Prequalification document

1
Prequalification document
Amsterdam, 3 April 2003

• Fred Hartjes, (Paul de Jong),
• Sandra Muijs, Joop Rövekamp

2
Overview

• This document contains the results of all QA
  tests performed during the assembly of the
  prequalification modules
• Description of the assembly and the QA procedure
  followed at NIKHEF can be found in separate
  documents
• In total 3 modules have been assembled and
  tested 1 mechanical dummy and 2 electrical
  dummies
• The first half of this document contains the
  reception test that are used for component
  selection
• The second half describes the test on the
  completed modules

3
Component reception tests

• Sensors
• IV curves
• (Quick check under microscope to look for obvious
  defects)
• Hybrids
• Visual inspection under microscope
• Metrology (z-plane)
• Electrical test confirmation sequence
• Spine
• Thickness verification

4
Sensor reception tests IV curves

• For both electrical modules sensors from batch
  5005 were used
• module K5-511 wafers 5 and 6
• module K5-512 wafers 1 and 13
• Wafer 5005-13 was selected even though it has an
  early breakdown,
• but this module is an electrical dummy so we did
  not want to use the
• best components we had in stock.
• Also Wafer 5005-1 has a breakdown, but only at
  very high voltages.

5
Wafers selected for module K5-511
Wafers 5005-5 and 5005-6. IV-curves measured at
NIKHEF, compared with those measured by the
manufacturer CiS (from the database info).
Results agree very well.
6
Wafers selected for module K5-512
Wafers 5005-1 and 5005-13 IV-curves measured at
NIKHEF, compared with those measured by the
manufacturer CiS (from the database info). Wafer
1 has a breakdown above 400V, which is obviously
not observed by the manufacturer Wafer 13 has a
clear early breakdown, but it remains inside the
manufacturers specs (lt 20 ?A at 350V) Wafers
are selected for dummy modules only.
7
Metrology results

• While assembling K5-511 we had a problem with the
  tooling while using for the first time a spine
  with a reinforced tongue. As a result the spine
  and far-end washer have been moved across about
  200 µm in X while the main washer is about 150 µm
  too high in Z. However the position of the
  relevant items (detectors, hybrid, fan-ins) has
  not been affected. The tooling has been modified
  to avoid this error in future.

8
Hybrid reception tests visual inspection K5-511

• K5-511 VDC surface looks scratched. Big pads in
  the middle are seriously damaged (probably not
  problematic). Some of the lines might be
  affected.
• weblink to picture of VDC
• DORIC has low bonds (weblink to picture of
  DORIC).
• ABCD S4 damaged edge (from dicing) on hybrid
  side, under bonds.
• ABCD S10 drop of spit (?) on input pad.
• One ABCD chip was not correctly glued. As a
  result, the bonds touch the edge
• of the chip (weblink to picture).
• None of these little defects caused problems in
  the electrical tests later.

9
Hybrid reception tests visual inspection K5-512

• VDC some damage (see picture of the VDC)
• DORIC surface looks scratched (see picture of
  the DORIC)
• ABCD S1 has a scratch
• None of the defects seems to cause problems in
  the electrical tests

10
Hybrid reception tests z metrology K5-511

• Hybrid is V-shaped around the cooling block with
  an amplitude of 150 200 µm, possibly due to the
  assymetry between the upper and the lower kapton
  flex. This appears to be a typical phemenon for
  most hybrids
• In contrast to the hybrids used for the NIKHEF
  miniproduction in 2002, the HV tongues do no
  longer curl up

11
Hybrid reception tests z metrology K5-512

• Hybrid is again V-shaped with an amplitude of 150
  200 µm

12
Module QA tests

• Metrology of the module after bonding
• Thermal cycling and again metrology
• Full electrical characterisation
• IV curve of the module after bonding

13
Module K5-511 z metrology

• Z metrology of the module before thermal cycling.
  

14
Module K5-511 xy metrology

• before thermal cycling

15
Module K5-511 z metrology after thermal cycling

• The module has been cycled 10 times from 30 to
  50 degrees and back, in 24 hrs
• The z metrology is much worse after thermal
  cycling.

16
Characterisation results k5-512
17
Characterisation result k5-511
18
IV curves module 511
IV curves at different stages during the
assembly In blue, the sum of the IV curves of
the individual wafers (reception tests) In
orange, the corresponding measurements made by
CiS In green, the sum of the front and back wafer
after they have been glued to the spine In red,
the IV curve of the module, after bonding The
breakdown behaviour of the unbonded glued wafers
is observed before in earlier modules made at
NIKHEF. It appears to be typical behaviour for
CiS wafers as long as they are unbonded, the
behaviour is rather unpredictable.
19
IV curves module K5-512
IV curves at different stages during the
assembly In blue, the sum of the IV curves of
the individual wafers (reception tests) In
orange, the corresponding measurements made by
CiS In green, the sum of the front and back wafer
after they have been glued to the spine In red,
the IV curve of the module, after bonding The
breakdown did not disappear after bonding, so
these wafers were indeed not good enough to build
modules (NB they were selected for dummy modules
only).
20
Study of bond heights in K5-511

• Bond heights all tuned not to exceed 0.5 mm above
  chip and fan-in surface
• Lower bonds all exceed 0.5 mm above detector edge