1 / 20

CCD-based Vertex Detector LCFI status report Nicolo de Groot RAL Conceptual design and goals

CCD-based Vertex Detector. CCD-based Vertex Detector LCFI status report Nicolo de Groot RAL Conceptual design and goals Detector R&D program at LCFI Development of Column Parallel CCDs Readout chips for the CPCCD Thin ladder program for mechanical support of the sensors Summary.

ophira
Télécharger la présentation

CCD-based Vertex Detector LCFI status report Nicolo de Groot RAL Conceptual design and goals

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. CCD-based Vertex Detector CCD-based Vertex Detector LCFI status report Nicolo de Groot RAL • Conceptual design and goals • Detector R&D program at LCFI • Development of Column Parallel CCDs • Readout chips for the CPCCD • Thin ladder program for mechanical support of the sensors • Summary

  2. Conceptual Design and Goals 5 layers at radii 15, 26, 37, 48 and 60 mm; Low power, gas cooled; High precision, low mass support mechanics; Encased in light foam cryostat; Minimum external connections. Thin detector (< 0.1% X0) for low error from multiple scattering; Close to the interaction point for reduced extrapolation error; Readout time: 8 ms for NLC/JLC (read between trains) 50 μs for TESLA inner layer (read 20 times during the train); Pixel size  20 μm20 μm, stand-alone tracking, radiation hard, etc.

  3. CCD Development Large area, high speed CCDs • Inner layer CCDs: 10013 mm2, 2500(V)650(H) pixels per CCD end; • Outer layers: 2 CCDs with size 12522 mm2 , 6250(V)1100(H) pixels; • 120 CCDs, 799106 pixels (20 μm square) in total; • For NLC/JLC: readout time  8 ms in principle sufficient, but not easy to achieve with standard CCDs, Column Parallel CCD is desirable; • For TESLA: • 50 μs readout time for inner layer CCDs : 50 Mpix/s from each CCD column • Outer layers: 250 μs readout, 25 MHz from each column • Column Parallel CCD is essential • Satisfy TESLA requirements, but thinking about NLC/JLC as well • CPCCD for JLC/NLC could be very advantageous

  4. CCD Ladder End • Electronics only at the ends of the ladders; • Bump-bonded assembly between thinned CPCCD and readout chip; • Readout chip does all the data processing: • Amplifier and ADC per column • Correlated double sampling built-in • Hit cluster finding with threshold • Data sparsification, time stamping • Memory and I/O interface • CPCCD is driven with high frequency, low voltage clocks; • Low inductance layout for clock delivery.

  5. CCD Development CPCCDs for TESLA: • Quality of 50 MHz clocks over the entire device (area = 13 cm2): • Power dissipation: • Large capacitive load (normally  2-3 nF/cm2), needs low clock amplitudes; • Low average power (  10 W) for the whole detector, but large peak power (TESLA duty cycle = 0.5%). • Feedthrough effects: • 2-phase drive with sine clocks – natural choice because of symmetry and low harmonics • Ground currents and capacitive feedthrough largely cancel CPCCDs for NLC/JLC: • Low readout frequency (780 kHz) – in principle few electrons noise could be achieved;

  6. Our First CPCCD Manufactured by E2V Direct outputs and 2-stage source followers (SF) Two phase, 750(H)400(V) pixels, 20 μm square; Wire/bump bond connections to readout chip and external electronics; Two charge transport regions. 1-stage source followers and direct outputs on 20 μm pitch

  7. First Signals First signals from one channel observed on 20 May 2003; Hits from 55Fe source clearly visible on the oscilloscope; Clocked with 1 MHz sine wave, full parallel readout; Particularly nice : small clock feedthrough. Great joy & happiness in the lab

  8. CPC-1 Tests Clock drivers 3 pre-amplifiers Low inductance clock delivery, short tracks; Clock symmetry; Pulsed bias voltages to CPCCD; Shutdown mode – everything in a gas-cooled cryostat; On-chip temperature monitoring diode .

  9. Stand-alone Performance 55Fe spectrum at 1 MHz readout Noise  60 electrons; Gain spread between channels  6%; CCD responsivity:  3.1 μV/electron. Single pixel hits spectrum using information from 3 adjacent columns

  10. Stand-alone Performance Minimum clock amplitude  1.9 Vpp ! • Two main CPC-1 variants (by design) : • inter-gate barrier  1 V • inter-gate barrier  2 V • CPC-1 optimized for small signals; • Measured barrier : 1.9 V and 2.8 V; • Difficult to control small dose implants in E2V process; • Future plans: • Reduce clock voltage even further; • No implant, better barrier definition by geometry (dielectric thickness).

