Present: Heinz, Carlos, Bernadette, James, Helene

Agenda: https://indico.cern.ch/event/972988/

MALTA modules for the BGV: Sensor and module geometries for simulations (Helene)

• Sensor
  • MALTA chip (23 x 21.3mm) + sensitive area (18.3 x 18.3 mm)
    • [Heinz] Die height might be a bit too much
  • Loaded geometry of sensor from Carlos' AllPix^2
    • But silicon box is 18um thick? [Heinz] this is the sensitive area -> but now the most recent (MALTA2) sensors is 30um. Heinz recommends:
      • 10 um circuitry
      • 30 um sensitive volume
      • 30 to 50 um substrate
• Module
  • 8+8 -> but inner radius restriction
  • 6+6 -> better, but then gaps on the outside
  • [Heinz] Internal discussions on MALTA3
    • Instead of 4 MATLA-2's -> could have 3 x MALTA-3's (?)
  • 7+7 -> better, inner radius "loss" = 1.2mm
    • [Heinz] Divide into 4 + 3 for the readout
    • [Heinz] Could avoid interfence by stagging in z
  • [Heinz] Connect silicon sensors with bump bond "bridge", avoids wire-bonds which would require space (3-4 mm) between the sensors. Wire-bonds require 3-4mm of space; suggest to optimise active area close to the beam pipe
  • [Heinz] Consider wedges rather than moons for modules with > 2 rings:
    • Minimise the material in the active area (each sub-module => cables);
    • Allows to construct in more "bite size" objects;
    • Wedges are easier to build rings

• Module positioning
  • First station = n_0 rings; Second station = n+0 + 1 rings' Third station = n_0 + 2 rings
  • Beam pipe inner diameter = 45. 0mm
  • Stations at 59cm; 87cm; 117cm (positions optimised for 7 TeV tracks distribution)
    • [Heinz] Comfortable distances; 60cm from first to last layer -> very comfortable distance, in particular, for a common carbon fiber support -> very practical distances (e.g. for cooling, cabling) and not too long for alignment
    • [Heinz] Why not consider 4 planes?
      • [Helene] Not decided -> to be studied with simulation
      • [Heinz] Chances of defining the track are much better with 4 x points
      • [Heinz] Rebuilt CMS & ATLAS pixel detectors with 4-layer pixel detector and now going to 5-layers
      • [Heinz] Use overlap between planes for alignment -> should have 20%, which is
      • [Heinz] Recommends to use a "safe" number of detector planes -> suggests to ask Andy for advice
      • [Heinz] Estimates for such a small detector; 50% mechanics & readout & 50% for the detector itself
      • [Heinz] 4th layer is usually for the pattern recognition and not the tracking itself

• Misc. consideration
  • Do we need CO2 cooling?
    • [Heinz] We do need some 2-stage cooling
    • [Carlos] ATLAS is moving away from C6F14 for environmental reasons
    • [Heinz] Both CO2 & C6F14 are used for the ATLAS pixel detector. CO2 has the advantage of requiring smaller diameter pipes (1.7mm) c.f. C6F14.
    • [James] Reason to operate at -20?
      • [Heinz] Leakage current doubles every 7 degrees
      • [Heinz] Anti-annealing effects -> bias voltage goes up
      • [Heinz] Noise improves as
      • [Heinz] Need an air tight (environmental) enclosure to flush with nitrogen or dry air to prevent condensation; need cable feedthoughs -> but quite simple
      • [Heinz] Not a showstopper to use C6F14
      • [Heinz] Need to calculate cooling power

MALTA3 (Carlos/Heinz)

• MALTA3 design meeting started last week
• Reviewed limitations of current MALTA / MALTA2 design
• Discussing list of possible applications & calling instiutes for a kick-off meeting

• New features
  • 1ns time resolution -> requires 1 GHz clock -> but chip power prop. clock freq.^2
  • Output speed
  • Serial powering (daisy chaining)
  • Masking -> 1 bit per pixel (enable / disable pixel)
  • Pixel shape -> hexagonal pixels to "round out" the corners; [Heinz] Tried out on a small test structure -> avoid some of the effects associated with corners (charge loss and long drift time), but technically difficult to design with current design tools; Will only be done if absolutely necessary
  • I2C -> allows communication with a simple device (e.g. Rasperry Pi)

• Active area = 30mm x 20mm
• Periphery = 30mm x 5mm
• Above limited by TowerJazz 128nm process
• [Heinz] Look of request for 10cm x 10cm modules
• [Heinz] Two years of development for all features (Covid delay included)
• [Heinz] In the mean time possibility start building prototype with MALTA1 (2cmx2cm) and move to MALTA3 once available
• [Heinz] We'll probably need 2 years anyway to understand module building (look at mechanics, cooling and bridge building)
• [Heinz] Kick-off meeting to listen to what people need; will invite BGV team to participate <-> suggests a 10 - 15 mins. introduction presentation

AOB