1. H.-G. Moser Max-Planck-Institut für Physik Semiconductor Laboratory 1 HLL Structure Devices Projects -MPE, CFEL -MPP: ATLAS upgrade SIPMs Belle II MPP Project Review 19.12.2011

2. H.-G. Moser Max-Planck-Institut für Physik 2 Structure and Organisation Joint Laboratory of MPI für Physik MPI für extraterrestrische Physik Directors: S. Bethke (MPP) K. Nandra (MPE) Head of laboratory:H.-G. Moser (MPP)L. Strüder (MPE) Personal MPP Project Review 19.12.2011 MPP Personal Senior Staff:L. Andricek, G. Liemann, H.-G. Moser, R. Richter, A. Wassatsch Postdocs:A. Macchiolo (ATLAS/SCT), J. Ninkovic (MINERVA) Technicians: Z. Albrechstkirchinger, G. Fuchs, A. Plis, M. Schnecke, C. Weber Secretaries: E. Fleischmann, P. Schmalhofer IT:J. Jakowitsch PhD-Students:C. Jendrysik, C. Koffmane, A. Ritter, M.Tesar Master Students: F. Müller, S. Petrovics MPP;21MPE: 30PNS45

3. H.-G. Moser Max-Planck-Institut für Physik Devices Strip and Pixel Detectors (ATLAS SCT and Pixel) Silicon Drift Detectors pn-CCD (fully depleted) (XMM-Newton, eRosita) Silicon Photomultipliers Thin Pixel detectors + 3D Integration (sLHC, rad hard) DEPFET Active Pixel Detectors 3

4. H.-G. Moser Max-Planck-Institut für Physik HLL Projects MPP Project Review 19.12.2011 4 Pixel sensorsPn-CCDsDEPFET APSSiMPl SIPM MPE (x-ray astronomy) CFEL/ASG (x-ray lasers) MPE (HEP) MPE (astro- particle) eRosita BepiColomboIXO/ATHENA CAMP DSSC Pixel upgrade Belle II PXD (ILC) ILC calorimeter CLIC tracking Cerenkov telescopes

5. H.-G. Moser Max-Planck-Institut für Physik 5 eROSITA XMM ROSAT Science Goals  First imaging all-sky survey up to 10 keV Investigation of dark energy and dark matter Launch: 2013 Next: BepiColombo (Mission to Mercury): 2014 ATHENA (with DEPFET sensor) -> 2020 eRosita (2013) MPP Project Review 19.12.2011

6. H.-G. Moser Max-Planck-Institut für Physik 6 Hard X-ray SASE Free Electron Lasers LINAC COHERENT LIGHT SOURCE LCLS SCSS SPring-8 Compact SASE Source European XFEL Facility 2010 MPP Project Review 19.12.2011

7. H.-G. Moser Max-Planck-Institut für Physik CFEL Projects MPP Project Review 19.12.2011 7 PN CDD for CAMP CFEL AGS Multi Purpose 2 CCDs with 1k x 1k pixels Operated at LCLS (SLAC) Readout time: 7ms (150 fps) LCLS frequency: 120Hz Mimi virus experiment: Janos Hajdu Nature 470, 78–81 (03 February 2011)

8. H.-G. Moser Max-Planck-Institut für Physik DEPFET for XFEL at DESY MPP Project Review 19.12.2011 8 Far too fast for CCDs => DEPFET sensor with single pixel readout Another interesting feature Signal compression in sensor => Extend dynamic range

9. H.-G. Moser Max-Planck-Institut für Physik MPP Projects: Need For Thin Silicon Thin pixel sensors for ATLAS -Good CCE after irradiation -Less material -Thickness: 75 – 150 µm Thin DEPFET active pixel sensors for Belle II (and ILC) -Low multiple scattering -Belle II 0.20% X 0 450  m 50  m Cut through the matrix n+n+ p+p+ n - non-depleted region n - depleted gap region n SiMPl SiPMs with bulk integrated quench resistors -Simple processing -Ideal for 3D integration -Need tuning of pitch, bulk resistivity and thickness => Thin silicon ~ 50 µm 9 MPP Project Review 19.12.2011

