Meeting UCSC

1
Stave Mechanical Issues
2
Basic Questions

• Preliminary Indications
• Radiation level increasing by factor of 10
• Potential silicon damage increases markedly
• Substantial increase in leakage current
• Suggesting potentially colder operation
• Radiation length issue
• Material in detector volume increased beyond
  original expectations
• Impetus for considering fundamental change to
  detector mounting, detector support, and
  electronic circuit design
• Long, individual stave structure with embedded
  cooling tubes
• Key questions, not all inclusive, could well be
• Is evaporative cooling still possible?
• If single-phase cooling were desirable, which
  coolant is rad-hard?
• Will detector thermal runaway require special
  cooling considerations?
• In satisfying the stringent requirements for
  detector cooling, thermal runaway, detector
  stability will a long stave structure be low
  mass?
• What are the best construction materials to
  withstand the radiation environment?

3
First Stave Geometry Studied
Electronic chip load 108W (Pixel stave 110W 80cm
length)
Module heating 1mW to 1W over life time
96cm
Recommend option being a 2m version of this
arrangement, with central support
Coolant in and out
4
Analysis

• Key initial considerations
• Stave stiffness (sag and thermal stability)
• Negative CTE materials
• Composite box (facings)
• Positive CTE materials
• Cooling tubes (Aluminum or PEEK)
• Silicon wafers
• Cable bus
• BeO hybrid
• Ceramic dielectric
• Stave Cooling
• Length and size of cooling tubes
• Thermal resistance of cable bus

Step A First order calculation of gravity
sag Step B Thermal model of cross-section
resulting from Step A
5
Gravity Sag

• Objective is to check proportions of composite
  structure
• Sag affected by orientationwe address worst case
• Sag affected by end support conditions and the
  unsupported length---we look at simple and fixed
  end conditions for 1m length
• (pins at each end provide something in between)
• Sag is affected by non-structural uniform mass
  distributed along length---an estimate is made
• Composite material-use a very modulus graphite
  fiber, quasi-isotropic layer
• For present ignore the negative CTE effect

simple support
Fixed support
6
Stave Sag
7
Sag Summary
If the separation between composite facings is
increased, the sag can be decreased within
acceptable bounds. This option will force the
consideration how best to package the cooling
tubes
8
Material Properties in Thermal Model
Properties used in 3D model
9
Thermal Model

• Model Makeup
• Array of 6 chips, two on back and one array on
  front, offset by the stagger in the silicon
  wafers
• Silicon wafers 3cm axial with 6.4cm width (280
  microns)
• Electronic chip, 7mm by 10mm (250microns)
• BeO, 2cm by 6.4cm (380 microns)
• Dielectric beneath chips, 229microns, k5W/mK
• Cable bus, 126microns, k0.12W/mK
• Aluminum cooling tubes, 12mil wall
• PEEK-fiber filled k0.9W/mK
• Objectives
• First order determination of gradient
• Effects of various material layers
• Compare PEEK versus Al coolant tubes
• Coolant film gradient

10
Thermal Solutions (No Wafer Heating)
Aluminum coolant tubes
Tube surface referenced to 0ºC
0.5W per chip
Side B
Side A
11
Thermal Solutions (No Wafer Heating)

• Solution---Aluminum Cooling Tubes
• Added convective cooling on interior of tube
  surface (3000W/m2K)
• Increase in peak chip temperature of 0.98ºC. An
  approximate calc of this component is 0.85ºC
• Notice small temperature variation for center
  wafer, however essentially uniform
• Side A electronic chip array is used to provide
  proper heating of the cooling tube between the
  two chip arrays on Side B

12
Thermal Solutions (No Wafer Heating)

• Solution---PEEK Cooling Tubes
• With convective cooling on interior of tube
  surface (3000W/m2K)
• Peak temperature 8.4ºC, above zero
• Increase in peak chip temperature of 2.2ºC.
• The small temperature variation in center wafer,
  among areas, suggest a more refined mesh should
  be considered in the final analysis

13
Comments

• Thermal Model
• Model is a slice from the stave
• Unfortunate but the heat load is unbalanced
  between Side A and Side B
• Gives the appearance that the silicon wafer
  temperature on Side A is not uniform
• The middle wafer on Side B has electronics on
  both sides of a wafer and here the surface
  temperature is nearly uniform
• Deviation of the chip temperature from wafer
  temperature is about 4ºC for both Al and PEEK
  cooling tubes.
• Heat flux is highest at the chip, but even here
  it is not very high (0.7W/cm2)
• BeO serves as heat spreader
• Flux between BeO and cable bus 0.23W/cm2 average
• Composite facing spreads heat along the cooling
  tubes
• Facing thickness is about the average thickness
  for the Pixel stave thermal management surface,
  but there are now two facings for distributing
  heat
• Thru thickness conductivity is lower, 1W/mK
  versus 10 to 20W/mK
• Heat flux at cooling tube wall reduced further
  over Pixel stave because of the two pass system
  within one stave

14
Wafer Heating (Next on List)

• Wafer Heating
• Two sources
• Leakage current peak estimated to be 1W per
  module from leakage current (temperature
  dependency)
• Free convection for surrounding gas
• Open for suggestions on gas temperature to be
  used
• Need to develop more confidence in results of
  thermal modeling before recommending the coolant
  inlet temperature
• Primary issues are portions chosen for stave
  structure and cooling tubes
• Secondary issues are the material thermal
  properties being used
• Must maintain reality check throughout analysis
• What follows
• Refinement in thermal model mesh
• Calculation of tendency for thermal runaway,
  although at the present no strong evidence this
  will be an issue within present geometry

15
Things to Watch For

• Stave cross section-2m length
• Need greater separation between facings for
  increased stiffness
• Poses interesting issue in integration of cooling
  tube- simple circular tube geometry not likely
• Wider Staves
• Again integration of coolant passages
• Possible wafer thermal runaway
• Temperature gradients
• Thermal distortion as a general issue
• CTE mismatch of materials
• Spatial temperature variations both lengthwise
  and transverse

16
Things to Watch For
One must be mindful that stave temperature varies
in Z-direction (internal pressure drop and
2-phase convection coefficient varies along stave
length)
17
Current Emphasis

• Structural model of 1m stave
• Compare to analytical solution
• Set up of thermal and structural coupling
  solution
• Predict CTE effects
• Axially
• Transverse
• Thermal gradient established by location of
  cooling tubes
• Tube geometry for increased separation between
  stave faces