irradiation (1.5 to 15 MeV) -10- [R.Radu, 22 nd RD50 Workshop, 3-5 June 2013] G. Casse - Sensing in Extreme Environments - London, 7 May 2015

11. Summary on defects with strong impact on device performance after irradiation Some identified defects – Trapping: Indications that E205a and H152K are important (further work needed) – Converging on consistent set of defects observed after p, , n,  and e irradiation. – Defect introduction rates are depending on particle type and particle energy and (for some) on material! G. Casse - Sensing in Extreme Environments - London, 7 May 2015 -11- positive charge (higher introduction after proton irradiation than after neutron irradiation) positive charge (higher introduction after proton than after neutron irradiation, oxygen dependent) leakage current & neg. charge current after  irrad, V 2 O (?) Reverse annealing (negative charge) Boron: shallow dopant (negative charge) Phosphorus: shallow dopant (positive charge) Leakage current E4/E5: V 3 (?) A table with levels and cross sections is given in the next slide (spare slide).

12. G. Casse - Sensing in Extreme Environments - London, 7 May 2015 12 Radiation tolerance prediction: “old” method “Good” operation of sensors was based on the ability to provide a bias voltage corresponding to 120-130% of the full depletion voltage. But the VFD would be well over 10000V at HL-LHC doses.... The Charge Correction Method (based on TCT) for determination of effective trapping times requires fully (over) depleted detector.

13. More relevant method: analogue readout with LHC speed electronics Mip signal from 90 Sr source Analogue information from the Alibava board (equipped with Beetle chip) 13 G. Casse - Sensing in Extreme Environments - London, 7 May 2015

14. Irradiated silicon sensors G. Casse - Sensing in Extreme Environments - London, 7 May 2015 14 Low density of charge carrier in “non depleted” bulk, electric field at any voltage in the “neutral bulk”, high level of trap concentration. Q tc  Q 0 exp(-t c /  tr ), 1/  tr = . Simulations, with appropriate description of the radiation damage, needed for modelling behaviours at high fluences.. Max,n (  =1e14)  2400µm Max,n (  =1e16)  24µm Max,p (  =1e14)  1600µm Max,p (  =1e16)  16µm

15. Simulation vs parameterisation 15 Parameterisation is very accurate to LHC strip fluences. After higher fluences concepts like full depletion do not apply anymore. Trapping and drift voltage penetration are the quantities that better describe the performance of heavily irradiated sensors. RD50 is working to wards making prediction tools for highly irradiated sensors. With both simulation and experimental methods. G. Casse - Sensing in Extreme Environments - London, 7 May 2015

16. RD50 Charge multiplication G. Casse - Sensing in Extreme Environments - London, 7 May 2015 16 Some times, things go better than planned, a good day out fishing......... Or more charge than you expect after irradiation......... G. Casse, et al., "Evidence of enhanced signal response at high bias voltages in planar silicon detectors irradiated up to 2.2×10 16 n eq cm -2 ", Nuclear Instruments and Methods in Physics Research A, Vol. 636 (2011), pp. 56–61, DOI 10.1016/j.nima.2010.04.085.

17. Results with proton irradiated 300  m n-in-p Micron sensors (up to 1x10 16 n eq cm -2 ) RED: irradiated with 24GeV/c protons Other: 26MeV protons 17 G. Casse - Sensing in Extreme Environments - London, 7 May 2015 Irradiated with reactor neutrons

18. 18 G. Casse, NSS-2012, Oct 29 - Nov 3, 2012, Anaheim, California Charge Collection vs Bias (CC(V)) for the 50, 100, 140, 300  m un-irradiated sensors. As expected the signal varies linearly with thickness Measurements before irradiation

19. 2x10 15 n eq cm -2 19 G. Casse, NSS-2012, Oct 29 - Nov 3, 2012, Anaheim, California Degradation of the CC(V) with fluence Notice that the CC(V) for the 140  m thick sensor exceeds the expected charge ionised by a mip in that thickness of silicon.

