F' Caspers CERN:Chopper development evolution

1
Chopper Development EvolutionF. Caspers
CERN-AB-RF-FB

• Brief review of present status
• Why to consider an alternative printing
  technology for the metallic meander on alumina
• Discussion of first results and difficulties
  experienced with moly-mangan structures
• An experimental approach to define the cross-over
  frequency S-parameters of our standing wave/
  traveling wave approach
• Conclusion and outlook

2
Present status

• Below our standard meander structure is shown
  again (as a reminder length depicted here
  200mm, width of the meander 42 mm) in reality
  the length will be 400 mm bluealumina, red
  metal

3
Why to think about an alternative printing
technology (1)

• So far we used silk screen thick film printing on
  alumina
• How does it work The alumina substrate is coated
  with a 20 micron thick layer of an emulsion
  containing silver and glass particles. After
  drying this layer at 150 C the photo-etching
  process is applied (removing all the undesired
  coating from the surface) and the remaining
  structure is fired in air (oxygen from air
  required during the firing process)_at_ 900 C
• Afterwards increase of silver-layer thickness by
  chem. electro-deposition

4
Why to think about an alternative printing
technology (2)

• Normally the structures produced by thick-film
  technology on alumina are extremely solid.
• However in our case we have to add some more
  silver by chemical methods (electrolytic bath)
  since after firing the silver traces are just 10
  micron thick and have a much lower bulk
  conductivity (presence of glass as binder) than
  normal massive silver.
• We have noticed that the chemical bath can weaken
  the normally very good bond between the fired
  silver layer and the alumina surface.

5
Why to think about an alternative printing
technology (3)

• Thus the question
• What is the best known ceramic (alumina) metal
  bond
• Answer
• Moly-mangan ( molybdenum - mangan)
• How does it work
• A paint made from fine molybdenum and mangan
  particles in an organic suspension is brought
  onto the ceramic surface, dried and fired at 1400
  C under hydrogen atmosphere
• Chemical etching is done afterwards

6
Results and difficulties seen with Moly-mangan (1)

• The meander structure produced with MM
  (moly-mangan) technology also needs an additional
  silver layer of about 30 micron, to be deposited
  in a chemical bath.
• Otherwise we would get intolerable conductor
  losses, since the MM layer is just 10 micron and
  its conductivity 10 times below copper.
• However, silver cannot be deposited directly onto
  MM by chemical methods. One uses normally an
  intermediate nickel layer of a few micron.
• But for the SPL chopper we are allergic to
  nickel, since the structure is in a magnet (quad)

7
Results and difficulties seen with Moly-mangan (2)

• Thus we have deposited as intermediate layer
  silver (few micron) by evaporation after the
  structure has been etched and subsequently went
  into the chemical bath for increase of the layer
  thickness by another 30 micron
• ?It worked well with nice results
• DC resistance over 40 cm slightly above1 Ohm
• Minor degradations (resistive loss) in RF
  parameters
• RF match not yet perfect (45 Ohm instead of 50
  Ohm) as in previous thick-film versions
• DC current test done with 10 Ampere (100 Watt
  loss!) result ok, but this again showed the need
  for water-cooling

8
Results and difficulties seen with Moly-mangan (3)

• In a subsequent test with slightly modified width
  of the meander line (0.35 mm instead of 0.45mm in
  order to get closer to 50 Ohm) we have added the
  silver layer by evaporation before the etching
  process of the MM and then did the addition of
  silver in the chemical bath.
• ? poor result very bad adhesion between silver
  and MM
• According to the experts the MM surface was
  perhaps not properly cleaned when doing the
  silver evaporation. But this process should work,
  if done properly. Thus after several discussions
  it was decided to deposit the silver by
  sputtering and doing an in situ gas discharge
  cleaning if the MM surface beforehand.
• We should know more in a few weeks from now

9
How to measure on the bench the cross-ver
frequency (1)
Global view of the chopper structure with
surrounding tank and quadrupole Note the
triaxial feed-throughs upstream and down stream
of each deflecting plate
The N-connector is the inner part of the triax
system
10
How to measure on the bench the cross-over
frequency (2)detailed view
Coaxial feed, center conductor,1st triax conductor
Aluminum support, 10mm thick, electrically
floating wrt the vacuum tank 2nd triax conductor
beam
Collimators 13.5 mm high
Meander structure on alumina (3 mm high)
Wall of vacuum tank (3rd triaxial conductor)
11
How to measure on the bench the cross-over
frequency (3)measurement concept

• The concept is to carry out a transmission
  measurement between both meander structures,
  using one meander structure in the traveling wave
  mode as a beam simulator
• The other electrode is operated either also in
  traveling wave mode (forward coupler, properly
  terminated, inner coax system of the triaxial
  structure) or in standing wave mode (like a
  capacitive PU (outer coax system of the triax
  structure)
• ResultsAre shown this afternoon in the lab.

12
Conclusion

• We have found very solid and promising technology
  for printing a meander structure on alumina
• High DC current load tolerated
• Very solid metal-ceramic contact (heat transfer)
• Good RF properties
• Excellent radiation hardness
• A simple in situ measurement technique is a good
  candidate for experimental determination of the
  RF transmission properties of the frequency
  diplexer formed by the meander in triax
  operation(floating ground plane)