  11. Stand-alone Performance Benefit of field enhanced implant  100 mV; CPC-1 clocked to 10 MHz with the same low clock amplitude of 1.9 V; 10 MHz is still low frequency for charge transfer; At > 10 MHz clock feedthrough rapidly increases. Clocking several nF of capacitance at high frequencies.

  12. Clock Feedthrough On-chip clock delivery in CPC-1 not fully symmetrical by design – technological issues; Faster clocking will be attempted soon. P1 P2 P1 P2 buslines buslines

  13. Plans In the next 2 months: Wire bond CPC-1 to CPR-1 : 1 in 3 channels, direct outputs and 1-stage SF; Bump bond CPC-1 to CPR-1, all channels (VTT Finland); Test both configurations; Study stand-alone performance at higher frequencies. After gaining enough confidence with CPC-1: Start design of CPC-2 : Much bigger chip : 8-10 cm in length (detector-scale); Better on-chip clock propagation; Test low clock operation in large area “stitched” CCD.

  14. Readout Chip CPR-1 First bump-bondable readout chip (CPR-1) delivered Designed by the Microelectronics Group at RAL; Size : 6 mm 6.5 mm; Voltage amplifiers for the 1-stage SF outputs; Charge amplifiers for the direct outputs; 250 5-bit flash ADCs; Everything on 20 μm pitch; 0.25 μm CMOS process; Scalable and designed to work at 50 MHz. Wire/bump bond pads Voltage Amplifiers Charge Amplifiers 250 5-bit flash ADCs 250(W)132(L)5-bit FIFO Wire/bump bond pads

  15. First Signals

  16. Readout Chip CPC-2 CPR-1 is now under intensive tests Work on next generation chip (CPC-2) is near completion Features of CPC-2: Amplifier and ADC per column Correlated Double Sampling for low noise Kernel threshold 22 cluster finding (limited by 0.25 μm feature size) Data sparsification Time stamping Memory and I/O interface

  17. Thin Ladder R&D A program to design CCD support structures with the following properties: • Very low mass (< 0.4% X0– SLD VXD3) • Shape repeatability to few microns when temperature cycled down to  –100 C; • Compatible with bump bonding; • Overall assembly sufficiently robust for safe handling with appropriate jigs; Three options: • Unsupported CCDs – thinned to  50 μm and held under tension • Semi-supported CCDs – thinned to  20 μm and attached to thin (and not rigid) support, held under tension; • Fully-supported CCDs – thinned to  20 μm and bonded to 3D rigid substrate (e.g. Be)

  18. Semi-supported Option Beryllium substrate with adhesive pads Thinned CCD (  20 μm) CCD brought down Shims Adhesive Assembly after shim removal and curing 0.2mm Beryllium substrate (250 μm) 1 mm FEA simulations continuing: • Distortions of only few μm, optimise adhesive pitch and size; • Silicone adhesive: (e.g. NuSil), excellent at low temperature • Layer thickness  0.12% X0 XY stage for 2-dimensional profiling being assembled: • Laser displacement meter • Resolution 1 μm • Models made from steel + unprocessed Si were studied

  19. A New Idea - Nanotechnology SLD had glue pads, which implies compression of silicon under cooldown. How to do better in 21st C? Micromechanical structure Maybe replace beryllium by some foam material – whatever gives best stiffness for least radiation length, regardless of thermal expansion properties

  20. Summary • Detector R&D work at the LCFI collaboration: • Development of fast column parallel CCDs and their readout chips; • Precision mechanical support of thinned CCDs. • Most aspects of the R&D are applicable to all proposed LC machines; • High speed CPCCDs are mainly for TESLA, however NLC/JLC likely to benefit from slow CPCCDs. On track to achieve TESLA-compatible detector; • Speed requirements for NLC/JLC easily exceeded with CPCCD – focus on optimisation of noise, radiation hardness, etc. End result is likely to be an excellent detector; • Significant work still required, challenging combination of chip size and speed; • More results to follow in a couple of months. More information is available from the LCFI’s web page: http://hepwww.rl.ac.uk/lcfi

More Related