Page tree

under construction

"Highest spatial and time resolution we can get, without noise, and as little material budget as possible"

Necessary Specifications

  • At least binary hit information
  • Material Budget → minimize multiple scattering
    • at 5 GeV, even up to 350 um is fine, if resolution is ~3 um
    • towards lower energies (higher rates) e.g. a hybrid w/ 100 um sensor, 50 um ASIC is fint
    • 50um is not a hard requirement!
  • Detection efficiency (for GeV electrons) >99%
  • Hit resolution: Linear Collider Vertex sensor requirement < 3 um single point resolution
    • → track resolution at DUT of same order
    • →  < 5.5 um single point of t'scope sensor (hard limit)
    • → Achieved by either small pitch binary or larger pitch + ToT charge weighting
  • Time resolution: <5 us ("Alpide" limit)

  • Read-Out:
    • no rolling shutter
    • frame based (global shutter)
  • Active Area
    • 2 x 2 cm
  • Availability after production: ~100 sensors
  • Radiation hardness: TODO (could be considered a NTH for longevity @ CERN)

Nice-to-have / Ideas

  • Time resolution:
    • For (future) DESY test beam: consider 2 GeV/c test beam particles (highest rate):
      • averaged: 4 kHz/cm2
      • due to time structure (magnet cycle) "high time" rate: (65ms-10ms) out of 80ms → 5.8 kHz/cm2
      • with multi-bunch: n = 10 → 58 kHz/cm2 → ~17 us integration time
      • → only little pile-up from 'fast' part of the source distribution
    • For future single bunch extraction: for up to 10^10 particles within 2ns, time resolution sub 2ns would be helpful
    • go for hit-driven = data-driven = event-based read-out ?
  • Sustainable hit rate off chip, assuming 2x2 cm: 240 kHz average (for n = 10)
  • Track resolution at DUT < 1.5 um for very aggressive designs (e.g. SOI 8x8 um2)
  • Select between "full speed" (and "full occupancy" (full frame)) mode ???
  • Larger area by 2 sides buttable for making 4-chip assembly per plane (2x2 cm2 → 4x4 cm2)

Homework to consolidate specifications:

  • Is 100 um thickness acceptable for chip with better single point resolution?
    → yes
  • How good needed to be as good as DATURA? Depends ...
    → even 350 um thickness at 5 GeV is fine, when the spatial resolution is ~ 3um. A thin low res (5um) sensor, is mostly worse (up to 5% X0) than a thick (350 um) high res (3 um) sensor
  • No labels