10. H.-G. Moser Max-Planck-Institut für Physik 10 Sensor Thinning ? Process backside e.g. structured implant Wafer bonding SOI process Thinning of top wafer (CMP) Processing etching of handle wafer (structured) diodes and large mechanical samples Belle II module Need thin (50µm-75µm) self supporting all silicon module MPP Project Review 19.12.2011

11. H.-G. Moser Max-Planck-Institut für Physik 11 ATLAS pixel upgrade: sensor production for FE-I4  6 SCM FE-I4 with IBL guard-rings  2 SCM FE-I4 with full guard-rings  7 DCM FE-I4 with IBL guard-rings  Production of IBL sensors on SOI p-bulk material with 5 wafers with an active thickness of 150  m  Some problems with phosphorous contaminations => repair needed and done  Outsourcing (VTT?) <- Bump bonded by IZM Source scan -> FE-I3 FE-I4 Pixel50x400µm² 50x200µm² Matrix18x160 80x336 Size7.4x11mm² 20x19mm² MPP Project Review 19.12.2011

12. H.-G. Moser Max-Planck-Institut für Physik 12 New pixel detector concept New pixel detector concept based on 3D integration  Thin sensors and ASICs  Backside connectivity for compact module design (eventually 4 side buttable)  High density interconnection. MPP Project Review 19.12.2011 Build demonstrator using ATLAS pixel chip (FE-I2/3) and thin pixel sensors made by MPI (complete wafers with FEI2, FEI3 chips available!) Interconnection with SLID and ICV technology by Fraunhofer EMFT Demonstration of post-processing of standard ASICs with via last test sensors (strip, pixel) on 75mm and 150µm SOI wafer

13. H.-G. Moser Max-Planck-Institut für Physik 13  SLID modules glued and wire-bonded to a modified version of the ATLAS pixel detector board (Bonn University)  Measurements performed with the ATLAS read-out system USBPix SLID Module MPP Project Review 19.12.2011

14. H.-G. Moser Max-Planck-Institut für Physik 14 d= 75  m Collected Charge with 90 Sr: compatible with CCE of a 75µm thick sensor Signal map: all channels connected and functional Module 10: Results Noise distribution: comparable to standard ATLAS module Charge collection after irradiation (2x10 15 n/cm²) MPP Project Review 19.12.2011

15. H.-G. Moser Max-Planck-Institut für Physik 15 R&D program Sensor/ASIC interconnection using SLID -ASIC thinned to 200 µm -No vias, integrated fanout on sensor for service connection Sensor/ASIC interconnection using SLID -ASIC thinned to 50 µm -vias for service connections (fanouts for redundancy) Future: SLID interconnection of sensors/ 3D FEI4 Sensor/ASIC interconnection using SLID -ASIC thinned to 200 µm -No vias, integrated fan-out on sensor for service connection Sensor/ASIC interconnection using SLID -ASIC thinned to ~50 µm -vias for service connections (fan-outs for redundancy) Future: SLID interconnection of sensors/ 3D FEI4 Sensor/ASIC interconnection using SLID -ASIC thinned to 200 µm -No vias, integrated fan-out on sensor for service connection Sensor/ASIC interconnection using SLID -ASIC thinned to ~50 µm -vias for service connections (fan-outs for redundancy) Future: SLID interconnection of sensors/ 3D FEI4 Supported by AIDA FP7 EU project (WP3) MPP Project Review 19.12.2011

16. H.-G. Moser Max-Planck-Institut für Physik 16 SiMPl: silicon photomultipliers with integrated bulk resistors - Components of a SIPM cellSIMPL(E) approach No need of polysilicon Unobstructed entrance window (nor resistors, Al-contacts) for high fill factor Simple technology, easy and cheap production Need to match wafer thickness, resistivity, and pixel size  Need for SOI or epi material  Only small variation of the pixel size on one wafer MPP Project Review 19.12.2011

17. H.-G. Moser Max-Planck-Institut für Physik 17 IEEE NSS/MIC 2009 Orlando, Florida 17 Prototype production >130 different chips 6mm 30x30 arrays 10x10 arrays MPP Project Review 19.12.2011