20. 20 G. Casse, NSS-2012, Oct 29 - Nov 3, 2012, Anaheim, California 5x10 15 n eq cm -2 Degradation of the CC(V) with fluence

21. 21 G. Casse, NSS-2012, Oct 29 - Nov 3, 2012, Anaheim, California 1x10 16 n eq cm -2 Degradation of the CC(V) with fluence 2x10 16 n eq cm -2

22. Role of thickness 22 Charge degradation vs fluence for silicon sensors with different thicknesses The onset of Charge Multiplication breaks a few rules, like the proportionality of the signal with thickness..... G. Casse - Sensing in Extreme Environments - London, 7 May 2015

23. 23 A great margin to gain with extremely high voltage Compiled by M. Moll..but good agreement among measurements from different groups and sensors from different manufacturers 900 V

24. 24 Array of electrode columns passing through substrate Electrode spacing << wafer thickness (e.g. 30  m:300  m) Benefits – V depletion (Electrode spacing) 2 – Collection time Electrode spacing – Reduced charge sharing More complicated fabrication - micromachining Different detector structure: 3D Detectors Planar3D 300 µmµm Proposed by S. Parker and C. Kenney of the University of Hawaii in 1995. G. Casse - Sensing in Extreme Environments - London, 7 May 2015

25. 25 3D sensors radiation hardness Super-LHC L=10 35 cm -2 s -1 charged particle dose 10 16 /cm2 ATLAS/CMS at 4cm in 3 years LHCb (lower L) at 8mm in 6 years ATLAS 3D has highest CCE and at relatively modest Voltages …but most difficult fabrication RD50 3D working group (Richard Bates) ATLAS 3D working group (Cinzia DaVia) G. Casse - Sensing in Extreme Environments - London, 7 May 2015

26. Poly trench P-type diffusion Trenched detector design n+ Oxide trench P+ implant under N electrode Centered, 5um wide Trench 5, 10 50um deep All 5um wide in center of N+ electrode Same as P-type diffusion but with trench through N+ 26 G. Casse, HSTD8, Taipei (TW) 4-7 Dec. 2011

27. More evidence that trenched detectors have enhanced CC(V) compared to reference All sensors have higher currents than reference 27 G. Casse, HSTD8, Taipei (TW) 4-7 Dec. 2011 CC(V) and IV after 5E15 n eq cm -2

28. 19th RD50 Workshop – CERN, 21-23 November 2011 Trenched detectors with higher charge collection than standard. 28 G. Casse, HSTD8, Taipei (TW) 4-7 Dec. 2011 CC(V) and IV after 1E16 n eq cm -2

29. 19th RD50 Workshop – CERN, 21-23 November 2011 Sill trenched sensors exhibit higher CC(V) than reference 29 G. Casse, HSTD8, Taipei (TW) 4-7 Dec. 2011 CC(V) and IV after 2E16 n eq cm -2

30. Current choices for hybridisation Microstrip detector assemblies Pixel detector assemblies G. Casse - Sensing in Extreme Environments - London, 7 May 2015 30

31. ATLAS Tracker Based on Barrel and Disc Supports Hybrid cards carrying read- out chips and multilayer interconnect circuit Sensor Microstrip sensors Current barrel-type sensors (768 strips of 80μm pitch per side) Future barrel-type sensors (1280x4 strips of ~75μm pitch per side)

32. Hybrid pixels 32 G. Casse - Sensing in Extreme Environments - London, 7 May 2015

33. Detector requirements for different collider types Super-B factories, Linear colliders, ALICE. These all will have similar requirements: Low mass (detector and services) High granularity Speed Radiation hard (to doses much lower than hadron colliders) Example:

34. G. Casse - Sensing in Extreme Environments - London, 7 May 2015 34

35. Monolithic Active Pixels MAPs, standard CMOS processing. Active pixel cell with an NMOS transistor. The N-well collects electrons from both ionization and photo- effect. 35 G. Casse - Sensing in Extreme Environments - London, 7 May 2015

36. 36 Renato Turchetta 2010 IEEE NSS&MIC (RAL), Ping Yang Pixel2014 (CCNU, CERN) QUADRUPLE WELL MAPS QUADRUPLE WELL MAPS pixel = isolation of electronics (water) from detector (oil), (potentially thicker active layer and operation in depletion), Attempt of: - increasing collection efficiency and collection speed  radiation hardness - making detector ”active” – processing of signals in situ  being able to cope with required timings Large area with stitching, but fetaures: radiation hardness, collection efficiency are not perfect technology of quadruple WELL MAPS (nwell, pwell, dnwell, dpwell) Good for 3T pixel design, no processing in pixel, uncomplicated pixel = good for yield of large devices,

37. A SOI detector consists of a thick full depleted high resitivity bulk and separated by a layer of SiO2 a low resistivity n-type material. NMOS and PMOs transistors are implemented in the low resitivity material using standard IC methods. Monolithic Active Pixels SOI: silicon on insulator 37 G. Casse - Sensing in Extreme Environments - London, 7 May 2015 SOI is being developed for > 15 y, with good results for x-ray applications. Does not exhibit yet sufficient radiation tolerance for tracking.