18. H.-G. Moser Max-Planck-Institut für Physik 18 Prototype production: results High homogeneity over big distances! 6 100 cells arrays placed over 6mm distance High homogeneity within the array! High linearity! MPP Project Review 19.12.2011

19. H.-G. Moser Max-Planck-Institut für Physik Fill factor & Cross Talk & Photon Detection Efficiency Pitch / GapFill factorCross talk meas. (  V=2V) PDE calc. (  V=2V) PDE calc. (  V=5V) 130  m / 10  m85.2%29%39%61% 130  m / 11  m83.8%27%38%60% 130  m / 12  m82.4%25%37%59% 130  m / 20  m71.6%15%32%52% 19 Produced SiMPl devices have the world record in the fill factors!  V=2V  V=1V Hamamatsu MPPC SiMPL  V=2V  V=1V SiMPl devices have extremely high fill factors and still lower cross talk! No special cross talk suppression technology applied just intrinsic property of SiMPl devices MPP Project Review 19.12.2011

20. H.-G. Moser Max-Planck-Institut für Physik @223K Particle Detection 20 Excellent time stamping due to the fast avalanche process (<1ns) MIP gives about 80pairs/  m  huge signal in SiPM  allows operation at small  V Reduction of dark rate and cross talk by order of magnitude 10% GE >98% MIP detection Dark rate: 1 MHz/mm² = 1 hit/µm²/s = O(Belle II) With 20 µm pitch and 12 ns time stamp: occupancy: 2.5 x10 -6 Power (analogue): ~ 5 µW/cm² (Dominated by dark rate) MPP Project Review 19.12.2011

21. H.-G. Moser Max-Planck-Institut für Physik Next generation SiMPl devices 21 n+n+ n - non-depleted region n - depleted gap region n TDC, Photon counter, active recharge Cell electronics Bond readout ASICs on top surface Blind for light Sensitive to MIPs -> tracker ‘easy’ flip chip bonding (depends on pitch) Bond to back surface (after thinning) Light sensitive Single photon imager 3D devices without TSV ! Bonding on thin devices n+n+ n - non-depleted region n - depleted gap region n TDC, Photon counter, active recharge Cell electronics MPP Project Review 19.12.2011

22. H.-G. Moser Max-Planck-Institut für Physik Pixels: 50 x 50(75) µm 75µm thick0.20% X 0 2 layers: @1.4(2.2) cm DPEFET PXD for Belle II 22 Each pixel is a p-channel FET on a depleted bulk Electrons accumulate in the internal gate and modulate the transistor current (g q ~ 400 pA/e - ) Fully depleted: => large signal, fast signal collection Low capacitance, internal amplification: => low noise High S/N even for thin sensors (75µm) Rolling shutter mode (column parallel) => 20 µs frame readout time => Low power (only few lines powered MPP Project Review 19.12.2011

23. H.-G. Moser Max-Planck-Institut für Physik Sensor Production at MPI semiconductor lab finished in 2011 -SOI material: 50 µm sensor wafer (production started before 75µm was decided) -6 wafers (+ 2 monitor wafers on standard material) -Small test sensors -Full size sensors for prototyping -Technology & design variations PXD6 production 23 Bad yield in 1 st sub batch; Al etch step: ~6% of large matrices Remaining wafers were processed differently Batch Nr.Overall yield of half-ladders 1 (3SOI +1 ref)1/16 =6.25% 2 (3SOI +1 ref)4/16=25% (afterAl1 100% - 16/16) 3 (2SOI)4/8=50% (after Al1 100% - 8/8) We still want to do better Special R&D project in order to improve yield further. Recipe will be transferred to prototype production MPP Project Review 19.12.2011

24. H.-G. Moser Max-Planck-Institut für Physik First Thin DEPFETs 24 450  m 50  m Protection cover for etchingWafer after etching Cut through a small test matrixSide view of a matrix cut through MPP Project Review 19.12.2011

25. H.-G. Moser Max-Planck-Institut für Physik Lab test of thin DEPFET 25 DCDB readout chip, Bump bonded on interface board PXD6 DEPFET matrix (64x32 ) Switcher control chip 450µm FZ 50µm SOI Signal scales with thickness (CURO readout) MPP Project Review 19.12.2011

26. H.-G. Moser Max-Planck-Institut für Physik Test Beam Test beam at CERN, October 2011 a)PXD6, 50µm, L=4µm, standard ox:S/N ~ 40 b)PXD6, 50µm, L=6µm, thin ox:S/N ~ 20 thin oxide not yet optimized => smaller S/N DCD V1 does not allow optimal I drain final sensors are 50% thicker => S/N > 40 expected for optimal settings Readout speed up to 320 MHz (100ns readout time) single pixel cluster all  =12.3µm single  =12.8µm Multiple  =8.6µm 26 MPP Project Review 19.12.2011

27. H.-G. Moser Max-Planck-Institut für Physik PXD9 Prototypes Slip of Belle II schedule gave us time for a intermediate prototype production: PXD9 Fully fledged protoypes! Final design and wafer layout! 4 outer and 2 inner modules 48 test matrices and devices + yield control structures Lean production: Two sub batches of 6 wafers each: Planed to be ready by Spring 2013  Test full functionality (incl. rad. hardness)  Yield determination  Material for trial assembly  System test  Backup modules for Belle II (outer layer) Production of final sensors scheduled 2013/2014 27 MPP Project Review 19.12.2011

28. H.-G. Moser Max-Planck-Institut für Physik Design Features thanks to Andrea s Liebel PNSen sor 3D simulation of potentials to optimize performance: charge collection, clear efficiency, rad. hardness Analyse PXD6 and avoid weak points (failure tolerant design) 3D model of layout Analysis of electrical network MPP Project Review 19.12.2011 28

29. H.-G. Moser Max-Planck-Institut für Physik 29 Conclusions MPE/CFEL: CCDs an DEPFETs for X-ray imaging very successful operation of CCDs at LCLS MPP Projects ALTAS upgrade thin Si-sensors work (better) as expected thin planar sensors are candidate for IBL and upgrade 3D interconnection progressing SiPMs SiMPL devices function! Project Seminar by Jelena Ninkovic: January 24 Belle II first thin DEPFET sensors tested Prototype production in 2012 Series production in 2013/2014 Change of organisation ahead (2013) MPP Project Review 19.12.2011

30. H.-G. Moser Max-Planck-Institut für Physik Change of Organisation Motivation : »Single administrative entity. »Re - organisation of commercial activities. »Be open to more customers from the MPG (CFEL..). Consequences: »HLL will become a MPG central unit (partly like digital library). One central administration (by MPE?). Personal will be (temporarily) transferred (n.b.: no lay-offs). »Control and financing by ‘stakeholders’ (MPP, MPE, ….) & GV. »Separation of PNSensor activities: full cost compensation for use of clean room for SDD production return will be shared by stakeholders (and GV) Change will (probably) happen by 2013 MPP Project Review 19.12.2011 30

31. H.-G. Moser Max-Planck-Institut für Physik Backup MPP Project Review 19.12.2011 31

32. H.-G. Moser Max-Planck-Institut für Physik Matrix Pretest 32 Unfortunately we made a mistake thinning the first 4 wafers. Some Al lines contacting the transistor drains were etched away in a developing step. 0 transistors connected 16 transistors connected 32 transistors connected 2 wafers repaired with improved process ~ 100% ok MPP Project Review 19.12.2011 Repaired by sputtering and structuring additional Al layer Some shorts introduced 2 wafers processed later: unaffected and etched without damage

33. H.-G. Moser Max-Planck-Institut für Physik MPP Project Review 19.12.2011 Implementation of a copper line for the bump bond process of the ASICs we need a solderable landing pad on the sensor substrate  install Cu (and Sn if needed) process at HLL (almost ready) the Cu layer is effectively a 3 rd metal on the sensor – ρ(Cu) / ρ(Al) ≈ 0.63  low impedance wiring layer and additional routing freedom 33 EMCM (electrical multi chip module)  Test 3 rd layer metal (Cu)  Exercise interconnection  Test module concept  Noise tests Design almost finished Fabrication next year