38. 38 Can be very thin (~25 µm of silicon in total) and still fully efficient! Problem: how to handle, interconnect and at the end built a low mass ladder with such a thin device? One of the main feature of MAPS G. Casse - Sensing in Extreme Environments - London, 7 May 2015

39. 39 Speeding up the CMOS: HV-CMOS 39 CMOS pixel flavors (five years ago) Standard MAPS INMAPS HVCMOSTWELL MAPS Peric - CPIX14, Bonn, Germany, 15-17 September 2014 G. Casse - Sensing in Extreme Environments - London, 7 May 2015

40. Possible HVCMOS for ATLAS Pixels 40 To be coupled to readout digital electronics Digital outputs of three pixels are multiplexed to one pixel readout cell HVCMOS pixel contains an amplifier and a comparator + TOT = sub pixel address Readout pixel Size: 50 µm x 250 µm Size: 33 µm x 125 µm Different logic 1 levels G. Casse - Sensing in Extreme Environments - London, 7 May 2015

41. CCPD detector (HV2FEI4) I. Peric, CPIX – Bonn, Sep. 2014 41 2 3 1 2 3 1 CCPD Pixels + Size: 33 µm x 125 µm G. Casse - Sensing in Extreme Environments - London, 7 May 2015

42. HV-CMOS is radiation hard. It could even be VERY radiation hard. Liverpool is designing two submission (with AMS.35 µm and one in AMS.18µm) where we have implemented new concepts for further hardening these types of devices. 42 G. Casse - Sensing in Extreme Environments - London, 7 May 2015

43. Ultimate Interconnection: Vertical Integration Ideal solution for reducing material and easing assembly in detector system is to integrate electronics and sensors into a single item … if affordable This has been a “dream” for many years More complex detectors, low mass Liberate us from bump/wire bonding Many different aspects of these new technologies such as SLID (solid liquid inter-diffusion), TSV (through silicon vias), ICV (inter-chip vias) as well as more highly integrated concepts. Commercial technologies becoming available for custom design: IBM, NEC, Elpida, OKI, Tohoku, DALSA, Tezzaron, Ziptronix, Chartered, TSMC, RPI, IMEC…….. But are they all, or even, are any technologies radiation hard? 43 G. Casse - Sensing in Extreme Environments - London, 7 May 2015

44. 44 G. Casse - Sensing in Extreme Environments - London, 7 May 2015 Demonstrator of 3D integration (G. Deptuch) Fusion D2W bonded 3D chip on sensor wafer VIPIC bonded to sensor can be tested through wire bonded and bump- bonded connections Vertically Integrated Photon Imaging Chip (VIPIC) detector: Si d=500  m, pitch 80×80  m 2, soft 8keV X-rays application: XPCS 44

45. G. Casse - Sensing in Extreme Environments - London, 7 May 2015 45 Summary Status of sensors for future supercolliders: silicon (3d and planar hybrid pixels, microstrips) provides the required performances (with adequate cooling and power). But radiation hardness to 2x10 16 n eq cm -2 has been proven, and we are now directing effort to crack the next challenge : 1x10 17 n eq cm -2. This is the requirement for mass flow calorimeters in the forward region of hadron colliders. Future requirements point very strongly on low mass, fast, high granularity sensors. This has strong implication on systems: reading out large numbers of channels in short times, small signals, minimise cooling requirements..... The integration between sensor and electronics needs robust R&D: the small analogue signal cannot be driven over significant distances, they will require local processing. Electronics will have to provide this with limited power consumption!

46. G. Casse - Sensing in Extreme Environments - London, 7 May 2015 46 Summary The various integrated designs are advanced and are already baseline for several applications: DepFET (Belle), CMOS-MAPS (ALICE and others), HV-CMOS (Mu- 3e). HV-CMOS in particular can be used in stand-alone or can be hybridised to readout electronics. It can offer much cheaper hybridisation (gluing instead of bump- bonding), and segmentation within a single readout channel (e.g. Analogue encoding). In this sense it is the technology that can replace hybrid pixels and micro- strip sensors for the HL-LHC, if radiation tolerance expectations are confirmed. The 3D integrated sensor might be come available in future, but the progresses in this direction have been slower than expected.